Promoting Green Urban Development in Africa: Enhancing the relationship between urbanization, environmental assets and ecosystem services PART I: A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY This page intentionally blank. Promoting Green Urban Development in Africa: Enhancing the relationship between urbanization, environmental assets and ecosystem services PART I: A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Authors Jane Turpie, Gwyneth Letley, Robynne Chyrstal, Stefan Corbella and Derek Stretch Prepared for AECOM on behalf of The World Bank by Anchor Environmental with support from The Nature Conservancy Prepared by Anchor Environmental Consultants 8 Steenberg House, Silverwood Close, Tokai 7945 www.anchorenvironmental.co.za 2017 COPYRIGHT © 2017 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. April 2017 Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to the Publishing and Knowledge Division, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Page iv A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY PREFACE AND ACKNOWLEDGMENTS This study forms one of the case studies of a larger study on Green Urban Development commissioned by the World Bank and co-funded by The Nature Conservancy. Anchor Environmental Consultants (Anchor) was subcontracted by AECOM to undertake case studies in three cities: Kampala, Uganda; Dar es Salaam, Tanzania; and Durban, South Africa. Each city was consulted as to the focus of the case study. In the case of Durban, the city requested a study to evaluate Durban’s natural capital and its role in Green Urban Development (GUD). The study is made up of two parts. The first part provides an updated, spatial estimate of the value of natural capital in the eThekwini Municipal Area and the second part is a scenario analysis that evaluates the potential returns to investing in GUD with a focus on the role of natural systems. This study builds on the preparation of an Environmental Profile for eThekwini Municipality by AECOM. The ecosystem valuation study was led by Dr Jane Turpie and Gwyn Letley of Anchor Environmental Consultants. Dr Robynne Chrystal and Prof Derek Stretch of CCS consulting undertook the hydrological modelling work, and Dr Stef Corbella of CCS consulting prepared the engineering cost estimates. Grant Benn assisted with the GIS aspects of the study. We are grateful to the eThekwini municipal staff for their interest and support of this project, in particular to eThekwini Municipality Real Estate Department for supplying the property transaction data and to the eThekwini Environmental Planning and Climate Protection Department for providing relevant GIS data and associated explanations for the Durban Metropolitan Open Space System. Thanks to Roland White and Chyi-Yun Huang of the World Bank and Diane Dale, Brian Goldberg and John Bachmann of AECOM for inputs and discussions during the project, and to Dr Timm Kroeger of TNC, Jeff Wielgus (independent), Mike Toman of World Bank, and Sanjay Strivastava of UNEP for inputs during the study and comments on an earlier draft. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page v This page intentionally blank. Page vi A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY EXECUTIVE SUMMARY Introduction The aim of this study was to provide estimates of the It is increasingly appreciated that ecosystems can make value of ecosystem services provided by natural open a significant contribution to human welfare through space areas within the eThekwini Municipal Area (EMA), the provision of raw materials and food, functions that and to map the geographic variation in these values as far influence air and water quality, climate, hydrology as possible so as to be able to compare both areas and and the abundance of useful and harmful organisms, types of value, using available and locally-relevant data. and the provision of opportunities for recreation and The Millennium Ecosystem Assessment classified entertainment, spiritual fulfilment, cultural, educational ecosystem services into provisioning, regulating, cultural and scientific activities. While these ecosystem services and supporting services. The values of these services are mostly associated with natural systems outside of are generally described using the Total Economic Value cities, including the services provided to cities such as framework, which classifies values into direct use, water quality amelioration, they can also be provided indirect, option and non-use values. Economic value by the natural systems occurring within urban areas. In of any asset or service can be defined in terms of the fact, urban ecosystems provide important amenities amount that people are willing to pay for it. Economic and contribute to the livelihoods and wellbeing of large welfare is a measure of the total benefit that society numbers of people, and ultimately to the resilience of derives from economic transactions, and is the sum cities. However, especially in developing countries, of the net benefit to consumers (consumers’ surplus) natural urban open spaces are rapidly becoming and the net benefit to producers (producers’ surplus). degraded or lost as a result of high rates of urbanisation Since some ecosystem services are not directly traded and a lack of city finances to retain and manage them. in markets, environmental economics has developed a Durban is located within a global biodiversity hotspot, suite of methods to value the non-market benefits of and still contains a wealth of biodiversity. Some of this is ecosystems and environmental quality. protected in nature reserves, but much of it is in private This study was carried out as a desktop study based hands or in communal lands on the city’s periphery. City on available data. Modelling assumptions were based managers are divided over the level of attention that on data from within the study area, drawing on the should be given to preserving these remaining natural regional, national or international literature only where areas. While it is argued that they make a significant necessary. The study focused on the direct values contribution to biodiversity conservation in the province, associated with the provision of natural resources, provide valuable ecosystem services and will contribute indirect use values associated with regulating services to the city’s resilience in the face of climate change, generated by ecosystem functioning, and the amenity the counter argument is that much of this area should values generated by ecosystem attributes. Estimation make way for development to alleviate the escalating of option values and non-use values associated with problem of unemployment. In the meantime, managers ecosystems require survey-based methods and were of these natural areas have to make do with a very small beyond the scope of this study. percentage of the city’s budget, reflecting the current priorities of the city. The eThekwini Municipal Area The study of ecosystem services and their value to society has made significant advances since an estimate The city of Durban is located in the eThekwini Municipal was made of Durban’s ecosystem services in the 1990s Area (EMA) in the province of KwaZulu-Natal on the east using early values from the international literature. This coast of South Africa. The EMA covers an area of just less has been partly due to the advent of vastly improved than 2300 km2 extending from the uTongati River in the spatial data which allows a better understanding of north to the aMahlongwa River in the south. The EMA is geographic variation in the supply of and demand for topographically diverse, characterised by winding river ecosystem services. However, relatively little work has valleys and steep hills rising from the narrow coastal been done on the role and value of ecosystems within plain. The area has a subtropical climate with humid urban areas, and most of this work has been limited to wet summers and mild dry winters, and a mean annual single types of services or value. precipitation of over 1000 millimetres. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 1 The EMA has a population of 3.44 million, the third Provision of natural resources largest of the six metropolitan municipalities in South The provisioning value of natural systems was estimated Africa. This represents just over one third of the as the value of the sustainable yield of natural resources, provincial population of KwaZulu-Natal (KwaZulu-Natal) up to the yield levels demanded. Terrestrial, freshwater and has been projected to rise to four million by 2020. and coastal ecosystems provide a number of living While the EMA covers only 2.5% of the area of KwaZulu- and non-living resources which are harvested for Natal it accounts for 66% of provincial and 11% of raw materials, food and medicine. In the eThekwini national GDP. municipality these resources are predominantly harvested by poorer households on a subsistence The EMA encompasses the urban area of Durban (25% basis or to generate some cash income. The greatest of the area) as well as significant peri-urban (30%) and harvesting pressure comes from the rural communities rural land use areas (45%). Only 55% of households that occupy the approximately 1500 km2 of rural areas are in formal dwellings, with 34% in informal dwellings in the hinterland of the EMA, as well as the people living (concentrated in the peri-urban areas) and 11% in in poor peri-urban communities, although markets for traditional or rural dwellings. The latter fall mainly in these resources will supply a much broader segment of communal land areas administered by the Ingonyama the community, with some of the demand coming from Trust. both poor and relatively wealthy households within the The EMA is situated in the centre of one of 34 Global urban area. Biodiversity Hotspots, the Maputaland-Pondoland- The types of resources that are harvested in the EMA Albany region, and contains an impressive array of include water for domestic use, reeds and thatching biological diversity. The natural assets of the EMA form grass, firewood, poles, food plants, bush meat and fish. part of the Durban Metropolitan Open Space System, There are almost no data on the production or use of or D’MOSS, which also includes agriculture, man-made these resources within the study area. Provisioning parks and sports areas. The D’MOSS covers 75 000 ha, values were estimated based on information on habitat or almost a third of the total municipal area, above the productive capacity, actual harvests from comparable high tide mark, and is under various types of ownership. areas, habitat condition, land ownership, accessibility The terrestrial landscape in the EMA is essentially a and proximity to main sources of demand. One of the mosaic of open grassland, woodland, thicket, thornveld commonly-used resources, sand, could not be supplied and forest. Much of this is degraded and very little sustainably. Provisioning services were estimated to be falls within protected areas. The EMA also has a vast worth in the region of R100 million per annum, with a network of rivers and wetlands that drain into 16 net present value of approximately R1.12-1.46 billion. estuaries along its coastline. The largest of these has Most of the value is made up by water and fuelwood. been developed as Durban harbour, one of South Africa’s most important ports. While most of the KwaZulu- ES Table 1 Estimated extent and cost of the proposed GUD Natal coast is a high energy environment, Durban interventions central beaches are relatively sheltered and popular for recreation. Natural Resource Value % of (R millions) total River water for domestic use 31.8 32.5% Fuelwood 46.5 47.5% Timber poles 6.6 6.7% Medicinal plants 4.7 4.8% Grass and reeds 1.4 1.4% Wild meat harvesting 0.6 0.6% Fishery resources 6.3 6.3% Total 97.9 Page 2 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Carbon sequestration degraded, such as the iSipingo and Little aManzimtoti, Climate change caused by increases in the atmospheric contribute very little. It was estimated that two thirds concentrations of greenhouse gases will carry an of the value that the EMA estuaries could provide to estimated cost of about 2 – 7% of Gross Domestic KwaZulu-Natal fisheries if they were in a high quality Product (GDP) in different parts of the world by 2050, state has been lost. with developing countries being expected to incur proportionally greater costs. Natural systems can make Agricultural support (pollination) a significant contribution to global climate regulation through the sequestration and storage of carbon. When Natural habitats support organisms that provide they are degraded or cleared, much of this carbon agricultural support services in the form of pollination is released into the atmosphere. These emissions and the control of agricultural pests. Crop pollination by contribute to global climate change, which is expected insects is an essential ecosystem service that increases to lead to changes in biodiversity and ecosystem both the yield and the quality of crops. Of the crops functioning, changes in water availability, more frequent grown within eThekwini Municipality, many are wind- and severe droughts and floods, increases in heat- pollinated, including sugar and maize. However, several related illness and mortality, and impacts on agriculture crops are directly dependent on insect pollination, and energy production. including subtropical fruit crops such as mangoes, papayas, avocados and litchis, and nut trees such as Based on a previous field-based study conducted in macadamia, cashews and almonds. These crops are the EMA, the total amount of carbon stored in all the likely to benefit from wild colonies of bees occurring major vegetation types of the EMA open space system in untransformed vegetation around the tree crops or is estimated to be 6.6±0.2 million tonnes of carbon gardens, saving on active pollination costs incurred by equivalent, or 24.3±0.9 million tons of carbon dioxide hiring of bee hives or dusting. equivalent, and some 8 400 – 9 800 tonnes of carbon are sequestered per annum. The global damage costs The fruit tree crops and market gardens are located that this amount of carbon could produce are over in small patches throughout the EMA, however the R9.8 billion, while the damage costs to South Africa largest areas requiring pollination services are located resulting from a loss of the carbon stocks within the predominantly in the outer-west and the northern EMA would be approximately R34.3 million per annum. planning regions. Assuming that these natural areas These avoided damage costs are the annual value of the supply the equivalent service provided by three beehives service. The net present value of this service is estimated per ha, the total cost of replacing the pollination service to be about R393-511 million. to 1144 ha of fruit and garden crops would be just over R1 million per year with a net present value of R11.5- 15 million. Assuming relationships between vegetation Fisheries support (nursery value) characteristics and bee density, the highest potential Estuaries provide nursery areas for numerous pollination value is associated with vegetation patches in fish species that are exploited by recreational and good condition that are surrounding market gardens and commercial fisheries in the inshore marine area. The tree crops. Coastal and scarp forest and open grassland estuaries in the EMA support several species of fish near agricultural areas could have a value of R65 - R90/ caught in marine areas that are dependent on estuaries ha/y, while thicket and woodland areas could be worth as nursery areas for at least their first year of life. Some R130/ha/y. of the larger estuary systems, such as the uMngeni Estuary, are also known to provide an important nursery Flow regulation habitat for penaeid prawns. The combination of weather-related (e.g. rainfall Based on the estimated value of the fisheries and intensity, extent and duration) and geophysical (e.g. the percentage contribution of estuary-dependent catchment size, geomorphology, soil and land use) fish, the estimated total contribution of estuaries to characteristics are the main factors that influence KwaZulu-Natal coastal and inshore marine fisheries flooding (Kareiva et al. 2011). Natural systems such as is R106.8 million per annum. Based on the sizes and wetlands and rivers or ecosystems with deep permeable relative health of KwaZulu-Natal’s estuaries, it was soils can regulate flows through the landscape by estimated that the overall contribution of estuaries slowing flows by means of storage and vegetative in the EMA is R11.4 million per annum, with a net resistance and facilitating infiltration into soils. In this present value of R131-170 million. This represents 11% way these systems ameliorate the potential impacts of of the contribution of nursery value in KwaZulu-Natal. flood events by smoothing streamflow peaks, reducing uMkhomazi, Durban Bay, oHlanga and uMngeni have the bank and streambed erosion (Vellidis et al. 2003), as well highest percentage contributions and associated nursery as reducing the risk of damage caused by flooding of values. Estuaries that are both small and severely downstream areas. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 3 For this study, a hydrological model was set up for the A hydrological model, set up using PC-SWMM software, entire catchment area of the eThekwini Municipality was used to simulate the hydrology and sediment using the PC-SWMM software. This model was set transport for the catchment area of the whole up to run design flood events in order to determine municipality. By comparing the modelled sediment the influence of natural vegetation areas on flood outputs per catchment under current land cover hydrographs at strategic points relating to the location versus fully transformed land use, it was possible to of existing flood conveyance infrastructure. The flood estimate the difference made by natural vegetation to hydrographs generated under current conditions were the sediment loads transported to estuaries and the compared with what they would be if the natural three main water supply reservoirs. The value of the systems were transformed to urban land use. This service was then estimated for dams in terms of avoided provides an indication of the impacts of loss of natural replacement cost, based on modelled cumulative areas on flooding and the difference can be construed as storage loss and for Durban Harbour in terms of avoided an estimation of the flood attenuation benefit obtained dredging cost. from (retaining) the natural systems. The additional flood volumes without these systems that would The loss of vegetation cover from dam catchment occur under different return period flood events (e.g. areas leads to a significant increase in the rate of 1:10 years), would require larger drains, culverts, etc., sedimentation, with the greatest impacts being felt by depending on the size of the event these constructed Hazelmere Dam. It was found that the replacement of flood management assets are designed to deal with. these natural areas with human settlements could result Thus a second model was developed in order to estimate in a nine-fold increase in sediments entering the dam. the capital costs of the structures required under the The total annual replacement cost associated with the present versus the without-vegetation scenarios. The loss in dam capacity as a result of sedimentation was difference, together with associated differences in estimated to be between R1.1 million and R2.9 million. discounted annual maintenance costs, is the total life- This equates to a net present value of between R12.25 cycle flood management cost avoided by the natural million and R33.03 million, with an average of R22.77 systems, which can be converted to the net present million. The annual TSS load into Durban Harbour value of the service. increased by an estimated 195% when vegetation was removed and by 206% when the vegetation was The avoided capital cost requirements for flood replaced with dense settlement, resulting in dredging conveyance were estimated to be R338 million. costs avoided of between R1.03 million and R1.15 million This represents a 0.7% - 3.5% capital cost saving in per year which translates to an average net present stormwater infrastructure. Including an estimated 6% of value of R12.5 million. Therefore the sediment retention capital costs as an annual repair and maintenance cost service was valued at R3.1 million per year with a net (eThekwini Municipality 2015), this suggests that the present value of R35.2-45.8 million. flood attenuation service provided by natural systems in the EMA leads to a life cycle cost saving in the order of R339 million in net present value terms (equivalent Water quality amelioration to R29.5 million per annum). The highest per hectare Anthropogenic introduction of nutrients into the values associated with D’MOSS are located in the landscape can lead to reduced water quality and the catchments that are situated above the built up areas of eutrophication of freshwater and marine ecosystems. Durban city centre and Durban North. This reduces the capacity of these systems to supply ecosystem services and increases water treatment costs. Natural vegetated systems can play an important Sediment retention role in the trapping of sediments and absorption and Erosion and sedimentation within watersheds can breakdown of organic and inorganic pollutants in become a major issue as it causes structural damage surface and sub-surface water runoff. Wetlands are to reservoirs, causes flooding, affects the quality particularly well known for their capacity for water of drinking water and increases water treatment quality amelioration, but the service is also provided by and maintenance costs at water treatment works. terrestrial landscapes. Phosphorus is removed through Vegetation can reduce erosivity by stabilising soils sediment trapping and plant uptake, nitrogen is removed and intercepting rainfall, thereby preventing erosion. through denitrification and plant uptake, and pathogens Vegetated areas, especially wetlands, also capture the are destroyed by UV radiation. sediments that have been eroded from agricultural and degraded lands and transported in surface flows, In this study, the impacts of natural open space areas preventing them from entering rivers. While some level on water quality were estimated using a hydrological of sedimentation of dams is expected and planned for model which was set up to estimate the production and under natural conditions, elevated catchment erosion transport of total suspended solids (TSS), phosphorus (P) either incurs dredging costs or shortens the lifespan of and total inorganic nitrogen (TIN) in the catchments of dams and related infrastructure. Page 4 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY the EMA. The model was set up to estimate the change very important influences on property prices, with well- in sediment and nutrient loads entering dams and managed natural open space areas attracting significant estuaries if the retention capacity of natural vegetation and positive price premiums and those in a degraded was eliminated (a hypothetical construct), or if natural condition resulting in lower property values. Well- vegetation was replaced with dense human settlement. managed green open space accounted for about 2% of These two scenarios provided the upper and lower overall property value, or R4.4 billion, while public parks bounds of the service provided by natural open space attracted premiums amounting to 6.5% of property areas in physical terms. The value of the water quality value, or R13.8 billion. These values were mapped to amelioration service was then estimated in terms of the the relevant green open space areas, and suggest that avoided costs to water treatment works and the avoided well-managed open space areas and parks within the loss of estuary value as a result of the eutrophication of urban edge have an average asset value of R1.4 million these systems. and R20.5 million per ha, respectively, and jointly account for about R356 million per annum in property Without the existing natural vegetation in their tax revenues to the city. The value of both types of catchment areas, phosphorous loads entering the open space areas was particularly high in wealthier Inanda, Hazelmere and Nungwane Dams would increase neighbourhoods, possibly because of safety issues as by at least 193-319%, 193-968%, and 200 - 509%, well as ability to pay. respectively. Using these estimates and the water treatment cost models developed for Wiggins WTW and Durban Heights WTW it was estimated that maintaining Tourism value the natural vegetation results in an annual saving of The year-round warm weather, extensive beaches and between R1 million and R8.68 million per annum, numerous outdoor activities make the city of Durban depending on the alternative land use. a leading tourism destination in South Africa. To It was estimated that maintaining the existing areas determine the nature-based tourism value associated of natural vegetation also avoids potential losses in with open space areas in the EMA we used a large estuary fishery and nursery values of between R2.19 and dataset of over 10 000 geo-tagged photographs R3.71 million per annum. Thus, overall, water quality uploaded to Google Earth’s Panoramio site, based on the amelioration services were estimated to be worth about premise that the numbers of photographs uploaded to R7.8 million per annum, with a net present value of specific sites are correlated with the recreational value R89.1-115.8 million. This is a conservative estimate as associated with these areas. The highest concentration it does not take all affected downstream services into of photographic uploads were associated with the account. The highest per hectare values associated with Durban beachfront, Durban city centre, sports grounds D’MOSS are located upstream of Inanda and Hazelmere and golf courses, protected areas, and shopping malls. Dams, in the Durban Harbour catchment, as well as Few photographs were taken in rural areas surrounding in the south of the EMA upstream of the Umkomaas, the main urban core of the EMA. Statistical analysis of Msimbazi and Umgababa estuaries. these patterns in relation to the features of the different grid cells showed that the numbers of photographs were significantly influenced by the presence of certain Amenity value to property owners land cover types, and that the interaction of man-made and natural features was important in determining the The amenity of green open space areas is reflected attractiveness of the latter. to a large extent in two markets that are observable – property and tourism. Urban residents often pay a The contribution of different attractions to tourism value premium to live close to the areas that they enjoy using, were estimated by analyzing the content (or setting) of or to have a good view. Similarly, visitors pay to travel the photographs with each photograph being placed to and stay in an area where they will have access to or into one of five categories; (1) built environment, (2) views of these amenities. natural open space, (3) man-made open space, (4) rural or agricultural, and (5) marine or coastal. The total value A hedonic valuation approach was used to estimate added by leisure tourism of R5.6 billion (based on a the value associated with different types of green previous study) was then assigned to the five categories open space within the EMA. Data on a total of 16 149 by relating the percentage breakdown of each category property sales over a two year period from January to the actual proportional land cover and the number 2012 to October 2014 were analysed in relation to the of photographs within each grid cell in the EMA. Note amount and condition of each type of green open space that this is an underestimate of tourism value since it within three distances of each property, as well as sea excludes consumers’ surplus. views and a range of other property, local population and neighbourhood characteristics. The results revealed The total tourism value assigned to natural habitats that both the type and condition of open space have (natural vegetation, freshwater systems, estuaries and A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 5 the coastal environment) was almost R2 billion and The provisioning value natural systems was estimated accounted for 40% of all photographic uploads with the to be R1.1-1.5 billion (NPV), with fuelwood and river coastal environment contributing just over half of this water contributing the most to this value. Considering value. Apart from beaches, natural areas with high values water is collected from rivers and streams by only 0.5% included Kranzkloof, Beachwood and Shongweni nature of the population in the EMA, this resource is thought to reserves, Umhlanga Lagoon, Durban Bay and uMngeni be the most valuable in terms of the service it provides Estuary. Man-made open space such as parks and golf per user household. The remaining values represent courses had a total leisure tourism value of R382 million approximately R225-292 million of the total provisioning at an average value of R234 000 per hectare. Therefore value of natural resources in the study area (ES Table 2). the combined tourism value of natural and man-made Regulating services have a net present value of about open space is approximately R2.4 billion per annum, with R1.0-1.2 billion with carbon storage and flow regulation a net present value of R27.5 billion. accounting for 39% and 34% of this value, respectively. This value is believed to be a conservative estimate. The many benefits derived from these supporting services Summary and conclusions are indirect and particularly difficult to value, and most Natural and semi-natural systems within the eThekwini of the regulating services also play the role of supporting Municipal Area give rise to flows of ecosystem services services that influence the outputs of other ecosystem worth at least R4.2 billion per year. The total asset value services. The amenity value associated with natural of these areas was therefore estimated to be at least and semi-natural open space areas in the EMA was R48 -62 billion (ES Table 2). estimated to be R45.7-59.4 billion, 96% of the overall value of ecosystem services (ES Table 2). ES Table 2 Total value of ecosystem services in the EMA. Values in R millions (2015) Ecosystem services Annual Value NPV (20 y, 6%) NPV (20 y, 3%) (R millions) (R millions) (R millions) Provisioning services River water for domestic use 31.8 364.7 474.1 Fuelwood 46.5 533.4 693.4 Timber poles 6.6 75.7 98.4 Food and medicinal plants 4.7 53.9 70.1 Grass and reeds 1.4 16.1 20.9 Bush meat 0.6 6.9 9.0 Fishery resources 6.3 72.3 94.0 Sub-total 97.9 1 122.9 1 459.9 Regulating services Carbon storage 34.3 393.4 511.4 Agricultural support 1 11.5 15.0 Fisheries support (nursery function) 11.4 130.8 170.0 Flow regulation 29.5 338.8 338.8 Sediment retention 3.1 35.2 45.8 Water quality amelioration 7.8 89.1 115.8 Sub-total 87.1 998.8 1 196.8 Cultural services Amenity value to property owners 1 586.8 18 200.0 23 660.0 Tourism value 2 400.0 27 527.8 35 786.1 Sub-total 3 986.8 45 727.8 59 446.1 Total 4 171.8 47 849.5 62 102.8 Page 6 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Within the city, and overall, amenity value is by far services is modest, due to the relatively small extent the most important value of natural and semi-natural of activities that benefit from this service. These are open space. This large value is derived from a small also generally outside of the urban edge. The nursery proportion of the D’MOSS. The patterns observed for service of estuaries, which supports the fisheries sector, property value (representing value to residents), and is comparatively large. The fisheries that are supported tourism value were similar, with certain areas in the include both recreational and small-scale commercial EMA contributing significantly to both. fisheries within and beyond the coastal areas of the EMA, though the majority of these beneficiaries are In contrast to amenity values, provisioning services are likely to reside within the EMA. more prominent in the peri-urban and rural areas of the EMA. The substantial area of rural landscapes around Three of the regulating services described are related the urban edge, much of which are communal lands to catchment hydrology. The capacity to perform these under the ownership of the traditional authority, deliver services tends to be greatest in the surrounding rural important provisioning services to the large numbers landscapes because of their location and size, but the of poor households residing in the relatively dense beneficiaries of these services are downstream, and the settlements in these areas. Within the urban edge, the value of hydrologic-related services is also dictated to a large natural areas tend to be under private ownership large extent by the location of infrastructure. Thus the or state protection. The value of these services is spatial variation of these services was more irregular highest in the outer-west and southern extents of in relation to the urban edge than for provisioning the municipality where there are still large tracts of and amenity services. The values of these services natural vegetation, such as woodlands and forests. The were found to be highest upstream of Inanda and provisioning value in the northern area of the EMA Hazelmere Dams in the northern and outer-west areas and in the urban core tends to be somewhat lower and of the EMA, where sediment retention and water largely restricted to the river systems and wetlands in quality amelioration were most valuable, and also in these areas. the catchments of the downtown and harbour areas, where flood attenuation was highly valuable. Within If urbanisation is properly managed, the spatial disjunct the lower urban areas, however, hydrologic-related of provisioning and amenity values is likely to track services tend to have been overwhelmed to the extent the future growth of the city and lead to the increased that they can no longer ameliorate the increased run- value of the remaining open space areas. If not properly off and pollution from urban areas to a significant managed, a dead zone could be created at the city’s extent. This is a general reality of urban systems which periphery in terms of ecosystem services, and valuable is increasingly being addressed through innovative opportunities will be lost. As Durban grows and the peri- engineering solutions. Indeed regulating services are urban areas become densely populated and urbanised, one area where technological innovation does show it is likely that the provisioning services from those promise of being able to augment or replace ecosystem landscapes will make way for amenity services to future services, and where increased efficiency may be also urban inhabitants, as the demands of urban inhabitants desirable to protect the supply of ecosystem services replace those of the former rural inhabitants. Both from aquatic ecosystems that, unlike terrestrial systems, within the planned urban edge and beyond, informal are more certain to remain in place, in one form or and rural settlements will continue to grow and densify, another. The optimal balance between relying on the and the city will continue to be under pressure to services provided by natural capital versus implementing provide housing for the poor for some time to come. It engineering solutions to neutralise the effects of will be important to consider the implications of how urbanisation on run-off and water quality is explored in this growth is allowed to happen. The findings of this more detail in a companion report (Turpie et al. 2016). study suggest that remnant green open space areas become increasingly valuable with urbanisation and increasing incomes. Thus ensuring the protection of key open space areas in the areas being settled would secure potentially valuable sources of amenity and spiritual and physical wellbeing for these future communities. The role and spatial variation of regulating services is particularly interesting. Carbon storage is the most important of these services from a local economy perspective, and is also important from an international perspective. This value and the global value need to be taken into consideration in South Africa’s and eThekwini Municipality’s commitments to combating climate change. The value of agricultural support A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 7 Although this was a desktop assessment which still has substantial room for improvement and refinement through further research, it represents the most comprehensive assessment of ecosystem services within an urban environment in Africa to date, and one of perhaps only a handful in the world. Overall, the study shows that the ecosystem services profile of an urban area is very different from that of non-urban landscapes or regional level assessments. While natural and semi- natural ecosystems within the study area provide a full suite of ecosystem services, the value is dominated by amenity services to residents and visitors, both of which make a considerable and tangible contribution to Durban’s economy. Even so, the estimate of amenity value does not capture all of the recreational, spiritual and health benefits associated with of green open space, since estimating these values requires further data collection. The findings of this study can be used to identify priority areas for supply of ecosystem services and will be useful for decision-making processes related to understanding the tradeoffs between spatial development and conservation. This is pertinent to both environmental and strategic development planning within the eThekwini Municipality in forming the development of a thriving, resilient city. While the study focuses on ecosystem services, it is important to note that it does not provide a full estimate of the value of biodiversity per se. Our study does not capture the existence value and some of the other intangible, cultural values attached to biodiversity, nor does it adequately capture the role of species richness and community structure on the capacity of ecosystems to supply ecosystem services. This is a shortcoming of most valuation studies, and it is for this reason that conservation and planning decisions cannot be made on the basis of economic values alone. Planning should also incorporate biodiversity targets that are determined on the basis of their intrinsic value and understood role in maintaining ecosystem integrity and resilience. Page 8 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY ACRONYMS AND ABBREVIATIONS CBD Central Business District IDP Integrated Development Plan CCP Cities for Climate Protection InVEST Integrated Valuation of Ecosystem Services and CPI Consumer Price Index Tradeoffs CPUE Catch-per-unit-effort IPCC Intergovernmental Panel on Climate Change CVM Contingent Valuation Method KZN KwaZulu-Natal DAFF Department of Agriculture, Forestry and LULC Land Use Land Cover Fisheries MEA Millennium Ecosystem Assessment DEARD Department of Agriculture, Environmental NPA National Ports Authority Affairs and Rural Development NTFP Non Timber Forest Products DLA Department of Land Affairs NTU Nephelometric Turbidity Units DWA Department of Water Affairs OLS Ordinary Least Squares D’MOSS Durban Metropolitan Open Space System SANBI South African National Biodiversity Institute EDGE Economic Development and Growth in SAT South African Tourism eThekwini SDF Spatial Development Plan EM eThekwini Municipality TEEB The Economics of Ecosystems and Biodiversity EMA eThekwini Municipal Area TEV Total Economic Value EESMP eThekwini Environmental Services TIN Total Inorganic Nitrogen Management Plan TLA Total Living Area EMSCP eThekwini Municipality Spatial Conservation TNC The Nature Conservancy Plan TOC Total Organic Carbon EMD Environmental Management Department TP Total Phosphorous EPCPD Environmental Planning and Climate Protection TSS Total Suspended Solids Department WMA Water Management Area GDP Gross Domestic Product WRC Water Resource Commission GIS Geographic Information System WTTC World Travel and Tourism Council HPM Hedonic Pricing Method WTW Water Treatment Works ICLEI International Council for Local Environmental WWTW Waste Water Treatment Works Initiatives A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 9 TABLE OF CONTENTS I. INTRODUCTION 15 1.1 Rationale............................................................................................................................................................................. 15 1.2 Aim of the study.................................................................................................................................................................. 16 1.3 Ecosystem services valuation framework......................................................................................................................... 16 1.3.1 Classification of ecosystem services...................................................................................................................... 16 1.3.2 Total Economic Value framework.............................................................................................................................17 1.3.3 Economic value ....................................................................................................................................................... 19 1.3.4 Present value ........................................................................................................................................................... 20 1.4 Overall approach and limitations...................................................................................................................................... 21 1.5 Structure of the report....................................................................................................................................................... 21 II. OVERVIEW OF THE ETHEKWINI MUNICIPAL AREA (EMA) 23 2.1 Geography and climate...................................................................................................................................................... 23 2.2 Land use and socio-economic context.............................................................................................................................. 25 2.3 Biodiversity and D’MOSS .................................................................................................................................................. 27 III. PROVISION OF NATURAL RESOURCES 31 3.1 Introduction......................................................................................................................................................................... 31 3.2 Approach............................................................................................................................................................................. 33 3.2.1 Sustainable yields.................................................................................................................................................... 33 3.2.2 Adjusting for habitat condition ............................................................................................................................... 33 3.2.3 Adjusting for expected demand .............................................................................................................................. 35 3.3 Water................................................................................................................................................................................... 36 3.4 Sand ................................................................................................................................................................................... 36 3.5 Woody resources................................................................................................................................................................ 38 3.6 Food and medicinal plants................................................................................................................................................ 42 3.7 Non-woody raw materials................................................................................................................................................... 45 3.8 Bush meat............................................................................................................................................................................47 3.9 Fishery resources............................................................................................................................................................... 49 Page 10 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY IV. REGULATING SERVICES 53 4.1 Introduction ........................................................................................................................................................................ 53 4.2 Carbon storage................................................................................................................................................................... 53 4.3 Fisheries support (nursery function)................................................................................................................................. 56 4.4 Agricultural Support (pollination)...................................................................................................................................... 59 4.5 Flow regulation................................................................................................................................................................... 61 4.6 Sediment retention............................................................................................................................................................ 67 4.7 Water quality amelioration................................................................................................................................................. 73 V. AMENITY VALUE 83 5.1 Introduction......................................................................................................................................................................... 83 5.2 Amenity value to property owners..................................................................................................................................... 83 5.3 Tourism value...................................................................................................................................................................... 86 VI. SUMMARY AND CONCLUSIONS 93 VII. REFERENCES 101 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 11 LIST OF FIGURES Figure 1.1 Relationships between urban natural assets, ecosystem services and their beneficiaries........................................................................15 Figure 1.2 The classification of ecosystem values that make up Total Economic Value (TEV) ....................................................................................18 Figure 1.3 Demand and supply curves for a good, showing the calculation of consumer and producer surplus.......................................................20 Figure 2.1 The eThekwini Municipal Area (EMA) is located on the east coast of South Africa and covers an area of 2297km2................................23 Figure 2.2 Topographical map of the eThekwini Municipal Area...................................................................................................................................24 Figure 2.3 Population density (people/km2) for the eThekwini Municipal Area...........................................................................................................26 Figure 2.4 Habitats and land cover in the EMA. Freshwater systems, grassland, thicket, woodland, forest, rocky outcrops, estuaries, and dams are all part of D’MOSS.......................................................................................................................28 Figure 2.5 The major and minor rivers found within the EMA, and location of the 16 estuaries................................................................................30 Figure 3.1 ...............................................................34 Habitat and land use map showing the condition of the natural habitat found within D’MOSS. Figure 3.2 Estimated amount of water extracted from rivers and streams per sub-place in the EMA (m3 per km per year)...................................37 Figure 3.3 Estimated annual sustainable fuelwood output (m3/ha) from different habitats in the EMA...................................................................40 Figure 3.4 Estimated annual sustainable timber pole output (m3/ha) from different habitats in the EMA............................................................... 41 Figure 3.5 Estimated annual sustainable wild food and medicinal plant output (kg/ha) from different habitats in the EMA.................................. 44 Figure 3.6 Estimated annual sustainable grass and reed output (kg/ha) in the EMA...................................................................................................46 Figure 3.7 Estimated annual sustainable hunting output (kg/ha) across the EMA.......................................................................................................48 Figure 4.1 Total carbon storage (tons/ha).......................................................................................................................................................................55 Figure 4.2 The location of market gardens and fruit tree crops and the associated pollination value (R/ha/y) . ......................................................60 Figure 4.3 The full eThekwini catchments showing flow paths......................................................................................................................................63 Figure 4.4 Flood attenuation value associated with natural systems (D’MOSS) (NPV, R/ha) in the EMA....................................................................66 Figure 4.5 Inanda, Hazelmere and Nungwane Dams are located within the EMA (Source: Umgeni Water)...............................................................68 Figure 4.6 The current, lower bound and upper bound cumulative loss of capacity (Mm3) over 50 years for (a.) Inanda Dam, (b.) Hazelmere Dam and (c.) Nungwane Dam ...................................................................................................................71 Figure 4.7 Sediment retention value associated with natural systems (D’MOSS) (NPV, R/ha) in the EMA.................................................................72 Figure 4.8 Schematic diagramme of the consequences of anthropogenic effect on water quality and their amelioration by natual systems: (Source: Author)........................................................................................................................................73 Figure 4.9 Summary of water quality amelioration services by natural systems (Source: Turpie 2015)..................................................................... 74 Figure 4.10 Schematic summary of the linkages from phosphorous loads in water supply dams to increased water treatment costs as a result of deteriorating water quality...............................................................................................76 Figure 4.11 Average phosphorous loads (kg) in the uMngeni River above Nagle Dam (top) and Inanda Dam (bottom) and corresponding water treatment costs (R/ML) at Durban Heights WTW and Wiggins WTW. . ...........................................77 Figure 4.12 Schematic summary of the linkages from water quality parameters to estuary ecosystem services for a case of deteriorating water quality.......................................................................................................................................................78 Figure 4.13 Relationship between water quality and fish scores for the 16 estuaries of the eThekwini municipality.................................................79 Figure 4.14 Relationship between modelled annual loads of TIN and TSS and the Water Quality Scores for estuaries of the eThekwini municipality........................................................................................................................................................79 Figure 4.15 Estimated avoided losses of estuary ecosystem services due to the water quality ...................................................................................................80 amelioration function of natural vegetation in the estuary catchments.. Figure 4.16 Water quality amelioration value associated with natural systems (D’MOSS) (NPV, R/ha) in the EMA....................................................81 Figure 5.1 Average value (R/ha) of (a) natural open space in a good condition and (b) parks within each main-place within the EMA. Note: the actual location and extent of open space areas within each EMA is not shown due to scale. .................................................85 Figure 5.2 ...............................88 Pattern of geo-tagged photo uploads in relation to major land cover types within the eThekwini Municipal Area.. Figure 5.3 Nature-based tourism value in the EMA shown for terrestrial natural open space as R/ha and for the coast as R/km..........................91 Figure 6.1 Summary of the provisioning, regulating and cultural service values (NPV) associated with ...........................................96 natural and semi-natural open space in the EMA, excluding estuary provisioning and regulating services.. Figure 6.2 Summary of the provisioning, regulating and cultural service values (NPV) associated with ...........................................97 natural and semi-natural open space in the EMA, excluding estuary provisioning and regulating services.. Figure 6.3 Summary of the provisioning, regulating and cultural service values (NPV) associated with ...........................................98 natural and semi-natural open space in the EMA, excluding estuary provisioning and regulating services.. Page 12 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY LIST OF TABLES Table 1.1 .........................................................................17 The main ecosystem goods and services generated by natural or semi-natural systems. Table 1.2 The linkages between the TEV framework and the supply of ecosystem goods, services and attributes ...............................................19 Table 3.2 Summary of available information in the supply and demand of natural resource use in the eThekwini municipality........................... 31 Table 3.3 Estuarine state ranking based on health and functionality used to assign condition to each estuary in the EMA (Source: Forbes & Demetriades 2008)...........................................................................................................................................................35 Table 3.1 Tree volumes calculated for forest and woodland/thicket habitats in the EMA. Volumes were calculated using data collected by Glenday (2007). . ...............................................................................................................................................................38 Table 3.4 ........................................................39 Estimated total value (R million/year) of fuelwood yields in different vegetation types in the EMA. Table 3.5 Estimated total value (R millions/year) of timber pole yields in different vegetation types in the EMA...................................................42 Table 3.6 Estimated total value (R million/year) of wild foods and medicinal plants in different vegetation types in the EMA..............................43 Table 3.7 Estimated total value (R million/year) of grasses and reeds in different vegetation types in the EMA.....................................................45 Table 3.8 Estimated total value (R million/year) of hunting in different vegetation types in the EMA......................................................................47 Table 3.9 Summary of the subsistence line and invertebrate fisheries along the KwaZulu-Natal coastline in terms of numbers, annual catch, annual effort and value in 2012 (Source: Everett 2014)................................................................................................................................50 Table 3.10 Summary of the subsistence line and invertebrate fisheries along the KwaZulu-Natal coastline in terms of numbers, annual catch, annual effort and value in 2012 (Source: Everett 2014)................................................................................................................................51 Table 4.1 Carbon densities (tC/ha) for land cover classes in the EMA (Source: Glenday 2007).................................................................................54 Table 4.2 The five major categories and subcategories of fish that utilize South African estuaries (Whitfield 1994). ............................................56 Table 4.3 Estimated participation, annual effort and catch in four KwaZulu-Natal near-shore fishing sectors based on data from (Dunlop & Mann 2012, 2013) and the associated KwaZulu-Natal fishery value and KwaZulu-Natal nursery value based on the amount and percentage of the total contributed by estuary-associated fish species.................................58 Table 4.4 Percentage contribution of estuaries in the EMA to KwaZulu-Natal nursery value and the estimated nursery value for each individual estuary based on size and fishery health scores.................................................................................................58 Table 4.5 Lower and upper bound flood attenuation values (R million) for each quaternary in the EMA. A zero value indicates that there is no stormwater infrastructure within that catchment. ...........................................................................................65 Table 4.6 Lower and upper bound flood attenuation values (R million) for each quaternary in the EMA. A zero value indicates that there is no stormwater infrastructure within that catchment. ...........................................................................................71 Table 4.7 Estimated maintenance dredging costs avoided due to the sediment retention function of natural vegetation in the Durban Harbour catchment...........................................................................................................................................................................71 Table 4.8 Annual water treatment cost savings associated with the three water supply dams in the EMA ............................................................80 Table 5.1 Property, neighbourhood and green open space characteristics included in the hedonic analysis..........................................................84 Table 5.2 .........................................................................86 Estimated expenditure and valued added by tourists visiting Durban in 2014 (R billion).. Table 5.3 Model estimation results................................................................................................................................................................................89 Table 5.4 The overall leisure tourism value (R million) and % photographs for the five land use categories............................................................90 Table 6.1 Total value of ecosystem services in the EMA. Values in R millions (2015).................................................................................................94 Table 6.2 Summary of the provisioning and regulating service values (NPV, 20 y, 6%) associated with estuaries in the EMA (R millions).............99 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 13 This page intentionally blank. Page 14 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY I. INTRODUCTION 1.1 Rationale It is increasingly appreciated that ecosystems can make Various types of environmental assets exist within cities, a significant contribution to human welfare through which yield a range of benefits, or ecosystem services, the provision of raw materials and food, functions that to different sectors of society (Figure 1.1). However, influence air and water quality, climate, hydrology there is generally a paucity of understanding of both and the abundance of useful and harmful organisms, the ecological functioning and the value of the existing and the provision of opportunities for recreation and natural capital in African cities in particular, or of the entertainment, spiritual fulfilment, cultural, educational trade-offs involved in developments that replace or and scientific activities (Barbier 1994, 2011, MEA 2003). degrade these assets. While these ecosystem services are mostly associated with natural systems outside of cities, including the African cities are faced with the rapid influx of rural poor services provided to cities such as water quality seeking better opportunities in addition to the intrinsic amelioration, they can also be provided by the natural growth of the cities’ populations. Because this rate of systems occurring within urban areas. In fact, urban growth often outpaces plans and the capacity of city ecosystems provide important amenities and contribute managers to provide the necessary services, it results to the livelihoods and wellbeing of large numbers in the development of informal settlements and results of people, and ultimately to the resilience of cities. in a burgeoning poor population and attendant social However, especially in developing countries, natural problems. This puts major financial pressures on cities urban open space areas are rapidly becoming degraded which are faced with provision of housing and services. and lost as a result of high rates of urbanisation and a One response is to plan urban economic development lack of city finances to retain and manage them. that will create employment opportunities. It is generally Figure 1.1 Relationships between urban natural assets, ecosystem services and their beneficiaries Source: Author A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 15 understood that economic development has to keep areas will also provide a better understanding of the pace with the growing demand for jobs in order for these potential opportunity costs of developing these areas. households to contribute financially to service provision They can help to evaluate conservation-development and avoid a downward spiral, and developers are trade-offs within urban areas and also to evaluate the therefore encouraged to bring such opportunities. Thus potential trade-off between allowing the loss of open the tandem growth of informal settlements and formal space within the urban edge, versus loss of natural areas developments steadily contributes to the conversion of through planned urban expansion or unplanned urban open space areas within and around cities. Added to this, sprawl (Elmqvist et al. 2013). Furthermore, it is often the lag in service provision and overwhelmed capacity extremely difficult and expensive to restore or rehabilitate for municipal functions also leads to the pollution and ecosystems once they have been degraded past a certain degradation of the natural areas that remain. point. Investing in the maintenance and protection of natural capital is therefore important and will translate These issues are as relevant to Durban as to any other into future cost savings and improved human well-being. African city. Unlike many other African cities, however, For these reasons there is a real need for cities to factor Durban is located within a global biodiversity hotspot, ecosystems and the services they provide into urban and still contains a wealth of biodiversity. Some of this is planning, management, budgets and policies (TEEB 2010). protected in nature reserves, but much of it is in private hands or in communal lands on the city’s periphery. City The study of ecosystem services and their value managers are divided over the level of attention that to society has made significant advances since the should be given to preserving these remaining natural 1990s. This has been partly due to the advent of areas. While it is argued that they make a significant vastly improved spatial data which allows a better contribution to biodiversity conservation in the province understanding of geographic variation in the supply (EPCPD 2012), provide valuable ecosystem services of and demand for ecosystem services. Nearly all of (Roberts et al. 1999) and will contribute to the city’s the research effort has taken place outside of urban resilience in the face of climate change (Cartwright et areas, where ecosystems are large, but increasingly al. 2013), the counter argument is that much of this fragmented and threatened by rural land uses and a host area should make way for much needed development of anthropogenic pressures. While some work has taken to relieve the escalating problem of unemployment. In place on the impacts of urban sprawl on ecosystems the meantime, managers of these natural areas have at regional scales (e.g. Eigenbrod et al. 2011), relatively to make do with a very small percentage of the city’s little work has been done on role and value of budget, reflecting the current priorities of the city. ecosystems within urban areas, and most of these have been limited to single types of services or value. It is commonly argued that these issues arise because people do not understand the value of ecosystems, or of green open space areas1 in cities (TEEB 2010). Based 1.2 Aim of the study on this premise, the value of Durban’s Metropolitan Open Space System (D’MOSS) was estimated in 2001 The aim of this study was to provide estimates of the (eThekwini Municipality 2002), resulting in an estimate value of ecosystem services provided by natural open of some R3.1 billion per annum. This study indeed raised space areas within the eThekwini Municipal Area (EMA), public sector awareness of the value of these areas, and and to map the geographic variation in these values as far resulted in improved allocation of resources for their as possible so as to be able to compare both areas and management (Mander, pers. comm.). However, the types of value, using available and locally-relevant data. effect of such a valuation can have a limited political lifespan (Roberts 2008), particularly if it is presented as a single large number. In reality, many decisions are 1.3 Ecosystem services valuation framework made at the margin. Therefore, part of the reason for the limited usefulness of this study was the poor spatial 1.3.1 Classification of ecosystem services resolution of its estimates, which were extrapolated from The concept of ecosystem goods and services stems international studies (mainly Costanza et al. 1997) on the from the perception of ecosystems as natural capital basis of total areas of different habitat types within the which contributes to economic production (Kareiva et city. More localised and locally-based estimates of value al. 2011). Goods are the resources that are harvested are required in order to understand the current value such as wood and fish, services contribute to economic of different open space areas and the resources that production or save costs such as waste assimilation, should be allocated to their protection and management. and attributes relate to the structure or organisation Understanding the relative value of different open space of biodiversity such as rarity, diversity and beauty. The Millennium Ecosystem Assessment (MEA 2003) defined ecosystem services as “the benefits people obtain from 1 comprising a mix of natural and semi-natural areas, including man- made outdoor recreational areas ecosystems” and categorised the services provided Page 16 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY by ecosystems into provisioning services, regulating 1.3.2 Total Economic Value framework services, cultural services and supporting services. The Total Economic Value (TEV) of ecosystem services Provisioning services relate to the living (e.g. fish and is comprised of different types of value which are grass) and non-living (e.g. water and sand) resources categorised into direct use, indirect, option and non-use that are harvested for building materials, food and values (Figure 1.2). medicines, regulating services include climate regulation, waste treatment and water purification, cultural Use values relate to the actual use of the good in services include the aesthetic, spiritual, educational question (e.g. a visit to a National Park), planned use and recreational benefits derived from ecosystems, and (e.g. a visit planned in the future) or possible use. supporting services include nutrient cycling, and services Actual and planned use are easily understood concepts, that maintain the conditions for life on Earth. The first but possible use is more ambiguous (but not any three align well with the definitions of goods, services less important) as it relates to what people would be and attributes described above. Although changes in willing to pay to maintain a good in existence in order supporting services influence the delivery of provisioning, to preserve the option of being able to use it in the regulating and cultural services, they are usually ignored future (Pearce et al. 2006). Non-use value refers to in valuation studies to avoid double counting. The main the willingness to pay to maintain a good in existence ecosystem services generated by natural or semi-natural even though there is no actual, planned or possible use. systems are described in Table 1.1. Theoretically, TEV is the sum of all use and non-use values, although depending on how they are measured they may not be additive. When adding values it is important to make sure that double-counting does not occur. Use and non-use value can be classified further and these values are explained below. Table 1.1 The main ecosystem goods and services generated by natural or semi-natural systems Ecological Economic Types of Services characteristics characteristics Fishery resources Sand Grazing Stocks of resources Goods Provisioning Fuelwood Woody raw materials (e.g. timber, poles) Non-woody raw materials (e.g. thatching grass) Food and medicinal plants Animals and birds (hunting) Carbon sequestration and storage Regulation of hydrological flows (infiltration, flood attenuation) Amelioration of water quality (detoxifying pollution, dilution Ecological functions & transport of pollution) Services Regulating and processes Erosion control and sediment trapping Habitat for organisms useful in pollinating and controlling pests of croplands Refugia/critical habitat for organisms used consumptively or non-consumptively beyond the study area Ecosystem Attributes Cultural activities and heritage characteristics (aesthetic, biodiversity, Recreational use and enjoyment Cultural and biodiversity rarity, physical features) Scientific and educational value composition Spiritual and religious activities and well-being A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 17 Figure 1.2 The classification of ecosystem values that make up Total Economic Value (TEV) Source: TEEB 2010 Consumptive use values are based on the provision of may also be important for local communities that have ecosystem goods such as crop production, fisheries, uncertain incomes and will rely on natural resource livestock grazing, wild plant harvesting and hunting. harvesting in future times of need. While option value Non-consumptive use values are based on ecosystem cannot be measured, it is possible to estimate a ‘quasi- attributes and include tourism, recreational activities, option value’, which is society’s willingness to pay research and study opportunities, and aesthetic, spiritual to retain the option for future use of the ecosystem and religious appreciation or use of ecosystems. (Perman et al. 2011). Indirect use values are derived from the regulation Non-use values from ecosystems are the values that services provided by species and ecosystems. Ecosystem do not include any direct or indirect uses of ecosystem functions may either generate outputs that form inputs services. They reflect the satisfaction that individuals in production processes elsewhere (i.e. the benefits derive from knowing that ecosystem services are are realised off-site), or they result in engineering cost conserved and that other people have or will have savings by performing functions that would otherwise access to them (TEEB 2010, Kolstad 2011). These may require costly infrastructure or man-made processes. arise from personal values or intergenerational equity These services include air quality regulation, water concerns. Estimating non-use values is far more difficult purification, erosion control and pollination. Their value than estimating use values, mostly because non-use is usually understood to be positively related to the level values relate to moral, aesthetic, charitable or religious of ecosystem health or integrity, and are seen as a public properties for which there are usually no markets (TEEB service which are generally not reflected in market 2010, Kolstad 2011). Nevertheless, non-use values transactions. are reflected to some extent in society’s willingness to pay to protect resources and ecosystems, and with Option value relates to the importance that people appropriate market mechanisms can be captured give to the future availability of ecosystem services for through transfers and converted to income. personal benefit (TEEB 2010). These include possible medicinal, leisure, agricultural and industrial uses. Table 1.2 provides an overview of the linkages between Option value is particularly important when there is the TEV framework and the supply of ecosystem goods, uncertainty regarding the potential use and value of the services and attributes. The TEV framework aligns ecosystem in the future (Perman et al. 2011). Although directly to the goods, services and attributes framework the ecosystem may be underutilised at present, it could and the Millennium Assessment (2003) concept of prove to be valuable for tourism, research or other ecosystem services. Supporting services are valued commercial enterprises in the future. Option value through the other categories of ecosystem services. Page 18 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table 1.2 The linkages between the TEV framework and the supply of ecosystem goods, services and attributes Types of services Direct use Indirect use Option Non-use value value value values Provisioning Fresh water supply, fisheries, wild foods, X - X - (Goods) natural medicines, fuel, biochemicals Climate regulation, water purification, Regulating nutrient recycling, pollination, air quality - X X - (Services) regulation etc. Cultural Recreation & tourism, aesthetic values, X - X X (Attributes) cultural and religious heritage Primary production, soil formation, Supporting services are valued through the other Supporting nutrient cycling categories of ecosystem services Much of the confusion and debate around categorising Economic value of the environment can be defined as and assessing the value of ecosystem services revolves the value of an asset, which lies in its role in achieving around the extent to which different services should be human goals, be it through aesthetic pleasure, the treated as intermediate versus final services, and the production of marketed commodities or spiritual extent to which the ecosystem is responsible for the fulfilment (Barbier et al. 2009, TEEB 2010). Value is benefits described (Barbier et al. 2011). For example recognised as a person’s willingness to pay for the recreational benefits are derived from a combination services that flow from an asset and this willingness of natural and man-made capital. These problems only depends on the socio-economic context in which really exist for static assessments of value such as those valuation takes place, such as human preferences, the by Costanza et al. (1997, 2014). To some extent, this can distribution of income and wealth, culture, production be solved by focussing only on the final services in order technologies and institutions (Barbier et al. 2009). A to avoid double counting. However, since it is often the change in any of these variables ultimately affects the supporting or intermediate services that are affected by estimated economic value. In a market economy, money policy changes, it is far more relevant to assess changes is a universally accepted measure of economic value, in welfare that will result from a change in the state of because the amount someone is willing to pay for an natural capital. That way, the fact that values depend asset tells how much of all other goods and services partly on man-made capital, such as hotels and boats, is they are willing to give up for that asset. Market prices, not problematic to the analysis. however, do not always accurately reflect economic value, since many people are actually willing to pay more than the market price. 1.3.3 Economic value Economic surplus or total economic welfare is measured The way in which values of ecosystem services are as the total benefit that society derives from an expressed also varies between studies. Different economic transaction. Economic welfare is represented measures of value are relevant to different decision- by ‘consumer surplus’ and ‘producer surplus’ (Figure makers. Individuals and firms make decisions on 1.3). Consumer surplus is calculated as the difference the basis of their own financial and/or utility gains. between an individual’s willingness to pay for a product Governments make decisions on the basis of overall or service and the actual amount paid. For society as a welfare gains (including contribution to income and whole this is measured as the area below the demand employment as measured in the national accounts). At curve for a good and above the price of that good. The a more local level, municipalities may make decisions value of the good changes if the price or quality of the based on the generation of revenues, e.g. from property good changes, or if demand for the good changes as a rates. It is important to understand value from both an result of changes in price and quality of substitute or individual/firm perspective and a government or social complementary goods. Similarly, producer surplus is the planner perspective, since the former constitute the additional benefit that producers gain, in terms of profit, market forces of change, and the latter are required to when the price they receive for a good is more than the make decisions that are in the overall interest of society. minimum that they would be prepared to supply it for. In this study, we take a social planner’s perspective as far This is measured as the area above the supply curve as possible, but also highlight financial implications for and below the market price (Figure 1.3). Therefore, the the municipality. total net economic benefit or costs of a change in an ecosystem is the sum of consumer and producer surplus, less any costs associated with the policy or initiative. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 19 1.3.4 Present value This study focuses on the values of natural systems within eThekwini municipality. Depending on the approach used, some values are estimated as capital values, such as the capital costs required to use an engineering design to replace the supply of a service, and others as annual flows of value, such as the annual value of fuelwood production. In order to estimate the asset (capital) value of these natural systems (natural capital), it is necessary to consider the value of all the benefits generated over time. This is the present value, which is the discounted sum of the flow of benefits generated. Estimating the present value of each service requires deciding on a time frame for the analysis and a rate of discount. Discounting of the flow of values gives greater weight to present value than to future values. It is used to estimate the amount of money one would have to have now in order to reach the value Figure 1.3 Demand and supply curves for a good, showing the generated in each future time period if it was subject calculation of consumer and producer surplus to compound growth. The choice of discount rate is controversial, since higher discount rates reduce future values more and therefore reduce overall present Market values usually exist for services such as timber, values. From an investment perspective, it makes fuelwood and carbon sequestration/storage. Where sense to choose a discount rate that reflects the rate markets are absent, estimates have to be obtained of return on capital. However, this does not necessarily indirectly. This can be done by looking at related reflect the social rate of time preference, and it is markets. For example, land which is more fertile will often argued that environmental valuation should trade at a higher price, the differential reflecting the employ lower or declining discount rates that better value of soil fertility. Alternatively, unpriced services reflect intergenerational preferences. In this study, we can be valued by estimating how much it would cost to have followed the World Bank guidelines in applying a replace them, or the damages that might be incurred if discount rate of 6%. Given the World Bank’s investment they were removed. perspective, this certainly leads to an underestimate of the value of ecosystem services to society. In the USA, the Environmental Protection Agency recommends using a social discount rate of 2 – 3%, and in the UK, rates of 2.4 – 3.5% have been suggested (Perman et al. 2011). A range of estimates of present value are therefore provided using discount rates of 6% and 3%. Page 20 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 1.4 Overall approach and limitations change under a range of scenarios will become possible. This study was carried out as a desktop study based Some other caveats need to be borne in mind in on available data. Modelling assumptions were based interpreting the results of this study. Firstly, South Africa, on data from within the study area, drawing on the irrespective of its value, has an obligation to a certain regional, national or international literature only where level of biodiversity protection as a signatory of the necessary. The study focused on the direct values Biodiversity Convention, and areas within the eThekwini associated with the provision of natural resources, Metropolitan Municipality may be critical to meeting indirect use values associated with regulating services this mandate. Secondly, while we can assess current and ecosystem functioning, and the amenity values value, it is difficult to assess the future role and value of generated by ecosystem attributes. Estimation of option ecosystems in a rapidly changing world. Natural systems values and non-use values associated with ecosystems are likely to provide an important buffer to the impacts require survey-based methods and were thus beyond of climate change, and will enhance the resilience of the scope of this study. cities. These relationships are not well understood because of the uncertainties involved, and therefore The values of different types of provisioning, regulating trade-offs should be made with a large margin as a and cultural services were estimated based on precautionary measure. consideration of demand and supply as appropriate, and in the case of provisioning services, based on estimated sustainable yields. Because of the broad scope of the study, a wide variety of methods has been employed 1.5 Structure of the report to tackle the various types of ecosystem services, and The following two chapters provide some context. includes both well established and relatively novel Chapter 2 provides an overview of the eThekwini approaches. These methods are described in detail in Municipal Area (EMA), including location, geography, the relevant chapters of the report. administration and land tenure, population and economy, and a brief description of the natural The annual flows of value were expressed in terms as systems and biodiversity of the study area, most of close to net economic value as possible. Because the which is counted as part of the D’MOSS. Chapter 3 latter is the sum of producer and consumer surplus, and provides an overview of ecosystem services, their estimating consumer surplus requires stated-preference classification and their values. The next three chapters surveys, the estimates generally provide a conservative focus on provisioning, regulating and cultural services, estimate of the welfare gains from natural capital. respectively. These chapters are divided into individual Annual flows were then used to estimate the value of services falling under each category. Chapter 4 estimates natural capital as a net present value. the provisioning value, within sustainable limits, of natural areas that are used for the extraction of water, Attempting to put a value on natural capital has some sand, plants, wild meat and fish. Chapter 5 estimates the inherent limitations that are still being grappled with value of natural systems in terms of their contribution in the field of natural resource accounting (www. to flood attenuation, water quality amelioration, carbon wavespartnership.org 2). The values of ecosystem storage, and support functions for agriculture and services are not simply measureable as final values as in marine fisheries. Chapter 6 focuses on the amenity value the national accounts. Most ecosystem services act as of natural and semi-natural systems, in terms of the inputs into various forms of economic production, the impact that this as on property values and domestic and outputs of which would not accrue without a range of foreign tourism expenditure. Chapter 7 draws together inputs including man-made capital. For example, tourism the findings and discusses their overall implications. value might be realised as a result of infrastructure and tourism facilities as well as biodiversity attractions. In these case, the values yielded by ecosystems are difficult to isolate as an asset value. To get around this, our approach has been to try and determine what would be lost if the natural asset (or its relevant function) was lost altogether. In reality, these are not usually all-or- nothing situations, as natural capital may be degraded or diminished rather than lost. Nevertheless, this study can provide the platform from which studies of marginal 2 The Wealth Accounting and the Valuation of Ecosystem Services (WAVES) partnership, led by the World Bank, aims to promote sustainable development by incorporating natural resources into development planning and national economic accounts. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 21 This page intentionally blank. Page 22 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY II. OVERVIEW OF THE ETHEKWINI MUNICIPAL AREA (EMA) 2.1 Geography and climate The topography is generally rugged across the EMA, especially inland in the valleys of the uMngeni River The city of Durban is located in the eThekwini Municipal catchment near to the Inanda Dam (EPCPD 2012). There Area (EMA) in the province of KwaZulu-Natal, South are some flatter areas towards the coast which include Africa and contains the largest port on the east coast the old alluvial coastal plain that extends from just north of Africa (Figure 2.1). The EMA covers an area of of the uMngeni Estuary through the Durban city centre approximately 2297 km2 extending from the Tongati to just south of the Mbokodweni Estuary (IDP 2014). River in the north to the aMahlongwa River in the south. The rugged interior is interspersed with flat sandstone The eastern edge of the EMA is bounded by 98 km of tabletops which range in size from large expansive Indian Ocean coastline and to the west the boundary areas like in the Kloof, Waterfall and Hillcrest areas, extends inland to 50 km at its widest point at Cato to the smaller more isolated uplands like the Inanda, Ridge. The EMA is bordered by three other district Matabatule and Fudu mountains (EPCPD 2012). The municipalities; iLembe to the north, uGu to the south highest point in the EMA is 873 m near Bartlett Estate in and uMgungundlovu to the west. Hammarsdale. Figure 2.1 The eThekwini Municipal Area (EMA) is located on the east coast of South Africa and covers an area of 2297km2 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 23 Figure 2.2 Topographical map of the eThekwini Municipal Area Page 24 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY The area has a subtropical climate with humid wet The extension of services to the outlying peri-urban summers and mild dry winters (EPCPD 2012). The warm and rural areas has improved significantly over recent Agulhas Current that flows southwards along the coast years, with more households gaining access to services has a moderating influence on the climate, keeping such as running water and electricity. Potable water winter temperatures mild and summers warm and (within 200 m from dwellings) is available to 92% of the humid with a maximum average daily temperature of EMA’s population, electricity to 71% and basic sanitation 28°C (Roberts & O’Donoghue 2013). The rainy season (above the minimum service level) is available to 80% falls between September and March with a mean of the population (eThekwini Municipality 2014a). annual precipitation of over 1000 millimetres (Roberts However, flush toilets connected to the sewerage & O’Donoghue 2013). Temperatures are usually highest mains or to septic tanks are available to only 50% of at the coast and get cooler as one moves inland. The the city’s population. In 2012 the percentage of people change in temperature between seasons is greatest in living under conditions associated with poverty was still the higher altitude areas in the west. Rainfall tends to significant at 32.3% (IDP 2014). As a result, almost all be highest in the south of the EMA and along the coast river systems and accessible natural areas have been but rainfall seasonality is greatest in the west, with most impacted by some form of anthropogenic pressure of the rainfall falling within the summer months. The (EPCPD 2012). predominant winds blow parallel to the coastline in a north-easterly and south-westerly direction. The rural areas of the EMA occupy approximately 1500 km2 of land in the surrounding areas of the southern and western regions, and include the peri-urban areas 2.2 Land use and socio-economic context alongside the N2 and N3 corridors (SDF 2014). These rural areas generally have rugged and hilly terrain The eThekwini Metropolitan Municipality is one of six and dispersed settlement patterns with most of the metropolitan municipalities in South Africa and has the population living in traditional dwellings on communal third largest population, after Johannesburg and Cape land holdings. A significant portion of the rural land Town (StatsSA 2011a). The total population of the EMA is in the southern and western regions is under the approximately 3.44 million, which is just over one third ownership of the Ingonyama Trust. The Ingonyama of the population of KwaZulu-Natal (StatsSA 2011a) and Trust is a corporate entity that was established in 1994 has been projected to rise to approximately four million to administer the 2.8 million hectares of traditionally- by 2020 (IDP 2014). The majority of the population in the owned land in KwaZulu-Natal for the benefit and welfare EMA are African (74%), followed by Indian (17%), white of the members of the tribes and communities living on (7%) and coloured (2%; SDF 2014). As well as the urban the land. area of Durban, the EMA also encompasses significant areas under peri-urban and rural land uses (EPCPD 2012). The rural and traditional authority areas are largely Approximately 45% of the EMA is considered rural, characterised by severe poverty and unemployment 30% is peri-urban and only 25% is urban (SDF 2014). with many of the households reliant on social grants and This pattern is reflected in the population distribution surrounding natural resources (SDF 2014). The heavy (Figure 2.3). reliance on the natural resource base in these areas has resulted in environmental vulnerability which is further Although only covering 2.5% of the area of the aggravated by fragmented service delivery, unresolved province EPCPD 2012), the EMA accounts for 65.5% land tenure and a history of poor planning (SDF 2014). of KwaZulu-Natal’s and 10.7% of national GDP (IDP Communities in these areas participate in some small- 2014). Key economic activities include wholesale scale maize and sugarcane farming and also farm some and retail, manufacturing, trade, transport, storage vegetables and fruits such as bananas, mangos, citrus, and communication, financial business services and amadumbes (a type of sweet potato) and ground nuts community services (IDP 2014). (Institute of Natural Resources 2004). There are approximately 945 910 households in the Approximately 29% of the land in the EMA has been EMA with an average household size of 3.2 (eThekwini designated for agricultural use, although not all of this Municipality 2014a, SDF 2014). Of these, 55% are in land is used for agricultural purposes (EPCPD 2012). formal dwellings, 34% in informal dwellings and 11% in Subsistence farming activities are found largely in the traditional or rural dwellings (eThekwini Municipality traditional authority areas and formal agriculture is 2014a). The unemployment rate in the EMA was dominated by sugarcane farming, especially in the reported to have decreased from 31.4% in 2006 to 20.4% northern and western areas (SDF 2014). Other formal in 2011, lower than the national average (StatsSA 2011a). agriculture includes timber, dairy, beef, piggeries, broilers, layers, sheep and goats, aquaculture, vegetables, fruit and cut flowers (Institute of Natural Resources 2004). A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 25 Figure 2.3 Population density (people/km2) for the eThekwini Municipal Area Page 26 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 2.3 Biodiversity and D’MOSS The EMA is situated in the centre of one of 34 Global Biodiversity Hotspots, the Maputaland-Pondoland- Albany region (EPCPD 2012). This region contains more than 7000 vascular plant species, 25% of which are endemic to the area (EPCPD 2012). The fact that the EMA is located in a transitional zone between warm tropical and cooler temperate features as well as its topographical and habitat diversity means that it supports relatively high levels of biodiversity within this region (EPCPD 2012). The EMA contains three of the country’s eight terrestrial biomes (savanna, forest and grassland), eight nationally recognised vegetation types (eastern valley bushveld, KwaZulu-Natal coastal belt, KwaZulu-Natal hinterland thornveld, KwaZulu- Natal sandstone sourveld, Ngongoni veld, scarp forest, northern coastal forest and mangrove forest), 4000 km of rivers, 98 km of coastline, 18 major river catchments and 16 estuaries. These natural assets fall within the EMA’s open space system, which also includes agriculture, man-made parks and sports areas. Known as the Durban Metropolitan Open Space System, or D’MOSS, this system of green open spaces covers 75 000 ha, or almost a third of the total municipal area, above the high tide mark (Figure 2.4; eThekwini Municipality 2012). The D’MOSS system incorporates areas of high biodiversity value that are linked together in a viable network (eThekwini Municipality 2012). These include nature reserves such as Paradise Valley and Burman Bush, large rural landscapes in the upper catchment areas, riverine and coastal corridors and some privately-owned land (eThekwini Municipality 2012). A significant proportion of D’MOSS land is located within the traditional authority areas in the outer-west and southern planning regions. The EMA’s natural open space areas can be categorised into terrestrial, freshwater, estuarine and marine ecosystems. These are distinguishable in terms of the flora and fauna that are found in and around them and the ecosystem goods and services that they provide (EPCPD 2012). These natural and semi-natural areas, found largely on the periphery of the EMA away from the city centre, provide a wide range of ecosystem goods and services across the whole study area. These ecosystem services provide resilience and can help to buffer the impacts of climate change and other pressing issues. The key environmental issues and a detailed description of the EMA’s ecosystem characteristics are outlined below in more detail. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 27 Figure 2.4 Habitats and land cover in the EMA. Freshwater systems, grassland, thicket, woodland, forest, rocky outcrops, estuaries, and dams are all part of D’MOSS Page 28 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY The terrestrial landscape in the EMA is essentially a The EMA also has a vast network of rivers and mosaic of open grassland, woodland, thicket, thornveld wetlands that drain into 16 estuaries along its coastline and forest (EPCPD 2012). Almost all of the vegetation (Figure 2.5). Many of the original wetlands have been types in the EMA have been significantly transformed, lost due to urbanisation. The remaining wetlands are threatened and in need of protection, with several are mainly valley bottom wetlands and floodplain being critically endangered. Most (60%) of Dry and Moist wetlands associated with the larger rivers (eThekwini Ngongoni Veld (endangered and critically endangered Municipality 2011). The main river systems are the in KwaZulu-Natal, respectively) and almost all (99%) uMkhomazi, uMngeni, uMdloti, oHlanga, uTongathi, of the near-threatened Eastern Valley Bushveld is in uMlazi, Mbokodweni and iLovu. The larger river systems the traditional authority areas, the latter occurring in originate in the Drakensberg Mountains, the medium deeply incised river valleys. The vulnerable KwaZulu- rivers in the KwaZulu-Natal Midlands and the smaller Natal Hinterland Thornveld, dominated by Acacia trees, rivers close to the coast (DWA 2013a) The health of usually lies between these vegetation types. None of the river systems has been well studied (Ground Truth these is in formally protected areas, but a lot of this 2006) and continues to be monitored. Most rivers are is relatively well protected from transformation by under increasing threat from impacts such as flow the nature of the terrain (EPCPD 2012). Scarp forest modification, sand mining and pollution from industrial which occurs in steep gorges and scarps is also fairly and sewage effluent. Unplanned and extensive informal protected by the terrain, although it is heavily exploited settlement development in the upper catchment areas is for resources such as medicinal plants and building also having a large impact on these systems. materials. 6% of this is in protected areas. More than two thirds of the critically endangered KwaZulu-Natal There are a 16 estuaries found within the study area Sandstone Sourveld has been transformed, and only 1% (Figure 2.5). These range from small and temporarily is formally protected, in Krantzkloof Nature Reserve. closed estuaries with poor floral and faunal diversity, to The remainder of this vegetation, which is rich in forbs large permanently open systems with very diverse faunal and has high levels of endemism, is under threat from and floral communities. More than half of the estuaries fire and alien invasive plants. Coastal Belt grassland and found in the study area are highly degraded or in a poor bushveld vegetation occurs along the coastal plain up to state of health. The biggest threats include habitat loss, 660 m altitude, and all four types are highly transformed organic and chemical pollution from WWTW, informal and threatened, with less than 1% in protected areas. settlements and industrial activities, unregulated sand Coastal Forests have also become highly fragmented as mining, overexploitation, and upstream freshwater a result of development. Most of the remaining forests diversions or abstractions. are in formal planning scheme areas, and about 1% is protected. Small patches of Swamp Forest and Mangrove Outside of D’MOSS, the EMA coastline and marine Forests also occur in the study area, with 72% of the ecosystem includes the sandy beaches, rocky shores remaining mangroves being protected in the Beachwood and in-shore marine environment. The coastline and Mangroves Nature Reserve (EPCPD 2012). The health of marine environment plays an important role in tourism all the terrestrial ecosystems has been assessed in terms and recreation activities and is also productive in terms of broad categories. of fisheries. While most of the KwaZulu-Natal coast is a high energy environment, Durban central beaches are There is high plant diversity within these habitats, with a relatively sheltered. total of 2267 species recorded from 204 different plant families (EPCPD 2012). This represents more than half Appendix 1 provides a more detailed description of the the known families found in South Africa. Endemicity biophysical characteristics of the terrestrial, freshwater, is also high in the EMA with 379 species, 16% of the estuarine and marine ecosystems found within the study total, being classified as South African endemics (EPCPD area that have a bearing on their value. In particular, this 2012). There are also 82 mammal species in the EMA, includes information about their health, extent, location, including several threatened species, 69 reptiles, 27 flora and fauna, and abundance of natural resources. amphibians and 526 bird species (EPCPD 2012). The populations of many of these species in the EMA are inevitably threatened by urbanisation and the activities and impacts that come with it. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 29 Figure 2.5 The major and minor rivers found within the EMA, and location of the 16 estuaries Page 30 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY III. PROVISION OF NATURAL RESOURCES 3.1 Introduction However, few studies have been carried out within the EMA or other metros. Furthermore, while such The provisioning value provided by natural systems studies can provide estimates of current use, this use was estimated as the value of the sustainable yield of is not necessarily sustainable. Snapshot estimates natural resources, as far as this is demanded. Terrestrial, can therefore provide distorted estimates of value. freshwater and coastal ecosystems provide a number Ideally, direct use value should be estimated based of living and non-living resources which are harvested on a combination of expected demand and estimated for raw materials, food and medicine. In the eThekwini sustainable yields. However, in some cases, the only municipality these resources are predominantly available data are on actual harvests. harvested by poorer households on a subsistence basis or to generate some cash income. The greatest Available information on sustainable yields, market harvesting pressure comes from the rural communities prices and input prices for different resource types that occupy the approximately 1500 km2 of rural areas were obtained from the literature, using locally-sourced in the hinterland of the EMA, where many people live in information as far as possible (see Table 3.1). The traditional dwellings within Ingonyama Trust lands, as vegetation types and condition of the natural areas, as well as the people living in poor peri-urban communities assessed in detail by eThekwini municipality, was taken such as those that run alongside the N2 and N3 highways into account in estimating sustainable yields of the (SDF 2014), although markets for these resources will different resources. The estimated supply capacity of supply a much broader segment of the community, natural areas was then mapped in GIS. Urban typologies with some of the demand coming from both poor and and characteristics of the urban populations were used relatively wealthy households within the urban area. to estimate the general level of demand, taking access The types of resources that are harvested in the EMA constraints into account. In other words, no value was include water for domestic use, reeds and thatching assigned to protected areas or areas that would not be grass, firewood, poles, food plants, bush meat and fish. accessible to potential users. A map of urban demand This chapter estimates the provisioning value of natural typologies (based on explicit assumptions outlined habitats, taking factors influencing supply and demand below) was superimposed onto the current supply into account as far as possible. capacity to determine the current provisioning value in terms of aggregate net income to households associated The use of natural resources has been studied to with specific habitats and land parcels within the study varying degrees in the province, particularly in the area. rural communal land areas under the Ingonyama Trust. Table 3.1 Summary of available information in the supply and demand of natural resource use in the eThekwini municipality Natural Resource Available Information Harvesting • Peterson et al. (2012) developed a compendium of local wild-harvested species used in the informal trade in Cape Town, South Africa. Wild flora and fauna • High & Shackleton (2000) estimated the value of wild and domestic plants in home gardens in a rural village in Bushbuckridge in the Lowveld of South Africa. Natural Capital/ • De Wit et al. (2009) looked at developing a financially motivated case for investing in natural capital in the city Ecosystem Services of Cape Town. • Lannas & Turpie (2009) valued the provisioning services (hunting, livestock, water use, resource harvesting) of Natural resource use in two wetlands, one in Lesotho and one in Cape Town. Information about hunting and harvesting rates (urban wetlands context) • Mander (1998) conducted a comprehensive study of medicinal plant harvesting in KwaZulu-Natal providing Medicinal Plants and supply and demand data. Grasses • Turpie et al. (2007) estimated the consequences of changing land use in Drakensberg grasslands. Value and sustainable harvesting rates estimated for medicinal plants and grasses. • Kaschula & Shackleton (2009) estimated the quantity and value of wild meat offtake in a rural village in the Eastern Cape. Wild meat • Grey-Ross et al. (2010) assessed the illegal hunting on farmland in KwaZulu-Natal, South Africa and the implications of this on the Oribi antelope. • Shackleton et al. (2002b) estimated the direct use values of non-timber forest products from three rural villages in the Kat River Valley. • Shackleton et al. (2007) estimated the direct use values of non-timber forest products from two villages on the Transkei Wild Coast. Non-timber forest • Cocks & Wiersum (2003) documented the significance of plant diversity in rural livelihoods in the Eastern products (NTFP) Cape looking at uses of NTFP’s, amounts harvested and their value. • Shackleton A SPATIAL VALUATION OF THE NATURAL & Shackleton AND (2004) reviewed SEMI-NATURAL OPENthe importance SPACE AREAS NTFP’s of IN in rural livelihood ETHEKWINI security in South Page Africa. MUNICIPALITY 31 • Twine et al. (2003) estimated the direct-use values of savannah resources used by rural households in Limpopo, South Africa. Natural Resource Available Information Harvesting • Peterson et al. (2012) developed a compendium of local wild-harvested species used in the informal trade in Cape Town, South Africa. Wild flora and fauna • High & Shackleton (2000) estimated the value of wild and domestic plants in home gardens in a rural village in Bushbuckridge in the Lowveld of South Africa. Natural Capital/ • De Wit et al. (2009) looked at developing a financially motivated case for investing in natural capital in the city Ecosystem Services of Cape Town. • Lannas & Turpie (2009) valued the provisioning services (hunting, livestock, water use, resource harvesting) of Natural resource use in two wetlands, one in Lesotho and one in Cape Town. Information about hunting and harvesting rates (urban wetlands context) • Mander (1998) conducted a comprehensive study of medicinal plant harvesting in KwaZulu-Natal providing Medicinal Plants and supply and demand data. Grasses • Turpie et al. (2007) estimated the consequences of changing land use in Drakensberg grasslands. Value and sustainable harvesting rates estimated for medicinal plants and grasses. • Kaschula & Shackleton (2009) estimated the quantity and value of wild meat offtake in a rural village in the Eastern Cape. Wild meat • Grey-Ross et al. (2010) assessed the illegal hunting on farmland in KwaZulu-Natal, South Africa and the implications of this on the Oribi antelope. • Shackleton et al. (2002b) estimated the direct use values of non-timber forest products from three rural villages in the Kat River Valley. • Shackleton et al. (2007) estimated the direct use values of non-timber forest products from two villages on the Transkei Wild Coast. Non-timber forest • Cocks & Wiersum (2003) documented the significance of plant diversity in rural livelihoods in the Eastern products (NTFP) Cape looking at uses of NTFP’s, amounts harvested and their value. • Shackleton & Shackleton (2004) reviewed the importance of NTFP’s in rural livelihood security in South Africa. • Twine et al. (2003) estimated the direct-use values of savannah resources used by rural households in Limpopo, South Africa. • Glenday (2007) study on carbon storage and sequestration in eThekwini. Information used to estimate tree volumes and sustainable yields in specific D’MOSS habitats in the study area. • Shackleton (1993) estimated fuelwood harvesting and sustainable utilisation in communal land and in protected areas in the lowveld, South Africa. Information about fuelwood harvesting rates and sustainable yields. • Luoga et al. (2000) categorised different uses of resources and quantified the amount of wood used for Fuelwood and poles firewood and building poles in miombo woodlands in eastern Tanzania. • Dovie et al. (2002) estimated direct use values of woodland resources consumed and traded in a rural village in South Africa. Information on fuelwood harvesting and prices. • Dovie et al. (2004) assessed the fuelwood crisis in Southern Africa by relating fuelwood use to livelihoods in a rural village. Information on fuelwood harvesting and prices. • Barnes et al. (2005) Forestry accounts from Namibia provide insight into calculating tree volumes and sustainable yield estimates. • Turpie et al. (2010) assessed the aquatic ecosystem services of the Olifants, Inkomati and Usutu to Mhlatuze Water use water management areas. Information about water use, water prices and the value of rivers. • GroundTruth (2006) conducted a study on the state of the rivers in eThekwini providing health status data. • Dunlop & Mann (2012 and 2013) recent assessment of participation, catch and effort in the KwaZulu-Natal shore-based and offshore boat-based linefisheries. • Lamberth & Turpie (2003) assessed the role of estuaries in South African fisheries. Information on prices, Fisheries values and estuaries by each region. • Forbes & Demetriades (2008) produced a report on Durban’s estuaries providing information on health and fish of each system in the EMA. • Demetriades (2007) investigated the location and size of sand mining operation on rivers and estuaries in Sand mining eThekwini.CSIR (2008) assessed the quantities and values of sand supply from eThekwini rivers. Page 32 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 3.2 Approach For example, in the case of medicinal plants, woodlands and grasslands in a good condition have the ability to supply plant material sustainably due the size and 3.2.1 Sustainable yields distribution of this habitat in the EMA. Forests however Data on sustainable yields for each resource and have been over utilised in the EMA and their ability habitat type were obtained from the literature as far as to supply material sustainably has been significantly possible. Where estimates of sustainable yields were impacted on. Usually production from forests would not available, explicit assumptions were made based on be greater than from woodlands and grasslands due to estimates from comparable habitats. Average prices for the variety of plant species found in forested habitats. each of the natural resources were obtained from the However, their capacity to supply sustainable yields is far literature or based on expert opinion and inflated to less than woodlands and grasslands due to most of the 2015 prices. forested areas being overexploited and degraded in the EMA. The criteria and assumptions used to determine sustainable yields and values are described in more The size of each habitat patch was also considered detail below. These were based on information collected to be an important factor in estimating the capacity from relevant studies carried out within the eThekwini to supply resources. Patch size and connectivity in Municipality and other areas within South Africa. Where urban environments has an important influence on data were not readily available for South Africa, studies biodiversity and ecosystem functioning, though this is from southern Africa and further afield were used. also influenced by management within the remaining habitat patch and land-uses in the surrounding landscape (Wilson et al. 2009). Several studies have tried 3.2.2 Adjusting for habitat condition to determine the minimum patch size to support certain The condition of terrestrial and freshwater habitats flora and fauna within urban environments (Wilson et was taken from the eThekwini Municipality Spatial al. 2009, Hennings & Soll 2010). Based on these studies, Conservation Plan (EMSCP; Figure 3.1, EPCPD 2012). only patches greater than 12 ha were considered large D’MOSS parcels classed as being in a good condition enough to be able to support resource harvesting. (i.e. mostly natural with little or no degradation evident) were assumed to have the capacity to supply natural resources at 100% of the potential sustainable yield for that vegetation type and resource. Those classed as being of intermediate condition1 were assumed to be able to supply resources at 50% of their potential. Parcels classed as degraded2 were assumed to have lost the capacity to support sustainable natural resource harvesting. 1 < 50% of the area has low (5% - 33%) to moderate (34% - 66%) estimated levels of alien plant infestation and/or limited soil exposure (< 33% of area). 2 > 50% of the area has moderate (34% - 66%) to high (67% - 100%) estimated coverage of alien invasive plants and/or extensive soil erosion (> 33% of area) A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 33 Figure 3.1 Habitat and land use map showing the condition of the natural habitat found within D’MOSS Page 34 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Information relating to the condition and functioning (uTongathi, iSipingo, eziMbokodweni, Amanzimtoti and of estuaries was taken from Forbes & Demetriades Little Amanzimtoti) that are highly degraded (Forbes & (2008; Table 3.2) and the National Estuary Biodiversity Demetriades 2008) and are known to have poor water Assessment (van Niekerk & Turpie 2012). The condition, quality and reduced fish abundance and diversity. size and type of each estuary was linked to their capacity Therefore it was assumed that the provisioning value for to supply natural resources. Small and/or degraded fishery resources in these estuaries was zero. The Ngane estuaries with poor water quality, severely impacted Estuary, recorded as being in a fair state, has low fish habitats and reduced or completely removed estuarine numbers and diversity as a result of its small size. The processes were assumed to lack the capacity to support fishery resources value associated with this estuary was natural resource harvesting. There are five estuaries adjusted to be one third of its assigned value. Table 3.2 Estuarine state ranking based on health and functionality used to assign condition to each estuary in the EMA (Source: Forbes & Demetriades 2008) Estuarine State Description Estuaries with high levels of habitat integrity, good water quality, high diversity and high provisioning levels of Excellent goods and services Estuaries with most of the core estuarine habitat and estuarine support habitats still present, good water Good quality, diversity of habitats and species, and estuarine processes in place Estuaries with core estuarine habitat intact, some estuarine support habitats, impacted water quality and Fair some loss of diversity and key estuarine process in place Estuaries with impacted core estuarine habitat. Substantially reduced or no estuarine support habitats, Poor polluted water, substantial loss of diversity and/or abundance and key estuarine processes impaired Estuaries which have had major impacts on core estuarine habitats through infilling, canalisation and Highly Degraded pollution, substantially reduced or no estuarine support habitats and major loss of key estuarine processes 3.2.3 Adjusting for expected demand protection was considered available for the harvesting The demand for natural resources was assumed to be of natural resources. A number of resources are known driven by the numbers of rural and poor households to be located on agricultural land and within protected living within close proximity to natural areas. There are areas. However farm land is privately owned land and a significant number of rural settlements, peri-urban the therefore the harvesting of resources from these settlements and informal settlements spread widely lands and within protected areas is illegal. Therefore across the EMA, all of which are in close proximity to all land parcels that fall outside of tribal land and state natural open space areas and could be assumed to undeveloped land were excluded from the analysis. have a moderate to high demand for natural resources. Remoteness was considered to be a harvesting Because these settlements were so widely spread and constraint and therefore resources located on steep almost all natural space in the EMA was accessible, slopes were considered inaccessible and unlikely to only land ownership was used to determine which be harvested due to their position in the landscape. natural areas could be harvested. All tribal authority Any natural vegetation located on slopes steeper than land and all state undeveloped land not under formal 50 degrees was considered inaccessible for resource harvesting. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 35 3.3 Water 3.4 Sand Rivers and streams are used as an important source of Sand is harvested commercially and for subsistence water for domestic purposes in rural, peri-urban and purposes from rivers and estuaries in the EMA for informal settlements, both in the home for drinking building and construction. A significant percentage of and bathing and also directly in the rivers themselves these mining operations are known to operate illegally. for the washing of clothes and as drinking water for Both the large number of sand permits issued over livestock. Households tend to collect water in containers recent years and the illegal and unregulated activities for household use but washing is usually done within have raised concerns over the long-term sustainability of the rivers and streams. Census data provided the sand extraction in the EMA and implications for coastal most comprehensive estimate of the degree to which erosion (CSIR 2008). The CSIR’s (2008) comprehensive households in the EMA rely on rivers and streams study on this issue found that the rivers of the EMA as their main source of water. The reliance on rivers would naturally yield some 480 -720 000 m3 of sand per for domestic water use is low, with 91% of the EMA year, which would have provided an important supply population receiving their water supply from municipal of sand to the beaches of the Durban Bight. However, connections and a further 4% relying on water tanks about a third of this yield is now trapped in 12 large and rain water collection tanks. Approximately 0.5% of dams, and an estimated 400 000 m3 is removed annually households in the EMA rely solely on rivers and streams in sand mining operations, leaving the beaches starved for their main source of water (StatsSA 2011b). Although of sand. While sand is valuable as a resource, their study the Census data provided information about the number found that its value was outweighed by the damage of households relying on rivers and streams, the actual costs being incurred, and it was recommended that quantity of water collected from rivers in the EMA was estuary and riparian sand mining be stopped altogether. not known and therefore an average of 25 litres/person/ Thus we have not included sand as an extractive day, based on the basic human needs rule, was used. resource in this valuation study. This is similar to the amount found in an empirical study in in the Olifants, Inkomati and Usutu to Mhlatuze water management areas (Turpie et al. 2010). The value of water usage was estimated based on the replacement cost of having to pay for water in containers from local vendors. This was estimated to be approximately R0.26 per litre or R260 per m3 (taken from Turpie et al. 2010). It was estimated that a total of 27 500 litres of water per year per household were collected from rivers and streams. The average water usage across the EMA was estimated to be 24 m3 per km with an average value of R6 280 per one kilometre stretch of river or stream. This equated to a total usage of just over 122 million litres of water per year extracted from rivers and streams within the EMA with a total value of R32 million. The households with the greatest reliance on rivers and streams are located predominantly in the sub-places located in the outer-west and southern planning regions (Tribal Lands) and sub-places adjacent to the larger river catchments such as the uMngeni and in poorer areas such as KwaMashu and Inanda where households in these sub-places are collecting up to approximately 750 m3 of water per kilometre stretch of river (Figure 3.2). In areas with the highest usage, this equates to approximately R361 000 per kilometre per year. Page 36 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure 3.2 Estimated amount of water extracted from rivers and streams per sub-place in the EMA (m3 per km per year) A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 37 3.5 Woody resources Fuelwood is a dominant source of energy used in Table 3.1 Tree volumes calculated for forest and woodland/thicket rural households in South Africa. Within the EMA the habitats in the EMA. Volumes were calculated using data collected by Glenday (2007). majority of households live in urban areas with a direct electricity supply. However, there are areas in the EMA General Detailed De- Volume (m3/ha) that have a significant rural and peri-urban population Description scription that rely heavily on fuelwood as a source of energy for cooking and heating. Due to limited financial resources Forest Coastal 154.7 in these rural and peri-urban areas, households rely Forest Dune scrub 57.3 on fuelwood as a cheaper alternative to electricity and paraffin. The dependence on fuelwood is known to Forest Feral Plantations 104.3 be especially high in the traditional authority areas in Forest Scarp 145.6 the outer-west and southern planning regions of the EMA. However, there is not much information on usage Forest Riverine 145.6 levels in these areas. Wooden poles are harvested predominantly in the traditional authority areas too, for Forest Swamp 212.8 the construction of houses, fences and for other building Forest Transitional 25.3 infrastructure. Again, there is not much information Dry valley thicket about pole harvesting rates. It was assumed that the Thicket or Broadleaved 45.9 vegetated areas in and around the tribal land are the woodland most impacted and also the most degraded due to the Thicket Transitional 10.5 demand for woody resources in these areas. Woodland Closed 14.4 In order to determine the sustainable supply of fuelwood and poles, stand volumes for each vegetation type Woodland Open 4.0 within the EMA were calculated by using basal area and Estuary Mangrove Forest 260.5 canopy height data collected by Glenday (2007; Table 3.3). Volume is generally estimated from dimensional Estuary Swamp Forest 212.8 variables such as diameter and height in the form of linear equations that take into account local vegetation Basal area and height data were not collected for form. However, in the absence of localised equations, mangrove forests in the Glenday (2007) study. Therefore cubic volume of wood for standing vegetation may volumes were estimated using data from associated be estimated using the following simplified equation literature for mangrove forests (Kairo 2001, Kairo et (Magnussen & Reed 2004): al. 2002, Ajonina et al. 2014) and an average volume of 260.5 m3/ha was used (Table 3.3). A sustainable yield Volume (m^3/ha) = basal area (m^2/ha)×canopy height of 3% of standing crop was applied for all vegetation (m)/3 classes. This estimate was based on the finding that the sustainable biomass harvesting rates for most of the While this equation may somewhat overestimate or ecosystems found in the peri-urban/urban areas of the underestimate the volume of certain woody vegetation EMA are 2-4% of standing biomass per annum (Glenday with different forms, it nevertheless provides a first 2007) and Shackleton (1993) found the same when approximation for determining stand volumes in the studying harvesting rates in Limpopo. EMA. The amount of standing woody volume that was physically utilisable was assumed to be 90% for fuelwood and 15% for poles (Barnes et al. 2005), allowing for a component of the standing volume to be assumed unsuitable for harvesting. Woody resources were valued by using average prices of fuelwood and poles recorded in surveys and in the literature from South Africa, Botswana and Namibia. Prices were adjusted to 2015 Rands where necessary. Prices for fuelwood and poles were obtained from LaFranchi (1996), Loxton, Venn & Associates (Botswana, 1986), Ntshona (2002), Shackleton et al. (2002a), and Barnes et al. (2005). An average of R1058 per m3 for fuelwood and R882 per m3 for poles were used. Page 38 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY A total of 44,000 m3 of fuelwood was estimated to R266 per hectare. The average value per hectare for be the sustainable output per year in the EMA, with fuelwood across the EMA is R1 707 (Table 3.4). The an estimated total value of R46.5 million (Table 3.4). largest vegetation patches where fuelwood can be Forested areas and thicket have the highest value harvested are located in the outer-west and southern and supply 60% and 37% of the fuelwood output planning regions on traditional authority land (Figure respectively. Estuarine forests have the highest per 3.3). hectare value at R3 400 and woodland the lowest at Table 3.3 Estimated total value (R million/year) of fuelwood yields in different vegetation types in the EMA Vegetation Vegetation sub- Available Total fuel- % of total Total Value Average Type types Area (ha) wood (m3) (R mil- value per lion/y) ha (R/ha) Estuarine Swamp & mangrove 114 345 <1 0.37 3 380 Forest Coastal, scarp, Forest transitional, dune and 8 765 26 308 60 27.84 2 470 swamp Wetland Riparian 8 16 <1 0.02 2 079 Sub-tropical Dune 11 13 <1 0.01 1 230 Vegetation Dry valley thicket, open, Thicket 15 884 16 216 37 17.16 816 closed and dune Closed, open, Woodland 9 247 1 112 3 1.18 266 transitional TOTAL 34 030 44 011 46.57 1 707 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 39 Figure 3.3 Estimated annual sustainable fuelwood output (m3/ha) from different habitats in the EMA Page 40 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY A total of almost 7500 m3 of poles can be sustainably shrubs are able to supply timber that can be used in harvested for construction purposes in the EMA each construction. It was also assumed that fewer people year. This equates to a total estimated value of R6.5 collect poles, whereas those relying on fuelwood would million (Table 3.5). The amount of wood harvested for be significantly higher. Similarly to fuelwood, forested poles to be used in the construction of houses and areas and thicket supply the majority of poles and these fences is much lower than the amounts of fuelwood vegetation types are predominantly located in the outer- harvested. This is because only larger trees and certain west and southern planning regions (Figure 3.4). Figure 3.4 Estimated annual sustainable timber pole output (m3/ha) from different habitats in the EMA A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 41 Table 3.4 Estimated total value (R millions/year) of timber pole yields in different vegetation types in the EMA Vegetation Vegetation sub- Available Total Poles % of total Total Value Average Type types Area (ha) (m3) (R mil- value per lion/y) ha (R/ha) Estuary Forest Swamp & mangrove 114 58 <1 0.05 470 Coastal, scarp, Forest transitional, dune & 8 766 4 487 60 3.96 360 swamp Wetland Riparian 8 3 <1 0.00 289 Sub-tropical Dune 11 2 <1 0.00 171 Vegetation Dry valley thicket, Thicket broadleaved woodland, 15 884 2 703 36 2.39 113 closed & dune Closed, open, Woodland 9 247 185 2 0.16 37 transitional TOTAL 34 030 7 437 6.57 240 3.6 Food and medicinal plants The demand for indigenous medicines in the eThekwini stock from sellers in other areas of KwaZulu-Natal, Municipality has always been high (Mander 1998) and Mpumalanga and the Eastern Cape. forms an important part of traditions and culture in many households. In 1998, Mander estimated that There is very little quantitative information on households were spending between 4 - 8% of their sustainable yields and even the harvesting of wild foods annual incomes on indigenous medicines in KwaZulu- and medicines. For medicinal plants, Mander (1998) Natal. The same study found that urban consumers estimated that yields of 0.78 kg/ha/y were possible in believed that the demand for indigenous medicines woodlands and grasslands of KwaZulu-Natal, and Turpie would not decrease but would remain stable or increase et al. (2007) estimated that medicinal bulbs and herbs in the future as it was not considered an alternative could be harvested in the Drakensberg grasslands at a to western medicine and was considered essential for sustainable rate of 20 kg/ha/y at a value of R9.40/kg. An everyday welfare (Mander 1998). average of the two sustainable yield estimates was used and applied to both forests and grassland/woodland Medicinal plants are collected from wild plant stocks habitats at 10.4 kg/ha/y at a price of R15.80/kg. that are located in a number of different vegetation types within the EMA. The harvesting of wild plants is Estimates of harvests of wild fruits and vegetables range not managed and little cultivation of these plants takes from 32 – 53 kg/household/y (Cocks & Wiersum 2003) place (Mander 1998). Bulbs, roots, whole plants, bark to 104 kg/hh/y (Shackleton & Shackleton 2004; fruit and leaves are collected from a variety of indigenous only), 128 kg/hh/year in Limpopo (Twine et al. 2003), wild plants. The wild stocks in the EMA have been 1254 kg/household/year in Lesotho (Lannas & Turpie severely depleted and degraded as a result of the high 2009) and 5 kg/hh/y in peri-urban Cape Town (Lannas demand and the lack of any resource management. & Turpie 2009). For this study it was assumed that only The unsustainable use of medicinal plants has led to a households in peri-urban and traditional authority areas decrease in supply and the loss of some certain scarce were harvesting wild fruits and vegetables and that the plants. Unsustainable commercial medicinal plant harvest rate would be similar to that of peri-urban areas harvesting has caused habitat degradation and species in Cape Town. Therefore a conservative estimate of 5 mortality in a number of areas within the EMA (EPCPD kg/hh/year was used for this study (Lannas & Turpie 2012). The most extreme form of this is the mass ring- 2009) at a price of R0.85/kg for wild vegetables and barking of trees, resulting in significant tree mortality R1.60/kg for wild fruits (Turpie et al. 2010). in woodland and forested areas. Although medicinal plants are still commercially sold in the eThekwini The sustainable output of wild food and medicinal Municipality, it is expected that a significant proportion plant products in the EMA each year is estimated to of the medicine being sold is brought in from outside of be approximately 300 000 kg with a total value of the EMA and commercial harvesters are having to travel R4.7 million (Table 3.6). Forest, thicket, woodland greater distances to collect plants and/or to purchase and grassland areas all contribute significantly to this Page 42 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY figure, with forests having the highest per hectare than the sustainable provisioning value for the same value at R101/ha. Freshwater wetlands also provide a area a decade ago. Figure 3.5 shows the sustainable number of medicinal plants but their supply in the EMA output (kg/ha) of medicinal plants across the study area. is compromised by the poor condition of a number of wetland areas. It is believed that this value is much lower Table 3.5 Estimated total value (R million/year) of wild foods and medicinal plants in different vegetation types in the EMA Vegetation Vegetation sub- Available Total % of total Total Value Average Type types Area (ha) Medicinal (R mil- value per Plants (kg) lion/y) ha (R/ha) Grassland Primary, secondary 8 097 33 499 11 0.53 62 Freshwater Floodplain mixed 3 818 9 926 3 0.16 41 wetland Estuary Forest Swamp & mangrove 114 595 <1 0.01 82 Coastal, scarp, Forest transitional, dune & 8 766 70 298 24 1.11 101 swamp Wetland Riparian 8 21 <1 0.00 41 Sub-tropical Dune 11 43 <1 0.00 62 Vegetation Dry valley thicket, Thicket broadleaved woodland, 15 884 138 019 47 2.18 90 closed, dune Closed, open, Woodland 9 247 44 552 15 0.71 78 transitional TOTAL 45 945 296 952 4.70 70 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 43 Figure 3.5 Estimated annual sustainable wild food and medicinal plant output (kg/ha) from different habitats in the EMA Page 44 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 3.7 Non-woody raw materials Grasses and reeds are harvested for the production on average 20 kg/ha. Mmopelwa et al. (2009) estimated of crafts and for the construction of houses in rural the direct use value of certain plants in three villages and peri-urban areas. Most of the grass and reeds are adjacent to the Okavango Delta and found that the harvested for the construction of roofs, and for crafts harvesting of thatching grass and reeds was sustainable such as mats and baskets. There is very little information with an estimated average harvest of 28 kg/ha for about the quantities and types of grasses and reeds thatching grass and 70 kg/ha for reeds. It was assumed that are harvested for these purposes within the EMA. that the harvesting of reeds and grasses in the EMA In recent times more houses in the tribal lands and would be less than the harvesting rates estimated by peri-urban areas tend to use corrugated iron and other Mmopelwa et al. (2009) for the Okavango Delta but building materials for roofs rather than the traditional similar to those estimated by Turpie et al. (2007) and thatching grass. This could be a result of a decreasing Twine et al. (2003). Therefore for the EMA a sustainable supply of this resource or it could be a result of cheaper, harvest of 24 kg/ha/y was applied at a price of R6.70 per more accessible building materials. kg, taken from Turpie et al. 2007. Estimated were based on information from Twine et al. It is estimated that the sustainable output of reeds (2003), Turpie et al. (2007) and Mmopelwa et al. (2009). and grasses in the EMA is just over 200 000 kg with a Twine et al. (2003) found that on average households total value of R1.4 million (Table 3.7). Grassland areas in a village in Limpopo were sustainably harvesting contribute 77% of the output and freshwater wetland just over 10 bundles of thatching grass per year, which areas 23%. Grassland habitats are scattered across the equated to 30 kg/ha. Turpie et al. (2007) found that EMA, with the largest patches being found in the outer- thatching grass was patchily distributed across the west region close to Cato Ridge and areas north of southern Drakensberg grassland region and provided Inanda (Figure 3.6). Table 3.6 Estimated total value (R million/year) of grasses and reeds in different vegetation types in the EMA Vegetation Vegetation sub- Available Total Grass % of total Total Value Average Type types Area (ha) & Reeds (R mil- value per (kg) lion/y) ha (R/ha) Grassland Primary, secondary 8 097 154 609 77 1.04 121 Freshwater Floodplain mixed 3 818 45 811 23 0.31 81 wetland TOTAL 11 914 200 419 1.35 101 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 45 Figure 3.6 Estimated annual sustainable grass and reed output (kg/ha) in the EMA Page 46 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 3.8 Bush meat Communities living in the rural areas of the EMA are River Valley (Shackleton et al. 2002b). In all cases the known to hunt for wild mammals and birds. However, harvesting rate did not appear to be unsustainable. a significant portion of this hunting is thought to be Although the EMA contains a significant area of rural illegal and takes place on protected and private land, landscapes, these tend to be densely populated and often with packs of dogs and the use of snares (Grey- the vegetation patches small and fragmented, and Ross et al. 2010). Small antelope, rodents and birds probably support lower densities of animals than in the are the most commonly caught wild meat (Kaschula areas that have been previously studied. We therefore & Shackleton 2009) but not much is known about the conservatively estimated a sustainable yield of 104 kg/ magnitude of off-take by local communities in the km2/y; an average of the lower estimates of wild meat rural areas within the EMA. Illegal hunting with packs harvesting from other regions in the country. An average of dogs is an increasingly concerning problem in many price of R21.90 per kg was used based on prices taken parts of KwaZulu-Natal where people with dogs are from Shackleton et al. 2007 and Turpie et al. 2010. hunting wild animals on private and protected lands for gambling purposes and not for subsistence (Grey-Ross The annual sustainable hunting offtake was estimated et al. 2010). These hunts involve groups of owners and to be 26 000 kg of wild meat and birds with a total their dogs taking bets on whose dog will kill the animal estimated value of R565 500 (Table 3.8). Forest, thicket first. These groups can be large and the money involved and woodland habitats are estimated to be able to quite substantial. This method of hunting is completely supply the majority of this output. These values are unselective and there are concerns about the impacts relatively low when compared to other studies, however overhunting is having on some of the more rare and they are based on sustainable offtake levels and do not endangered mammals, such as the Oribi antelope consider the illegal hunting of mammals on private and (Ourebia ourebi). protected lands. Therefore these levels are assumed to be realistic in a predominantly urban and fragmented Studies of wild meat harvesting (mainly mammals and area. The spatial representation of hunting across the birds) have found offtake rates in the region of 209 kg/ EMA is shown in Figure 3.7 and as with the other natural km2/y in coastal regions of South Africa (Shackleton resources, the highest levels of output are associated et al. 2007), 268.6 kg/km2/y or 3 kg/person/y in the with habitats located in the outer-west and southern inland communal areas of the Eastern Cape (Kaschula planning regions where natural habitat patches are & Shackleton 2009) and 151 kg/km2/y in the Kat larger, less fragmented and on communal land. Table 3.7 Estimated total value (R million/year) of hunting in different vegetation types in the EMA Vegetation Vegetation sub- Available Total Hunt- % of total Total Value Average Type types Area (ha) ing (kg) (R mil- value per lion/y) ha (R/ha) Grassland Primary, secondary 8 097 1 353 5 0.03 3.50 Freshwater Floodplain mixed, 3 826 191 <1 0.00 1.10 wetland riparian Estuary Forest Swamp & mangrove 88 4 <1 0.00 1.10 Coastal, scarp, Forest transitional, dune and 8 766 6 639 26 0.15 7.70 swamp Dry valley thicket, Thicket broadleaved woodland, 15 884 13 728 53 0.30 12.10 closed, dune Closed, open, Woodland 9 247 3 901 15 0.09 8.00 transitional TOTAL 45 907 25 817 0.57 5.60 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 47 Figure 3.7 Estimated annual sustainable hunting output (kg/ha) across the EMA Page 48 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 3.9 Fishery resources A variety of fishery resources are harvested from Everett 2014). These figures relate only to the managed rivers, estuaries and the marine environment both component of the fishery, it is expected that the catch commercially and for subsistence. A number of fisheries could be higher as the true numbers of subsistence operate along the KwaZulu-Natal coastline including fishermen are not known. There is limited information the commercial line fisheries, subsistence line fishery, about the value of this fishery. Everett (2014) reported recreational line fisheries, small scale commercial net that the total annual catch value ranged from R150 000 fisheries, recreational net fisheries, illegal gill and seine to R920 000 based on estimates from studies conducted net fishery, and inshore invertebrate fisheries (Everett in 2010 and WIOFish Annual Report (2013). 2014). There has been relatively few traditional or artisanal fisheries along the KwaZulu-Natal coastline Illegal gill net fishing has been taking place in the rivers due to the high-energy nature of the coastline making and estuaries of KwaZulu-Natal for decades (Everett fishing with low technology difficult, and the fact 2014). The gill-nets used in the region range in length that historically the Zulu people who moved to the from 10 – 1000 metres depending on location (Everett coast were traditionally cattle farmers (Everett 2014). 2014). The nets are usually set along estuary margins However, there is a section of society living in the EMA or across estuary or river channels targeting a wide that do rely on fishery resources. There are three main range of different fish species. The majority of the subsistence fisheries found in the EMA; the subsistence people involved in this fishery are unemployed, poor line fishery, the illegal gill net fishery and the small-scale people that live within close proximity to estuaries or rocky shore and sandy beach invertebrate fisheries. large river systems. Most of the illegal gill netting in KwaZulu-Natal takes place in the St Lucia iSimangaliso Subsistence fishers are usually referred to as poor, area and associated systems and there are only a few unemployed people who harvest fish and other aquatic estuaries in the EMA where illegal gill netting has been organisms in close proximity to where they live as a known to occur, namely Durban Bay and the uMngeni means of meeting their basic needs of food security Estuary, however Everett (2014) notes that there has (Branch et al. 2002). Usually these fishers fish along been periodic netting in many of the smaller estuaries the sea or estuary shoreline and cannot afford vessels. along the KwaZulu-Natal coast too. Because the fishery They use basic equipment and generally catch their is illegal, effort and participation data is difficult to own bait, such as sand prawns, red bait and mussels estimate or collect. The fishery has high level impacts (Everett 2014). Most of the fish that are caught are on resources, is considered to be unsustainable and was for personal or family consumption and are not sold. therefore not valued as part of this study. Only when catches are large will they either be sold or bartered (Everett 2014). Estimating the total number Small-scale and subsistence fishers living along the of “true” subsistence fishers along the coast is difficult. KwaZulu-Natal coastline collect a variety of intertidal In recent years there has been an increase in the and sandy beach organisms, including both mobile number of people operating under the pretence of and sessile invertebrates (Everett 2014). In total there being subsistence line fishermen so as to exceed daily are approximately 556 invertebrate fishers along the bag limits and to sell their fish at local markets (Everett coast but 300 are estimated to be living within the 2014). As explained by Everett (2014) this has resulted in iSimangaliso Wetland Park and the remainder (256) serious challenges in trying to successfully identify and are spread along the rest of the KwaZulu-Natal coast manage true subsistence line fisheries in KwaZulu-Natal. (Everett 2014). The harvesting of intertidal and sandy beach organisms is usually undertaken by women and Roving creel surveys conducted along the coast of children. WIOFish (2013) reported that the small-scale KwaZulu-Natal have shown that a relatively small invertebrate fishery caught in total 8 043 kg of mangrove percentage (3-6%) of the total number of shore fishers crabs which equates to over 400 000 individual crabs, are “true” subsistence line fishers (Dunlop 2011). Using 200 kg of ghost crab, 100 kg of mole crabs, 1 700 kg this data Everett (2014) predicts that the best estimate of mixed invertebrates and 9 000 kg of mussels. The of total subsistence line fishers along the KwaZulu-Natal mixed invertebrates that are harvested include redbait, coast is approximately 4000 people. It is estimated that urchins, whelks, octopus, sea cucumbers and limpets the annual amount of fish harvested in estuarine and (Everett 2014). marine environments by subsistence line fishermen along the KwaZulu-Natal coast is around 23 tons (Everett 2014). The main species that are caught include shad Pomatomus saltatrix (22%), grey grunter Pomadasys urcatum (15%), stonebream Kyphosus lithophilus (14%), largespot pompano Trachinotus botla (12%), blacktail Diplodus capensis (11%), karanteen Sarpa salpa (8%) and stumpnose Rhabdosargus spp. (6%; WIOFish 2013, A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 49 Table 3.8 Summary of the subsistence line and invertebrate fisheries along the KwaZulu-Natal coastline in terms of numbers, annual catch, annual effort and value in 2012 (Source: Everett 2014) Fishery Number of fishers Total Annual Catch Total Annual Effort Total value of catch Subsistence Line 4000 23 tons 48 000 days R920 000 Fishery Subsistence invertebrate 556 19.3 tons unknown R445 000 fisheries There is very scant information on subsistence fishing the value of estuarine subsistence fishing in northern in the rivers of the EMA. Most of the information KwaZulu-Natal. This was used to estimate potential available relates to fishing in estuaries and along value, and adjusted using estuarine health scores the coastline. In the Olifants, Inkomati and Usutu to from the National Estuary Biodiversity Assessment Mhlatuze Water Management Areas (Turpie et al. 2010), (van Niekerk & Turpie 2012), to provide an order-of- fishing was found to be marginal activity with only 4% of magnitude estimate of the provisioning value of these households in the former homeland areas sampled being resources in the EMA. involved in fishing. Freshwater catches in this survery region comprised mostly of cichlid Tilapia rendalli and The subsistence fishery value associated with estuaries Sarotherodon mossambicus, tiger fish Hydrocynus in the EMA was estimated to be close to R6.3 million vittatus, barbel Clarias gariepinus and mud suckers (Table 3.10), with Durban Bay and the uMngeni Estuary Labeo rosae and L. rubropunctatus (Turpie et al. 2010). contributing the most to this value. Whilst these two While some freshwater fishing is likely to occur in the estuaries are considered to be degraded, they are both EMA, its value is likely to be negligible. large open estuarine systems that are known to have diverse and resilient fish communities. A number of Fishing in the estuaries of the EMA is more common. smaller estuaries which are in a highly degraded state Subsistence fishing effort and values have been such as the iSipingo and aManzimtoti were considered estimated for a number of South African estuaries in a unlikely to support subsistence fishing activities. The fish rapid national assessment (Branch et al. 2002), but data communities in these estuaries have been reported to were patchy and no comprehensive studies have taken be poor with very low numbers recorded in the system place. Using these data, Turpie et al. (2010) determined and low carrying capacities (Forbes & Demetriades 2008, a relationship between fishing value and area (R125 000 van Niekerk & Turpie 2012). per Ln(A), where A = area, in ha) in order to estimate Page 50 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table 3.9 Summary of the subsistence line and invertebrate fisheries along the KwaZulu-Natal coastline in terms of numbers, annual catch, annual effort and value in 2012 (Source: Everett 2014) Estuary Health category Ecological health Total Annual Effort Total value of catch (Forbes & Demetria- category des 2008) uTongati Highly degraded E 0 0 uMdloti Poor D 617 700 4 412 oHlanga Poor D 603 500 4 828 uMngeni Highly degraded D 679 800 2 955 Durban Bay Highly degraded D 851 700 936 iSiphingo Highly degraded F 0 0 eziMbokodweni Highly degraded E 0 0 aManzimtoti Highly degraded D 0 0 Little aManzimtoti Highly degraded D 0 0 iLovu Fair C 677 000 3 009 uMsimbazi Good B 569 200 5 992 uMgababa Good B 593 100 5 158 Ngane Fair B 99 900 9 083 uMkhomazi Fair C 616 800 4 437 uMahlongwane Good B 397 300 16 552 iMahlongwa Fair C 555 300 6 533 TOTAL 6 261 300 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 51 This page intentionally blank. Page 52 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY IV. REGULATING SERVICES 4.1 Introduction impacts will affect economies and human wellbeing on a global scale, but more so in developing countries that are Regulating services relate to the capacity of natural more reliant and land and natural resources (Tol 2012). and semi-natural ecosystems to regulate essential Adaptation to these changes could come at a high cost. ecological processes and life-support systems through The conservation and restoration of natural systems bio-geochemical cycles and biospheric processes (de thus helps to reduce the rate at which greenhouse gases Groot et al. 2002). Regulating services enable natural accumulate in the atmosphere and the consequent ecosystems to supply provisioning services such as food impacts of climate change. This is a global benefit. and water as they maintain the capacity of system to continue to function over a range of conditions (Simonit Concerns about the loss of natural systems and the & Perrings 2011). They have been identified by the impacts of climate change have motivated efforts to Millennium Ecosystem Assessment (MEA) as the least quantify the role and value of these ecosystems in the understood but potentially most valuable of services global carbon cycle, and have also encouraged political provided by ecosystems (MEA 2005, Simonit & Perrings efforts (Glenday 2007, Davies et al. 2011, Timilsina et al. 2011). In addition to maintaining ecosystem health, 2014). The eThekwini Municipality has made a strong these regulation functions provide numerous services commitment to addressing the causes and impacts of that have direct and indirect benefits, such as clean climate change and the subsequent loss of ecosystem water and air, soil development and stabilisation, climate goods and services (Glenday 2007). The eThekwini regulation, agricultural support and biological control. Municipality’s Environmental Management Department However, because many of the benefits are indirect, (EMD) has established the eThekwini Environmental they are often not recognised until they are degraded, Services Management Plan (EESMP) and is a member disturbed or completely lost (de Groot et al. 2002). of the International Council for Local Environmental Initiatives (ICLEI) Cities for Climate Protection (CCP) Amongst the most important examples of regulating Program. As part of this program the municipality has services are those that relate to water. The hydrological been working towards climate change mitigation and services valued as part of this study are flow regulation adaptation, focusing on land cover management and and water quality amelioration (nutrient and sediment urban land use planning. retention). The valuation and mapping of these two services required complex hydrological modelling, which The capacity for carbon sequestration and storage varies is outlined in detail in Appendix 2. Other regulating between different types of ecoystems and in different services valued and mapped in this study include carbon locations, and can be estimated based on a combination storage, nursery function (i.e. contribution to marine of satellite data and ground-level sampling. In a study fisheries), and agricultural support. conducted in 2005/6, Glenday (2007) estimated that the total amount of carbon stored in all the major vegetation types of the EMA open space system is 6.6±0.2 million 4.2 Carbon storage tonnes of carbon (Mt C) equivalent, or 24.3±0.9 million Climate change caused by increases in the emissions of tons of carbon dioxide (Mt CO2). Some 8 400 – 9 800 greenhouse gases will carry a cost of about 2 – 7% of tonnes of carbon are sequestered per annum (tC/y). The Gross Domestic Product (GDP) in different parts of the spatial distribution of this service is shown in Figure 4.1. world by 2050 (Fankhauser & Tol 1997). Natural systems are understood to make a significant contribution to Carbon densities (tons carbon per hectare, tC/ha) for all global climate regulation through the sequestration the major landcover classes in the EMA, as calculated and storage of carbon. About half of the biomass of by Glenday (2007), were used for this study (Table 4.1). vegetation, both above and below ground, comprises Glenday (2007) determined the mean carbon density carbon. Furthermore, carbon accumulates in the soils per carbon pool for all vegetation types sampled in as a result of leaf litter. When natural systems are the EMA. For example, forest carbon densities were degraded or cleared, much this carbon is released into determined for above ground biomass, below ground the atmosphere, especially if the degradation is for biomass, coarse deadwood, litter, herbaceous vegetation fuel wood production or due to burning for grazing and soil to a depth of 30cm, whereas grassland carbon (Hoffa et al. 1999). These emissions contribute to global density was calculated based on above ground biomass, climate change, which is expected to lead to changes below ground biomass, and soil to a depth of 30cm. The in biodiversity and ecosystem functioning, changes in carbon densities were assigned and mapped to specific water availability, more frequent and severe droughts vegetation types and to more general land cover types and floods, increases in heat-related illness, and impacts using the most recent D’MOSS and landuse GIS layers. on agriculture and energy production (IPCC 2007). These Heavily urbanized areas were assumed to not have any carbon storage. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 53 A recent estimate puts this value at US$29 (US$1 equal occur to South Africa would be proportional to its GDP to R11.84 in May 2015) per tonne of carbon in 2015, contribution to Africa, scaled by level of vulnerability and this is expected to increase by about 2% per year to climate change. The Notre Dame Global Adaptation (Tol 2012). While developed countries emit more Initiative (ND-GAIN) vulnerability index was used to carbon, developing countries are expected to incur scale GDP contributions across Africa. The vulnerability proportionally greater costs in terms of percentage of index measures a country’s exposure and sensitivity GDP. to negative impacts of climate change. The overall vulnerability index is scored based on six life-supporting Estimates of the social cost of carbon are based on the sectors; food, water, health, ecosystem services, human impacts of climate change on country GDP outputs habitat and infrastructure. Based on this index, it was aggregated at a global scale. The most recent estimate estimated that South Africa’s share of the social cost of placed the social cost of carbon at US$34 per ton of CO2 carbon borne by Africa would be 11.7%. (in 2015 USD; Nordhaus 2017;US$1 = R11.84). While developed countries emit more carbon, developing Thus, while the global damage costs that this amount of countries are expected to incur proportionally greater carbon could produce are over R9.8 billion, the damage costs in terms of percentage of GDP. Nordhaus (2017) costs to South Africa resulting from a loss of the carbon estimated that of the total global cost of carbon, only 3% stocks within the EMA would be approximately R34.3 would be borne in Africa. Therefore the cost that would million per annum. Table 4.1 Carbon densities (tC/ha) for land cover classes in the EMA (Source: Glenday 2007) Landcover Sub-Type Carbon density (tC/ha) Woodland Dry Valley Thicket/Broadleaf 121 Coastal Bushclump Grassland 82 Wooded Grasslands Acacia Savanna, Protea and Faurea Woodland 66 Coastal Scarp Forest 199 Transitional Forest 103 Forest Coastal Lowland Forest 166 Dune Scrub and Dune Forest 133 Riverine Forest 165 Floodplains and freshwater wetlands 149 Wetland (non-woody) Estuarine wetland 244 Swamp Forest 287 Wetland Forest Mangrove Forest 375 Barringtonia racemosa & Hibiscus tiliaceus Forest 301 Secondary grassland 75 Grassland Primary grassland 62 Recreational Parks, golf course & sports fields 102 Sparse rural (disturbed woodland) 59 Settlements Dense rural 15 Fallow crop lands 75 Field crops Commercial market gardening & cropland 37 Plantations 85 Tree crops Fruit trees 150 Cemetery 102 Utility Grassed road verges & reservoirs 46 Alien Thicket 53 Alien Vegetation Alien Woodland 92 Feral Plantation 121 Page 54 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure 4.1 Total carbon storage (tons/ha) A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 55 4.3 Fisheries support (nursery function) South African estuaries support several species of fish There are 16 estuaries in the EMA with a total estuarine that are dependent on nursery areas for at least their open water area of approximately 972 ha, or 2.6% of the first year of life (Lamberth & Turpie 2003). The capacity total KwaZulu-Natal estuarine open water environment. of estuaries to function as nursery areas is dependent More information about these estuaries can be found on the condition of their habitats and their fish stocks. in Appendix 1. In KwaZulu-Natal there are 71 estuary- These, in turn, are dependent on the quantity and associated fish species that are caught in fisheries quality of freshwater inflows, the management of (Lamberth & Turpie 2003). Fish diversity and abundance habitats, and fishing pressures within the estuaries. differs between estuaries of different sizes and types, Estuarine fish species have been categorised according with higher species richness associated with larger and to their level of dependence on estuaries (Table 4.2), permanently-open systems (Lamberth & Turpie 2003), with the estuaries playing an important role for spceis in such as Durban Bay and the uMngeni Estuary. However, categories I and II. Most estuary-dependent fish species there are a number of estuaries withi the EMA that have enter estuaries as larvae or post larvae (Whitfield & become severely degraded as a result of significant Marais 1999) and once the estuarine dependent phase flow modifications, very poor water quality, habitat is complete, they leave for the marine environment destruction and overfishing (DWA 2013a). This has where they become available to marine fisheries, and affected their current contribution of each estuary to upon maturity contribute to the spawning stock (Wallace nursery value. 1975a,b). Some of the larger estuary systems, such as the uMngeni Estuary, are also known to provide an This study used catch data and associated economic important nursery habitat for penaeid prawns. Adults value of recreational and commercial fisheries as spawn offshore, and once the eggs have hatched and described by Lamberth & Turpie (2003) in conjunction developed to the post-larva stage, they enter estuaries with more recent (2009/10) fishery survey data for where they reside as juveniles until they are ready to KwaZulu-Natal fisheries that included updated estimates return to the sea (Forbes & Forbes 2013). on participation, effort and average catches published Table 4.2 The five major categories and subcategories of fish that utilize South African estuaries (Whitfield 1994). Category Description Estuarine species that breed in southern African estuaries: I Ia. Resident species, no record of spawning in marine or freshwater environments Ib. Resident species that do have marine and freshwater breeding populations Euryhaline marine species that normally breed at sea, with juveniles showing varying degrees of dependence on southern African estuaries: II IIa. Juveniles dependent on estuaries as nursery areas IIb. Juveniles occur mainly in estuaries, but are also found at sea IIc. Juveniles occur in estuaries, but are usually more abundant at sea. III Marine species that occur in estuaries in small numbers, but are not dependent on these systems IV Freshwater species, whose penetration into estuaries is determined mainly by salinity tolerance Catadromous species that use estuaries as transit routes between marine and freshwater environments, but may also occupy estuaries in some regions V Va. Obligate catadromous species that require freshwater in their development Vb. Facultative catadromous species that do not require a freshwater phase in their development A number of estuarine dependent fish species are by Dunlop & Mann (2012, 2013). The fish species exploited by recreational and commercial fisheries in included in the catch data were categorised based on the inshore marine area. KwaZulu-Natal is a sought their level of dependence on estuaries (Table 4.2). In the after location for recreational shore- and boat-based commercial fisheries, fish were valued at their market angling. Many thousands of visitors come to KwaZulu- prices, whereas in the recreational fisheries, a value was Natal annually to fish along the shore. This industry is of assigned to fish equal to their market price multiplied immense importance to the economy of the province by ratio of total expenditure to total market value of the and each fish caught effectively brings in much more catch. This was based on the assumption that differences revenue than would be obtained from the commercial in average market prices also reflected the relative value catching and selling of the fish. of the different species to recreational anglers. For the Page 56 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY recreational fisheries, the gross output value of these 2012, 2013) St Lucia, the largest estuarine system in fisheries was taken to be the expenditure by the fishers the country, was closed off from the sea and had been on their activities, which include expenditure in the closed for almost ten years. It was therefore assumed accommodation and retail sectors, as well as on travel, that St Lucia over this period did not contribute to the which translates into direct gross income (turnover) in nursery value along the KwaZulu-Natal coastline and was those sectors. In the absence of available data on the therefore removed from the analysis when determining expenditure by recreational anglers, this was estimated individual estuarine contribution to nursery value. based on the comprehensive study of McGrath et The percentage contribution of Kosi Bay to the overall al. (1997). Current expenditure was estimated on nursery value in KwaZulu-Natal was also reconsidered the basis of average value generated per angler day. based on the knowledge that the increase in traditional The economic value of the commercial fisheries was fish traps and their efficiency in trapping fish has a estimated based on the recent catch estimates and the significant effect on the numbers of fish able to leave landed catch value of the fisheries as a measure of gross the estuary for the sea. It was assumed that Kosi Bay output. contributes only 10% of what it would if traditional fish traps were not in the estuary. The data collected by Dunlop & Mann (2012, 2013) showed a 60% decrease in annual fishing effort within Updated catch data were not available for the the recreational fishery compared to data collected in recreational spearfish sector or the commercial net the earlier Lamberth & Turpie (2003) study. Effort and fishery in KwaZulu-Natal, and these fisheries were not catch was also significantly lower for the commercial included in this analysis. The estuarine contribution to line fishery, but this was a result of implemented policy the catch value in these two sectors was estimated to changes. Using the updated effort data and subsequent be comparatively small (R2.7 million) or just 1.3% of update for average value generated per angler per day near shore fishery value attributed to KwaZulu-Natal an overall current value for recreational and commercial estuaries by Lamberth & Turpie (2003). Commercial fishing was estimated. (legal) gill netting in KwaZulu-Natal was phased out with the allocation of medium-term commercial fishing rights Based on updated estimates of participation, effort and in 2003 and the remaining commercial beach-seine catch in KwaZulu-Natal fisheries, it is estimated that the operation near Durban lands very low volumes of fish, recreational and commercial fisheries in K waZulu-Natal estimated at 7 tons by Beckley & Fennessy (1996). are worth R429 million (Table 4.3). Of this R350 million (or 82%) is attributable to recreational shore angling Based on the estimated value of the fisheries and and only R16.6 million to the commercial line fishery. the percentage contribution of estuarine fish, the Recreational boat angling and charter angling is worth estimated total contribution of all KwaZulu-Natal approximately R62.3 million. estuaries to coastal and inshore marine fisheries was R106.8 million (Table 4.3). Applying estuary size and fish To determine the value of the estuary contribution, the health scores to all estuaries in KwaZulu-Natal it was values of the species categorised by Whitfield (1994) estimated that the overall contribution of estuaries in as Category II were multiplied by their estimated level the EMA was R11.4 million (Table 4.4). This represents of dependence on estuaries (Category IIa = 100%, IIb a 11% contribution to KwaZulu-Natal nursery value. = 90% and Category IIc = 30%). The overall nursery uMkhomazi, Durban Bay, oHlanga and uMngeni have value for estuaries in KwaZulu-Natal was estimated by the highest percentage contribution and associated multiplying the total fishery value for KwaZulu-Natal nursery value. Estuaries that are severely degraded and by the percentage contribution of estuary fish caught small in size, such as the iSipingo and Little aManzimtoti, within each fishery. contribute almost nothing (Table 4.4). The overall functionality of the EMA estuaries is approximately 35%, In order to assign a nursery value to each individual therefore two thirds of the value that these estuaries estuary in KwaZulu-Natal the size of each estuary was could provide to KwaZulu-Natal fisheries has been lost. multiplied by the fish health score (taken from van Niekerk & Turpie 2012) for each estuary to determine an overall “effective nursery area” for each estuary. The value for each estuary was then divided by the sum of all the KwaZulu-Natal healthy estuarine area and a percentage contribution of each estuary to nursery value was determined. From this a total nursery value for the EMA could be estimated by multiplying the overall nursery value in KwaZulu-Natal by the percentage contribution for each estuary. It should be noted that during the most recent comprehensive surveys and data collection for fisheries in KwaZulu-Natal (Dunlop & Mann A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 57 Table 4.3 Estimated participation, annual effort and catch in four KwaZulu-Natal near-shore fishing sectors based on data from (Dunlop & Mann 2012, 2013) and the associated KwaZulu-Natal fishery value and KwaZulu-Natal nursery value based on the amount and percentage of the total contributed by estuary-associated fish species Fishery Participation 1 Annual effort 2 Catch (t) Value of Total Value Average Fishery (R million/y) value per ha (R/ha) Shore- 54,685 801,692 263 349.5 30.5 106.60 angling Recreational 2,768 30,435 457 1.3 1.7 0.02 boat angling Charter boat ~100 5,898 245 61.0 0.11 0.07 angling Commercial 51 3,331 785 16.6 0.85 0.14 line fishing Total 1750 428.4 106.8 1: Shore angler participation is number of anglers; Boat fisheries are number of boats 2: Effort is shore angler-days.y-1, and number of boat launches. y-1 Table 4.4 Percentage contribution of estuaries in the EMA to KwaZulu-Natal nursery value and the estimated nursery value for each individual estuary based on size and fishery health scores. Estuary Fish Health Category % contribution to Nursery value R/ha/y (van Niekerk & KwaZulu-Natal (R millions) Turpie 2012) nursery value uTongati E 0.03 0.04 10 084 uMdloti E 0.22 0.24 8 403 oHlanga E 0.14 0.15 13 445 uMngeni D 1.46 1.56 18 487 Durban Bay E 5.19 5.54 8 403 iSipingo F 0.03 0.03 3 361 Mbokodweni E 0.11 0.12 13 445 aManzimtoti E 0.05 0.05 10 084 Little aManzimtoti F 0.02 0.02 6 723 iLovu C 0.78 0.84 23 529 Msimbazi B 0.58 0.62 30 252 uMgababa B 0.46 0.49 28 571 Ngane C 0.04 0.04 23 529 uMkhomazi D 1.22 1.30 18 487 uMahlongwane C 0.15 0.16 25 210 iMahlongwa D 0.14 0.15 20 168 TOTAL 10.6 11.4 Page 58 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 4.4 Agricultural Support (pollination) natural vegetation as an alternative food source for domesticated honeybee pollinators. The distance honey Natural habitats support organisms that provide bees travel to collect pollen will be dependent upon a agricultural support services in the form of pollination number of factors including season, habitat complexity and the control of agricultural pests. Crop pollination by and colony size (Abou-Shaara 2014). Distance recorded insects is an essential ecosystem service that increases for honey bees foraging in lucerne fields in the USA both the yield and the quality of crops (Melin et al. ranged from under 50 m to almost 6 km (Hagler et al. 2014). It has been estimated that as much as 30% of 2011). While there is no information on movements of worldwide food production is reliant upon pollination by honey bees in the EMA, we can assume that there is insects that rely on natural vegetation (De Groot et al. some demand for pollination service by farmers within 2 2002). However, there has been relatively little empirical km of their farms. research carried out on natural systems’ contribution to pest control. There are two subspecies of the indigenous The total value of the pollination service to the fruit honeybee Apis mellifera; A. m. scutellata occurs in trees and commercial market gardens was estimated the summer rainfall regions of South Africa and A. m. based on the number of hives required per hectare capensis is found in the winter rainfall regions. These and the replacement cost of hiring beehives in South honeybees are the most important pollinators of many Africa (R306; Allsopp et al. 2008). The value of the South African crops because they can be managed by pollination service was mapped by distributing the total humans at the scale needed for large-scale commercial value across the natural habitats within 2 km distance pollinator services. from fruit tree plots and commercial garden markets based on scores of the ability to provide habitat for Of the crops grown within eThekwini Municipality, bees. We assumed that intact natural habitats were many are wind-pollinated, including sugar and maize. best at delivering this service and degraded habitats In the case of root crops such as sweet potato and were not as effective. The fruit tree crops and market cassava, production is not directly dependent on gardens are located in small patches throughout the pollination, since the plants are usually propagated with EMA, however the largest areas requiring pollination cuttings, although pollinators are required in breeding services are located predominantly in the outer-west programmes. However, several crops are directly and the northern planning regions, as indicated in green dependent on insect pollination, including subtropcial in Figure 4.2. The highest potential pollination value is fruit crops such as mangoes, papayas, avocados and associated with vegetation patches in a good condition litchis, and nut trees such as macadamia, cashews and that are surrounding market gardens and tree crops. almonds. These crops are likely to benefit from wild Coastal and scarp forest and open grassland were colonies of bees occurring in untransformed vegetation estimated to have a value of between R65.00 to R90.00 around the tree crops or gardens, saving on active per ha. Thicket and woodland areas were estimated to pollination costs incurred by hiring of bee hives or have a higher value at R130.00 per ha. The total cost dusting. In addition, domesticated bees kept by farmers of replacing the pollination service to 1144 ha of fruit for own purposes, for rental to other farmers and for and garden crops would be just over R1 million per year honey production are at least partly dependent upon assuming a conservative estimate of 3 beehives required naturally-occurring vegetation as a source of forage, per hectare. especially during winter. The number and location of these beehives is however, unknown. Most rural households mainly grow their own maize, dry beans and sweet potato which provide the majority of their home-grown food (Institute of Natural Resources 2004). While none of these crops require pollination there are certain households that grow their own fruit trees and have small garden crops which benefit from the pollination service provided by natural pollinators. These crops and fruits only make up a small proportion of the home-grown produce. While we cannot account for the service provided to individual households, we assume that it is small. However, within the EMA there are just over 1000 ha of known commercial fruit tree plantations and commercial market gardens (eThekwini GIS department). We can assume that these areas do benefit from pollination services from natural pollinators, or rely on nearby A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 59 Figure 4.2 The location of market gardens and fruit tree crops and the associated pollination value (R/ha/y) Page 60 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 4.5 Flow regulation Hydrological models can be used to predict the magnitude of particular floods and to quantify the link The combination of weather-related (e.g. rainfall between changes in land use and land cover, and flood intensity, extent and duration) and geophysical (e.g. risk (Kareiva et al. 2011). Using this approach the severity catchment size, geomorphology, soil and land use) of flooding in terms of water volumes and flow rates, characteristics are the main factors that influence and corresponding damages from the storm event can flooding (Kareiva et al. 2011). Natural systems such as be estimated. The hydrological model provides the wetlands and rivers or ecosystems with deep permeable opportunity to define the capacities of different land soils can regulate flows through the landscape by slowing cover types to supply flood regulation in the EMA. High flows by means of storage and vegetative resistance rainfall events and localised flooding in the EMA has in and facilitating infiltration into soils. In this way these the past caused damage to infrastructure and property systems ameliorate the potential impacts of flood events as well as resulted in the loss of life. Large quantities by reducing the flood peaks and lengthening the flood of solid waste and refuse in rivers and stormwater period at a lower level, reducing bank and streambed pipes exacerbate the flooding problem in the EMA. erosion (Vellidis et al. 2003), as well as reducing the risk Stormwater pipes become overloaded and culverts and of damage caused by flooding of downstream areas. grates become blocked. As a result, water backs up The landscape capacity for infiltration of rainfall also rapidly, overtopping barriers and bursting river banks, contributes to groundwater supply and/or dry season inundating adjacent properties and roads. surface flows in areas downstream. Dry-season flows are critical to aquatic ecosystem health, as well as to rural The flood attenuation service can be valued using the populations that are directly reliant on rivers or springs lower of either flood damages avoided the avoided for agriculture, domestic use and livestock watering. The costs of replacing the natural systems with alternative key factors influencing storm peak mitigation are canopy flood mitigation options. The avoided damage costs are interception, soil infiltration, soil water storage and the extra costs that would be incurred in the form of location in the landscape. incremental losses from increased flooding if the natural ecosystems were transformed and also includes the Ecosystems, such as wetlands, floodplains and forests, opportunity cost of having to increase setback lines in affect the water balance within a river catchment greenfields areas. The replacement cost method involves through interception, evaporation and infiltration estimating the costs of infrastructure that would be (Nedkov & Burkhard 2012). Interception depends on the required to provide the same level of flood mitigation structure of the land and vegetation above ground (i.e. as the natural systems. In the urban context, as more land cover) and infiltration is strongly influenced by soil land becomes transformed, cities such as Durban tend properties (Brauman et al. 2007, Nedkov & Burkhard to respond to the resultant increased flood risk by 2012). Surface runoff, which is the main factor for flood implementing engineering solutions such as changes to formation, is also strongly influenced by other abiotic the stormwater infrastructure. Indeed, the city is already factors such as topography. Hydrologic ecosystem on a path to increasing the capacity of its infrastructure services can have preventive or mitigating functions in preparation for anticipated increases in the size of (Nedkov & Burkhard 2012). Certain land cover, such as storm events as a result of climate change (Schulze et al. forests and deep permeable soils, are able to redirect or 2010). Therefore a replacement cost approach was used. absorb incoming water and rainfall ultimately reducing surface runoff and the amount of water entering river For this study, a hydrological model was set up for the systems. Other ecosystems, such as floodplains and entire catchment area of the eThekwini Municipality wetlands, provide storm peak mitigation services using the PC-SWMM software (see Box 4.1). This (Nedkov & Burkhard 2012). These ecosystems provide model was set up to run design flood events in order retention space for any excess water, thereby slowing to determine the influence of natural vegetation areas flows and reducing the impact and power of the flood. on flood hydrographs at strategic points relating to the Therefore the conversion of forests and wetlands into location of existing flood conveyance infrastructure. The agricultural or developed land increases the amount flood hydrographs generated under current conditions of hardened surface thereby increasing the volume were compared with what they would be if the natural of runoff and flooding associated with storm events systems were transformed to urban land use. This (Kareiva et al. 2011). This tends to be valid for medium provides an indication of the impacts of loss of natural and small return period storm events as vegetation areas on flooding and the difference can be construed as captures water as it flows through the landscape an estimate of the flood attenuation benefit obtained by through canopy interception and enhanced infiltration. retaining the natural systems. However, vegetation has limited ability to mitigate flooding associated with large return period storm events because enhanced soil infiltration only captures a small fraction of total precipitation depth for such storm events (Kareiva et al. 2011). A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 61 Box 4.1 Summary of the flood modelling method (details in Appendix 2) The full eThekwini catchment system was modelled using the US-EPA SWMM5 hydrology and hydraulics engine, interfaced by the PC-SWMM software. The calibration methodology adopted focused heavily on reducing the uncertainty during model set up. GIS subcatchment data available for this project from the eThekwini Municipality (EM) flood studies were in the order of 1 km2 and larger. The subcatchments were further discretised into smaller subcatchments, in the order of 0.2 km2, within the eThekwini Municipal Area (EMA). For the region outside of EMA, Shuttle Radar Topography Mission (SRTM – 30 x 30m cell size resolution) data were acquired and special analysis tools were used to discretise the model into 0.5 to 1 km2 sized subcatchments, with larger subcatchments closer to the source areas (Drakensberg). A spatial analysis tool was then used to process the flow paths, watershed boundaries, and river centre lines. The outlet points for the model were then identified and selected; i.e. dams, estuaries and stormwater infrastructure. EM flood models, based on geometric HecRAS files, were imported into the PC-SWMM model. These datasets account for approximately 30% of the EMA and contain limited stormwater infrastructure (e.g. culverts, bridges). The stormwater network shapefile for the EMA was incomplete and contained numerous errors and inconsistencies. These networks were amended where possible, however due to the extent of the required revisions, a more detailed stormwater network was only included in the U60F quaternary catchment. This quaternary catchment encompasses the main developed part of the city, and is also the study area for the green urban development scenario analysis, which is in an accompanying report. Available current GIS landuse files (e.g. zoning files, D’MOSS) were collated and reviewed. These files were concatenated into one consistent landuse polygon shapefile and the landuse classifications were reviewed and summarised into a common set of landuse conventions. The resulting shapefile was ‘groundtruthed’ using aerial imagery for the whole of the EMA. The Southern African National Land Cover dataset (2013/14) was used for the catchments outside of the EMA. The 72 different landuse descriptions were summarised into a more manageable list of landuse categories consistent with those used within the EMA. A number of input parameters are required for SWMM5. These include hydraulic parameters, soil infiltration properties, rainfall and water quality parameters. The determination of the catchment characteristics was estimated using a spatial analyst tool for zonal statistics. Raster files were generated to represent the information required for the hydraulic and hydrological models, with reference to each subcatchment. The most significant input hydraulic parameter is the percentage of impervious area (Imperv. %). The hydraulic parameters were assigned to each landuse classification based on literature. The largest proportion of rainfall losses over pervious areas generally occur due to soil infiltration. The Green-Ampt method was adopted for this study. This method provides a soil memory as opposed to a broad brush coefficient approach. Three user-specified soil parameters were used; i.e. capillary suction head, saturated hydraulic conductivity, and the maximum available moisture deficit. Average daily abstractions and return flows/discharges were added as point sources at the appropriate junctions. There are a number of large dams in the EMA. For simplicity, the model was split by routing all the flows entering the dam to an outfall. A new flowpath was created downstream of the dam. A user-defined hyetograph was used as the precipitation input into the model. The hyetograph was created using the total daily mean-areal precipitation depths derived by Smithers & Schulze (2000) for a 24-hour design storm. The temporal distribution was derived using a synthetic SCS Type II distribution for 2-, 5-, 10- and 20-year return periods. The final SWMM model of the full EMA comprised about 30 000 subcatchments. The eThekwini catchment system was divided into 3 separate models, representing the northern, central and southern catchments to reduce simulation running times (Figure 4.3). Page 62 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure 4.3 The full eThekwini catchments showing flow paths The additional flood volumes that would occur under different return period flood events (e.g. 1:10 years), would require larger drains, culverts, etc., depending on the size of the event they are designed to deal with. Thus a second model was developed in order to estimate the capital costs of the structures required under the present versus the without-vegetation scenarios (see Box 4.2). The difference, together with associated differences in maintenance costs, is the total life-cycle cost avoided, which can be converted to a net present value which is the value of the service. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 63 Box 4.2 Summary of the method used to estimate engineering cost savings (details in Appendix 3) An analysis was undertaken to determine the difference in the replacement cost of stormwater infrastructure (including both conveyance and storage) between the existing stormwater infrastructure in the EMA and infrastructure that would be designed to different specifications depending on the amount of natural vegetation within the EMA. The adopted methodology for this analysis focused on estimating the cost difference of the infrastructure with and without natural/undeveloped areas in the EMA. The relationship between cost and flow for the status quo was established and used to estimate the cost of infrastructure based on changes caused to the flow as a result of having natural areas within the EMA developed into the average land use, i.e. what happens to the flows if the current D’MOSS is replaced with the average land use type within that catchment? The methodology used to determine stormwater infrastructure engineering costs is outlined as follows: 1. Identify all stormwater infrastructure within the EMA 2. Divide the infrastructure into four major categories (bridges, canals, culverts and pipes) 3. Determine the dimensions of all infrastructure in the EMA 4. Estimate the costs of the existing infrastructure based on these dimensions 5. Assign rainfall design return periods to each category of infrastructure 6. Calculate the maximum open channel flow from the Mannings equation (i.e. the threshold flow) 7. Estimate a scaling relationship between infrastructure dimensions and flow using theoretical uniform flows 8. Simulate the design rainfall return periods and estimate the maximum design flows for each scenario (status quo versus average land use) 9. Use the maximum flows, the existing infrastructure dimensions and the threshold flows to scale the existing infrastructure to required infrastructure dimensions under the average land use scenario 10. Use the new infrastructure dimensions to estimate the cost of the scaled infrastructure 11. Compare the cost of the existing infrastructure to the cost of the scenario infrastructure Because much of the flood conveyance infrastructure is overdesigned for various reasons including the problems of blockages by litter, the estimation initially yielded a small cost to increase the size of the structures to deal with the difference in flows. Since this is clearly downward biased, a correction was then applied to adjust for this overdesign and produce a more comparable estimate from which to derive the realistic difference in value. The savings in the capital cost requirements for flood conveyance were estimated to be in the range of R63 – R338 million (Table 4.5, Figure 4.4). This represents a 0.7% - 3.5% capital cost saving in stormwater infrastructure. Including an estimated 6% of capital costs as an annual repair and maintenance cost (eThekwini Municipality 2015), this suggests that the flood attenuation service provided by natural systems in the EMA has a net present value in the order of R107 – R571 million (average R339 million; Table 4.5). We consider the upper bound value to be the closest approximation of the value of the service. The highest per hectare values associated with D’MOSS are located in catchment U20L and U60F, catchments that are situated above the built up areas of Durban city centre and Durban North (Figure 4.4). These D’MOSS areas in upper catchment areas therefore provide a significant flood attenuation service. Page 64 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table 4.5 Lower and upper bound flood attenuation values (R million) for each quaternary in the EMA. A zero value indicates that there is no stormwater infrastructure within that catchment. Quaternary Lower bound Upper bound Catchment Capital cost Maintenance Flood Capital Maintenance Flood savings Cost (NPV attenuation cost Cost (NPV 20 attenuation 20 yrs, 6%) value (R savings yrs, 6%) value (R million) million) U10M - - - - - - U20J - - - 0.1 0.1 0.2 U20K 0.1 0.1 0.2 0.1 0.1 0.2 U20L 0.3 0.2 0.4 106.2 73.1 179.3 U20M 18.0 12.4 30.4 25.7 17.7 43.4 U30B 20.0 13.8 33.8 31.7 21.8 53.6 U30D 2.3 1.6 3.9 77.7 53.5 131.2 U60C 3.4 2.4 5.8 3.5 2.4 5.8 U60D 2.1 1.4 3.5 3.5 2.4 5.9 U60E 0.0 0.0 0.1 0.4 0.3 0.7 U60F 13.9 9.5 23.4 85.3 58.7 144.0 U70B - - - - - - U70C - - - - - - U70D 2.1 1.4 3.5 2.1 1.5 3.6 U70E 0.0 0.0 0.0 0.1 0.1 0.2 U70F 1.2 0.8 2.0 1.7 1.2 2.9 Total 63.3 43.6 106.9 338.1 232.7 570.7 The average estimated flood attenuation value for natural vegetation in each catchment was mapped to the green open space areas within that catchment using an estimate of the relative contribution of different land parcels to flood attenuation modelled using InVEST modelling software (Glenday 2012). This provided the relative flood attenuation score given to each grid cell within the EMA based on rainfall, soil and landcover. The value per pixel was then determined for all green open space areas (based on the latest GIS data of D’MOSS), total flood retention value for each catchment, the retention score given to the pixel and the sum of all pixel flood retention scores. The value per pixel was then divided by the area of the pixel to arrive at a flood retention value per hectare estimate for all green open space areas in the EMA. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 65 Figure 4.4 Flood attenuation value associated with natural systems (D’MOSS) (NPV, R/ha) in the EMA Page 66 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 4.6 Sediment retention Human activities within the landscape can lead to In the EMA, natural sediment transport from catchments increased soil erosion and the introduction of nutrients is the main source of beach sand. Most of the sand that into river systems from agricultural activities and human is supplied to the coast comes from bed load in the wastes. Agricultural expansion, encroachment into river, with very little sand contributed by the suspended natural wetlands and the removal of natural vegetation load (CSIR 2008). However, both dams and sand mining result in elevated levels of erosion and subsequent have all but cut off the supply of sand to the coast, with increases in sediment loads being carried downstream. dams trapping almost 100% of coarse sediments that The total sediment load being transported in rivers is flow into them (CSIR 2008). This is not a problem that made up of bed load, suspended load and dissolved can be solved by the conservation of landscapes in the load. The bed load is the portion that is transported EMA, but can only be solved by expensive engineering along the river bed, is coarse and generally moves at solutions such as off-channel dams, and the control velocities slower than the flow. The suspended load is or elimination of sand mining. Notwithstanding their particulate sediment that is held in the water column, contribution to beach erosion problems, the fact that and is made up of smaller particles such as clay and fine the dams trap sediments is also costly, and this cost silt. The dissolved sediment load is the material that is elevated when sediment yields from the catchment is chemically carried in the water. When flow speeds are elevated by human activities. Here, ecosystem drop—when rivers enter dams, lakes, wetlands or services do play a role in reducing the potential extent estuaries—the loads that are carried tend to drop out of of these costs due to increasing human activity in their suspension and accumulate, with the smallest particles catchments. In the EMA, sedimentation problems are taking longest to settle out. In this section, we focus on mainly associated with the water supply dams (Box 4.3) the problem of sedimentation of man-made structures and Durban Harbour. Because of the steep gradients of and harbours. Elevated loads of suspended sediments much of the EMA, sedimentation of drainage networks is also contribute to water quality problems, which are fairly limited (Geoff Tooley, pers. comm.). addressed in Section 4.7. The extent to which sediments end up in river systems is determined by a number of factors including soils, rainfall patterns (amount and intensity), slope and the type and amount of vegetative cover. Vegetative cover prevents erosion by stabilizing soil and by intercepting rainfall, thereby reducing its erosivity (De Groot et al. 2002). This is particularly valuable where soils are highly erodible. Vegetated areas, especially wetlands, may also capture the sediments that are eroded from agricultural and degraded lands and transported in surface flows, preventing them from entering streams and rivers (Blumenfeld et al. 2009, Conte et al. 2011). This protects downstream areas from the impacts of sedimentation, which can include impacts on water storage capacity, hydropower generation and navigability of rivers (Pimentel et al. 1995). While some level of sedimentation of dams is expected under natural conditions and planned for, elevated catchment erosion either incurs dredging costs or shortens the projected lifespan of dams and related infrastructure. Globally, anthropogenic sedimentation has been estimated to account for about 37% of the annual costs of dams (i.e. $21 billion) in terms of replacement costs (Basson et al. 2009). In urban contexts, elevated sediment loads also have to be removed from sewerage systems, storm water drainage systems and harbours. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 67 Box 4.3 The main water supply dams of the EMA There are three large water supply dams in th EMA that have part of their catchment areas falling within the EMA, and that therefore potentially benefit from sediment reduction services delivered within the EMA as well as in the rest of their catchment areas1. Inanda Dam, located on the uMngeni River, is the largest of three dams and supplies water to Wiggins WTW the second largest treatment works in the EMA. It is the largest of the three water supply dams. Inanda Dam is located 25 km upstream of the uMngeni Estuary and is known to cut off the main source of coarse material to the estuary (Cooper 1993), trapping 99.6% of coarse sediments (CSIR 2008). Hazelmere Dam is located on the uMdloti River and supplies water to Hazelmere WTW on the north coast and also supplies water for agricultural irrigation. CSIR (2008) found that present sediment yields in this catchment are much lower than the historic levels, the result of significant land use changes (less subsistence agriculture and more natural vegetation). Nungwane Dam is the smallest of the three dams and supplies water to the Amanzimtoti WTW on the south coast. While the Nungwane Dam has a sediment trapping efficiency of 99.1%, the dam is situated on a tributary of the Lovu River and thus does not trap the main river sediment load moving to the coast. Details relating to the three water supply dams located within the EMA Water Supply Dam Inanda Hazelmere Nungwane River uMngeni uMdloti Nungwane Year completed 1989 1977 1978 Associated WTW Wiggins Hazelmere Amanzimtoti Total catchment area (km2) 1547 377 58 Sediment yield in catchment (t/km /y) 2 314 381* 340* Initial dam capacity (million m3) 259 23.9 2.4 Current dam capacity (million m ) 2 3 252 17.9 2.24 Coarse sediment trapping efficiency* (%) 99.6 98.7 99.1 Capital Replacement Cost (2008, R million) 769.4 473.4 ? Source:  * CSIR 2008; (1) (2) Umgeni water website Figure 4.5 Inanda, Hazelmere and Nungwane Dams are located within the EMA (Source: Umgeni Water) 1 Note: Two other dams, the Shongweni Dam (Mlazi River) and Dudley Pringle Dam (Wewe River) were not included in this analysis for the following reasons: (1) the Shongweni Dam (capacity of 3.8 million m3) has been decommissioned as far as water supply is concerned. It is used only for recreational purposes; and (2) the Dudley Pringle Dam (capacity of 2.3 million m3) is owned by Tongaat Hulett and supplies water to Maidstone Sugar Mill. It is also used for irrigation purposes and the eThekwini Municipality have the rights to extract up to 15 Ml/day to supply Tongaat WTW. However, the dam is located on the northern border of the EMA with almost all of the catchment falling outside of the EMA. Page 68 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY A model was set up in PC-SWMM software to simulate The costs of sedimentation of dams could be estimated the hydrology and sediment transport for the catchment either as the replacement cost of lost storage capacity area of the whole municipality, and to estimate the (e.g. through raising the dam wall or constructing a extent to which natural vegetation prevented sediments substitute dam at a new site to make up the reduction from entering the main reservoirs and estuaries, by in capacity), or as the cost of dredging to remove the comparing rates of sedimentation with and without accumulated sediment. For dams, we estimated the natural vegetation (Box 4.4). value of the service in terms of dam construction costs (being the lower cost option), based on modelled cumulative storage loss over 20 years (Box 4.5). For Durban Harbour, we used the avoided dredging cost, Box 4.4 Summary of sediment modelling under the assumption that these would be lower than approach (see Appendix 4 for details). the damage costs avoided as a result of ecosystem degradation (Box 4.6). For the rest of the estuaries, Annual sediment loads were estimated by the impacts of changes in TSS loads were considered simulating the total suspended sediment load together with water quality (see Section 4.7). (TSS) using the PC-SWMM model described in Appendix 2. TSS loads were simulated for one The loss of vegetation cover from their catchments year from August 2013 to August 2014. The would lead to a significant increase in the rate of pollutant washoff from a given landuse during sedimentation of all three of the main water supply periods of wet weather was characterized in dams (Figure 4.6). The greatest impacts would be felt the model by using a user defined Event Mean by Hazelmere Dam, which has a large area of natural Concentration (EMC). The EMCs were derived vegetation in its catchment. The replacement of these from literature and applied to the different natural areas with human settlements could result in landuse categories. Model subcatchment a nine-fold increase in sediments entering the dam parameters were derived by area weighting (Figure 4.6, Table 4.6). The total annual replacement cost the various land use parcels within each associated with the loss in dam capacity as a result of subcatchment. The modelled runoff flows sedimentation is estimated to be between R1.1 million were coupled with EMC values to estimate the and R2.9 million (average R1.97 million; Table 4.6). Over concentration and total load of TSS at different a 20 year period from 2015 and using a discount rate points in the study area. of 6% this equates to a net present value of between R12.25 million and R33.03 milllion (average R22.77 The percentage contribution of bed load and million). suspended sediments to the total sediment load is different for each catchment. For this study suspended sediment loads were multiplied by a factor of 1.25 in order to account for bed load. This is the factor generally applied in South Africa (Msadala et al. 2010, after Rooseboom 1992). By comparing the modelled sediment outputs per catchment under current land cover versus fully transformed land use, it was possible to estimate the difference made by natural vegetation to the sediment loads transported to dams and estuaries. Two scenarios were used to estimate the sediment retention service provided by existing natural vegetation in the EMA. The first scenario involved removing the trapping capacity of natural vegetation (a hypothetical construct) and the second involved replacing natural vegetation with dense rural settlement as a likely alternative land use. These two scenarios provided lower and upper bound estimates of the magnitude of the service. A detailed description of the model is provided in Appendix 4. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 69 Box 4.5 Estimation of costs avoided for dams The sedimentation rate over a dam’s lifetime can be projected from 50 = � a measurement at some point in time, using the Vt/V50 relationship �0.376ln( )� developed by Rooseboom (1975, see also 3.5 WRC 1992): Where V50 = sediment volume after 50 years, Vt = sediment volume measured after t yea 50 years, Where V50 = sediment volume after years. Vt = sediment volume Sedimentation measured of the three damsafter wast years, and t using = modelled data on average sedi period in years. Sedimentation of the was modelled three damsstarting catchment, using capacity anddata on average current (2015) sediment capacityyields (Box in 4.3). The percentage current (2015) was the catchment, starting capacity andsedimentation capacity taken(Box from4.3). the The percentage sediment modelling change in rate study. The of actual rate of sedimen sedimentation was taken from the sediment modelling the sediment study. The modelling actual study was rate of sedimentation than the rate estimated lower estimated based on the above lower than to in the sediment modelling study wasconsidered the berate theestimated based less reliable ofabove on the estimate actualmethod, rates. but was considered to be the less reliable estimate of actual rates. 3 The volume of sediment was estimated from mass using a density of 1.35 t/m (Rooseboo The volume of sediment was estimated from mass using a density of 1.35 t/m3 (Rooseboom 1992, Haarhoff Cassa 2009). The loss of capacity over a 20 year period was estimated, and costed in t & Cassa 2009). The loss of capacity over a 20 year period was estimated, and costed in terms of the cost of replacing the equivalent capacity, based on the capital replacement costs of the dams. replacing the equivalent capacity, based on the capital replacement costs of the dams. Box 4.6. Estimation of costs avoided for dredging of Durban harbour The avoided sedimentation of Durban harbour was taken as the difference in annual sedim Box 4.6 Estimation of costs avoided for dredging of Durban harbour the modelled scenarios. The avoided costs were estimated using dredging data provid Durban Harbour for the period 1 April 2015 – 31 March 2016. Maintenance dredging invo The avoided sedimentation of Durban harbour was sediments fromtaken as thebasins channels, difference and in annual berths sediment within input the harbour. Dredging of channels an between the modelled scenarios. The avoided 3 costs were estimated using dredging data provided3 by per m on average and dredging of berths costs R636 per m . A total of just under 182 000 Transnet for Durban Harbour for the period 1 April 2015 – 31 March 2016. Maintenance dredging involves dredged from the harbour over the period 2015/16, at an overall average cost of R229 per the removal of sediments from channels, basins and berths within the harbour. Dredging of channels and However, most of the sediment removed from the harbour through maintenance dredging basins costs R85 per m3 on average and dredging of berths costs R636 per m3. A total of just under 182 000 river inputs but is from exisiting estuarine sediments that are shifted by the movement of m3 of sediment was dredged from the harbour over the period 2015/16, at an overall average cost of R229 the harbour (Transnet NPA). The “silt canal” at the top end of the estuary is not dredged per m3 (Transnet NPA). However, most of the sediment removed from the harbour through maintenance and it is estimated that less than 5% of the annual volume of dredged sediment is of dredging is not derived from river inputs but is from exisiting estuarine sediments that are shifted by the Greyling, Transnet NPA pers. comm.). This corresponds to the modelled status quo TSS loa movement of large ships through the harbour (Transnet NPA). The “silt canal” at the top end of the estuary 2.6% of the annual volume of dredged sediment. is not dredged on a regular basis and it is estimated that less than 5% of the annual volume of dredged sediment is of fluvial origin (Clive Greyling, Transnet NPA pers. comm.). This corresponds to the modelled The annual TSS loads generated under current conditions were compared with what th status quo TSS load which represents 2.6% of the annual volume of dredged sediment. natural systems within the catchment were removed or transformed to a different land 3 The annual TSS loads generated under dense settlement. current conditionsThe annual were TSS loads compared with were whatconverted they wouldtobe volume if the (1.35 t/m ) and multi natural systems within the catchment average dredging were removed cost, or with the to transformed difference in costs a different between land use, the such as scenarios representing th rural dredging of Durban harbour. dense settlement. The annual TSS loads were converted to volume (1.35 t/m3) and multiplied by the overall average dredging cost, with the difference in costs between the scenarios representing the costs avoided for dredging of Durban harbour. It was estimated that the annual TSS load into Durban Harbour increased by 195% current load when vegetation in the Umhlatuzana-Umbilo catchment was remov It was estimated that the annual TSS load into Durban Harbour increased by 195% compared to the current load when vegetation in the Umhlatuzana-Umbilo by 206% was and catchment when the vegetation removed (lower bound)inand thebycatchment 206% when the was replaced with den vegetation in the catchment was replaced with dense (upper settlement bound, rural informal Table 4.7). settlement Thisbound, (upper resultsTable in dredging costs 4.7). This avoided of betw results in dredging costs avoided of between R1.03 million and R1.15 million (average R1.1 million) and R1.15 million (average R1.1 million) per year. per year. The overall net present value of the sediment retention service was estimated to be in the order of R35.17 million. The portion of the value pertaining to each dam and the harbour was mapped to the green open space areas within the relevant catchments using an estimate of the relative contribution of different land parcels to sediment retention modelled using InVEST modelling software (Glenday 2012). This provided the relative sediment retention score given to each grid cell within the EMA based on rainfall, soil and landcover. The value per pixel was then determined for all green open space areas (based on the latest GIS data of D’MOSS), total sediment retention value for each catchment, the retention score given to the pixel and the sum of all pixel retention scores. The value per pixel was then divided by the area of the pixel to arrive at a sediment retention value per hectare estimate for all green open space areas in each catchment. Page 70 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 80 Figure 4.6 The current, lower bound and upper bound cumulative loss of capacity (Mm3) over 50 years for (a.) Inanda Dam, (b.) Hazelmere Dam and (c.) Nungwane Dam Table 4.6 Lower and upper bound flood attenuation values (R million) for each quaternary in the EMA. A zero value indicates that there is no stormwater infrastructure within that catchment. Dam Replacement cost NPV 20y from 2015, 6% (R millions per year) (R millions) without vegetation with settlement without vegetation with settlement (lower bound) (upper bound) (lower bound) (upper bound) Inanda 0.79 1.42 9.09 16.31 Hazelmere 0.26 1.42 3.01 16.33 Nungwane 0.01 0.03 0.15 0.38 TOTAL 1.07 2.88 12.25 33.02 Table 4.7 Estimated maintenance dredging costs avoided due to the sediment retention function of natural vegetation in the Durban Harbour catchment. Durban Harbour Average Change in % change in Annual dredging NPV dredging cost annual TSS annual costs avoided (R/m3) load (m3) TSS load (R millions) Lower bound 229 4 511 195% 1.033 11.85 Upper bound 229 5 029 206% 1.152 13.21 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 71 Figure 4.7 Sediment retention value associated with natural systems (D’MOSS) (NPV, R/ha) in the EMA Page 72 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY 4.7 Water quality amelioration The introduction of nutrients into the landscape (anthropogenic eutrophication) can lead to reduced water quality and the eutrophication of freshwater and marine ecosystems. This reduces their capacity to supply ecosystem services as well as increasing water treatment costs (Graham 2004, Rangeti 2014). Nutrient enrichment of water resources often occurs in highly populated urban areas where agricultural activities, use of fertilizers, untreated sewage, organic wastes and water-borne sewage systems lead to increased nutrient loads in surface waters. Where excessive nutrients end up in water supply reservoirs, the resultant algal growth leads to increased water treatment costs. A number of studies have analysed the effect of water quality variables on water treatment costs in South Africa (e.g. Dennison & Lynne 1997, Graham 2004, Friedrich et al. 2009, Gebremedhin 2009, Graham et al. 2012, Rangeti 2014) as well as globally (e.g. Dearmont et al. 1998, Nkonya et al. 2008, Telles et al. 2013, McDonald & Shemie 2014, TNC 2015). In the study area, Graham (2004) showed that the costs of water treatment for four WTWs operated by Umgeni Water increased as a result of increases in turbidity, aluminium, iron, suspended solids, nitrates, total organic carbon (TOC), total dissolved solids, silicon, coliform numbers, potassium and algae in their respective reservoirs. Rangeti (2014) also found that algae, turbidity and TOC in the Inanda and Nagle Dams were the key drivers in chemical dosage and chemical costs in treating potable water at Durban Heights and Wiggins WTWs. Figure 4.8 Schematic diagramme of the consequences of anthropogenic effect on water quality and their amelioration by natual However, none of these studies or studies from further systems: (Source: Author) afield have developed models to relate water treatment costs to the pollutant loads entering the reservoirs (which can be related to land use and anthropogenic inputs) and only one (Vincent et al. 2015) has related treatment costs removal is mainly through denitrification and also to directly to the state of the catchment. These kinds of some extent by plant uptake (Hill 2000). Nutrients relationships need to be estimated in order to estimate that are introduced in dissolved form can be taken up the value of water quality amelioration services. directly by plants and incorporated into plant tissue as they grow. Most of the phosphorous that is transported The anthropogenic deterioration of water quality by flows is attached to sediment and settles out, where also has a negative impact on the health of aquatic it can remain inactive (Brinson 2000). However, if ecosystems and their capacity to deliver ecosystem sediments that settle in aquatic systems are stirred services. For example, increases in nutrient and up again then some of this phosphorous can go back suspended sediment loads entering estuaries lead into solution and become available for use by plants. to impacts on fish stocks, estuarine fishery values Phosphorus is usually naturally limiting (in short supply) and the export of fish to marine fisheries. Thus water in freshwater systems, whereas nitrogen is typically quality amerlioration by natural areas can also play an limiting in marine systems and estuaries. Therefore plant important role as a supporting service. uptake uses up different nutrients in different systems. The uptake of nutrients will continue as long as there is Water quality amelioration is the removal of some of room for further plant growth (in terms of space, oxygen the excess nutrients and sediments that are generated or plant size limits), after which the system will reach through anthropogenic processes in the landscape and some kind of equilibrium in which the uptake is balanced transported in surface water runoff and/or groundwater by the senescence, death and rotting of plant material systems. There are a number of different processes which reintroduces nutrients into the water column through which natural systems remove pollutants from (remineralisation). At this point there would be no surface and sub-surface flows (Figure 4.9). Nitrogen further net uptake of nutrients by the ecosystem unless A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 73 Figure 4.9 Summary of water quality amelioration services by natural systems (Source: Turpie 2015) nutrients are being exported out of the system (e.g. by Accurate valuation of this service requires a much better harvesting plants or dredging and removal of sediments), understanding of the removal rates of natural systems or unless there is a natural process of peat formation. or better evidence of its impact on the downstream costs or benefits. This usually requires the use of both Wetlands are generally regarded as the most efficient biophysical as well as economic analytical and modelling natural system for removing pollutants, partly because tools (Keeler et al. 2012). they have much greater capacity for traping sediments, but forests and other terrestrial vegetation types In this study, the impacts of natural open space areas also have the capacity for water quality amelioration on water quality were estimated using a hydrological (Asmussen et al. 1979). Terrestrial systems have been model which was set up to estimate the production and shown to improve water quality at a landscape scale transport of total suspended solids (TSS), phosphorus (Dixon & Rowlands 2007), and it has also been shown (P) and total inorganic nitrogen (TIN) in the catchments that natural vegetation along streams acts as an of the EMA. The model was set up to estimate the important buffer between agricultural landscapes and changes in loads entering dams and estuaries if the river systems, removing a high percentage of sediments retention capacity of natural vegetation was eliminated, and nutrients from surface and subsurface flows (Mayer or if natural vegetation was replaced with dense human et al. 2007, Liu et al. 2008, Yuan et al. 2009, Zhang et al. settlement (see Box 4.7). These two scenarios provided 2010, Weller et al. 2011, Sweeney & Newbold 2014). the upper and lower bound of the service provided by natural open space areas in physical terms. The value of Water quality amelioration services only have value the service was then estimated in terms of the avoided downstream of where wastes are generated and costs to water treatment works and the avoided loss of wherever downstream water and ecosystem services estuary value as a result of the eutrophication of these occur that would be impacted by a loss of water quality. systems. Water quality amelioration is thus largely a supporting service of terrestrial and aquatic ecosystems that Water treatment cost savings were estimated for the influences the output of final goods and services from three water treatment plants based on data supplied by downstream aquatic ecosystems as well as reducing Umgeni Water and the eThekwini Water and Sanitation the potential costs of water storage and treatment. Department on the Durban Heights WTW and Wiggins Studies that put a monetary value on this service WTW and the Nagle and Inanda Dams. Two separate frequently make the assumption that natural systems models were developed to relate phosphorous loads such as wetlands remove the total pollution input entering the water supply dams to the water treatment load (e.g. Emerton et al. 1998, Qian & Linfei 2012). costs at each of the WTWs (see Box 4.8). While there Page 74 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY was detailed data available for the uMngeni River system and associated dams and water treatment plants, data for the Hazelmere Dam (and associated Hazelmere WTW) and Nungwane Dam (and associated aManzimtoti WTW) were not available. The model developed for Nagle Dam and Durban Heights WTW was therefore applied to Hazelmere and Nungwane. Nagle Dam is much smaller in size than Inanda Dam, with the dimensions of the reservoir being more comparable to Hazelmere and Nungwane Dams. The technology and chemicals used to treat water at Durban Heights WTW and at Hazelmere and aManzimtoti WTW are also comparable. The modelled impact of the removal/ replacement of natural habitats on phosphorus loads and water treatment costs was used to estimate the water quality amerlioration value of natural open space areas in the catchment areas of each dam. Box 4.7 Summary of the hydrological - nutrient modelling methods (detail in Appendix 4) The impacts of natural open space areas on water quality were estimated using a hydrological model set up in PC-SWMM software for the catchment area of the whole municipality. The model, described in Appendix 4, was set up to estimate the average annual loads of nutrients entering raw water reservoirs of water treatment plants within the EMA. The EMC for nutrients, TIN and P, were derived from literature and applied to the different landuse categories within the study area. Simulated catchment runoff was coupled with the EMC values to estimate total annual nutrient loads. TIN and P loads were simulated for one year from August 2013 to August 2014 (total annual precipitation of 572 mm for the Durban city area, lower than MAP). The annual phosphorous loads generated under current conditions were compared with what they would be if the natural systems were removed or transformed to a different land use, such as rural dense settlement. This provides an indication of the impacts of loss of natural areas on nutrient loads entering surface waters and the percentage change can be construed as an estimation of the water quality amelioration benefit obtained by retaining the natural systems. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 75 retaining the natural systems. Box 4.8. Summary of method used to estimate avoided water treatment costs (detail in Appendix 5) A series of water treatment cost models were used to estimate the water quality amelioration value associated with natural open space in the catchments of water supply dams in the EMA. Water treatment cost quality data data and waterBox 4.8 were provided Summary of method for a five used to year period estimate from 1water avoided July 2010 – 30 June treatment 2015 costs for in (detail the two Appendix 5) largest WTW in the EMA; Durban Heights and Wiggins which receive water from Nagle Dam and Inanda Dam on the uMngeni River. Together, these two treatment works supply more than 60% of the potable water in A series of water treatment cost models were used to estimate the water quality amelioration value the EMA. Treatment data natural cost with associated and water quality open space data were in the analysed of catchments and it was water expected supply damsthat higher in the EMA. nutrient Water treatment loads, in particular phosphorous, in the uMngeni River would result in increased water treatment costs cost data and water quality data were provided for a five year period from 1 July 2010 – 30 June as a 2015 for result of increased algal growth and associated changes in water colour and odour. the two largest WTW in the EMA; Durban Heights and Wiggins which receive water from Nagle Dam and Inanda Dam on the uMngeni River. Together, these two treatment works supply more than 60% of the potable water in the EMA. Treatment cost data and water quality data were analysed and it was expected that higher nutrient loads, in particular phosphorous, in the uMngeni River would result in increased water treatment costs as a result of increased algal growth and associated changes in water colour and odour. Figure 4.10. Schematic summary of the linkages from phosphorous loads in water supply dams to increased Figure 4.10 Schematic summary of the linkages from phosphorous loads in water supply dams to increased water treatment costs as a water treatment costs result as a result of deteriorating of deteriorating water quality. water quality. The first set of regression models investigated the relationship between treatment costs 3 (R/m3 of treated regression The first set ofwater) range ofinvestigated and a models water quality the relationship variables in the between treatment water being abstractedcosts (R/m from of treated the supply dams. The water) and a range waterthese of from results quality variables models in thethat indicated water thebeing abstracted increase from the in suspended supply solids and dams. The results algal blooms, especially from these models duringindicated the summer the increase that rainfall months infrom suspended Novembersolids and algal through blooms, to March, are especially during for the the main reason heightened summer rainfall months from November through to March, are the main reason for heightened treatment treatment costs. The rising costs associated with these factors are a result of increased usage of coagulants and costs costs. The rising associated disinfectants neededwithtothese removefactors are a sediments suspended and algae, usage result of increased of coagulants and associated odour and and colour disinfectants needed to remove suspended sediments and algae, and associated odour issues, during the treatment process. All regression models took the following form: and colour issues, during the treatment process. All regression models took the following form: = ( , , ) 3 where is the water treatment cost associated with treating 1m of water and , , are the nutrient, where TCw is the water treatment cost associated with treating 1m3 of water and xn,xc,xa are the nutrient, chemical and algal water quality paramters related to TCw. 86 The next set of models linked treatment costs to phosphorous loads and other water quality variables such as coliforms, colour, temperature and alkalinity in the water entering the water supply dams. The results revealed that the phosphorous loads in the uMngeni River entering Nagle Dam and Inanda Dam were positively and significantly correlated with water treatment costs at both Durban Heights WTW and Wiggins WTW (see Table A5. 5 and Table A5. 7 in Appendix 5). The model results had a reasonable fit to the actual treatment cost data supplied for Durban Heights WTW and Wiggins WTW. Page 76 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure 4.11 Average phosphorous loads (kg) in the uMngeni River above Nagle Dam (top) and Inanda Dam (bottom) and corresponding water treatment costs (R/ML) at Durban Heights WTW and Wiggins WTW. These water treatment cost models were used to predict A third model was developed to estimate the avoided the outcome of catchment land-use changes, such as loss of estuary values as a result of water quality the impact that natural vegetation on reducing nutrient amelioration by natural systems (see Box 4.9). The runoff into surface waters. Using a simple scenario model was developed based on statistical relationships approach, comparisons between the status quo and between nutrient and TSS loads, water quality scores modelled outputs for phosphorous loads into water and the status of fish communities described by experts supply dams can be related to the treatment costs in these fields for the EMA’s 16 estuaries, and the using the above models and overall cost savings can be nursery values described in Section 4.3. The modelled estimated. impact of the removal/replacement of natural habitats on nitrogen and TSS loads, estuary water quality, fish The detailed water treatment cost analysis, methodology populations and nursery value was used to estimate the and model results can be found in Appendix 5. water quality amerlioration value of natural open space areas in the catchment areas of each estuary. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 77 Box 4.9 Relationships between increased N and TSS loads, estuary water quality and the status of estuarine fish communities Anthopogenic increases in nitrogen and TSS have a negative impact on estuarine health and the capacity of estuaries to deliver services such as recreational fishing value and nursery value (Figure 4.12). Estimation of the water quality amelioration service on estuary values required the development of a model to estimate the impacts of changes in N and TSS loads on water quality, and the impacts of changes of water quality on fish populations. Figure 4.12 Schematic summary of the linkages from water quality parameters to estuary ecosystem services for a case of deteriorating water quality. The relationships between water quality and fish abundance have been well studied in KwaZulu-Natal (Cyrus & Blaber 1987, 1988). Expert understanding of these and other relationships is used in the quantification of estuarine responses to changes in in the quantity and quality of freshwater inflows to estuaries. A substantial amount of work has also been carried out by groups of estuarine specialist scientists to describe the present status of estuaries throughout South Africa, following standardised methods of describing estuarine health developed for the setting of environmental flow requirements as well as for estuary management more generally (Turpie et al. 1999, 2012). These require the description and scoring of all the abiotic and biotic components of estuaries, including water quality variables and fish communities. These scores have been collated and summarised for the most recent National Biodiversity Asssement of estuaries (van Niekerk & Turpie 2012). Analysis of the health scores for the 16 estuaries in the study area revealed a strong linear relationship between a sub-element of the overall water quality score (based on nutrients, TSS, oxygen and heavy metals, but not salinity), hereafter referred to as the Water Quality Score (excluding salinity), and the Fish Score (Figure 4.13). The effect of nutrient and TSS loads on the Water Quality Score was then determined based on the relationship between modelled annual loads for the EMA estuaries and the Water Quality Score. The annual loads were first standardised to account for differences in catchment size, characteristics and rainfall by expressing them in terms of natural Mean Annual Runoff. After exclusion of two outliers, there was a significant negative logarithmic relationship between the standardised N, P and TSS loads and the Water Quality Score (Figure 4.14). These relationships were used to estimate changes in the Water Quality Score as a result of loss of the natural vegetation in the catchment, and to estimate the consequent impacts on fish populations and values. The modelled loads under the status quo and the two alternative scenarios (removing the ecosystem service) were then used to generate modelled water quality scores. The percentage difference in these scores was used to adjust the actual water quality scores. The latter were then used to estimate diffrences in the fish scores. Again, these differences were used to adjust the actual fish scores. Finally, the proportional change in fish scores was used to estimate changes in the fishery and nursery values of each estuary, as a measure of the damage costs avoided. Page 78 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure 4.13 Relationship between water quality and fish scores for the 16 estuaries of the eThekwini municipality. Figure 4.14 Relationship between modelled annual loads of TIN and TSS and the Water Quality Scores for estuaries of the eThekwini municipality. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 79 Table 4.8 Annual water treatment cost savings associated with the three water supply dams in the EMA Dam Without vegetation (min) With settlement (max) Increase in Total annual water Increase in Total annual water treatment cost treatment cost treatment cost treatment cost (R per ML) savings (R millions) (R per ML) savings (R millions) Inanda 7.45 0.734 44.96 4.43 Hazelmere 8.12 0.154 193.80 3.68 Nungwane 8.77 0.067 74.61 0.57 TOTAL 0.955 8.68 Phosphorous loads entering the water supply dams as a In addition, it was estimated that the maintaining the result of the loss of natural vegetation were estimated to existing areas of natural vegetation avoids potential increase by 193%-319%x for Inanda Dam, 193%-968% for losses in estuary fishery and nursery values of between Hazelmere and 200%-509% for Nungwane Dam, with the 2 and 48% (Figure 4.15). The total avoided loss as a lower values being an estimate of actual current levels of result of the water quality amelioration function of removal and the higher values incorporating the potential natural vegetation in the estuary catchments was additional inputs that would occur under changed land estimated to be in the order of R2.19 to R3.71 million use, which is effectively the cost avoided by conserving per annum. Overall, water quality amelioration services the current land use. Using these estimates and the water were estimated to be worth between R3.15 and R12.39 treatment cost model for Wiggins WTW and Durban million per annum, the average of which is R7.77 million. Heights WTW it was estimated that the cost saving to This translates to a NPV of R89.3 million. water treatment works is between R1 million and R8.7 million per annum (Table 4.8). The Changes were greatest for Hazelmere dam, which has a large area of natural vegetation in its catchment. Figure 4.15 Estimated avoided losses of estuary ecosystem services due to the water quality amelioration function of natural vegetation in the estuary catchments. Values derived from the above analyses were mapped total phosphorous and nitrogen retention value for to the green open space areas within each catchment each catchment, the retention score given to the pixel using an estimate of the relative contribution of different and the sum of all pixel retention scores. The value per land parcels to phosphorous and nitrogen retention pixel was then divided by the area of the pixel to arrive modelled using InVEST modelling software (Glenday at a phosphorous and nitrogen retention value per 2012). This provided the relative phosphorous and hectare estimate for all green open space areas in each nitrogen retention score given to each grid cell within catchment. The values for phosphorous and nitrogen the EMA based on rainfall, soil and landcover. The retention were then summed to get a total value per value per pixel was then determined for all green open hectare for nutrient retention (Figure 4.16). space areas (based on the latest GIS data of D’MOSS), Page 80 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure 4.16 Water quality amelioration value associated with natural systems (D’MOSS) (NPV, R/ha) in the EMA A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 81 This page intentionally blank. Page 82 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY V. AMENITY VALUE 5.1 Introduction in an area where they will have access to or views of different types of amenities. Thus one can determine Durban has been rated as the city with the highest amenity value by using revealed preference methods, quality of living in South Africa (based on the Mercer in which actual data from related markets are analysed rankings of 230 cities in the world). Many factors in order to estimate the contribution made by a feature contribute to quality of living, including environmental of interest, in this case green open space. It is fair factors such as cleanliness and air quality, and the to assume that the two values derived are from two amenity value provided by natural and semi-natural separate user groups and can be added together. These systems. two types of value are analysed separately below. The amenity value of urban green open space areas is derived from their individual combinations of natural attributes such as size, beauty and rarity, and man- 5.2 Amenity value to property owners made enhancements such as pathways, lawns and The value that residents place on open space is security. These attributes determine the extent to which reflected, to an extent, in private property and each area is suitable or attractive for recreational use, real estate markets. When buying a home, certain religious use or spiritual fulfilment. Their value, or actual preferences for different characteristics are revealed contribution to human welfare also ultimately depends through the amount that each homebuyer is willing to on factors that influence demand for these services, pay for the property, with homes that have a higher such as the number and income levels of people living in number of desirable characteristics usually selling the vicinity, as well as by people living elsewhere. Users for a higher price. Property attributes include the of urban parks have described them as rejuvenating, physical characteristics of the home such as size of relaxing, break from daily routine and spiritual the living area, number of bathrooms and condition, reconnection with the natural world (Chiesura 2004). neighbourhood characteristics such as access to Other studies have shown that public parks reduce amenities, and environmental characteristics such stress (Ulrich & Addoms 1981) and provide a sense of as scenic views and the amount of green open space peacefulness (Kaplan 1983). Therefore, public parks surrounding a property. Therefore if residents do value benefit emotional and psychological health and are a key open space then it would be expected that these values component to sustainable development (Chiesura 2004). should be revealed in property prices. In fact, the lack of green open spaces is the primary reason why people left the city of Leuven, Belgium, and A hedonic property value approach was used to estimate sought properties on the urban fringe (Van Herzele & the value associated with different types of green Wiedemann 2003). open space within the EMA. A total of 16 149 property transactions, provided by the eThekwini Municipality Because of the intangible nature of these values, the Real Estate Department, were analysed over a two year welfare gains generated by the supply of cultural period from January 2012 to October 2014. The dataset services are difficult to estimate and map (Milcu et included several characteristics of each property. Each al. 2013). In theory, the value of the cultural services property in the dataset had a unique PIN which allowed provided by existing green open space areas is what us to determine its location by matching each property people would demand in compensation for giving up sale with a property boundary (erf) in the eThekwini GIS the benefits they receive from those spaces. This can cadastral layer. Using GIS data, the areas of different be estimated through the use of stated preference types of natural open space such as forests, woodlands, methods, involving surveys of users that elicit their rivers and grassland as well as recreational green open willingness to pay to retain or willingness to accept space such as public parks and golf courses within compensation to forgo a certain state of the world. different distances of each property were determined. However, the inherent methodological biases of these The condition (good, intermediate, degraded) of the methods can be extremely challenging, and require natural open space was also considered as it was very comprehensive studies. No such studies have been assumed that degraded patches of natural vegetation carried out within the Durban area. may influence property price differently to patches in a good condition. Distances to the CBD, private schools Nevertheless, the welfare gains associated with the and the coast were also calculated. In addition, the amenity of green open space areas are reflected to characteristics of each neighbourhood (or suburb), such a large extent in two markets that are observable – as population density and average household income, property and tourism. Urban residents often pay a were collated from Census 2011 data. Average tree premium to live close to the areas they enjoy, or to have cover was also calculated for the streets in each suburb. a good view. Similarly, visitors pay to travel to and stay The property prices were then analysed in terms of the A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 83 property characteristics, neighbourhood characteristics informal settlement sites. As a result, degradation and access to green open space (Table 5.1) using and a general decrease in attractiveness of these areas ordinary least squares (OLS) regression. The analysis is occurs. Degraded open space patches are also often described in detail in Appendix 6). associated with crime, making them even less desirable in terms of their proximity to properties and their Table 5.1 Property, neighbourhood and green open space accessibility. Unexpectedly, river systems, regardless of characteristics included in the hedonic analysis their condition, had a negative impact on housing prices. While the percentage decrease on property value was Property Neighbourhood Green relatively small, it is still surprising that even when in characteristics characteristics open space a good condition the proximity to rivers is considered characteristics negative. This could be that people prefer to have views Total Living area Distance to CBD Amount of natural over river gorges, such as in the areas of Hillcrest and (m2) (km) vegetation (ha) Kloof, but consider rivers in close proximity to property View from Population Density Length of river/ as a negative influence as similarly to natural open space property (ppl/km2) stream (m) areas, they are known to attract informal settlements, Amount of are prone to flooding and can be considered breeding Modal household Security income (R) sugarcane grounds for unfavourable insects such as mosquitos. farmland (ha) Man-made green open space, such as golf courses and Distance to nearest Sub-place park areas both had a significant positive effect on house Condition greenness: tree prices. school (m) cover (%) Presence of a Amount of Amount of golf The total premium associated with natural open space commercial land in a good condition was 2% of overall property value, garage course area (ha) (ha) which amounted to R4.4 billion with an average value of Presence of a Amount of Amount of park R108 900 per ha (Figure 5.1). This is only part of the asset swimming pool industrial land (ha) area (ha) value of these areas, which also provide other ecosystem Distance to services. The highest values were situated within the Amount of major nearest coastline road network (ha) main urban core of the EMA (R1.4 million per ha in (km) Durban MP) and along the coastline where high quality Distance to nearest coastline (km) natural coastal forests are still intact (e.g. Umhlanga – R3.4 million per ha). The values were also higher in and around the suburbs of Hillcrest and Kloof (about R1 In order to spatially assign a value to the green open million per ha), both being affluent areas, much like the space in the EMA it was necessary to calculate the coastal suburbs of Umhlanga and Durban MP, whereas premium associated with natural open space and park they were lower in inland suburbs such as Cato Ridge land and aggregate the model results. This was achieved (R37 500 per ha) and Pinetown (R423 000 per ha). by assigning a census level sub-place and main-place to each property and the effect of open space on property The total premium associated with public parks was values was calculated using the estimated model approximately 6.4% of overall property value, amounting coefficients, which provided a percentage change in to a total of R13.8 billion with an average value of R14.7 property value given a unit change in the value of the million per ha (Figure 5.1). and an average value of open space variable under consideration. The aggregate R14.7 million per hectare. This is more than three times effect was then estimated by by applying the modelled the total value associated with natural open space in a regression results to the entire stock of residential good condition. Park value is highest in and around the houses within each sub-place of the EMA. The value (R/ urban core of Durban city but is also important in other ha) associated with each area of open space or park land densely populated areas such as Chatsworth, Phoenix was calculated based on these total premiums and the and Pinetown. With an average population density in total amount of green open space present within each Chatsworth of 5500 people per km2 and in Phoenix 6400 main-place census unit. people per km2, both higher than the EMA average, it is clear why residents in these areas may value public open The results revealed that the type of open space and the space more than residents living in less dense areas condition of open space have very important influences with larger-sized properties. Park areas offer more on the amenity value associated with these areas, with recreational opportunities when compared to natural natural open space areas in a good condition attracting open space areas and it is also thought that safety significant and positive price premiums and those in a and security play an important role in where residents degraded condition attracting negative price premiums. choose to enjoy green open space in the city, with public Natural open space areas that are not formally protected park areas often more safe than natural dense vegetated are often targeted by people moving into the city areas. with no place to live, often being transformed into Page 84 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure 5.1 Average value (R/ha) of (a) natural open space in a good condition and (b) parks within each main-place within the EMA. Note: the actual location and extent of open space areas within each EMA is not shown due to scale. The combined value of green open space areas in the EMA, both natural and man-made, was estimated to be R18.2 billion. Based on the these values, the presence of well managed green open space in the eThekwini Municipality contributes an estimated R356 million per annum in property tax revenues to the city. This would account for more than 5% of total property rates income to the municipality. These results provide a useful understanding for future development in Durban and highlight the importance of maintaining healthy natural open space areas. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 85 5.3 Tourism value The city of Durban is a leading tourism destination in D’MOSS area (eThekwini Municipality 2012). These South Africa. The year-round warm weather conditions range in size from small bird sanctuaries to large nature encourage leisure tourists to enjoy the many outdoor reserves, and they vary in terms of their visitor facilities. tourist attractions such as the extensive beaches and Two of the most popular are the Durban Botanical beachfront promenade, the uShaka Marine World and Gardens and the 584 ha Krantzkloof Nature Reserve the SkyCar and bungee swing at the Moses Mabida situated on the coastal escarpment between Pinetown World Cup Stadium, as well as sporting events. The and Hillcrest. Krantzkloof features scenic river gorges, beach and marine/coastal environment is the core excellent views, indigenous coastal forest and grassland leisure tourism experience offered in Durban (eThekwini and an abundance of bird life, and has well-utilised Municipality 2014b). According to recent surveys, paths for walking and trail running. There are also going to the beach was the main activity for 50% of nature reserves associated with the water supply dams international tourists and 83% of domestic tourists to (Shongweni, Hazelmere and Inanda) within the EMA KwaZulu-Natal (SAT 2014). and these are managed by Msinsi Resorts and Game Reserves. In 2012, some 15.5 million people visited the city, resulting in a total expenditure of about R5.7 billion Determining the tourism value of particular areas such (eThekwini Municipality 2014b). Some 15% of these as parks or nature reserves usually involves on-site visitors are foreign, and 57% of all visitors are holiday surveys of tourists and the collection of user statistics visitors (eThekwini Municipality 2014b). Based on data in to determine expenditure on visiting these areas, and WTTC on tourism growth and multipliers (WTTC 2015), using revealed preference methods such as travel total expenditure by domestic and foreign tourists was cost analysis in order to determine their consumers’ estimated to be R6.718 billion in 2014. This expenditure surplus, or the net welfare gains resulting from the would have resulted in a total GDP contribution of use of these areas (Willis & Benson 1989, Shrestha et R9.75 billion. Of this, leisure tourists were estimated to al. 2007, Wood et al. 2013). These kinds of methods contribute som R5.56 billion (Table 5.2). are impractical for estimation of the value of multiple sites across an extensive area. Most efforts to quantify Table 5.2 Estimated expenditure and valued added by tourists and map cultural services such as recreational value, visiting Durban in 2014 (R billion). have used the number of visitors to an area as a proxy (Hill & Courtney 2006). Other proxies include number Expenditure Direct Total of tourist attractions, tourist expenditure, sightings (R billion) Value Value of rare species, accessibility to natural areas and days Added Added spent fishing (Raudsepp-Hearne et al. 2010; Chan et (R billion) (R billion) al. 2011; O’Farrell et al. 2011). More comprehensive All tourists 6.718 3.138 9.753 studies typically rely on household or off-site user Leisure surveys (e.g. from a database of people with fishing 3.829 1.788 5.559 licences) or on-site choice decisions to estimate values tourists using a random utility modelling framework. However, these methods have only really been applied to the use There are no survey data on the extent to which visitors of local or regional amenities by residents of a city or are attracted by or make use of the natural assets within region. Determining the contribution of multiple natural Durban, however. While it is acknowledged that many amenties to tourism value might involve broader-based of the scenic and nature tourism attractions are not tourism surveys with dedicated questions on the use of well known or well developed for tourism (eThekwini these amenities. For example, Tyrvӓinen & Vӓӓnӓnen Municipality 2014b), these areas do already make some (1998) used the Contingent Valuation Method (CVM) to contribution to the value of tourism in the EMA. For determine the benefits of urban forests and Fleischer & example, the uMngeni River plays host to the annual Tsur (2000) used a combination of these two methods Dusi Canoe Marathon, and the Inanda Dam is used for to establish the recreational value of the agricultural fishing competitions, endurance events and a number of landscape. Where many such studies have taken place, other watersports. it is possible to use the findings to map recreational value using extrapolation techniques. For example, Nature-based activities such as walking and bird Sen et al. (2011) did an economic assessment of the watching are also important in certain areas, such as on recreational value of ecosystems in Great Britain using estuaries and wetlands and in nature reserves within a recreation valuation model which they developed, the EMA, and undoubtedly add value to the tourism using recreational survey data as well as land use and experience. There are 21 Municipal Parks and Nature population data. However, these methods generally Reserves within the EMA and six Ezemvelo KwaZulu- require a large amount of data which is expensive and Natal Nature Reserves, making up about 14% of the time consuming to collect especially at large geographic Page 86 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY scales. In Durban, none of the above such surveys have used the number of individuals (expressed per 1 km2) been carried out, and user statistics are only collected at who uploaded photographs onto Google Earth (via a handful of paying nature reserves. Panoramio) as a measure of recreation rather than the total number of photographs. They made the argument The recent emergence of various social media tools that this method was more appropriate when analysing provides an alternative to assess how people respond hotspots as the total number of photographs in an area to nature and open space areas (Wood et al. 2013). reflects the level of activity of individual photographers One such way of doing this is to analyse the pattern of rather than the overall value placed on a site by visitors. georeferenced photographs taken by the public and Turpie et al. (2014) used a kriging method to analyse uploaded onto the Internet. For example, Panoramio patterns of Panoramio photo uploads to map tourism hosts photographs from all over the globe, focusing on value in Zambia. Alfaro (2015) used photographs taken images of landscapes, natural features and animals in from Panoramio and Flickr to estimate the distribution their natural environment. Images that focus on people, of recreational services in Nebraska. By determining the interiors, paintings, logos and events are excluded from location of clusters of photographs and comparing these the website (Panoramio 2015). Geo-tagged imagery can clusters to population density, the researcher was able provide information about the places depicted in the to identify new areas of high recreational value. photographs, as well as the interests, behaviours and mobility of the people who took them (Andrienko et al. In this study, photograph data were extracted from 2009). Panoramio (wwww.panoramio.com) for the study area using a grid of 0.025 degrees (roughly 2.75 x 2.4 The number of people using the internet has increased km) which was created in ArcGIS® 10.1 (Esri® 2012). to nearly 100% on a global scale (International The API programming command was set to count all Telecommunication Union 2012). Several researchers photographs, including duplicates of the same subject have begun to use social media data for mapping of by different photographers. The size of the grid cell was recreational ecosystem services, and to determine the selected in order to achieve a sufficiently fine resolution appeal of a particular natural asset to both tourists and without being too small, given that the subject of locals (Alfaro 2015, Kachkaev & Wood 2013, Howarth photographs can be expected to be at a distance of up 2012). An advantage of this approach is that data are to 100 m or more from the camera, and given the overall often free and quicker to access than by traditional numbers of photographic uploads. means (e.g. surveys and hedonic pricing). Furthermore, it holds “big data” (Goes 2014), because an enormous A total of 10 016 photo uploads were obtained from amount of information is submitted by millions of users Panoramia within the EMA boundary. The highest worldwide. This “big data” is available through virtual concentration of photographic uploads were in areas sources which can be utilised by academics, consultants that were easily accessible, such as along the Durban and organisations that are looking to evaluate people’s and Umhlanga beachfront where there are a number preference for certain natural commodities (Alfaro, of hotels, guesthouses, walkways, restaurants and 2015). For instance, six billion images have been other tourist activities and facilities (Figure 5.2). A uploaded onto Flickr by Yahoo, to its public database large proportion of photographs were taken along the alone. Social media websites such as Flickr and National Route 2 (N2) highway that runs parallel to the Panoramio (by Google) can therefore provide a wealth coastline and the National Route 3 (N3) highway that of information through their numerous geo-tagged runs west towards Pietermaritzburg from the Durban photographs for example location, distribution and types city centre. High concentrations of photographs are also of interests, in addition to users’ demographics (Wood found in the high-income suburbs along the N3 route, et al. 2013). such as Kloof and Hillcrest. Protected areas also had large numbers of photographs associated with them. The use of geo-tagged photo data method is based on Other popular attractions noted during the photograph the premise that the numbers of photographs uploaded analysis were golf courses and sports grounds. A to these sites are correlated with the recreational value significant proportion of the photographs taken in of different areas (Casalegno et al. 2013). Several studies and around the Durban city centre are of cultural and have used this method. Turpie (2011) used photographs commercial content, such as photos of museums, from Panoramio.com to estimate landscape contribution historic statues and shopping malls. The high density to tourism value in an undeveloped area of Nambia. Li et of photographs along the length of the EMA coastline al. (2011) used parameters recorded by tourists’ digital highlights the importance of the marine environment in cameras and stored in Panoramio to determine tourists’ attracting tourists to the area. Few photographs were temporal variations, length of stay, daily average taken in rural areas surrounding the main urban core number of tourists, individual movement traces and a of the EMA, probabloy because these areas are less flow map. A non-parametric density estimation called accessible to visitors and much of this land is private or a kernel density estimation was then used to generate tribal. tourism hotspots (Li et al. 2011). Casalegno et al. (2013) A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 87 Figure 5.2 Pattern of geo-tagged photo uploads in relation to major land cover types within the eThekwini Municipal Area. Page 88 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Statistical analysis of these patterns in relation to the features of the different grid cells showed trhat the numbers of photographs were significantly influenced by the presence of certain land cover types, and that the interaction of man-made and natural features was important in determining the attractions of the latter (see Box 5.1 for details). Box 5.1 Analysis of photo densisty A generalized linear model was then used to analyse the numbers of photographs per grid cell in relation to the extent of different land use/land cover types within each grid cell. This was to determine the factors influencing the relative values of the cells. A semi-log model was used for the analysis where the dependent variable was the natural log of the photo count within each grid cell. The independent variables included in the model were based on a number of different land uses and land cover types as adapted from the combined D’MOSS and LULC map. These included the amount of commercial/retail land, industrial land, roads, urban settlement, informal settlement, rural settlement, agricultural land, parks, golf courses, natural land not within protected areas, land within protected areas, dams, estuarine and marine areas. Interaction terms were included e.g. to account for accessibility in reaching certain land cover types. Data were analysed using ordinary least squares (OLS) regression in R Project for Statistical Computing (ver. 3.2.0). Collinearity amongst variables was tested and those that displayed high levels were removed from the model as suggested by the software analysis through stepwise regression techniques. The model estimation results are shown in Table 5.3. Table 5.3 Model estimation results co-efficient std. error t. value Pr(>|t|) (Intercept) 1.09 0.11 10.25 < 2e-16 *** Commercial 0.02 0.0042 4.52 8.02E-06 *** Industrial 0.0027 0.0012 2.29 0.0228 * Urban Settlement 0.0010 0.00 2.64 0.0087 ** Rural Settlement -0.0013 0.00 -3.27 0.0012 ** Golf Course 0.0143 0.01 1.86 0.0631 . Horse Racetrack 0.0340 0.01 2.29 0.0223 * Protected Areas 0.0111 0.00 3.89 0.0001 *** Dam 0.0035 0.00 1.80 0.0728 . Estuary 0.0138 0.00 5.45 8.64E-08 *** Natural Land*Road 0.0003 0.00 5.69 2.49E-08 *** Marine*Built Env. 0.0007 0.00 6.93 1.74E-11 *** Sample size 414 R-squared 0.5 Adjusted R-squared 0.48 Notes: (1) ***p<0.001, **p<0.01, *p<0.05, .p<0.10. The results indicated that accessibility is an important factor in the value of natural open space. Natural open space outside of protected areas only had a significant effect in interaction with roads. Similarly, coastal environment had a significant effect in interaction with the built environment (commercial and urban settlement areas only), reflecting the importance of beach front accommodation and amenities. Photographs often included beachfront activities or views of the beach from hotels and restaurants along the coastal strip. Natural open space within protected areas was highly significant, with these areas having a high number of photographic uploads. Dams and estuaries were both positive and highly significant, with high concentrations of photographic uploads of these features. Rural settlements had a negative influence on photograph counts due to the low number of photgraphic uploads in these areas as a result of visitor inaccessibility. Urban, commercial and industrial land cover had a positive and significant influence on photographic counts due to the high numbers of architectural and cultural photographs taken in the urban areas across the EMA. Variables that did not influence model results included informal settlements, parks, sports A SPATIAL grounds and VALUATION agricultural OF THE NATURAL land. AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 89 The statistical distribution of the number of photographs average value of R234 000 per hectare. Therefore the was negatively binomial. In order to test whether the combined tourism value of natural and man-made open relationship between tourism value and photo densities space is approximately R2.4 billion. should be linear or non-linear, we conducted a broad- scale test at the national scale, using photo uploads at the 0.25 degree scale and national parks visitor data for 2014. On the basis of this, it was assumed that Table 5.4 The overall leisure tourism value (R million) and % leisure tourism value was directly proportional to photo photographs for the five land use categories. numbers. Therefore the proportion of leisure tourism value derived from any grid cell was assumed to be equal Category Total Leisure Overall % Tourism Value Photographs to the total value multiplied by the percentage of photo (R million) uploads in that grid cell. Built 2 929 42.3 The relative importance of places of interest to leisure- Environment based tourists (whether for cultural, architectural, Rural/ 232 15.2 nature-based or other reasons) was assumed to be Agricultural reflected by the spatial distribution of georeferenced Natural Open 928 31.8 photograph uploads on Panoramio.com. In order to Space estimate the contribution of different attractions to Man-made 384 2.6 tourism value, the contents of the photographs were Open Space analysed. The content (or setting) of each photograph Marine/ 1 085 8.2 in each grid cell was examined1 and put into one of five Coastal categories; (1) built environment, (2) natural open space, TOTAL 5 559 100 (3) man-made open space, (4) rural or agricultural, and (5) marine or coastal. The built environment category included urban, dense rural and informal settlements, as well as commercial/retail land, rail and industrial land. Following this, the percentage contribution of different settings to tourism value was determined for each grid cell. The corresponding value was then assigned to the actual land cover within each grid cell in the EMA based on the combined LULC/D’MOSS map. In this way, the total value added by leisure tourism (R5.6 billion; Table 5.2) was assigned to five different land use classes within each grid cell. The total nature- based tourism value was mapped using GIS and is shown in (Figure 5.3). A number of the natural areas with the highest values were located within nature reserves, such as Kranzkloof, Shongweni, Beachwood and Umhlanga Lagoon. Estuaries, such as Durban Bay and uMngeni also had high values. The total tourism value assigned to natural habitats (natural vegetation, freshwater systems, estuaries and the coastal environment) was almost R2 billion and accounted for 40% of all photographic uploads (Table 5.4) with the coastal environment contributing just over half of this value. The tourism value associated with the coast was very high, with the most popular coastal areas having values of up to R66 million per kilometer stretch of coastline (Figure 5.3). Man-made open space such as parks and golf courses had a total leisure tourism value of R382 million at an 1 This analysis was achieved by downloading the details of each photograph from each grid cell using code developed by Panoramio. The output for each grid cell produced information about the number of photographs, the number of users taking the photographs, the title of each photograph and a url link to each photograph in Panoramia. Page 90 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure 5.3 Nature-based tourism value in the EMA shown for terrestrial natural open space as R/ha and for the coast as R/km. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 91 This page intentionally blank. Page 92 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY VI. SUMMARY AND CONCLUSIONS Natural and semi-natural systems within the eThekwini Generally, ecosystem services in urban areas are Municipal Area give rise to flows of ecosystem services characterised by high demand/use due to the large worth in the order of R4.2 billion per year. In 2002 the number of immediate local beneficiaries, compared value of Durban’s Metropolitan Open Space System to ecosystem services generated in rural sparsely (D’MOSS) was estimated to be R3.1 billion per annum populated areas (Elmqvist et al. 2015). The spatial excluding tourism value (eThekwini Municipality 2002). representation of the relative value of ecosystem This earlier study raised public sector awareness of services, or the mapping of geographic variation in these the value of these areas, and resulted in improved values, provides the opportunity to compare areas and allocation of resources for their management. These types of values. The relative value, expressed as net value estimates were extrapolated from international present value per hectare, can be compared across studies (mainly Costanza et al. 1997) on the basis of total groups of services and can be used to identify areas of areas of different habitat types within the city. Wetlands importance based on the relative values associated with and forests were valued at about R160 000 and R21 000 different areas in the EMA. The geographic variation per hectare per year, respectively. The Costanza-derived between groups of services is shown in Figure 6.1. A values were higher than our estimates, as expected, due summary of the total net present value associated with to the inherent bias in the data from which they were estuaries (subsistence fishery value and nursery value) is derived. The updated study has shown that cultural included in Table 6.2. These values were not included in services are the dominant values of urban green open Figure 6.1 due to scale issues associated with mapping spaces and make a substantial difference to estimates of smaller estuaries. The provisioning and regulating overall value. services provided by estuaries have a net present value of R203-263 million with Durban Bay, uMngeni Based on a 20-year time horizon and a discount rate Estuary and uMkhomazi Estuary accounting for 36.2%, of 6%, the total asset value of the natural and semi- 12.7% and 10.9%, respectively (Table 6.2). However, natural capital in the eThekwini Municipal Area that the seven estuaries along the south coast from iLovu encompasses the city of Durban and surrounding peri- to iMahlongwa have the highest relative values, largely urban and rural lands, was estimated to be at least R48 a result of being in a better condition compared to the billion (Table 6.1). The true value is probably much higher. estuaries further north. For instance, applying a social discount rate of 3% would yield a value about 30% higher, or around R62 billion. Within the city, and overall, amenity value is by far the The results from this study have shown that ecological most important economic value of natural and semi- infrastructure in cities is not only environmentally and natural open space. This large value is derived from socially desirable, but is also beneficial from an economic a small proportion of the D’MOSS (Figure 6.1). The point of view with numerous benefits accruing from patterns observed for property value (representing natural open space areas. value to residents), and tourism value were similar, with certain areas in the EMA contributing significantly to The provisioning value natural systems was estimated both. These areas include the natural open space areas to be R1.1-1.5 billion (NPV), with fuelwood and river in Kloof and Hillcrest, the nature reserves along the coast water contributing the most to this value. Considering such as Beachwood, and the central areas of Durban city water is collected from rivers and streams by only 0.5% centre, Durban Bay and the beachfront promenade. of the population in the EMA, this resource is thought to be the most valuable in terms of the service it provides per user household. The remaining values represent approximately R225-292 million of the total provisioning value of natural resources in the study area (Table 6.1). Regulating services have a net present value of about R1.0-1.2 billion with carbon storage and flow regulation accounting for 39% and 34% of this value, respectively. This value is believed to be a conservative estimate. The many benefits derived from these supporting services are indirect and particularly difficult to value, and most of the regulating services also play the role of supporting services that influence the outputs of other ecosystem services. The amenity value associated with natural and semi-natural open space areas in the EMA was estimated to be R45.7-59.4 billion, 96% of the overall value of ecosystem services (Table 6.1). A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 93 Table 6.1 Total value of ecosystem services in the EMA. Values in R millions (2015). Ecosystem services Annual Value NPV (20 y, 6%) NPV (20 y, 3%) (R millions) (R millions) Provisioning services River water for domestic use 31.8 364.7 474.1 Fuelwood 46.5 533.4 693.4 Timber poles 6.6 75.7 98.4 Food and medicinal plants 4.7 53.9 70.1 Grass and reeds 1.4 16.1 20.9 Bush meat 0.6 6.9 9.0 Fishery resources 6.3 72.3 94.0 Sub-total 97.9 1 122.9 1 459.9 Regulating services Carbon storage 34.3 393.4 511.4 Agricultural support 1 11.5 15.0 Fisheries support (nursery function) 11.4 130.8 170.0 Flow regulation 29.5 338.8 338.8 Sediment retention 3.1 35.2 45.8 Water quality amelioration 7.8 89.1 115.8 Sub-total 87.1 998.8 1 196.8 Cultural services Amenity value to property owners 1 586.8 18 200.0 23 660.0 Tourism value 2 400.0 27 527.8 35 786.1 Sub-total 3 986.8 45 727.8 59 446.1 TOTAL 4 171.8 47 849.5 62 102.8 The values associated with natural and semi-natural football, activities that cannot be done easily in natural, open space provide more of an understanding as to how densely vegetated areas, and also due to the fact residents in different areas value open space and these that residents seek a safer and more easily accessible results suggest that city planners and developers need environment in which to enjoy open space in the city. to consider the spatial context of open space areas and how best to provide and protect these areas across the In contrast to amenity values, provisioning services are EMA. The condition of the natural open space areas in largely only significant in the peri-urban and rural areas the EMA is extremely important, and was found to have of the EMA. The substantial area of rural landscapes a significant influence on amenity value in particular. around the urban edge, much of which are communal The study found that residents in the EMA are willing lands under the ownership of the traditional authority, to pay for access to or views of open space, but only deliver important provisioning services to the large when these areas are in good condition. The results numbers of poor households residing in the relatively have also shown that residents attach a higher amenity dense settlements in these areas. Within the urban value to urban parks than they do to natural open space. edge, the large natural areas tend to be under private It is assumed that this is due to the many recreational ownership or state protection. The value of these opportunities that parks offer, such as jogging or playing services is highest in the outer-west and southern extents of the municipality where there are still large Page 94 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY tracts of natural vegetation, such as woodlands and to the urban edge than for provisioning and amenity forests. In these areas, outside of the urban edge, there services (Figure 6.1 through 6-3). The values of these is a significant proportion of the population that relies services were found to be highest upstream of Inanda on natural resources. The provisioning value in the and Hazelmere Dams in the northern and outer-west northern area of the EMA and in the urban core tends to areas of the EMA, where sediment retention and water be considerably lower and largely restricted to the river quality amelioration were most valuable, and also in systems and wetlands in these areas. the catchments of the downtown and harbour areas, where flood attenuation was highly valuable. Within the If urbanisation is properly managed, the spatial disjunct lower urban areas, however, hydrologic-related services of provisioning and amenity values is likely to track tend to have been overwhelmed to the extent that they the future growth of the city and lead to the increased can no longer ameliorate the increased run-off and value of the remaining open space areas. If not properly pollution from urban areas to a significant extent. This is managed, a dead zone could be created at the city’s a general reality of urban systems which is increasingly periphery in terms of ecosystem services, and valuable being addressed through innovative engineering opportunities will be lost. As Durban grows and the peri- solutions. Indeed regulating services are one area where urban areas become densely populated and urbanised, it technological innovation does show promise of being able is likely that the provisioning services provided by those to augment or replace ecosystem services, and where landscapes will make way for amenity services to future increased efficiency may be also desirable to protect the urban inhabitants, as the demands of urban inhabitants supply of ecosystem services from aquatic ecosystems replace those of the former rural inhabitants. Both that, unlike terrestrial systems, are more certain to remain within the planned urban edge and beyond, informal and in place, in one form or another. The optimal balance rural settlements will continue to grow and densify, and between relying on the services provided by natural the city will continue to be under pressure to provide capital versus implementing engineering solutions to housing for the poor for some time to come. It will neutralise the effects of urbanisation on run-off and water be important to consider the implications of how this quality is explored in more detail in the accompanying growth is allowed to happen. The findings of this study report (Turpie et al. 2016). suggest that remnant green open space areas become increasingly valuable with urbanisation and increasing Very few comprehensive studies have quantified and incomes. Thus ensuring the protection of key open space spatially evaluated the benefits supplied by urban areas in the areas being settled would secure potentially ecosystems (Elmqvist et al. 2015). Although this was a valuable sources of amenity and spiritual and physical desktop assessment which still has substantial room for wellbeing for these future communities. improvement and refinement through further research, it represents the most comprehensive assessment of The role and spatial variation of regulating services ecosystem services within an urban environment in is particularly interesting. Carbon storage is the most Africa to date, and one of perhaps only a handful in the important of these services from a local economy world. perspective, and is also important from an international perspective. This value and the global value should be Overall, the study shows that the ecosystem services taken into consideration in South Africa’s and eThekwini profile of an urban area is very different from that of Municipality’s commitments to combating climate non-urban landscapes or regional level assessments. change. The value of agricultural support services is While natural and semi-natural ecosystems within the modest, due to the relatively small extent of activities study area provide a full suite of ecosystem services, the that benefit from this service. These are also generally value is dominated by amenity services to residents and outside of the urban edge. The nursery service of visitors, both of which make a considerable and tangible estuaries (Table 6.2), which supports the fisheries sector, contribution to Durban’s economy. Even so, the estimate is comparatively large. The fisheries that are supported of amenity value does not capture all of the recreational, include both recreational and small-scale commercial spiritual and health benefits associated with of green fisheries within and beyond the coastal areas of the open space, since estimating these values requires EMA, though the majority of these beneficiaries are further data collection. likely to reside within the EMA. The findings of this study can be used to identify Three of the regulating services described are related priority areas for supply of ecosystem services and to catchment hydrology. The capacity to perform these will be useful for decision-making processes related services tends to be greatest in the surrounding rural to understanding the tradeoffs between spatial landscapes because of their location and size, but the development and conservation. This is pertinent to both beneficiaries of these services are downstream, and the environmental and strategic development planning value of hydrologic services is also dictated to a large within the eThekwini Municipality in in forming the extent by the location of infrastructure. Thus the spatial development of a thriving, resilient city. While the study variation of these services was more irregular in relation focuses on ecosystem services, it is important to note A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 95 Figure 6.1 Summary of the provisioning, regulating and cultural service values (NPV) associated with natural and semi-natural open space in the EMA, excluding estuary provisioning and regulating services. Page 96 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure 6.2 Summary of the provisioning, regulating and cultural service values (NPV) associated with natural and semi-natural open space in the EMA, excluding estuary provisioning and regulating services. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 97 Figure 6.3 Summary of the provisioning, regulating and cultural service values (NPV) associated with natural and semi-natural open space in the EMA, excluding estuary provisioning and regulating services. Page 98 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY that it does not provide a full estimate of the value of biodiversity per se. Our study does not capture the existence value and some of the other intangible, cultural values attached to biodiversity, nor does it adequately capture the role of species richness and community structure on the capacity of ecosystems to supply ecosystem services. This is a shortcoming of most valuation studies, and it is for this reason that conservation and planning decisions cannot be made on the basis of economic values alone. Planning should also incorporate biodiversity targets that are determined on the basis of their intrinsic value and understood role in maintaining ecosystem integrity and resilience. Table 6.2 Summary of the provisioning and regulating service values (NPV, 20 y, 6%) associated with estuaries in the EMA (R millions) Estuary Provisioning Services Regulating Services (subsistence fisheries) (nursery value) NPV @ 6% NPV @3% NPV @ 6% NPV @3% uTongati - - 0.4 0.5 uMdloti 7.1 9.2 2.7 3.6 oHlanga 6.9 9.0 1.7 2.2 uMngeni 7.8 10.1 17.9 23.3 Durban Bay 9.8 12.7 63.6 82.7 iSiphingo - - 0.3 0.4 eziMbokodweni - - 1.4 1.8 aManzimtoti - - 0.6 0.8 Little aManzimtoti - - 0.2 0.3 iLovu 7.8 10.1 9.6 12.5 uMsimbazi 6.5 8.5 7.1 9.2 uMgababa 6.8 8.8 5.6 7.3 Ngane 1.2 1.5 0.5 0.7 uMkhomazi 7.1 9.2 14.9 19.4 uMahlongwane 4.6 5.9 1.9 2.5 iMahlongwa 6.4 8.3 1.8 2.3 TOTAL 71.8 93.4 130.8 169.3 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 99 This page intentionally blank. Page 100 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY VII. REFERENCES Abou-Shaara, H. F. 2014. The foraging behaviour of Ecosystem Functioning, and Human Wellbeing: An honey bees, Apis mellifera: a review. Veterinarni Ecological and Economic Perspective. Oxford: Oxford Medicina 59(1): 1–10. University Press, UK, pp. 248–262. Ajonina, G. J. G. Kairo, G. Grimsditch, T. Sembres, G. Barnes, J. I., Nhuleipo, O., Muteyauli, P. I. & MacGregor, Chuyong, D. E. Mibog, A. Nyambane and C. FitzGerald J. 2005. Preliminary economic asset and flow accounts 2014. Carbon pools and multiple benefits of mangroves for forest resources in Namibia. DEA Research Discussion in Central Africa: Assessment for REDD+. 72pp. Paper, Ministry of Environment and Tourism. Anderson, S. T. 2000. The effect of open space on single- Beckley, L.E. & Fennessy, S.T. 1996. The beach-seine family, residential home property values. Macalester fishery off Durban, KwaZulu-Natal. South African Journal College, St Paul Minnesota. of Zoology 31: 186–192. Andrienko, G., Andrienko, N., Bak, P., Kisilevich, S. & Behrer, P. 2010. Building in the mountains: a hedonic Keim, D. 2009. Analysis of community-contributed analysis of the value of degraded mountain views using space- and time-referenced data: example of Flickr GIS modelling. Discussion Paper 2009-15, Cambridge, photos. Proceedings of the IEEE Symposium on Visual Mass.: Harvard Environmental Economics Program, May Analytics Science and Technology (VAST 2009), Poster 2010. Paper, 2009. Beron, K., Murdoch, J. & Thayer, M. 2001. The benefits Alfaro, R.W.F. 2015. Evaluation of cultural ecosystem of visibility improvement: new evidence from the Los aesthetic value of the state of Nebraska by mapping Angeles metropolitan area. Journal of Real Estate geo-tagged photographs from social media data of Finance and Economics 22: 319-337. Panoramio and Flickr. Community and Regional Planning Program: Student Projects and Theses. Paper 34. Blumenfeld, S., Lu, C., Christophersen, T. & Coates, D. 2009. Water, Wetlands and Forests. A Review of Allanson, B. & Baird, D. 1999. Estuaries of South Africa. Ecological, Economic and Policy Linkages. Secretariat of Cambridge University Press. the Convention on Biological Diversity and Secretariat of the Ramsar Convention on Wetlands, Montreal and Allsopp, M. H., de Lange, W. J. & Veldtman, R. Gland. CBD Technical Series No. 47. 2008. Valuing insect pollination services with cost replacement. PLoS ONE 3(9): e3128. doi:10.1371/journal. Bolund, P. & Hunhammar, S. 1999. Ecosystem services in pone.0003128. urban areas. Ecological Economics 29: 293 – 301. Anderson, S. T. & West, S. E. 2006. Open space, Branch, G. M., May, J., Roberts, B., Russell, E. & Clark, residential property values, and spatial context. Regional B. M. 2002. Case studies on the socio-economic Science and Urban Economics 36: 773-789. characteristics and lifestyles of subsistence and informal fishers in South Africa. South African Journal of Marine ASCE, 1992. Design and Construction of Urban Science 24: 439–462 Stormwater Management Systems, New York, USA. Brauman, K.A., Daily, G.C., Duarte, T.K. & Mooney, H.A. Banzhaf, S. & Jawahar, P. 2005. Public benefits of 2007. The nature and value of ecosystem services: an undeveloped lands on urban outskirts: non-market overview highlighting hydrological services. Annual valuation studies and their role in land use plans. Review of Environment and Resources 32: 67-98. Resources for the Future Report. Washington, DC. Broadley, D.G. 2010. Scelotes inornatus. The IUCN Barbier, E.B. 1994. Valuing environmental functions: Red List of Threatened Species. Version 2015.2. www. tropical wetlands. Land Economics 70: 155-173. iucnredlist.org. Downloaded on 21 July 2015. Barbier, E.B. 2011. Capitalizing on nature: ecosystems as Brouwer, S.L., Mann, B.Q., Lamberth, S.J., Sauer, W.H.H. natural assets. Cambridge University Press, Cambridge & Erasmus, C. 1997. A survey of the South African shore- and New York. angling fishery. South African Journal of Marine Science 18: 165–177. Barbier, E.B., Baumgartner, S., Chopra, K., Costello, C., Duraiappah, A. Hassan, R., Kinzig, A., Lehman, M., Pascual, U., Polasky, S. & Perrings, C. 2009. The valuation of ecosystem services. In: Naeem, S., Bunker, D., Hector, A., Loreau, M. & Perrings, C. (eds) Biodiversity, A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 101 Cartwright, A., Blignaut, J., De Wit, M., Goldberg, Costanza. R., de Groot, R. Sutton, P., van der Ploeg, S., K., Mander, M., O’Donoghue, S. & Roberts, D. 2013. Anderson, S.J., Kubiszewski, I., Farber, S., Turner, R.K. Economics of climate change adaptation at the local 2014. Changes in the global value of ecosystem services. scale under conditions of uncertainty and resource Global Environmental Change 26: 152-158. constraints: the case of Durban, South Africa. Environment & Urbanization 25(1): 1–18. CSIR 2008. Sand supply from rivers within the eThekwini jurisdiction: impllications for coastal sand budgets Casalegno, S., Igner, R., DeSilvey, C. & Gaston, K.J. and resource economics. Theron, A. K., de Lange, W., 2013. Spatial covariance between aesthetic value and Nahman, A. & Hardwick, D. (eds). Report published by ecosystem services. PloS ONE 8: e68437. CSIR. Report No: CSIR/NRE/ECO/ER/2008/0096/C. Cavender-Bares, J., Polasky, S., King, E. & Balvanera, P. Cyrus, D.P. & Blaber, S.J.M. 1987. The influence of 2015. A sustainability framework for assessing trade-offs turbidity on juvenile marine fishes in estuaries. Part 1. in ecosystem services. Ecology and Society 20(1): 17. Field studies at Lake St. Lucia on the southeastern coast of Africa. Journal of Experimental Marine Biology and Chan, K.M.A., Goldstein, J., Satterfield, T., Hannahs, Ecology 109: 53-70. N., Kikiloi, K, Naidoo, R., et al. 2011. Cultural Services and non-use values. In: Karieva, P., Daily, G., Ricketts, Cyrus, D.P. & Blaber, S.J.M. 1988. The potential effects T., Tallis, H., Polasky, S. (Eds.), Natural Capital: Theory of dredging activities and increased silt load on the St. & Practice of Mapping Ecosystem Services. Oxford Lucia system, with special reference to turbidity and University Press Inc., New York, pp. 206–228. estuarine fauna. Water SA 14(1): 53-70. Chiesura, A. 2004. The role of urban parks for the Davies, Z. G., Edmondson, J.L., Heinemeyer, A., Leake, sustainable city. Landscape and Urban Planning. 68: 129- J. R. & Gaston, K. J. 2011. Mapping an urban ecosystem 138. service: quantifying above-ground carbon storage at a city-wide scale. Journal of Applied Ecology 48: 1125– Chmura, G.L., Anisfeld, S.C., Cahoon, D.R. & Lynch, 1134. J.C. 2003. Global carbon sequestration in tidal, saline wetland soils. Global Biogeochemical Cycles 17 (4): 1111. De Groot, R.S., Wilson, M.A. & Boumans, R.M.J. 2002. A typology for the classification, description and valuation Cho, S-H., Bowker, J. M. & Park, W. M. 2006. Measuring of ecosystem functions, goods and services. Ecological the contribution of water and green space amenities Economics 41: 393– 408. to housing values: an application and comparison of spatially weighted hedonic models. Journal of Dearmont, D., McCarl, B. & Tolman, D. 1998. Costs of Agricultural and Resource Economics 31: 485-507. Water Treatment Due to Dimished Water Quality: A Case Study in Texas. Water Resources Research 34 (4): Cho, S-H., Poudyal, N.C. & Roberts, R.K. 2008. Spatial 849–854. analysis of the amenity value of green open space. Ecological Economics 66: 403-416 Department of Agriculuture, Environmental Affairs and Rural Development (DAEARD), 2010. KwaZulu-Natal Cocks, M.L. & Wiersum, K.F. 2003. The significance of State of the Environment 2004: Marine and Coastal plant diversity to rural households in the Eastern Cape Environment Specialist Report. KwaZulu-Natal Provincial province of South Africa. Forests, Trees and Livelihoods Government, Pietermaritzburg. 13: 39–58. De Wit, M., Van Zyl, H., Crookes, D., Blignaut, J., Jayiya, Conte, M., Ennaanay, D., Mendoza, G., Walter, M. T., Goiset, V. & Mahumani, B. K. 2009. Investing in T., Wolny, S., Freyberg, D., Nelson, E. & Solorzano, L. natural assets: A business case for the environment in 2011. Chapter 6: Retention of nutrients and sediment the City of Cape Town. Report prepared for the City of by vegetation. In: Kareiva, P., Tallis, H., Ricketts, T. H., Cape Town, Cape Town, South Africa. Daily, G. C. & Polasky, S. (eds) Natural Capital: theory and practice of mapping ecosystem services. Oxford Demetriades, N. 2007. An inventory of sand mining University Press Inc., New York. operations in KwaZulu-Natal estuaries: Thukela to Mtamvuma. Investigational report for Coastwatch, Costanza, R., D’Arge, R., de Groot, R., Farber, S., Grasso, WESSA KwaZulu-Natal. M., Hannon, B., Limburg, K., Naeem, S., O’Neill, R. Paruelo, J., Raskin, R.G., Sutton, P. & van den Belt, M. Dennison, D. B. & Lyne, M. C. 1997. Analysis and 1997. The value of the world’s ecosystem services and prediction of water treatment costs at the DV Harris natural capital. Nature 387: 253-260. Plant in Umgeni Catchment Area. Agrekon 36: 40. Page 102 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Do, A.Q. & Grudnitski, G. 1995. Golf courses and Environmental Planning and Climate Protection residential house prices: an empirical examination, Department (EPCPD) 2012. eThekwini Municipality Journal of Real Estate Finance and Economics 10: 261- Systematic Conservation Plan. eThekwini Municipality. 270. eThekwini Municipality 2002. Durban Environmental Doss, C. R. & Taff, S. J. 1996. The influence of wetland Services Mangement Plan. Prepared by: The type and wetland proximity on residential property Environmental Management Branch of eThekwini values. Journal of Agricultural and Resource Economics Municipality and Urban Explorations. 21:120-129 eThekwini Municipality 2008. Design Manual: Guidelines Dovie, D. B., Shackleton, C. M. & Witkowski, E. T. 2002. And Policy For The Design Of Stormwater Drainage Direct-use values of woodland resources consumed and And Stormwater Management Systems. Engineering traded in a South African village. International Journal of Unit, Coastal Stormwater and Catchment Management Sustainable Development and World Ecology 9: 269-283. Department. Dovie, D. B. K., Witkowski, E. T. F. & Shackleton, C. M. eThekwini Municipality 2009. Durban’s beaches: more 2004. The fuelwood crisis in southern Africa – relating than just sand. Compiled by Natasha Govender at the fuelwood use to livelihoods in a rural village. GeoJournal Environmental Management Department. 60: 123–133. eThekwini Municipality 2011. Freshwater Wetlands of Dunlop, S. W. & Mann, B. Q. 2012. An assessment of eThekwini Municipality: What they are and why they are participation, catch and effort in the KwaZulu-Natal important. Compiled by the Environmental Planning and shore-based marine linefishery, with comments on Climate Protection Department. management effectiveness. African Journal of Marine Science 34: 479 – 496. eThekwini Municipality 2012. State of Biodiversty Report 2011/2012. Compiled by the Environmental Planning and Dunlop, S. W. & Mann, B. Q. 2013. An assessment of Climate Protection Department. participation, catch and effort in the offshore boat- based linefishery in KwaZulu-Natal, South Africa. African eThekwini Municipality 2014a. eThekwini Municipality Journal of Marine Science 35: 79 – 97. Annual Report 2013/2014. Department of Water Affairs (DWA), 2013a. Classification eThekwini Municipality, 2014b. Tourism in eThekwini: of Water Resources and Determination of the positioning Durban globally. Economic Development and Comprehensive Reserve and Resource Quality Objectives Growth in eThekwini 8: 2-57. in the Mvoti to Umzimkulu Water Management Area: Status quo assessment, IUA delineation and biophysical eThekwini Municipality 2014c. State of Biodiversty node identification. Prepared by: Rivers for Africa eFlows Report 2013/2014. Compiled by the Environmental Consulting (Pty) Ltd. Planning and Climate Protection Department. Department of Water Affairs (DWA), 2013b. The eThekwini Municipality 2015. Medium Term Revenue uMkhomazi Water Project Phase 1. Technical Feasibility And Expenditure Framework 2015/2016 to 2017/2018. Study: Raw Water. Water Resources Yield Assessment As part of the Annual Budget 2015/16. Report Supporting Document 1: Sediment Yield Report. Everett, B. I. (ed) 2014. Marine and estuarine fisheries Prepared by: Aecom. along the KwaZulu-Natal Coast: an inventory and brief Eigenbrod, F., Bell, V.A., Davies, H.N., Heinemeyer, description. Oceanographic Research Institute. Special A., Armsworth, P.R. & Gaston, K.J. 2011. The impact Publication no. 11. ISBN 978-0-620-64146-3. of protected increases in urbanization on ecosystem Fankhauser, S. & Tol, R.S.J. 1997. The Social Costs of services. Proceedings of the Royal Society – Biological Climate Change: The IPCC Second Assessment Report Sciences 278: doi: 10.1098/rspb.2010.2754. and Beyond. Mitigation and Adaptation Strategies for Elmqvist, T. et al. (eds). 2013. Urbaniszation, biodiversity Global Change 1(4): 385–403. and ecosystem services: challenges and opportunities: a Fleischer, A. & Tsur, Y. 2000. Measuring the recreational global assessment. value of agricultural landscape. European Review of Elmqvist, T., Setala, H., Handel, S. N., van der Ploeg, S., Agricultural Economics 27: 385-398. Aronson, J., Blignaut, J. N., Gomez-Baggethun, E., Nowak, D. J., Kronenberg, J. & de Groot, R. 2015. Benefits of restoring ecosystem services in urban areas. Current Opinion in Environmental Sustainability 14: 101-108. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 103 Forbes, A. T. & Demetriades, N. T. 2008. Estuaries of Haase, D., Schwarz, N., Strohbach, M., Kroll, F. & Seppelt, Durban, KwaZulu-Natal, South Africa. Report for the R. 2012. Synergies, Trade-offs, and Losses of Ecosystem Environmental Management Department, eThekwini Services in Urban Regions: an Integrated Multiscale Municipality. Framework Applied to the Leipzig-Halle Region, Germany. Ecology and Society 17(3): 22. Forbes, T. & Forbes, N. 2013. Penaeid prawns. In Perissinotto, R. Stretch, D. & Taylor, R. (eds.) Ecology and Hagler, J.R., Mueller, S., Teuber, L.R., Machtley, S.A. & Conservation of Estuarine Ecosystems: Lake St Lucia as a Van Deynze, A. 2011. Foraging range of honey bees, Apis global model. Cambridge University Press, New York. mellifera, in alfalfa seed production fields. Journal of Insect Science 11: 144. Friedrich, E., Pillay, S. & Buckley, C. A. 2009. Environmental life cycle assessments for water Hansford, J. R. 2003. Raising of Hazelmere Dam treatment processes – A South African case study of an Feasibility Study. Hydrology Report. Prepared by Knight urban water cycle. Water SA 35 (1): 73 – 84. Piesold Consulting for the Department of Water Affairs and Forestry and Umgeni Water. DWAF Report No. PB Gebremedhin, S. K. 2009. Analysis and Prediction of U300-00-0301, UW Report No. U3/02F/10. Chemical Treatment Cost of Potable Water in the Upper and Middle Vaal Water Management Areas. MSc Thesis, Hennings, L. & Soll, J. 2010. Wildlife corridors and Agricultural Economics. University of KwaZulu-Natal, permeability: A literature review. Metro Sustainability Pietermaritzburg. Centre, Oregan Metro, Portland. Gibbons, S., Mourato, S. & Resende, G. 2011. The High, C. & Shackleton, C. M. 2000. The comparative amenity value of English nature: a hedonic price value of wild and domestic plants in home gardens of approach. Spatial Economics Research Centre (SERC). a South African rural village. Agroforestry Systems 48: Discussion Paper 74. London School of Economics. 141–156. Glenday, J. 2007. Carbon Storage and Sequestration Hill, G.W. & Courtney, P.R. 2006. Demand analysis Analysis for the eThekwini Environmental Services projections for recreational visits to countryside Management Plan Open Space System. Report prepared woodlands in Great Britain. Forestry 79: 185–200. for eThekwini Municipality Environmental Management Department. Howarth, C. 2012. Where to go? Using social media to assess the spatial distribution of recreation on the Great Goes, P.B. 2014. Big Data and IS Research, MIS Quarterly, Barrier Reef, A thesis submitted in partial fulfilment of Academic Search Premier, 38(3): iii-viii. the requirements for the degree of Master of Science and the Diploma of Imperial College London, UK. Graham, P. 2004. Modelling the water quality in dams within the Umgeni Water operational area with Institute of Natural Resources. 2004. eThekwini emphasis on algal relations. PhD Thesis, North West Agricultural Status Quo. Rural agricultural land potential University. assessment and agribusiness policy for eThekwini Municipality. Graham, M., Blignaut, J., de Villiers, L., Mostert, D., Sibande, X., Gebremedhin, S., Harding, W., Rossouw, N., Integrated Development Plan (IDP), 2014. eThekwini Freese, N. S., Ferrer, S. & Browne, M. 2012. Development Municipality Draft Integrated Development Plan: Five of a generic model to assess the costs associated with year plan 2012/13 to 2016/17. Annual Review 2014/2015. eutrophication. WRC Report No. 1568/1/12. ISBN 978-1- 4312-0265-2. International Telecommunications Unions website, 2012. Percentage of individuals using the internet Grey-Ross, R., Downs, C. T. & Kirkman, K. 2010. An 2000–2010. Accessed August 2015. Available at: http:// Assessment of Illegal Hunting on Farmland in KwaZulu- www.itu.int/ITUD/ict/statistics/material/excel/2010/ Natal, South Africa: Implications for Oribi (Ourebia IndividualsUsingInternet_00-10.xls. ourebi) Conservation. South African Journal of Wildlife Research 40: 43-52. IPCC 1996. Climate Change 1995: The science of climate change. Report of the Intergovernmental Panel on Ground Truth Consultants 2006. eThekwini Municipality Climate Change. Cambridge University Press, Cambridge, State of Rivers Report. Final Summary Report, 111pp. United Kingdom and New York, NY, USA. Haarhoff, J. & Cassa, A. 2009. Introduction to flood hydrology. Juta Legal and Academic Publishers. Page 104 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY IPCC, 2007. Climate Change 2007: The Physical Science Lamberth, S. J. & Turpie, J. K. 2003. The role of estuaries Basis. Contribution of Working Group I to the Fourth in South African fisheries: economic importance and Assessment Report of the Intergovernmental Panel on economic implications. African Journal of Marine Science Climate Change. Cambridge University Press, Cambridge, 25: 131-157. United Kingdom and New York, NY, USA. Lannas, K. S. M. & Turpie, J. K. 2009. Valuing the Jensen, C. U., Panduro, T. E. & Lundhede, T. H. 2014. provisioning services of wetlands: contrasting a rural The vindication of Don Quixote: the impact of noise and wetland in Lesotho with a peri-urban wetland in South visual pollution from wind turbines. Land Economics 90: Africa. Ecology and Society 14: 18. 668-682. Lee, A.C.K., Jordan, H. C. & Horsley, J. 2015. Value of Kachkaev, A. & Wood, J. 2013. Crowd-sourced urban green spaces in promoting healthy living and photographic content for urban recreational route wellbeing: prospects for planning. Journal of Risk planning. Paper presented at the University Transport Management and Healthcare Policy 8: 131-137. Study Group UK Annual Conference, Oxford, UK. Leiva, D.F. & Page, D.R. 2000. The use of cluster Kairo, J. G. 2001. Ecology and Restoration of Mangrove analysis in distinguishing farmland prone to residential Systems in Kenya. PhD Thesis Laboratory of Plant development: a case study of Sterling, Massachusetts. Sciences and Nature Management, Univeristy of Environmental Management 25: 541-548. Brussels. Li, C., Yang, Z., Sun, X., Su, X., Zheng, S., Dong, R. & Shi, L. Kairo, J. G., Dahdouh-Guebas, F., Gwada, P. O., Ochieng, 2011. Photography-based analysis of tourists’ temporal- C. & Koedam, N. 2002. Regeneration Status of Mangrove spatial behaviour in the old town of Lijiang. International Forests in Mida Creek, Kenya: A Compromised or Journal of Sustainable Development and World Ecology Secured Future? Ambio 31: 562-568. 18: 523-529. Kaplan, R. 1983. The analysis of perception via Long, F., Paez, A. & Farber, S. 2007. Spatial effects in preference: a strategy for studying how the environment hedonic price estimation: a case study in the city of is experienced. Landscape and Urban Planning 12: Toronto. Working Paper 020, Centre for Spatial Analysis 161–176. (CSpA), McMaster University. Kareiva, P., Tallis, H., Ricketts, T. H., Daily, G. C. & Polasky, Loxton, Venn & Associates (Botswana) (Pty) Ltd. 1986. S. 2011. Natural Capital: Theory and practice of mapping Feasibility study for a large scale afforestation project ecosystem services. Oxford University Press Inc., New in Botswana. Gaborone: Botswana Development York. Corporation Ltd, 122 pp. Kaschula, S. A. & Shackleton, C. M. 2009. Quantity and Luoga, E. J., Witkowski, E. T. F. & Balkwill, K. 2000. significance of wild meat off-take by a rural community Subsistence use of wood products and shifting in the Eastern Cape, South Africa. Environmental cultivation within a miombo woodland of eastern Conservation 36: 192–200. Tanzania, with some notes on commercial uses. South African Journal of Botany 66: 72 – 85. Kolstad, C. D. 2011. Intermediate Environmental Economics. Oxford University Press, Inc. New York. Magnussen, S. & Reed, D. 2004. Knowledge reference for national forest assessments - modeling for estimation Kong, F., Yin, H. & Nakagoshi, N. 2007. Using GIS and and monitoring. FAO Forestry. Downloaded from www. landscape metrics in the hedonic price modelling of fao.org/forestry January 2015. the amenity value of urban green space: a case study in Jinan City, China. Landscape and Urban Planning 79: 240- Mahan, B. L., Polasky, S. & Adams, R. M. 2000. Valuing 252. urban wetlands: a property price approach. Land Economics 76: 100-113. Kroeger, T. 2008. Open space property value premium analysis. Conservation Economic Program. National Mander, M. 1998. Marketing of indigenous medicinal Council for Science and the Environment 2006 Wildlife plants in South Africa: a case study in KwaZulu-Natal. Habitat Policy Research Program. Food and Agriculture Organisation of the United Nations. LaFranchi, C. 1996. Small-scale and subsistence use Mann, B.Q., Beckley, L.E. & Van Der Elst, R.P. 1996. of natural resources in Namibian communal areas. Evaluation of linefishery participation and management Windhoek: World Wildlife Fund (US), Living in Finite along the KwaZulu-Natal coast. Unpublished Report, Environment (LIFE) Programme, 93 pp. Oceanographic Research Institute, South Africa 134: 17 pp. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 105 Mann, B.Q., Scott, G.M., Mann-Lang, J.B., Brouwer, S., Nordhaus, W. D. 2017. Revisiting the social cost of Lamberth, S.J., Sauer, W.H.H. & Erasmus, C. 1997. An carbon. Proceedings of the National Academy of Science evaluation of participation in and management of the 144 (7): 1518-1523. South African spearfishery. South African Journal of Marine Science 18: 179-193. Ntshona, Z. 2002. The contribution of communal rangelands to rural livelihoods in the Maluti District: McConnell, V. & Walls, M. 2005. The value of open Valuation of fuelwood. In Benjaminsen, TA, Cousins, B space: evidence from studies of nonmarket benefits. & Thompson, L (eds.). Contested resources: Challenges Resources for the Future Report. Washington, DC. to the governance of natural resources in southern Africa. Cape Town: Programme for Land and Agrarian McDonald, R.I. & Shemie, D. 2014. Urban Water Studies (PLAAS), School of Government, University of the Blueprint: Mapping conservation solutions to the global Western Cape, pp 317–327. water challenge. The Nature Conservancy: Washington, D.C. O’Farrell, P.J., De Lange, W. J., Le Maitre, D. C. et al. 2011. The possibilities and pitfalls presented by a pragmatic Melin, A., Rouget, M., Midgley, J. & Donaldson, J. S. approach to ecosystem service valuation in an arid 2014. Pollination ecosystem services in South African biodiversity hotspot. Journal of Arid Environments 75: agricultural systems. South African Journal of Science 612–623. 110: 9 pp. Owusu-Edusei, K. & Espey, M. 2003. Does proximity to a Milcu, A. I., Hanspach, J., Abson, D. & Fischer, J. 2013. golf course matter? Working Paper 012203. Department Cultural ecosystem services: a literature review and of Agricultural and Applied Economics, Clemson prospects for future research. Ecology and Society University. South Carolina Agriculture and Forestry 18(3):44. Research. Millennium Ecosystem Assessment, 2003. Ecosystems Palmer, B., van der Elst, R. & Parak, O. 2011. and human well-being: a framework for assessment. Understanding our Coast: A synopsis of KwaZulu-Natal’s Millennium Ecosystem Assessment. Island Press, coastal zone. KwaZulu-Natal Department of Agriculture, Washington D.C. Environmental Affairs and Rural Development, Cedara, Pietermaritzburg. 32 pp. Millennium Ecosystem Assessment, 2005. Ecosystems and Human well-being: Biodiversity Synthesis. World Panoramio website. 2015. Panoramio acceptance policy Resources Institute, Washington, DC. for Google Earth and Google Maps. Available: http:// www.panoramio.com/help/acceptance policy. Accessed Mmopelwa, G., Blignaut, J. N. & Hassan, R. 2009. 2015 August 3. Direct use values of selected vegetation resources in the Okavango Delta Wetland. South African Journal of Pearce, D., Atkinson, G. & Mourato, S. 2006. Cost-Benefit Economic and Management Sciences 12: 242-255. Analysis and the Environment: Recent Developments. Organisation for Economic Co-Operation and Murray, B. C., Jenkins, W. A., Sifleet, S., Pendleton, L. & Development (OECD). Baldera, A. 2010. Payments for Blue Carbon: Potential for Protecting Threatened Coastal Habitats. Policy Brief, Perman, R., Ma, Y., Common, M., Maddison, D. & NI PB 10-05. Nicholas Institute, Duke University. McGilvray, J. 2011. Natural Resource and Environmental Economics. Fourth Edition. Pearson Education Limited. Nedkov, S. & Burkhard, B. 2012. Flood regulating ecosystem services – mapping supply and demand, in Petersen, L. M., Moll, E. J., Collins, R. & Hockings, M. the Etropole municipality, Bulgaria. Ecological Indicators T. 2012. Development of a compendium of local, wild- 21: 67-79. harvested species used in the informal economy trade, Cape Town, South Africa. Ecology and Society 17: 26. Nicholls, S. & Crompton, J. L. 2007. The impact of a golf course on residential property values. Journal of Sport Pimentel, D., C. Harvey, P. Resosudarmo, K. Sinclair, D. Management 21: 555-570. Kurz, M., McNair, S. Crist, L. Shpritz, L. Fitton, R. Saffouri, and R. Blair. 1995. Environmental and economic costs Nkonya, E., Gicheru, P., Woelcke, J., Okaba, B., Kilambya, of soil erosion and conservation benefits. Science 267: D. & Gachimbi, L. N. 2008. On-site and off-site long term 1117–1123. economic impacts of soil fertility management practices: the case of maize based cropping systems in Kenya. Rangeti, I. 2014. Determinants of Key Drivers for Potable International Food Policy Research Institute (IFPRI) Water Treatment Cost in uMngeni Basin. Master of Discussion Paper 00778. Technology Thesis. Environmental Health in the Faculty of Health Sciences, Durban University of Technology. Page 106 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Rawls, W.J., Brakensiek, D. L. & Miller, N. 1983. Green- Shackleton, C, Hassan, R, De Wit, M, Shackleton, S & Ampt infiltration parameters from soils data. Journal of Beukman, R. 2002a. Natural woodlands and forests: Hydaulic Engineering 109: 62-70 Contribution to national income and economic welfare. In Hassan, R (ed.). Accounting for stock and flow Reid, H., Sahlén, L., Stage, J. & MacGregor, J. 2007. The values of woody land resources: Methods and results economic impact of climate change in Namibia. How from South Africa. Pretoria: Centre for Environmental climate change will affect the contribution of Namibia’s Economics and Policy (CEEPA), University of Pretoria, pp natural resources to its economy. International Institute 53–72. for Environment and Development. Discussion Paper 07- 02, November 2007. Shackleton, C. & Shackleton, S. 2004. The importance of non-timber forest products in rural livelihood security Raudsepp-Hearne, C., Peterson, G.D. & Bennett, E.M. and as safety nets: a review of evidence from South 2010. Ecosystem service bundles for analyzing tradeoffs Africa. South African Journal of Science 100: 658 – 664. in diverse landscapes. Proceedings of the National Academy of Sciences of the United States of America. Shackleton, C. M. 1993. Fuelwood harvesting and http://www.pnas.org/content/107/11/5242.full.pdf. sustainable utilisation in a communal ggrazing land and protected area of the eastern Transvaal lowveld. Roberts, D. 2008. Thinking globally, acting locally – Biological Conservation 63: 247 – 254. institutionalizing climate change at the local government level in Durban, South Africa. Environment and Shackleton, C. M., Shackleton, S.E., Ntshudu, M. & Urbanization 20(2): 521–537. Ntzebeza, J. 2002b. Direct use values of non-timber forest products from three rural villages in the Kat River Roberts, D. & O’Donoghue, S. 2013. Urban Valley, Eastern Cape Province, South Africa. Journal of environmental challenges and climate change action in Tropical Forest Products 8: 45-65. Durban, South Africa. Environment and Urbanisation 25(2): 299 – 319. Shackleton, C. M., Timmermans, H. G., Nongwe, N., Hamer, N. & Palmer, R. 2007. Direct-use values of non- Roberts, D., Seppings, K., Voortman, BN., Harigobin, timber forest products from two areas on the Transkei S., English, M., Nichols, G., Mander, M. 1999 Durban Wild Coast. Agrekon 46: 135 – 156. Metropolitan Open Space System Framework Plan. Shrestha, R.K., Stein, T.V. & Clark, J. 2007. Valuing Rooseboom, A. 1975. Sediment Transport in Rivers and nature-based recreation in public natural areas of Reservoirs. D Eng dissertation, University of Pretoria the Apalachicola River region, Florida. Journal of (in Afrikaans). Later published by the Water Research Environmental Management 85: 977-985. Commission as Report No 297/1/92, 1992 (in English). Smith, V.K., Poulos, C. & Kim, H. 2002. Treating open Rosen, S. 1974. Hedonic prices and implicit markets: space as an urban amenity. Resource and Energy product differentiation in pure competition. The Journal Economics 24: 107–129. of Political Economy 82: 34-55. Smithers, J.C. & Schulze, R.E, 2002. Design Rainfall and Sauer, W.H.H., Penney, A.J., Erasmus, C., Mann, B.Q., Flood Estimation in South Africa. School of Bioresources Brouwer, S.L., Lamberth, S.J. & T.J. Stewart. 1997. An Engineering and Environmental Hydrology, University of evaluation of attitudes and responses to monitoring and Natal, Pietermaritzburg, South Africa. WRC Project No: management measures for the South African boat-based K5/1060 linefishery. South African Journal of Marine Science 18: 147–163. Smithers, J.C., 2002. The Estimation for Design Rainfall for Durban Unicity. Draft Report to eThekwini Schulze, R.E., Knoesen, D.M., Kunz, R.P. & van Niekerk, Municipality. School of Bioresources Engineering L.M. 2010. Impacts of protected climate change on and Environmental Hydrology. University of Natal, design rainfall and streamflows in the eThekwini Metro Pietermaritzburg, South Africa. Area. ACRUcons Report 62. University of KwaZulu-Natal. South African Frog Re-assessment Group (SA-FRoG) Sen, A., Darnell, A., Crowe, A., Bateman, I., Munday, & IUCN SSC Amphibian Specialist Group. 2012. P. & Foden, J. 2011. Economic Assessment of the Natalobatrachus bonebergi. The IUCN Red List of Recreational Value of Ecosystems in Great Britain. Threatened Species. Version 2015.2. www.iucnredlist. Report to the Economics Team of the UK National org. Downloaded on 21 July 2015. Ecosystem Assessment. The Centre for Social and Economic Research on the Global Environment (CSERGE), University of East Anglia. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 107 South African Frog Re-assessment Group (SA-FRoG), Turpie, J.K., de Wet, J., Clark, B., Orr, K., Brown, C., IUCN SSC Amphibian Specialist Group. 2010. Hemisus Joubert, A., Reinecke, K., Magoba, R., Beuster, H., Clarke, guttatus. The IUCN Red List of Threatened Species. T. & Kotze, D. 2010. The nature, distribution and value of Version 2015.2. www.iucnredlist.org. Downloaded on 21 aquatic ecosystem services of the Olifants, Inkomati and July 2015. Usutu to Mhlatuze Water Management Areas. Contract Report for Department: Water Affairs and Forestry, South African Tourism (SAT) 2014. 2014 Statistics of our 378pp. Tourism Sector, KwaZulu-Natal. Turpie, J.K. & Marx, D. 2012. Economic value of the Spatial Development Framework (SDF) 2014. eThekwini Uranium mining landscape in Namibia. Report to Flora Municipality. & Fauna International, on behalf of the Namibian Government. Stats SA. 2011a. Understanding our people: a journey through Census 2011. eThekwini Municipality. Turpie, J.K., Warr, B., & Ingram, J.C. 2014. The economic value of Zambia’s forest ecosystems and potential Stats SA. 2011b. Census 2011 Municipal Report KwaZulu- benefits of REDD+ in Green Economy transformation in Natal. Report No. 03-01-53. Zambia. UNEP Report. 142pp. SWWM, 2009. Stormwater Management Model Turpie, J.K., Day, L., Letley, G., Chrystal, R. & Corbello, S. Applications Manual. Report: EPA/600/R-03/000. United 2016. The value of Durban’s natural capital and its role States. in Green Urban Development. Part 2: Scenario Analysis, evaluating the potential returns to investing in Green Tarrant, J. & Armstrong, A. J. 2013. Biodiversity Urban Development. World Bank Report. management plan for the Pickersgill’s Reed Frog: Background information for stakeholders workshop. Twine, W. C., Moshe, D., Netshiluvhi, T. R. & Siphungu, Endagered Wildlife Trust (EWT) and Ezemvelo KwaZulu- V. 2003. Consumption and direct-use values of savanna Natal Wildlife. bioresources used by rural households in Mametja, a semi-arid area of Limpopo Province. South African TEEB, 2010. The Economics of Ecosystems and Journal of Science 99: 467-473. Biodiversity Ecological and Economic Foundations. Edited by Pushpam Kumar. Earthscan, London and Tyrvӓinen, L. & Vӓӓnӓnen, H. 1998. The economic Washington. value of urban forest amenities: an application of the contingent valuation method. Landscape and Urban Telles, T. S., Dechen, S. C. F., de Souza, L. G. A., de Fatima Planning 43: 105-118. Guimaraes, M. 2013. Valuation and assessment of soil erosion costs. Scientia Agricola 70 (3): 209 – 216. Tyrvӓinen, L. & Miettinen, A. 2000. Property prices and urban forest amenities. Journal of Environmental Timilsina, N., Escobedo, F. J., Staudhammer, C. L. & Economics and Management 39: 205-223. Brandeis, T. 2014. Analyzing the causal factors of carbon stores in a subtropical urban forest. Ecological Ulrich, R.S. & Addoms, D.L. 1981. Psychological and Complexity 20: 23–32. recreational benefits of a residential park. Journal of Leisure Research 13: 43-65. TNC, 2015. Upper Tana-Nairobi Water Fund Business Case. The Nature Conservancy, Nairobi, Kenya. Van Herzele, A. & Wiedemann, T. 2003. A monitoring tool for the provision of accessible and attractive urban Tol, R.S.J. 2012. On the Uncertainty About the Total green spaces. Landscape and Urban Planning 63: 109– Economic Impact of Climate Change. Environmental 126 Resource Economics 53: 97–116. Vellidis, G., Lowrance, R., Gay, P. & Hubbard, R. K. 2003. Treg, C. 2010. A multilevel property hedonic approach Nutrient transport in a restored riparian wetland. Journal to valuing parks and open space, Graduate College of Environmental Quality 32:711–726. Dissertations and Theses, University of Vermont, Paper 230. Wallace, J.H. 1975a. The estuarine fishes of the east coast of South Africa. I. Species composition and length Turpie, J.K., O’Connor, T. Mills, A. Robertson, H. 2007. distribution in the estuarine and marine environments. The ecological and economic consequences of changing II. Seasonal abundance and migrations. Durban, South land use in the southern Drakensberg grasslands, Africa: Investigational Report, Oceanographic Research South Africa. South African Journal of Economic and Institute, 40: 1-72. Management Sciences 10: 423 – 441. Page 108 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Wallace, J.H. 1975b. The estuarine fishes of the east coast of South Africa. III. Reproduction. Durban, South Africa: Investigational Report, Oceanographic Research Institute 41: 1-51. Whitfield, A.K. 1998. Biology and Ecology of Fishes of Southern African Estuaries. Ichthyological Monograph of the J.L.B.Smith Institute of Ichthyology 2: 223. Whitfield, A.K. 1992. A characterization of Southern African estuarine systems. South African Journal of Science 18: 89 - 103. Whitfield, A.K. & Marais, J.F.K. 1999. The ichthyofauna. In Estuaries of South Africa Allanson, B.R., Baird, D. (eds.). Cambridge: Cambridge University Press, pp. 209- 233. Willis, K.G. & Benson, J. F. 1989. Recreational Values of Forests. Forestry 62: 93-110. Wilson, B., Valentine, L., Kinloch, J., Sonneman, T. & Swinburn, M. 2009. Habitat Loss and Fragmentation. In Wilson, B. & Valentine, L. (eds.) Biodiversity values and threatening processes of the Gnangara groundwater system. Department of Environment and Conservation, Western Australia. Water Research Commission, 1992. The Development of the New Sediment Yield Map of Southern Africa. WRC Report No 297/2/92. Compiled by Sigma Beta Consulting Engineers. Water Research Commission, 2002. State-of-Rivers Report: uMngeni River and neighbouring rivers and streams. WRC report no.TT 200/02. Water Research Commission, Pretoria. ISBN No: 1 86845 899 7. Wood, S. A., Guerry, A. D. & Lacayo, M. 2013. Using social media to quantify nature-based tourism and recreation. Scientific Reports 3: 2976. www.nature.com/ scientificreports Yoo, W., Mayberry, R., Bae, S., Singh, K., He, Q. & Lillard, Jr., J.W. 2014. A study of effects of multicollinearity in the multivariable analysis. International Journal of Applied Science and Technology 4(5): 9-19. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 109 This page intentionally blank. Page 110 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY APPENDIX 1. ECOSYSTEMS AND BIODIVERSITY OF THE EMA A1.1 Terrestrial ecosystems private collections and museums (EPCPD 2012). More information is available for plants and birds which have been studied in more detail across the EMA. A1.1.1 Introduction This section describes the extent, biodiversity attributes, location and abundance of natural resources found in A1.1.2 Vegetation types terrestrial ecosystems across the EMA. There is very A variety of forest, woodland, bush and grassland limited historical research for a number of terrestrial habitats are found in the EMA (Table A1. 1). Almost taxa, such as invertebrates, mammals, amphibians all of these have been significantly transformed, are and reptiles, with most studies focusing on a few focal threatened and in need of protection, with several being taxa and records of the distribution and occurrence of critically endangered (Table A1. 1; SDF 2014, SANBI; species within the EMA being incidental coming from EPCPD 2012). Table A1.1 The vegetation types (KwaZulu-Natal and SA) found within the EMA and their conservation status as used by eThekwini Municipality and SANBI (Source: EPCPD Spatial Conservation Plan 2012) SA Vegetation Type KwaZulu-Natal Conservation Conservation Status Vegetation Type Status (eThekwini (National) Municipality) Dry Ngongoni Veld Endangered Ngongoni Veld Vulnerable Moist Ngongoni Veld Critically Endangered Eastern Valley Bushveld Eastern Valley Bushveld Near Threatened Least Threatened KwaZulu-Natal Hinterland KwaZulu-Natal Hinterland Thornveld Vulnerable Vulnerable Thornveld KwaZulu-Natal Sandstone KwaZulu-Natal Sandstone Sourveld Critically Endangered Endangered Sourveld North Coast Bushland Critically Endangered South Coast Bushland Endangered KwaZulu-Natal Coastal Belt Endangered North Coast Grassland Critically Endangered South Coast Grassland Critically Endangered Southern Coastal Scarp Scarp Forest Vulnerable Least Threatened Forest KwaZulu-Natal Coastal Endangered Forest Northern Coastal Forest Least Threatened KwaZulu-Natal Dune Critically Endangered Forest Mangrove Forest Mangrove Forest Critically Endangered Critically Endangered Voacanga thouarsii Swamp Forest Endangered Critically Endangered Swamp Forest A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 111 As part of the eThekwini Municipality Spatial have been transformed in the EMA respectively, and Conservation Plan the national, regional and local of the remaining untransformed grassland 52% of it extent (ha) of the different vegetation types were is considered to be degraded (Table A1. 2). General listed with estimated original extents, current levels of descriptions for each vegetation type are given transformation and remaining areas within the EMA below and include information about extent, levels classified according to their condition as assessed for of transformation and land use based on data and the EMSCP (Table A1. 2). Table A1. 2 highlights the information presented in the EMSCP (2012). extent of transformation for each vegetation type and the levels of degradation within the EMA. For example, 77% and 72% of the North and South Coast Grassland Table A1.2 The vegetation types (KwaZulu-Natal and SA) found within the EMA and their conservation status as used by eThekwini Municipality and SANBI (Source: EPCPD Spatial Conservation Plan 2012) Original Extent (ha) Transformed % Condition of untransformed (EMA) KwaZulu- EMA EMA (%) KwaZulu- EMA Good Inter. Degr. Natal Natal (%) (%) Dry Ngongoni 268 024 18 109 7 73 33 69 7 23 Veld Moist Ngongoni 442 424 12 394 3 83 31 33 9 58 Veld Eastern Valley 291 207 20 080 2 ? 19 68 6 26 Bushveld KwaZulu-Natal Hinterland 113 341 6 824 6 ? 32 80 3 16 Thornveld KwaZulu-Natal Sandstone 160 819 15 681 10 ? 67 61 10 29 Sourveld North Coast 88 811 32 758 37 79 44 32 15 53 Bushland South Coast 89 103 1 953 2 71 37 12 33 55 Bushland North Coast 291 877 82 979 28 95 77 22 27 52 Grassland South Coast 153 568 24 184 16 94 72 19 29 52 Grassland Southern Coastal 33 750 8 878 26 NA NA 85 14 1 Scarp Forest KwaZulu-Natal 21 089 2 193 10 NA NA 59 36 4 Coastal Forest KwaZulu-Natal 12 396 1 287 10 NA NA 80 12 9 Dune Forest Mangrove Forest 2 297 65 3 NA NA 100 0 0 Voacanga Swamp 3 022 55 2 NA NA 87 6 7 Forest Page 112 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Ngongoni Veld KwaZulu-Natal Sandstone Sourveld Within the EMA Ngongoni Veld is found inland from This vegetation type is typically species rich with high the upper uMngeni catchment in the north to the levels of forb diversity and high rates of plant endemism upper iLovu catchment in the south and can be found (EPCPD 2012). Woody species are generally fire tolerant at elevations of 180-870 metres above sea level. This such as Protea species and low shrubs. Within the ecosystem is characteristically species-poor due to the EMA KwaZulu-Natal Sandstone Sourveld tends to be dominance of the tall grass Aristida junciformis. Within found in insular pockets from Inanda Mountain in the KwaZulu-Natal there are two vegetation types found north to Umbumbulu in the south, usually between within this group, namely Dry and Moist Ngongoni 130-870 metres above sea level (EPCPD 2012). This Veld. Regionally both of these ecosystems are highly vegetation type is classified as Critically Endangered and transformed and of the remaining extent of the has experienced significant transformation across its Ngongoni Veld in the EMA just more than one third has range with more than two thirds of its original extent undergone some form of transformation (EPCPD 2012). transformed. The remaining areas of KwaZulu-Natal Sandstone Sourveld continue to be under further Of the untransformed areas of Ngongoni Veld in the threat with 29% in a degraded state as a result of EMA, 60% is located in traditional authority areas, 18% unmanaged fire regimes and the spread of alien invasive is under formal town planning schemes and 22% is plant species (EPCPD 2012). Only 1% of the original designated for non-scheme agricultural areas (EPCPD distribution of this vegetation type is formally protected 2012). There are no Ngongoni Veld areas that are in the EMA in Krantzkloof Nature Reserve (EPCPD 2012). formally protected in the EMA (EPCPD 2012). KwaZulu-Natal Coastal Belt Eastern Valley Bushveld The KwaZulu-Natal Coastal Belt exists in a dissected and This vegetation type is generally confined to deeply rolling landscape which in the past would have consisted incised river valleys with plant communities dominated of a mix of coastal forest, grassland and woodland plant by semi-deciduous woody species and succulent species. communities. Within the EMA this habitat is located Within the EMA Eastern Valley Bushveld is confined to along the entire coastal plain between sea level and the upper reaches of the uMngeni and Umlaas Rivers 660 metres above sea level. There are four vegetation occurring between 100-800 metres above sea level. types classified within the KwaZulu-Natal Coastal Belt, The areas where this vegetation is predominantly found namely North Coast Grassland, South Coast Grassland, are steep resulting in low levels of transformation with South Coast Bushland and North Coast Bushland. All less than a third having been transformed and of the of these vegetation types are found in the most highly remaining habitat the majority still in a good condition transformed areas of KwaZulu-Natal and as a result have (EPCPD 2012). Of the untransformed areas of Eastern become highly impacted and threatened. North and Valley Bushveld in the study area 99% is found in South Coast Grassland are considered to be particularly traditional authority areas and only 1% is covered by important for the EMA as more than two-thirds of their formal planning schemes. There are no areas of this original extent was located within the boundaries of vegetation that are formally protected in the EMA the EMA (EPCPD 2012). Untransformed areas have also (EPCPD 2012). become highly degraded and there are very few original patches remaining. More than half of the remaining KwaZulu-Natal Hinterland Thornveld areas are located in traditional authority areas (52%) KwaZulu-Natal Hinterland Thornveld is usually situated and the rest in formal planning schemes (30%) and non- upslope from Eastern Valley Bushveld in steep river scheme agricultural areas (18%). Within the EMA only valleys and is intermediate in its characteristics between North Coast Bushland and North Coast Grassland are this type of vegetation and Ngongoni Veld where there formally protected, and these areas make up less than are exchanges of common biota (EPCPD 2012). It is 1% of their original extent in the EMA (EPCPD 2012). characterised by open thornveld dominated by Acacia species. Within the EMA this vegetation is limited to Scarp Forest the upper catchments of the uMngeni and Umlaas Scarp Forests are species rich, have high levels of Rivers occurring between 300-825 metres above sea endemicity and tend to be found in steep gorges level. It is estimated that roughly one third of KwaZulu- and scarps. They are transitional in nature between Natal Hinterland Thornveld has been transformed Afromontane and Coastal Forests characterised by with 80% of the remaining areas being in a good species from both of these vegetation types. Scarp condition. The steepness of the localities and lack of Forest is structurally tall with high basal area and woody any vehicular access contributes preventing any further stems, low levels of multi-stemming and a poor herb transformation. Untransformed areas of this type of layer (EPCPD 2012). Within the EMA these forests are vegetation are located predominantly in traditional highly fragmented but occur predominantly inland of authority areas (68%) and some in non-scheme 4km from the coast between 300-875 metres above sea agricultural areas (30%). Only a small portion (2%) is level. under formal planning schemes. There are no areas of this vegetation that are formally protected in the EMA (EPCPD 2012). A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 113 Scarp Forest has been highly exploited in the EMA for the area and the associated diversions of the Umlaas natural resources such as medicinal plants and various and Isipingo rivers (EPCPD 2012). All three remaining building materials. However, there are still some mangrove stands fall within the formal planning good quality patched remaining in areas that are less scheme zoned for open space (62%) and utility (22%). accessible (EPCPD 2012). The remaining Scarp Forests Approximately 72% of mangroves are formally protected are located in traditional authority areas (44%), formal in the EMA in the Beachwood Mangroves Nature planning schemes (35%) and non-scheme agricultural Reserve (EPCPD 2012). areas (21%). Of the remaining forest, only 6% (531 ha) is formally protected in the Krantzkloof, North Park, Swamp Forest Kenneth Stainbank and Palmiet nature reserves (EPCPD Swamp Forests are grow in waterlogged soils and are 2012). medium height forests with well-developed canopy layers and a poorly developed understorey layer (EPCPD Northern Coastal Forest 2012). Within the EMA Swamp Forest is located in only Coastal Forests are medium to tall forests with low basal a few small fragments in four localities in the central area of woody stems, high levels of multi-stemming and southern planning regions occurring between 140- and a well developed herb layer (EPCPD 2012). In the 530 metres above sea level (EPCPD 2012). Only a small EMA these forests are located along the entire coastline portion (55ha) is found within the EMA and surprisingly on the undulating coastal plain from the dunes to most of the areas are reported to be in a good condition approximately 580 metres above sea level. Northern despite occurring in very populated regions of the Coastal Forest is made up of two forest types; KwaZulu- EMA. Most of the remaining Swamp Forest is situated Natal Coastal Forest and KwaZulu-Natal Dune Forest. in traditional authority areas (71%) and commercial KwaZulu-Natal Dune Forest occurs largely along the farming areas (25%) and only a small portion (4%) is in border of the dunes beyond the salt spray zone (EPCPD formal planning schemes. The latter being Glenholme 2012) and has become highly fragmented as a result Private Nature Reserve that is zoned for open space of coastal development. It is classified as Critically (EPCPD 2012). Endangered. Approximately 60% of the remaining Dune Forest area is in formal planning schemes (open space, residential and amenity), 34% in non-scheme agricultural A1.1.3 Plants areas and 6% in traditional authority areas. Only 2% There is high plant diversity within the EMA with a of Dune Forest is formally protected in the Umhlanga total of 2267 species recorded from 204 different plant Lagoon Nature Reserve and Beachwood Mangroves families (EPCPD 2012). This represents more than half Nature Reserve. Of the remaining untransformed the known families found in South Africa. Endemicity is KwaZulu-Natal Coastal Forest, 56% is within formal also high in the EMA with 379 species, 16% of the total, planning scheme areas (residential, open space and being classified as South African endemics (EPCPD 2012). utility), 28% is in non-scheme agricultural areas and 16% Biogeographically the plant species in the EMA are made is in traditional authority areas. Only about 1% of the up of a large tropical and subtropical component, a Coastal Forest is formally protected in the Happy Valley smaller temperate component and a localised endemic Nature Reserve. group (EPCPD 2012). Mangrove Forest A total of 180 plant species significant for conservation Mangroves are typically limited to the intertidal areas of in the EMA were assessed during the EMSCP process and permanently open estuaries and are species poor usually of these 31 were classified as threatened, including 11 consisting of one to three dominant species of mangrove Endangered and seven Critically Endangered species. Of tree (EPCPD 2012). Mangroves are able to tolerate a those that are Critically Endangered only three species wide range of conditions that they are subjected to in are known to have existing populations in the EMA the intertidal zone. The forests are structurally simple with the remaining four species not recorded in the with a limited vertical variation when compared to EMA for decades and assumed to be locally or globally terrestrial forest types. In the EMA mangrove stands extinct (EPCPD 2012). A significant proportion of the are located in three estuaries; Isipingo, Durban Bay plant species are classified as being rare and although (Bayhead) and uMngeni (Beachwood). Approximately they are not yet threatened, localised and unregulated 97% of mangrove area in Durban Bay has been lost since pressures in the EMA could impact on these species the beginning of the twentieth century and the current too. The rare status of these species has been driven by extent in this estuarine bay is 14 ha. Similarly in Isipingo habitat displacement, habitat degradation and the direct there has been a significant reduction in mangroves exploitation of some species for medicinal purposes, as a result of residential and industrial development in fuelwood and construction purposes (EPCPD 2012). Page 114 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY A1.1.4 Mammals ƒƒ Southern African seasonal migrants, mostly of a more A total of 82 mammal species have been recorded in the temperate origin, including altitudinal migrants and EMA and of these three are alien and five of the larger migrants from the south western regions of South mammals have been reintroduced in the area (EPCPD Africa; 2012). Historically the EMA was inhabited by a number ƒƒ African migrants from the tropical regions of the of large mammal species that are now locally extinct, continent; and such as elephant, lion, buffalo, leopard, hippopotamus and spotted hyena (EPCPD 2012). Detailed mammal data ƒƒ Palearctic migrants from Eurasia. in the EMA is limited, however the EPCPD (2012) Spatial Conservation Plan outlined the current status of some Of the bird species assessed and used in the EPCPD of the known mammal species and the most common (2012) Spatial Conservation Plan 12 were classified threats to these species. as threatened including two Endangered species, the Spotted Ground-Thrush and the Black-rumped Threatened species included two bats, Kerivoula Buttonquail. The Spotted Ground-Thrush is a forest argentata and Otomops martiensseni that were specialist that migrates into southern KwaZulu-Natal classified as Endangered and Vulnerable respectively. from the Eastern Cape over the winter months and the Kerivoula argentata, the Damara woolly bat, is forests within the EMA make up a significant proportion insectivorous and has its southern most distribution of its non-breeding habitat (EPCPD 2012). Rates of records in the EMA, however record are outdated and endemism in the selected species were found to be low its current extent is not known (EPCPD 2012). Otomops with only two species being near endemics to South martiensseni, the large-eared free-tailed bat has a highly Africa; the Bush Blackcap, a non-breeding migrant, disjointed population within the EMA with its nearest and the Knysna Turaco. There are four species that known populations being in Zimbabwe and Madagascar have their southernmost distributions within the EMA and as a result has been recognised as an evolutionary and a further two species that have distributions that significant species in the EMA (EPCPD 2012). There terminate just south of the EMA boundary (EPCPD are two threatened species of antelope that occur 2012). A number of birds found within the EMA are in the EMA, the Oribi, Ourebia ourebia, and the Blue indicators of ecosystem health, the majority of these are Duiker, Philantomba monticola, which are classified as waterbird species that are found only in a particular type Endangered and Vulnerable respectively. Both these or condition of habitat that they use. There are specific antelope species are under significant pressure in the raptors and owls that are also used to assess ecosystem EMA as a result of fragmented populations, illegal health due to their high trophic position, including the hunting and other anthropogenic influences (EPCPD Fish Eagle, African Marsh Harrier, Martial Eagle, and 2012). There are five bat species and two rodent species Lanner and Peregrine Falcons (EPCPD 2012). that have their southernmost distribution records in the EMA with one species in particular the Angoni vlei rat, Otomys angoniensis, is considered an indicator of A1.1.6 Reptiles ecosystem health due to its association with wetlands (EPCPD 2012). Other mammals found in the EMA include A total of 69 reptile species have been recorded mice, rats, servals, weasels and shrews. within the EMA including 26 lizards, 42 snakes and two terrapins (EPCPD 2012). However, many of the records for these species predate the 1990’s and the lack of A1.1.5 Birds current up to date data means many of these species may no longer occur in the EMA. One snake species, a A total of 526 bird species have been recorded in the harmless blind snake, Rhamphotyphlops braminus is an EMA (EPCPD 2012) and of these at least 13 species alien invasive species unintentionally introduced into have gone locally extinct, including Saddle-billed the EMA some time before 1985 (EPCPD 2012). There Stork, Lesser Flamingo, Greater Flamingo, Black Heron, are two species classified as threatened, the Natal Chestnut-banded Plover, Southern Banded Snake- rock python which is Vulnerable and Smith’s dwarf Eagle, Swamp Nightjar, Eastern Bronze-naped Pigeon, burrowing skink which is Endangered. Smith’s dwarf Ground Woodpecker, Black Coucal, Blue Swallow, Knysna burrowing skink has a very restricted distribution and Warbler and Rosythroated Longclaw. Most of these is endemic to the EMA, found only from the mouth local extinctions are a result of habitat degradation of the uMngeni River south to Amanzimtoti covering and transformation (EPCPD 2012). The EPCPD (2012) approximately 431km2 (Broadley 2010). It is found in broadly classified the bird species into the following sandy soils of coastal thicket, grassland and dune forest biogeographic groups: from Berea Red Sands from the Canelands in the north to Clansthal in the south, and only 10% of its known ƒƒ A tropical group of widespread species that are mostly habitat is formally protected (Broadley 2010, EPCPD resident throughout the year; 2012). Endemic to KwaZulu-Natal and with most of its A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 115 range in the EMA, the black-headed dwarf chameleon, A1.2 Rivers and Wetlands Bradypodion melanocephalum, is a flagship species for the EMA and is predominantly found in coastal A1.2.1 Introduction habitat (EPCPD 2012). The ranges of both Smith’s dwarf burrowing skink and the black-headed dwarf chameleon Rivers and wetlands provide a wide range of ecosystem intersect some of the most developed coastal areas of goods and services such as drinking water, fish, flood the EMA where habitat transformation and degradation attenuation and water purification. There is little is high (Broadley 2010, EPCPD 2012). information about natural resource use in freshwater systems within the EMA, however there is information Other notable species include those that reach their about the goods and services provided by wetlands and southernmost distribution in the EMA; the Natal purple rivers and use of these in other areas of KwaZulu-Natal. glossed snake and the Nyasa file-snake and two species There is also a general lack of fine scale biodiversity that are indicators of ecosystem health within the EMA; data for the rivers and wetlands of the EMA (EPCPD the Natal black water-snake (riverine) and the green 2012) and most of the available information has focused mamba (dune forest) (EPCPD 2012). around the health of the river systems and habitat type data collection. Ground Truth (2006) conducted a study on the state of the rivers in the EMA providing health A1.1.7 Amphibians status data for 33 rivers and streams. The health status There are a total of 37 amphibian species recorded for each river was assessed based on bioindicators such in the EMA, however five of these species have not as habitat integrity, water quality, plant communities been recorded since 1996 and are considered to be and invertebrate assemblages. The rivers and locally extinct (EPCPD 2012). Amphibians are very wetlands of the EMA fall into the Mvoti – Umzimkulu sensitive to changes in their surroundings, especially to Water Management Area (WMA) and the status quo anthropogenic influences in urban environments such as assessment and delineation analysis report (DWA 2013a) water and air pollution, habitat degradation, isolation, provides some detailed information about the tertiary pathogens and climate change (Hamer & McDonnell catchments and their associated rivers and wetlands. 2008). Freshwater systems in the EMA are under increasing The Kloof frog, Natalobatrachus bonebergi, is listed as threat from impacts such as flow modification, sand Endangered and its numbers are thought to be further mining and pollution from industrial and sewage decreasing as a result of habitat loss, exploitation and effluent. Unplanned and extensive informal settlement pollution (SA-FRoG & IUCN 2012). This species is found development in the upper catchment areas is also having in pristine forested gorges where it is usually found along a large impact on these systems. streams, where it breeds (EPCPD 2012, SA-FRoG & IUCN 2012). Its habitat in the EMA is becoming increasingly threatened by water pollution and over-exploitation of A1.2.2 Rivers forested areas (EPCPD 2012). Another species of concern Within the EMA there are 18 river catchments and all the is the spotted shovel-nosed frog, Hemissus guttatus, rivers located in the EMA fall into the Mvoti – Umzimkulu which has its southern most records in the EMA and is WMA. This WMA consists of eight tertiary catchments listed as Vulnerable (SA-FRoG & IUCN 2010). It occurs in of which four are included in the eThekwini Municipality; wooded and open habitat adjacent to wetlands, seasonal the uMdloti, uMngeni, uMlazi and iLovu, and uMkhomazi pans, swampy areas and pools near rivers (SA-FRoG & tertiary catchments (Figure A1. 1, DWA 2013a). The IUCN 2010). main river systems found in these tertiary catchments and in the EMA are the uMkhomazi, uMngeni, uMdloti, oHlanga, uTongathi, uMlazi, Mbokodweni and iLovu. The larger river systems originate in the Drakensberg Mountains, the medium rivers in the KwaZulu-Natal Midlands and the smaller rivers close to the coast (DWA 2013a). Page 116 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure A1.1 The major and minor rivers found within the EMA A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 117 uMdloti tertiary catchment quality in the catchment is poor, especially the uMlazi The uMdloti tertiary catchment includes the uMdloti, River. uTongathi and oHlanga Rivers. There are two major dams on these rivers within the EMA; the Hazelmere Upstream of the Shongweni Dam, the uMlazi river Dam on the uMdloti River and the smaller Dudley Pringle catchment is impacted by changes to flow from farm Dam on the Tongati River. Land use in this catchment dams and irrigation and non-flow impacts such as consists mainly of dryland and irrigated sugarcane, invasive alien plants, and point source pollution from predominantly on communal land (DWA 2013a). The forestry, agriculture and settlement activities. Low small urban centres of Tongaat, Umhlanga and Verulam density settlements, semi-urban and industrial areas are located within this catchment. The water quality are located in the lower half of the upstream section of the rivers in this catchment are generally poor as a and discharges from the Hopewell and Hammersdale result of point source pollution, especially along the WWTW into the river systems contribute to the poor coastal strip which is highly developed. Impacts both water quality found in this area (DWA 2013a). The lower above and below Hazelmere Dam on the uMdloti River uMlazi is in a poor state due to degraded water quality include pollution from informal settlements, subsistence and riparian vegetation removal as a result of animal agriculture and animal grazing. grazing and wood harvesting by communities living along this stretch of river. Discharges from WWTW in this Numerous Waste Water Treatment Works (WWTW) section of river affects both the flow and the wwater discharge into the uMdloti and Tongati Rivers from quality of the river and there is a hazardous landfill site Phoenix, Umhlanga, Tongaat and the King Shaka in the upper reaches of the tributaries that affects water International Airport WWTWs (DWA 2013a) affecting quality. The lower end of the uMlazi River is canalised both the flow and water quality of the rivers. There and therefore there is no estuary (DWA 2013a). are a number of low density rural settlements spread throughout the catchment. There are no dams on the Mbokodweni River. The upper and middle reaches are dominated by some dryland uMngeni tertiary catchment sugarcane farming and are occupied by scattered rural The uMngeni River system is largely regulated and villages. The middle to lower areas are dominated by developed (DWA 2013a). This river functions as the semi-urban and urban areas (uMlazi and Isipingo) as well main source of water for the Durban to Pietermaritzburg as industrial areas by the coast. Discharges from WWTW area and is serviced by four major dams; Midmar in the lower reaches also impact water quality of this Dam, Nagle Dam, Albert Falls Dam and Inanda Dam. river negatively. However, Inanda Dam is the only one falling within the eThekwini Municipal boundary. Water quality in the The upper iLovu catchment is situated in an area with lower uMngeni is generally poor as a result of the dense extensive forestry and sugarcane and rural areas with human population in and around Durban. A vast number high density townships, all of which have contributed of informal settlements found along this river system are to the poor water quality found in these rivers. The not serviced with adequate sanitation resulting in point coastal rivers (Umsimbazi, uMgababa, Manzimtoti, source pollution contributing to poor water quality. Little Manzimtoti) associated with the iLovu catchment are also mostly in a poor state with rural settlements, The area above Inanda Dam is largely rural with some development and agricultural activities contributing to subsistence agriculture activities and dryland sugarcane poor water quality. and forestry. The lower uMngeni River below Inanda Dam is considered to be in a particularly poor state The Mhlatuzane and Umbilo Rivers, located upstream as a result of flow regulation and extensive urban and of Durban Bay, and the Sipingo River located south of industrial areas. The Palmiet River located below the Durban Bay, are highly developed and are surrounded dam is also in a poor state as a result of impacts from by industrial and urban development. The main impacts industrial and urban developments along this stretch of include sedimentation, solid waste pollution, effluent river. Discharges from WWTW into the uMngeni and run-off, alien vegetation and the removal of riparian associated tributaries, such as the Umhlangane River, vegetation. also contribute the poor health of this system. uMkhomazi tertiary catchment uMlazi & iLovu tertiary catchment Forestry and dryland sugarcane farming are the This catchment is dominated by irrigation agriculture predominant activities in this catchment. The upper and afforestation and is largely unregulated (DWA Mkomazi River is located outside of the EMA with only a 2013a). The main rivers associated with this catchment small section of the middle reaches and all of the lower are the uMlazi, iLovu, Mhlatuzana, Mbokodweni and reaches located within the EMA in the southern planning Umbilo. The Shongweni Dam on the uMlazi River has region. Low density rural settlements are found in the silted up and is now only used for recreational and middle to lower reaches and land use activities include educational purposes (DWA 2013a). Generally the water community water use, grazing, subsistence agriculture Page 118 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY and riparian vegetation removal for wood and grazing. A number of biophysical drivers shape the structure, High levels of sedimentation from forestry and composition and functioning of wetland ecosystems and agricultural activities occur throughout this catchment. the vegetation that is found in these systems. These At the mouth of the uMkhomazi River approximately 44 drivers include climate, geology, hydrology, water quality million m3 per annum of water is abstracted by SAPPI- and geomorphology. Vegetation characteristically SAICCOR (DWA 2013a). found in wetlands in the EMA include reeds such as the Common reed (Phragmites australis), rushes such as Bullrush (Typha capensis), certain species of water A1.2.3 Wetlands loving sedges, such as Giant sedge (Cyperus dives), Wetlands are highly variable systems and range from Dwarf papyrus (Cyperus prolifer), and Finger sedge permanently wet shallow lakes to periodically wet valley (Eleocharis dregeana), water loving grasses such as Wild bottoms (eThekwini Municipality 2011). In the EMA rice grass (Leersia hexandra) and Rat tail drop seed grass there are typically five hydrogeomorphic wetland types: (Sporobilis africana), certain swamp forest tree species such as Umdoni (Syzigium cordatum), Swamp fig (Ficus ƒƒ Channelled valley bottom wetlands are situated in the trichopoda) and Wild frangipani (Voacanga thouarsii) bottom of a valley, have small channels and are fed by (eThekwini Municpality 2011). ground water from slopes and overbank flooding. There is one Critically Endangered species of frog found ƒƒ Unchannelled valley bottom wetlands are situated in within the EMA. The Pickersgill’s Reed Frog is endemic the bottom of a valley but have no channel and are fed to KwaZulu-Natal coast of South Africa, with all of the by ground water and surrounding slopes. southernmost records for the species found within the EMA (EPCPD 2012, Tarrant & Armstrong 2013). This frog ƒƒ Floodplains are extensive flat areas that are usually on is found in densely vegetated (reed-bed), stagnant valley either side of a large river channel, fed by overbank bottom wetlands, which are also Critically Endangered, flooding. There are usually oxbow lakes and meander from the coast to approximately 200 metres above scars present. sea level (EPCPD 2012). The majority of the wetland habitat is located on privately or community owned land ƒƒ Hillslope feeding a watercourse. These are wet patches and is threatened by habitat degradation and coastal that are located on a hill slope of hilltop rather than in development (Tarrant & Armstrong 2013). a valley and there is a natural water course joining the system to other wetlands, streams or rivers. In the rural and traditional authority areas of the EMA local communities rely of wetlands for building and ƒƒ Hillslope not feeding a watercourse. These wetland craft materials such as sedges and reeds. They also patches are located on a hillside or hilltop that does harvest certain tree materials, such as the fruits of the not have any direct natural link to other wetlands, wild frangipani tree for medicinal purposes. Other plant streams or rivers. materials harvested include the roots and seeds of grasses for food and flour, and stems, tubers, leaves and As a result of the undulating nature of the EMA, the bark from other species for food, drink and medicinal most common wetland type are the valley bottom uses. Wetlands are also important for water provision wetlands (eThekwini Municipality 2011). The largest for communities that live in rural areas and do not have wetlands in the EMA are the floodplains associated access to piped water. Water is used for household with the larger rivers such as the uMdloti, uMngeni and needs such as drinking water, bathing, washing of uMlazi, but they are also associated with some of the clothes and also for livestock. The fertile soils in and smaller rivers found in the EMA (eThekwini Municipality around wetlands are also used for the growing of food 2011). High priority wetlands with high conservation crops such as banana and madumbi (sweet potato). importance include those found on the Mdloti, uMngeni Details about the quantities and extent of natural and oHlanga Rivers which are mainly floodplain and resource use in wetlands in the EMA are not well known. channelled valley bottom wetlands (DWA 2013a). These Estimates of resource use were based on studies from areas are characterised by unique, high diversity wetland other areas of KwaZulu-Natal and on expert opinion. types and high species richness (DWA 2013a). Wetlands with moderate importance and lower priority status are located on the Mkomazi, iLovu and Tongati and associated rivers and are mostly small, narrow valley bottom wetlands. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 119 A1.3 Estuaries A1.3.1 Introduction Estuaries constitute the region where fresh water from Estuaries in South Africa fall into three biogeographical terrestrial drainage meets and mixes with sea water zones, with all of the KwaZulu-Natal estuaries falling (Allanson & Baird 1999). They are well known to be within the Subtropical Zone. The Subtropical Zone one of the most biologically productive ecosystems on extends from Mbashe Estuary to Maputo Bay. Upwelling Earth and are highly variable in terms of their geology, seldom occurs along this stretch of coastline and hydrology, salinity and sedimentation (Allanson & Baird estuarine water temperatures usually range from 14- 1999, Forbes & Demetriades 2008). Estuaries supply 28°C, with coastal sea water conditions often above 20°C a wide range of provisioning, regulating and cultural due to the close proximity of the warm Agulhas Current ecosystem services. (Whitfield 1998). The topography of the area is broadly similar to that immediately to the south and north of There are a 16 estuaries found within the study area. the municipal area in that it is hilly and rises steeply From north to south these are the uTongati, uMdloti, from the coast. This has resulted in the development oHlanga, uMngeni, Durban Bay, Isipingo, eziMbokodweni, of many of the smaller estuaries, although in cases such aManzimtoti, Little aManzimtoti, iLovu, uMsimbazi, as the uMkhomazi, the catchment extends well beyond uMgababa, Ngane, uMkhomazi, uMahlongwane and the western municipal boundaries all the way to the iMahlongwa (Figure A1. 2). These range in size from only Drakensberg mountains (Forbes & Demetriades 2008). 9 ha to 910 ha, with a total estuarine area of about 2300 ha (Table A1. 3). The areas of the estuaries are taken South African estuaries can be classified into five major from Forbes & Demetriades (2008) and represent the types, namely; estuarine bay, permanently open, river total estuarine area, including the area up to the flood mouth, estuarine lake, and temporary open/closed lines and development set back lines. (Whitfield 1992). The characteristics of each type of estuary are shown in Table A1. 3. Table A1.3 Physical classification of estuaries (Source: based on Whitfield 1992) Type Typical size Tidal Prism Mixing Process Average Salinity (PSU) Estuarine Bay Large Large (>10 x 106m3) Tidal 20-35 Moderate (1 - 10 x Permanently Open Medium – Large Tidal/Riverine 10->35 106m3) River Mouth Small – Large Small (<1 x 106m3) Riverine <10 Negligible (<0.1 x Estuarine Lake Large Wind 1->35 106m3) Temporarily open/closed Small – Medium Absent Wind 1->35 Page 120 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure A1.2 The location of the 16 estuaries found within the EMA and their corresponding condition A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 121 A1.3.2 Estuaries of the EMA All five of the estuarine types are represented within the KwaZulu-Natal Province but only four are found within the boundary of the EMA (Table A1. 4). The only type not represented in the EMA is that of large estuarine lakes, such as St Lucia. More than half of the estuaries found in the study area are highly degraded or in a poor state of health (Table A1. 4). The biggest threats to estuaries in the EMA include habitat loss, organic and chemical pollution from WWTW, informal settlements and industrial activities, unregulated sand mining, overexploitation, and upstream freshwater diversions or abstractions. Table A1.4 Physical classification of estuaries in the eThekwini Municipal Area Estuary Classification Total estuarine area (ha) Health uTongati Temporarily open/closed 150.8 Highly degraded uMdloti Temporarily open/closed 99.5 Poor oHlanga Temporarily open/closed 86.3 Poor uMngeni Permanently Open 373.3 Highly degraded Durban Bay Estuarine Bay 2792.2 Highly degraded Isipingo Temporarily open/closed 598.8 Highly degraded eziMbokodweni Temporarily open/closed 57.1 Highly degraded aManzimtoti Temporarily open/closed 24.4 Highly degraded Little aManzimtoti Temporarily open/closed 22.1 Highly degraded iLovu Temporarily open/closed 62.0 Fair uMsimbazi Temporarily open/closed 69.6 Good uMgababa Temporarily open/closed 74.5 Good Ngane Temporarily open/closed 14.8 Fair uMkhomazi River Mouth 174.5 Fair uMahlongwane Temporarily open/closed 19.3 Good iMahlongwa Temporarily open/closed 76.7 Fair Most of the estuaries in the EMA have been well studied and the main characteristics of each system are described in the Table A1. 5, based on information available in the literature. A comprehensive technical report of the estuaries of the eThekwini Municipality was produced in 2008 by Forbes & Demetriades and much of the below information comes from this study. Page 122 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A1.5 Physical classification of estuaries in the eThekwini Municipal Area Estuary Water Quality Vegetation Fish Birds Major Impacts Poor WQ. High nutrient & poor oxygen levels. High Proliferation of marginal & floating macrophytes. Habitat loss. Regional importance to the waterbird community. levels of ammonia, polyaromatic hydrocarbons The nothern bank dominated by coastal dune scrub Dominated by estuarine dependent species. Poor water quality. Chemical contamination. uTongati Dominated by common open-water birds & waders (PAHs) & certain pesticides. & some swamp forest. Southern bank dominated Decline in fish species diversity & abundance. Invasive alien plants. and species utilising estuaries as roosting areas. Algal blooms. by coastal forest and reeds. Unregulated sand mining. WQ severely degraded as a result of ineffective Highly modified vegetation. Cultivation of waste water treatment and land use activities. sugarcane in riparian zone. Narrow reed fringe Largely dominated by estuarine & freshwater spp Locally diverse community of waterbirds. Low uMdloti Habitat loss. Low salinity and oxygen levels. System is artificially remains. Some coastal forest on south bank. due to mouth closure for extended periods. numbers of fish & invertebrate feeding birds. eutrophic. Proliferation of alien invasive species. Eutrophication. Freshwater diversions. Poor WQ. High nutrient & poor oxygen levels. Algal Broad areas of reed swamp and coastal forest. Low species richness and abundance. Numerous Fairly diverse system, especially in terms of the Sewage (WWTW). oHlanga blooms. Some secondary grassland and woody vegetation. fish kills as a result of low oxygen in system. piscivores. Habitat Loss. Illegal gill net fishing Sewage (WWTW). Chemical contamination. Beachwood mangroves important & relatively Poor WQ. Low oxygen levels in upper reaches. Richest fish community in EMA but decline in Highly abundant and diverse waterbird community Habitat Loss. intact. Most of the other riparian vegetation lost uMngeni Nitrogen and Phosphorus levels fluctuate with open conditions suggest impacts on fish diversity and but declines in certain species have been noted. Freshwater diversions. through coastal development and canalisation in mouth conditions. High bacterial counts. abundance. Intertidal sandbanks attract numerous waders. Litter and debris. upper reaches. Eutrophication. Overexploitation (fishing and bait collecting) Sewage. Chemical contamination. Habitat Loss. Diverse and abundant fish community. Dominated Waterbird community has declined significantly WQ severely degraded. Extremely high levels of Mangroves. Diatoms found in the polluted areas. Freshwater diversions. Durban Bay by marine fish at mouth of bay and estuarine over the last few decades. Sandbanks do attract faecal coliform bacteria in certain areas of the bay. Seaweeds also occur certain areas. Litter and debris. bottom feeders further in. numerous waders. Eutrophication. Overexploitation (fishing and bait collecting) Introduced species. Solid waste pollution. Sewage (WWTW) Extremely poor water quality. Extremely high Some mangroves on northern bank. Riparian Highly impoverished fish community (lowest in the Habitat loss. Isipingo levels of faecal coliform bacteria, nitrogen and Low abundance and diversity of bird assemblages. vegetation reduced significantly. Water hyacinth EMA). Sedimentation. phosphorus. Freshwater diversions. Chemical and organic pollution. Canalisation of both banks of the eziMbokodweni Eutrophication. has resulted in a complete loss of estuarine Extremely poor water quality. Organic pollution, Carrying capacity of estuary small but relatively Relatively low abundance and diversity but high Habitat loss. eziMbokodweni floodplain habitats. The riparian zone is confined high levels of bacteria, nitogen and phosphorus. high abundance and diversity of fish. numbers of piscivores. Chemical and organic pollution. to a narrow fragmented strip on the northern bank Increased flows from WWTW. and is almost non-existent on the southern bank. Eutrophication. Habitat loss. Extremely poor water quality as a result of a number Riparian vegetation is very limited along the entire aManzimtoti Relatively high abundance and diversity of fish. Low abundance and diversity of bird assemblages. Chemical and organic pollution. of anthropogenic influences. course of the system. Weir at mouth. Alien species. Eutrophication. Poor water quality as a result of sewage effluent and Riparian vegetation is very limited and alien Relatively high diversity for a small estuary, Low abundance and diversity of bird assemblages. Habitat loss. Little aManzimtoti other anthropogenic influences. invasive species have taken over. restricted to the mouth area. White-backed night heron recorded. Organic pollution. Increased flows from WWTW. Riparian vegetation has been transformed and/or Nutrient concentrations low and water quality Diverse fish community, a result of favourable open Strong presence of wading birds (intertidal sand Habitat loss. iLovu removed and has been invaded yby alien invasive considered to be fair. mouth condition. Second most diverse in the EMA. banks). Low counts of piscivores. Sand mining. species. Relatively good water quality with low bacteria Dense stand of P.australis in the middle of the One of the most diverse bird communities in the Canalization of floodplain. uMsimbazi Reasonably diverse fish assemblages levels. Some litter recorded in estuary. estuary important. EMA, with a large proportion of piscivores. Habitat loss due to bridge construction. The extensive sedge and reed beds along the Relatively good water quality with low bacteria uMgababa course of the system represent a unique feature Reasonably diverse fish assemblages Reasonably diverse bird assemblages Freshwater diversions (uMnini Dam) levels. along the KwaZulu-Natal coastline. Reed bed at head of estuary and a few black Relatively good water quality with low bacteria Low diversity and abundance as a result of the Low diversity and abundance as a result of the Floodplain encroachment. Ngane mangroves are the only important botanical levels. small size of the estuary. small size of the estuary. Habitat loss due to bridge construction. feature. Relatively low diversity and Eutrophication. WQ is fair with some organic pollution and Very little natural riparian vegetation is left along Diverse fish community as a result of the open numbers of waterbirds considered surprising Organic pollution. uMkhomazi eutrophication. the course of the system mouth of the system. given the size of the estuary and the presence of Sand mining. intertidal sandbanks. Habitat loss. Water quality is fair. High levels of nitrogen and Dense stand of P.australis in the middle of the Relatively low diversity and Relatively low diversity and Eutrophication. uMahlongwane phosphorus. estuary important. numbers of fish numbers of waterbirds Habitat loss in the floodplain (sugarcane). Organic pollution. Water quality is fair. Bacteria levels higher in Large island of P.australis is prominent and and Relatively low diversity and Relatively low diversity and iMahlongwa A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE Eutrophication. AREAS IN ETHEKWINI MUNICIPALITY Page 123 summer – thought to be from septic tank intrusion. important feature in this system. numbers of fish numbers of waterbirds Habitat loss in the floodplain (sugarcane). This page intentionally blank. Page 124 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY A1.4 Marine Environment (eThekwini Municipality 2009). Climate change and associated pressures such as sea level rise and increased storm events also threaten these ecosystems. The A1.4.1 Introduction loss of beaches and dune systems along this stretch of The coastline of KwaZulu-Natal extends for 640km and coastline threatens coastal property and infrastructure supports a greater number of species than any other and will affect the overall functioning of the coastal stretch of South African coast (DAEARD 2010). The rich ecosystem (eThekwini Municipality 2009). tropical biodiversity of the region is, in part, a result of the warm Agulhas Current that runs along the edge of the east coast of South Africa influencing species A1.4.3 Rocky shores abundance, distribution and diversity, weather patterns Rocky shores are dynamic ecosystems and dominate and storm events (Palmer et al. 2011). The EMA marine sections of the KwaZulu-Natal coastline, especially the environment covers a 98km stretch of this KwaZulu- section of coast south of Durban Bay. Rocky shores can Natal coastline and is characterised by offshore rocky take the form of headlands, wide wave-cut platforms or reefs, rocky terraces and large sandy areas (DAEARD rocky outcrops separated by sandy beaches (Palmer et 2010). Beaches along the KwaZulu-Natal coast are al. 2011). Rocky shores are inhabited by a wide range generally unprotected, high-energy systems with fine to of sessile and mobile organisms found at different coarse grained sand and characteristically have coastal vertical zones based tidal levels and inundation. Sessile dune forests that flank the top of the shore (DAEARD organisms include barnacles, mussels, limpets, seaweed 2010). However, within the EMA these dunes and and redbait and mobile organisms include crabs, fish, coastal environments have been severely degraded or sea urchins, sea cucumbers and sea stars (Palmer et al. completely transformed as a result of extensive coastal 2011). development. Other threats include overexploitation and unregulated fishing, changes to the sediment and The invertebrate subsistence fishers collect a variety nutrient transport from estuaries to the sea and coastal of rocky shore organisms which are a source of food pollution. secuirty. The main organisms harvested are mussels, redbait, oysters and octopus (DAEARD 2010). Rocky shore organisms are also harvested by recreational and A1.4.2 Sandy beaches subsistence line fishermen who use them for bait if they The KwaZulu-Natal coast is typically dominated by sandy cannot afford sardines. A significant proportion of this beaches which are dynamic and mobile ecosystems harvesting is unregulated as a result overexploitation (Palmer et al. 2011) constantly changing as waves, wind is a constant threat with easily accessible sections of and tides influence the shifting of sand both inshore rocky shore being completely devoid of many intertidal and offshore. During rough conditions or storm events, organisms. large amounts of sand can be removed from the beaches by the sea and deposited offshore and during calmer periods, waves transport sand back onto the beach A1.4.4 In-shore marine environment (Palmer et al. 2011). Durbans central beaches are The coastal waters off the KwaZulu-Natal coast are sheltered by Durban Bay and are maintained through very diverse with approximately 2500 species of fish sand pumping operations. The sand on these central recorded, five species of turtle, 28 types of whales and beaches is finer, the slope flatter and the waves more dolphins, more than 46 seabird species and thousands gentle when compared to other beaches of the region, of invertebrate species such as shellfish and octopus making these beaches the most popular for tourism and (Palmer et al. 2011). The diving and snorkelling sites recreation (eThekwini Municipality 2009). associated with these coastal waters are popular tourist attractions and are also extremely popular A variety of organisms such as small crustaceans, with recreational fishermen (e.g. boat and land based burrowing worms, molecrabs and ghost crabs, clams linefishery and spear fishing). and plough snails all inhabit sandy shores and are adapted to be able to deal with harsh conditions. Sandy The annual ‘Sardine Run’ which occurs in winter (usually shores are also important feeding grounds for birds, between May and July) is a major tourism attraction such as the sand plover. Beaches and dunes provide for the region. Enormous shoals of sardines migrate storm protection and are coastal buffers protecting the from the Cape up the coast into the seas off KwaZulu- coastline from storms and absorbing and dissipating Natal attracting enormous numbers of other predatory the energy produced by waves (eThekwini Municipality gamefish such as elf, garrick, yellowtail, geelbek and 2009). However, sandy beaches and associated dune dusky kob (Palmer et al. 2011). Sharks such as dusky systems are under constant pressure especially in the and copper sharks, dolphins, seals, whales and seabirds EMA where population growth and coastal development are also attracted to the the vast food supply. When has lead to the excavation of beaches and dunes the sardines move close inshore licensed commercial A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 125 fishermen attempt to catch the shoals using beach seine nets and recreational anglers attempt to catch gamefish and sharks (Palmer et al. 2011). Although the east coast supports a wide range of fish and invertebrate species there are few commercial fisheries that actually operate. This is largely due to the fact that the KwaZulu-Natal coast is very nutrient poor when compared to the Atlantic coast which is dominated by upwelling events and subsequent high productivity. The continental shelf along the east coast is also very narrow, limiting the habitat available to support productive commercial fisheries (DAEARD 2010). Commercial fisheries operating on the east coast and within the EMA include commercial linefishing, prawn trawling, pelagic longlining, beach seine-netting, and oyster harvesting (DAEARD 2010, Everett 2014). Recreational and subsistence fisheries such as shore and boat-based linefishing are also important both ecologically and economically (DAEARD 2010). Page 126 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY APPENDIX 2: FLOOD MODELLING Investigating the hydrologic ecosystem services provided ƒƒ Various GIS datasets from the eThekwini Municipality, by Durban’s catchments (i.e. flow regulation) required including: 0.5m Rasta .IMG files for surface elevation, comprehensive and complex hydraulic surface runoff 2m Contour, Landuse Zonal, Durban Metro Open models for all catchments that feed the full EMA. In Space System (D’MOSS), High resolution Aerial order to achieve this, the US-EPA SWMM5 hydrology and imagery; hydraulics engine, interfaced by the PCSWMM software was selected. SWMM5 is an integrated, physically based ƒƒ Shuttle Radar Topography Mission (SRTM) 30m model that was selected based on discussions with resolution surface elevation data; eThekwini Municipality’s Catchment, Stormwater and Coastal Management (CSCM) department. CSCM has ƒƒ Soil type classification maps; recently completed the migration of completed HEC-RAS ƒƒ Geometric HecRAS hydraulic files for EMA rivers hydrologic models of the EMA to SWMM5. The outputs (where available); from these models were used to estimate and map the value of hydrologic services. This appendix provides ƒƒ Stormwater networks (where available); the methodology and model setup for the hydrological modelling used to determine the value of flow regulation ƒƒ Relevant point source data (e.g. WWTW); in the EMA. ƒƒ Design Rainfall Estimation (HydroRisk, http://ukzn- The aims of the hydrological modelling for ecosystem iis-02.ukzn.ac.za/unp/beeh/hydrorisk); services valuation included: ƒƒ eThekwini Design Rainfall (Smithers, 2002); and ƒƒ Estimating the influence of natural systems on flood hydrographs at selected points (e.g. above flood-prone ƒƒ Water quality parameters and landuse change areas/relevant infrastructure) for a range of return shapefiles. periods (RPs) by comparing outputs from the status quo (i.e. current baseline situation) with a hypothetical scenario where the infiltration and roughness of A2.1.2 Software natural systems are modified to a “fully developed” The following specialised software was used for the situation. model setup and calibration: Therefore the estimated effects of particular landscape ƒƒ QGIS; units due to the modelled land use changes represent the principle outcome delivered by this hydrological ƒƒ US-EPA SWMM5 interfaced by the PCSWMM GUI; modelling exercise. ƒƒ HecRAS, Hec-Geo-RAS; and A2.1 Model setup ƒƒ Anaconda – Spyder –Python 2.7 – Data Analysis/ Management. The full eThekwini catchment system was modelled using the US-EPA SWMM5 hydrology and hydraulics engine, interfaced by the PCSWMM software. The model A2.1.3 GIS land use layers required a variety of baseline information, the details of Available current GIS landuse files (e.g. zoning files, which are outlined below. The calibration methodology landcover and D’MOSS) were collated and reviewed. adopted focused heavily on reducing the uncertainty These files were concatenated into one consistent during model set up. landuse polygon shapefile. Numerous errors and irregularities were found in the landuse zoning A2.1.1 Baseline information descriptions. Landuse files were merged from five different zones that make up the EMA and each of The modelling of landuse changes required a these files had their own set of landuse classifications. comprehensive complex hydraulic surface runoff model. This meant that there was significant inconsistency in The setup of the model for ecosystem service valuation the landuse descriptions of the EM Hydrological Study. required the integration of several GIS layers and post- Further post-processing was performed to dissolve these processing for the hydraulic model input parameters. descriptions into a common set of landuse conventions The following baseline information was used in (see Figure A2. 1). In order to address any errors, the developing the model: resulting shapefile was later ‘groundtruthed’ using aerial imagery for the whole of the EMA. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 127 It is important to note that the original landuse The Southern African National Land Cover dataset (2013- descriptions from the D’MOSS landuse file (Figure A2. 14) was used for the catchments outside of the EMA. 1) were maintained throughout the landuse description These data have been generated from multi-seasonal, process and have been classified under “Nature and 30 metre resolution Landsat 8 satellite imagery. The 72 Conservation Areas”. The D’MOSS classifications were different landuse descriptions were summarised into a further discretised to provide an indication of the more manageable list of landuse categories (Table A2. hydraulic parameters required for the hydraulic model, 1). For example ‘grass_medium’ and ‘grass_high’ were i.e. grassland could be described as ‘open_grass’ or labelled as ‘grassland’. ‘open_grass_soil’ which indicates a higher soil erodibility. Figure A2.1 Landuse Categories and D’MOSS subcategories for all Conservation and Nature Areas within the EMA Page 128 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A2.2 Landuse Categories used for the catchments outside of the EMA. These were then linked to the naming convention as shown in Figure 1. LANDUSE CATGORIES 1 Indigenous Forest 10 Cultivated subsistence crops 2 Thicket/dense bush 11 Settlements 3 Woodland/open bush 12 Wetlands 4 Low shrub land 13 Grasslands 5 Plantations/woodlots 14 Mines 6 Cultivated commercial annual: non-pivot 15 Waterbodies 7 Cultivated commercial annual: pivot 16 Bare ground 8 Cultivated commercial permanent orchards 17 Degraded 9 Cultivated commercial permanent vines A2.1.4 Subcatchment delineation The study area was divided into subcatchments and the outlet points were identified (subcatchment runoff is routed to a single discharge point). Outlet points can be defined as nodes of the drainage system or they can be routed to other subcatchments. The GIS subcatchment (watershed basins) data made available for this project were derived from EM flood studies and are in the order of 1km2 and larger. Although appropriate for flood studies, the information relevant to this scope of works required discretisation into smaller, more appropriate subcatchments, in the order of 0.2km2, within the EMA (Figure A2. 2). These new subcatchments were processed from high resolution, .IMG raster files (DEM files). The raster files were converted to .flt float files which can be used as a TIN (Triangulated Irregular Network). A spatial analysis tool was then used to process out the flow paths, watershed boundaries, and river centre lines. For the region outside of the EMA, Shuttle Radar Topography Mission (SRTM – 30 x 30m cell size resolution) data were acquired and used to process the same special analysis tools to delineate watershed information. These two data sets were merged, as required for the scope of works. The SRTM data is not intended to provide detailed analysis results, but will allow for a qualitative assessment on a larger scale than just the EMA zone. The region outside of the EMA was discretised into approximately 0.5 to 1 km2 subcatchments with larger subcatchments closer to the Drakensberg Mountains. Both subcatchment files were merged into one file. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 129 Figure A2.2 Information required to delineate subcatchments: topographical aerial survey, DEM and river centre lines based on river flow paths Page 130 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY A2.1.5 Flow lines The EM flood models, based on geometric HecRAS files, were imported into the PCSWMM model (Figure A2. 3). The geometric file contains more information with regards to stormwater infrastructure (culverts, bridges etc.), however, it does not contain the stormwater network. These data sets only account for approximately 30% of the EMA and do not include the catchments beyond its boundaries. Flow paths simulated using a watershed delineation tool (WDT) were appended to the HecRAS files in order to represent the required study area. Figure A2.3 Illustration of HecRAS files merged and converted to SWMM5 layers for the study A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 131 There are a number of large dams in the eThekwini catchments, including Midmar Dam, Albert Falls Dam and Inanda Dam. Routing river flows through a dam is a fairly complex exercise which requires bathymetric data or hypsometric curves, time stamped/current water level data, knowledge of siltation rates and updates on whether the gates are open/closed, if there is overflow. In order to provide continuity through these dams and best represent the impacts of dams on flood flows, the model was split by routing all the flows entering the dam to an outfall (Figure A2. 4). Inflows were subsequently input downstream of the dams based on the analysis of data collected from the relevant water authorities. Figure A2.4 An example of flows entering and exiting a dam. Page 132 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY A2.1.6 Point sources Average daily abstractions and return flows/discharges were added as point sources at the appropriate junctions. A list of the wastewater treatment works (WWTWs) located within the EMA is given in Table A2. 2. Note that the design capacity was used where the operating capacity was not given. Table A2.3 WWTWs located within the study area. WWTW Name Latitude Longitude Design Capacity Operating Subcatchment (Ml/d) capacity (Ml/d) Umkomaas -30.20315 30.794653 1 U10M Verulam -29.64487 31.063288 13 U30B Umhlanga -29.69723 31.081528 6.8 U30B Phoenix -29.67962 31.037052 25 U30B Umdloti -29.6501 31.10941 3 U30B Tongaat Central -29.56028 31.137734 6 7.33 U30D Gennazzano -29.60671 31.156189 1.8 U30D Fredville -29.70166 30.644582 2 U20L Northern Works -29.79581 30.997692 70 U20M New Germany -29.80605 30.896683 7 U20M Kwadabeka -29.76397 30.929854 U20M KwaMashu -29.72957 31.008914 65 U20M Magabeni -30.16583 30.781316 1.3 U70E Hammarsdale -29.80025 30.66339 13 U60C Cato Ridge 0.95 U60C Mpumalanga -29.80374 30.592571 6.4 U60C KwaNdengezi -29.86829 30.768762 2.4 U60D Dassenhoek -29.87836 30.793295 6 U60D Southern -29.95855 30.97295 230 U60D Amanzimtoti -30.00776 30.916413 27 U60E Isipingo -29.99021 30.906703 18.8 10.98 U60E Umhlatuzana -29.87713 30.884036 14.8 U60F Umbilo -29.84561 30.891653 23.2 U60F Hillcrest -29.7941 30.75635 1.2 U60F Craigieburn 1.78 U80L Central -29.87683 31.060138 135 Bluff Kingsburgh -30.07451 30.856273 7.2 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 133 A2.1.7 Stormwater network A shapefile of the stormwater networks was provided by the eThekwini Municipality. The current available stormwater shapefile is incomplete and contains numerous errors and inconsistencies (see Figure A2. 5). Invert levels and pipe sizes are often missing and connections are incorrect and/or missing. For this reason, the stormwater network was only included in the U60F model. Available networks were amended where possible, i.e. a standard circular pipe size of 0.375m was allocated to pipes with missing geometry and tools were applied to either fill in missing invert levels (from the DEM) or to apply slopes within the network. Where necessary, main pipelines had to be added to these networks. The pipe profiles were later checked to ensure reasonable slope gradients and the continuity of flows. Figure A2.5 Current available stormwater shapefile (the red lines represent the flow paths, yellow lines are stormwater conduits and blue dots represent stormwater junctions). The eThekwini Municipality is presently carrying out an SMS (Stormwater Management System) audit of all stormwater infrastructure. The audit entails a visual inspection and assessment of all stormwater infrastructure from which shapefiles of existing junctions, stormwater pipes and culverts are generated. Where available (see Figure A2. 6), these new shapefiles were imported into the model and connected to existing stormwater networks and flow paths. Continuity in these networks were checked and errors/connections were corrected where necessary. The original stormwater shapefile was merged with the new SMS shapefiles and both were connected to the HECRAS and WDT flow paths. Again profiles and connections were checked. Page 134 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure A2.6 Newly available stormwater network shapefiles from the EM’s SMS audit. A2.2 Catchment characteristics A2.2.1 Hydraulic parameters A number of input parameters are required for SWMM5. The determination of the catchment characteristics were These include hydraulic parameters, soil infiltration estimated using a spatial analyst tool for zonal statics. properties, rainfall and water quality parameters. Raster files were generated to represent the following information required for the hydraulic and hydrological models, with reference to each subcatchment. These were used to estimate the many runoff characteristics outlined in Table A2. 3. Table A2.4 Hydraulic input properties required for each subcatchment. Hydraulic Parameter Description Source Area (ha) Area of subcatchment GIS tool Width (m) Width of overland flow path GIS tool Flow Length (m) Length of overland sheet flow GIS tool Slope (%) Average slope along the pathway of overland flow to inlet locations. GIS tool Imperv. (%) Percent impervious area RGB colour extraction Manning’s roughness coefficient, N, for overland flow for impervious N Imperv Rossman, 2015 (Table 5) area. N Perv Manning’s roughness coefficient, N, for overland flow for pervious area. Rossman, 2015 (Table 5) Dstore Imperv (mm) Depth of depression storage on impervious areas ASCE, 1992 (Table 4) Dstore Perv (mm) Depth of depression storage on pervious areas ASCE, 1992 (Table 4) SWMM default setting of 25% Zero Imperv (%) Percent of impervious area with no depression storage based on literature Percent Routed (%) Percent of runoff routed between sub-areas Outfalls A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 135 Subcatchment areas were measured and the width category based on standard values for different landuses of the subcatchment is defined as the physical width found in the literature and 2) RGB colour extraction tool of the overland flow and in an idealised, rectangular was applied to differentiate between impervious and catchment, the total width would be twice the length pervious areas based on aerial imagery (in the EMA) and of the drainage channel (assuming both sides of the Google Earth images (outside of the EMA). The EM has subcatchment are symmetrical). been working on the second approach, however we were not satisfied with the results and decided to use the first The most significant input hydraulic parameter is the approach. Figure A2. 7 shows sections of two different percentage of impervious area (Imperv. %). There areas of contrasting landuse i.e. residential and industrial. are a number of methods that can be employed to The top value represents the %Imperv using approach estimate the percent imperviousness of a subcatchment. 1 and the bottom value represents the %Imperv using Ideally the percent imperviousness could be measured approach 2. The estimation using approach 2 was accurately from aerial photos or land use maps, however, reasonable except where the colour spectrum was this can be tedious for large study areas such as this a mixture of green and brown, e.g. recently harvest one. Two approaches were investigated: 1) a percent sugarcane and rural sandy areas. impervious area was associated with each landuse Figure A2.7 Percent impervious area for two different areas of contrasting landuse. The top value represents the %Imperv using approach 1 and the bottom value represents the %Imperv using approach 2. Page 136 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY The impervious and pervious N values were taken from literature. Estimates of Manning’s roughness coefficient (N values) for overland flow are taken from literature. A summary from three different sources are given in Table A2. 4. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 137 Table A2.5 Estimates of Manning’s roughness coefficient (N values) for overland flow. A summary from three different sources (Source: Rossman 2015) Hydraulic Description Source Parameter Smooth asphalt 0.01 Asphalt and concrete paving 0.014 Packed clay 0.03 Crawford and Linsley (1966)a Light turf 0.20 Dense turf 0.35 Dense shrubbery and forest litter 0.4 Concrete or asphalt 0.011 0.010-0.013 Bare sand 0.010 0.01-0.016 Graveled surface 0.02 0.012-0.03 Bare clay-loa, (eroded) 0.02 0.012-0.033 Engman(1986)b Range (natural) 0.13 0.01-0.32 Bluegrass sod 0.45 0.39-0.63 Short grass prairie 0.15 0.10-0.20 Bermuda grass 0.41 0.30-00.48 Smooth asphalt pavement 0.012 0.010-0.015 Smooth impervious surface 0.013 0.011-0.015 Tar and sand pavement 0.014 0.012-0.016 Concrete pavement 0.017 0.014-0.020 Rough impervious surface 0.019 0.015-0.023 Smooth bare packed soil 0.021 0.017-0.025 Moderate bare packed soil 0.030 0.025-0.035 Rough bare packed soil 0.038 0.032-0.045 Gravel soil 0.032 0.025-0.045 Mowed poor grass 0.038 0.030-0.045 Yen (2001)c Average grass, closely clipped sod 0.050 0.040-0.060 Pasture 0.055 0.040-0.070 Timberland 0.090 0.060-0.120 Dense grass 0.090 0.060-0.120 Shrubs and bushes 0.120 0.080-0.180 Business land use 0.022 0.014-0.035 Semi-business land use 0.035 0.022-0.050 Industrial land use 0.035 0.020-0.050 Dense residential land use 0.040 0.025-0.060 Suburban residential land use 0.055 0.030-0.080 Parks and lawns 0.075 0.040-0.120 a Obtained by calibration of Stanford Watershed Model b Computed by Engman (1986) by kinematic wave and storage analysis of measured rainfall-runoff data Page c 138 on A Computed SPATIAL basis VALUATION of kinematic OF THE NATURAL AND SEMI-NATURAL OPEN wave analysis SPACE AREAS IN ETHEKWINI MUNICIPALITY The depression storage is the volume that must be filled the dependence of infiltration capacity on soil prior to the occurrence of runoff on both pervious and characteristics and the present soil capacity during a impervious areas. Values for depression storage were storm event. There are five options that can be used in taken from the SWMM Manual (EPS 2015 after ASCE, SWMM, namely Horton’s method, the modified Horton 1992; Table A2. 5). In SWMM, depression storage may be method, the Green-Ampt method, the modified Green- treated as a calibration parameter, particularly to adjust Ampt method and the Curve Number method. With all runoff volumes. Therefore obtaining accurate values in of these models, the parameters depend on the type the setup may be unnecessary as these value may change and condition of the soil of interest. during calibration. Depression storage is most sensitive for small storms; as the depth increases it becomes a It is worth noting that the Flood Line Delineation studies smaller component of the water budget (EPA, 2015). for EM use the Soil Curve Number (SCN) to represent the runoff co-efficient for catchment routing. Although suitable for flood studies (as a conservative approach), Table A2.1 Values used for the depression storage based on landuse this investigation will use the Green-Ampt method. This (ASCE, 1992) methods provides a soil memory as opposed to a broad WWTW Name Latitude brush coefficient approach. Umkomaas -30.20315 For the Green-Ampt infiltration method, the model requires three soil parameters that the user must specify Verulam -29.64487 for each of the subcatchments: Umhlanga -29.69723 1. Capillary suction head, Ψs (mm); Phoenix -29.67962 2. Saturated hydraulic conductivity, Ks (mm/hr); and 3. The maximum available moisture deficit, θdmax (volume of dry voids per volume of soil). A2.2.2 Soil Infiltration These parameters were taken from the SWMM Manual The largest proportion of rainfall losses over (Table A2. 6). Figure A2. 8 is a map of the green-ampt pervious areas generally occur due to soil infiltration. parameters for the study area that has recently been Theoretically the Richards equation is the most developed and applied during a current Water Research representative, however its highly nonlinear partial Commission (WRC) study by Pegram and Sinclair at the differential equations make is unsuitable for continuous University of Kaw-Zulu Natal (UKwaZulu-Natal). These long-term simulations. Simpler algebraic infiltration parameters will be used and the results compared with models have been developed that represent those using the parameters given in Table A2.6. Table A2.6 Values used for the depression storage based on landuse (ASCE, 1992) Soil Texture Suction Hydraulic Initial Deficit Porosity Field Wilting Class Head (mm) Conductivity (fraction) (fraction) Capacity Point (mm/hr) (fraction) (fraction) Sand 49.02 120.34 0.413 0.437 0.062 0.024 Loamy Sand 60.96 29.97 0.39 0.437 0.105 0.047 Sandy Loam 109.98 10.92 0.368 0.453 0.19 0.085 Loam 88.9 3.3 0.347 0.463 0.232 0.116 Silt Loam 169.93 6.6 0.366 0.501 0.284 0.135 Sandy Clay Loam 219.96 1.52 0.262 0.398 0.244 0.136 Clay Loam 210.06 1.02 0.277 0.464 0.31 0.187 Silty Clay Loam 270 1.02 0.261 0.471 0.342 0.21 Sandy Clay 240.03 0.51 0.209 0.43 0.321 0.221 Silty Clay 290.07 0.51 0.228 0.479 0.371 0.251 Clay 320.04 0.25 0.21 0.475 0.378 0.265 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 139 Figure A2.8 Map of Green-Ampt Parameters developed by UKwaZulu-Natal (Source: Sinclair 2015) A2.2.3 Water quality parameters The landuses that generate these pollutants were The water quality parameters chosen for this study are defined and the pollutant buildup, pollutant washoff nutrients (nitrogen, phosphorous) and total suspended and street cleaning parameters were assigned to each solids (TSS). The landuses that generate these pollutants landuse. Note that no data was available to estimate were defined and the pollutant buildup, pollutant the pollutant buildup and street cleaning parameters washoff and street cleaning parameters were assigned and therefore these features were not considered. to each landuse. The pollutant removal functions for The pollutant washoff from a given land use occurs nodes within the drainage system that contain storage/ during periods of wet weather and can be characterized treatment facilities were also defined. The input in SWMM5 by either using an exponential or rating parameters for each pollutant are as follows: curve relationship. The Event Mean Concentration is a case of Rating Curve Washoff where the exponent ƒƒ the pollutant name; is 1.0 and the coefficient represents the washoff pollutant concentration in mg/L. In each case buildup is ƒƒ the concentration units (i.e. mg/L, μg/L, counts/l); continuously depleted as washoff proceeds, and washoff ceases when there is no more buildup available. The ƒƒ oncentration in rainfall; EMCs were derived from literature (Table A2. 7). These ƒƒ concentration in groundwater; data were applied to the different landuse categories across the study area. The data below can be applied in ƒƒ concentration in direct infiltration/inflow; and determining the water quality volume to be catered for in stormwater management devices (e.g. SUDS). ƒƒ first-order decay coefficient. Page 140 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A2.7 Event Mean Concentration (EMC) data for different landuse types Landuse Description TSS (mg/l) BOD (mg/l) TIN (mg/l) P (mg/l) Settlement - urban 100 15 3.41 0.79 Commercial / Retail / Institutional 166 9 2.1 0.37 Industrial / Road & Rail 166 9 2.1 0.37 Extractive / Utility 166 9 2.1 0.37 Farming / plantations & woodlots 201 4 1.56 0.36 Recreational open space 201 4 1.56 0.36 Settlement - rural 201 4 1.56 0.36 Natural vegetation (D'MOSS) 70 6 1.51 0.12 Settlement - informal 497 22 6.7 A2.3 Storm design events Standard techniques for flood estimation generally include the analysis of observed peak discharges and event modelling using rainfall-runoff techniques, however observed streamflow data are often not available in South Africa. The most common technique for many engineering and conservation design decisions is therefore to use rainfall event-based methods. Smithers and Schulze (2000) compiled tables of design rainfall depths at selected rainfall stations for different durations based on frequency analysis of the annual maximum precipitation for a given duration from historical rainfall data. This data is listed based on a latitude and longitude grid for a variety of recurrence intervals from 2 to 1000 years for storm durations from 5 minutes to several days. The duration of design rainfall may range from five minutes for small urban catchments (with a rapid hydrological response) to a number of days for large regional flood studies. The design rainfall depth used in design flood estimation should be based on the critical storm duration or time of concentration (TC) of a catchment. The TC is defined as the time required for rainfall based runoff, with a spatially and temporally uniform distribution, to contribute to the peak discharge at the outlet of the catchment outlet. The eThekwini Municipality design rainfall spreadsheet based on Smithers (2002) analysis of the region was assigned to each subcatchment using an area-weighting tool. One method is to assign a rainfall distribution to the time of estimated concentration (Tc) for each subcatchment. The SCS Type II distribution is the most relevant in general, and the total rainfall for a given Tc will be used to calculate the intensity for each RP (Figure A2.9). A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 141 Figure A2.9 SCS 24-hour rainfall distributions (not to scale) (Source: SCS 1984) The Design Rainfall for the areas outside of the EMA were taken from the tables derived by Smithers and Schulze (2002) for a 5, 10, 20 and 50-year return period (Table A2. 8). The RPs relevant to this study have been summarized below. Table A2.8 Design rainfall for area outside of the EMA (Smithers and Schulze, 2002). Station ID Station MAP Altitude Years Duration Return Period (years) name (days) 2 5 10 20 50 100 0239482W CEDARA 876 1134 40 1 56 78 95 114 142 167 Emerald 0238806W 902 1209 40 1 61 82 98 114 137 155 Dale Page 142 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY A2.4 Final Model The final SWMM model of the full EMA comprised about 30 000 subcatchments. The eThekwini catchment system was divided into 3 separate models, representing the northern, central and southern catchments in order to reduce simulation running times (Figure A2. 10, Figure A2. 11, Figure A2. 12, and Figure A2. 13). Figure A2.10 The full eThekwini catchments showing flow paths A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 143 Figure A2.11 Snapshot of the SWMM model of the Northern EMA catchments. Page 144 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure A2.12 Snapshot of SWMM model of the Central EMA catchments. Figure A2.13 Snapshot of the SWMM model of the Southern EMA catchments. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 145 A2.5 Model calibration and validation river water levels at several locations. For the region outside of the EM, real-time rainfall data (5 minute Real-time data was used for model calibration and intervals) were obtained from the SAWS for a number of validation. This was done using the data and methods rainfall stations (Table A2. 9). Real-time rainfall data for the outlined below. Flows and water quality within the EMA was obtained from the EM database (Figure A2.14). EMA are relatively well monitored, however there are limitations with using these data for calibration purposes. These limitations are discussed below. Table A2.9 Real-time rainfall data was obtained for the following stations outside of the EMA (Source: SAWS) A2.5.1 Rainfall Selection and Application Rainfall Latitude Longitude Station ID Phase 1 of the calibration and validation was focused on the hydraulic flows and volumes. Real-time rainfall data 1 Emerald Dale -29.940300° 29.959700° available from the EM and the South African Weather 2 Mooi River -29.218000° 30.002500° Services (SAWS) was used in conjunction with measured 3 Cedara -29.541700° 30.265000° 4 Shaleburn -29.352500° 29.786900° Figure A2.14 Locations of the available rain gauge stations in the EMA (Source: eThekwini Municipality) Page 146 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY The Thiessen polygon method was applied to the rainfall stations and each rain gauge was assigned to a certain area (Figure A2. 15). Figure A2.15 Thiessen polygons determined for the full EMA catchments. A2.5.2 Available measured data The efficacy of calibration depends entirely on the of the sensors were washed away. Two sensors were availability of measured data. Flow and/or water level retrieved - one in the Umbilo River and the one in the data are most useful for the calibration of the hydraulics. Umhlatuzana River. This data provides an indication Several real time rainfall data sets were applied in order of baseflows which help to validate the accuracy of to calibrate the models. Calibration points were chosen abstractions and return flows in these rivers. on rivers without dams and where flow data were available. It was not possible to address every river in The eThekwini Municipality’s Water and Sanitation such a large system. The calibration parameters were Department measure river, outfall, stormwater and applied to similar landuse zones in general, and as such beach water quality at numerous monitoring stations were used in all of EM during the process. within the EMA. Of these points, many do not measure nutrient concentrations. Table A2. 10 provides a list of The EM recently deployed a number of water sensors the available monitoring stations and measured data for in the U60F catchment (in the Umhlatuzana and Umbilo the EMA and Table A2. 11 provides a list of the sampling Rivers) over two separate time periods. However, points in the EMA that measure Ammonia (free), Nitrates unusually heavy rainfall was experienced during both and Nitrites (DIN), orthophosphates (P) and turbidity. periods (including an almost 100-year flood) and most A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 147 Table A2.10 DAvailable hydrology monitoring stations within the full EMA catchments from the Department of Water and Sanitation. Gauge Gauge Name/Location Catchment Latitude Longitude Available number Area (km²) data 1960-08-14 U1H005 Mkomazi River @ Lot 93 1821 1744 29.74369 29.90494 2016-01-19 2004-05-06 U1H009 Mkomazi River 4328 30.13561 30.67372 2016-05-12 1950-11-01 U2H005 Mgeni River @ Table Mountain 2519 29.57603 30.60258 2016-01-29 1954-01-04 U2H006 Karkloof River @ Shafton 339 29.38175 30.27775 2016-05-30 1954-07-16 U2H007 Lions River (Mpofana River) @Weltevreden 358 29.44258 30.14853 2016-05-30 1957-12-24 U2H011 Msunduze River @ Henley Dam 176 29.64708 30.25975 2016-01-26 1960-08-11 U2H012 Sterk River @ Groothoek 438 29.42306 30.48828 2016-01-26 1960-08-10 U2H013 Mgeni River @ Petrus Stroom 299 29.51261 30.09442 2016-05-30 1983-09-07 U2H022 Msunduze River @ Inanda Loc. 881 29.66086 30.63617 2016-01-27 1996-01-31 U2H041 Msunduze River @ Hamstead Park 534 29.60772 30.45025 2016-02-09 1989-10-26 U2H055 Mgeni River @ Inanda Loc. 2624 29.64244 30.68861 2016-01-14 1995-06-02 U2H057 Slang Spruit @ Pietermaritzburg 48 29.63072 30.35322 2016-02-09 1995-04-25 U2H058 Msunduze River @ Masons Mill 327 29.64144 30.36544 2016-02-09 2013-02-26 U2H061 Mpofana River 29.39162 30.06297 2016-02-23 1966-10-07 U3H001 Tongati River @ Riet Kuil 236 29.53368 31.08922 2016-02-25 1999-06-29 U4H010 Kleinspruit 16 29.04139 30.56306 2016-02-04 1981-07-13 U6H002 Mlazi River @ Nooitgedacht 105 29.749 30.317 2016-05-23 1981-11-13 U6H003 Mlazi River @ Umlaas 417 29.80395 30.51587 2016-04-15 1949-07-09 U7H001 Zwateni River @ Highlands 16 29.84733 30.23531 2016-05-23 1964-10-13 U7H007 Lovu River @ Beaulieu Estate 114 29.86244 30.24417 2016-05-23 1997-08-20 U7H012 Nungwana River @ Umbumbulu 51 30.00597 30.71214 2016-02-11 1987-05-27 U8H003 Mpambanyoni River @ Umbeli Belli 378.8 30.27403 30.69603 2016-01-20 Page 148 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A2.11 Water quality monitoring stations within the EMA. Sampling station River Sampling station River R-MLAAS_39 R_Zana_10 R-MLAAS_75 R_Zana_28 R-MLAAS_76 R_Zana_29 Umhlatuzana R-NDENGEZI uMlaas River R_Zana_34 R-DASSENHOEK R_Zana_35 R-STERK_03 R-STERK_05 R_Umbilo_04 R-MKOMAZI_03 R_Umbilo_13 Umbilo uMkomazi River R-MKOMAZI_S R_Umbilo_27 R-BOKODWENI_02 R-BOKODWENI_03 R_NGANE_02 South Durban Ngane River R-HLONGWANA_01 R_NGANE_03 R-HLONGWANA_02 R-ISPINGO_00 R-OHLANGA_08 R-ISPINGO_01 R-OHLANGA_07 R-ISPINGO_03 Isipingo River R-OHLANGA_05 oHlanga River R-ISPINGO_04 R-OHLANGA_02 R-ISPINGO_05 R-OHLANGA_01 R-LTOTI_00 R_THONGATI_02 R-LTOTI_02 Little aManzimtoti R_THONGATI_03 uThongati River R-LTOTI_05 R_THONGATI_04 R-MDLOTI_01 R-MDLOTI_02 R-MDLOTI_03 uMdloti River R-ILLOVU_02 iLovu River R-MDLOTI_04 R-MDLOTI_05 R-MGENI_08 R-MGENI_71 R-MGENI_75 R-MGENI_80 R-MGENI_71 uMngeni River R-ALLER_01 R-GANE_04 R-GANE_18 R-NKUTU_01 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 149 A2.5.3 Calibration of flows and water levels The models were calibrated using available measured flow There is reasonable agreement between the simulated and water quality data. Note that in some cases gauging water depths and the measured water depth at the DWA stations did not provide adequate data for calibration gauging station on the Nungwana River at Umbumbulu purposes, for example there is no measured flow data (Figure A2. 17). The simulated peak flows were, however, for the Umbilo and Umhlatuzana Rivers. Therefore, predicting much higher flows than those measured at emphasis was placed on setting up the model as the gauging stations. This indicates that the actual losses accurately as possible. Simulated peak flows for the rivers were higher than in the modelled flows (the volume were compared with those estimated in other studies of the simulated flows is greater than the volume of conducted in the EMA by Jezewski (1984) and Mkwananzi measured flows as shown in (Figure A2. 18). In addition, & Pegram (2004). Although the calibration period allowed the simulated flows respond much quicker to rainfall. In for this study was very time constrained, we feel the order to smooth out the simulated peak flows and reduce results of the model are within an acceptable range. The the simulated volumes, the flow paths were increased results certainly highlight the changes between between to include more tributaries, thereby increasing the time different river systems with regards to subcatchment area of concentration (see Figure A2. 19). As you can see in size and shape. For the purposes of this study there is Figure x the weir is located downstream of a small dam reasonable certainty in the model outputs. which may also be smoothing out measured peak flows. The full eThekwini catchment system was divided into 3 separate models, representing the northern, central and southern catchments in order to reduce simulation running times. Models were calibrated separately. Measured flows and water depths were compared with simulated flows and water depths on the uMsunduze (uMgeni) River. Simulated water levels were lower than measured water levels. Return flows from the wastewater treatment works were added as a constant ‘baseflow’. There appears to have been an irregularity in the measured data with the large rapid increase in flow and water level, this was neglected (Figure A2. 16). Figure A2.16 Measured (gauge U2H022, orange line) vs simulated flows (blue line) and water levels on the Msunduze (Umgeni) River. Page 150 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure A2.17 Picture of the weir and small dam on the Nungwana River Figure A2.18 Measured (red line) vs simulated flows (blue line) and water levels on the Nungwane River in the south. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 151 Figure A2.19 The southern EThekwini catchments showing the additional conduits that were included in the model represented by the red lines. The sensitivity of the model to various hydraulic parameters was tested using the SRTC tool on PCSWMM (Table A2. 12). The sensitivity of the results was tested for the various parameters. The results were most sensitive to changes in the depression storage and % imperviousness. Simulated flood peaks were compared with those estimated by Jezewski in 1984 (Table A2. 13). Jezewski (1984) used a simple rational method approach to estimate the 2-year flood peak flows for the different rivers. Table A2.12 Sensitivity of runoff volume and peak flow to surface runoff parameters (EPA, 2015). Parameter Typical Effect of Effect of Comments effect on increase on increase on hydrograph runoff volume runoff peak Area Significant Increase Increase Less effect for a highly porous catchment Less effect when pervious areas have low Imperviousness Significant Increase Increase infiltration capacity For storms of varying intensity, increasing the width tends to produce higher and earlier hydrograph peaks, a generally faster response. Width Affects shape Decrease Increase Only affects volume to the extent that reduced width on pervious areas provides more time for infiltration. Same as for width, but less sensitive, since flow is Slope Affects shape Decrease Increase proportional to square root of slope. Roughness Affects shape Increase Decrease Inverse effects as for width. Depression storage Moderate Decrease Decrease Significant effect only for low-depth storms. Page 152 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A2.13 Sensitivity of runoff volume and peak flow to surface runoff parameters (EPA, 2015). Source Jezewski (1984) Current SWMM Model (2016) Peak Discharge Q (m3/s) River Area Tc (h) 2yr Flood 2yr 5yr 10yr 20yr (km2) Peak Return Return Return Return Discharge Period Period Period Period Q (m3/s) uTongati 436 8.5 69 190 494 831 1151 uMdloti 527 12.5 76 57 197 203 212 oHlanga 118 5.8 36 40 82 113 130 uMgeni 4432 33.6 223 302 668 1033 1463 Durban Bay 242 6.1 Umbilo 230 364 539 680 Umhlatuzana 366 718 992 1128 Mlazi 972 12.7 1050 2293 3384 4204 eziMbokodweni 295 8.3 57 617 1051 1505 1902 Manzimtoti 44.5 2.8 22 82 160 1 49 Little Manzimtoti 12.5 1.9 11.5 43 64 74 85 iLovu 893 14.3 100 676 1045 1709 1774 uMsimbazi 36.4 2.4 19.5 48 102 132 188 uMgababa 37 2.6 19.8 34 125 198 330 Ngane 16.5 1.8 13.3 35 64 84 109 uMkhomazi 4310 39.0 220 20 103 242 404 uMahlongwane 17 1.4 13.4 6 10 13 16 A2.5.4 Calibration of water quality Monthly water quality data collected by the EM Water and Sanitation Division were used for calibration and validation of the model. Most of the rivers within the EM are monitored The water quality measured at the outfalls of the WWTWs were analysed. These values were significantly lower than those provided in the General Effluent Limits. These values were replaced with the average TIN, P and SS values measured at each outfall of the WWTWs (see Table A2. 14). A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 153 Table A2.14 Sensitivity of runoff volume and peak flow to surface runoff parameters (EPA, 2015). WWTWs Tongaat Umdloti Verulam Phoenix Mhlanga Genazzano KwaMashu (River) (Tongati) (Umdloti) (Umdloti) (Mhlanga) (Mhlanga) (Beachbums) (Mgeni) TIN (Ammonia, Nitrates and Nitrites) (mg/L) min 0.6 0.6 0.0 1.1 0.6 0.0 0.6 max 30.0 25.5 14.5 43.1 48.0 52.9 24.5 average 8.5 4.2 2.5 10.7 11.8 9.2 8.9 summer 9.2 5.1 2.4 10.1 10.3 9.0 7.9 winter 7.8 3.2 2.6 11.3 13.2 9.3 9.7 Orthophosphates (mg/L) min 0.0 0.2 0.0 0.5 0.0 0.1 0.2 max 30.0 10.0 14.0 13.0 12.0 11.0 22.0 average 1.4 4.4 1.2 4.0 4.6 4.2 3.4 summer 1.6 4.6 0.8 4.0 4.2 4.1 3.6 winter 1.3 4.4 1.5 4.0 4.9 4.4 3.2 Suspended Solids (mg/L) min 0 0 1 0 0 0 0 max 1674 85 2173 56 104 249 55 average 24 6 68 5 9 7 11 summer 36 6 63 6 7 10 10 winter 13 6 74 4 10 4 12 Turbidty (NTU) min 3 1 1 1 2 1 3 max 1409 31 4001 57 343 129 4001 average 21 4 170 5 12 5 67 summer 31 4 227 6 9 7 65 winter 11 4 113 5 14 3 67 Simulations were run for a one-week period from the 1 – 7 July 2014 at 2-second time interval. This period was chosen because there was one rainfall event experienced throughout the subcatchment. The results from these simulations are given in Figure A2. 20 and Figure A2. 21 below. The corresponding water quality parameters measured at the same points are provided in Figure A2. 22. Note that the simulated values are within reasonable range of the measured values. Page 154 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure A2.20 Measure rainfall and simulated flow, depth, P, TIN, TSS concentrations for monitoring station (R_ZANA_10) on the Umhlatuzana River. Figure A2.21 Measure rainfall and simulated flow, depth, P, TIN, TSS concentrations for monitoring station (R_UMBILO_13) on the Umbilo River. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 155 Figure A2.22 Figure A2. 22. Measured water quality from water sampling stations on the Umbilo and Umhlatuzana Rivers at the same locations as the simulations. A2.6 Assumptions and Limitations A number of assumptions were made during the setup found in these networks. Missing data were entered of the SWMM model where input data were either based on the following assumptions: insufficient or unreliable. These assumptions may be regarded as limitations of the model and therefore ƒƒ pipe sizes: a default value of 0.375 m should be considered when analysing the results. ƒƒ invert levels: levels were taken from the DEM and the The use of design rainfall profile was altered in order to acquire a reasonable Note that design rainfall assumes that rainfall is equally slope distributed over the whole catchment at the same time. Realistically, a specific design rainfall does not imply an The model was run numerous times in order to resolve equal design runoff, however this is general practice flooding and continuity issues resulting from problems when performing flood studies. with these data. Invert levels were manually adjusted in order to correct negative slopes. Groundwater and baseflows Groundwater was not included in the modelling. Calibration data Insufficient data were available to incorporate any The water quality monitoring program takes accurate representation of groundwater flows. measurements on a monthly basis. This data therefore Therefore, measured flow/water level data was used to provides a snapshot of the water quality at a point at infer the groundwater as baseflow. Groundwater inputs a specific time. TSS concentrations were inferred from vary seasonally, therefore the ‘baseflow’ was estimated turbidity data collected by the EM. The relationship during summer and winter periods and incorporated into between TSS and turbidity was taken from data the model as a time series where possible. collected by Newman (2015) in the Durban Bay, which has a high salinity. TSS concentrations are temporally Stormwater network data and spatially dependant and therefore this is not a The current stormwater network data were inconsistent true representation of actual TSS concentrations. We and incomplete. Only certain areas of the current SMS recognise that this a major limitation in the estimation of audit have been completed and were included in the measured TSS loads used for calibration. model, however inconsistencies and errors were also Page 156 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY TIN, P and suspended solids concentrations from the WWTWs was averaged and added into the model as a point source at the outfall site of the relevant WWTWs. The discharge rates were assumed to be equal to the design capacity and is therefore not a true representation of actual discharge rates and quality. Generalised event mean concentrations (EMC) were based on values derived in the United States. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 157 This page intentionally blank. Page 158 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY APPENDIX 3: INFRASTRUCTURE COST ESTIMATE METHOD A3.1 Overview Excavation was taken as R90/m3, selected backfill as R80/m3 and backfill at R60/m3. The supply and placing This section describes the cost estimation method. The of concrete was estimate at R2400/m3, shuttering was method estimates the existing infrastructure costs from assumed to be R650/m2 and the supplying and fixing their dimensions. It then uses a scaling relationship of steel was taken as R12000/t. Concrete blinding was between flow and the infrastructure dimensions to estimated as R1600/m3. A 10-20% allowance on the total estimate the stormwater requirements under the cost was provided for preliminary and general items different land use scenarios. The infrastructure required (P&G) such as site establishment and supervision. The to satisfy the various scenarios are then costed. The cost individual rates for each infrastructure category are difference between the existing infrastructure and the summarised in additional information. scenario infrastructure is indicative of the value of the natural areas. A3.4.1 Bridges Bridges have two subcategories: bridge culverts and A3.2 Identifying Existing Infrastructure pipe culverts. Bridges differ from the culvert and pipe An inventory of all the stormwater infrastructure is categories as they are positioned within watercourses. identified and categorised into four major categories: To allow for the complications of dealing with water the bridges; canals; culverts and pipe networks. The bridges P&G was set to 20%. With the exception of not having are divided into a further two subcategories: bridge manholes the bridge culverts and the pipe bridges culverts and bridge pipes. The bridge category excludes are priced the same as the culverts and the pipes major bridges as their size is insensitive to flows. respectively. A3.3 Assigning Rainfall Return Periods A3.4.2 Culverts Each infrastructure category is assigned a design return Culverts are defined as any non-circular structure not period based on the eThekwini Design Guidelines acting as a bridge. The culvert costs are estimated (eThekwini Municipality 2008). Table A3. 1 shows from the cross-sectional area and the length. It is the return periods associated with the relevant assumed that all the culverts are constructed from 0.3 infrastructure category. The design rainfalls are then m thick insitu reinforced concrete with steel reinforcing modelled to estimate the peak flows for each structure. attributing to 4% of the total volume. It is assumed that all the ground conditions are the same and that the Table A3.1 Return periods assigned to each of the infrastructure structures are founded on 200 mm of concrete blinding. categories Excavation quantities are based on 600 mm of cover and a payment width as defined in Clause 5.2 of SANS Category Return Period 1200DB. Manholes were priced as R25 000 each and one (years) was assumed every 60 m. Bridges Culverts 20 Bridges Bridges Pipes 20 A3.4.3 Canals Canals 10 Canals are priced the same as culverts except there are no roof slabs, cover material or manholes. Culverts 5 Pipes 2 A3.4.4 Pipes The pipe costs are calculated similarly to the culvert costs. A3.4 Cost Estimate of the Infrastructure It is assumed that all the ground conditions are the same and that a Class B bedding (Drawing LB-1, SANS 1200LB) is Costs are estimated for each of the four categories as used throughout. Excavation quantities are based on 600 each contains different assumptions. Material costs are mm of cover and a payment width as defined in Clause 5.2 based on 2016 prices with delivery to central Durban. of SANS 1200DB. Manholes were priced as R25 000 each All prices include a 10% mark-up and exclude value and one was assumed every 60 m. added tax (VAT). Labour rates are legislated for the Civil Construction industry and were taken as R27/hr. All infrastructure was assumed to be under road ways and the reinstatement was estimated to be R420/m2. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 159 and the scaling relationship for pressurised flows is � = 0 � 0 ,where is the scaled area, 0 is the existing culvert area, is the scenario flow and 0 is the flow under the existing conditions (status quo). The culvert width is then determined by dividing the A3.5 Flow vs Dimension scaled area, Relationship A, by the culverts original height. To estimate the changes in cost a relationship between A3.5.3 Estimating the Flow Type flow and the infrastructure dimensions needs to be To determine which scaling relationship is to be used established. These flow relationships can be established A3.5.3 and theoretically for diameter Estimating the Flow area from uniform flowType the flow type needs to be estimated. The flow type is estimated by calculating the infrastructure’s maximum conditions. The relationship is then used to scale the To determine open channel flow from the Mannings equation. infrastructure dimensions for the which scaling different relationship scenario is to be used the flow type needs to be estimated. The flow This flow is referred to as the threshold flow and it is flows. A scaling relationship open exists forby type is estimated channel the infrastructure’s maximum open channel flow from the Mannings calculating calculated from flows and pressurised flows. Both of these flow equation. This flow is referred toare types as the threshold flow and it is calculated from estimated for each flow scenario. 5 1 3 1 2 0 2 ℎℎ = A3.5.1 Pipe Scaling 3 Where , the Mannings The scaling relationship for pipe diameters in openroughness, is assumed to represent concrete at a value of 0.015. is the Where n, the Mannings channel flow is cross-sectional area, is the perimeter and 0 is the slope of the infrastructure. roughness, is assumed to ipe diameters in open channel flow is represent concrete at a value of 0.015. a is the cross- 2 sectional area, P is the perimeter and S_0 is the slope of 5 If the scenario flow, , is less than the threshold flow the open channel scaling is used. If the the infrastructure. ipe diameters in open = 0 channel � � flow is 0 scenario flow exceeds the threshold flow then the pressurised scaling is used. 2 If the scenario flow, Q, is less than the threshold flow or pressurised flows is 5 the open channel scaling is used. If the scenario flow ipe diameters in open = 0 channel � �1 Two flow is and the scaling relationship other for conditions pressurised are included flows is exceeds that to ensure the threshold the scaling flow thennot does artificially inflate the pressurised scaling the benefit 2 0 or pressurised flows = is 0 � �5 of the natural areas. If the scenario flow is used. does not exceed the threshold flow and the existing flow ipe diameters in open = 0 channel �0 �1 flow is then no scaling is applied. If the scenario flow exceeds the threshold flow but does not exceed the eter, 0 is the existing pipe 0 2 diameter, is the scenario flow and 0 is other conditions are included to ensure that the Two or pressurised flows = is 0 � �5 existing flow then the open channel scaling scaling is does applied. not artificially These conditions inflate the benefit ensure ofthat the scenario flows onditions (status = quo). 0 �0 �1 0 2 diameter, natural that are larger than the existing flows but that do not require larger infrastructurethe areas. If the scenario flow does not exceed are not scaled. eter, 0 is the existing pipe is the scenario flow and 0 is threshold flow and the existing flow then no scaling is or pressurised flowsD = is 0 � � onditions (status ,where quo).is the 0 scaled diameter, D_0 is the existing pipe applied. If the scenario flow exceeds the threshold flow 1 eter, 0 is the existing pipe diameter, Q is the scenario flow and Q_0 is the flow but does not exceed the existing flow then the open 2 diameter, is the scenario flow and 0 is = the under 0 � existing � conditions (status quo). channel scaling is applied. These conditions ensure that onditions (status quo). 0 174that are larger than the existing flows but ulvert area in open channel flow is scenario flows eter, 0 is the existing pipe 4 diameter, is the scenario flow and 0 is do not require larger infrastructure are not scaled. that A3.5.2 Culvert onditions (status quo). Scaling 5 This means that artificial benefits are not attributed to ulvert area in The open = 0 � relationship channel � scaling 0 flow is for culvert area in open channel This the natural areas. It also ensures that the cost of over means that artificial benefits are not attributed to the natura flow is 4 design and future capacity are not penalised. or pressurised flows is 5 of over design and future capacity are not penalised. ulvert area in open = channel 0 � � flow is 0 � The followings is a summary of all the conditions = 0 � 4 The followings is a summary of all the conditions relevant to the s or pressurised flows is 0 5 relevant to the scaling: ulvert area 0 is the in open existing = culvert channel 0 � � area, flow isis the scenario flow and 0 is the flow 0 4 ns = 0 � � Condition 1: or (status quo). pressurised flows The is culvert 0 5 width is then determined by dividing If the > ℎℎ s original height. and = the scaling 0 � �relationship for pressurised flows is 0 is the existing culvert 0 area, is the scenario flow and 0 is the flow 1 = 0 � � then = � 2 � or = 0 � � ns or (status quo). pressurised flows The is culvert 0 width is then determined by dividing the 0 0 0 s existing culvert 0 is the height. original area, is the scenario flow and 0 is the flow ow Type = 0 � � ns (status quo). The culvert 0 width is then determined by dividing Condition the 2: If < ℎℎ and < 0 s 0 is the height. original existing culvert area, is the scenario flow and 0 is the flow relationship ow Type is to be used the flow type needs to be estimated. The flow 2 4 ns (status quo). ,where TheAculvert is the scaled width A_0 isdetermined is then area, by dividing then the existing culvert the = 0 � �5 or = 0 � �5 ing the infrastructure’s maximum open channel flow from the Mannings 0 0 s original height. area, Q is the scenario flow and Q_0 is the flow under d toType relationship ow as the threshold isthe to existing be used flow theandflow conditions it is (statuscalculated type needs quo).from to be The estimated. culvert width isThe flow Condition 3: then determined 5 by dividing the scaled area, A, by the ing the infrastructure’s maximum 1 3height. 1 open channel flow from the Mannings If < ℎℎ and ≥ 0 relationship is culverts to be used original = the flow 2 type needs to be estimated. The flow d toType ow as the thresholdℎℎ flow and 2 it is calculated from 0 then = or = ing the infrastructure’s maximum 53 open channel flow from the Mannings ghness, is assumed to 1 3 1 concrete at a value of 0.015. is the represent d to as the relationship threshold is ℎtobe used ℎ flow = the and flow it2is type 2 0 calculated needsfrom to be estimated. The flow perimeter and is the 5 slope of the infrastructure. A3.6 Cost Comparison ing the infrastructure’s maximum 0 3 open channel flow from the Mannings 1 3 1 ghness, is assumed to =represent it2isconcrete a value of 0.015. is the atfrom d to as the threshold ℎℎ flow and 2 0 calculated The difference between the existing infrastructure costs and th ess than the perimeter and 0 is the slope threshold flow 5 the open 3 of the infrastructure. channel scaling is used. If the theindicative value of the natural areas. ghness, flow reshold is assumed to 1 represent 3 A SPATIAL concrete 1 at aTHE value of 0.015. is the then the pressurised 2 scaling is used. Page 160 VALUATION OF NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY ℎℎ = 2 0 perimeter and 0 is the slope ess than the threshold flow 3 of the the infrastructure. open channel scaling is used. If the ghness, cluded reshold toisensure flowassumed thenthat the to represent the pressurised scaling does concrete scaling not at a value inflate artificially is used. of 0.015.the isA3.7 benefit the Additional information A3.6 Cost Comparison The difference between the existing infrastructure costs and the scenario infrastructure costs are the indicative value of the natural areas. A3.7 Additional information The items included for the costing of each category are shown in Table A3. 2, Table A3. 3, Table A3. 4, Table A3.5. Table A3. 6 shows the linear meter cost of concrete stormwater pipes. Table A3.2 Construction rates applied to the bridge culvert category Table A3.4 Construction rates applied to the canal and culvert category Description Unit Rate Description Unit Rate Excavation m3 R 90.00 Excavation m3 R 90.00 Selected fill m3 R 80.00 Selected fill m3 R 80.00 Backfill m 3 R 60.00 Backfill m 3 R 60.00 Blinding m3 R 1 600.00 Blinding m3 R 1 600.00 Shuttering m2 R 650.00 Shuttering m2 R 650.00 Steel t R 12 000.00 Steel t R 12 000.00 Concrete m3 R 2 400.00 Concrete m3 R 2 400.00 Reinstatement m2 R 420.00 Reinstatement m2 R 420.00 Preliminary and Preliminary and % 20 % 10 general items general items Table A3.3 Construction rates applied to the bridge pipe category Table A3.5 Construction rates applied to the pipe category Description Unit Rate Description Unit Rate Refer to Error! Refer to Error! Pipe m Reference source Pipe m Reference source not found. not found. m3 R 80.00 Excavation m3 R 90.00 Excavation m 3 R 90.00 Bedding m 3 R 320.00 Bedding m 3 R 320.00 Selected fill m3 R 80.00 Concrete m 3 R 2 400.00 Backfill m3 R 60.00 Reinstatement m 2 R 420.00 Manholes m2 R 25 000.00 Preliminary and % 10 Reinstatement m2 R 420.00 general items Preliminary and % 10 general items A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 161 Table A3.6 The cost of pipes in 2016 delivered to central Durban Diameter (m) Rate (R/m) Diameter (m) Rate (R/m) 0.3 313.50 0.9 1383.80 0.375 424.60 1.05 1777.60 0.45 555.50 1.2 2284.70 0.525 641.30 1.35 2649.90 0.6 1136.30 1.5 3634.40 0.75 1060.40 1.8 5195.30 0.825 1204.50 Page 162 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY APPENDIX 4: SEDIMENT AND NUTRIENT MODELLING A4.1 Introduction The contribution and retention of natural areas to nitrogen (TIN), phosphorous (P) and sediment loads (TSS) was determined under two different scenarios. These two scenarios were chosen in order to provide a range of what the water quality changes would be considering a conservative and a worse case assumption. A4.2 Model setup The water quality parameters chosen for this study were nutrients (nitrogen, phosphorous) and total suspended solids (TSS). The input parameters for each pollutant are as follows: ƒƒ the pollutant name ƒƒ the concentration units (i.e. mg/L, µg/L, counts/l) ƒƒ concentration in rainfall ƒƒ concentration in groundwater ƒƒ concentration in direct infiltration/inflow pollutant washoff from a given land use occurs during periods of wet weather and can be characterized in ƒƒ first-order decay coefficient SWMM5 by either using an exponential or rating curve relationship. The Event Mean Concentration (EMC) is ƒƒ It is possible to assign a co-pollutant as a fixed fraction a case of Rating Curve Washoff where the exponent is of the runoff concentration of the main pollutant. 1.0 and the coefficient represents the washoff pollutant However, this was feature was not used. concentration in mg/L. In each case pollutant buildup is continuously depleted as washoff proceeds, and washoff The landuses that generate these pollutants were ceases when there is no more buildup available. The defined and the pollutant buildup, pollutant washoff EMCs were derived from literature (Table A4. 1). These and street cleaning parameters were assigned to each data were applied to the different landuse categories landuse. Note that no data was available to estimate across the study area. The data below can be applied in the pollutant buildup and street cleaning parameters determining the water quality volume to be catered for and therefore these features were not considered. The in stormwater management devices (e.g. SUDS). Table A4.1 Event Mean Concentration data used in water quality modelling. Landuse Description TSS (mg/l) BOD (mg/l) TIN (mg/l) P (mg/l) Settlement - urban 100 15 3.41 0.79 Commercial / Retail / Institutional 166 9 2.1 0.37 Industrial / Road & Rail 166 9 2.1 0.37 Extractive / Utility 166 9 2.1 0.37 Farming / plantations & woodlots 201 4 1.56 0.36 Recreational open space 201 4 1.56 0.36 Settlement - rural 201 4 1.56 0.36 Natural vegetation (D'MOSS) 70 6 1.51 0.12 Settlement - informal 497 22 6.7 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 163 100 15 3.41 0.79 166 9 2.1 0.37 166 9 2.1 0.37 166 9 2.1 0.37 201 4 1.56 0.36 201 4 estimated in The TSS load PCSWMM accounts 1.56 0.36 for the A4.2.2 Rainfall total suspended sediments generated due to catchment Real-time rainfall data was applied to the models as 201 runoff and4 does not account 1.56 0.36 of for the proportion described in Appendix 2. The simulations were run from 70 sediment6transport activated 1.51 from the river0.12bed (i.e. the 1 August 2013 until 30 July 2014. Note that sediment 497 bedload). While bedload transport 22 is the dominant 6.7 mode yields vary spatially and temporally and therefore a one for low velocity flows and/or large grain sizes, suspended year simulation is not indicative of the mean annual load transport is the dominant mode for high velocity sediment yield. The annual rainfall (572 mm for Durban M accounts for the total and/ro suspended fine grain sediments sizes (Chadwick generated et al., due to 2013). In South city central) experienced during this period was below ount for the proportion of sediment Africa, a factor transport of 1.25 is generally activated applied from to cater forthe the MAP of Durban (1000 mm) and therefore simulated bed load and non uniformity in suspended bedload transport is the dominant mode for low velocity flows sediment results are conservative. Note that Rooseboom and concentrations in order to estimate the mean annual Lotriet (1992) suugest that six years of continuous d load transport is the dominant mode for high velocity and/ro sediment load (Msadala et al. 2010, after Rooseboom mintitoring is requied to obtain a reasonable estimate Africa, 013). In South1992). a factor Cooper of 1.25 (1993) is generally referenced applied estimates of theto cater of the average sediment load of a typical South African proportion of bedload in KwaZulul-Natal n suspended sediment concentrations in order to estimate the rivers from river. other studies to range from 12 to 50%. ala et al. 2010, after Rooseboom 1992). Cooper (1993) referenced load in KwaZulul-Natal rivers from other studies to range from 12 A4.2.1 Pollutant reduction through dams A4.3 The impact of natural systems on water quality Routing river flows through dams is a fairly complex exercise. There are a number of large dams in the Two scenarios were simulated to provide an indication of the impacts of the loss of natural areas on sediment and rough dams EMA. Therefore, in order to provide continuity through nutrient loads entering surface waters and an estimation these dams and best represent the impacts of dams on flood flows, the model was split by routing all of the water quality amelioration benefit obtained by s a fairly complex exercise. There are a number of large dams in the flows entering the dam to an outfall. The model retaining these natural systems. provide continuity startedthese through was then dams and best again downstream of therepresent dam. The the The first hypothetical scenario describes a situation simulated outputs for estuaries with dams situated e model was split by routing all the flows entering the dam to an without the pollutant retention benefits of natural upstream therefore ted again downstream of the only dam. represent the catchment The simulated yield outputs for vegetation. The pollutant retention function of green downstream of the dam. eam therefore only represent the catchment yield downstream of areas was ‘switched off’ to provide an estimate of the The TIN and P loads were assumed to be conserved ecosystem services provided by natural elements of through the dams and therefore the total TIN and P the landscape. This is considered to be a conservative loads into the dams were added to the estimated loads approach as it did not require any assumptions of how at the through d to be conserved respectivethe dams and estuaries therefore situated the total downstream TIN of these the landuse would change under different development dded to the estimated dams. Cooper loads at the (1993) respective suggested estuaries that dams situated in uMgeni scenarios. River catchment reduce sediment yield, and r (1993) suggested that dams in uMgeni River catchment reduce in particular The EMC of the different pollutants were altered for bedload, to the coast. TSS removal through dams was edload, to the coast. TSS estimated basedremoval through given on the equation dams bywas Fair estimated et al. natural vegetation based on the pollutant loading r et al. (1958) (1958) and tested by Kuo reduction factor (Table A4. 2). For example, if grass areas and tested by (1976), where Kuo (1976), where are estimated to reduce pollutant loadings by 50%, we assumed that the EMC from that type of landuse was − double the EMC ascribed to that landuse. = 1 − �1 + � ∗ where 178 EFF is the Removal Efficiency Factor i.e. fraction of TSS load which is deposited in the tank/basin or removed by other means, V_s is the particle settling velocity (m/s), typically between 0.00011 and 0.00058 m/s, n is the turbulence factor, where n = 1 means significant, n = 5 is low turbulence, Q is the inflow and A is the surface area of the basin. Water quality data for the WWTWs were taken from measurements taken by the EM at the respective WWTWs outfalls, where TIN = ammonia + nitrates + nitrites, P = orthophosphates and TSS = suspended solids. The average concentration was calculated for each pollutant. Page 164 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A4.2 Assumed pollutant reduction (%) based on 0.5-year return period event (Georgia 2015). Approach TSS TIN Total P Grass channel 50 20 25 Gravity oil / grit separator (industrial areas) 40 54 5 Permeable paving 80 50 50 Stormwater pond 80 30 50 Revegetation / reforestation 80 25 50 Vegetated filter strip 60 20 20 The results from the second hypothetical scenario determine the impacts on water quality if all conservations areas (D’MOSS areas) were replaced by dense rural informal settlement. This scenario provides a worst case scenario as changing the landuse from natural vegetation to dense informal settlement has a dual negative impact, i.e. informal settlements are associated with a high event mean concentration of pollutants and the removal of natural vegetation reduces the uptake of nutrients and sediments. The landuse file was edited and all natural vegetation was replaced with rural informal settlement. The area weighting tool was used to import the new shapefile into SWMM with the corresponding changes in hydraulic, soil and water quality properties associated with the change in landuse. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 165 This page intentionally blank. Page 166 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY APPENDIX 5: WATER TREATMENT COST ANALYSIS A5.1 Introduction ƒƒ Wasted water on more frequent backwashing to clean clogged filters; and Both high sediment and nutrient loads lead to increased water treatment operation and maintenance costs. ƒƒ Increased nutrients require the use of more complex Increases in sedimentation and associated increases and costly treatment technologies. in turbidity in river systems have become a common challenge facing many cities worldwide (McDonald & Shemie 2014). Changes in turbidity affect more than A5.2 Water treatment in eThekwini just the river ecosystem but also shorten the life span of storage dams, increase water treatment costs and Municipality change the nature of coastal areas (McDonald & Shemie The eThekwini Municipality receives bulk potable water 2014). Cities with lower than average forest cover and from Umgeni Water, a National Government Business a higher amount of agriculture within the catchment Enterprise and the largest water supplier in KwaZulu- generally have the highest levels of sediments and Natal accounting for a total bulk water sales volume of nutrients entering aquatic systems (McDonald & 423 million kilolitres per annum and serving a total of Shemie 2014). An accumulation of the essential plant 6.1 million people (Umgeni Water 2015). Umgeni Water nutrients, nitrogen and phosphorous, in rivers and dams extracts the raw water from a number of dams within can cause excessive algal growth and can lead to algal the uMngeni Basin that are linked to water treatment blooms. In most freshwater systems, phosphorous is the works throughout the province. Two of the largest water limiting nutrient for plant growth and large increases treatment works in the province are located in the EMA. in phosphorous entering waterways can lead to algal Raw water from Nagle Dam is supplied to The Durban blooms. Phosphorous is found in a number of common Heights WTW (Figure A5. 1) which produces 523 ML items such as fertilizer, manure, detergent, human or 42% of Durban’s potable water and raw water from waste and decaying plants, all of which enter waterways Inanda Dam supplies water to Wiggins WTW which through agricultural and urban runoff, poor sanitation produces 267 ML or 21% of the city’s potable water. within the catchment and industrial discharges. During Once the water is treated, it is reticulated to domestic high rainfall events the amount of total phosphorous and industrial consumers within the EMA. Although the entering rivers and dams increases resulting in the bulk water is purchased from Umgeni Water who run the potential for algal blooms and increased water treatment WTW’s, the bulk distribution and downstream pipelines costs. are owned and operated by the eThekwini Municipality. In the context of treating water the biggest effects of increased sediment and nutrient loads on the cost of treating water are listed below (Source: Graham 2004, McDonald & Shemie 2014, Rangeti 2014): ƒƒ Increased usage of coagulants; ƒƒ Increased amount of time water spends in settling ponds; ƒƒ Increased wastewater sludge (costly to treat and transport); ƒƒ Increased sediment loads preventing adequate filtration and disinfection of other pathogens and algae; ƒƒ Increased occurrence of algal blooms and toxic algae; ƒƒ Increased dominance by blue-green algae; ƒƒ Clogging of reticulation systems by filamentous algae; ƒƒ Increased occurrence of taste and odour issues in potable water and the need for activated carbon usage to eliminate these; A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 167 Figure A5.1 Durban Heights WTW with a capacity of 615 Ml/day is the largest WTW operated by Umgeni Water and is located within the EMA (Source: Umgeni Water) The water treatment process at Durban Heights and calcite or CaCo3) is most often used to adjust the pH. Wiggins WTW are very similar and involves a number Following this, a process of mixing, coagulation and of different physical and chemical treatment methods, flocculation processes occur whereby a coagulant is such as filtration and coagulation. The principle objective used to coagulate unwanted particles and algae into of treating water is to produce water at a reasonable larger flocs, which then sink to the bottom of the cost and in a reliable manner that is fit for domestic sedimentation tank (Graham 2004, Rangeti 2014). At use (Rangeti 2014). However, chemical dosage and Durban Heights and Wiggins WTW, a polyelectrolyte (a operating costs at WTWs depend primarily on the man-made organic compound) is used as a coagulant in quality of the raw water abstracted from the dams with the treatment process because of its effectiveness over temporal variation having large influences on these a wide range of pH thus reducing pH adjustment costs costs (Graham et al. 2012). Within a conventional water and because it has low water retention capacity it also treatment process coagulants, disinfectants, oxidants reduces floc disposal costs (Rangeti 2014). The clear and pH adjusters are used throughout. In the first water is then passed through large filters that remove step of the treatment process the water abstracted any remaining suspended matter. The final step is the from the dams passes through wire screens to remove post-treatment disinfection process that uses chlorine to any solid objects. The water is then pre-treated using kill any remaining pathogens and unwanted algae. The chlorine as a disinfectant, inactivating any disease treated water is then held in reservoir tanks and pumped causing pathogens and reducing odour, colour and taste to industrial and domestic consumers throughout the problems (Rangeti 2014). The level of the pH in the raw EMA. Table A5. 1 outlines the operating characteristics water is adjusted to create an optimum environment of Durban Heights and Wiggins WTWs. for effective coagulation. Lime (calcium carbonate, Page 168 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A5.1 Operating characteristics for Durban Heights and Wiggins WTWs (Source: Umgeni Water) Landuse Description TSS (mg/l) BOD (mg/l) River system Lower uMngeni Lower uMngeni Northern areas of EMA: e.g. Newlands, Southern areas of the EMA: e.g. Areas serviced KwaMashu, Ntuzuma, Phoenix, Durban Woodlands, Isipingo, Lotus Park, North & Umhlanga Amanzimtoti & KwaMakuta areas Maximum capacity 615 Ml/day 350 Ml/day Current utilisation 497 Ml/day 258 Ml/day Pre-oxidation type Prechlorination Ozone Primary pre-treatment chemical Polymeric coagulant Polymeric coagulant Clarifier type Pulsator clarifier Pulsator clarifier Number of clarifiers 18 4 Filter type Constant rate rapid gravity filters Constant rate rapid gravity filters Number of filters 100 24 Capacity of backwash water tanks 2326 m3 - Capacity of sludge treatment plant 30 000 kg/day thin sludge - Post disinfection type Chlorine gas Hypochlorite A5.3 Current state of the uMngeni River agricultural runoff and alien vegetation in the riparian Catchment Area zone (Dennison & Lyne 1997, Graham 2004, Rangeti 2014). Many of these issues are a result of rapid The uMngeni River is one of the most developed urbanisation throughout the catchment and all influence catchments in South Africa and regionally is an area of nutrient enrichment and sedimentation levels in the major economic, cultural and ecological importance uMngeni River and associated water supply dams. The (WRC 2002). The uMngeni River originates in the Lower uMngeni System serves the greater eThekwini uMngeni Vlei in the highlands of the KwaZulu-Natal Municipal Area, deriving potable water from Nagle and Midlands, approximately 1900m above sea level Inanda Dams which are supported by Albert Falls and (Graham 2004). The river flows through the KwaZulu- Midmar Dams further upstream (Umgeni Water 2015). Natal Midlands, Howick, Pietermaritzburg, the Valley of a Nagle Dam with a capacity of 24.6 million m3, although Thousand Hills and down into Durban emptying into the located outside of the EMA, supplies water to Durban Indian Ocean just north of Durban Bay. Major tributaries Heights WTW and Inanda Dam with a capacity of 251.6 include the Karkloof, Lion, uMsunduze and the Mqeku million m3 supplies water to Wiggins WTW, both of rivers. There are approximately 130 registered dams in which are located within the boundaries of the EMA the uMngeni catchment but there are five that are large (Umgeni Water 2015). and of significance in terms of water supply; Midmar, Albert Falls, Nagle, Henley and Inanda Dam having a combined capacity of 753 million cubic meteres of A5.4 Data and methods water (WRC 2002). These resevoirs supply the major Pietermaritzburg-Durban complex, home to about 45% Umgeni Water has an extensive water quality monitoring of the provincial population (Graham 2004). programme that covers all of its WTWs and supply dams. The modelling of nutrient and sediment levels against The water quality in the uMngeni River is generally various cost data was based on water quality, volume, fair upstream and deteriorates significantly as the chemical dosage and cost data supplied by Umgeni river moves towards the coast passing through a Water for Durban Heights and Wiggins WTW for a five number of large urban and peri-urban centres. The year period from 1 July 2010 – 30 June 2015. Water major management concerns include pollution from quality data was supplied for the raw water entering the overflowing sewers, pollution from informal settlements, treatment works as well as water quality data for the illegal discharges, industrial discharges, solid waste final potable water. This data included all water quality pollution, overgrazing causing increased erosion, parameters measured at the WTWs such as turbidity, A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 169 alkalinity, pH, algal counts, iron, nitrates etc. Chemical River and water supply dams and this data was used to dosage and related costs were also supplied for the determine the link between Nitrogen and Phosphorous same period and outlined the amount of each type of and algae in the treatment water so as to determine the chemical used during different stages of the treatment treatment cost savings involved in the reduction of these process and the associated individual and overall costs. nutrients into the water supply dams. Water quality Daily water volume data for both treatment works was data for monitoring sites on the uMngeni River above supplied for the same period. Nagle and Inanda Dam were provided for the same five year period to assess the water treatment cost savings The water quality parameters measured in the raw associated with reduced nutrients entering the supply water at the WTWs did not include phosphorous or total dams. River volume and flow data was downloaded nitrogen. These parameters are measured bi-monthly at from the Department of Water Affairs (DWA) interactive a monitoring station in the water supply dams and in the website for water monitoring stations above the dams. uMngeni River above the dams and only the quantity of Table A5. 2 and Table A5. 3 provide the summary different types of algae are measured in the raw water statistics for the Durban Heights and Wiggins WTW entering the WTW. Total Phosphorous (TP), Nitrate, model sample data. Nitrite and Ammonia are measured in the uMngeni Table A5.2 Summary statistics of water quality, volume and cost data from July 2010 – June 2015 from Durban Heights WTW. Variable Min Mean Max Variable Min Mean Max Treatment Cost (R/ML) 40 72 148 Oscillatoria 0.0 47.5 523.0 Volume (ML) 11 482 15 434 17 992 Ceratium 0.0 10.7 51.5 Total Algal Count (cells/mL) 126 955 8 016 Chlamydomonas 0.0 25.3 111.8 Alkalinity (mg CaCO3/L) 29 36 44.7 Chlorela 0.0 5.8 43.0 Colour (°H (mg Pt-Co/L) 1.1 4.8 36.3 Coelastrum 0.0 133.1 545.0 E.coli (MPN/100mL) 0.0 8.3 115.3 Crucigenia 0.0 28.3 455.0 Coliforms (MPN/100mL) 77 1 146 4 839 Cryptomonas 0.0 22.9 137.0 Turbidity (NTU) 2.7 6.1 21.1 Cyclotella 0.0 62.3 515.1 TOC (mg/L) 1.6 2.8 4.7 Cymbella 0.0 7.7 54.0 Suspended Solids (mg/L) 2.1 6.5 42 Dictyosphaerium 0.0 16.7 547.0 pH 7.4 7.8 8.2 Fragileria 0.0 446.2 6869.9 Conductivity (mS/m) 10 13.1 17.4 Pandorina 0.0 47.0 339.7 Temperature (°C) 15 21.1 31 Oocystis 0.0 26.2 127.0 Fe (mg/L) 0 0.1 0.9 Scenedesmus 0.0 41.3 241.2 Algal genera (cells/mL): Quadrigula 0.0 27.6 290.7 Anabaena 0 53.2 1571.7 Sphaerocystis 0.0 81.5 717.0 Microcystis 0 121.4 1331.3 Melosira 0.0 89.8 362.0 Page 170 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A5.3 Summary statistics of water quality, volume and cost data from July 2010 – June 2015 from Wiggins WTW. Variable Min Mean Max Variable Min Mean Max Treatment Cost (R/ML) 34 67 130 Total Algal Count 20 1954 14601 Volume (ML) 7093 9075 17992 E.coli (MPN/100mL) 0.0 1.4 30.4 Coliforms Alkalinity 48.6 60.5 68.1 66.5 1731.8 5000.0 (MPN/100mL) Algal genera (cells/ Colour (°H (mg Pt-Co/L) 1.4 3.1 10.9 mL): TOC (mg/L) 1.91 3.23914 5.79 Microcysitis 0.0 512.2 12404.6 Turbidity (NTU) 0.5 1.2 2.3 Anabaena 0.0 16.4 388.5 TOC (mg C/L) 1.9 3.2 5.8 Chlamydomonas 0.0 21.5 159.0 Suspended Solids (mg/L) 2.2 4.6 9.3 Cryptomonas 0.0 22.2 80.0 pH 7.3 7.8 8.8 Cyclotella 0.0 23.1 267.5 Conductivity (mS/m) 19.1 24.9 27.9 Fragilaria 0.0 874.3 7038.3 Temperature (°C) 15.5 21.3 26.3 Melosira 0.0 39.8 268.0 Iron (mg/L) 0.0 0.1 0.8 Quadrigula 0.0 14.6 372.0 Manganese (mg/L) 0.01 0.03 0.12 Ceratium 0.0 19.9 147.0 Fluorine (mg/L) 99.0 140.4 183.0 Scenedesmus 0.0 100.7 967.8 Water treatment cost models were estimated for turbidity, suspended solids, E.coli and total organic Durban Heights WTW and Wiggins WTW based on data carbon (TOC). Suspended solids and turbidity were also provided by Umgeni Water. Durban Heights is the largest highly correlated. In order to link sediment loads (kg) to water treatment works and Wiggins is the second largest water treatment costs it was more practical to include in the EMA supplying two thirds of the municipalities suspended solids in the model which is measured in mg/L water and therefore formed the focus of this study. The rather than turbidity which is measured in Nephelometric benefits of improved water quality were estimated using Turbidity Units (NTU). A number of algal genera were regression models that link turbidity and nutrients to included in the model, rather than just the total algal water treatment costs at Durban Heights and Wiggins count. Stepwise regression methods were used to WTWs. Changes in sanitation and land use within the eliminate variables that were not contributing to the uMngeni catchment influences the amount of sediment, overall fit of the model. Month was included in the model phosphorous and nitrogen entering river systems and to capture seasonal trends in water treatment costs. water supply dams. Using hydrological modelling, sediment and nutrient export output can be established A separate set of models were developed in order to for the catchments within the EMA. These values can relate the phosphorous loads entering the water supply then be used to determine water treatment cost savings dams to water treatment costs. Higher phosphorous with changes in catchment land use by applying the levels in the uMngeni River entering Nagle and Inanda water treatment cost model and values established for Dams were expected to be related to increased water sediment and nutrient loads. treatment costs as a result of increased algal activity. This model was used to link changes in catchment land The water treatment cost generalised linear models were use to changes in nutrients entering water supply dams analysed by entering all variables into an ordinary least and the subsequent decrease in water treatment costs. squares (OLS) regressoion using R Project for Statistical Treatment costs (Rands per Mega-litre) were related Computing (ver. 3.2.0). All costs were inflated to 2015 to phosphorous loads in the river and other water Rands. Collinearity was a concern with this model due quality variables such as coliforms and water colour. to the large number of water quality variables that are Because phosphorous loads were highly correlated with related to one another. As a result a number of stepwise TSS, turbidity, iron, nitrogen, pH and E.coli they were collinearity methods were used for determining which removed from the model through AIC stepwise analysis. variables to retain and which to remove from the model. For example, iron was highly correlated with A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 171 A5.5 Results A5.5.1 Durban Heights WTW The total water treatment costs at Durban Heights WTW in terms of chemical cost composition were similar to those described by Graham (2004). Caogulant (and flocculation) costs and disinfection costs contributed the most to the overall treatment costs with coagulant costs contributing 41.4% and disinfection costs 42.9% (Figure A5. 2). The coagulant and disinfection processes during water treatment remove unwanted sediments, algae, odours and colour from the raw water. Figure A5.2 Average treatment cost breakdown at Durban Heights WTW. Treatment costs increased steadily between 2010 and 2015 (Figure A5. 3), with clear seasonal fluctuations. Treatment costs were higher over the summer months from November through to March when rainfall is at its highest in the EMA. Figure A5.3 Total water treatment costs for Durban Heights WTW from July 2010 – June 2015 Page 172 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure A5. 3. Total water treatment costs for Durban Heights WTW from July 2010 – June 2015 The regression estimates for the water treatment model where treatment cost (R/ML) is the dependent variable and a number of water quality parameters are the independent variables is shown in Table A5. 4. The model was highly significant with an R-squared of 0.922, representing a strong relationship between actual and predicted treatment costs (Figure A5.4). Table A5.4 Estimated coefficients of variables affecting water treatment costs at Durban Heights WTW Coefficient Std. Error t.value Pr(>|t|) Intercept -109.0000 103.2000 -1.0560 0.2990 Suspended Solids 0.9473 0.2600 3.6430 0.0010 *** pH 21.2400 13.1200 1.6180 0.1157 Coliforms 0.0049 0.0016 2.9860 0.0055 ** Anabaena 0.0172 0.0059 2.8970 0.0068 ** Ceratium 0.2282 0.1235 1.8470 0.0742 . Chlorella 0.5319 0.1270 4.1870 0.0002 *** Coelastrum 0.0173 0.0094 1.8420 0.0751 . Crucigenia -0.0903 0.0374 -2.4160 0.0218 * Cyclotella -0.0866 0.0257 -3.3700 0.0020 ** Dictyosphaerium 0.1690 0.0229 7.3780 0.0000 *** Scenedesmus -0.1602 0.0380 -4.2190 0.0002 *** Quadrigula 0.0809 0.0282 2.8660 0.0074 ** Sphaerocystis 0.0142 0.0091 1.5680 0.1270 Melosira -0.0418 0.0173 -2.4130 0.0219 * Month dummies significant R-squared 0.9220 Notes: (1) ***p<0.001, **p<0.01, *p<0.05, .p<0.10. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 173 Figure A5.4 Actual versus predicted water treatment costs fitted using the treatment cost model estimates for Durban Heights WTW Water treatment costs increased significantly in the in water. Difficult and costly treatment methods are summer months (November – March) as a result of needed to remove tastes and odours from water, such as increased rainfall and runoff resulting in increased powdered activated carbon (Graham 2004). Algal levels suspended solids and algal blooms within the raw water tend to increase when nutrient levels rise, especially being abstracted from Nagle Dam. Suspended solids when phosphorous levels are high in the inflow have a significant impact on treatment costs, as do the water, and this is often during summer months when level of coliforms in the water. High coliform counts are phosphorous levels increase due to high rainfall. These often associated with high rainfall events as a result factors are all directly related to water pollution as a of human and other wastes entering rivers and dams result of both anthropogenic inputs such as accelerated from urban, peri-ubran and rural settlement runoff. The erosion and sewage, and natural inputs such as siltation. rising costs associated with these factors are a result of Runoff from agricultural activities such as cattle lots increased usage of coagulants and disinfectants when and from informal settlements increase nutrient levels coliform and suspended solid concentrations are high. in rivers and dams. Figure A5. 5 shows the general A number of algae were found to significantly influence trend of increasing phosphorous and nitrogen loads treatment costs, with some having a positive impact and entering Nagle Dam and the seasonal peaks in nutrient others a negative impact on overall costs. Anabaena is concentrations over a five year period (data taken from a genus of filamentous cyanobacteria that can release the sampling station on the uMngeni River upstream of harmful toxins and also cause taste and odour problems Nagle Dam). Figure A5.5 (a) Total phosphorous concentrations (mg/L) Dec 2010 – April 2015 and (b) Total nitrogen concentrations (mg/L) Jan 2010 – April Page 174 A SPATIAL 2015 VALUATION at the inflow OF THE to Nagle Dam NATURAL on the AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY uMngeni River Phosphorous loads in the uMngeni River entering Nagle Dam were positively and significantly correlated with water treatment costs at Durban Heights WTW (Figure A5. 6, Table A5. 5). Phosphorous loads were strongly correlated with Turbidity, TSS, E.coli, Iron and pH levels in the raw water abstracted for treatment. Phosphorous loads (kilograms of TP) were the highest over the summer rainfall months from November through to March resulting in increased water treatment costs as a result of increased turbidity, algae and coliforms within Nagle Dam and the subsequent need for higher coagulant and disinfection chemical dosages at Durban Heights WTW. Figure A5.6 Average phosphorous loads (kg) in the uMngeni River above Nagle Dam and corresponding water treatment costs (R/ML) at Durban Heights WTW. Table A5.5 Estimated coefficients of variables affecting water treatment costs at Durban Heights WTW Coefficient Std. Error t.value Pr(>|t|) Intercept 47.780 23.007 2.077 0.045 * Phosphorous Load 0.0152 0.0045 3.381 0.002 ** Temperature 0.6847 0.6740 1.016 0.316 Coliforms 0.0056 0.0025 2.261 0.029 * Colour 0.5012 0.4034 1.242 0.222 Conductivity 0.9800 1.2397 -0.791 0.434 R-squared 0.481 Notes: (1) ***p<0.001, **p<0.01, *p<0.05, .p<0.10. The model relating water treatment costs to phosphorous loads in the uMngeni River was significant with a reasonable fit to the actual treatment cost data provided for Durban Heights WTW (Table A5. 5, Figure A5. 7). The model relates changes in sanitation and land use within the uMngeni catchment to changes in water treatment costs. It is assumed that improved sanitation would decrease the phosphorous loads entering Nagle Dam, decreasing the frequency and magnitude of algal blooms, coliform numbers and TSS and therefore ultimately decreasing water treatment costs at Durban Heights WTW. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 175 Figure A5.7 Actual versus predicted water treatment costs (R/ML) fitted using the treatment cost model estimates for Durban Heights WTW and the uMngeni River phosphorous loads. A5.5.2 Wiggins WTW The percentage breakdown of water treatment costs at Wiggins WTW are similar to those at Durban Heights WTW, with disinfection and coagulant costs making up 88% of total costs (Figure A5. 8). Water treatment costs increased between 2010 - 2015 (Figure A5. 9). On average water treatment costs are cheaper at Wiggins WTW than at Durban Heights WTW with a mean of R67 per ML compared to R72 per ML at Durban Heights. This is similar to the findings presented in Graham (2004). One explanation for this could be the size difference in the dams supplying the raw water to the treatment works. Inanda Dam is approximately ten times larger than Nagle Dam allowing for increased mixing and diluting of nutrients resulting in significant differences between the quality of water entering the dam and the Figure A5.8 Percentage treatment cost breakdown at Wiggins WTW. quality of water being abstracted from outlet point for treatment. Rangeti (2014) noted that although Inanda Dam is considered badly polluted, the sheer size of the dam allows for dilution and mixing which improves the water quality before abstraction. Page 176 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure A5.9 Total water treatment costs for Wiggins WTW from July 2010 – June 2015 The water treatment cost model results for Wiggins WTW are shown in Figure A5. 10 and Table A5. 6. The model was highly significant with Algae, Manganese, Alkalinity and the algal genera Anabaena having significant effects on water treatment costs. From the model it seems that algae has the largest influence on treatment costs at Wiggings WTW. Algal levels in Inanda Dam tend to increase during the summer rainfall months when nutrients entering waterways increase. Algae tend to increase filtration time and increase coagulation dosage during treatment (Rangeti 2014). Certain genera, such as the filamentous cyanobacteria, Anabaena, can have a significant impact on the taste and odour of the water being treated. When Anabaena cells are present in high numbers water treatment costs increase substantially. Figure A5.10 Actual versus predicted water treatment costs fitted using the treatment cost model estimates for Wiggins WTW A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 177 Table A5.6 Estimated coefficients of variables affecting water treatment costs at Wiggins WTW Coefficient Std. Error t.value Pr(>|t|) Intercept 122.70 23.010 5.33 4.38e-06 *** Alkalinity -1.33 0.390 -3.44 1.42e-03 ** Algae 0.004 0.001 7.70 2.37e-09 *** Manganese (Mn) 286.00 55.910 5.12 8.70e-06 *** Anabaena 0.047 0.020 2.09 4.32e-02 * R-squared 0.682 Notes: (1) ***p<0.001, **p<0.01, *p<0.05, .p<0.10. Phosphorous concentrations in the uMngeni River above Inanda Dam have increased since 2010 (Figure A5. 11a). Phosphorous loads (kg of TP) were calculated using these concentrations and flow data from the same monitoring station. It was found that phosphorous loads at this site follow the same pattern as those found in the uMngeni River above Nagle Dam (Figure A5. 11b). However, the TP loads above Inanda Dam are significantly higher than those above Nagle Dam. It is expected that the higher phosphorous loads are a result of the Msunduzi River which joins the uMngeni River before this monitoring point. The Msunduzi River is known to be highly polluted. Although the phosphorous loads entering Inanda Dam are higher than those entering Nagle Dam, the size of Inanda Dam has a large dilution and settling influence on the water in the dam. Phosphorous loads entering Inanda Dam were found to have a significant impact on water treatment costs at Wiggins WTW (Figure A5. 12). Figure A5.11 (a) Total phosphorous concentrations (mg/L) Dec 2010 – March 2015 in the uMngeni River above Inanda Dam, (b) Phosphorous loads (kg) in the uMngeni River above Nagle Dam and above Inanda Dam. Page 178 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Figure A5.12 Average phosphorous load (kg) in the uMngeni River just above Inanda Dam and corresponding water treatment costs (R/ML) at Wiggins WTW (lag adjusted). The model relating water treatment costs at Wiggins WTW to phosphorous loads in the uMngeni River was significant, although the fit of the model to the actual treatment cost data was not as strong as the model for Durban Heights WTW (Table A5. 7). The reason for this can be explained once again by the size of Inanda Dam and the impact that dilution and mixing has on the incoming phosphorous loads as well as the direct correlation that phosphorous loads have with algae, TSS and temperature. Although the fit of the model is not strong, the relationship between treatment costs and phosphorous loads is significant. The analysis showed that there was a possible slight lag in phosphorous loads entering Inanda Dam and the treatment costs at Wiggins WTW. There are unknown variables within the system that would influence phosphorous levels such as temperature, wind, degree of dilution and mixing and the amount of phosphorous entering the system directly through runoff from rural settlements surrounding Inanda Dam. Understanding and capturing these complex interactions is difficult but from the analysis it is clear that phosphorous loads in the uMngeni River do have a significant influence on water treatment costs. Table A5.7 Estimated coefficients of the water treatment cost - phosphorous load model for Wiggins WTW Coefficient Std. Error t.value Pr(>|t|) Intercept 135.10 37.38 3.615 0.0009 *** Phosphorous load 0.0049 0.002 2.807 0.0079 ** Alkalinity -1.11 0.61 -1.813 0.0779 Manganese (Mn) -334.70 1.04 -3.217 0.0026 ** R-squared 0.331 Notes: (1) ***p<0.001, **p<0.01, *p<0.05, .p<0.10. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 179 A5.6 Conclusions From this analysis it is clear that water quality had a significant impact on the cost of treating water at each of the WTWs studied. The models developed for Durban Heights WTW and Wiggins WTW are useful tools that can be used to predict the impacts of changes in water quality on water treatment costs. Regression coefficients, estimated using ordinary least squares, quantified the independent effects of each of the water quality parameters, such as phosphorous, as significantly effecting water treatment costs. These models can therefore be used to predict the outcome of catchment land-use changes, such as the impact that natural vegetation has on reducing nutrient runoff into surface waters. Using a simple scenario approach, comparisons between the status quo and modelled outputs can be related to the treatment costs and overall cost savings can be estimated. Page 180 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY APPENDIX 6: CONTRIBUTION OF GREEN OPEN SPACE TO PROPERTY VALUE Introduction A6.1 and across suburbs. In general it is expected that the proximity to well-maintained open space and views of It is well known that green open space and natural natural areas lead to increases in surrounding property areas in cities provide a number of benefits, such as prices, which ultimately manifests in greater property opportunities for recreation and tourism, attractive tax revenues for municipalities. views, habitat for wildlife and other ecological benefits such as improved air quality and biodiversity The policy implications of understanding and conservation (Beron et al. 2001, Anderson & West 2006, quantifying the economic value of green open space Kong et al. 2007, Gibbons et al. 2011, Li et al. 2015). can be beneficial to municipalities and urban planners Open space may also be valued based on an absence (Anderson & West 2006, Behrer 2010). The benefits of unpleasant qualities associated with development associated with environmental amenity are often in cities, such as noise, traffic congestion and pollution underestimated by policy makers and urban planners (Irwin 2002, Anderson & West 2006, Kroeger 2008). (Kong et al. 2007, Lee et al. 2015) which has resulted in As a result of rapid urbanisation in African cities and urban sprawl and the subsequent degradation and loss the subsequent increase in and sprawl of informal of green open spaces (Leiva & Page 2000, Kong et al. settlements, city managers and planners are constantly 2007). Local municipalities receive a significant portion weighing up between developing and preserving open of their revenue from property taxes and therefore have space areas. There is constant pressure to provide an incentive to maximise the value of properties within additional housing and to develop commercial and their jurisdiction. If policy makers are provided with industrial areas further. However, the way in which information about how open space influences property residents value open space areas and the benefits that sales and how their residents value open space areas, they provide in urban environments are often not well more informed decisions can be made in terms of land understood or are underestimated by urban planners use planning and future development. resulting in poor zoning and land-use policy decision making and the ensuing loss of green open space Literature looking at the effects of open space on (Anderson & West 2006, Behrer 2010). residential property values is large and growing. McConnell & Walls (2005) conducted a broad review of The value that residents place on open space is studies that have valued the effects of open space on reflected, to an extent, in private property and real property markets and found that although the results estate markets (Kroeger 2008). When prospective tended to be case specific, they were able to conclude homebuyers purchase a home they reveal certain information about the direction of certain effects and preferences for different characteristics of the property how values vary by location. Other reviews include through the amount that they are willing to pay for it Banzhaf & Jawahar (2005) who focused on undeveloped (Taylor 2003, Behrer 2010). Homes that have a higher land on the outskirts of cities and Kroeger (2008) who number of desirable characteristics are usually sold for developed an open space property value tool through a higher price when compared to homes with fewer a comprehensive review of open space valuation of these characteristics. Housing attributes include studies. Numerous studies have estimated the amenity physical characteristics of the property such as size of value of particular types of open space such as golf the living area, number of bathrooms, security, and courses, neighbourhood parks, wetlands, forests and condition of the property, neighbourhood characteristics agricultural lands (e.g. Do & Grudnitski 1995, Doss & such as schools and crime levels, and environmental Taff 1996, Mahan et al. 2000, Tyrvainen & Miettinen characteristics such as scenic views and the amount of 2000, Owusu-Edusei & Espey 2003, Nicholls & Crompton green open space surrounding a property. If residents 2007, Gibbons et al. 2011), whilst others have focused do value open space and associated amenities then on a spatial context of open space and estimated the it would be expected that these values should be effects of proximity, density and quality of open space revealed in property prices. Property values associated on property values using land cover data and spatial with open spaces are therefore expected to vary analysis (e.g. Smith et al. 2002, Anderson & West 2006, according to factors such as the size of the open space, Cho et al. 2006, Kong et al. 2007, Long et al. 2007, Cho et its distance from property, the type of vegetation al. 2008, Behrer 2010, Jensen et al. 2014, Li et al. 2015). cover and its overall appeal (Kong et al. 2007, Kroeger Anderson & West (2006) found, as expected, that the 2008). Household income is also considered to be an value of proximity to open space in the Minneapolis - St. important variable (Kroeger, 2008), and this is especially Paul metropolitan area was higher in dense, high-income relevant in the eThekwini Municipality in Durban where areas that were close to the Central Business District annual income varies significantly amongst households (CBD) and were home to many children. Surprisingly, A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 181 neighbourhood parks and the amount of natural open space surrounding each p forests, grassland and woodland). The condition of each natural open space pa into consideration property and forms value of a number ofan important different part of types of the study. open Geographic space; Informa proximity to was used to determine spatial variable estimates. The effects of neighbourhood parks and the amount of natural open space surrounding each propepopulation den household forests, grasslandincome and other covariates and woodland). believed The condition to influence of each the space natural open of ope value patch w into included the study found that residents in high-crime in theand consideration areas also analysis. forms A6.2 an important part of the study. Geographic Information Methods openused placed a higher value on proximity to was space areas to determine spatial variable estimates. The effects of population density suggesting parks buffer against the negative effects of household income and The covariates other A6.2.1 econometric believed model to influence the value of open sp crime (Anderson & West 2006). A study conducted in A6.2 included Jinan City, China found that the size-distance Methods of analysis. in the index Property value associated with environmental assets is forest, the accessibility to green parks and plazas, and generally estimated using the Hedonic Pricing Method the percentage of green urban space wereA6.2.1 of multiple regression analysis. The HPM (HPM), a form model The econometric all significant and confirmed the positive amenity impact of proximate assumes that the final price of a good is a function of green space on property prices (KongA6.2et al. Methods 2007). Property Troy the values of value associated theenvironmental with individual attributes assets(Rosen 1974). estimated using t is generally & Grove (2008) found that park proximity is positively Therefore the method is used to measure the implicit A6.2.1 valued by the housing market in Baltimore Method City, The (HPM), a form econometric value of of a multipleunderlying model property’s regression analysis. The characteristics HPM assumes that (i.e. Maryland where the combined robbery and goodrapeis a function rates of the values structural, of the locational individual attributes and environmental; Gibbons et al. 1974). There (Rosen for a neighbourhood were below a certain threshold usedvalue Property to measure 2011). associated theA HPM relates implicit with the environmental market value ofassets price a property’sof a property is generally to underlying estimated characteristi using the H rate but negatively valued when above that threshold. these attributes, with each property owner choosing Methodlocational (HPM), and environmental; a form of multipleGibbons regressionet al . 2011). The analysis. their property based on utility maximization given A HPMHPM relates the market assumes that the prf Gibbons et al. (2011) estimated the amenity value of nature on property prices in the Unitedgood these Kingdom attributes, is a function and ofbywith the the eachfunction values price property of owner the individual (Taylor 2003,choosing attributes Anderson their (Rosen property & West 1974). based on u Therefore found that gardens, green space and areas given of by water the all 2006). price Using function standard (Taylor hedonic 2003, theory Anderson (Rosen & used to measure the implicit value of a property’s underlying characteristics ( 1974) West 2006). Using standa attract a considerable positive price premium. They also each property in in thisthis study was defined by its structural, (Rosen locational and1974) each environmental;property Gibbons et al.study 2011). was A HPM defined relates bythe structural, itsmarket price no found that increasing distance to natural amenities such neighbourhood and environmental characteristics and these as rivers and national parks is associated environmental with attributes, a decreasewith characteristics theeach property general the andowner hedonic general price choosing function hedonic was their price function property applied: wason based applied: utility in property prices. given by the price function (Taylor 2003, Anderson & West 2006). Using standard h (Rosen 1974) each , = ( ,in property ) study was defined by its structural, neighb this In Durban it is expected that the socio-economic variables such as household income and population characteristics and the general hedonic price function was applied: environmental where density could have a significant influence on property is thewhere sales P_p price is the of sales the propertyprice of the and property , and , x ,x ,xe structural, n s n the are values and the amenity value associated with open are the structural, neighbourhood and environmental environmental = ( , , )related to . characteristics space. One would expect that people with higher characteristics related to Pp. incomes are more willing to pay for open space or where is the structure sales Beyond price the of structure the property ofthere the and model athere isto a need to structural, scenic views when compared to people Beyond with very little of the model is ,need , are determine the which neigh funct determine which functional form to use, something that disposable income. People living in denser areas environmental may something characteristics related to that is is not not specified specified inin thethe . economic economic theory theory surrounding surrounding hedonic pric place a higher value on open space and the recreational 1974, McConnell opportunities that they provide, such as open space hedonic & Walls price 2005, equations Kong (Rosen et al. 1974, 2007,McConnell Behrer 2010, & Walls Gibbons et al. number Beyond the of structure functional 2005, of Kong forms the model et al. that 2007, can there be Behrer used is a 2010, for need Gibbons the to regression determine et al. 2011). analysis, which such as li functiona for practicing football, however those living in dense There are a number of functional forms that can be used lower surroundings usually form part of thesomething log income and that quadratic is not forms (see specified Taylor in the economic 2003). theory The standard surrounding in more hedonic recent studie price eq for the regression analysis, such as linear, semi-log, log- group in the population with their willingness to pay for 1974,Troy McConnell & Grove& 2008, Walls log and Poudyal 2005,quadratic et al. Kong et 2009, forms al. (see Conway 2007, Taylor Behrer et al.The 2003). 2010, 2010, standard Gibbons et al. al. 2011 2011 open space often not reflected in property prices. been number ofto use the in functional semi-log forms more that recent regression studies can be used (Cho model for etwhere al. the 2008, regression the Troy dependent & Grove analysis, variable such as islinear the n This study aims to determine the contribution of the quadratic green sale price for 2008, each Poudyal property. et al. 2009, This is the Conway et al. 2010, Gibbons log and forms (see Taylor 2003). Thefunctional standard in form more applied recent to studies the model (Ch open space to residential property value in the EMA. et al. 2011, Li et al. 2015) has been to use the semi-log TroyThis & Grove approach 2008, allows Poudyal regression the et coefficients model al. 2009, where Conway of thea dependent et al. 2010, semi-log model variable Gibbons toisbe the interpretedet al. 2011, Liae as Using a hedonic analysis of property sales data from been the EMA we estimate the effects on property in toprices use the value of(Kong semi-log et natural al. 2007, regression logarithm Gibbons model et al of the . 2011). where sale pricethe for dependent each property. variable is the natur a number of different types of open space; proximity This the sale price for each property. This is the functional form applied to the model use is the functional form applied to the model used to the coastline, neighbourhood parks andInthe amount in this study. This approach allows the coefficients of This this approach study, allows the thesemi-log coefficients a semi-log model to be interpreted as a percentage model of was a semi-log specifiedmodel by a natural to be interpreted log transformation as a perc o of natural open space surrounding each property (e.g. price in prices rivers, forests, grassland and woodland). The condition in the (Kong ethedonical. change 2007, regression. in prices et Gibbons The (Kong al.model2011). et al. 2007, applied Gibbons is as et follows: al. 2011). of each natural open space patch was also taken into In this study, the semi-log model was specified by a consideration and forms an important Inpart this of the the semi-log study, log =was + 1 0 specified +byhousing 2 + sales 3 price + naturalmodel transformation of the a natural log transformation in of the study. Geographic Information Systems (GIS) was used price in the hedonic regression. the hedonic The model The regression. applied model is applied as follows: is as follows: to determine spatial variable estimates. The effects of population density, amenity size, household income and other covariates believed to influence the value of open = 0 + 1 + 2 + 3 + space were also included in the analysis. 198 where the dependent variable (Ln Ppt) is the natural logarithm of the sale price for each property transaction ‘p’ in period ‘t’. Spt represents the structural housing characteristics, Npt the measure of neighbourhood 198 characteristics and E_pt represents the environmental variables of interest. Similarly β0,β1,β2,β3 represent the Page 182 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY corresponding parameters to be estimated, whereas εpt transformation was applied to the distance, household captures the stochastic error term. income, and area-related variables. A total of 44 variables were chosen in the construction of the hedonic regression. See Table A6. 1 for A6.2.2 Property and locational data descriptions of these variables. The modelling procedure Table A6. 1 lists the variables used to estimate the involved entering all variables into an ordinary least property analysis model equation and Table A6. 2 squares (OLS) regression using R Project for Statistical provides a summary of descriptive statistics for our Computing (ver. 3.2.0). Presence of multicollinearity sample data. Our units of analysis for this study are was tested using the Variance Inflation Factor (VIF). The individual residential properties located across the VIF specifies the strength of linear dependencies and eThekwini Municipal Area. Property sales data from identifies how much of the variance of each coefficient January 2012 to October 2014 provided by the eThekwini is inflated due to collinearity when compared to the Municipality Real Estate Department were analysed. independent variables which are not linearly related The data set contained details on 17,476 residential (Cho et al. 2008, Poudyal et al. 2009, Yoo et al. 2014). free standing properties within the EMA, including There is no formal rule associated with VIF scores (Yoo details about location, type, size, condition, number of et al. 2014). The generally accepted approach is that bathrooms, and the presence/absence of a garage and multicollinearity is not an issue when the VIF is less than swimming pool. Approximately 8% of the observations 10 (Cho et al. 2006, Cho et al. 2008, Poudyal et al. 2009, were omitted from the analysis because of missing or Treg 2010, Yoo et al. 2014). However, other studies have implausible data, reducing the sample size to 16,149 applied a VIF of less than 5 (Anderson 2000; Troy & property transactions. Grove 2008) and this was the threshold applied during this study. A score of 1 denotes perfectly correlated Each property sale transaction in the dataset had a variables (Troy & Grove 2008). unique PIN which allowed for matching each sale with a property boundary (erf) in the GIS eThekwini cadastral Certain variables, such as the number of bathrooms, layer. Figure A6. 1 shows the location of all 16,149 were found to be insignificant as well as correlated properties included in the sample set and indicates if with other variables in the model and were therefore their sales price was above or below the study sample removed from the specification. To test for clustered mean of R998 302. error terms within the robust standard error OLS baseline model, a cluster robust estimator was The dependent variable in the model is the natural implemented to account for cluster-error at the suburb logarithm of the sales price for each residential property. level. If the errors of the baseline model are in fact The date at which each property was sold was included clustered then the cluster-adjusted model helps to as a dummy variable in the model by grouping the sale of improve standard error estimates. Incorrect standard the property annually (2012, 2013, and 2014). Structural errors can lead to incorrect t-statistics and p-values. If housing characteristics include property size in the form the errors are not clustered then the cluster-adjusted of Total Living Area (TLA, m2), and a number of dummy model yields the same standard error estimates as if variables; garage (yes, no), swimming pool (yes, no), the cluster had not been specified. The results from level of security (medium-high, none-low), type of view the two additional cluster-adjusted models that were (sea, partial sea, panoramic, surrounds, commercial, implemented were the same as the baseline OLS model industrial, informal settlement), and condition of the suggesting that clustering concerns at the suburb level property (good, average, poor). These variables were are not warranted. measured by the eThekwini Real Estate Department as part of the housing assessment process. The level Previous studies have found that a log transformation of security is based on how secure the property is in of distance variables, monetary value variables and terms of safety structures such as security gates, alarm area variables generally perform better than a linear systems and boundary walls. The view from the property functional form because the log transformation captures is based on the main setting or visual from the front of the declining effect of these variables (Taylor 2003, each property. Bin & Polasky 2004, Cho et al. 2008, Poudyal et al. 2009, Li et al. 2015). For example, the price effect of The data used for the neighbourhood variables came moving one distance unit closer to the coastline should from a number of sources. The distance to CBD attribute decrease as distance to the coastline increases; i.e. a was calculated using GIS and land use layers provided move from 50m to 200m from the coastline is expected by the eThekwini Municipality and was based on the to have a large impact on property price, while a move distance to the Durban CBD from each individual from 2000m to 2150m from the coastline is expected property. Schools are an important local public service to have only a negligible effect on price. The log and it is expected that housing prices are effected by the transformation also corrects for heteroscedasticity in location and quality of schools within an area. For this the dataset (Wooldridge 2003). Therefore, a natural log A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 183 analysis only distance to the nearest independent (or space were considered in the model; (1) golf courses private) primary and secondary schools were included. and (2) park land. With regards to local environmental Distance to the nearest school was estimated for each characteristics, five broad habitat categories were used property using GIS locational data for schools as provided based on the D’MOSS and the land cover map: (1) Forest; by the municipal GIS department. Population density and (2) Estuaries and rivers; (3) Grasslands; (4) Broad-leaved/ household income variables were estimated using Census mixed woodland and thicket; and (5) enclosed sugarcane 2011 data provided by StatsSA. Census data was available farmland. These were further grouped together into for every sub-place (i.e. suburb) across the EMA allowing (1) Natural Vegetated Areas (forest, grassland, mixed us to estimate population density and modal household woodland); (2) Natural Aquatic Areas (estuaries and income per census sub-place. rivers); and (3) Farmland. For each property, the total amount of each type of open space was determined within the 300m, 1500m and 5000m buffer radius. For A6.2.3 Land cover data rivers, the length of river located within each buffer GIS was used to calculate the environmental and was used. In addition to these natural amenities, the neighbourhood spatial variables used in the model. distance to coastline from each property was also A land use land cover (LULC) map produced by the included. Distance was measured as the straight line eThekwini GIS Department was combined with the distance in kilometres to the nearest edge of coastline Durban Metropolitan Open Space System (D’MOSS) map from the property. The “greenness” of each sub-place produced by the eThekwini Environmental Planning and was determined by estimating the percentage tree Climate Protection Department which considers the cover in each residential neighbourhood by overlaying condition of each natural green open space parcel in the census sub-place layer in Google Earth. Within each the municipal area. In order to account for potentially sub-place five stretches of residential roadway inside of important areas that are more distant than within each residential area were randomly selected and the immediate proximity to a property, the total quantity of percentage tree canopy cover along the verge of each open space surrounding each property was measured of the five selected roads was determined. The overall within three given buffer distances. Once the single percentage tree cover for each sub-place was based on land use map was created the spatial variables were an average of these five estimates. calculated using buffer analysis in GIS by calculating The condition of the natural open space was also the total quantity of different types of open space considered as it is assumed that degraded patches within three different sized buffer areas; an immediate of natural vegetation may influence property price buffer with a 300m radius surrounding each property, differently to patches in a good condition. Three a neighbourhood buffer with a radius of 1500m condition levels were used (good, intermediate, surrounding each property and a more regional buffer degraded) to determine if condition does in fact affect with a 5000m radius surrounding each property. In amenity value. The condition data were provided by doing so the effect of distance and locality of open space the eThekwini Environmental Planning and Climate on property prices could be determined. The buffers Protection Department based on their assessment of the surrounding each property were exclusive of each other ecological condition of natural open space systems in the in that the amount of open space within the smaller EMA. These condition factors were applied to all natural 300m buffer was not included in the larger 1500m buffer vegetated and aquatic environments within the EMA. and so on. In doing so the buffers remained independent The condition of the rivers and estuaries in the EMA was of one another and the impacts of collinearity were based on water quality data provided by the eThekwini significantly reduced. Department of Water and Sanitation and included over Industrial areas, commercial areas and main roads are 100 sampling stations. The water quality data from each assumed to have an influence on property prices and station was used to allocate a “fitness for use” condition these variables were included in the analysis as the (based on E.coli counts and total permanganate total amount of major road network, industrial and levels) to each stretch of river as recommended by the commercial land surrounding each property within each Department of Water and Sanitation (DWS) guidelines. of the three buffers. With our analysis come certain limitations. It is Different types of green open space may influence very difficult to control for all relevant local scale housing prices in a number of different ways and characteristics because we do not have sufficient data this was considered when developing the model. on all variables, such as data on local crime levels, local Recreational open space such as a golf course may air quality and the extent or type of domestic gardens have a very different influence on property price when and their influence on “greening” neighbourhoods. compared to natural vegetated areas. As such, the model Including all possible variables in the regression analysis developed for this study included a number of different becomes impractical even with access to all data and types of open space. Two types of urban green open therefore we rely on the set of control variables, such as Page 184 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY population density, household income and the amount of industrial and major road networks described above which would be correlated to other relevant neighbourhood characteristics that are not included in our model, such as pollution or noise levels. Table A6.1 Definitions of variables used in the model Variable name Definition Structural variables Sales Price Property transaction price (Rands) Date Year of sale: 2012, 2013, 2014 Total Living Area Total area of main living space (m2) Garage Presence/absence of a garage Pool Presence/absence of a swimming pool Security Level of security: Med-high or None-low. View from property: Sea, Partial Sea, Panoramic, Surrounds, Commercial, Industrial, Informal View Settlements Condition Condition of property: Good, Average, Poor Neighbourhood variables: Population density Number of persons per km2 in census sub-place CBD Distance to Central Business District (km) School Distance to nearest independent school (m) Income Modal household income per census sub-place (Rands) Industry Amount of industrial land within property radius (ha) Commercial/Retail Amount of commercial/retail land within property radius (ha) Road Amount of major roadway within property radius (ha) Coastline Distance to nearest coastline point (km) Tree cover Percentage neighbourhood tree cover per census sub-place (%) Open space amenities: Golf course Amount of golf course within property radius (ha) Park Amount of park land within property radius (ha) Sugarcane farmland Amount of sugarcane farmland within property radius (ha) Amount of natural vegetated open space in a good (1), intermediate (2) or degraded (3) Natural vegetation condition within property radius (ha) Length of river in a good (1), intermediate (2) or degraded (3) condition within property Rivers radius (m) A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 185 Figure A6.1 Location and values of properties in the EMA. Figure is based on the sample size of 16149 property sales from January 2012 – September 2014. Red indicates properties that exceed the sample mean value of R998 302 and the purple indicates properties with sales prices below the study sample mean. Page 186 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A6.2 Summary statistics based on estimation sample of 16149 properties in the EMA between January 2012 and October 2014. (1), (2) and (3) adjacent to natural vegetation and rivers indicate the condition where (1) is good, (2) is intermediate and (3) is degraded. Variables Mean Std Dev Min Max Property Price (Rands) 998,302 1,037,171 30,000 16,000,000 Total Living area (m2) 131.79 79.97 10.00 918.00 Distance to CBD (km) 14.68 8.06 0.64 47.27 Population Density (ppl/km2) 3867.89 3130.25 21.12 26214.57 Modal HH Income 136,497 111,748 7,200 921,400 Distance to nearest school (m) 2645.86 2098.98 18.35 15148.38 Distance to nearest coastline (km) 9.69 7.72 0.06 44.6 Tree Cover (%) 30 22 0 80 Within 300m radius: Natural vegetation (1) (ha) 0.88 2.51 0.00 27.42 Natural vegetation (2) (ha) 0.87 2.14 0.00 27.51 Natural vegetation (3) (ha) 0.94 2.14 0.00 27.33 Main rivers (1) (m) 25.30 120.81 0.00 1342.78 Main rivers (2) (m) 57.89 193.58 0.00 1630.64 Main rivers (3) (m) 1.94 32.26 0.00 1195.34 Golf (ha) 0.12 0.99 0.00 18.40 Park (ha) 0.39 1.30 0.00 17.39 Sugarcane farmland (ha) 0.12 1.00 0.00 21.02 Commercial/Retail (ha) 0.21 0.94 0.00 18.61 Industry (ha) 0.34 1.47 0.00 21.59 Roads (ha) 0.25 0.71 0.00 10.74 Within 1500m radius: Natural vegetation (1) (ha) 35.47 46.26 0.00 426.39 Natural vegetation (2) (ha) 30.26 30.96 0.00 214.44 Natural vegetation (3) (ha) 29.99 32.51 0.00 305.73 Main rivers (1) (m) 814.41 1456.22 0.00 8537.70 Main rivers (2) (m) 1835.22 2153.05 0.00 11895.11 Main rivers (3) (m) 99.24 474.33 0.00 4430.78 Golf (ha) 3.50 12.41 0.00 111.83 Park (ha) 8.29 13.09 0.00 87.81 Sugarcane farmland (ha) 11.41 41.27 0.00 414.97 Commercial/Retail (ha) 9.44 15.41 0.00 146.59 Industry (ha) 26.67 44.10 0.00 301.23 Roads (ha) 8.05 10.06 0.00 72.14 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 187 Variables Mean Std Dev Min Max Within 5000m radius: Golf (ha) 29.30 42.47 0.00 180.73 Park (ha) 55.95 55.25 0.00 248.01 Sugarcane farmland (ha) 215.30 510.74 0.00 3396.17 Commercial/Retail (ha) 101.26 99.61 0.00 540.64 Industry (ha) 395.69 332.97 0.04 1423.34 Roads (ha) 88.77 73.89 0.00 335.86 A6.2.4 Assigning values to green open space areas Results and discussion A6.3 In order to assign a value to green open space it was necessary to calculate the premium associated with A6.3.1 Model estimations natural open space and park land, aggregate the model results, and determine spatially the value associated Table A6. 3Error! Reference source not found. presents with each patch of green open space. Each property the regression estimates from the hedonic property was assigned to a census level sub-place (roughly model. A White test of heteroscedasticity rejected the equivalent to a suburb) and the effect of open space null hypothesis of homoscedasticity at the 1% level in (natural and parkland) on property values was obtained the hedonic regression. White’s heteroscedasticity- from the estimated model coefficients, which provide correlated covariance matrix was used to make inference the percentage change in property value given a unit and the results from the regression based on the change in the value of the open space variable under updated standard error terms. Computed values of VIF consideration. The aggregate effect of open space in did not exceed the threshold value of 5, with none of the EMA was then estimated by applying the regression the variables being higher than 3. Conventional adjusted results to the entire stock of residential houses with a R2 (0.81) reveals a good fit of the data into the specified property extent of less than 5000 m2 within each sub- model. Coefficients on most of the variables were place of the EMA. Sub-place units were found to be too statistically significant at the 1% or better level and most small and as a result the total premiums per sub-place of them had the expected signs. were combined into a slightly larger scale of census Certain variables were dropped from the model if they main-place level. Using GIS, the total amount of natural did contribute to the overall fit. These variables were vegetation in a good condition and the amount of public found to be insignificant and by removing them, the park land was established for each sub-place and then for model fit improved and the VIF values improved for each main-place. The value (R/ha) associated with each remaining variables. For example, distance to the CBD area of open space or parkland could then be calculated was unexpectedly found to have a positive influence on based on total premiums and the total amount of green housing prices, was insignificant and was subsequently open space present within each main-place. dropped from the model. This may be a result of the EMA having more than one central business district with areas like Westville, Pinetown and Umhlanga having important and relatively large business centres and industrial parks themselves making the centre of Durban less important as a CBD. Page 188 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Table A6.3 Model estimation results Variable Co-efficient Standard error t-value Pr(>|t|) p (Intercept) 7.21820 0.12430 58.07 < 2.2e-16 *** Ln(TLA) 0.59723 0.01038 57.54 < 2.2e-16 *** View Industrial 0.03184 0.05975 0.53 0.59416 View Informal Settlement -0.07022 0.06300 -1.11 0.26504 View Panoramic 0.01284 0.05649 0.23 0.82018 View Partial Sea 0.17231 0.05858 2.94 0.00327 ** View Sea 0.25139 0.07092 3.54 0.00039 *** View Surrounds 0.01480 0.05536 0.27 0.78927 Security - Low/None -0.09165 0.00821 -11.17 < 2.2e-16 *** Condition - Good 0.08011 0.00822 9.74 < 2.2e-16 *** Condition - Poor -0.29831 0.01365 -21.86 < 2.2e-16 *** Garage - Yes 0.09924 0.00897 11.06 < 2.2e-16 *** Pool - Yes 0.07163 0.00827 8.67 < 2.2e-16 *** Density -0.00001 0.00000 -4.28 0.00002 *** Ln(Income) 0.30389 0.00807 37.65 < 2.2e-16 *** Ln(DistanceSchool) -0.03265 0.00540 -6.05 1.498E-09 *** Ln(DistanceCoast) -0.02660 0.00415 -6.42 1.443E-10 *** Golf - 300m 0.02886 0.00489 5.90 3.695E-09 *** Park - 300m 0.01432 0.00283 5.07 4.065E-07 *** Road - 300m -0.02788 0.00500 -5.57 2.555E-08 *** Industrial - 300m -0.01618 0.00313 -5.17 2.335E-07 *** Golf - 1500m 0.00269 0.00034 7.99 1.392E-15 *** Park - 1500m 0.00205 0.00037 5.59 2.288E-08 *** Sugarcane Farmland - 1500m 0.00070 0.00010 6.83 8.714E-12 *** Commercial - 1500m 0.00293 0.00028 10.57 < 2.2e-16 *** Industrial - 1500m -0.00077 0.00010 -7.60 3.098E-14 *** Golf - 5000m 0.00209 0.00010 20.55 < 2.2e-16 *** Park - 5000m 0.00077 0.00009 8.66 < 2.2e-16 *** Commercial - 5000m 0.00030 0.00005 6.52 7.095E-11 *** Industrial - 5000m -0.00006 0.00001 -3.72 0.00020 *** NatVeg300m_1 0.00470 0.00167 2.82 0.00478 ** NatVeg300m_3 -0.00878 0.00213 -4.13 0.00004 *** NatVeg1500m_1 0.00049 0.00010 4.80 1.587E-06 *** NatVeg1500m_2 -0.00118 0.00015 -7.95 1.998E-15 *** Neighbourhood Tree Cover 0.00169 0.00029 5.79 7.113E-09 *** A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 189 Variable Co-efficient Standard error t-value Pr(>|t|) p Rivers300m_1 -0.00025 0.00004 -6.99 2.905E-12 *** Rivers1500m_1 -0.00001 0.00000 -3.53 0.00041 *** Rivers1500m_2 -0.00001 0.00000 -6.27 3.728E-10 *** Rivers1500m_3 -0.00008 0.00001 -7.49 7E-14 *** R-squared 0.8075 Notes: (1) Continuous independent variables are interpreted as 100(b) to get the percentage change in Y for a one-unit change in X. (2) Categorical independent variables (View, Security, Condition, Garage, Pool) are calculated using the difference in sub-group means between the designated group and the reference group, applying the equation 100[EXP(X)-1]. (3) ***p<0.001, **p<0.01, *p<0.05. A6.3.2 Structural and neighbourhood effects price, whereas those with views of the surrounding The estimated coefficients for the property structural landscape, the sea or partial sea had a strong positive characteristics are all highly significant with the expected influence on property price. However, only sea and signs. Structural variables including square meterage partial sea views were found to be significant. Views of of living area, the presence of a garage and swimming partial sea increased property prices by 18.8% and direct pool, the condition of the property and the security of views of sea by 26%. the property were all positively related and significant The amount of commercial (or retail), industrial and at the 0.1% level, as expected. Since the size of the major road area had significant impacts on property living area in metres squared were log-transformed, prices at the 300m, 1500m and 5000m scale. The as was the dependent variable, this coefficient can be amount of major road network and industrial area had interpreted as elasticity. It revealed that a 1% increase a negative impact on property prices with every extra in square meterage of the living area increased the hectare of industrial area within 300m decreasing sales price of property by 0.59%. Density (people per prices by 1.6% (R16 150) and for major roadways by km2) has a significant negative effect on housing prices, 2.8% (R27 900). Commercial areas within 300m were with property prices decreasing by 0.001% for every found to have no impact on property prices but had a extra person per km2. Modal household income was positive and significant impact at the 1500m scale. A positive and significant with a 0.3% increase in home one hectare increase in commercial land within 1500m value for every 1% increase in modal household income; increased house prices by 0.3% (R2900), indicating the households with higher incomes being able to afford importance of accessibility to shops, shopping centres more expensive property. Distance to the nearest and restaurants. At the 5000m scale, commercial land independent school was negative and significant with a was also found to be a contributing factor with property 1% increase in distance to school lowering house prices prices increasing by 0.03% (R300) for every extra hectare by 0.03% per km. of commercial or retail area and decreasing property The condition of the house was found to be extremely prices for every extra hectare of industrial area by important, as was security. Properties in a good and 0.006% (R55). average condition have significantly higher prices when compared to those in a poor condition. The percentage A6.3.3 Environmental effects decrease in price between properties in an average condition compared to those in a poor condition was Many of the environmental variables included in the found to be 26% (R258 000) and for those in a good model are highly significant and represent relatively condition the change in price compared to those in an large economic effects. As expected, distance to the average condition was 8.3% (R83 000). Property price coastline was negative and highly significant. The logged decreased by 9% (R87 500) for houses with poor or no distance to coastline variable indicates, with all else security compared to houses with good security. Having constant, price decreases with increasing distance from a swimming pool and a garage also has a significant the coastline, but the rate of decrease lessens with positive impact on property price. Property prices increasing distance from the coast. As a logged variable increase by 10.4% (R105 000) for houses that have a with a logged dependent, the coefficient of -0.026 can garage compared to those that don’t and 7.4% (R75 be interpreted as an elasticity, indicating that for each 000) for houses that have a swimming pool. The view 1% increase in distance from the coastline, there is a from the property was found to have a significant effect 2.6% decrease in value. This highlights the importance on property prices. Views of informal settlements and that residents place on accessibility and proximity to the commercial areas had a negative impact on property coastal environment. Page 190 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY What became clear from the analysis was that the impact on property prices at the neighbourhood scale condition of natural open space areas is very important with every extra hectare increasing house prices by and does have a significant influence on property sales 0.07% (R700). At the 5000m radius golf courses and prices within the EMA, especially at the immediate and park land still had a significantly positive effect, with neighbourhood scale. At the immediate scale (300m golf courses increasing house prices by 0.21% (R2000) radius) natural vegetated open space areas in a good for every extra hectare and parks by 0.08% (800) per condition had a significant positive impact on property hectare. The “greenness” of each neighbourhood, price. For every extra one hectare of natural vegetated represented by the percentage tree cover within open space in a good condition, house prices increase residential areas was found to be positive and significant. by 0.5% or R4700. However, for every extra hectare For every extra 1% increase in tree cover, property value of degraded vegetated open space within a 300m increases by 0.17% (R1700). radius, house prices decreased by 0.9% or R8700. At the neighbourhood scale (1500m) natural vegetated open space in a good condition increases house prices A6.3.4 The value of well-managed natural open space by 0.05% (R500) for every extra hectare. At this scale, The total premium associated with natural open space intermediate and degraded natural vegetated open in a good condition was 2% of overall property value, space has a significant negative impact, with every one which amounted to R4.4 billion. When assigned to the extra hectare decreasing house prices by 0.12% (R1200). relevant open space areas, these areas were estimated Natural open areas that are not protected formally are to contribute R108 900 per ha on average within the often targeted by people moving into the city with no EMA (Figure A6. 2). This is only part of the asset value formal place to live. These areas are often transformed of these areas, which also provide other ecosystem into informal settlement sites or are used for gathering services. The highest values were located within the natural resources such as fuelwood. Overuse and main urban core (R1.4 million per ha in Durban MP) and unregulated pressures from increasing urbanisation along the coastline where high quality natural coastal results in the degradation of these natural open space forests are still intact (e.g. Umhlanga – R3.4 million per areas, resulting in a general decrease in attractiveness of ha). The values were also higher in and around Hillcrest these areas. Degraded open space patches are also often and Kloof (about R1 million per ha), both being affluent associated with crime, making them even less desirable areas, much like the coastal suburbs of Umhlanga and in terms of their proximity to properties and their Durban MP, whereas they were lower in inland areas accessibility. It was therefore no surprise that degraded such as Cato Ridge (R37 500 per ha) and Pinetown (R423 natural open space areas had a negative impact on 000 per ha). house prices within 300m and 1500m. Unexpectedly, river systems, regardless of their condition, had a negative impact on housing prices at the immediate and neighbourhood scale. Whilst the percentage decrease on property value is relatively small ranging from 0.001-0.025% (R12 – R250), it is still surprising that even when in a good condition the proximity to rivers is considered negative. This could be that people prefer to have views over river gorges, such as in the areas of Hillcrest and Kloof, but consider rivers in close proximity to property as a negative influence as they are known to attract informal settlements, are prone to flooding and can be considered breeding grounds for unfavourable insects such as mosquitos. Other types of green open space, such as golf courses and park areas both had a significant positive effect on house prices at the 300m, 1500m and 5000m scale. At the immediate scale, for every extra one hectare of golf course within a 300m radius house prices increased by 2.9% (R29 000) and 1.4% (R14 000) for parks. Similarly golf courses and parks influenced property prices positively at the neighbourhood scale with every extra hectare of golf course increasing house prices by 0.27% (R2700) and every extra hectare of park land by 0.21% (R2000). Sugarcane farmland had a significant positive A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 191 Figure A6.2 Average value (R/ha) of (a) natural open space in a good condition and (b) parks within each main-place within the EMA. Note: the actual location and extent of open space areas within each EMA is not shown due to scale. A6.3.5 The value of public park land our understanding that urban parks provide a number of The total premium associated with public parks was opportunities for every day recreation such as jogging, approximately 6.4% of overall property value, amounting walking dogs and safe areas for children to play. Natural to a total of R13.8 billion. When assigned to the relevant open spaces may provide a beautiful view or a place to parks, these areas were estimated to contribute R14.7 walk, but it is assumed that these areas are not utilised million per ha on average (Figure A6. 2). Park value is in the way that parks can be and as a result are valued highest in and around the urban core of Durban city but less by residents who are seeking a safer and easily is also important in other densely populated areas such accessible environment in which to enjoy green open as Chatsworth, Phoenix and Pinetown. With an average space in the city. population density in Chatsworth of 5500 people per km2 and in Phoenix 6400 people per km2, both higher A6.3.6 Contribution to property rates income than the EMA average, it is clear why residents in these areas may value public open space more than residents Based on the above, the presence of well managed living in less dense areas with larger property sizes. The green open space in eThekwini Municiaplity contributes suburbs surrounding the city centre such as Morningside an estimated R356 million per annum in property tax and Musgrave have a number of public park areas that revenues to the city. This would account for more than contribute significantly to the premiums in this area. It is 5% of total property rates income to the municipality. Page 192 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY A6.4 Conclusions The eThekwini property market analysis yielded three The results from this research have several management important insights. The first was that the effect of open implications. The premium values associated with green space on sales price depends on property location and open space provide more of an understanding as to how neighbourhood characteristics, the second was that the residents in different areas value green open space and condition of the natural open space areas is extremely these results suggest that city planners and developers important, and the third being that the type of open need to consider the spatial context of open space space has a significant influence on property price. The areas and how best to provide and protect these areas analysis showed that natural open space environments across the EMA. In Durban it is clear that the type and in a good condition have a significant positive impact condition of green open space has a large influence on on property prices and natural areas in a degraded the value associated with these areas, as does household condition have a strong negative impact on property income and population density. This information can be prices. This was evident at both the 300m and 1500m used to better understand, plan and budget for future buffer radius highlighting the importance of good quality development across the EMA and can also be used to natural open space and the positive impact it has on highlight the importance of these areas and the need to house premiums in the EMA. Residents therefore attach maintain them. Degraded open areas, however, have a amenity value and are willing to pay for access to or significant negative impact on house prices. This result views of open space, but only when these areas are in a needs to be examined further in terms of how possible healthy condition. restoration or rehabilitation of degraded areas can increase property value in these areas, or the possibility The combined value of green open space areas in the of developing certain degraded areas into other green EMA, both natural and man-made, was estimated open space areas such as public parkland. There are to be R18.2 billion. This study has shown that at the however costs involved in maintaining public park areas immediate and neighbourhood property scale residents and therefore a detailed cost benefit analysis is needed of Durban attach a higher amenity value to man-made to fully understand how best to provide good quality parks than they do natural green open space. It is our green open space for the residents of Durban who understanding that safety plays an important part clearly do value these spaces. in how people value certain types of open space. It is assumed to be because of the many recreational opportunities that parks offer, such as jogging or playing football, activities that cannot be done easily in natural, densely vegetated areas and that residents seek a safer and more easily accessible environment in which to enjoy green open space in the city. A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY Page 193 This page intentionally blank. Page 194 A SPATIAL VALUATION OF THE NATURAL AND SEMI-NATURAL OPEN SPACE AREAS IN ETHEKWINI MUNICIPALITY