62642 CLIMATE RISK AND BUSINESS PORTS Terminal Marítimo Muelles el Bosque Cartagena, Colombia Appendices Acknowledgements © 2011, International Finance Corporation Authored by Vladimir Stenek, International Finance Corporation Jean-Christophe Amado, Richenda Connell and Olivia Palin, Acclimatise Stewart Wright, Ben Pope, John Hunter, James McGregor, Will Morgan and Ben Stanley, WorleyParsons Richard Washington and Diana Liverman, University of Oxford Hope Sherwin and Paul Kapelus, Synergy Carlos Andrade, EXOCOL José Daniel Pabón, Universidad Nacional de Colombia The authors wish to thank the owners, management and staff of Terminal Marítimo Muelles el Bosque (MEB) for their support and cooperation in this study, especially Gabriel Echavarría, Alberto Jimenez, Carlos Castaño Muñoz, Rafael Zorrilla Salazar, Alan Duque, Humberto Angulo, Manuel Otálora Gomez, Andres Burgos, Elizabeth Pedroza Arias and Juan Casilla Vergara. The authors also wish to thank the following institutions for their valuable contributions to the study: Alcaldía de Cartagena de Indias - Secretaría de Infraestructura; Centro de Investigación de la Caña de Azúcar de Colombia (Cenicaña); Centro de Investigaciones Oceanográficas e Hidrográficas de la Dirección General Marítima (CIOH); Centro Internacional de Agricultura Tropical (CIAT); Centro Nacional de Investigaciones de Café (Cenicafé); Corporación Autónoma Regional del Canal del Dique (CARDIQUE); Corporación Colombiana de Investigación Agropecuaria (CORPOICA); Departamento Nacional de Planeación (DNP); Dirección General Marítima (DIMAR); Federación Nacional de Cafeteros; Fundación Natura; Instituto Colombiano Agropecuario (ICA); Instituto de Hidrología, Meteorología y Estudios Ambientales de Colombia (IDEAM); Instituto de Investigaciones Marinas y Costeras (INVEMAR); Ministerio de Agricultura y Desarrollo Rural; Ministerio de Ambiente, Vivienda y Desarrollo Territorial (MAVDT); Puerto de Mamonal; Sociedad Portuaria Regional de Cartagena (SPRC); Universidad de Cartagena; Universidad de los Andes - Centro Interdisciplinario de Estudios sobre Desarrollo (CIDER); and Universidad Nacional de Colombia. Reviewers We thank Lisa Wunder (Port of Los Angeles), Ahmed Shaukat (IFC) and an anonymous reviewer for their critical comments and suggestions. This work benefited from support provided by the Trust Fund for Environmentally & Socially Sustainable Development (TFESSD), made available by the governments of Finland and Norway. CLIMATE RISK AND BUSINESS PORTS Terminal Marítimo Muelles el Bosque Cartagena, Colombia Appendices List of Acronyms AOGCM Atmosphere-Ocean General Circulation Model CHOCO Chorro del Occidente Colombiano CIOH Centro de Investigaciones Oceanográficas e Hidrográficas CORDEX Coordinated Regional climate Downscaling Experiment CRU Climate Research Unit CSIRO Commonwealth Scientific and Industrial Research Organisation DJF December, January, February DTR Diurnal Temperature Range ENSO El Niño Southern Oscillation GCM General Circulation Model or Global Climate Model IDEAM Instituto de Hidrología, Meteorología y Estudios Ambientales de Colombia IPCC Intergovernmental Panel on Climate Change ITCZ Intertropical Convergence Zone JJA June, July, August MAM March, April, May MCS Mesoscale Convective System MJO Madden-Julian Oscillation MRI Meteorological Research Institute, Japan NCEP National Center for Environmental Prediction, USA NOAA National Oceanic and Atmospheric Administration, USA PRECIS A Regional Climate Model developed by the UK Hadley Centre RCM Regional Climate Model SCAR Scientific Committee on Antarctic Research SDII Simple Daily Intensity Index SLR Sea Level Rise SON September, October, November SRES Special Report on Emissions Scenarios v Contents Appendix 1. Supplementary Information to Section 4 on ‘Observed and projected Future Climate Conditions’ ................................................................................................................. 1 1. Introduction ...................................................................................................................................... 2 2. Characteristics of the Colombian climate ......................................................................................... 2 3. Observed climatic conditions ........................................................................................................... 4 3.1 Gridded climate data ............................................................................................................... 4 3.2 Meteorological station data .................................................................................................. 15 3.2.1 Monthly data for Cartagena .......................................................................................... 16 3.2.2 Monthly data for Buenaventura.................................................................................... 20 3.2.3 Monthly data for Barranquilla ....................................................................................... 23 3.2.4 Daily met data for Cartagena ........................................................................................ 24 3.3 Tropical cyclones ................................................................................................................... 32 4. Future climatic conditions .............................................................................................................. 33 4.1 Multi-model projections from the IPCC AR4 over Colombia and the surrounding area ...... 33 4.1.1 Data, methods and uncertainties .................................................................................. 33 4.1.2 Multi-model GCM projection ........................................................................................ 36 4.2 Projections from empirical downscaling ............................................................................... 67 4.2.1 Temperature ................................................................................................................. 68 4.2.2 Precipitation .................................................................................................................. 69 4.3 Projections from a regional climate model: PRECIS .............................................................. 73 4.3.1 Temperature ................................................................................................................. 76 4.3.2 Precipitation .................................................................................................................. 76 4.3.3 Wind .............................................................................................................................. 76 4.4 Estimates of changes in future extreme rainfall ................................................................... 76 4.5 Tropical cyclone projections ................................................................................................. 77 5. Further information on climate patterns and variability over Colombia ....................................... 79 5.1 The Intertropical Convergence Zone ..................................................................................... 79 5.2 The Chorro del Occidente Colombian jet .............................................................................. 80 5.3 El Niño Southern Oscillation and interannual climate variability ......................................... 80 Appendix 2. PRECIS Projections for Climate Change Over Cartagena ...................................... 86 1. Introduction ..................................................................................................................................... 87 2. Physiographical features of the region............................................................................................ 87 3. Data and methods ........................................................................................................................... 89 4. Recent trends in air temperature, relative humidity, and precipitation ......................................... 91 5. The future climate (2011-2040 and 2070-2100 periods) ................................................................ 91 vi Appendix 3. Supplementary Information to Section 5 on ‘Vehicle Movements Inside the Port’: Sea Level Variability............................................................................................................... 96 1. Introduction ..................................................................................................................................... 97 2. Definition of a datum for the flood risk assessment ....................................................................... 97 3. Harmonic analysis on astronomical tides in Cartagena................................................................... 97 4. Analysis of annual variations in sea level due to meteorological factors ....................................... 98 Appendix 4. Supplementary Information to Section 6 on ‘Demand, Trade Levels and Patterns’ .............................................................................................................................. 103 1. Introduction ................................................................................................................................... 104 2. Further information on the methodology, assumptions and results of the ‘Stern Review’ ......... 104 3. Supplementary analysis on the impact of climate change on grain imports at MEB.................... 108 3.1 General, non-climatic factors affecting global grain imports.............................................. 108 3.2 Impact of climate change on agricultural yields and demand for grain ............................. 109 3.3 Impact of climate change on Colombian grain imports ...................................................... 112 Appendix 5. Supplementary Information to Section 6 on ‘Demand, Trade Levels and Patterns’, on Climate Change Impacts on Selected Colombian Agricultural Exports.......................... 116 1. Introduction ................................................................................................................................... 117 2. Coffee ............................................................................................................................................ 117 2.1 Sensitivity of coffee crop to climate and climate change ................................................... 117 2.2 Sensitivity of coffee’s pests and diseases to climate and climate change .......................... 121 2.3 Possible adaptation options ................................................................................................ 122 3. Bananas......................................................................................................................................... 123 3.1 Sensitivity of banana crop to climate and climate change ................................................. 123 3.2 Sensitivity of banana’s pests and diseases to climate......................................................... 125 3.3 Possible adaptation options ................................................................................................ 125 4. Sugar Cane .................................................................................................................................... 126 4.1 Sensitivity of sugar cane crop to climate and climate change ............................................ 126 4.2 Sensitivity of sugar cane’s pests and diseases to climate and climate change ................... 127 4.3 Possible adaptation options ................................................................................................ 127 5. Plantain ......................................................................................................................................... 128 5.1 Sensitivity of plantain crop to climate and climate change ................................................ 128 5.2 Possible adaptation options ................................................................................................ 128 Appendix 6. Supplementary Information to Section 7 on ‘Goods Storage’ ............................ 130 1. Introduction .................................................................................................................................. 131 2. Assumptions and detail of the technical analysis on surface flooding risk .................................. 131 3. Assumptions and detail of the technical analysis on refrigeration .............................................. 135 vii Appendix 7. Supplementary Information to Section 12 on ‘Social Performance’ .................. 137 1. Introduction .................................................................................................................................. 138 2. Overview of the social context in Colombia ................................................................................. 138 3. Overview of the social context in Cartagena ................................................................................ 138 4. Climatic vulnerabilities of Colombian communities ..................................................................... 142 viii Climate risk case study: Terminal Marítimo Muelles El Bosque Appendix 1: Supplementary information to Section 4 ‘Observed and Projected Future Climate Conditions’ 1. Introduction This appendix focuses on observed climate variability and future climate change in Colombia. The baseline characteristics of the Colombian climate, especially over MEB and competing ports (such as Buenaventura, Santa Marta and Barranquilla) are presented first based on two gridded climate data sets and data from meteorological stations, followed by an assessment of recent climate trends. Projections of future climate are provided in Section 4. As part of this review of present-day and future climatology, original analysis of observed and simulated climatic data was performed. 2. Characteristics of the Colombian climate Colombia covers an area of about 1,140,000km2 extending from the Caribbean in the north (12°N) to the Amazon in the South (4°S) and from the Pacific Ocean at 79°W to the Orinoco-Negro River at 66°W (Marin and Ramirez, 2006). The Andean Cordillera is the name given to the mountain range running through Colombia and dominating the temperate highland landscape. The range is divided into three and runs through the western half of the country, extending beyond the borders to both the north and south. The central mountain range is the highest (higher than 3500m above sea level with an average width of 90km) with many glacial peaks greater than 5000m elevation. The western Cordillera (Cordillera Occidental) is the lowest of the three ranges (2000m above sea level and about 40km wide). The eastern range is approximately 2500m high and 150km wide (Parsons, 1982). By cutting through Colombia from south to north, the Andes act as both barriers and channels for atmospheric flow, leading to their own local atmospheric circulations (with significant influence on climate in regions, such as Colombia, where wind speeds are generally low). Climate patterns in Colombia can be explained by several driving factors such as the Intertropical Convergence Zone (see Section 0), the Andes, the distance to the Caribbean Sea or the Pacific Ocean, and the vegetation and land surface feedbacks of the Amazon Basin (Poveda et al, 2001). The El Niño Southern Oscillation (ENSO) is the strongest driver of natural interannual climate variability in Colombia (Mantilla et al, 2009). ENSO is a cyclic phenomenon occurring every two to seven years. The impact of ENSO in Colombia is felt earlier and stronger in western and central Colombia than in the east. The strongest impacts occur during the months December to February, of the year following the onset of an El Niño episode. In Colombia, El Niño events are associated with reduced rainfall (by an average of 22%), whereas La Niña events are associated with above average precipitation (increased by 6 to 40%). Overall, average precipitation is 48% higher in La Niña years compared to El Niño years. Furthermore, El Niño events raise sea level along the Pacific coast of Colombia by between 20 and 35cm (Restrepo et al, 2002). Marin and Ramirez (2006) postulate that the Colombian climate can be subdivided into five distinct subtypes: x The Caribbean with average precipitation of 1500mm per year. Some regions (such as the coastal Sierra Nevada of Santa Marta, as high as 5800m above sea level) have an associated rainfall of 4000mm per year. Minimum rainfall in this region is found in the northeast Guajira Peninsula (300mm per year). 2 x The Andes Mountains, with an average of 1500mm per year. The northwest mountains are the wettest with precipitation of 6000mm per year. There is significant variation within the region, with windward slopes receiving twice as much precipitation as leeward slopes. x Llanos, in the eastern part of Colombia, with average rainfall of 2800mm per year. This region is formed by lowlands and contains a river network of high drainage density extending into the Orinoco River. However, the high average precipitation is deceptive as intense rainfall is concentrated in the southwest. x The Amazon which refers to the zone in the southern part of Colombia, adjacent to the Brazilian border. Annual average precipitation here is high, measuring 3500mm per year, with an average yearly maximum 500mm greater and a minimum 2500mm lower, next to the Amazon River. x The Pacific coast, located between the shorelines of the Pacific Ocean and the western range of the Andes. This is the wettest region of Colombia with rainfall of 5500mm per year. Indeed, the wettest area of the American continent is located within this zone and receives precipitation measuring up to 13300mm per year. In equatorial latitudes, such as in Colombia, the main difference between seasons is in the amount of rainfall, given that there is little difference in temperature between the warmest and coolest months. Temperatures in Colombia are consistently high throughout the year and are consistent in magnitude with much of the rest of the tropics. The average temperature difference between the warmest and the coolest months of the year is only 5°C. By comparison, the increase in altitude corresponds with a decrease in temperature of approximately 0.06°C for each 100m (Poveda et al, 2005). Except for the Western range of the Andes, the coolest season in the country is JJA (with average maximum temperatures of 30°C and average minimum temperatures between 10 and 24°C). In contrast, the warmest season is MAM and sees average maximum temperatures of between 34 and 38°C in the North and North East of the Andes (temperatures off the Pacific coast remain more consistent with the rest of the year, with an average maximum temperature of around 32°C).The Colombian average difference between the lowest temperature at night and the highest temperature during the day (diurnal temperature range) is less than 10°C. Broadly, the Colombian climate is characterized by one wet season (from May to November) and one dry season (from December to April). The Colombian dry season is characterized by strong winds and weak rains, alongside the possibility of storm surges. The Colombian wet season has weaker winds and is accompanied by significant rainfall, especially with tropical cyclones. Areas of northern Colombia, close to the Caribbean, have one single dry season centered on the early months of the calendar year (Parsons, 1982). Colombia as a whole has a mean annual precipitation of 3000 mm, almost three times the global average of 900 mm per year. The country is the fourth country in the world with the most available surface water. Russia, Canada and Brazil have more water runoff, though Colombia is on average less than an eighth of the area of the four preceding countries (Gutierrez and Dracup, 2001). The Pacific coast of Colombia, where Buenaventura port is located, has extreme precipitation rates, which makes it one of the wettest places on earth (Poveda et al, 2005). Annual average precipitation reaches 12000 mm a year in some places (Poveda, Waylen and Pulwarty, 2006). 3 Additional background material on the controls over the Colombian climate and the influence of ENSO is provided in Section 5. 3. Observed climatic conditions This section provides details on observed climate variability and trends in Colombia from three different sources: x Gridded climate data sets (namely the Climate Research Unit (CRU) 0.5° by 0.5° seasonal dataset and the NCEP ’reanalysis dataset’ from the National Oceanic and Atmospheric Administration), x Data from four meteorological stations in Colombia, and x Published literature on extreme weather events. 3.1 Gridded climate data In order to gain some idea of the baseline climate over Colombia, and particularly MEB, a series of figures have been created to illustrate the range of current climatic conditions. Figure 3-1 shows the CRU 0.5° by 0.5° average temperature dataset for each season (DJF, MAM, JJA and SON) between 1961 and 1990 over Colombia. Temperatures in Colombia are consistently high throughout the year and are consistent in magnitude with much of the rest of the tropics. There are nevertheless some seasonal differences in temperature. Overall, the coolest season is June to August, where maximum temperatures measure 30°C to the east of the Andes (Figure 3-1 to Figure 3-3). In JJA, the northern Caribbean coast of Colombia sees maximum temperatures of between 32°C and 34°C, with maximum temperatures of up to 32°C along the Pacific Ocean coast to the west of the Western Cordillera. In contrast, the warmest season is between March and May, during which time there is a different spatial pattern of high temperatures compared to the following cooler months. Whilst the Pacific coast remains consistent with a maximum temperature of 32°C, to the northeast of the Andes there is a region of maximum temperature of between 34°C and 38°C. This very warm region extends downwards from Venezuela. The Caribbean coast, where Cartagena is located, sees maximum temperatures of between 34°C and 36°C at this time. The remaining six months of the year (September to February) show very similar spatial patterns and magnitude of maximum temperature. This pattern is similar to that of JJA, with just a decreased magnitude over northeast Colombia where maximum temperatures now reach up to 36°C. However, in addition there is a further region of maximum temperature in southern Colombia, extending upwards from the maximum over Brazil and northern Chile. Maximum temperatures over the Western range of the Andes are largely consistent between seasons, measuring at between 22°C and 28°C, with cooler temperatures being found in the centre of each range and corresponding with higher topography. In contrast over lower lying land, whilst temperatures are consistently high, there is more variation between seasons. 4 Figure 3-1 – Average seasonal temperature over Colombia for 1961-1990. Source: Climate Research Unit 0.5° by 0.5° data set DJF SON MAM JJA . 5 Figure 3-2 – Mean maximum temperature for 1961-1990 for each season over Colombia. Source: Climate Research Unit 0.5° by 0.5° data set. DJF SON MAM JJA 6 Figure 3-3 – Mean minimum temperature 1961-1990 for each season (anticlockwise from top left, DJF, MAM, JJA, SON). Source: Climate Research Unit 0.5° by 0.5° data set. DJF SON MAM JJA The pattern of average minimum temperature for each month (Figure 3-3) follows very similar spatial distribution to the pattern of average maximum temperatures. As expected, the coolest minimum temperatures are found over the higher terrain of the Andes (minimum mean temperatures of between 6 and 8°C for each season at the highest point). Conversely, the highest minimum temperatures are found over northern Colombia, along the Caribbean coast and the northern border with Venezuela. The mean minimum temperature here measures between 20°C and 24°C. Mean minimum temperatures of between 18°C and 22°C are found over much of the rest of the country. In terms of seasonal differentiation, the coolest months are again found in the austral winter (JJA), where the extent of minimum temperatures of between 18°C and 20°C is much larger than in the other seasons, where minimum temperatures between 20 °C and 22°C warmer dominate. 7 As with many regions in the tropics, the seasonal cycle of precipitation is pronounced (Figure 3-4). During December to February, maximum daily precipitation values of up to 24mm per day are found in a small region on the Pacific coast of Colombia, with much of the rest of the country seeing relatively low precipitation rates corresponding to this being the driest of Colombia’s seasons. The driest area of Colombia during these months is found to the north, including Cartagena on the Caribbean coast, where daily precipitation values of between 0 and 2mm per day are found. In the next three months (MAM), daily mean precipitation values are generally increased over much of the country with the exception of the Caribbean coast where they remain low (up to 2mm per day with up to 4mm per day slightly further south). The maximum on the Pacific coast seen in the previous season is still present, although much decreased. Central Colombia has approximately 4 to 6mm precipitation per day during this time with slightly more present along the southern and eastern borders. June to August sees a precipitation peak over central southern Colombia. Here, values of up to 26mm per day are found, with a similar peak re-emerging on the Pacific coast. The Caribbean coast, whilst still being the driest area of Colombia, also sees considerable precipitation during this period, measuring between 4 and 12mm per day for the 1961-1990 baseline climatology period. Finally, the boreal spring (September to November) sees a spatial distribution of precipitation very similar to that found six months earlier. The region of maximum is on the Pacific coast (being present in each season, this highlights the region as one of the wettest globally). The only noticeable difference in the patterns of precipitation from MAM during this time is the slightly higher rates of between 2 and 8mm per day present on the Caribbean coast in comparison to the lower rates of up to 2mm per day earlier in the year. Figure 3-5 highlights that the diurnal temperature range (DTR) in Colombia is relatively consistent throughout all seasons, measuring between 10°C and 13°C over much of the country for the majority of the time. Exceptions to this pattern can be found. For example, a higher DTR of up to 18°C is found over a small region in the northeast of the country on the Colombian border with Venezuela during both December to February and March to May. In contrast, a narrower DTR of between 8°C and 10°C is seen from June to August over much of central and eastern Colombia. A more general pattern (present all year round) can also be identified. This is the lower DTR (between 9° and 10°C) that is present around both the Pacific and Caribbean coastlines, and can be explained due to the moderating effects of a large body of water on temperature. Figure 3-6 and Figure 3-7 show the trend in mean annual temperature and precipitation between 1948 and 2009 based on the NCEP reanalysis data. Concurring with trends reported by the Intergovernmental Panel on Climate Change (IPCC), temperature over a northern South American domain centered on Colombia shows an upwards trend over the second half of the twenty-first century. This is in conjunction with considerable interannual and seasonal variations in temperature, due to factors such as ENSO and sea surface temperatures, among others (for more information, see Section 0). With respect to precipitation rate, it is much harder to discern a pattern from the NCEP reanalysis data. There does appear to be some increase in precipitation but the increase is very small when set against natural variability. 8 Figure 3-4 – Mean Precipitation (mm/day) 1961-1990 for each season (anticlockwise from top left, DJF, MAM, JJA, SON). Source: Climate Research Unit 0.5° by 0.5° data set. DJF SON MAM JJA 9 Figure 3-5 – Diurnal Temperature Range for 1961-1990 each season (anticlockwise from top left DJF, MAM, JJA, SON). Source: Climate Research Unit 0.5° by 0.5° data set. DJF SON MAM JJA 10 Figure 3-6 – Annual mean air temperature over a Colombian domain (1948-2009) Figure 3-7 – Annual precipitation rate over a Colombian domain (1948-2009) 11 In terms of climate extremes, namely the warmest, coolest and wettest months, Table 3.1 to Table 3.4 show several of these variables over Colombia: x Table 3.1 shows the largest values of maximum, mean and minimum temperatures of the warmest month of the year, for each calendar month between 1961 and 1990. From this, it can be seen that the mean maximum temperature is very consistent year round, with values ranging from 30.998°C in June to 32.333°C in September, and showing that the highest maximum temperatures were reached in September and October between 1961 and 1990. x A similar pattern exists for the largest values of maximum, mean and minimum temperatures of the coolest month of the year, as shown in Table 3.2. The range of mean minimum temperatures is very narrow. The highest mean minimum temperature over the Colombian domain reaches 19.315°C in April. In contrast, the coolest mean minimum temperature is 17.875°C in July. x In terms of average temperatures (Table 3.3), the two coolest months are June and July. This corresponds with these two months being where the lowest mean maximum and minimum temperatures are found. x Table 3.4 shows mean daily precipitation over the same domain and time period (1961-1990). Because of the wide range of precipitation values and changing spatial pattern throughout the seasons, the mean values here are of less value in depicting any precipitation patterns. A single region of large precipitation can have a skewing effect each month creating average figures that are non-representative. Furthermore, the variation in rainfall amounts over the country means that a single figure cannot capture a reasonable representation of precipitation over Colombia in any month, so this table is included just for reference. Table 3.1 – Maximum temperature of each month over Colombia for 1961-1990 Date Time Code Level Size Miss : Minimum Mean Maximum 1 : 1990-01-15 00:00 -1 0 3720 1585 : 16.901 31.850 38.901 2 : 1990-02-15 00:00 -1 0 3720 1585 : 16.699 32.181 38.001 3 : 1990-03-16 00:00 -1 0 3720 1585 : 18.201 32.098 37.800 4 : 1990-04-16 00:00 -1 0 3720 1585 : 16.501 31.563 37.601 5 : 1990-05-16 00:00 -1 0 3720 1585 : 18.000 31.448 36.301 6 : 1990-06-16 00:00 -1 0 3720 1585 : 17.700 30.998 35.499 7 : 1990-07-16 00:00 -1 0 3720 1585 : 17.301 31.447 38.599 8 : 1990-08-16 00:00 -1 0 3720 1585 : 17.499 31.882 37.201 9 : 1990-09-16 00:00 -1 0 3720 1585 : 17.499 32.333 37.800 10 : 1990-10-16 00:00 -1 0 3720 1585 : 16.901 32.208 36.201 11 : 1990-11-16 00:00 -1 0 3720 1585 : 16.999 31.770 35.001 12 : 1990-12-16 00:00 -1 0 3720 1585 : 17.401 31.614 35.999 12 Table 3.2 – Minimum temperature of the coolest month over Colombia for 1961-1990 Date Time Code Level Size Miss : Minimum Mean Maximum 1 : 1990-01-15 00:00 -1 0 3720 1585 : -1.0986 18.312 22.699 2 : 1990-02-15 00:00 -1 0 3720 1585 : -0.90028 18.783 22.501 3 : 1990-03-16 00:00 -1 0 3720 1585 : -3.7995 19.029 23.001 4 : 1990-04-16 00:00 -1 0 3720 1585 : -2.3987 19.315 23.600 5 : 1990-05-16 00:00 -1 0 3720 1585 : -4.0009 19.252 24.302 6 : 1990-06-16 00:00 -1 0 3720 1585 : -5.8015 18.676 24.500 7 : 1990-07-16 00:00 -1 0 3720 1585 : -5.3010 17.875 24.601 8 : 1990-08-16 00:00 -1 0 3720 1585 : -4.3000 18.452 24.302 9 : 1990-09-16 00:00 -1 0 3720 1585 : -4.0009 18.660 24.302 10 : 1990-10-16 00:00 -1 0 3720 1585 : -1.8006 18.876 23.999 11 : 1990-11-16 00:00 -1 0 3720 1585 : -2.3987 19.142 23.401 12 : 1990-12-16 00:00 -1 0 3720 1585 : -2.3011 18.819 22.501 Table 3.3 – Mean monthly temperature over Colombia for 1961-1990 Date Time Code Level Size Miss : Minimum Mean Maximum 1 : 1990-01-15 00:00 -1 0 3782 1637 : 6.8940 24.934 28.586 2 : 1990-02-15 00:00 -1 0 3782 1637 : 6.8635 25.175 29.343 3 : 1990-03-16 00:00 -1 0 3782 1637 : 6.6438 25.398 29.895 4 : 1990-04-16 00:00 -1 0 3782 1637 : 6.8696 25.425 29.657 5 : 1990-05-16 00:00 -1 0 3782 1637 : 6.9459 25.161 29.136 6 : 1990-06-16 00:00 -1 0 3782 1637 : 6.3142 24.698 29.526 7 : 1990-07-16 00:00 -1 0 3782 1637 : 6.1555 24.589 29.584 8 : 1990-08-16 00:00 -1 0 3782 1637 : 6.4698 24.987 29.807 9 : 1990-09-16 00:00 -1 0 3782 1637 : 6.6194 25.288 29.776 10 : 1990-10-16 00:00 -1 0 3782 1637 : 7.3426 25.398 28.901 11 : 1990-11-16 00:00 -1 0 3782 1637 : 7.2023 25.281 28.507 12 : 1990-12-16 00:00 -1 0 3782 1637 : 7.0314 25.016 27.875 13 Table 3.4 – Mean daily precipitation over Colombia for 1961-1990 Date Time Code Level Size Miss : Minimum Mean Maximum 1 : 1961-12-16 00:00 -1 0 3720 1585 : -0.00068665 5.9408 21.459 2 : 1962-12-16 00:00 -1 0 3720 1585 : -0.00068665 6.1334 18.759 3 : 1963-12-16 00:00 -1 0 3720 1585 : -0.00068665 5.9973 16.057 4 : 1964-12-15 00:00 -1 0 3720 1585 : 0.00011444 6.0432 20.972 5 : 1965-12-16 00:00 -1 0 3720 1585 : 0.0017166 5.8730 19.452 6 : 1966-12-16 00:00 -1 0 3720 1585 : -0.00068665 6.0690 19.557 7 : 1967-12-16 00:00 -1 0 3720 1585 : 0.0033188 5.8638 24.773 8 : 1968-12-15 00:00 -1 0 3720 1585 : 0.00038147 5.9925 22.277 9 : 1969-12-16 00:00 -1 0 3720 1585 : -0.00015259 5.9833 19.263 10 : 1970-12-16 00:00 -1 0 3720 1585 : 0.00038147 6.5629 22.318 11 : 1971-12-16 00:00 -1 0 3720 1585 : 0.00038147 6.3476 20.180 12 : 1972-12-15 00:00 -1 0 3720 1585 : 0.00064850 6.2817 18.206 13 : 1973-12-16 00:00 -1 0 3720 1585 : 0.0025177 6.5663 23.318 14 : 1974-12-16 00:00 -1 0 3720 1585 : -0.00015259 6.2037 28.920 15 : 1975-12-16 00:00 -1 0 3720 1585 : 0.00091553 6.6548 29.240 16 : 1976-12-15 00:00 -1 0 3720 1585 : 0.00038147 6.2338 19.136 17 : 1977-12-16 00:00 -1 0 3720 1585 : -0.00015259 6.0240 18.399 18 : 1978-12-16 00:00 -1 0 3720 1585 : -0.00015259 5.9675 19.722 19 : 1979-12-16 00:00 -1 0 3720 1585 : -0.00068665 5.9898 21.008 20 : 1980-12-15 00:00 -1 0 3720 1585 : 0.00038147 5.6367 19.512 21 : 1981-12-16 00:00 -1 0 3720 1585 : 0.0014496 6.5923 22.036 22 : 1982-12-16 00:00 -1 0 3720 1585 : 0.0011826 6.1087 19.367 23 : 1983-12-16 00:00 -1 0 3720 1585 : 0.0030518 5.7606 19.974 24 : 1984-12-15 00:00 -1 0 3720 1585 : -0.00015259 6.3001 25.378 25 : 1985-12-16 00:00 -1 0 3720 1585 : 0.00038147 5.4170 17.783 26 : 1986-12-16 00:00 -1 0 3720 1585 : -0.00068665 5.8553 20.400 27 : 1987-12-16 00:00 -1 0 3720 1585 : 0.00064850 5.9744 21.866 28 : 1988-12-15 00:00 -1 0 3720 1585 : -0.00015259 6.1863 22.423 29 : 1989-12-16 00:00 -1 0 3720 1585 : -0.00068665 6.0904 19.624 30 : 1990-12-16 00:00 -1 0 3720 1585 : -0.00068665 5.6471 20.306 14 3.2 Meteorological station data Monthly station data has been obtained for four Colombian stations. Two of them (the airport station ‘Apt Nune Rafael’ and the station ‘Esc Naval’ at CIOH) are located in Cartagena. The station ‘Flores Las’ is located on the same Caribbean coast as Cartagena, further north in the municipality of Barranquilla. Finally, the fourth station ‘Apto Buenaventura’ is located in the far south of Colombia on the Pacific coast, in the port of Buenaventura. For each station, the trends in key variables (average precipitation, average maximum and minimum temperatures) have been calculated using least-squares regression. The station details are included in Table 3.5. Analysis of the station datasets reveals a common trend over the second half of the twentieth century: the increase in annual precipitation which was found to be statistically significant at the 95% confidence level in three out of the four stations. For temperature variables, there was much less consistency, making it hard to draw any clear conclusions. For none of the stations was there a significant trend in the mean annual temperatures. There were two stations which showed a statistical increase in minimum temperatures and two stations showing a decrease in maximum temperatures, with one station having both statistically significant trends present. It is difficult from these data to draw any conclusions regarding trends in temperature in Colombia over the second half of the twentieth century. Table 3.5 – Details (name, location and coordinates) and available variables for the four met stations used in this study. Source: IDEAM, Colombia. Station Name Municipality Latitude Longitude Variables (including length of record) Apt Nune Rafael Cartagena 10.26N 75.30W Monthly precipitation (1941- 2009) Mean monthly temperature (1941-2009) Mean maximum monthly temperature (1943-2009) Mean minimum monthly temperature (1943-2009) Esc Naval CIOH Cartagena 10.23N 75.32W Monthly precipitation (1947- 2009) Mean monthly temperature (1953-2009) Mean maximum monthly temperature (1953-2004) Mean minimum monthly temperature (1953-2004) Flores Las Barranquilla 11.02N 74.49W Monthly precipitation (1980- 2008) Mean monthly temperature (1980-2008) Mean maximum monthly temperature (1980-2008) 15 Station Name Municipality Latitude Longitude Variables (including length of record) Mean minimum monthly temperature (1980-2008) Apto Buenaventura 03.49N 76.59W Monthly precipitation (1961- Buenaventura 2008) Mean monthly temperature (1962-2009) Mean maximum monthly temperature (1962-2009) Mean minimum monthly temperature (1962-2009) 3.2.1 Monthly data for Cartagena Average monthly temperature and precipitation recorded at the two meteorological stations in Cartagena (Apt Nune Rafael and Esc Naval CIOH) show similar seasonal patterns and relatively similar values, thus increasing confidence in the characteristics of the observed climate over MEB. The range of average monthly temperatures over MEB throughout the year is relatively narrow (between 26.7 and 28.5°C). The coldest months on average are January and February, while the warmest are between May and September. Variability in average monthly temperatures exists from year-to-year, and they range from a minimum of 25.5°C (recorded in January) to close to 30°C (recorded in June) (Figure 3-8). As shown in Figure 3-9, average monthly precipitation is characterized by: x A dry season (from December to April) with average monthly precipitation below 40mm and often close to 0mm. The maximum average precipitation in a month recorded during the time period was up to 200mm. x A transitional season (from May to July) with average precipitation below 100mm per month, but characterized by high inter-annual variability. In some years average monthly precipitation goes above 200mm (with a maximum close to 400mm in the month of July) while in others it is close to 0mm. x Wet months with relatively higher average rainfall (from August to November). Average rainfall in those months is rather consistently above 100mm. The month of October stands out with an average close to 200mm. Maximum average precipitation values go above 400mm (with an extremely wet October recorded in 2007 of approximately 600mm). However, in some years, monthly average precipitation is near 0mm. The met station ‘Apt Nune Rafael’, located at Cartagena airport, has the longest meteorological record of the four stations, dating from the early 1940s to present-day. The trends in the key variables were calculated to see if any of them were significant at the 95% confidence level. In contrast to the global analysis of trends in climatic variables over past decades, it is the annual precipitation record which shows a significant upward trend at the 5% level (Figure 3-10). Over the period of the observed record (68 years), annual precipitation has increased by about 525mm, or 7.7mm/year on average. 16 Neither the mean or maximum yearly temperatures show any sign of a significant trend over the second half of the twentieth century. However, the minimum temperatures (Figure 3-11) do show an increasing trend that is significant at the 95% confidence level. The station ‘Esc Naval CIOH’ in Cartagena shows similar patterns to the airport met station. For instance, annual precipitation from 1947 to 2009 shows a significant increasing trend (Figure 3-12). Over the period of the observed record, annual precipitation has increased by about 6mm/year on average. As with the airport station, there is no uniform pattern between the three temperature variables (minimum, maximum and mean) over the same period. Annual average maximum temperature shows a significant trend at the 95% confidence level. However, somewhat unexpectedly, this is a decreasing trend (Figure 3-13). Figure 3-8 – Average monthly temperature (°C) over Cartagena for the meteorological stations ‘Apt Nune Rafael’ (blue bars) and ‘Esc Naval CIOH’ (red bars) respectively for 1941-2009 and 1953-2009. The error bars represent the maximum and minimum average temperatures recorded in each month over the time series. 17 Figure 3-9 - Average monthly precipitation (mm) over Cartagena for the meteorological stations Apt Nune Rafael (blue bars) and Esc Naval CIOH (red bars) respectively for 1941-2009 and 1947- 2009. The error bars represent the maximum and minimum average precipitations recorded in each month over the time series. Figure 3-10 – Annual average precipitation from 1941 to 2009 for the ‘Apt Nune Rafael’ (airport) meteorological station. The pink line represents the linear trend during the time period, which was found to be statistically significantly at the 5% level. 1800 1600 Precipitation in mm 1400 1200 1000 800 600 400 200 0 1920 1940 1960 1980 2000 2020 Year 18 Figure 3-11 – Annual average minimum temperature (°C) from 1943 to 2009 for the ‘Apt Nune Rafael’ (airport) meteorological station. The pink line represents the linear trend during the time period, which was found to be statistically significantly at the 5% level. 25 23 Minimum Temp (ºC) 21 19 17 15 1920 1940 1960 1980 2000 2020 Year Figure 3-12 – Annual average precipitation from 1947 to 2009 for the ‘Esc Naval CIOH’ meteorological station. The pink line represents the linear trend during the time period, which was found to be statistically significantly at the 5% level. 2000 1500 Precipitation mm 1000 500 0 1940 1960 1980 2000 2020 Year 19 Figure 3-13 – Annual average maximum temperature (°C) from 1953 to 2009 for the ‘Esc Naval CIOH’ meteorological station. The pink line represents the linear trend during the time period, which was found to be statistically significantly at the 5% level. 3.2.2 Monthly data for Buenaventura This station is located in the Buenaventura municipality on the western Pacific coast of Colombia in the south of the country. Monthly average temperatures are very similar throughout the year (around 26°C) with no clear seasonal patterns (Figure 3-14). They are lower than average temperatures in Cartagena all year around. Interannual variability exists, and average temperatures can be as low as 22.5°C (recorded in January and September) and as high as 28.2°C (recorded in July). Average monthly precipitation over Buenaventura is high, and the region is considered one of the wettest areas in the world. Months from January until March are relatively less wet than other months (with average precipitation between 300 and 400mm). Between April and August, precipitation is between 400 and 600 mm on average each. The three wettest months are September to November, which are characterized by average precipitation above 600mm (and up to 800mm for October). Year-to-year variability is relatively high, as precipitation in a month can be near 50 to 100mm between December and March, while it can be as high as 1200mm in October to December (Figure 3-15). In common with the two met stations in Cartagena, the record here (from 1961 onwards) shows a significant increasing trend in annual precipitation at the 95% confidence level. As with the other stations there is no significant change to mean temperatures, but statistically significant trends are found in both the annual mean maximum and minimum temperature. The directions of these trends differ though, with minimum temperatures increasing at the 95% confidence level (Figure 3-16) and maximum temperatures decreasing at the 95% confidence level (Figure 3-17). This signals a narrowing of the diurnal temperature range over the past 50 years in this region. 20 Figure 3-14 – Average monthly temperature (°C) over Buenaventura for 1962-2009. The error bars represent the maximum and minimum average temperatures recorded in a month in the time series. Figure 3-15 – Average monthly precipitation (mm) over Buenaventura for 1961-2008. The error bars represent the maximum and minimum average precipitation recorded in a month in the time series. 21 Figure 3-16 – Annual average minimum temperature (°C) from 1962 to 2009 for Buenaventura. The pink line represents the linear trend during the time period, which was found to be statistically significantly at the 5% level. 24 22 Minimum Temperature (ºC) 20 18 16 14 12 10 1950 1960 1970 1980 1990 2000 2010 2020 Year Figure 3-17 – Annual average maximum temperature (°C) from 1962 to 2009 for Buenaventura. The pink line represents the linear trend during the time period, which was found to be statistically significantly at the 5% level. 38 Maximum Temperature (ºC) 37 36 35 34 33 32 31 30 1940 1960 1980 2000 2020 Year 22 3.2.3 Monthly data for Barranquilla The station ‘Flores Las’ is located on the Caribbean coast of Colombia, approximately 100km north of Cartagena, at Barranquilla. Average monthly temperatures over Barranquilla show slight seasonal patterns: between December and April they are below 28°C, while for all other months they are between 28 and 29°C (Figure 3-18). Depending on the year, average temperatures have been as high as 30.5°C (recorded in June) and as low as 24°C (recorded in January). Average monthly precipitation is significantly lower in Barranquilla compared to Cartagena, though they show similar seasonal patterns (Figure 3-19): x Dry months (between January and April) with average precipitation near 0mm; x Transitional months (May-August and November-December) with modest rainfall (between 30 and 80mm); and x Wet months (September and October) of average precipitation between 125 and 150mm, which have reached 380mm in the month of October. There are no significant trends in any of the four variables at this station. This may be due to the short length of the station record (only 28 years). Figure 3-18 – Average monthly temperature (°C) over Barranquilla for 1980-2008. The error bars represent the maximum and minimum average temperatures recorded in a month in the time series. 23 Figure 3-19 – Average monthly precipitation (mm) over Barranquilla for 1980-2009. The error bars represent the maximum and minimum average precipitation recorded in a month in the time series. 3.2.4 Daily met data for Cartagena 3.2.4.1 Average daily precipitation The record of daily precipitation in Cartagena shows a very high frequency of days with 0mm of rainfall (representing about 80% of the time, over the period 2000-2009). Days of more than 0mm of rainfall are much less frequent and together represent less than 20% of all days between 2000 and 2009 (Figure 3-20). 24 Figure 3-20 – Frequency of days of rainfall between 2000 and 2009 of a certain average amount (as recorded by the meteorological station ‘Esc Naval CIOH’) (blue bars) and cumulative percentage of days with average rainfall equal to or below a certain amount (pink line) 3.2.4.2 Average daily wind speed and direction Table 3.6 presents wind speed distribution frequencies for Cartagena, Santa Marta, Barranquilla and Buenaventura, based on the wind roses provided in Figure 3-21 to Figure 3-24. Cartagena has the highest percentage of ‘calms’ (wind speeds of 0m/s) though Buenaventura, on the Pacific coast, is the least windy location overall, seeing wind speeds of less than 5.5m/s for 93% of the time. Of the three Caribbean ports, Barranquilla is slightly windier overall. For all three locations, high wind speeds, in excess of 14m/s, occur only about 1% of the time. For Cartagena and Santa Marta, both of which are situated on the coast, the most common wind direction is from the north (32% and 26% of the time respectively), with north easterly winds being the next most common. At Barranquilla, located approximately 22 km upstream of the mouth of the Magdalena River, winds are most commonly from the north east, whereas at Buenaventura, south westerlies are most common. 25 Table 3.6 – Wind speed distribution frequency (%) by threshold*. Source: IDEAM Wind speed (m/s) Cartagena Santa Marta Barranquilla Buenaventura 0 (calm) 21 14 14 3 0.0-1.5 7 36 6 48 1.6-3.3 26 24 28 33 3.4-5.4 21 17 29 9 5.5-7.9 10 4 12 1 8.0-10.7 7 3 7 1.5 10.8-13.9 5 1 3 0.5 14.0-19.9 1 0.5 1 0.1 20.0-41.9 0.1 0.1 0.1 0.01 > 42 0.01 0.01 0.01 0.001 *Note that percentage frequencies in each wind speed range have been measured from the wind roses below and are therefore approximate Figure 3-21 – Wind rose for Cartagena using data for 19 years in the period 1961-1990. Source: IDEAM * Note: 1. Percentage frequencies in each wind speed range have been measured from the wind rose above and are therefore approximate; for instance, the total amounts to 98.1% 26 Figure 3-22 – Wind rose for Santa Marta. Source: IDEAM * Note: 1. Percentage frequencies in each wind speed range have been measured from the wind rose above and are therefore approximate; for instance, the total amounts to 99.6%. 27 Figure 3-23 – Wind rose for Barranquilla. Source: IDEAM *Note: 1. Percentage frequencies in each wind speed range have been measured from the wind rose above and are therefore approximate; for instance, the total amounts to 100.1%. 28 Figure 3-24 – Wind rose for Buenaventura. Source: IDEAM * Notes: 1. Percentage frequencies in each wind speed range have been measured from the wind rose above and are therefore approximate; for instance, the total amounts to 96.1%. 3.2.4.3 Heavy precipitation As discussed in Section 3.2.1, monthly average observed climate data for Cartagena reveal the statistically significant trends in average annual precipitation shown in Table 3.7. If it was assumed that the observed increases were distributed evenly across the wet days in the year, these data indicate increases in daily precipitation amounts on wet days of 0.6% per year. The record of daily precipitation from the ‘Esc Naval CIOH’ meteorological station obtained from IDEAM, for the period 2000-2008, was too short to establish any long-term trends in daily precipitation amounts. 29 Table 3.7 – Change in annual average precipitation for meteorological stations in Cartagena (mm/year) Change in average % change in Change in annual precipitation for average average Station Period wet days only* precipitation for precipitation (mm/year) wet days only* (mm/year) (%/year) Apt Nune Rafael 1941-2009 +7.7 (7.7/65) = 0.1 Esc Naval CIOH 1947-2009 +6.0 (6.0/65) = 0.1 (0.1/15.4) = 0.6 * The average number of wet days (defined as days with precipitation amountsшϭŵŵͿŝƐϲϱнͬ - 20 days per year. This was estimated based on daily precipitation data from the station ‘Esc Naval CIOH’ for the period 2000-2008. For both data series, ŝŶĐƌĞĂƐĞŝŶĂǀĞƌĂŐĞƌĂŝŶĨĂůůƌĞĂĐŚĞĚƚŚĞϱйƐŝŐŶŝĨŝĐĂŶĐĞůĞǀĞů͘EŽƚĞƚŚĂƚŝŶƐŽŵĞLJĞĂƌƐ͕ƐĞǀĞƌĂůŵŽŶƚŚƐŽĨĚĂƚĂĂƌĞŵŝƐƐŝŶŐ͘ A study which analyzed long-term daily precipitation records for the whole of Central America and northern South America found a general trend of intensifying rainfall events and an increasing contribution of wet and very wet days to total annual precipitation (Aguilar et al., 2005). Despite regional variability, across the region as a whole the study found statistically significant trends in extreme rainfall indices (see Figure 3-25). The results of the study are summarized in Table 3.8. Table 3.8 – Extreme precipitation trends for 1961-2003 based on data from various meteorological stations in Central and South America (Aguilar et al., 2005). Both trends passed the test of statistical significance at the 5% level. Indicator of extreme Unit Annual trend (mm/year) precipitation Annual maximum 1-day mm for the maximum 1-day +0.26 precipitation amount precipitation amount in the (i.e. the rainfall amount on the year heaviest day is increasing by Ϭ͘ϮϲŵŵĞĂĐŚLJĞĂƌͿ Average precipitation per wet mm per wet day +0.03 day (defined as a day with (i.e. on average, each wet day is rainfall amounts шϭŵŵͿ getting wetter by 0.03mm each LJĞĂƌͿ 30 Figure 3-25 – Recent trends in observed precipitation extremes: Trends from 1961–2003 in SDII (simple daily intensity index, or average precipitation amount per wet day). Source: Aguilar et al., 2005 Notes: Red large triangles indicate positive significant trends; red small triangles indicate positive, non- significant trends; blue large triangles indicate negative significant trends; and blue small triangles indicate negative non-significant trends. Note: This figure shows trends from 1961–2003 in SDII (simple daily intensity index or average precipitation amount per wet day) averaged over the area shown in the top graph. Note: This figure shows trends from 1961–2003 in annual maximum 1-day precipitation amount averaged over the area shown in the top graph. 31 3.3 Tropical cyclones Cartagena is located south of the tracks of most tropical cyclones affecting the Caribbean (Figure 3-26). The chance of a tropical cyclone impacting Colombia is between 1% and 5% each year. Tropical cyclones occur from June to November, with September and October experiencing the greatest frequencies. Historically, Colombia has not been very affected by tropical cyclones or downgraded tropical cyclones. Examples of events that have affected the country are summarized in the box below. MEB reported during the site visit to have been only rarely affected by high winds (which are associated with such weather events). Figure 3-26 – Observed tracks of tropical cyclones over the Caribbean. Source: NASA, 2009 Hurricane impacts on Colombia Hurricane Lenny (1999) Affected coastal villages and the fishing industry in Colombia, killing two fishermen and leaving 540 homeless after destroying half a coastal village (BBC, 1999). At least 1,200 homes and businesses were flooded, causing extensive damage to goods and crops (ADRC, 1999). Hurricane Beta (2005) Category 1 storm, with winds of 120 km/hr causing pervasive damage to the Colombian Islands of Providencia and San Andres (IFRC, 2005). Hundreds of people were forced to evacuate San Andres or take up residence in temporary accommodation provided by the government (Associated Press, 2005). The Colombian Red Cross, with assistance from the Colombian government provided non-food aid to 600 families on Providencia, at a cost of CHF 50,000 (ICRC, 2005). Hurricane Felix (2007) A Category 2 hurricane that hit the Caribbean coast of Colombia, affecting the La Guajira peninsula and the Santa Marta mountains in particular (IDEAM, 2007). Warnings issued by NOAA predicted rainfall of 2 to 4 inches with a maximum of 6 inches over the La Guajira peninsula (NOAA, 2007). 32 4. Future climatic conditions 4.1 Multi-model projections from the IPCC AR4 over Colombia and the surrounding area This section aims to present and analyze the results of a set of global climate model (GCM) projections centered over Colombia. Colombia covers an area of approximately 1,140,000km2, extending from the Caribbean in the north (12°N) to the Amazon in the south (4°S). The country borders the Pacific Ocean to the west at 79°W and the Orinoco-Negro River marks its easternmost border at 66°W (Marin and Ramirez, 2006). 4.1.1 Data, methods and uncertainties For this study, the mean projections from a set of ten GCMs employed in the IPCC Fourth Assessment Report (2007) were analyzed. The ten GCMs used in the study are: x CGCM3.1 (T47) from the Canadian Centre for Climate Modeling and Analysis www.cccma.bc.ec.gc.ca/eng_index.shtml x CM3 from the Centre National de Recherches Météorologique, France www.cnrm.meteo.fr/gmme/ x MK3.0 from CSIRO, the Australian Commonwealth Scientific and Industrial Research Organisation www.csiro.au/science/EMM.html x CM2.0 from the US NOAA Geophysical Fluid Dynamics Laboratory www.gfdl.noaa.gov/climate-change-variability-and- prediction/?_rewrite_sticky=research/climate/ x ECHAM5 from the Max Planck Institute for Meteorology, Germany www.mpimet.mpg.de/en/home.html x GCCM2.3.2a from MRI, the Japanese Meteorological Research Institute www.mri- jma.go.jp/Dep/cl/cl.html x GISS-ER from the NASA Goddard Institute for Space Studies, US www.giss.nasa.gov/research/modeling/ x MIUB from the German Meteorological Institute of the Rheinische Friedrich-Wilhelms Universität Bonn www.meteo.uni-bonn.de/index.en.html x HadCM3 from the UK Meteorological Office Hadley Centre www.metoffice.gov.uk/climatechange/science/hadleycentre/ x HadGEM1 from the UK Meteorological Office Hadley Centre www.metoffice.gov.uk/climatechange/science/hadleycentre/ The GCMs were forced with various different greenhouse gas emissions scenarios to represent a spread of possible different emission futures, with the results then archived for analysis. An ensemble (or averaging of the results of several models) is used to create more defensible climate change scenarios, given that no one model can perfectly represent all aspects of the current climate. The models selected represent a leading group of research centers producing appropriate models of suitable complexity and representation of the current climate. Given that it is not known how the world is going to develop in the future socially, demographically, economically or technologically, various scenarios are used to represent possible different levels of 33 greenhouse gas emissions over the coming century. The IPCC has developed the most widely used emission scenarios, three of which are used in this study: x A2: This is the highest marker scenario of all the SRES scenarios, reaching 1862 gigatons (Gt) of carbon by 2100. The best estimate of global temperature change at 2100 is 3.4°C with a range of 2.4-5.0°C. A2 is frequently referred to as a ‘high’ or ‘medium-high’ emissions scenario. x A1B: This is a more middling scenario in terms of emissions with total carbon emissions of 1499Gt by 2100 as a best estimate. Temperature change of 2.8°C (1.7-4.4°C) is projected. A1B is frequently referred to as a ‘medium’ emissions scenario. x B1: This is the lowest carbon marker scenario (983Gt by 2100). This is reflected in the lower best estimate of global temperature increase by 2100 of 1.8°C with an accompanying projected range of 1.1-2.9°C. B1 is frequently referred to as a ‘low’ emissions scenario. Figure 4-1 shows the time series for projected future emissions for each scenario. Figure 4-2 highlights the projected relationship between the increase in emissions for each scenario and global temperature increase over the course of the 21st century. All of these scenarios are considered by the IPCC to be equally likely, and no probabilities are attached to them. However, some scientists have warned that GHG emissions are in fact rising more rapidly than expected. Figure 4-3 shows that the trend of current emissions is towards the top of the range offered by the IPCC scenarios. Figure 4-1 – Greenhouse gas emissions associated with four of the IPCC SRES scenarios. Source: Met Office Hadley Centre 34 Figure 4-2 – Relationship between a sample of IPCC emissions scenarios and global average temperature changes between 1990 and 2100. Figure 4-3 – Comparison of recent observed GHG emissions (black squares) with the IPCC emission scenarios out to 2020 (solid lines). Source: Legett et al., 2008. 35 4.1.2 Multi-model GCM projections In this section, climate change results for three ten-year time slices (2020-2029, 2050-2059 and 2080- 2089 or respectively 2020s, 2050s and 2080s) are discussed. Results are presented for all three emissions scenarios (A2, A1B and B2) described above, based on the ensemble mean of the ten IPCC AR4 models. Three key variables are analyzed: x seasonal average temperature (°C), x seasonal average precipitation (mm/day), and x seasonal average wind speed (m/s). Projections are provided for four seasons: December to February (DJF), March to May (MAM), June to August (JJA) and September to November (SON). Results are shown as both the absolute value for each variable for each time slice, and also, more importantly, the change in each variable relative to the baseline climatology period (1961-1990), taken from the 20C3M run of the climate models. The 20C3M run represents the ‘climate of the twentieth century’ and is forced using observed values of greenhouse gas emissions and aerosols from throughout the twentieth century, in order to replicate the observed twentieth century climate within the IPCC models. An advantage of using the 20C3M values for the climatology period is that it removes some of the model bias, since it is only the change component that is being considered. For reference, and to consider in conjunction with the change components, the observed climatology for 1961-1990 using CRU observed 0.5 x 0.5° data for the same three variables is also shown. Before considering the specifics of climate change projections over Colombia, it is first helpful to briefly look at climate change over the whole South American continent. Figure 4-4 and Figure 4-5 show temperature and precipitation changes between the period 1961-1990 and the 2080s, under the A2 emissions scenario. From the plots, it is clear that although temperatures in Colombia are projected to increase in common with the rest of the continent, the country is on the western fringe of the region of the continent where the amplitude of increased temperature is greatest. This is especially true during the months of June to November, where the greatest increases are seen over the interior northern part of the continent and the difference is greater than 5.0°C in respect to the baseline climatology period. The large size of the temperature rise here in comparison to Colombia is due to this region being remote from the moderating influence of the oceans. Conversely, the southernmost regions of Chile and Argentina can expect smaller temperature increases than Colombia, reaching less than half (up to 2.2°C) of the projected increases further north. Precipitation over South America is much more variable in general, with some areas expected to see decreases and some to see increases in future. However, a general three band pattern running diagonally across the continent from the northwest to southeast can be picked out. This consists of decreased rainfall in the top band, increased in the middle and decreased precipitation again projected in the lower band. The pattern can be picked out in all the seasons but the magnitude of change of each band varies depending on time of year. For example, the increased precipitation of the central band is projected to be felt most strongly during DJF, whilst the strongest decrease in precipitation is felt during JJA. Colombia straddles the border between the regions of mainly decreased and increased precipitation over the next century, creating varying patterns of climate change, which will be examined in more detail later. 36 Figure 4-4 – Multi-model ensemble mean projections for South American average temperature change (°C) for the A2 emissions scenario. Anticlockwise from top left: annual, DJF, MAM, JJA and SON. 37 Figure 4-5 – Multi-model ensemble mean projections for South American average precipitation change (mm/day) for the A2 emissions scenario. Anticlockwise from top left: annual, DJF, MAM, JJA and SON 38 4.1.2.1 Temperature Plots of projected temperature change (°C) and future absolute change (in Kelvin) for all the study seasons are shown in Figure 4-6 to Figure 4-13. In all cases, the changes are calculated for three time slices (2020s, 2050s and 2080s) and three IPCC SRES scenarios (A2, A1B and B1) with all resulting plots shown. The projections ranges for MEB, across the three emissions scenarios are provided in Table 4.1. Table 4.1 – Average temperature increase (°C) relative to 1961-1990 over MEB across the A2, A1B and B1 emissions scenarios from the IPCC multi-model ensemble (see Figure 4-6, Figure 4-8, Figure 4-10 and Figure 4-12) Season 2020s 2050s 2080s Projections are the same across all 0.7 to 1.2 1.2 to 2.2 1.7 to 3.7 seasons Warming is seen unequivocally in all four seasons and all time slices, with no cooling evident. The warming occurs incrementally during the twenty-first century and the magnitude of the warming is closely related to levels of greenhouse gas emissions associated with each scenario. Therefore the greatest temperature change from the baseline period is seen for the 2080s time slice for the high A2 scenario, with maximum warming seen over the eastern fringes of Colombia, measuring between 4.7 and 5.2°C during the SON season. A very similar spatial pattern of warming is observed year round across Colombia in all scenarios and in all times slices. This pattern develops most clearly from the 2050s time slice onwards and involves greater warming in the interior and east of Colombia, with a slower increase in temperatures seen to the west and in both the Pacific and Caribbean coastal regions. The December to February season follows the general pattern of warming over Colombia described above (Figure 4-6). The 2020s sees very little difference in the magnitude of warming between scenarios, with warming of 0.7 to 1.2°C over much of the west of the country and slightly higher warming of up to 1.4°C over the eastern region, emanating from the Amazon (the region of highest potential warming over the coming century). If anything, there is a greater area of slightly increased warming in the lower emissions B1 scenario. By the 2050s, the difference in magnitude of warming between the two lower emissions scenarios is starting to become more pronounced. The greatest warming under the B2 scenario at this point is 2.2°C, with a uniform increase across Colombia with the exception of the northern tip, which sees a temperature rise of just 1.7°C. In contrast, the A1B and A2 scenarios see greater warming of up to 2.7°C by the 2050s. Under the A1B scenario this increase is confined to eastern Colombia whereas it is more widespread in the A2 scenario, extending west to the Pacific coast in the central region. By the 2080s, the magnitude of warming under the three scenarios has separated out completely. Greatest warming (of up to 4.7°C) is observed under the A2 scenario in eastern Colombia, in contrast with warming of just 2.7°C over the same region in the B2 scenario. The following season, March to May, sees a very similar spatial pattern to the warming that occurs in the DJF period (Figure 4-8). However the magnitude of warming during the austral autumn is diminished; over Colombia a maximum increase of 4.2°C is seen for the 2080s under the A2 scenario, 0.5°C less than the equivalent warming during austral summer. This difference in magnitude of warming translates down through the A1B scenario. However, under the B1 scenario a very similar magnitude of warming (a maximum of 2.2°C in western Colombia and 2.7°C to the eastern border with Venezuela) to the DJF season is present. 39 Between June and August, a slightly different pattern of warming occurs by the end of the twenty- first century (Figure 4-10), with the region of maximum warming over the Amazon extending into Colombia from further southeast, confirmed by the plots of South America as a whole (Figure 4-4). The magnitude of temperature increase is in line with that for DJF, but due to the altered positioning of the area of greatest increase, a maximum difference from the climatology under the A2 scenario of between 3.7 and 4.2°C is seen over eastern and southern Colombia, with the greater projected rise of up to 5.2°C remaining across the border in Brazil. The A1B scenario gives projected temperature increases of between 2.7 and 3.2°C over western Colombia and 3.2 to 3.7°C in the east. There are small regions with other projected increases, namely the cooler Pacific coast (2.2 to 2.7°C by the 2080s) and the southernmost region, with a greater projected rise of up to 4.2°C. Under the B1 scenario, by the 2080s there is a northeast to southwest divide in temperature increase, with the cooler northwest (including the Caribbean and Pacific coastlines, moderated by the ocean) warming between 1.7 and 2.2°C and the more continental southeast seeing a temperature increase of between 2.2 and 2.7°C. The final season, September to November, sees some of the most extensive warming over Colombia (Figure 4-12). As with the other three seasons, almost identical warming occurs by the 2020s irrespective of scenario. The magnitude of warming occurring in SON across Colombia under the A1B scenario by the 2080s is very similar to that seen under the A2 scenario in JJA. By the 2080s, the A2 scenario sees increased warming of up to 5.2°C at the easternmost fringes with a more extensive region of a 4.2-4.7°C rise covering the east of the country. On the whole, western Colombia, including both coast regions sees temperatures warming in this season by 3.2-3.7°C. In general, Cartagena, being situated on the western Caribbean coast of Colombia, sees its temperature increase being moderated by the oceans. This means that, as illustrated above, it is in the region of lowest temperature increases compared to Colombia as a whole. 40 Figure 4-6 – Ensemble mean difference in average temperature (°C) for DJF based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. The scale bar is fixed for each map and subsequent mean difference temperature maps. 41 Figure 4-7 – Ensemble absolute temperature (in Kelvin, which corresponds to -273.15°C) for DJF based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. The scale bar is fixed for each map and subsequent absolute temperature maps. 42 Figure 4-8 – Ensemble mean difference in average temperature (°C) for MAM based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 43 Figure 4-9 – Ensemble absolute temperature (in Kelvin, which corresponds to -273.15°C) for MAM based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 44 Figure 4-10 – Ensemble mean difference in average temperature (°C) for JJA based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 45 Figure 4-11 – Ensemble absolute temperature (in Kelvin, which corresponds to -273.15°C) for JJA based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 46 Figure 4-12 – Ensemble mean difference in average temperature (°C) for SON based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 47 Figure 4-13 – Ensemble absolute temperature (in Kelvin, which corresponds to -273.15°C) for SON based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 48 4.1.2.2 Precipitation Plots of projected precipitation change and absolute precipitation (in mm/day) for all the study seasons, December to February, March to May, June to August and September to November are shown in Figure 4-14 to Figure 4-21. In all cases, the changes are calculated for three time slices (2020s, 2050s and 2080s) and three IPCC SRES scenarios (A2, A1B and B1) with all resulting plots shown. Unlike projected temperature changes, precipitation changes are much less uniform, with both increases and decreases projected over different regions of Colombia. Indeed, with the exception of the austral spring (SON), the remaining three seasons see projected changes to precipitation of both signs (i.e. positive and negative) within the Colombian borders. One general pattern that can be highlighted is the previously mentioned three bands running from northwest to southeast South America. Colombia appears on the boundary between the decreased precipitation of the top band and the increased precipitation projected across central South America. During austral summer (DJF), very similar patterns of equal magnitude are seen across all three emissions scenarios for the 2020s. In the southern third of Colombia, precipitation increases by 0- 0.4mm/day with the rest of Colombia seeing either no change or a very slight projected decrease in precipitation of up to 0.1mm/day. This pattern of only a very small change in either direction continues throughout the twenty-first century regardless of emissions scenario over Colombia. Greater changes in precipitation are seen in countries around Colombia – for example a projected increase of up to 0.65mm/day in Peru under the A2 scenario in the 2080s. The small changes in Colombia, irrespective of scenario or time slice, during the December to February season could be due to the nature of the season, which is generally the dry period in Colombia. During March to May, differences in precipitation change from the baseline are more pronounced between scenarios by the 2080s time slice. A very similar pattern of change to DJF is seen in the 2020s and to some extent the 2050s. But by the 2080s, the northern half of Colombia sees a quite considerable decrease in precipitation, of up to -0.85mm/day under the A2 scenario, contrasting with a projected precipitation increase of between 0.15 and 0.65 in the south of the country. June to August sees the greatest decrease in precipitation by the 2080s under the A2 scenario of all four seasons. It also sees the greatest projected increase (of up to 0.9mm/day) and this region of increased precipitation extends over a greater area than during the previous 3 months. It is the final season of September to November which does not follow the pattern described above of projected increases in precipitation to the south of the country. Here, from the 2020s, onwards the change in rainfall is negative over the whole of Colombia. In the 2020s this decrease is small (up to a 0.35mm/day decrease under A2 and A1B scenarios, and up to 0.6mm/day under the B1 scenario. However by the 2080s the decrease (particularly in surrounding countries) in precipitation is more noticeable in both the A2 and A1B plots. 49 Figure 4-14 – Ensemble mean difference in average precipitation (mm/day) for DJF based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. The scale bar is fixed for each and subsequent mean different precipitation maps. 50 Figure 4-15 – Ensemble absolute average precipitation (mm/day) for DJF based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. The scale bar is fixed for each and subsequent absolute precipitation maps 51 Figure 4-16 – Ensemble mean difference in average precipitation (mm/day) for MAM based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 52 Figure 4-17 – Ensemble absolute average precipitation (mm/day) for MAM based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 53 Figure 4-18 – Ensemble mean difference in average precipitation (mm/day) for JJA based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 54 Figure 4-19 – Ensemble absolute average precipitation (mm/day) for JJA based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 55 Figure 4-20 – Ensemble mean difference in average precipitation (mm/day) for SON based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 56 Figure 4-21 – Ensemble absolute average precipitation (mm/day) for SON based on 10 GCMs. Left column 2020s, middle 2050s, right 2080s time slices, top row A2, middle A1B and bottom B1 scenarios. 57 4.1.2.3 Wind speed Plots of projected change in wind speed in meters per second for December to February, March to May, June to August and September to November are shown in Figure 4-22. This is accompanied by plots of projected absolute wind speed in future (Figure 4-23). These are the results of just one SRES emissions scenario, A2, as this is the only scenario for which wind speed data are readily available. Projections over MEB, taken from these plots, are presented in Table 4.2. Table 4.2 – Average projected wind speed changes (meters/second) from 1961-1990 over MEB for the A2 emissions scenario from the IPCC multi-model ensemble. Months 2020s 2050s 2080s DJF 0 to 0.1 -0.1 to 0 0 to 0.1 MAM 0 to 0.1 0.1 to 0.2 0.2 to 0.3 JJA 0.1 to 0.2 0.1 to 0.2 0.4 to 0.5 SON 0 to 0.1 0.1 to 0.2 0.2 to 0.3 As shown in Figure 4-22, for all four seasons (DJF, MAM, JJA, SON) there is little change in wind speed evident under the A2 scenario. Although there is some fluctuation around 0, this does not exceed 0.1 m/s in either a positive or negative direction during the 2020-2029 period over Colombia. This is a pattern which continues into the 2050s, although by this point there are some noticeable changes in wind speed in the region surrounding Colombia, which vary depending on season and could impact on the climate over Colombia. For December to February, off the northern Caribbean coast of Colombia wind speeds are projected to decrease by speeds of up to -0.1 to -0.2m/s during the 2050s, a pattern which is maintained through into the 2080s. This is accompanied by an increase in wind speed of up to 0.3m/s over eastern northern South America. The following season, the pattern of change over the Caribbean is the opposite of that seen from December to February. During MAM in the 2050s a pattern of increases of up to 0.4m/s emerges. In the 2080s this pattern intensifies leading to an increase in wind speeds of between 0.5 and 0.6m/s off the Caribbean coast of Colombia. In contrast, off the southeast Colombian coast a slight decrease in wind speeds of 0.1-0.2m/s has emerged by the 2080s. During austral winter (June to August) wind speed changes under the A2 scenario follow a very similar spatial pattern to the changes of March to May. However, by the 2080s the region of increased wind speed off the north Colombian coast reaches a peak of 0.9m/s with an increase of up to 0.6m/s also evident in northern Colombia. This pattern also curves round into Venezuela and northern Brazil. As with JJA and MAM, the months of September to November follow the same pattern of increased wind speed of the northern coast of Colombia. The magnitude of the increase is less than during austral winter but greater than during March to May, with the maximum increases compared to the baseline of 0.7m/s. Again, there are also regions of slight projected decrease in wind speed of 0.1 to 0.2m/s off the western Pacific coast of Colombia. 58 Figure 4-22 – Ensemble mean difference in wind speed (m/s) based on 10 GCMs for the A2 emissions scenario. Left column 2020s, middle 2050s, right 2080s time slices, top row DJF, second MAM, third JJA and bottom SON seasons. 59 Figure 4-23 – Ensemble mean absolute wind speed (m/s) based on 10 GCMs for the A2 emissions scenario. Left column 2020s, middle 2050s, right 2080s time slices, top row DJF, second MAM, third JJA and bottom SON seasons. 60 4.1.2.4 Sea level Sea level changes could have a significant impact on MEB. In general, estimates of sea level rise include four components: x Thermal expansion (calculated from the climate models, specifically the ocean component), x Glaciers and ice caps (computed from simple empirical formula which links global mean temperature to mass loss based on observed data from 1963 to 2003), x Ice sheet surface mass balance, and ice sheet surface balance model with snowfall amounts and temperature (computed from a high resolution model scaled to the coupled models), and x Dynamical imbalance (computed from extrapolation of observed rates from 1993 to 2003 and contributing up to 0.7mm during this period) which corresponds to changes in the Greenland and Antarctic ice sheets. IPCC AR4 projections for sea level rise The IPCC AR4 Summary for Policy Makers gives global sea level rise (SLR) projections for the end of the century, as shown in Table 4.3. Contributions to sea level rise are as follows for the warmest emissions scenario, A1FI: x 28 cm derived from thermal expansion, x 12 cm from glaciers, x -3 cm from ice sheet mass balance, and x 3 cm from ice flow. Table 4.3 – Global sea level rise projections (in meters) in 2090-2099 relative to 1980-1999 for six emissions scenarios from an IPCC multi-model ensemble. Note that these projections exclude potential future rapid dynamical changes in ice flow. Emissions scenarios Range of projections B1 0.18 – 0.38 A1T 0.20 – 0.45 B2 0.20 – 0.43 A1B 0.21 – 0.48 A2 0.23 – 0.51 A1FI 0.26 – 0.59 Spanning emissions scenarios, the range of sea level rise according to the IPCC AR4 is therefore 18 to 59 cm by the end of the twenty-first century relative to the last two decades of the twentieth century. It is important to note the following about the projections from the IPCC AR4: x Compared with the projections from the IPCC Third Assessment Report (2001), the ranges of sea level rise are narrower (Figure 4-24). This is because the IPCC AR4 based its projections on revised information about some uncertainties in the projected contributions to sea level rise. However, as shown in Figure 4-24, when the potential contribution of increased ice- sheet melting due to the dynamic response of the Greenland and Antarctic ice sheets to global warming is included (this contribution is currently not well understood and is poorly quantified), the IPCC projections of 2001 and 2007 are comparable. 61 x Important uncertainties remain in sea level rise projections. Models used to date do not include uncertainties in climate-carbon cycle feedback nor do they include the full effects of changes in ice sheet flow (including dynamical processes), because a consensus in the published literature was lacking at the time of writing the IPCC AR4 in 2007. The projections include a contribution due to increased ice flow from Greenland and Antarctica at the rates observed for 1993-2003. Current global model studies project that the Antarctic ice sheet will remain too cold for widespread surface melting and may gain in mass due to increased snowfall. However, net loss of ice mass could occur if dynamical ice discharge dominates the ice sheet mass balance. Although the IPCC does not provide global sea level rise projections for nearer time periods than the end of the century, these can be extrapolated from Figure 4-24. Figure 4-24 – Sea level rise projections from the IPCC Third Assessment (green lines) (TAR) and Fourth Assessment Reports (AR4) (red bars). (Source: Church et al., 2008) Notes: (a) The dark green shading is the average envelope for all emissions scenarios; the light shading is the average envelope for all climate models and all emissions scenarios; the outer lines include an allowance for additional land-ice uncertainty. (b) The red bar plotted at 2095 represents the IPCC AR4 projections; the dark red bar is the extended sea level rise range to allow for additional contribution to sea level rise from potential ice-sheet dynamic processes. (c) The inset graph pots the observed rate of sea level rise from tide gauges (blue line) and satellite altimeters (orange line) for 1990-2006 against the IPCC sea level rise simulations of 2001. Estimates of sea level rise from recent research Since the IPCC AR4 was published in 2007, there have been a number of studies which conclude that sea level rise by 2100 could exceed 100 cm. These include: x The assessment of the Dutch Delta Commission, 62 x The Synthesis Report of the Copenhagen Climate Congress, x The Copenhagen Diagnosis report, and x The Scientific Committee on Antarctic Research (SCAR) report on Antarctic Climate Change. In addition there are several peer-reviewed papers which point to revised estimates (e.g. Rahmstorf et al 2007, Pfeffer et al. 2008, Vermeer and Rahmstorf 2009, Grinstead et al 2010). Rahmstorf et al (2007) note that since 1990 sea level has been rising faster than projected by climate models, as shown both by tide gauge data and satellite altimeter data (see Figure 4-24, inset). The satellite data show a linear trend of 3.3mm (± 0.4) per year between 1993 and 2006 and the tide gauge data trend is slightly less, whereas the IPCC projected a best-estimate rise of less than 2 mm per year. The largest contributions to the rapid observed rise in global sea levels come from ocean thermal expansion and the melting of non-polar glaciers. Although the ice sheet contribution has been small, observations indicate that it is increasing rapidly, with contributions both from Greenland and Antarctica. Recently, Vermeer and Rahmstorf (2009) have provided estimates of sea level rise based on a simple relationship linking global sea level variations to global mean temperature. Using this relationship, they project global average sea level rise by 2100 (compared to 1990) in the range 75 to 190 cm (Figure 4-25). Figure 4-25 – Sea level rise estimates from Vermeer and Rahmstorf (2009) Sea level rise for Cartagena To consider risks from sea level rise at MEB, a range of scenarios of future changes needs to be assessed. Sea level has been measured in the Bay of Cartagena for more than 50 years at the locations shown in Figure 4-26. Between 1951 and 1993 a tide gauge was located at the Naval Club located at the tip of Castillogrande; in 1993 it was moved to the pier of CIOH. 63 By far the most comprehensive local study on observed sea level change in Cartagena is that included in the assessment by Alexandre et al. (2008). Using sea level data from the tide gauges in the bay over the period 1951-2000, sea level rise was found to be around 5.6mm (± 0.008mm) per year (see Figure 4-27). This includes the effects of land subsidence, which, according to another study, is sinking at a rate of 2.7mm per year, based on GPS observations (Sutherland et al., 2008). This observed rate of sea level rise makes Cartagena an area of particular concern (though the geology of the area suggests that similar patterns could be present in the surrounding coastline). Another sea level rise study based on the tide gauges in the Bay of Cartagena found rates of 4.5 and 4mm per year (for 1952-1992 and 1951-2000 respectively) (Andrade, 2008). Observations from satellite altimetry, which do not include the effects of local land subsidence, show the Caribbean coast of Colombia experiencing sea level rise of between 2 and 5 mm per year (Figure 4-28). The quality of the Bay of Cartagena tide gauge dataset has been questioned by some researchers, due to: the high percentage of missing data in certain years; the relatively high spread around the mean of the dataset (standard deviation); and the movement of the tide gauge in 1993 (Torres et al., 2006). As a result, some researchers have instead inferred sea level rise in Cartagena using a tide gauge located at Cristobal in Panama. Sea level rise in Panama is reported to be 2mm per year over the period 1907-1997, and 3.6mm per year from 1952 to 1997 (Torres et al., 2006). As part of this study, the observed sea level data from the tide gauges in the Bay of Cartagena (1951- 2000) were reviewed and analyzed, to assist in the development of credible SLR scenarios for MEB. The dataset was compared to sea level observations from Cristobal, Panama, to identify any potential time periods for which the data might be erroneous. No anomalies were identified. The rate of sea level rise for the dataset was calculated using linear regression and a linear trend line was fitted to the daily mean data using the least squares method. While globally the rate of sea level rise appears to be accelerating, the trend shown by the dataset is linear, as shown in Figure 4-29. In agreement with other studies (Sutherland et al., 2008 and Alexandre et al., 2008), the average annual increase for 1952-1993 was found to be 5.6mm. Note that the data from Cartagena tide gauge for the period 1993-2000 were not included in the analysis to avoid any issues that could arise from joining the two datasets. Based on this analysis, the study of MEB used the observed rate of 5.6mm per year as one of the scenarios of future sea level rise, to assess flood risk at the port. This rate includes the observed influence of local oceanographic and meteorological factors. An upper estimate for sea level rise is difficult to pinpoint because there is little consensus as to what this should be, as outlined above. So, a second ‘Accelerated SLR’ scenario has been used in the flood risk assessment for MEB. This scenario applies a rate of SLR which increases from the current rate of 5.6 mm per annum up to 20 mm per annum by 2100. 64 Figure 4-26 – Locations of tidal sea level measurements in Bay of Cartagena (indicated by the red triangles) Figure 4-27 – Sea level time series (which combines two tide gauges in Bay of Cartagena) from 1951 to 2000 (Alexandre et al., 2008) Measured Sea Level (mm) Time (days from 1 Jan 1951) 65 Figure 4-28 – Recently observed regional patterns of sea level rise from satellite altimeters (explained mostly by regional variability in thermal expansion, ocean density, circulation and salinity). Source: Cazenave et al. 2008 Figure 4-29 – Sea level time series from the tide gauge in the Bay of Cartagena from 1951 to 1993, including linear trend (+5.6mm per annum) calculated using least squares method. 1 0.9 0.8 Sea-level (Daily Mean m) 0.7 0.6 0.5 0.4 0.3 1949 1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 66 Comparison between rates of sea level rise in Cartagena and Buenaventura Sea level rise due to climate change on the Pacific coast of Colombia (where the port of Buenaventura is located) may be slightly lower than on the Caribbean coast. For instance, the IPCC projects an increase in sea level on the Pacific coast which is up to 5cm less than the global average by the end of the century, while sea level rise projections for the Caribbean coast are 5cm higher than the global average. In Cartagena itself, high observed rates of sea level rise can be partly explained by land subsidence rates of 2.7mm per year. (Land subsidence rates on the Pacific coast are not known). However, sea level on the Pacific coast is affected by decadal variations: during El Niño events, sea levels are raised by between 20-35cm (Restrepo et al., 2002), which will add to increased sea level due to climate warming. 4.1.2.5 Thunderstorms and lightning Thunderstorms and lightning primarily depend on convective clouds such as the cumulonimbus, which are not adequately resolved by GCMs due to their coarse spatial resolution. Thunderstorms form when an air mass becomes unstable due to a layer becoming unusually warm and humid and/or unusually cool and starts overturning (or convecting) violently. Lightening occurs when strong wind updrafts within thunderstorms carry small ice particles to the upper region of the cloud which collide and produce electrical discharges (Trapp et al., 2007). Climate models generally do not simulate thunderstorms and lightning directly. Researchers have attempted to derive estimates of future thunderstorms and lightning from empirical relationships between observed variables. For instance, in the case of lightning, some scientists have derived estimates of future peak lightning flash rate from daily data on wind velocity during the season when lightning is most prevalent, because convective thunderstorms can also produce high wind speeds (Hulme et al., 2002). A study using one set of RCM projections for one emission scenario found a net increase in the number of days in which the meteorological conditions that lead to severe thunderstorms occur and suggests that the number of days of severe thunderstorm environmental conditions could increase two-fold in many locations in the US by the end of this century (such as Atlanta and New York) (Trapp et al., 2007). Attempts to estimate future changes in lightning have not been undertaken for this study in light of their low confidence and the lack of evidence that lightning pose a risk to MEB. However, in general, increased global mean surface temperatures and higher atmospheric water vapor would be expected to increase rather than decrease lightning incidence (Trapp et al., 2007 and Hulme et al., 2002). 4.2 Projections from empirical downscaling Climate projections from the IPCC multi-model ensemble, discussed in Section 4.1, offer the most comprehensive summary of future climate. However, at approximately 2.5 x 2.5 degrees, the spatial resolution of these models is coarse and in the case of a mountainous country such as Colombia, downscaling to finer resolution is desirable. There are two main routes to downscaling: dynamical or statistical (empirical). Dynamical downscaling involves running regional models over a limited domain at resolutions typically at or better than 50 km (see Section 4.3 for results from one regional model, PRECIS). Regional climate models are very demanding computationally; to complete a suite of runs forced by several global 67 climate models, they require several years of computing time on fast workstations. Therefore, statistical downscaling offers a feasible alternative to dynamical downscaling. Empirical downscaling consists of deriving climate change information from a present-day statistical model which relates large-scale climate variables to regional and local variables, based on site- specific observed climatic data. It is a technique that is computationally inexpensive and thus can be applied to output from different global climate model experiments (which is the most robust approach for impact analyses). Statistical downscaling assumes that the present-day statistical relationship holds true in future climate conditions. Statistical downscaling has been undertaken for Colombia for individual IPCC AR4 global climate models (Jarvis Pers. Comm.). The results of 14 IPCC AR4 models downscaled over the Colombia region to 10 minute spatial resolution have been averaged to create an ensemble mean product, which is discussed next. 4.2.1 Temperature The statistically downscaled plots of maximum temperature change over Colombia (see Figure 4-30) do not show the unequivocal warming that the maps of mean temperatures from the GCMs showed. Instead, there is a distinctive pattern of change to the maximum temperatures by the 2050s, which is largely consistent through the year. Areas which are projected to see a decrease in maximum temperatures under increasing greenhouse gases are the Andes mountain ranges running through Colombia from south to north, with the largest decreases in maximum temperature (up to 6°C) corresponding with the highest peaks in all seasons (DJF, MAM, JJA, SON). The smallest decreases of between 0 and 1°C in maximum temperature are found at the fringes of the mountains. In contrast, the lowland areas of Colombia are projected to see increases in maximum temperature by the mid twenty-first century. As with areas of decrease, the spatial pattern of the increase stays relatively consistent throughout the seasons. The regions which see the greatest increase in maximum temperature (of up to 6°C) are the Caribbean coastal region and the northern Pacific coastal region, with both regions of this maximum increase extending inland somewhat. This contrasts with the projections for mean temperature using the AOGCM ensemble where the coastal region was projected to see some of the lowest increases in temperature. From the statistical downscaling, the rest of Colombia is projected to see maximum temperatures rise between 1 and 4°C by the 2050s. The boreal winter months of December to February see the most extensive warming over eastern inland Colombia, especially in the north. The lowest projected increases in maximum temperature are consistently found in south-east Colombia, and range from no change in maximum temperatures to an increase of 1.5-2°C by 2050. In this region, the smallest warming is found in DJF and the maximum warming is found between June and November. Figure 4-31 shows the downscaled minimum temperature projections over Colombia for the 2050s. The spatial pattern of areas where minimum temperature is projected to increase and decrease is very similar to that for maximum temperatures. Once again, minimum temperatures are projected to decrease over the Andes by up to 6°C during all seasons. The greatest projected increases in minimum temperatures (again by up to 6°C) are found in the coastal regions, in particular the Caribbean coast. Minimum temperatures here are also projected to increase by the same amount further inland during the months of June, July and August particularly, extending the region of this maximum projected increase. As with maximum temperatures, the region of lowest projected increase in the minimum temperatures is found in south-east Colombia, where the projected increase measures between 1.5°C in the far south, increasing to up to 4°C further north. 68 Given that the spatial patterns for projected changes in both minimum and maximum downscaled temperatures are similar, it can be inferred that it is likely that under these downscaled projections the diurnal temperature range over Colombia will be projected to remain fairly consistent. Averaging the projected changes in maximum and minimum temperatures over MEB indicates increases in mean temperature of about 6°C (±0.5) by the 2050s across all seasons. 4.2.2 Precipitation In contrast with the projected changes to maximum and minimum temperatures, projected changes in precipitation from empirical downscaling show considerable variation between seasons (Figure 4-32) and so each season will be considered in turn. The empirical downscaled projections over MEB are summarized in Table 4.4. Table 4.4 – Projections of annual rainfall (mm/day) in 2050 from empirical downscaling Season Projected change in precipitation (mm/day) DJF -0.6 MAM -0.6 JJA +0.6 or more SON +0.6 or more (though areas very close to Cartagena show projections of decreased precipitation of at least 0.6 mm/day) From December to February (dry season), according to the empirical downscaling, it is projected that most of Colombia will see a decrease in precipitation. This region of decrease covers the Andes, the Caribbean coast and central and eastern inland Colombia and is about 50mm over the season (or decrease of 0.6 mm per day) The only areas which are projected to see an increase in precipitation during these boreal winter months are the Pacific coastal region and far southeast of Colombia, both of which are projected to see increases of up to 50mm (or increase of up to 0.6 mm per day). These distinctions between projected increased and decreased precipitation during December to February are very clear cut, with just a narrow transitional band running across South America between the region of projected decrease to the north and increase in the south. The pattern of projected precipitation change from March to May (transitional rainfall season) is much more varied than in the preceding months. However, one consistent feature is the projected decrease in precipitation over the Caribbean coast where Cartagena is located and the projected increase over the Pacific coast, both changes to the magnitude of up to 50mm (or 0.6 mm per day). The inland regions of Colombia see a variety of projected changes in precipitation. It can be noted that the southeast corner is again projected to see a seasonal increase in precipitation (0-50mm) with decreases projected further north. During JJA (wet season), rainfall is projected to increase over the Caribbean coast and Eastern Colombia by at least 0.6 mm per day (or 50 mm over the season). Other parts of the country, including parts of the Andes and the South East will experience decreases of at least 0.6 mm per day (or 50 mm for the season). For the months of September to November, projected precipitation shows much more variable patterns. Over MEB, precipitation is projected to increase by at least 0.6 mm per day, though areas close to Cartagena are projected to see decreases in precipitation of at least 0.6 mm per day. Apart 69 from the Northern Pacific coast where increased precipitation is clearly projected, changes in precipitation are very variable across Colombia. Figure 4-30 – Projections of future changes in average maximum temperatures (°C) over Colombia from the downscaling of 14 GCMs under the A2 emissions scenario for the 2050s (compared with 1961-1990). From left to right and top to bottom: DJF, MAM, JJA and SON 70 Figure 4-31 – Projections of future changes in average minimum temperatures (°C) over Colombia from the downscaling of 14 GCMs under the A2 emissions scenario for the 2050s (compared with 1961-1990). From left to right and top to bottom: DJF, MAM, JJA and SON (average of model ensemble) 71 Figure 4-32 – Projections of future changes in average precipitation (mm per season) over Colombia from the downscaling of 14 GCMs under the A2 emissions scenario for the 2050s (compared with 1961-1990). From left to right and top to bottom: DJF, MAM, JJA and SON (average of model ensemble) 72 4.3 Projections from a regional climate model: PRECIS PRECIS is a regional climate modeling system from the UK Hadley Centre. It has been used as part of Colombia’s Second National Communication on Climate Change, as it can resolve spatial features down to about 25km2. Regional climate models use outputs from GCM simulations to drive high- resolution simulations for selected time periods. Overall, regional climate models allow the representation of local surface features because of their finer spatial resolution. They are reliant on the forcing from the parent model and therefore tend to reproduce any systematic bias in those models. They are best used in collaboration with GCMs to add confidence in future climate. However, they have important limitations. Because they are hugely demanding computationally, the state of science at the moment typically permits the analysis of one regional model (e.g. PRECIS) driven by one global model (GCM) only, for a range of emissions scenarios. Since the regional model output is highly dependent on global model input, the sampling limitations of this approach (i.e. the limitation of using a single global model) normally outweigh the advantages. Figure 4-33 shows the precipitation change under climate change for a variety of global climate models. Clearly there is a wide spread of results from wet to dry. Forcing a regional climate model with a single global model from the set shown in Figure 4-33 would clearly lead to a biased downscaling experiment. Ideally, a range of regional models, each forced with a range of global climate models, themselves forced by several emissions scenarios, provides a more robust set of projections, assuming the regional models cope with the regional climate reasonably well (indications are that this is not the case for Colombia). Regional modeling and dynamical downscaling needs to be coordinated as a global effort (an initiative which is being taken forward in IPCC AR5 under CORDEX). Further, some scientists have found that PRECIS performs poorly over Colombia compared with other high-resolution climate models, such as the Japanese MRI model for example, (Jarvis, Pers. Comm.). A summary of PRECIS annual projections for Cartagena for the 2020s and 2080s (compared to the 1961-1990 baseline), based on two emissions scenarios (A2 and B2), is presented below (Table 4.5). It is intended that presenting the results from PRECIS for Cartagena should facilitate consideration of the conclusions from this study in the context of climate assessments being undertaken by the government. Appendix 2 provides a full report of the PRECIS analysis. 73 Figure 4-33 – GCM projections of average precipitation change (mm/day) over Colombia for 18 models CCCMA-CGCM3.1 BCCR-BCM2.0 CCCMA-CGCM2 CCCMA-CGCM3.1-T63 CNRM-CM3 IAP-FGOALS-1.0G T47 GISS-AOM GFDL-CM2.1 GFDL-CM2.0 CSIRO-MK3.0 IPSL-CM4 MIROC3.2-HIRES MIROC3.2-MEDRES MIUB-ECHO-G MPI-ECHAM5 MRI-CGCM2.3.2A NCAR-PCM1 UKMO-HADCM3 74 Table 4.5 – Summary of PRECIS projections for Cartagena for the 2020s and 2080s. (The grid point close to MEB is coloured in yellow). TEMPERATURE COORDINATES OBSERVED SIMULATED SCENARY A2 SCENARY B2 2071 - 2100 LAT LONG 1961-1990 1961-1990 1971-2000 2011-2020 2021-2030 2031-2040 2011-2040 2011-2020 2021-2030 2031-2040 2011-2040 SCENARY A2 SCENARY B2 10,705 -75,76 27,9 27,8 27,5 27,7 27,8 27,7 27,5 27,9 28,0 27,8 29,8 29,0 10,705 -75,54 27,8 27,8 27,4 27,7 27,8 27,6 27,5 27,8 27,9 27,7 29,8 29,0 10,705 -75,32 27,5 27,6 28,7 28,9 28,9 28,9 28,7 29,0 29,2 29,0 30,7 29,9 10,486 -75,76 27,8 27,8 27,4 27,6 27,7 27,6 27,4 27,8 27,9 27,7 29,6 28,8 10,486 -75,54 27,5 27,2 27,3 28,5 28,7 28,7 28,6 28,5 28,8 28,9 28,8 30,6 29,8 10,486 -75,32 26,2 26,5 29,4 29,7 29,6 29,6 29,4 29,8 29,8 29,7 31,5 30,6 10,267 -75,76 27,8 27,8 27,3 27,5 27,6 27,5 27,3 27,7 27,8 27,6 29,3 28,6 10,267 -75,54 27,0 27,1 28,6 28,8 28,8 28,7 28,6 28,9 29,1 28,8 30,4 29,6 10,267 -75,32 25,9 26,2 29,6 29,8 29,8 29,7 29,5 29,9 30,0 29,8 31,5 30,7 RELATIVE HUMIDITY COORDINATES OBSERVED SIMULATED SCENARY A2 SCENARY B2 2071 - 2100 LAT LONG 1961-1990 1961-1990 1971-2000 2011-2020 2021-2030 2031-2040 2011-2040 2011-2020 2021-2030 2031-2040 2011-2040 SCENARY A2 SCENARY B2 10,705 -75,76 83,5 83,9 89,4 89,6 89,9 89,6 89,6 89,7 89,7 89,7 85,8 86,2 10,705 -75,54 84,6 84,8 89,1 89,3 89,6 89,3 89,3 89,4 89,4 89,4 85,4 85,9 10,705 -75,32 85,4 85,2 84,8 84,8 85,4 85,0 84,9 85,1 85,1 85,0 81,9 82,5 10,486 -75,76 83,8 84,2 90,4 90,6 91,0 90,7 90,7 90,7 90,7 90,7 87,2 87,8 10,486 -75,54 80 86,1 85,9 86,8 87,0 87,6 87,1 87,0 87,1 87,3 87,2 83,9 84,5 10,486 -75,32 88,2 87,3 81,6 81,8 82,7 82,0 82,0 82,0 82,4 82,1 79,3 79,9 10,267 -75,76 84,3 84,7 91,4 91,6 91,9 91,6 91,6 91,6 91,7 91,6 88,6 89,1 10,267 -75,54 86,5 86,3 86,0 86,3 86,9 86,4 86,4 86,4 86,6 86,5 84,1 84,5 10,267 -75,32 88,0 87,0 79,4 79,8 80,6 79,9 80,1 80,0 80,3 80,1 77,6 78,0 PRECIPITATION COORDINATES OBSERVED SIMULATED SCENARY A2 SCENARY B2 2071 - 2100 LAT LONG 1961-1990 1961-1990 1971-2000 2011-2020 2021-2030 2031-2040 2011-2040 2011-2020 2021-2030 2031-2040 2011-2040 SCENARY A2 SCENARY B2 10,705 -75,76 2797,9 2593,9 352,1 337,6 373,7 354,4 294,9 289,7 412,9 332,5 1148,7 904,3 10,705 -75,54 1786,4 1641,6 217,8 194,4 214,5 208,9 177,9 171,5 240,8 196,7 782,4 579,1 10,705 -75,32 1292,8 1180,6 332,5 339,4 364,3 345,4 317,7 339,8 358,3 338,6 668,7 599,5 10,486 -75,76 1969,2 1878,3 290,6 286,6 300,0 292,4 253,5 254,5 363,3 290,4 543,7 454,9 10,486 -75,54 1020,5 1986,5 1898,7 502,6 525,1 548,6 525,4 481,4 499,5 574,3 518,4 1061,2 947,9 10,486 -75,32 2911,9 2763,2 1113,6 1209,2 1251,4 1191,4 1116,0 1171,2 1255,9 1181,0 1900,2 1740,7 10,267 -75,76 1919,3 1831,9 293,9 292,2 296,2 294,1 276,7 277,7 372,9 309,1 418,2 307,6 10,267 -75,54 2046,4 1949,9 488,8 514,6 530,8 511,4 491,2 501,3 547,8 513,4 813,1 684,9 10,267 -75,32 2814,0 2670,2 1104,6 1187,4 1230,9 1174,3 1137,4 1179,4 1210,9 1175,9 1630,2 1460,2 75 4.3.1 Temperature As shown in Table 4.5, PRECIS projects temperature increases over MEB for the 2020s and 2080s of 1.1-1.2°C and 2.3-3.1°C respectively. These projections are in line with the results of the IPCC multi- ensemble GCM, but are significantly lower than the empirically downscaled projections for the 2050s presented in Section 4.2. 4.3.2 Precipitation As shown in Table 4.5, simulated baseline precipitation from PRECIS for 1961-1990 (1,987mm) is approximately double the observed precipitation amount (1021mm). PRECIS also simulates a very large future reduction in precipitation over MEB by the 2020s – of about 50% compared to the observed baseline, or 75% compared to the PRECIS simulated baseline. By the 2080s, PRECIS projects precipitation amounts similar to the observed baseline, but representing only about 50% of the simulated baseline. The full PRECIS report (Appendix 2), recommends that observed trends in precipitation be used in place of the PRECIS projections, in the light of these results. 4.3.3 Wind PRECIS projects a future increase in the frequency of wind speeds in the range 3 -10 meters per second, across both the A2 and B2 emissions scenarios by the 2080s, compared to the 1961-1990 period. 4.4 Estimates of changes in future extreme rainfall Overall global climate models predict that: x The intensity of precipitation events will increase, particularly in tropical areas that experience increases in average precipitation (IPCC WG2, 2007), x In most tropical areas, precipitation extremes will increase more than average precipitation (IPCC WG2, 2007). Data from six-hourly time steps are available from the IPCC AR4 data set for several global climate models. It is therefore possible to calculate changes in rainfall intensity from these data. However, daily rainfall from global models is generally regarded as unreliable and extreme rainfall even less reliable, unless comprehensive work has been done on evaluating whether the climate mechanisms or weather systems simulated by the models are realistic for the region of interest. In the case of Colombia, which is characterized by extreme topographical gradients, there is an even greater need for this extensive background work. Where it is not feasible to undertake this extensive work or where preliminary work has pointed to substantial difficulties with the global and or regional climate models (such as in the case of this study), it is helpful to consider the trend in observed rainfall characteristics, not least because these are likely to be the best guide until 2025 even in the presence of a climate model that performs well (given that the greenhouse gas forcing only becomes pronounced with respect to natural variability by about 2025). 76 While acknowledging limitations 1, it is therefore sensible to test the capacity of MEB’s facilities (e.g. drainage systems) to cope with the observed extreme precipitation trends shown in Table 3.7 and Table 3.8, namely: x 0.6% per year increase in precipitation amounts on wet days; and x 0.26mm per year increase in the precipitation amount on the wettest day. 4.5 Tropical cyclone projections Knutson et al. (2010) provide the most recent comprehensive assessment of both recent trends in tropical cyclone activity and projections of future possible trends in such activity over the course of the next two centuries. Generally, the detection of trends of tropical cyclone activity over the past decades is complicated by large natural fluctuations in their frequency and intensity, in addition to limits in the availability and quality of historical records of cyclone activity. Similarly, projections of future changes in tropical cyclone activity are limited through uncertainty about potential changes to large-scale tropical climate under anthropogenic greenhouse gas increases and how this will impact on cyclone activity. However, despite this, there has been significant progress in assessments of tropical cyclone activity, which the authors document (Knutson et al. 2010). Tropical cyclone frequency data appear to show statistically significant trends of an increase in tropical storm frequency over the twentieth century. However, this trend is much decreased after accounting for fewer observations (and therefore the possibility of ‘missing’ cyclones) prior to 1966, when satellite data became available. Furthermore, the increased trend is mostly due to an increase in two-day events which are most likely now being sampled more reliably, owing to improvements in observing systems. Notably, the trend is not significant at the 5% level after taking into account these considerations. Similarly, both hurricane counts from 1850 to the present, and land falling tropical storm and hurricane activity over the US show no long term increases. Knutson et al (2010) conclude that it is uncertain if past changes to tropical cyclone frequency have exceeded the envelope of natural variability and therefore whether there is a discernible trend of anthropogenic climate change. Projections of changes in tropical cyclone frequency over the next two centuries show some convergence on a global level, although individual basin activity shows a distinct lack of agreement. It is concluded that it is likely that global mean tropical cyclone frequency will either decrease or remain essentially unchanged. Modeling studies indicate global average decreases of 6% to 34% by the end of the twenty-first century (under the SRES A1B scenario, as with all projections made by Knutson et al. 2010). Mean frequency in the southern hemisphere is projected to decrease more robustly than in the northern hemisphere. A second property of tropical cyclone activity which is examined by Knutson et al (2010) is intensity. Theoretically, future sea surface warming and associated changes in the atmospheric state will lead to increases in the upper limits of tropical cyclone intensity. This is reinforced by high resolution models which indicate an increase in both mean intensity and the number of cyclones of high intensity. Satellite data since 1981 supports an increase in tropical cyclone intensity, with the most significant increases found in the tropical Atlantic (although uncertainties exist in some regions such 1 To build confidence in analyses of the extremes, it is necessary to establish what kind of climate system is responsible for the extreme events, to evaluate how well the models simulate these systems in the control run and to assess what happens to these systems in the future. This is beyond the scope of the analysis that can be performed for this study. 77 as the Indian Ocean). However, due to the short length of the observational record, it is not possible to distinguish whether this is due to natural variability or to increases driven by man-made greenhouse gas emissions. Several projections have been made of changes in the future intensity of tropical cyclones in individual tropical cyclone basins. In general, model projections at the basin scale are not in agreement about whether cyclone intensity might increase or decrease (see Figure 4-34). At a global scale, however, Knutson et al (2010) estimate increases in the mean maximum wind speed of tropical cyclones, of between 2 and 11% over the twenty-first century. In relation to tropical cyclone rainfall amounts, Knutson et al (2010) conclude that so far there is no detectable change, even though atmospheric moisture content has increased in many regions in recent decades. However, they conclude that it is likely that tropical cyclone rainfall rates will increase in the future with global warming, stating that this is a robust projection across all available model experiments. The range of projections from existing studies is for precipitation increases of between +3 to +37% over the twenty-first century. The final characteristics of tropical cyclones examined are the genesis, tracks, duration and storm flooding. There is no conclusive evidence that there have been recent observed changes in any of these variables that exceed natural variability. Similarly, the authors state that there is little confidence in projected future changes to these tropical cyclone features. Figure 4-34 – Downscaling projections forced by seven GCMs of future change (in %) by the end of the 22nd century in tropical cyclone intensity (measured by a power dissipation index) for five tropical cyclone basins (The scale is the logarithm of the power dissipation ratio index multiplied by 100). Source: Knutson et al., 2010. 78 To summarize, at this time it cannot be concluded whether there is an identifiable anthropogenic signal in tropical cyclone activity over the past decades, although equally it cannot be ruled out. Progress made in projections of twenty-first century cyclone activity indicate that globally, the frequency of all tropical cyclones may remain the same or decrease. In relation to the most intense tropical cyclones, Knutson et al (2010) conclude: ‘We judge that a substantial increase in the frequency of the most intense storms is more likely than not globally, although this may not occur in all tropical regions. Our confidence in this finding is limited, since the model-projected change results from a competition between the influence of increasing storm intensity and decreasing overall storm frequency’. Cartagena, however, is unlikely to experience an increase in tropical cyclone activity owing to its position on the landward side of the cyclone tracks, and its low latitude. Over the longer term, MEB’s competitor ports further north in Colombia (Santa Marta) and in the wider Caribbean region may experience increased disruption from more intense tropical cyclones. This could marginally improve MEB’s competitive advantage. 5. Further information on climate patterns and variability over Colombia 5.1 The Intertropical Convergence Zone The Intertropical Convergence Zone (ITCZ) is probably the major control factor over precipitation in Colombia: the positioning of the ITCZ around the equator contributes greatly to the perennial humidity of southern Colombia (Poveda et al, 2005). The ITCZ describes a region near the Equator where trade winds converge and is comprised of ascending air, low pressure, deep convective clouds and heavy precipitation. The regional placement of the ITCZ can be linked to the seasonal distribution of outgoing long-wave solar radiation: during the austral summer (DJF) the ITCZ is located at 5°N over Colombia, while during the boreal summer (JJA) the ITCZ travels northwards over Colombia (Poveda, Waylen and Pulwarty, 2006). The ITCZ shifts northwards during June and July and consequently southwards six months later from December-January, with maximum precipitation responding slowly to the seasonal migration of the sun (Paegle and Mo, 2002). The migration of the ITCZ is inseparable from atmospheric circulation features over the Caribbean Sea, easternmost Pacific Ocean and Amazon River basin. At the start of the dry season (towards the end of December), the ITCZ is stable at 8°N, leading to considerable cloudiness and dominance of the near surface winds by the easterly (trade) winds as a result of two high pressure systems (the North Atlantic Azores and North Colombian Caribbean). Central and west Colombia has a bimodal annual cycle of precipitation. There are two distinct rainy seasons from April to May and October to November and two ‘dry’ (or less rainy) seasons from December to February and June to August, explained by the double passage of the ITCZ. In contrast the north Caribbean coast of Colombia has a single rainy season from May to October reflecting the northernmost position of the ITCZ (Poveda et al, 2005). The region where MEB is located, over the Caribbean coast of Colombia, lies by the Magdalena, Sinu and Atrato Rivers. Its climate is modified by the geographical positioning of the ITCZ. This leads to a windy and dry December to April season and rainy August to October season, with the rest of the year being transitional. During the dry season, the ITCZ is located at its southernmost position (0- 5°S) and so trade winds dominate the Colombian climate with average speeds of 8m/s and diurnal 79 peaks of up to 15m/s. In contrast, during the rainy season the ITCZ moves over southwest Colombia decreasing wind speeds of the southerlies there and promoting rainfall rates higher than seen anywhere else in the western hemisphere (Restrepo and Lopez, 2008). In addition to the meridional migration of the ITCZ and its associated trade wind convergence, the climate of tropical South America is heavily controlled by a number of other factors (Poveda, Waylen and Pulwarty, 2006). These include sea-surface temperatures in the Pacific and Atlantic Oceans, the Amazon basin, the Andes mountains, and the dominant mode of interannual climate variability (namely the El Niño Southern Oscillation) which is responsible for many large climate anomalies over Colombia (Marin and Ramirez, 2006). The temporal distribution of rainfall in Colombia is controlled by the seasonal movement of the trade winds which meet at the ITCZ. This mechanism is the main driving force of moisture transfer over Colombia (Marin and Ramirez, 2006). The spatial distribution of precipitation is controlled by topographic factors such as the Andes Mountains and Amazon basin and also adjoining and remote sea-surface temperatures such as in the Pacific Ocean. All of the above factors points to Colombian precipitation that is highly variable in both time and space (Marin and Ramirez, 2006). 5.2 The Chorro del Occidente Colombian jet The southerly trade wind region over the equatorial tropical Pacific generally converges into the ITCZ mean position near the equator. However, near the South American coast the flow becomes westerly (except during February), due to the changing sign of the Coriolis force and the land-sea temperature gradient (Amador et al, 2006). It enters Colombia from the Pacific at 5°N.This westerly trade wind – known as the Chorro del Occidente Colombiano (CHOCO) jet – is a key feature of the Colombian climate and responsible for the high precipitation values in the Pacific lowlands. Wind speeds within the jet measure a maximum of 6 to 8 meters per second near 1km above the surface from October to November (Amador et al, 2006). The CHOCO jet is strongest between the months of September and November and practically non-existent from February to March. This strength in September through November offers one explanation as to why the concurrent rainy season is more intense than the April to May wet season over central and western Colombia. The strength of the CHOCO jet is strongly linked to the gradient of surface air temperature between western Colombia and the tropical Pacific near the date line and so has strong annual and interannual variations. The gradient is greatest during October to November and weakest in February to March, corresponding to the strength of the jet. The CHOCO jet has most of the characteristics of a low level jet including a maximum velocity at 900-1000hPa, an association with a strong land-ocean temperature gradient, lots of horizontal and vertical shear, a relation to strong convection, an association with strong moisture transport and a link to the topographic ‘gap’ in the western Andes between 5°N and 5°30’N that acts to converge the flow (Poveda and Mesa, 2000). Clusters of thunderstorms which last for tens of hours are responsible for delivering heavy moisture to the region, particularly during night time hours and over the continent. Between May and October, and particularly from August to October, tropical cyclones can also impact on the Colombian climate after forming in the Caribbean and East Pacific Oceans (www. cioh.org.co). 5.3 El Niño Southern Oscillation and interannual climate variability El Niño events are characterized by anomalously warm sea surface temperatures over the eastern and central equatorial Pacific, the deepening of the warm ocean water layer known as the thermocline in the eastern Pacific and a decrease in surface easterly trades. Negative values of the Southern Oscillation Index (defined as the pressure gradient between Darwin and Tahiti) are 80 associated with warm events, while positive values are associated with cool climate episodes known as la Niña (Poveda, Waylen and Pulwarty, 2006). Reduced precipitation during El Niño events can be explained by the decreasing sea surface temperature gradient in the eastern Pacific along with the strength of the CHOCO jet. This leads to decreased moisture transport, and negative Colombian precipitation anomalies. Additionally, changes to the onshore flow of moisture are associated with increased atmospheric pressure, especially between December and February. In turn this induces a displacement of thunderstorms within the ITCZ to southwest of the normal position and so over tropical South America (Poveda, Waylen and Pulwarty, 2006). ENSO’s impact on river discharge in Colombia happens progressively later as one moves to the east and the impacts of El Nino are more pronounced than the impacts of La Niña. ENSO anomalies lead to hydrological anomalies by one month in western Colombia, by 2 to 4 months in central Colombia and by up to 6 months in the east of the country. However, other factors also impact on Colombian hydrology. The Caribbean experiences positive sea surface temperatures during and after El Niño events, as does the tropical North Atlantic. The reasons for this are uncertain, although it could be due to a ‘land-atmosphere bridge’ transferring the signal to the tropical Atlantic and Caribbean. Dry conditions induced by El Niño events are self-sustaining through feedback mechanisms (Poveda and Mesa, 1997). There are also other modes of variation that impact consistently on precipitation over Colombia. A decrease in high frequency activity during July though October and February to April during El Niño years is associated with lower tropical easterly wave activity over the Caribbean, with the opposite occurring during La Niña years. For example, rainfall rates generally show increases during the westerly phase of the Madden Julian Oscillation (MJO) and a concurrent decrease during the easterly phase. During the westerly phase of the MJO tropical storm activity over the Caribbean increases which can disrupt atmospheric patterns and induce heavy rainfall events in northern South America (Poveda et al, 2005). The MJO occurs in a thirty to sixty day band (Poveda, Waylen and Pulwarty, 2006) and as such variability associated with this can occur on a seasonal basis. In addition to the expected patterns, it must be noted that ENSO anomalies produce a complex and non-linear response in Colombian climate meaning that even ENSO events of the same phase can produce highly variable anomalies (Marin and Ramirez, 2006). For example, the very strong 1982-3 El Niño event did not lead to the intense dry anomalies that may have been expected and conversely there was a prolonged dry spell from 1957-60 even through the associated 1957-8 ENSO event wasn’t especially strong (Poveda and Mesa, 1997). A further feature of Colombian climate is the presence of ‘paramos zones’. These can be defined as desolate regions found only between areas of high Andes forests and ‘permanent snow’ areas in the tropics. The climatic conditions of these zones are characterized by average temperatures inferior to 10°C, a strong diurnal temperature range, cloudy skies and foggy days, strong winds and light rains. A 2002 NCI report concluded that Colombian glaciers had lost almost 80% of their area since 1850 and 78% of the remaining glaciers were expected to be seriously affected by temperature increase in the coming four decades (Ruiz et al, 2008). Further, a study by Ceballos et al. (2006) concluded that global warming is threatening glacier ecosystems. Here the authors found that 50% of glacial areas were lost in the past fifty years. They concluded that this strong loss has continued in the last fifteen years, which has seen 10 to 50% of the remaining glacial areas melting. The primary control of such ice melt has been an increase of 1°C measured at high altitude Colombian weather stations over the past three decades (Ceballos et al, 2006). 81 References Aguilar, E., et al. 2005. “Changes in precipitation and temperature extremes in Central America and northern South America, 1961–2003”, J. Geophys. Res., 110. Pp1-15. Alexandre, N, Francois, T, Paul, D, Torres, R, Andrade, C. 2009. ^ĞĂůĞǀĞůsĂƌŝĂďŝůŝƚLJĨƌŽŵϭϵϱϬƚŽ 2000 and Hazards linked to Storm Surge Episodes in Bocagrande and Castillogrande Peninsulas, Cartagena de Indias, Colombia. CIOH Scientific Bulletin, No. 26, ISSN 0120-0542, 71-84. Amador et al. 2006. Atmospheric forcing of the eastern topical Pacific: a review, Progress in KĐĞĂŶŽŐƌĂƉŚLJ͕ϲϵ;Ϯ-ϰͿ͕ƉϭϬϭ-ϭϰϮ Andraded, C. A. 2008. “Recent changes in sea level. In: Deltas of Colombia, Morphodynamics and Vulnerability to Global Change”. JD Restrepo (eds). ISBN 978-958-720-020-1. Fondo Editorial Universidad EAFIT, Colciencias. 101-121pp. Asian Disaster Reduction Centre. 1999. ŽůŽŵďŝĂ,ƵƌƌŝĐĂŶĞϭϵϵϵͬϭϭĚĞƚĂŝůƐŽĨĚŝƐĂƐƚĞƌŝŶĨŽƌŵĂƚŝŽŶ, accessible at www.adrc.asia/view_disaster_en.php?NationCode=170&lang=en&KEY=92 (accessed 20/06/2010) Associated Press. 2005. “Beta Sweeps into Nicaragua”, 30th October 2005, article accessible at www.cbsnews.com/stories/2005/10/27/world/main984848.shtml (accessed 20/06/2010). BBC (British Broadcasting Corporation) 1999. “Hurricane Lenny threatens Puerto Rico”, November 17th 1999, news article accessible at news.bbc.co.uk/1/hi/world/americas/523905.stm (accessed 20/06/2010). Blanco et al. 2007. “ENSO and the rise and fall of the tilapia fishery in northern Colombia”. Fisheries ZĞƐĞĂƌĐŚ͕ϴϴ;ϭ-ϯͿ͕ƉϭϬϬ-ϭϬϴ Bouma et al. 1997. “Predicting high risk years for malaria in Colombia using parameters of ENSO”, dƌŽƉŝĐĂůDĞĚŝĐŝŶĞĂŶĚ/ŶƚĞƌŶĂƚŝŽŶĂů,ĞĂůƚŚ͕Ϯ;ϭϮͿ͕ƉϭϭϮϮ-ϭϭϮϳ Ceballos et al., 2006.” Fast shrinkage of tropical glaciers in Colombia”, Annals of Glaciology, 43 (-Ϳ͕ Ɖϭϵϰ-ϮϬϭ Christensen, J. H., Hewitson, B., Busuioc, A., Chen, A., Gao, X., Held, I., Jones, R., Kolli, R. K., Kwon, W.- T., Laprise, R., Rueda, V. M., Mearns, L., Menéndez, C. G., Räisänen, J., Rinke, A., Sarr, A. & Whetton, P. (2007) Regional Climate Projections. IN Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B., Tignor, M. & Miller, H. L. (Eds.) ůŝŵĂƚĞŚĂŶŐĞϮϬϬϳ͗dŚĞ Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom and New York, NY, USA, Cambridge University Press. Church, J., A., White, N., J., Aarup, T., Wilson, W., S., Woodworth, P., L., Domingues, C., M., Hunter, J., R., and Lambeck, K. 2008 “Understanding global sea levels: past, present and future”. Sustainability Science. 3:1. pp. 9-22. Grinsted, A; Moore, JC; Jevrejeva, S. (2010). Reconstructing sea level from paleo and projected temperatures 2000 to 2100 AD. Climate Dynamics. 34 4 461-472 82 Gutierrez and Dracup (2001) An analysis of the feasibility of long range streamflow forecasting for Colombia using ENSO indices, :ŽƵƌŶĂůŽĨ,LJĚƌŽůŽŐLJϮϰϲ;ϭ-ϰͿ͕Ɖϭϴϭ-ϭϵϲ Hulme, M, Jenkins, G. L., Lu, X., Turnpenny, J. R., Mitchell, T. D., Jones, R. G., Lowe, J., Murphy, J. M., Hassell, D., Boorman, P., McDonald, R. and Hills, S. (2002) Climate Change Scenarios for the hŶŝƚĞĚ<ŝŶŐĚŽŵ͗ƚŚĞh</WϬϮ^ĐŝĞŶƚŝĨŝĐZĞƉŽƌƚ͘Tyndall Centre for Climate Change Research. University of East Anglia. Norwich. 120pp. ICRC (International Committee of the Red Cross). 2005: “Minor Emergency – Colombia and Nicaragua: Hurricane Beta”, 31st October 2005, accessible at www.ifrc.org/docs/appeals/rpts05/BetaME1.pdf (accessed 20/06/2010). IDEAM (Instituto de Hidrología, Meteorología y Estudios Ambientales de Colombia). 2007. “Colombia, Venezuela bracing for Hurricane Felix”, 3rd September 2007, accessible at news.xinhuanet.com/english/2007-09/03/content_6653156.htm (accessed 20/06/2010). Knutson, McBride, Chan, Emanuel, Holland, Landsea, Held, Kossin, Srivastava, and Sugi, S. 2010. “Tropical cyclones and climate change”. EĂƚƵƌĞ'ĞŽƐĐŝĞŶĐĞϯ͗ϭϱϳ-ϭϲϯ͘ Legett, J., A., and Logan, J. 2008. Are carbon dioxide emissions rising more rapidly than expected? CRS Report for Congress: Washington, USA. Mantilla et al. 2009. “The role of ENSO in understanding changes in Colombia’s annual malaria burden by region 1960-2006”, DĂůĂƌŝĂ:ŽƵƌŶĂů͕ϴ;ϭͿ͕ĂƌƚŝĐůĞŶŽ͘ϲ Marin, S and Ramirez, J, A. 2006. “The response of precipitation and surface hydrology to tropical macro-climate forcing in Colombia”, ,LJĚƌŽůŽŐŝĐĂůWƌŽĐĞƐƐĞƐ͕ϮϬ;ϭϳͿ͕Ɖϯϳϱϵ-ϯϳϴϵ Meehl and Stocker (lead authors) .2007. Chapter 10, Global Climate Projections, /WZϰt'ϭ͕ Ɖϳϰϳ-ϴϰϲ Mitchell, T. P. and J. M. Wallace (1992) “The annual cycle in equatorial convection and sea surface temperatures”, :ŽƵƌŶĂůŽĨůŝŵĂƚĞ͕ϱ;ϭϬͿ͕ƉϭϭϰϬ-ϭϭϱϲ Mogil, H, M. 2007. džƚƌĞŵĞǁĞĂƚŚĞƌ͗ƵŶĚĞƌƐƚĂŶĚŝŶŐƚŚĞƐĐŝĞŶĐĞŽĨŚƵƌƌŝĐĂŶĞƐ͕ƚŽƌŶĂĚŽĞƐ͕ŚĞĂƚǁĂǀĞƐ͕ snow storms, global warming and other atmospheric disturbances. Blue Red Press. First Edition. 61pp. Nakicenovic and Swart (lead authors, 2007) Summary for Policy Makers, /WZϰt'ϭ͕Ɖϭ-ϭϵ NASA (National Aeronautics and Space Administration). 2009. “Always Something Brewing Year 'Round on NASA's Hurricane Web Page”. Map of the cumulative tracks of all tropical cyclones during the 1985-2005 time period. 02.17.09. Accessed from www.nasa.gov/mission_pages/hurricanes/features/hurricane_brew.html (23/06/10) NOAA (National Oceanic and Atmospheric Administration). 2007. “Hurricane Felix Intermediate Advisory Number 8”, 2rd September 2007, accessible at www.nhc.noaa.gov/archive/2007/al06/al062007.public.008.shtml (accessed 20/06/2010). Paegle, J, N and Mo, K, C. 2002 “Links between summer rainfall variability over South America and sea surface temperature anomalies”, :ŽƵƌŶĂůŽĨůŝŵĂƚĞ͕ϭϱ;ϭϮͿ͕Ɖϭϯϴϵ-ϭϰϬϳ 83 Parsons, J, J. 1982. “The North Andean Environment”, Mountain Research and Development, 2 (3), p253-262 Pfeffer, W.T., Harper, J.T. and O'Neel, S 2008. “Kinematic constraints on glacier contributions to 21st- century sea-level rise”. Science 321 5894 1340-1343 Poveda et al 2001. Coupling between annual and ENSO timescales in the malaria-climate association in Colombia, EŶǀŝƌŽŶŵĞŶƚĂů,ĞĂůƚŚWĞƌƐƉĞĐƚŝǀĞƐ͕ϭϬϵ;ϱͿ͕Ɖϰϴϵ-ϰϵϯ ———. 2001. “The diurnal cycle of precipitation in the tropical Andes of Colombia”, Monthly tĞĂƚŚĞƌZĞǀŝĞǁ͕ϭϯϯ;ϭͿ͕ƉϮϮϴ-240 Poveda and Mesa 2000. Feedbacks between hydrological processes in tropical South America and large scale ocean-atmosphere phenomenon, :ŽƵƌŶĂůŽĨůŝŵĂƚĞ͕ϭϬ;ϭϬͿ͕ƉϮϲϵϬ-ϮϳϬϮ Poveda, Waylen and Pulwarty 2006. Annual and interannual variation of the present climate in northern South America and South Mesoamerica, Palaeogeography, Palaeoclimatology and WĂůĂĞŽĞĐŽůŽŐLJ͕Ϯϯϰ;ϭͿ͕Ɖϯ-Ϯϳ Rahmstorf, S, Cazenave, A and Church, J.A, et al. 2007. Recent climate observations compared to projections. Science. 316 5825 709-709 Restrepo et al 2002. “Morphodynamics of a high discharge tropical delta, San Juan River, Pacific Coast of Colombia, DĂƌŝŶĞ'ĞŽůŽŐLJ͕ϭϵϮ;ϰͿ͕Ɖϯϱϱ-ϯϴϭ Restrepo and Lopez. 2008. “Morphodynamics of the Pacific and Caribbean deltas of Colombia, South America”, :ŽƵƌŶĂůŽĨ^ŽƵƚŚŵĞƌŝĐĂŶĂƌƚŚ^ĐŝĞŶĐĞƐ͕Ϯϱ;ϭͿ͕Ɖϭ-Ϯϭ Ruiz et al. 2008. “Changing climates and endangered high mountain ecosystems in Colombia”, Science of the Total EŶǀŝƌŽŶŵĞŶƚ͕ϯϵϴ;ϭ-ϯͿ͕ƉϭϮϮ-ϭϯϮ Sutherland et al 2008. “Monitoring sea level change in the Caribbean”, 'ĞŽŵĂƚŝĐĂ͕ϲϮ;ϰͿ͕ƉϰϮϴ-ϰϯϲ Torres R, Otero L, Afanador F Marriaga L. 2008. Sea Level Behaviour on Colombia’s Caribbean Coastline. Boletín Científico CIOH. 26: 8-21. ISSN 0120-0542 Torres R, Gómez J, Afanador F. 2006. Mean Sea Level Variation at the Colombian Caribbean. Boletín Científico CIOH. 24: 64-72. ISSN 0120-0542 Trapp, R. J., Diffenbaugh, N. S., Brooks, H. E., Baldwin, M E., Robinson, E. D., and Pal, J. S. 2007 Changes in severe thunderstorms environment frequency during the 21st century caused by anthropogenic enhanced global radiative forcing. Proceedings of the National Academy of Sciences. 104: 50. 19719-19723p. Trenberth and Jones (lead authors, 2007) Chapter 3, Observations - surface and atmospheric climate change, /WZϰt'ϭ͕ƉϮϯϱ-ϯϯϲ Vermeer, M and Rahmstorf, S. 2009. “Global sea level linked to global temperature”. Proceedings of the National Academy of Sciences in the United States of America. 106: 51 21527-21532 84 Vernekar et al .2003. “Low level jets and their effects on South American Summer Climate as simulated by the NCEP Eta model”, :ŽƵƌŶĂůŽĨůŝŵĂƚĞ͕ϭϲ;ϮͿ͕ƉϮϵϳ-ϯϭϭ Villar et al .2009. “Spatio-temporal rainfall variability in the Amazon Basin countries (Brazil, Peru, Bolivia, Colombia and Ecuador)”, International Journal oĨůŝŵĂƚŽůŽŐLJ͕Ϯϵ;ϭϭ͕ͿƉϭϱϳϰ-ϭϱϵϰ 85 Climate risk case study: Terminal Marítimo Muelles El Bosque Appendix 2: PRECIS projections for climate change over Cartagena 1. Introduction Information on the changes in climate variables such as mean values of air temperature, relative humidity, precipitation and wind, is important for understanding climate change impacts that could affect MEB. This report presents a brief overview of observed climatic conditions in Cartagena, along with scenarios of future changes in the city’s climate. The future scenarios are based on observed trends and on projections of the PRECIS regional climate model (RCM). PRECIS is also being used by IDEAM in developing the Colombian government’s national assessments of climate change impacts. Using it in the context of the current study for MEB should facilitate consideration of the results of this study in the context of work underway nationally 1. 2. Physiographical features of the region Cartagena City is located on the northern coast of Colombia at 10°25' North, 75°32' West (10.41667, -75.5333) in Cartagena Bay (Caribbean Sea). A vast extension of land, swamps, and lagoons extend to the south east of the city. To the North West is the Caribbean Sea. Figure 2-1 – Location of Cartagena (Colombia) on Caribbean coast. 1 This appendix was written by Professor Jose Daniel Pabon, Department of Geography, National University of Colombia, who also produced the PRECIS projections for the national government. 87 Mountainous terrain is located approximately 150 kilometers to the east (the high mountain of Sierra Nevada de Santa Martha with elevation more than 5000 meters above sea level) and 60 kilometers to the South (the Serranía de San Jacinto with an average elevation of between 1000 and 2000 meters above sea level). The Sierra Nevada de Santa Marta blocks the flow of trade winds (dominantly easterlies). As a result, on the western side of the Sierra, where Cartagena is located, wind flow is disturbed and the direction of the winds is mainly N and NE. The climate is tropical with a mean annual air temperature between 26.5-28°C and a low annual amplitude (Figure 2-2). The minimum temperature recorded is 18°C and the maximum is 40°C. The average value of monthly maximum temperature is 33.6°C. Figure 2-2 – Annual cycle of the mean air temperature (°C) in Cartagena (at the airport Rafael Nuñez meteorological station for the period 1961-1990) 28,5 MONTHLY AVERAGE TEMPERATURE, °C 28 27,5 27 26,5 26 25,5 1 2 3 4 5 6 7 8 9 10 11 12 MONTHS The relative humidity in Cartagena oscillates between 78 and 83 % (see Figure 2-3). Figure 2-3 – Annual cycle of relative humidity (%) in Cartagena (at the airport Rafael Nuñez meteorological station for the period 1961-1990) 83 RELATIVE HUMIDITY, % 82 81 80 79 78 77 76 75 1 2 3 4 5 6 7 8 9 10 11 12 MONTHS Precipitation is approximately 1000 millimeters/year distributed in an annual cycle showed in Figure 2-4. 88 Figure 2-4 – Distribution of precipitation (millimeters) through the year in Cartagena 300 MONTHLY PRECIPITATION (mms) 250 200 150 100 50 0 1 2 3 4 5 6 7 8 9 10 11 12 MONTHS Winds in the region are mainly N and NE, as shown in Figure 2-5, however with a low frequency other directions are observed. The dominant wind speed is between 3.5 and 8 m/s. Figure 2-5 – The wind rose for Cartagena (using data for 19 years in the period 1961-1990). It is necessary to point out that due to its location the city experiences daily oscillations of climatic variables, specifically caused by breezes from coastal zones. 3. Data and methods Observational data of air temperature, relative humidity, precipitation and winds is available for the period 1940-2008. In spite of this, the period 1960-2008 was used, as data before 1960 presented some inconsistencies (see Figure 3-1, where time sequences of annual data are shown). 89 Figure 3-1 – Time series of annual values of air temperature (top), relative humidity (middle), and precipitation (bottom) for Cartagena Airport (period: 1941-2008). 29 28,8 ANNUAL MEAN TEMPERATURE, °C 28,6 28,4 28,2 28 27,8 27,6 27,4 27,2 27 41 44 47 50 53 56 59 62 65 68 71 74 77 80 83 86 89 92 95 98 01 04 07 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 YEARS 85 RELATIVE HUMIDITY, % 84 83 82 81 80 79 78 1941 1944 1947 1950 1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 YEARS 1800 1600 ANNUAL PRECIPITATION, mms 1400 1200 1000 800 600 400 200 0 41 44 47 50 53 56 59 62 65 68 71 74 77 80 83 86 89 92 95 98 01 04 07 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 YEARS To describe regional climate (presented above) multiannual averages for the 1961-1990 period were used. This period serves as a reference to compare to future conditions and calculate changes in mentioned variables. To estimate long term trends, the linear trend was calculated for the period 1961-2008. To estimate the future values of climatic variables that were used in this study, an extrapolation of long term trends and data of scenarios from simulations made with PRECIS for Colombia in A2 and B2 IPCC scenarios, using the outputs of the HadCM3 general circulation model as boundary conditions. To have an idea of the approximation of the PRECIS simulations to the real climate of Cartagena sector, a comparison between observed 1961-1990 pattern and that generated for the model for the same period was made. 90 PRECIS is a regional climate model, developed by the Hadley Center of the United Kingdom Meteorological Office (UK Met Office), that allows dynamic downscaling of global model outputs to produce for a given region climate data in a 25 x25 kilometers spatial resolution (a detailed description of PRECIS may be obtained from Jones et al, 2004; technical details are presented by Wilson et al., 2008 which can be downloaded from http://precis.metoffice.com/user_support.html). The baseline climate (1961-1990) was generated by PRECIS using ERA40 data set (see Uppala et al., 2005). To assess the approximation of the PRECIS simulations to the real climate of Cartagena sector a comparison between the observed 1961-1990 pattern and that generated for the model for the same period was made. Following this, the HadCM3 model for global A2 and B2 climate scenarios was used as the boundary conditions for PRECIS generating scenario outputs at a regional 25x25 kilometers scale. The changes in climate variables were calculated as the difference between average values for the period 2070-2100 (scenarios A2 and B2) and the average of the reference climate (1961- 1990). 4. Recent trends in air temperature, relative humidity, and precipitation Using the time sequences of annual values of air temperature, relative humidity and precipitation for the 1961-2008 period the following long term (almost 50 years) trends were found: x Air temperature: +0,11°C/decade x Relative humidity +0,65%/decade x Precipitation: +20 mm/decade or an increasing of approx. 2% of annual amounts by decade. 5. The future climate (2011-2040 and 2070-2100 periods) Based on an extrapolation of long term trends, the values for future decades could be as shown in Table 5-1. Table 5-1 – Estimation of decadal values based on extrapolation of actual trends Decade Temperature (oC) RH (%) Precipitation (mm) 1961-1970 27.5 80.3 1030.7 1971-1980 27.7 81.0 1050.7 1981-1990 27.8 81.6 1070.7 1991-2000 27.9 82.3 1090.7 2001-2010 28.0 82.9 1110.7 2011-2020 28.1 83.6 1130.7 2021-2030 28.2 84.2 1150.7 2031-2040 28.3 84.9 1170.7 2041-2050 28.5 85.5 1190.7 2051-2060 28.6 86.2 1210.7 2061-2070 28.7 86.8 1230.7 2071-2080 28.8 87.5 1250.7 2081-2090 28.9 88.1 1270.7 2091-2100 29.0 88.8 1290.7 91 Data from scenarios A2 and B2 for different decades produced by PRECIS are presented in Table 5-2 (the coordinates of the grid point close to the port is colored in yellow). Comparing the results of both the extrapolation and simulations obtained for air temperature a warming between 2-2.5°C appears to be most likely. Data obtained for the B2 scenario could be used. This means that at the end of the century with annual mean temperature of 30°C maximum temperatures above 33.5°C will be observed more frequently and the absolute maximum could rise up to 42.5°C. In the decades between 2010 and 2040 the actual absolute maximum (40°C) will became more frequent. Relative humidity is increasing in simulations at a lower rate than observed trends. A rate between 0.55 and 0.6% by decade could be applied to actual decadal means in order to estimate the values for future periods. For precipitation the results of the simulation with PRECIS contradict the observed trends. In PRECIS, precipitation for the region will decrease, while actual trends show an increase in the region where Cartagena is located. In this case the use of the observed trends is recommended. Figure 5-1 presents the increase in wind as constructed using data from PRECIS simulations. The increase that reproduces the 1961-1990 period differs from that observed in the same period (see Figure 2-5). The simulated increase is 45° to the east of the actual wind direction and shows a high frequency of north-east, not of north winds as observed; other directions close to NE also are shown. Wind speed is frequently between 3 to 10 m/s. However, the changes between the observed period and A2 and B2 scenarios for 2070-2100 could be noted. For example, an increase in the frequency of NE winds and reduction of other directions is expected. Also, it may be seen that the frequency of the range of wind speed 3-10 m/s increases in A2 and B2 scenarios compared to the 1961-1990 period. Adjusting these increases to the observed trends, it can be concluded that in Cartagena the frequency of northerlies will increase, with speeds in the range of 3-10 m/s becoming increasingly frequent. 92 Table 5-2 – Decadal and normal values of climate variables as produced by PRECIS for reference period (1961-1900) and for decades from 1971 to 2040, as well as for the 30-year period 2070-2100 (under A2 and B2 IPCC SRES scenarios). 93 Figure 5-1 – Wind rose (speed and direction) simulated by PRECIS for 1961-1990 period (left) and for 2070-2100 in A2 (center) and B2 (right) IPCC 94 References Jones, R.G., Noguer, M., Hassell, D.C., Hudson, D., Wilson, S.S., Jenkins, G.J. and Mitchell, J.F.B. 2004. Generating high resolution climate change scenarios using PRECIS, Met Office Hadley Centre, Exeter, UK, 40pp UK Met Office. 2010. United Kingdom Meteorological Office. PRECIS Regional Climate Model Outputs. http://precis.metoffice.com/user_support.html (accessed 23/05/2010). Uppala S.M., Kållberg P.W., Simmons A.J., Andrae U., da Costa Bechtold V., Fiorino M., Gibson J.K., Haseler J., Hernandez A., Kelly G.A., Li X., Onogi K., Saarinen S., Sokka N., Allan R.P., Andersson E., Arpe K., Balmaseda M.A., Beljaars A.C.M., van de Berg L., Bidlot J., Bormann N., Caires S., Chevallier F., Dethof A., Dragosavac M., Fisher M., Fuentes M., Hagemann S., Hólm E., Hoskins B.J., Isaksen L., Janssen P.A.E.M., Jenne R., McNally A.P., Mahfouf J.-F., Morcrette J.-J., Rayner N.A., Saunders R.W., Simon P., Sterl A., Trenberth K.E., Untch A., Vasiljevic D., Viterbo P., Woollen J. 2005. “ The ERA-40 re- analysis”. Quart. J. R. Meteorol. Soc., 131, 2961-3012. Wilson S., Hassell D., Hein D., Jones R., Taylor R., 2008: Installing and using the Hadley Centre regional climate modeling system, PRECIS. Met Office, 155Pp. 95 Climate risk case study: Terminal Marítimo Muelles El Bosque Appendix 3: Supplementary information to Section 5 ‘Vehicle Movements Inside the Port’: Sea Level Variability 1. Introduction Section 5 of the main report analyses the potential climate change risks to vehicle movements at MEB as a result of seawater flooding on the port and increased damage to unpaved areas of the port. This appendix presents the detail of some of the analysis underpinning Section 5, on the various components of sea level rise and flood risk. 2. Definition of a datum for the flood risk assessment This flood risk assessment used a variety of information sources, including sea level and topographic data. However, these sources do not contain a consistent datum. For this assessment, elevation has been set to port datum, which is the datum used in the port plans. The predicted tidal series is to chart datum (CD), which is approximately mean low water on a spring tide 1. The chart datum is approximately 0.1m below mean sea level (MSL) in the predicted tidal series. All other factors contributing to sea level are added to the tidal signal. The elevations at the port have been taken from MEB’s port plans 2. The elevations in the port plans are quoted to MSL. However, MSL is not a fixed datum in space and time, as sea level changes. In order to calibrate MSL to real sea level, observations of flooding events which have already occurred have been used to adjust the elevations in the port plans to the measured and predicted sea surface elevations. The causeway between the island and the mainland sites of the port was flooded on a number of occasions during September and October 2008 3. Although storm surge and precipitation may have played a role in these flooding events, it is assumed that high waters must have been close to the level of the causeway during these events. 3. Harmonic analysis on astronomical tides in Cartagena Sea level variations due to astronomical tides can be accurately predicted using data available from the UK Hydrographic Office (UKHO), which keeps a database of tidal sequences globally. To predict the tides at a specific location, such as Cartagena, a process of harmonic analysis is used. This process involves extracting sine waves of known amplitudes, frequencies and phase from the tidal time series. From this analysis it is possible to reconstruct the tide for any point in time. The longer the record used for analysis, the more harmonic constituents that can be extracted and the more accurate the predictions. CIOH used a similar technique to assess the non-tidal components of sea level in Cartagena. Using harmonic analysis the astronomical tidal and mean sea level components were extracted from the measured sea level time series. The residual components of sea level obtained through this process are shown in Figure 3-1. Although termed ‘storm surge’, freshwater input from the Canal del Dique into the bay was considered to be a potential cause of the residual variability in sea level between 1951 and 2000. From a visual inspection of Figure 3-1, it is clear that the annual sea level variations that this assessment uncovered are included within the CIOH residual sea level time series. 1 The spring neap cycle is 14 days and is due to the gravitational pull of the sun and the moon coming in and out of phase. 2 Provided by Andres Burgos during the MEB site visit 3 Email from Andres Burgos of 17 February 2010 97 Figure 3-1 – Residual sea level time series between 1951 and 2000 (Alexandre et al. 2008). Note: a version of this graph with English axis titles is provided in the main report. 4. Analysis of annual variations in sea level due to meteorological factors Water run-off from the Canal de Dique discharges to the sea at four locations, as shown in Figure 4-1. It is estimated that 24% of water run-off from the canal enters the Bay of Cartagena, 40% discharges to the south of Baru and 36% is lost through abstraction and to groundwater. The Canal del Dique annual mean water discharge at the Santa Helena station is 397 m3/s. Discharges as high as 800m3/s often occur during November (Restrepo et al., 2005). It was not possible to obtain water run-off data from the Santa Helena hydrological station, but data was obtained from the Calamar station located upstream of Santa Helena on the canal. Because the estimated run-off at Santa Helena is 79% of the run-off recorded at Calamar on average (see Figure 4-1), the estimated mean and peak river flows into Cartagena Bay are approximately 120m3/s and 240m3/s respectively. 98 Figure 4-1. Analysis of water run-off in the Canal del Dique. Source: Ordóñez et al., 2001. A comparison between monthly mean sea level and flow in the Magdalena river for a period of 15 years from 1985 to 2000 suggests that the seasonal variation is similar in the two datasets but that the inter-annual variability differs (see Figure 4-2) 4. 4 Note that if data was available, it would be preferable to compare monthly mean sea level against river flow in the Canal del Dique. 99 Figure 4-2 – Comparison of mean monthly flow in the Magdalena and sea level in Cartagena Bay from 1985 to 2000 (normalized about the mean) 2 Magdalena Mean Flow Cartagena Bay Sea-level 1.5 Normalised Flow & Sea-level 1 0.5 0 -0.5 0 20 40 60 80 100 120 140 160 180 200 Months To evaluate the role of rainfall through water run-off from the Canal del Dique, a hydrodynamic model is used. The model is run using the modelled tidal signal (astronomical tide plus annual variation) and a range of hypothetical water flows (from no flow to 2000m3/s). The model is based on charted bathymetry and is not tested against measured data (calibrated and validated). Water flows through Bocachica and Bocagrande are fundamental to this assessment and the modelled flows used in this analysis have not been compared to measured data, beyond a simple comparison of peak current velocities reported on the admiralty charts. The results of the model are presented in Figure 4-3 and Figure 4-4. These show no increase in sea level inside Cartagena Bay during times of peak flow (~200m3/s); at 800m3/s there is only a 10mm difference in sea level (Figure 4-3). The difference is 50 mm with a flow of 2000m3/s, but such a flow does not occur in the canal (Figure 4-4). 100 Figure 4-3 – Surface elevation (sea level) in and outside of the Bay of Cartagena, for a flow from Canal del Dique of 800m3/s. 0.45 0.4 0.35 0.3 Sea-level (m above CD) 0.25 Offshore MEB 0.2 0.15 0.1 0.05 0 04/10/2009 06/10/2009 08/10/2009 10/10/2009 12/10/2009 14/10/2009 16/10/2009 18/10/2009 Figure 4-4 – Surface elevation (sea level) in and outside the Bay of Cartagena, flow from Canal del Dique of 2000m3/s. 0.45 0.4 0.35 0.3 Sea-level (m above CD) 0.25 Offshore MEB 0.2 0.15 0.1 0.05 0 04/10/2009 06/10/2009 08/10/2009 10/10/2009 12/10/2009 14/10/2009 16/10/2009 18/10/2009 101 References Alexandre, N., Francois, T., Paul, D., Torres, R, and Andrade, C. 2008. Sea level Variability from 1950 to 2000 and Hazards linked to Storm Surge Episodes in Bocagrande and Castillogrande Peninsulas, Cartagena de Indias, Colombia. CIOH Scientific Bulletin, No. 26, ISSN 0120-0542, 71-84. Ordóñez, J., Cubillos Peña, C. and Medina Bello, E. 2009. Alternativas para el control sedimentológico del canal del Dique y sus efectos sobre el balance ecológico de la región. Universidad Nacional de Colombia. Sede Bogotá. Laboratorio de Ensayos Hidráulicos. Restrepo, J., Zapata, P., Díaz, J., Garzón-Ferreira, J, and García, C. 2005. “Fluvial fluxes into the Caribbean Sea and their impact on coastal ecosystems: the Magdalena River, Colombia”. Global and Planetary Change 50 (1-2). Pp 33-49. 102 Climate risk case study: Terminal Marítimo Muelles El Bosque Appendix 4: Supplementary information to Section 6 ‘Demand, Trade Levels and Patterns’ 1. Introduction Section 6 of the main report demonstrates how climate change impacts could affect imports and exports at MEB by changing the patterns of global trade, grain imports or agricultural exports. In analyzing climate change implications for global trade, and consequently trade at MEB, projections from the Stern Review on the Economics of Climate Change (‘Stern Review’) of climate change effects on global GDP were used. More information on the methodology and assumptions of Stern Review and on the projections used in this study is provided below. Supplementary information on the implications of climate change for food prices and demand for grain is also provided here below, as part of the analysis of climate change risks to grain imports presented in Section 6. 2. Further information on the methodology, assumptions and results of the ‘Stern Review’ The Stern Review adopts various techniques in its analysis, a key one of which is the use of economic models. Integrated assessment models (IAMs) estimate the economic impacts of climate change, while macro-economic models represent the costs and effects of a transition to an economy-wide low carbon energy system. Reviewing Stern’s economic models enables an analysis of how costs of climate change may change incrementally over time. While the Stern Review headlines report losses in per capita consumption of 5% to 20% now and forever under business as usual scenarios, these figures are aggregated over time (i.e. from 2006 to 2200 and even beyond). To estimate climate change impact costs over the next century, one should turn directly to Stern’s modelled outputs. As shown in Figure 2-1, projected reductions in per capita GDP up to 2100 are significantly lower than the headline figures. Figure 2-1 – Stern Review (2006) projections of the economic impacts of climate change. Grey shading represents 5-95% impacts range for each scenario 1. [Results are given as percentage change in global per-capita consumption]. Source: Stern Review, 2006a, p157 1 Dark grey shading shows the impact range for the Stern baseline climate scenario, medium grey shows the impact range for the high climate + market impacts + risk of catastrophe scenario, and light grey shows the impact range for the high climate + market impacts + risk of catastrophe + non-market impacts scenario. 104 Figure 2-1 illustrates three Stern ‘with climate change’ cases. These include: (i) the ‘baseline climate + market impacts + risk of catastrophe’ case, (ii) the ‘with high climate + market impacts + risk of catastrophe case’ and (iii) the ‘with high climate + market impacts + risk of catastrophe + non-market impacts case’. The baseline climate scenario derives from assumptions presented in the Intergovernmental Panel on Climate Change Third Assessment Report (IPCC TAR) 2, while the high climate scenario is designed to explore impacts associated with the action of amplifying climate system feedbacks (which increase temperature change) 3. In terms of economic impact, both market and non-market impacts are addressed. The ‘risk of catastrophe’ refers to accounting for the impacts of gradual climate change as well as the risks of catastrophic climate impacts at higher temperatures. Table 2-2 summarises the results shown in Figure 2-1 above, targeting information relevant to this study, i.e. changes in per capita consumption (comparable to global GDP), for the 2020s and 2050s. Table 2-2 also includes one extra Stern ‘with climate change’ case, referred to as ‘high climate + market impacts + risk of catastrophe + non-market impacts + value judgements for regional distribution’. This is Stern’s most extreme ‘with climate change’ case which includes the value judgements in regional distribution which account for the disproportionate burden of climate change falling on the poor 4. It should be noted that all losses shown in Table 2-1 are benchmarked against projected GDP growth in Stern’s ‘baseline case’ - a world without climate change 5. 2 This scenario “produces a mean warming of 3.9°C relative to pre-industrial in 2100 and a 90% confidence interval of 2.4- 5.8°C for the A2 emissions scenario...This is in line with the mean projection of 4.1°C given by the IPCC TAR. The IPCC does not give a probability range of temperatures. It does quote a range across several models of 3.0-5.3°C” (Stern Review, 2006, p154). 3 The latter scenario, which represents a more extreme climate future, includes two types of amplifying feedback, a weakening of natural carbon absorption and increased natural methane releases from, for example, thawing permafrost. The high climate scenario, by incorporating two feedbacks, pushes “the mean temperature change up by around 0.4°C and give(s) a high probability of larger temperature increases. Accordingly, the 90% confidence interval increases to 2.6 – 6.5°C” (Stern Review, 2006, p154). 4 The Stern Review (2006) did not calculate the ‘Balanced Growth Equivalent’ cost of climate change after including value judgements for regional distribution (as it did for other issues), due to limited time. However, Nordhaus and Boyer (2000) estimate that, when giving more weight to impacts in poor regions, the global cost of climate change increases from 6% to 8% of GDP for 5 ńġŸŢųŮŪůŨĭġŪįŦįġŰůŦġŲŶŢųŵŦųġũŪŨũŦųįġŊůġŰųťŦųġŵŰġŤŢűŵŶųŦġŵũŦġŸŰųŴŵġŤŢŴŦġŤŭŪŮŢŵŦġŤũŢůŨŦġŴŤŦůŢųŪŰĭġŔŵŦųůġ applied this 25% extra cost. 5 Note: “Income in the ‘no climate change’ scenario is conventionally measured in terms of GDP – the value of economic output. The difficulty is that some of the negative effects of climate change will actually lead to increases in expenditure, which increase economic output. Examples are increasing expenditure on air conditioning and flood defences. But it is correct to subtract these from GDP in the ‘no climate change’ scenario, because such expenditures are a cost of climate change. As a result, the measure of the monetary cost of climate change that we derive is really a measure of income loss, rather than output loss as conventionally measured by GDP” (Stern Review, 2006, p145). 105 Table 2-1 – Stern Review (2006) projections of the economic impacts of climate change for 2025 and 2055. [Results are given as percentage change in global per-capita consumption and are intended to resemble the 2020s and 2050s]. Scenario Losses in per capita GDP per annum (%) (2dp) Climate Economic Year Mean 5th percentile 6 95th percentile Baseline Market impacts + risk 2025 -0.20 NA* NA climate of catastrophe 2055 -0.25 NA NA High climate Market impacts + risk 2025 -0.25 NA NA of catastrophe 2055 -0.50 NA NA Market impacts + risk 2025 -0.50 NA NA of catastrophe + non- market impacts 2055 -1.00 -0.25 -1.75 Market impacts + risk 2025 -0.63 NA NA of catastrophe + non- market impacts + 2055 -1.25 NA NA value judgements for regional distribution† * NA = not available †This analysis has followed Stern’s logic by raising the ‘High climate + market impacts + risk of catastrophe + non-market impacts’ scenario 25% to estimate the for the most extreme climate change scenario, including value judgements for regional distribution. See footnote 6 for more information. In addition to Stern’s ‘with climate change’ projections (presented in Figure 2-1 and Table 2-1), Table 2-2 presents modelled losses in global GDP associated with Stern’s ‘with mitigation and adaptation’ case over time. This is directly comparable with changes in per capita consumption presented in Table 2-1. The stabilisation target Stern uses for his mitigation calculations is 500-550ppm CO2e (parts per million carbon dioxide equivalent). Also, Stern implies that there are adaptation and mitigation costs up to 2050, but only mitigation costs from 2050 onward (as all significant impacts are assumed to be avoided by this point). While Stern acknowledges that a stabilisation point of 500- 550ppm CO2e does not avoid all risks from climate change (Stern, 2006b), for the purposes of assessing cost, the Review has not attempted to quantify the cost of climate change impacts from now onward or the costs of adapting to them after 2050 7. 6 Some 5th and 95th percentile data are ‘not available’ (NA) for the baseline climate and first two high climate scenarios as data points are very similar to mean values and thus cannot be read from Figure 2-1. . Some 5th and 95th percentile data are also ‘not available’ (NA) for the final high climate scenario as these points are also derived from (i.e. uplifted by 25% from the high climate + market impacts + risk of catastrophe + non-market impact scenario), in which some datapoints cannot be read. 7 The Stern Review estimates costs of stabilisation at 500-550ppm CO2e as “stablisation at 450ppm CO2e is already almost out of reach, given that we are likely to reach this level within ten years and that there are real difficulties of making the sharp reductions required with current and foreseeable technologies” (Stern Review, Executive Summary, 2006, xv). Climate sensitivity estimates from the IPCC 2001 report however indicate that stabilisation at 550ppm CO2e may lead to global temperature change of approximately 1.5°C-4.7°C (based on the 5-95% range). See Stern Review Exectuve summary for further information on the range of impacts expected at different levels of warming. 106 Table 2-2– Stern Review (2006a) projections of the economic impacts of mitigation and adaptation action 8. [Results are given as percentage change in global GDP and are intended to resemble the 2020s and 2050s]. Scenario Losses in global GDP per annum (2dp) Year Mean Mitigation & Adaptation 2025 -0.05 to -0.50% GDP (adaptation costs of making new infrastructure and buildings Adaptation to climate change impacts resilient to climate change in OECD before 2050 (‘before mitigation countries). measures can have effect’). Note that there is limited quantitative information on this cost 2055 -1.00% of GDP from 2050 (range of -5% to +2%, depending on scale of mitigation plus required, pace of technological innovation, efficiency at which policy is Mitigation to stabilise at 500-550ppm applied globally). CO2e. Assumption that mitigation leads to no climate change impact costs or adaptation costs. Thus, as shown in Table 2-1 and Table 2-2 the mean costs of climate change range from 0.20% to 1.25% of global per capita consumption in the ‘with climate change’ cases and 0.05% to 1.00% of global GDP in the ‘with mitigation and adaption’ cases. It is important to note, however, that the Stern Review was criticised by some economists for its choice of discount rate (1.4%) to set against future consumption. Stern used a ‘social discount rate’ rather than one based on observations consistent with current financial market predictions. This has led to criticism that the study overstates the costs to the future, despite the effects of uncertainty, better technology and decreasing marginal utility, which would lead some to conclude that future consumption should be considered significantly less valuable than consumption today (Nordhaus, 2007) 9. Of course it is worth noting that the discount rate that MEB applies to its investment decisions at the port is much higher, at 16%. 8 This is not a formal Stern scenario (as the climate impact scenarios are), however are issues that Stern has quantitatively analysed. 9 Nordhaus, though agreeing with Stern on the need for urgent climate action, suggests a discount rate of 3% - more than double the rate used in the Stern Review. 107 3. Supplementary analysis on the impact of climate change on grain imports at MEB 3.1 General, non-climatic factors affecting global grain imports While analysis in the main report focuses on climate change impacts on grain imports specifically, it is worth briefly examining other non-climate related factors that affect the grain trade. This is due to the complex nature of international commodity markets and trading, in which a range of factors interrelate. Below is a list of these key factors with some preliminary information on how they operate: x State of the global economy – The state of the global economy influences the amount of total trade in the world, and thus imports of grain. x Comparative Advantage – The global trading position that any one country holds at any one time will affect the price elasticity of demand for grain. An improvement in a country’s terms of trade, a net comparative advantage in other sectors of the import/export market, and low domestic interest rates would also be expected to affect the interplay between price and demand for agricultural commodities. x Global grain supply/production – Increases/decreases in production efficiency, or changing weather patterns, can increase/decrease grain yields, and in turn lead to variations in price. This is relevant as trade in agricultural goods is highly price sensitive. x Global grain supply available for food and feed – Increased demand for biofuels may lead to reduced grain available to import for human consumption and livestock feed. Fischer (2009) projects cereal prices to increase substantially in the future as a result of growing biofuel demand. (Note that the extent of price increase depends on the biofuel scenario selected). x Local grain production – If a country is able to produce its own grain, fewer imports are required (though importing may be cheaper). As well as land suitability, a country’s grain production is strongly linked to agricultural policy. For example, currently Colombia imports 70% of its maize. However, during the recent recession, the Ministry of Agriculture announced incentives for farmers to increase production of basic grain products (USDA FAS, 2008, USDA FAS, 2009). While this did not lead to an immediate reduction in grain imports, events such as this may lead to changes in import requirements in the future. x Cost of grain production – According to the UN FAO (2009), one of the key factors affecting agricultural commodity price is crude oil price. If oil prices increase (as they are projected to, rebounding to US$100 by 2020 and US$115 by 2030 – IEA, 2009 10), this is likely to translate into higher grain prices (via higher chemical, fertiliser and transportation costs). x Global grain demand – As noted by Fischer, in the long term, increased demand for agricultural products “is largely driven by population and economic growth, both foremost in developing countries” (Fischer, 2009, p3). Demand for cereal is also affected by whether it is being used as a staple food crop or for animal feed, and in turn the meat industry. Agrimonde (2009) relates an increase in meat eating with growing income. Global population and GDP is projected to increase up to 2050 (Fischer, 2009), in turn implying greater demand for grain (both for livestock and direct human consumption). A growing number of dairy diets will also lead to increased demand for grain (Gregory & Ingram, 2008). x Local grain demand – Related to the state of the importing economy, increased population and wealth in the importing country is likely to lead to increased demand and potentially increased grain imports. 10 These are in year-2008 dollars. 108 x Grain storage – Availability of crop stocks affects price, and in turn imports (Gregory & Ingram, 2008). x Tariffs and trade agreements - High import tariffs can reduce the level of imports from one particular country to another, while favourable trade agreements can facilitate trade. The US-Colombian Trade Promotion Agreement (CPTA), which is awaiting Senate approval, is an example of a favourable trade agreement which provides for a tariff-rate quota of 2.1 million tons for US yellow maize and 136,500 tons for US white maize to enter Colombia duty free (USDA FAS, 2009). x Additional factors include financial speculation and export restriction, which can affect commodity price (Gregory & Ingram, 2008), which in turn affects buying power and ability to import. It should be noted that only climate change impacts on grain imports have been assessed here, i.e. this analysis does not investigate the effects on grain imports of climate change mitigation policies. Such policies could have significant effects on grain imports, for instance if they affect oil prices or competition for land use between food crops and biofuels. 3.2 Impact of climate change on agricultural yields and demand for grain Figure 3-1 illustrates Muller et al.’s (2009) projections for climate change impacts on agricultural yields in 2050, while Figure 3-2 and Figure 3-3 show Cline’s (2007) projections for climate change impacts on agricultural productivity in 2080 (with and without carbon fertilization effects, respectively 11). Each figure shows a general global trend of decreasing productivity, though Figure 3-2 and Figure 3-3 also illustrate increasing productivity in parts of North America (the source of MEB’s grain), Europe, Northern and Eastern Asia and China. Figure 3-1 – Map showing projected climate change impacts on agricultural yields for 2050, given current agricultural practices and crop varieties (Source: Muller et al, 2009). 11 Carbon fertilization refers to the effect of additional amounts of carbon dioxide in the atmosphere on increasing plant growth (IFPRI, 2010). 109 Figure 3-2– Map showing projected climate change impact on agricultural productivity for 2080, with carbon fertilization. Results are shown in percentage change (Source: Cline, 2007). Figure 3-3 – Map showing projected climate change impact on agricultural productivity for 2080, without carbon fertilization. Results are shown in percentage change. (Source: Cline, 2007). It is worth noting that while Figure 3-1, to Figure 3-3 project climate change impacts on agricultural productivity, they do not incorporate adaptation action by farmers. This could mean that impacts may not be as extreme as illustrated. Increased investment in agricultural research (e.g. into drought resistant crops) and in developing suitable land for agriculture may lead to more land being productive than expected (UN FAO, 2009). Nevertheless, Fischer (2009), in his projections of cereal production (which did incorporate adaptation and price projections), also projects an overall decline in 2020, 2050 and 2080 at the global scale 12. 12 Note that Fischer (2009) provides 3 projections for each timestep (the Hadley A2 climate model, the CSIRO A2 climate model and the Hadley A2 model without CO2 fertilisation). All projections, apart from Hadley A2 for 2020, illustrate a decline in world cereal production relative to the reference scenario without climate change. 110 Climate change impacts on grain production in turn are likely to affect global grain price. Figure 3-4 illustrates the changing monthly FAO price index for cereals (in green), showing recent increases associated with the 2007-2008 food crisis. Various reasons have been put forward for the crisis, including financial speculation, the falling value of the US dollar, increased demand for grains (related to growing global wealth and population), increased demand for feed for livestock and dairy diets, export bans on selected foodstuffs, inadequate grain stocks, higher oil prices, the use of crop lands for production of biofuels and poor harvests. However, the direct impacts of climate change on crop production are also considered to have made a ‘slight’ contribution (Gregory & Ingram, 2008). While wheat produced in Australia is only a small proportion of global production (mean 3.4% from 1995 to 2007), exports from Australia over this period, on average, were 13.9% of wheat that was traded globally (Gregory & Ingram, 2008). With increasing global demand from 2007-2008, a significant reduction in supply from Australia at this time would in turn relate to increasing wheat prices. Figure 3-4 – Monthly FAO price indices for basic food commodity groups. (Source: FAO, 2009b, p10.) In addition to climate change affecting global grain production generally, and in turn global price, a changing climate may also affect the production of grain in the regions that Colombia and MEB import from. This may in turn lead to changes in export and import behaviour. For example, maize production in the USA is projected to increase relative to current yield in both 2030 and 2090 in dryland cultivation, and remain steady or decrease slightly in irrigated cultivation (Figure 3-5). This is unlikely to have negative impacts for MEB’s grain import regime. 111 Figure 3-5 – Model simulations of average changes in USA crop yields for 16 crops, including maize (corn). Yield changes are given as percentages and represent the differences between current yields and those projected for 2030 and 2090 for two different climate scenarios (Source: USA National Assessment Synthesis team, 2000). Finally, climate change may have a direct impact on imports if it affects demand for grain. In addition to the effects of a changing climate on the economy in general, and in turn on buying power and demand, climate change may also affect buyers more directly. According to the IPCC, increases in heat stress may enhance the mortality of pigs and broiler chickens grown in intensive conditions. Additionally, increasing temperatures may increase the risk of livestock diseases by supporting the dispersal of insect vectors, enhancing the survival of viruses from one year to the next, and improving conditions for new insect vectors that to date have been limited by temperature (IPCC, 2007). As these livestock are often dependent on grains such as maize for feed, negative climate impacts on the livestock industry may in turn lead to reduced demand for grain. 3.3 Impact of climate change on Colombian grain imports Figure 3-6 shows that Colombian maize imports have grown substantially since 1990, rising in most years, though plateauing in the late 1990s/early 2000s, and again in the late 2000s. Interestingly grain imports did not decrease below previous years in 2007 and 2008 (during the high prices of the food crisis), though did decrease a little in 2009, a mark of the recent recession. Figure 3-7 supports this observation, showing a low correlation between maize imports and maize price. It even shows that imports were highest when prices were highest. Thus, while maize price has fluctuated since 1990, rising notably in 1996 and since 2007, maize imports have continued to rise and/or have remained high. This suggests that Colombian maize imports are not sensitive to price, but rather are driven by growing demand. As the majority of Colombia’s maize is used for animal feed13, this is a reflection of its “dynamic and growing economy” (USDA FAS, 2008, p2) in which increasing wealth is leading to increasing meat demand. As is shown 13 From 2007 to 2009, 73-75% of Colombian maize was used for animal feed rather than for human or industrial consumption (USDA FAS, 2009, p11). 112 in Section 65 of the report, Colombia’s GDP is largely tracking the overall increasing trend of global GDP. It is also noted that there are other drivers behind increasing grain import for livestock, such as the introduction of cold chains, in turn allowing greater levels of meat production. Figure 3-6 – Total Colombian maize imports and international price of yellow maize from 1990- 2009. Source: Colombian Ministry of Agriculture (2010) 14 Figure 3-7 – Total Colombian maize imports plotted against international price of yellow maize, 1990-2009 data. Source: Colombian Ministry of Agriculture (2010) Similarly Figure 3-8 illustrates a general increasing trend of Colombian wheat imports since 1990. This trend does not correlate with the more variable global wheat price, which rose significantly in 1996 and in 2007 and 2008. Again, this is supported by Figure 3-9, showing the lack of relationship between global wheat price and Colombian wheat imports. This suggests that wheat imports have been influenced more strongly by Colombian demand than by international price. While less than 2% 14 Maize imports include all varieties of maize registered on the Colombian Ministry of Agriculture website. The international price of yellow maize was used rather than white maize, as yellow maize is used mostly for livestock (the key type of maize imported at MEB). The maize price is sourced from the Chicago Stock Exchange via Reuters, listed as ‘Andean Community’. 113 of wheat in Colombia is used for animal feed 15 (USDA FAS, 2009), a growing economy will nevertheless lead to increased demand for staple food products such as wheat. Figure 3-8 – Total Colombian wheat imports and international price of wheat from 1990-2009. Source: Colombian Ministry of Agriculture (2010) 16 Figure 3-9 – Total Colombian wheat imports plotted against international price of wheat, 1990- 2009 data. Source: Colombian Ministry of Agriculture (2010) 17 15 From 2007 to 2009, under 2% of Colombian wheat was used for animal feed(USDA FAS, 2009, p11). 16 Wheat imports include all varieties of wheat registered on the Colombian Ministry of Agriculture website. The wheat price is sourced from the Kansas Bag via Reuters, listed as ‘Andean Community’. 17 Wheat imports include all varieties of wheat registered on the Colombian Ministry of Agriculture website. The wheat price is sourced from the Kansas Bag via Reuters, listed as ‘Andean Community’. 114 References Agrimonde. 2009. Agrimonde Scenarios and Challenges for Feeding the World in 2050. Paper prepared for the FAO Expert Meeting on “How to Feed the World in 2050,” FAO, Rome, 24- 26 June 2009. Provisional version produced June 2009. Cline, W. R. 2007. Global Warming and Agriculture: Impact Estimates by Country. Center for Global Development, Washington, DC, USA, and Peterson Institute for International Economics, Washington, DC, USA. pp 201. Colombian Ministry of Agriculture. 2010. Information from website. Available at: http://www.agronet.gov.co/agronetweb/ (accessed 04/05/2010). Fischer, G. 2009. World Food and Agriculture to 2030/50: How do climate change and bioenergy alter the long-term outlook for food, agriculture and resource availability? Paper prepared for the FAO Expert Meeting on “How to Feed the World in 2050,” FAO, Rome, 24-26 June 2009. Final draft produced August 2009. Gregory, P. J. & Ingram, J. S. I. 2008. Climate change and the current ‘food crisis’. CAB Reviews: Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources, 3, No. 099. IEA (International Energy Agency). 2009. World Energy Outlook Executive Summary. Paris, France. 13pp. Müller, C., Bondeau, A., Popp, A., Waha, K & Fader, M. 2009. Climate Change Impacts on Agricultural Yields. Background note for the WDR 2010. Stern, N. 2006a. The Economics of Climate Change: The Stern Review. Cambridge University Press. [Online]. Available at: http://www.hmtreasury.gov.uk/independent_reviews/stern_review_economics_climate_ch ange/sternreview_index.cfm (accessed 2/03/10). ———2006b. Gurukul Chevening Lecture: The Stern Review on the Economics of Climate Change at the London School of Economics. Nov 2006. UN FAO (UN Food and Agriculture Organization) 2009. Global agriculture towards 2050. Issue brief prepared for the FAO Expert Meeting on “How to Feed the World in 2050,” FAO, Rome, 12- 13 Oct 2009. 4pp. USDA FAS. (US Department of Agriculture Foreign Agricultural Service). 2008. Colombia Agricultural Situation, Food and Grains Price Report. Global Agriculture Information Network Report, Foreign Agricultural Service, United States Department of Agriculture, USA. GAIN Report Number: CO8005. ———2009. Colombia Grain and Feed Annual 2009 Report. Global Agriculture Information Network Report, Foreign Agricultural Service, United States Department of Agriculture, USA. GAIN Report Number: CO9006. 115 Climate risk case study: Terminal Marítimo Muelles El Bosque Appendix 5: Supplementary information to Section 6 ‘Demand, Trade Levels and Patterns’, on climate change impacts on selected Colombian agricultural exports 1. Introduction As part of assessing the risks of climate change for imports and exports, it can be useful to consider the vulnerability to climate change of export crops going through ports. Because agricultural exports only represent a small portion of MEB’s revenues, their vulnerability to climate change was not assessed in detail. However, in the light of the importance of agricultural exports to the revenues of other ports in developing countries and to provide an indication of the level of information that is available for climate change impact assessments, this Appendix presents information on climate change impacts on a number of crops which have gone through MEB in the past: coffee, bananas, sugar cane and plantain. 2. Coffee 2.1 Sensitivity of coffee crop to climate and climate change Ideal growing conditions for Arabica coffee (which is grown in Colombia) areȗ 19 - 21.5ȗ Θ 1800mm - 2800mm rainfall (Andrew Jarvis, 2009). Cenicafe, the Colombian National Centre for Coffee Research, (Pers. Comm., 2009) suggest a wider range for temperature (18 - 22ȗͿ ĂŶĚ precipitation (1500 - 3000mm), while the International Coffee Organisation (ICO) (2009) present an even wider range for optimum mean annual air temperature (18 - 23ȗͿ͘ĐĐŽƌĚŝŶŐƚŽ/K;ϮϬϬϵͿ͕ above a mean of 23ȗ͕ƚŚĞĚĞǀĞůŽƉŵĞ nt and ripening of cherries is accelerated which can often lead 18 ůŽǁ ƚĞŵƉĞƌĂƚƵƌĞƐ ĐĂŶ ŚĂŵƉĞƌ ŐƌŽǁƚŚ 1. Also, continual to loss in quality, whereas below ȗ͕ exposure to daily temperatures as high as 30ȗĐĂŶůĞĂĚƚŽƌĞĚƵĐĞĚŐƌŽǁƚŚĂŶĚƚŚĞĚĞǀĞůŽƉŵĞŶƚŽĨ abnormalities such as leaf yellowing (ICO, 2009). In addition to these annual average optima, varying climatic conditions are preferable at different stages of the coffee plant’s life cycle. For example, heavy rain during flowering can result in low productivity, as it did in 2008/9 in Colombia (Jarvis et al., 2009). This is due to it limiting pollination. Heavy precipitation during the harvest season can also lead to losses as cherries drop to the ground, mixing with unripe/rotten berries, in turn attracting the coffee berry borer (Andres Guhl, Centre for Interdisciplinary Studies, Universida de los Andes (CIDER), Pers. Comm., 2009). Coffee planted in different production systems also responds to climatic conditions differently. Sun grown coffee requires more water (as well as management and fertiliser) and is more at risk from rising temperatures as there is no protection for the plant. Shade grown coffee conversely provides protection as it reduces temperatures by approx 1-2ȗ͘dŚŝƐƐLJƐƚĞŵŝƐŝĚĞĂůĨŽƌƌĞŐions north of 6ȗE where dry seasons are prolonged. Shade grown coffee also reduces the variability of temperatures from day to night (Jarvis et al., 2009). Note that while sun grown is generally more profitable than shade grown coffee, organic specialist shade grown coffee can be competitive by charging a premium price. The latter however is usually only open to medium/large landowners due to the high cost of certification/auditing (Andres Guhl, 1 Note however that “selected cultivars under intensive management conditions have allowed Arabica coffee plantations to spread to marginal regions with mean annual air temperatures as high as 24’C or 25’C, with satisfactory yields, such as in the Northeast and North regions of Brazil” (ICO, 2009, p10). 117 CIDER, Pers. Comm.). Also note that in Colombia all coffee systems are rainfed, which provides a competitive advantage against countries where coffee is irrigated. In Colombia, coffee is grown across 875,673ha (MADR, 2009 In: Cadenas-Lopez, 2009) and between 1200m - 2000m altitude (Cenicafe, Pers. Comm., 2009). Climatic conditions for now and projected for the future (in 2050 under an A2 climate scenario), for different elevations, are shown in Table 2-1. As climate change can lead to the shifting of climatic bands “upslope”, it in turn may lead to the shifting of optimal growing regions (see Figure 2-1) 2. Table 2-1 – Current and future (2050, A2 climate scenario) temperature and precipitation change for Colombia. Source: Jarvis et al., 2009. Current Annual Future Annual Annual Average Current Annual Future Annual Total Altitudinal Average Average Temperature Total Total Precipitation Range Temperature Temperature Change Precipitation Precipitation Change 190-500 25.54 27.70 2.16 5891 6002 1.88 501-1000 23.47 25.66 2.19 3490 3597 3.04 1000-1500 21.29 23.50 2.21 2537 2641 4.10 1500-2000 18.36 20.58 2.22 2519 2622 4.08 2000-2500 15.60 17.82 2.22 2555 2657 4.00 2500-3000 13.33 15.54 2.21 2471 2575 4.20 2 o Note temperature reduces by 0.51 C per 100m increase in altitude. 118 Figure 2-1 – Current and future (2020 and 2050, A2 climate scenario) coffee suitability at different elevations. Source: Jarvis et al., 2009. Overall, according to research by CIAT, in Colombia suitable areas for coffee cultivation are expected to decline by 2050 under the A2 GHG emissions scenario (Table 2-2). Table 2-2 – Coffee crop suitability in Colombia – current and projected for 2050 under an A2 climate scenario. Change in suitable land area is shown for land of more and less suitable land. Source: Ramirez & Jarvis, 2009a. Crop suitability (for crops grown in ideal conditions, suitability is 100%; less Current suitable conditions are expressed Area Future Area Change in Area relative to this) (km2) (km2) (km2) 50 – 60 202,331.9 11,097.3 -191,234.6 61 – 70 47,142.1 17,979.4 -29,162.7 71 – 80 29,764.9 28,646.5 -1,118.4 81 – 90 28,560.6 28,560.6 0.0 91 - 100 65,895.6 65,895.6 0.0 Research by CIAT has investigated climate change impacts on coffee production in the Cauca region. In this region, the current area suitable for coffee growing (see Figure 2-2, top left) is projected to decline marginally by 2020 (Figure 2-2 top right) and significantly by 2050 (Figure 2-2, bottom). However, by 2050, some new parts of the Cauca region are projected to become highly suitable for coffee production. 119 Figure 2-2 – Current and future (2020 and 2050, A2 climate scenario) coffee suitability across Colombia. Source: Jarvis et al., 2009. Reduced suitability – need to find alternative livelihood to coffee Mixed suitability – requires management Improved suitability – new market opportunities 120 2.2 Sensitivityofcoffee’spestsanddiseasestoclimateandclimatechange  The two major coffee pests/diseases in Colombia are the coffee berry borer (or ‘broca’) and coffee rust.  According to Cenicafe and Andrew Jarvis (CIAT), broca is more prolific at higher temperatures, especially above 21.5ȗC (Figure 2Ͳ3), which relates temperature and the intrinsic rate of increase of broca, illustrates this point. This is one of the reasons why coffee cannot be grown at low altitudes. Also, relative humidity is thought to be a driving factor for broca (Andres Guhl, CIDER, Pers. Comm., 2009). Angela Arcila (CORPOICA) is currently working with the Ministry of Agriculture in Colombia to map current and future pest distribution (including broca) under climate change based on this kind ofsensitivityanalysis.  Figure 2Ͳ3 – Hypothenemus hampei intrinsic rate of increase (rm) estimated as function of temperature(oC).Source:Jaramilloetal.,2009.    Coffee rust is also exacerbated by high temperatures (21.5ȗC+) as well as heavy precipitation conditions and high relative humidity (ICO, 2009). However, keeping coffee in low precipitation regions to avoid rust, can also lead to coffee growth being impeded (Andes Guhl, Pers. Comm.). Given the relative humidity found in shade grown coffee plantations, Guhl considers that rust is more of a problem for shade grown than sun grown coffee (though Cenicafe do not consider this clear). Also, from a management perspective, it is easier to manage and treat coffee pests and diseasesinthemoreintensivesungrownplotsthaninshadegrownplots.  Theseclimatesensitivitiesinturnputcoffeeinavulnerablepositioninthefaceofclimatechange.   121 2.3 Possible adaptation options Although there is high confidence that temperatures will increase in the future in Colombia’s coffee growing regions, it is less clear what the patterns of precipitation change will be. Nevertheless, below are some proposed adaptation solutions which may be revised once further climatic analysis has been performed. According to Andrew Jarvis (CIAT), the first adaptation option to consider for coffee is management. This includes the use of shade in coffee growing (as it reduces temperatures), as well as irrigation (in the face of declining rainfall supplies). It is worth noting however that to date coffee in Colombia has been rainfed. This strategy may also include managing for changing pest/disease conditions. In the case of increasing numbers of broca (say due to increasing temperature), the use of ecosystem services provided by natural habitats as a control mechanism may be very effective, e.g. the provision of Astecca ant species/ various types of spider which are predators of the broca (Tomos Leon, IDEA, Pers. Comm., 2009; Lorena Vidal, Fundacion Natura, Pers. Comm., 2009). Jarvis’s second adaptation option is variety change. Here the planting of new varieties can be considered – those which are more able to cope with changing climatic conditions. (Cenicafe also mentioned that they are considering this). Moving production to higher altitudes can also be considered as an adaptation strategy (though there will be competition for land/land ownership issues to contend with). While some areas can be managed to cope with a changing climate and new areas will become viable for the first time, some areas will no longer be viable. Clearly, one possible strategy for farmers with land in areas where suitability declines significantly is to change livelihood – e.g. from coffee to a more climate change- resilient agricultural product. When considering adaptation options for coffee, it is important to note that “the vulnerability of smallholder farmers to coffee and crop suitability decrease is very site-specific” and that therefore “site specific adaptation strategies are needed” (Laderach et al., 2008). Figure 2-4 demonstrates the high proportion of small farms producing coffee in Colombia. Figure 2-4 – Percentage of Colombian farms under 10ha. Source: MADR, 2005. In: Jarvis et al., 2009. 100 90 80 Porcentaje de fincas <10ha 70 60 50 40 30 20 10 0 Palma Banano Café Caña Arroz Cacao 122 3. Bananas 3.1 Sensitivity of banana crop to climate and climate change According to the EcoCrop model used by Ramirez and Jarvis (CIAT), growth parameters for the banana plant are as shown in Table 3-1. Table 3-1 – Banana crop growth parameters. In: Ramirez & Jarvis., 2009b. Growing season (days) 365 Freezing temperature (oC) 0 Absolute minimum temperature (oC) 16 Minimum optimal temperature (oC) 24 Maximum optimal temperature (oC) 27 Absolute maximum temperature (oC) 35 Absolute minimum rainfall (mm) 700 Minimum optimal rainfall (mm) 1000 Maximum optimal rainfall (mm) 1300 Absolute maximum rainfall (mm) 5000 In work conducted by Ramirez and Jarvis, the areas that are currently least suitable for banana growth are projected to experience greater losses of suitable land by 2050, while overall, gains are expected (Table 3-2 and Figure 3-1). Table 3-2 – Banana crop suitability in Colombia – current and projected for 2050 under an A2 climate scenario. Change in suitable land area is shown for land of more and less suitable land. Source: Ramirez & Jarvis, 2009a. Suitability (for crops grown in ideal conditions, suitability is 100%; less suitable conditions are expressed Current relative to this) Area (km2) Future Area (km2) Change in area (km2) 50 – 60 31,485.4 25,979.7 -5505.7 61 – 70 30,367.1 27,098.1 -3269.0 71 – 80 30,367.1 31,657.5 1290.4 81 – 90 31,313.4 42,754.7 11441.4 91 - 100 836,083.6 859,396.0 23312.4 In more recent work, incorporating a better calibrated model, ZĂŵŝƌĞnj Θ :ĂƌǀŝƐ ŚĂǀĞ ƉƌŽũĞĐƚĞĚ country-wide banana suitability to decline. This is in spite of the banana crop moving upslope to adapt to climate change (see Figure 3-1). Like coffee, under a changing climate, suitable climatic/altitudinal bands for growth are expected to change/move upslope. 123 Figure 3-1 – Banana crop suitability in South America – current (left) and projected (right) for 2050 under an A2 climate scenario (right). Source: Ramirez & Jarvis, 2009b. Figure 3-2 – Banana crop suitability in Latin America under baseline (1960-1990) conditions and 2020, 2050 and 2080 A2 climate change scenarios. Data are from WorldClim. In: Ramirez & Jarvis., 2009b. 0.35 2080 (2070-2099) 0.30 2050 (2040-2069) 2020 (2010-2039) 0.25 Baseline Suitability 0.20 0.15 0.10 0.05 0.00 50 300 550 800 1050 1300 1550 1800 2050 2300 2550 2800 3050 Altitude (m) 124 3.2 Sensitivity of banana’s pests and diseases to climate The black sigatoka, a leaf spot disease of banana plants (caused by ascomycete fungus Mycosphaerella fijiensis Morelet) is the main disease currently threatening the banana industry (J Ramirez, CIAT). In Colombia, Cenibanano have been actively researching this disease due to its negative economic impacts and demands by consumers for pesticide-free products (black sigatoka has traditionally been controlled through the application of fungicides in aerial spraying, Cenibanano, Dec 2009). Using information on its climatic tolerances together wiƚŚ 'D ĚĂƚĂ͕ ZĂŵŝƌĞnj Θ Jarvis, show that black sigatoka severity is expected to marginally decline in Colombia under future climatic conditions. Figure 3-3 – Black sigatoka severity change under projected 2050 (2040-2069) A2 scenario climate change conditions. Source: Ramirez & Jarvis., 2009b. 10 8 BLS disease severity change 6 4 2 0 -2 -4 -6 -8 Mexico Venezuela Myanmar Cen. Africa Vanuatu South Africa Brazil Ecuador Haiti Cuba Japan China Taiwan Burundi Laos Costa Rica Colombia Bangladesh 3.3 Possible adaptation options ĐĐŽƌĚŝŶŐƚŽǁŽƌŬďLJ:ƵůŝĂŶZĂŵŝƌĞnjΘŶĚƌĞǁ:ĂƌǀŝƐ͕ƚŚĞĂďŝůŝƚLJŽĨƚŚĞŽůŽŵďŝĂŶďĂŶĂŶĂƚŽĂĚĂƉƚŝƐ variable (see Figure 3-4). 125 Figure 3-4 – Black sigatoka adaptability change under projected 2050 (2040-2069) A2 scenario climate change conditions. Source: Ramirez & Jarvis., 2009b. 40 30 Banana adaptability change 20 10 0 -10 -20 -30 -40 Mexico Venezuela Myanmar Cen. Africa Vanuatu South Africa Brazil Ecuador Haiti Cuba Burundi Japan China Taiwan Laos Costa Rica Colombia Bangladesh As climate change is expected to have a direct impact on banana crop growth, but less of an impact on pests and diseases (i.e. the black sigatoka is not expected to be a major problem), adaptation strategies may be better directed at managing direct impacts on growth. Ramirez & Jarvis propose the genetic improvement of banana crops as a way to build resilience to climate change. They can either be improved to be more drought or flood tolerant. Wild relatives and Creole cultivars provide a source of genes for improvement of current varieties. They also advocate making decisions about migration and management (using a CBA approach to adaptation decisions). 4. Sugar Cane 4.1 Sensitivity of sugar cane crop to climate and climate change Colombia is the eighth largest producer of sugar cane in the world (UN FAO, 2009). In Colombia, sugar cane is the fourth largest agricultural commodity produced (following cattle meat, cow milk, and chicken meat). Sugar cane is used to produce both sugar and panela (an unrefined product). Cane is grown across 220,000ha and is concentrated in one region – the Cauca river valley in the east of the country. Farms are large, with only about 1800 farmers in the business (Cenicana, Pers. Comm., 2009; see Figure 2-4). The main climatic components that control cane growth, yield and quality are temperature, light and moisture availability (Netafim, 2009). According to the UN FAO (2009), sugar cane flourishes under a long, warm growing season (with a high incidence of radiation and adequate moisture), followed by a dry, sunny and fairly cool but frost-free ripening and harvesting period. Optimum mean daily temperature for growth is between 22ȗĂŶĚϯϬȗ͕ǁŚŝůĞŵŝŶŝŵƵŵƚĞŵƉĞƌĂƚƵƌĞĨŽƌĂĐƚŝǀĞŐƌŽǁƚŚŝƐ about 20 ȗ C. For ripening, temperatures betweenȗ 10ĂŶĚ ϮϬȗ ĂƌĞ ĚĞƐŝƌĂďůĞ ĂƐ ƚŚŝƐ ĂůůŽǁƐ enrichment of sucrose content (UN FAO, 2009). Also, 1600mm of precipitation per annum is ideal according to Cenicana. (Thus in parts of Colombia where rainfall falls below this, significant irrigation is required). 126 Incorporating the climate sensitivities of sugar cane into the EcoCrop model, ZĂŵŝƌĞnjΘ Jarvis (CIAT) calculated the future suitability for sugar cane in Colombia (for 2050, under the A2 scenario). As shown in Table 4-1, suitability is expected to decline overall, particularly in the currently most suitable areas. Table 4-1 – Sugar cane crop suitability in Colombia: current and projected for 2050 under an A2 climate scenario. Change in suitable land area is shown for land of more and less suitable land. Source: Ramirez & Jarvis, 2009a. Suitability (for crops grown in ideal conditions, suitability is 100%; less suitable conditions are Current area Future area Change in area 2 2 expressed relative to this) (km ) (km ) (km2) 50 – 60 44,991.3 38,195.3 -6,796.0 61 – 70 34,754.4 32,345.6 -2,408.8 71 – 80 30,539.1 42,410.6 11,871.5 81 – 90 26,581.9 52,733.7 26,151.8 91 - 100 119,833.7 45,335.5 -74,498.2 4.2 Sensitivity of sugar cane’s pests and diseases to climate and climate change Sugar cane blight and the Diatrae borer pest are the most significant pest/diseases affecting sugar cane growth in Colombia (Cardenas-Lopez, 2009; Cenicana, Pers. Comm., 2009). The froghopper Aeneolamia varia is another pest, though not as problematic for production (Peck et al., 2001; Cardenas-Lopez, 2009). The Instituto Colombiano Agropecuario (ICA), the Colombian institute responsible for monitoring and controlling pest and disease outbreaks, is currently investigating impacts of climate change on these sugar cane pests and diseases. 4.3 Possible adaptation options With regard to changing suitability, as with coffee, producers will need to consider the range of options – management, new varieties and relocation/replacement crop. As sugarcane is a large scale industrial commodity however, approaches may be different to those of the small-scale coffee/banana farmer. According to Cenicana, if precipitation declined, more irrigation would be required whereas if it increased, better drainage would be needed. In terms of addressing any possible changes in pest/disease, Cardenas-Lopez (2009), on behalf of ICA, recommends the implementation of pest-monitoring plans, forecasting and early warning systems. 127 5. Plantain 5.1 Sensitivity of plantain crop to climate and climate change Colombia is the fifth largest producer of plantain in the world (FAO, 2009). With their knowledge of plantain’s climatic sensitivities, Ramirez and Jarvis (CIAT) used the EcoCrop model to project changes in area of suitable land for plantain in 2050 (under the A2 scenario) for Colombia. As shown below, the areas that are currently most suitable will become least suitable in future. Nevertheless, areas currently less suitable are projected to improve in their ability to support plantain growth, are growth in projected overall (Table 5-1). Table 5-1 – Plantain crop suitability in Colombia – current and projected for 2050 under an A2 climate scenario. Change in suitable land area is shown for land of more and less suitable land. Source: Ramirez & Jarvis, 2009a. Suitability (for crops grown in ideal conditions, suitability is 100%; less suitable conditions are expressed Current area Future area Change in area relative to this) (km2) (km2) (km2) 50 - 60 30,195.0 105,381.3 75,186.3 61 - 70 27,442.2 296,358.1 268,915.9 71 - 80 53,766.0 299,884.9 246,118.9 81 - 90 127,059.3 123,446.8 -3,612.5 91 - 100 644,934.1 80,520.0 -564,414.1 As for banana production, black sigatoka is the main disease affecting plantain production in Colombia (Cardenas-Lopez, 2009). As shown for the banana, black sigatoka is projected to become slightly less of a problem under future climate change conditions (see Figure 3-3). 5.2 Possible adaptation options As with the banana, climate change is expected to have a direct impact on crop growth and less of an impact on pests and diseases (i.e. the black sigatoka is not projected to be a major problem). Therefore, adaptation strategies may again be better focused on managing direct impacts on growth. 128 References Cardenas-Lopez, J. 2009. Future needs in SPS Technical Assistance. Presentation for ICA at ‘Climate change and agricultural trade: risks and responses’ conference, The World Bank, Washington D.C. ICO. 2009. Climate change and coffee. A report on the effects of climate change on producing countries to assist Members of the ICO with preparation for the UNFCCC COP15 Conference, 14 Sept 2009. Jaramillo, J., Chabi-Olaye, A., Kamonjo, C., Jaramillo, A., Vega, F., E., Poehling, H.-͕D͕͘Θ Borgemeister, C. 2009. “Thermal tolerance of the Coffee Berry Borer Hypothenemus hampei: Predictions of climate change impact on a tropical insect pest”. PLoS ONE. 4(8). Bristol, UK. Jarvis, A., Ramirez, J., Laderach, P., Guevara͕͘ΘZapata, E. 2009. Escenarios de Cambio climático en Colombia y la agricultura. CIAT presentation, November 2009. >ĂĚĞƌĂĐŚ͕W͕͘:ĂƌǀŝƐ͕ΘZĂŵŝƌĞnj͕:͘2008. The impact of climate change in coffee-growing regions: the case of 10 municipalities in Nicaragua. Project results for Cooperación Publico-Privada “AdapCC” Netafim. 2009. Netafim website. Available at: http://www.netafim.com/. Last accessed: 24/05/2010. ZĂŵŝƌĞnj͕:Θ:ĂƌǀŝƐ͕͘ϮϬϬϵĂ. Changes in crop adaptation for 28 major crops of Colombia, CIAT report to Departamento de Planeacion Nacional. ———2009b Impactos e implicaciones del cambio climatico para la produccion de musáceas en Latinoamérica y el Caribe. In: Sandoval JA (Ed.) Resúmenes 3er Congreso Científico-Técnico Bananero Nacional: del Laboratorio al Campo. Guápiles, Costa Rica, 10-13 November 2009. UN FAO. 2009. Food and Agriculture Organization of the United Nations website. Available at: http://www.fao.org/ (accessed: 15/12/2009). 129 Climate risk case study: Terminal Marítimo Muelles El Bosque Appendix 6: Supplementary information to Section 7 on ‘Goods Storage’ 1. Introduction Goods storage capacity has been highlighted by MEB staff as one of the key limiting factors to MEB’s growth. There are five key climate-related factors that have been identified as potentially affecting goods storage and are addressed in Section 7 of the main report: x Surface flooding as a result of heavy precipitation and sea level rise, x Seawater flooding of storage areas, x Changes in rainfall affecting the volume of water required for spraying the coke stored at the port, x Increased temperature reducing the efficiency of refrigerated containers, and x Increases in temperature and moisture affecting grain storage. Additional information on the methodology and assumptions used in assessing the risks of increased surface flooding and increased electricity requirements for refrigeration can be found below. 2. Assumptions and detail of the technical analysis on surface flooding risk The drainage system on the Isla del Diablo (or island site) is composed of a sequence of pipes laid approximately 1.5m below the ground surface. The diameter of the pipes varies, but the majority appears to be 254mm (10’’) PVC construction pipes. These pipes discharge directly to sea through a number of drainage outlets on the side of the quay. Flow through these pipes is restricted by major losses due to friction along the length of the pipe and minor losses, which occur at orifices or fittings. Major losses can be described by the Hazen-Williams formula. It is assumed that minor losses are zero as there is no quantifiable constriction at the entrance or the exit of the pipe. In reality the drainage grates may be the crucial factor in the drainage capacity of the port (see Figure 2-1). Figure 2-1 – Drainage grate on Quay 2 1 1 Terminal Marítimo Muelles El Bosque Cartagena, Colombia, Technical Review prepared for IFC, 2007. 131 The resulting equation for the maximum flow through pipes is the following (Weiner and Matthews, 2003): Q = 0.278CD2.63S0.54, where Q is the flow rate through the pipe (in m3s-1), C is the roughness co-efficient (unitless), D is the diameter of the pipe (in m) and S is the gradient of the pipe (unitless). The gradient is calculated by dividing the height of the drainage channel discharge point on the wall below the quay by the length of the pipe 2. In this analysis, the following assumptions are made: C = 130 (appropriate for PVC pipes), D = 0.254m and S = 0.015 (1.5 m / 100 m). The resulting estimated maximum flow through a single 100m long pipe is 0.1 m3s-1, 367 m3h-1; or 8796 m3d-1. The total area to be drained on the island site is approximately 41,300m3. There are three drainage outlets for this area. For simplicity it is assumed that all pipes are 100 m in length, which is approximately the length of the longest pipe along the quay, when calculating the head loss (this is a conservative estimate). The head is the pressure exerted by the weight of water above a given point. In this case, it is calculated as the difference in height between the drainage pipe inlet and the outlet or sea level (whichever is higher). The greater the head is, the greater the maximum water flow through the pipe can be. The relationship for a 100m pipeline between maximum water flow rate and the head is shown in Figure 2-2 for the drainage pipes of the general cargo yard. At present, the difference in elevation between water entering the pipe and the drainage outlet is approximately 1.5m. The height of the drainage outlets varies throughout the island site of the port. The following are estimated heights of the drainage outlets above mean sea level in December 2009 3: x 33 cm for quays 1 and 2, x 50 cm for quay 3, x 50 cm for the container yard, and x 25 cm for the general cargo yard. 2 The information on the drainage pipe is extracted from the port and drainage plans provided by MEB during the site visit. 3 Email from Andres Burgos of 14 December 2009 132 Figure 2-2 – Maximum flow rate through a drainage pipe with reducing head In the port plans the drainage outlets on quay 3 are at 0.56 m and 0.8 m, which is greater than the heights measured in December 2009 4, which are quoted to mean sea level and provided in Table 2- 1. It is possible that this is due to the difference between mean sea level and the port plan datum which was 0.18 m in 2000 or 0.24 m in 2010. Table 2-1 – Height of drainage outlets above MSL Location Height above MSL – 2009 (m) Estimated height above port plan datum (m) Quay 1 and 2 0.33 0.56 Quay 3 0.5 0.73 Container yard 0.5 0.73 General cargo yard 0.25 0.48 The storage area and warehouses on the mainland site have a different drainage system (Figure 2-3). There is a low wall surrounding the mainland area, which at frequent intervals has narrow plastic drainage pipes through it. A photograph of one of the drainage pipes is shown in Figure 2-4. There is insufficient information available to calculate the capability of these drainage pipes to deal with the runoff from the storage area and warehouses. Given the limited information and the low risk on the island site, the risk is not assessed further for the mainland site. 4 Email from Andres Burgos of 14 December 2009 133 Figure 2-3 – Drainage plan for mainland area of MEB 134 Figure 2-4 – Drainage pipe on mainland storage area 3. Assumptions and detail of the technical analysis on refrigeration Saidur et al. 2000 found for domestic refrigerators that the increase in energy consumption with higher temperature gradients is exponential: as an example, for a 1°C increase in ambient temperature energy consumption generally increases by approximately 5% (Figure 3-1). As explained in the climate review in Appendix 2, there is a wide range of temperature projections above Cartagena across climate models. As an upper bound a temperature increase of 6°C by 2050s is here considered (this is projected across seasons by an ensemble of 14 downscaled GCMs). If the relationship between temperature and energy consumption observed for domestic fridges and reported in Figure 3-1 is assumed to apply in the case of MEB’s reefers, a 6°C increase in average temperatures would lead to a 30% increase in reefer energy consumption. 135 Figure 3-1 – Energy consumption of two 80 litre capacity domestic fridges (model E and model S) at a range of different ambient temperatures. Consumption increases by 47 Wh/ ȗ;ϱ͘ϬйͿĨŽƌŵŽĚĞů ĂŶĚϱϯtŚͬȗ;ϱ͘ϰйͿĨŽƌŵŽĚĞů^ (Source: Saidur et al., 2002). MEB do not record electricity use specifically for reefers, therefore it is not possible to assume a simple annual electricity consumption increase. As a general average for all container and cargo types, under regular ambient conditions, a value of 3.6 kW per TEU can be used. A 20 ft container will use approximately 4 kW and a 40 ft container approximately 7 kW (GDV, 2010). Assuming that MEB has an average of 40 reefers stored (of 40 ft in length) and that the increase in electricity consumption is 30%, the resulting increase in annual electricity consumption at MEB is: 30% x 7kW x 40 = 84 kW. References GDV, 2010. Container Handbook. http://www.containerhandbuch.de/ (accessed 15/05/2010). Saidur, R., Masjuki, H.H., Choudhury, I.A. 2002. “Role of ambient temperature, door opening, thermostat setting position and their combined effect on refrigerator-freezer energy consumption”. Energy Conversion and Management 43 Pp845–854. 136 Climate risk case study: Terminal Marítimo Muelles El Bosque Appendix 7: Supplementary information to Section 12 ‘Social Performance’ 1. Introduction This appendix complements the analysis presented in Section 12 of the main report which reviews climate change risks to MEB’s social performance by considering impacts on local communities living in the area surrounding MEB and impacts on MEB’s employees. It provides background information on the social context in Colombia and Cartagena, which is useful to understand the inherent vulnerabilities that characterize Colombian communities and explain their sensitivity to climate change. 2. Overview of the social context in Colombia Over the past thirty years, social and economic development in Colombia has been hampered by violence and the struggle against drugs cartels and their militia. The violence has badly impacted certain regions of Colombia and the economy as a whole. The election of President Uribe in 2002 on a tough anti-violence ticket and his subsequent re-election in 2006 has brought some positive change to the security situation and to the economy. The 2009 Human Development Report (UNDP 2009) places Colombia towards the lower end of the bracket of countries with high human development, behind neighbouring Brazil and Venezuela but ahead of Peru, Bolivia and Ecuador. The economy has strengthened with increased foreign direct investment, increased GDP and reduced unemployment, but although Colombia’s growth has had a positive overall effect on poverty, stark inequalities and vulnerabilities in Colombian society still exist. Poverty in rural areas remains high (68%) compared with urban areas (42%), and particular groups including Afro-Colombians, indigenous peoples and female headed households are particularly vulnerable. In Colombia, as elsewhere in the world, the people least able to adapt and the most vulnerable to climate change are the poorest and marginalised sections of society. The First National Communication on climate change published by the Government of Colombia highlights coastal erosion and flooding as potential impacts of climate change on the coastal human population. Malaria and dengue fever are highlighted as threats to human health. While these climate change- related impacts may present threats to human health and livelihoods elsewhere in Colombia, as will be shown later, they are not considered threats to the population in the area surrounding MEB. 3. Overview of the social context in Cartagena The city of Cartagena where MEB is located has played and continues to play an important role in national and regional trade. The area around the port is of mixed use, containing both industrial and residential buildings. The port shares a small part of its perimeter fence with a residential area. As indicated in the 2009 Human Development Report, Colombia is ranked as a high human development country, which should mean that its population is less vulnerable to climate change than some countries of low human development. Education spending in Colombia is high by international standards, but has yet to be translated into better outcomes. In Cartagena District a rate of gross coverage in basic education of 94% was reached in 2003. In the case of primary basic education, the cover is 100% (World Bank, 2008). Colombia has seen important progress in the health sector in recent years. Some 64% of the population has health insurance, there is 93% coverage in immunizations, the infant mortality rate is 24.4 per thousand, maternal mortality is 100.1 per one hundred thousand live births, and 89% of the population has access to drinking water. Waterborne bacterial diseases and ambient and indoor air 138 pollution are the most important health effects of environmental degradation. Diarrhea, which accounts for approximately 7.3% of child mortality, is the most common and widespread waterborne disease and most prevalent in rural areas populated by the poorest segments of the population (Word Bank, 2008). Public health in Cartagena has experienced difficulties following the closure of the University Hospital and the Lions Club Clinic. These closures considerably reduced the supply of health services. In 2003 the district had 1,360 beds available (public and private), to cater for a population of 978,000 inhabitants, a ratio of one bed per 720 inhabitants (The National Administrative Department of Statistics (DANE); Invemar 2005). The poorest and most vulnerable sections of society in Colombia are the Afro-Colombian and Indigenous Peoples. These populations face particular challenges in achieving legal recognition, rights and access to basic services. According to the World Bank, illiteracy in Colombia is 16%, and that of the Indigenous and Afro-Colombian populations is 24% and 31% respectively; 14% of Afro- Colombian students attend secondary and superior schools compared to a 26% national average; and, the infant mortality rate among indigenous groups is at 63.3 deaths for every 1,000 live births, compared to a national average of 41.3. (World Bank, 2007). Indigenous communities are predominantly found along the Pacific coast, to the very north of Colombia and in the interior. Black communities are found along the Pacific coast, as indicated in Figure 3-1. There are a small number of people living in neighbourhoods in the vicinity of the port, as highlighted in Table 3.1. To the south of the port are two residential neighbourhoods, Zapatero, categorised as low in socioeconomic development (stratas 1 and 2), and the island of Manzanillo. Over time, residents of these communities have migrated to other sectors of the city giving way to the installation of more industrial and commercial activities. Today the island of Manzanillo houses customs warehousing and various other shipping and marine organisations as well as officers’ housing, the Naval Cadet School, and the Centre for Hydrographic and Oceanographic Research (CIOH). To the east are two neighbourhoods, el Bosque and Alto Bosque. The former is industrial with companies including Servientrega, Indufrial and Papas Margarita (see Table 3.2); the latter is categorised as socioeconomic level 4 and is residential. MEB itself has limited impacts on surrounding communities, discussed below, which points at limited potential risks associated with climate change for MEB’s social and financial performance, as well as for its reputation. 139 Figure 3-1– Distribution of indigenous reserves and black communities in Colombia (SIGOT – Sistema de información geográfica para la planeación y el ordenamiento territorial) 140 Table 3.1 – Demographics of the local area around MEB (MEB) Neighborhood Strata 1 Population No. Education Medical centers dwellings establishments 1 primary 1 naval Isla de Manzanillo 4 1520 380 school dispensary 1 primary and Bosque industrial 1,2 & 3 16113 2183 secondary Bosquecito 2 1080 147 Zapatero 1&2 1496 200 Cartagenita 1&2 1690 230 Manzanillo 1 176 31 Alto Bosque 4 2351 320 Table 3.2 – Businesses and institutions in the local area around MEB (MEB) Company/ Institution No. employees/ students Sociedad Portuaria 220 Papas Margarita 35 Indufrial 100 Algranel 23 Naval college/ Center for Hydrographic and Oceanographic Research 1220 (CIOH) 1 SISBEN (System for the Selection of Beneficiaries of Social Programs) is a proxy means test index widely used as a targeting system for social programs in Colombia. The SISBEN index is a function of a set of household variables related to the consumption of durable goods, human capital endowment and current income. SISBEN was created by the Colombian government with the purpose of simplifying, expediting and reducing the cost of targeting individual beneficiaries of social programs at the various government levels. In the Cartagena case this index is applied to the people belonging to the 1 and 2 social strata and served as a source of data for the social characterization of the population and identification of household welfare. (Invemar, 2005, p21) 141 Figure 3-2 – Neighborhood boundaries in Cartagena (Consorcio Consultores Cartageneros) MEB 4. Climatic vulnerabilities of Colombian communities Studies in other coastal cities of Colombia (Lampis, 2009) have assessed how certain sections of society, already vulnerable to disease and poverty, could be affected by climate change, and have highlighted the potential for civil unrest and violence disrupting socio-economic activities. In particular, communities living close to the shore, those reliant on fishing for their livelihood, and those marginalised by society are considered likely to be worst affected. Andrea Lampis (2009) has undertaken studies of Afro-Colombian black communities living in coastal areas around the city of Tumacao, on the Pacific coast. These communities are particularly vulnerable to climate change because they are marginalised and have limited capacity to adapt to any kind of change; their livelihoods are linked to the sea; and their homes are built directly on the coast, many in the traditional style on wooden poles. As described above, this community has difficulty accessing basic services and gaining recognition for their situation. Communities in Cartagena of similar social standing, with a close association to the marine environment could also be vulnerable to climate change. However, our understanding of the population and communities surrounding the port suggests that they would not be more vulnerable to climate change than the general population for the following reasons: x They are not thought to include any particularly vulnerable groups x The populations of low social development have migrated elsewhere x They are not dependent on artisanal fishing for their livelihoods x They do not live in traditionally constructed homes 142 With the exception of some naval officers’ housing on the island of Manzanillo, they do not live on the shoreline. Figure 4-1 – Map indicating human development variations across Colombia (SIGOT – Sistema de información geográfica para la planeación y el ordenamiento territorial) Bolivar state is categorized as medium human development 143 Figure 4-2 – Unmet basic needs across Colombia, 2005 (SIGOT – Sistema de información geográfica para la planeación y el ordenamiento territorial) Cartagena: 5-30% of people with unmet basic needs 144 References Lampis, A. 2009. Climate Change Vulnerability and Adaptation in Colombia. Presentation for the Institute for the Studies of the Americas (CIDER). Environment and Development in Latin America. 20th January 2009. UNDP (United Nations Development Programme). 2009. Human Development Report Colombia: Overcoming Barriers: Human Mobility and Development. http://hdrstats.undp.org/en/countries/country_fact_sheets/cty_fs_COL.html (accessed 22/ 06/2010). World Bank. 2007. Colombia 2006-2010: A Window of Opportunity. Policy Notes presented by the World Bank. Prepared By The International Bank for Reconstruction and Development and the World Bank. USA; Washington, D.C. 68 pp. ———2008. Country Partnership Strategy for the Republic of Colombia 2008-2011. http://web.worldbank.org/WBSITE/EXTERNAL/COUNTRIES/LACEXT/COLOMBIAEXTN/0,,cont entMDK:21738915~pagePK:1497618~piPK:217854~theSitePK:324946,00.html (accessed 22/06/2010). 145 Climate Risk and Business: Ports, Terminal Marítimo Muelles el Bosque C International Finance Corporation 2121 Pennsylvania Ave. NW Washington, DC 20433 Tel. 1-202-473-1000 www.ifc.org/climatechange The material in this publication is copyrighted. IFC encourages the dissemination of the content for educational purposes. Content from this publication may be used freely without prior permission, provided that clear attribution is given to IFC and that content is not used for commercial purposes. The findings, interpretations, views, and conclusions expressed herein are those of the authors and do not necessarily reflect the views of the Executive Directors of the International Finance Corporation or of the International Bank for Reconstruction and Development (the World Bank) or the governments they represent, or those of Terminal Marítimo Muelles el Bosque, S.A.