Water Global Practice Discussion Paper Water, Poverty, and the Economy State of Knowledge on Climate Change, Water, and Economics Anil Markandya About the Water Global Practice Launched in 2014, the Word Bank Group's Water Global Practice brings together financing, knowledge, and implementation in one platform. By combining the Bank's global knowledge with country investments, this model generates more firepower for transformational solutions to help countries grow sustainably. Please visit us at www.worldbank.org/water or follow us on Twitter at @WorldBankWater. 3 Introducing Commercial Finance into the Water Sector in Developing Countries State of Knowledge on Climate Change, Water, and Economics Anil Markandya © 2017 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 19 18 18 17 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Please cite the work as follows: Markandya, Anil. 2017. “State of Knowledge on Climate Change, Water, and Economics.” Discussion Paper. Washington, DC: World Bank. All queries on rights and licenses should be addressed to the World Bank Publishing, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights​ @­worldbank.org. Cover design: Jean Franz, Franz & Company, Inc. Abstract T he current state of knowledge on climate change and water points to predominantly negative effects on the economy and society. This paper reviews the literature on these effects by geographical region and notes the differences as well as the uncertainties. An important feature of the projections is the fact that the climate effects will occur on top of a water scarcity sit- uation that currently prevails in many parts of the world. The impact of climate change on scarcity is generally small compared to the impact of the socioeconomic factors. Adopting steps to increase the efficiency of water use could make a big contribution to addressing water problems, including those caused by climate change. In-depth estimates of damages from climate change related to water have been made to 2060 and, less accurately, to 2100. The 2060 estimates indicate that the negative impacts from changes in water supply or changes in water-­ related extreme events and marine flows could amount to about 1.5 percent of GDP in 2060 in the absence of mitigation or adaptation. This average figure, however, may be an underestimate for a number of reasons. ­ Estimates to 2100 of potential damages in economic terms are even more uncertain, but there are strong reasons to believe they will be greater as a percentage of GDP—perhaps around 10 percent globally, and possibly even higher. Adaptation can make a major contribution to reducing damages from climate change in all mitigation scenarios, and more so when mitigation is absent or limited. Adaptation will require both private and public actions. Adaptation measures need further analysis to include more of the softer options (such as those involving the use of ecosystems) and to incorporate steps to increase efficiency in the use of scarce water, as well as other resources. In terms of next steps, work is needed on how future economic growth could be affected by the effects of climate change on water and on water-related extreme events. In addition, a better understanding of how increases in the efficiency of water use could affect the water-energy-economic nexus in the context of climate change is needed. Finally, a better estimate of the likely reduction of damages from adaptation is needed, based on a detailed bottom-up assessment rather than a top-down one. ­ Introduction snowmelt and glacier melt. This discussion paper aims to show how climate impacts channeled Water is the central driver of the impacts of climate through  water will affect the economy and society change on society. It touches almost all areas and sec- coming decades, how these impacts depend in in the ­ tors of economic activity, through sea level rise and part on the evolution of greenhouse gas (GHG) storm surges; and through changes in precipitation, in ­emissions, and what measures can be taken to address evaporation due to higher temperatures, and in the impacts. As the title indicates, the perspective taken is an economic one: impacts are assessed as far  as p ­ ­ossible in e conomic terms, as are the This discussion paper was authored by Anil Markandya, Ikerbasque Professor, Basque Centre for Climate Change, and Honorary Professor, ­ measures  to reduce them through mitigation and University of Bath. adaptation. State of Knowledge on Climate Change, Water, and Economics 1 Studies of the effects of climate change focus on the on  a background of current water scarcity in many sectors shown in table 1, which highlights the links to ­ parts of the world. changes in water. Changes in the availability of water or its  rate of flow affect agriculture, coastal zones, Uncertain Water Supplies ecosystems, extreme events, health, energy, water stress, human security, and Climate model simulations suggest that, overall, global tipping points (large-scale dis- average precipitation will increase as global tempera- ruptive events). tures rise. As a result, total water availability is Throughout the next century, climate effects on water will expected to increase with climate change, but large be characterized by increasing The table clearly illustrates how regional differences are expected. At high latitudes variability in rainfall, increased water plays a key role in the and in some wet tropical areas, river flow and water uncertainty of river flows, and way climate change affects availability are projected to increase; however, they major changes in groundwater societies. Furthermore, these are expected to decrease in some dry regions at recharge. effects will be superimposed mid-latitudes and in the dry tropics (Calzadilla 2010). Table 1. Categories of Climate Impacts and the Role of Water in Producing those Impacts Sector Impacts Role of Water Agriculture Changes in crop yields (including cropland productivity and water stress) Changes in precipitation, surface Livestock mortality and morbidity from heat and cold exposure runoff, snowfall, groundwater. Variability of rainfall. Increased Changes in pasture- and rangeland productivity demand due to higher Changes in aquaculture productivity temperatures. Coastal zones Loss of land and physical capital from sea level rise Sea level changes, storm surges. Nonmarket impacts in coastal zones Ecosystems Loss of ecosystems and biodiversity As for agriculture, through changing Changes in forest plantation yields river flow regimes and water quality. Changes in potential fisheries catch Extreme events Mortality, land, and capital damages from hurricanes All such events involve increased Mortality, land, and capital damages from floods variability of water. Health Mortality and morbidity from heat and cold exposure (including heatwaves) Increase in water-borne diseases Mortality and morbidity from infectious diseases and cardiovascular and from extreme events. respiratory diseases Energy Changes in energy demand for cooling and heating Changes in demand for water. Changes in snowfall. Tourism Changes in tourism flows and services Water stress Changes in energy supply Directly related to water availability. Changes in irrigation water availability Changes in drinking water to end users (including households) Human security Civil conflict Water can be a source of conflict. Migration Tipping points Large-scale disruptive events Melting of polar ice sheets. Source: Adapted from OECD 2015a. 2 State of Knowledge on Climate Change, Water, and Economics In many regions, the positive effects of higher annual River flow is a useful indicator of freshwater availability runoff and total water supply are likely to be offset by for agricultural production. Irrigated agriculture, which the negative effects of changes in precipitation pat- currently accounts for 90 percent of global water con- terns, intensity, and extremes, as well as shifts in sea- sumption, relies on the availability of water from sur- sonal runoff. By itself, an increase in precipitation face and groundwater sources, which depend on the would increase soil moisture. However, even with seasonality and inter-annual variability of river flow. higher precipitation, surface runoff may decrease in Consequently, when river flow limits a region’s water some river basins because of greater evaporation in a supply and hence constrains its ability to irrigate crops, warmer atmosphere (IPCC 2001). Temperature is par- the impacts can be severe. ticularly important in regions dominated by snow, determining the timing of snowmelt and thus the sea- Calzadilla (2010) provides a map of predicted changes in sonality of available water. Therefore, the overall river flow relative to the 1961–90 period for two time global impacts of climate change on freshwater sys- periods (the 2020s and 2050s) and for the two emission tems are expected to be negative (Bates et al. 2008; scenarios (A1B and B2).1 The map (reproduced as map 1) Calzadilla et al. 2013). shows large regional differences, which do not change Map 1. Percentage Change in Annual Average River Flow for 2020 and 2050 under Two Emissions Scenarios Relative to the 1961–90 Average a. A1B – 2020 (2006–35) b. A1B – 2050 (2036–65) c. A2 – 2020 (2006–35) d. A2 – 2050 (2036–65) Percent change –100 –50 –30 –15 –9 –3 0 3 9 15 30 50 250 no data IBRD 42867 | APRIL 2017 Source: Calzadilla 2010. State of Knowledge on Climate Change, Water, and Economics 3 notably across the emissions scenarios. For both emis- (decrease) (see figure 1 for the regions). Changes in pre- sion scenarios and time periods, the number of coun- cipitation intensity may also decrease groundwater tries subject to decreasing river flow is projected to be recharge as infiltration capacity is exceeded (typical in higher than those with increasing river flow. Significant humid areas), but it may also increase it as a result of decreases in river flow are predicted for northern regions fast percolation (typical in semi-arid areas). in South America, southern Europe, the Middle East, Less snowfall is expected to lead to less groundwater North Africa, and southern Africa. By contrast, substan- recharge even if precipitation remains constant. tial increases in river flow are predicted for boreal Coastal groundwater is affected not only by runoff but regions of North America and Eurasia, western Africa also by sea level rise. Unconfined aquifers are likely to and southern Asia. Some exceptions are parts of eastern suffer salt water intrusion over a long period. Africa and the Middle East, where changes in river flow vary depending on the scenario and time period. There is also some evidence that changes in groundwa- Information on the effects of climate change on ground- ter recharge can also affect stream flows. A study from water is more limited than for surface water, but work Uganda indicates that high temperature increases are since 2007 has been increasing the knowledge base expected to decrease groundwater outflow to the (Jiménez Cisneros et al. 2014). In general, areas where Upper Nile Basin so much that the spring discharge will total runoff is expected to increase (decrease) are also disappear and the flow regime will change from ones where groundwater resources will increase bimodal to unimodal (Jiménez Cisneros et al. 2014). Figure 1. Percent of Population Facing Severe Water Stress 0.8 Percent of global population under severe stress 0.7 conditions (WSI>0.4) in year 2095 0.6 0.5 0.4 0.3 0.2 0.1 0 2000 2020 2040 2060 2080 2100 Time Human (8.8 w/m2) Climate (8.8 W/m2) Climate (lower bound) Both (5.5 W/m2) Human (7.7 w/m2) Climate (7.7 W/m2) Climate (upper bound) Both (4.2 W/m2) Human (5.5 w/m2) Climate (5.5 W/m2) Both (8.8 W/m2) Both (lower bound) Human (4.2 W/m2) Climate (4.2 W/m2) Both (7.7 W/m2) Both (upper bound) Source: Hejazi et al. 2014b. Courtesy of the Pacific Northwest National Laboratory, operated by Battelle for the U.S. Department of Energy. Note: The figure is derived using Ensemble Mean GCM. W/m2 = watts per square meter. 4 State of Knowledge on Climate Change, Water, and Economics Higher Levels of Runoff and Lower quality from toxins, such as Water Quality those produced by algae (OECD Higher levels of runoff are likely 2012, 2014; IPCC 2014). to lead to increased erosion and Heavy rainfall events are expected to increase in fre- sediment loads in rivers, and quency, and are likely to lead to an increase in erosion. Linkages between observed a reduction in water quality in Yang et al. (2003) estimate a 14 percent increase in ero- effects on water quality and many parts of the world. sion rates in the 2090s relative to the 1980s, with an water variability need to be increase of 40–50 percent in Australia and Africa. The interpreted with care, as many largest amount is expected in semi-arid areas, where other factors also must be taken into account. extreme events may account for up to 40 percent of Nevertheless, there is a medium level of confidence the erosion. In temperate regions, the impacts of such that if observed deterioration of water quality contin- events are less clear. They could result in a sharp ues, measures already in place to control pollution increase in erosion or a decline in areas where rainfall may be inadequate in light of the negative impacts of is predicted to fall in the erosion-sensitive months. In climate change (Jiménez Cisneros et al. 2014). general, land management practices are critical to the rate of erosion. With conservation tillage, these rates Reductions in Meltwater and Constraints can be reduced significantly. on Year-Round Water Supplies Climate change is also likely to increase sediment loads As with precipitation, changes in rivers through soil erosion processes coupled with in snowfall in the past cen- Meltwater from accumulated land use changes. This in turn can reduce flow rates. tury  are indeterminate; how- snowfall and glaciers feed The phenomenon is more pronounced in rivers ever, consistent with observed many of the world’s rivers and affected by melting glaciers and permafrost degrada- warming, shorter snowfall sea- help replenish groundwater tion, such as the Ganges. sons are observed over most of stocks. Hotter temperatures Literature on the impacts of climate change on the the Northern Hemisphere, with will reduce snowfall seasons quality of water is limited and uncertainty concerning snowmelt seasons starting ear- and shrink glaciers, constraining the effects is high. Studies indicate that impacts are lier. Decreases in the extent year-round water supplies. highly dependent on local conditions and on the cur- of permafrost and increases in rent state of pollution of the water body. In general, its average temperature are widely observed in observed impacts are likely to continue. some  regions of the Arctic and Eurasia (IPCC 2013, chapter 4). In most parts of the world, glaciers are los- Current data indicate that for lakes and rivers, the most ing mass. For example, almost all glaciers in the tropi- frequent impacts are more eutrophication and higher cal Andes have been shrinking rapidly since the 1980s; nutrient loads from increased storm runoff. Higher similarly, Himalayan glaciers are losing mass at present runoff also results in higher salts, pathogens, and (Bolch et al. 2011). heavy metals in the water. For rivers, the reported impacts point to reduced water quality, even when All projections for the twenty-first century (WGI AR5, runoff increases. Instead of diluting the pollution, the chapter 13) show continued mass loss from glaciers. process sweeps the pollutants from the soil into the As the glaciers shrink, their relative contribution to watercourses. Some of these flows also reduce oxygen summer flows decreases, and the annual runoff peak concentrations. In this context, it is important to note shifts toward spring (Jiménez Cisneros et al. 2014, that the availability of freshwater can be reduced by chapter 4). This shift is expected with very high confi- the negative impacts of climate change on water dence in most regions, although not in the eastern State of Knowledge on Climate Change, Water, and Economics 5 Himalayas, where the monsoon and the melt season confidence in the changes in frequency of fluvial floods coincide. The relative importance of high summer gla- (floods from overflowing rivers caused by intense rain- cier meltwater can be substantial. Glacier meltwater fall). Recent work on global flood projections indicate also increases in importance during droughts and heat flood hazards increasing in more than half the regions waves. If the warming rate is constant, and if, as of the world, but with great variability at the scale of expected, ice melting per unit area increases and total individual river basins. The ensemble of hydrology ice-covered area decreases, the total annual yield models reviewed in Jiménez Cisneros et al. (2014) indi- passes through a broad maximum, known as “peak cate that flood hazards will increase in parts of South meltwater.” Peak-meltwater dates have been pro- Asia, Southeast Asia, East Africa, Central and West jected between 2010 and 2050 for different parts of the Africa, northeast Eurasia, and South America. In con- world. In fact, declining yields relative to various trast, a decrease in flood frequency is projected in parts dates in the past have been detected in some observa- of Northern and Eastern Europe, Anatolia, Central tional studies; that is, the peak has already passed. Asia, central North America, and southern South There is medium confidence that the peak response to America. While there is agreement on the broad twentieth and twenty-first century warming will fall regions, there can be major differences at the local within the twenty-first century in many inhabited gla- level, where even the direction of change may be sub- cierized basins, where society is currently benefitting ject to dispute. from a transitory “meltwater dividend.” When they Droughts. There is medium confidence that, since the are in equilibrium, glaciers reduce the year-to-year 1950s, droughts have intensified and become longer in variability of water resources by storing water during some regions of the world, in particular in southern cold or wet years and releasing it during warm years. Europe and West Africa, SREX (2012) concludes. As glaciers shrink, their diminishing influence may However, in some regions, droughts have become make the water supply less dependable. (Jiménez less frequent, less intense, or shorter, such as central Cisneros et al. 2014). North America and North Western Australia. There is medium confidence that anthropogenic influence has contributed to some changes in the drought patterns Increased Risk of Floods and Droughts observed in the second half of the twentieth century, Although global trends in pre- based on its attributed impact on precipitation and Hydrological extremes—floods cipitation from several datasets temperature changes. There is, however, low confi- and droughts—are expected covering 1901 to 2005 are sta- dence in the attribution of changes in droughts at the to increase in frequency and tistically insignificant (Jiménez level of single regions because of inconsistent or insuf- intensity in some parts of the Cisneros et al. 2014), regional ficient evidence. world, affecting increasingly observations show that most vulnerable populations. Looking to the future, studies conducted since the droughts and extreme rainfall 4th  assessment by the Intergovernmental Panel on events of the 1990s and 2000s have been the worst Climate Change (IPCC) indicate a medium confidence since the 1950s (Arndt et al. 2010). Certain trends in in a projected increase in duration and intensity of total and extreme precipitation amounts emerge from droughts in some regions of the world, including the data. southern Europe and the Mediterranean region, cen- Floods. Temperature and precipitation changes could tral Europe, central North America, Central America result in increased frequency of floods, according and Mexico, northeast Brazil, and southern Africa. to  the SREX report (IPCC 2012), but there is low Confidence is generally low for other regions because 6 State of Knowledge on Climate Change, Water, and Economics of insufficient agreement of projections of drought of rise is projected to be 12.0 [8.5 to 16.0] mm/yr. under changes (dependent both on the model and dryness the most optimistic scenario, RCP2.6. Under the least index). Definitional issues and lack of data limit confi- optimistic  scenario (RCP8.5), the rate goes up to dence to a medium level in observations of drought 14.5 [11.0 to 19.0] mm/yr. These figures now include a changes, while these issues, plus the inability of mod- likely range of ice sheet flow contributions from els to include all the factors likely to influence Greenland and Antarctica (Wong et al. 2014).3 droughts, preclude anything more than medium Sea level rise is not expected to be uniform in space confidence in the projections. ­ and time. Natural modes of climate variability influ- Economic impacts of floods and droughts. Irrespective ence sea levels in different regions of the globe. This of the projected frequency and intensity of floods will affect the rate of rise from year to year and decade and droughts, their economic impacts are projected to decade. For example, in the equatorial Pacific, levels to increase even when the hazard remains constant can vary from the global mean by up to 40 cm due to El because of increased exposure and vulnerability Niño-Southern Oscillation. Regional variations in the (Jiménez Cisneros et al. 2014). In Europe, for ­example, rate of sea level rise on the coast can also arise from a from 1961 to 1990, flood damage was €6.4 billion number of other climate and ocean dynamic processes ($7.3  billion) and 200,000 people were exposed, (Wong et al. 2014). on  average, each year. This is expected to double by  the  2080s under scenario B2 and triple under Severe storms such as tropical and extratropical ­scenario A2.2 Drought impacts at continental and cyclones (ETCs) can generate storm surges over smaller scales are difficult to assess because they coastal seas. There is low confidence regarding changes vary greatly with the local hydrological setting and in tropical cyclone activity around the world during water management practices. More frequent droughts the twentieth century because of changes in observa- due to climate change may challenge existing water tional capabilities, although it is virtually certain management systems; together with an increase of there has been an increase in the frequency and inten- population, this may place even the domestic supply sity of the strongest tropical cyclones in the North at risk in parts of Africa. Atlantic since the 1970s (WGI AR5, Section 2.6). In the future, the frequency of tropical cyclones globally is  likely either to decrease or remain unchanged, Increased Risks to Coastal Areas but  global mean tropical cyclone precipitation rates  and maximum wind The 5th Assessment Report WGII chapter on coastal speed will likely increase (WGI systems and low-lying areas concludes that such areas The trend of global average sea AR5, Section 14.6). will increasingly experience adverse impacts such as level rise is likely to continue submergence, coastal flooding, and coastal erosion Extreme sea levels arise throughout the twenty-first from  combinations of factors, century, and at an accelerated due to relative sea level rise. It is very likely that global pace. Sea level rise, combined mean sea level rose at a mean rate of 1.7 [1.5 to 1.9] including astronomical tides, with more intense tropical mm/yr. between 1900 and 2010 and at a rate 3.2 [2.8 to storm surges, wind waves and cyclones, will place coastal areas 3.6] mm/yr. from 1993 to 2010 (WGI AR5, Section swell, and year-to-year vari- at greater risk. 13.2.2). Future rates are projected to exceed the ability in sea levels. To date, observed rate for the period 1971–2010 of 2.0 [1.7 to 2.3] observed trends in extreme sea levels have been mm/yr. for all Representative Concentration Pathway mainly consistent with mean sea level (MSL) trends. (RCP) scenarios. From 2046 to 2065, the expected rate Regarding future changes to storm surges, studies State of Knowledge on Climate Change, Water, and Economics 7 show strong regional variabil- Worsening Water Scarcity Regions currently facing ity and sensitivity to the choice Despite the fact that the world only uses 10 percent of water scarcity are expected to of Global Climate Model (GCM) the available fresh surface and groundwater, water scar- experience worsening scarcity, or Regional Climate Model city occurs because water availability is highly variable with socioeconomic factors (RCM). over time and space. The United Nations estimates that like population and economic The effect of future tropical about 1.2 billion people, or almost one-fifth of the world’s growth as leading contributors. cyclone changes on storm population, live in areas of physical water scarcity, and surges has also been investi- 500 million people are approaching this situation. An gated in a number of regions using a range of different additional 1.6 billion people, or almost one-quarter of methods. In general, the small number of regional the world’s population, face economic water shortage storm surge studies together with other uncertainties (where countries lack the necessary infrastructure means there is low confidence in projections of storm to take water from rivers and aquifers).4 By 2025, about surges due to changes in storm characteristics. 1.8 billion people will be living in countries or regions However, observed upward trends in mean sea level, with absolute scarcity, according to the United Nations. together with projected increases for 2100 and beyond, Hejazi et al. (2013) have estimated water scarcity using indicate that there is a high confidence that coastal sys- Raskin’s definition of scarcity as the ratio of total water tems and low-lying areas will increasingly experience withdrawal (TWW) to total water availability (TWA). extreme sea levels and their adverse impacts (see WGI The index indicates no scarcity when it is below 0.1, AR5, Section 13.7). low scarcity for values of between 0.1 and 0.2, moder- Changes in sea water are also expected to have an ate scarcity for values between 0.2 and 0.4, and severe impact on fisheries. Cheung et al. (2010) estimate scarcity for values greater than 0.4. They analyze future that climate change may lead to a large-scale redistri- demand under a range of socioeconomic scenarios bution of global catch potential, with an average of without climate change using the GCAM Integrated 30–70 percent increase in high-latitude regions and a Assessment Model,5 and find that by 2050, regions in drop of up to 40 percent in the tropics. Moreover, China, India, the Republic of Korea, and the Middle East maximum catch potential could decline considerably will be severely water stressed. When climate change is in the southward margins of semi-enclosed seas, included in the analysis, Hejazi et al. (2014a) show that while it is likely to increase in poleward tips of conti- by 2095 similar or elevated water scarcity conditions nental shelf margins. Such changes are most appar- are expected in these regions. More specifically, regions ent in the Pacific Ocean. Among the 20 most experiencing some level of scarcity are projected to important fishing Exclusive Economic Zone (EEZ) experience even more scarcity, primarily due to mount- regions in terms of their total landings, EEZ regions ing demands and changes in water availability. The larg- with the highest increase in catch potential by 2055 est increases in scarcity in the twenty-first century include Norway, Greenland, the United States include regions in eastern China, India, Western Europe, (Alaska), and Russia (Asiatic part).By contrast, EEZ and the Middle East. Some of that change is attributed regions with the biggest loss in maximum catch to the large increase in population and the income potential include Indonesia, the United States effect in these regions, which raises water demand (excluding Alaska and Hawaii), Chile, and China. under the assumed socioeconomic developments. A Many highly impacted regions, particularly those in number of other studies also predict an increase in the tropics, are socioeconomically vulnerable to demand for water due to climate change. Döll (2002) these changes. estimates that climate change may cause the global 8 State of Knowledge on Climate Change, Water, and Economics total irrigation requirement to increase by 5–8 percent The other important conclusion from the analysis is until the 2070s. About two-thirds of the area equipped the small effect of climate change on scarcity relative for irrigation in 1995 will experience an increase in irri- to human factors. Figure 1 shows the percent of popu- gation demand. Fischer et al. (2007) find an even larger lation under severe water stress due to climate (with increase of 20 percent in global irrigation needs by different mitigation targets), to human factors, and to 2080. About two-thirds of the increase results from both. Clearly, the effect of climate is minor compared higher irrigation intensity. These data form the basis of to that from socioeconomic factors.6 adaptation estimates discussed later in this paper. The analysis by Hejazi et al. (2014a) is relevant in two The Growing Need for Careful Management of Water Resources respects. The first is the role of mitigation policies on water scarcity. The model considers two kinds of Water pricing can play an taxes: a uniform carbon tax (UCT), which includes all important role in managing Increasing water scarcity implies carbon emissions in all sectors (including land use water resources. At present, the strong need to manage emissions) and all regions of the world; and a fossil the marginal value of water in water resources more carefully fuel and industrial emissions tax (FFICT) regime, different uses varies a great over the rest of this century. which does not include a tax on land-use-related car- deal because the prices paid bon emissions. Under both regimes, the carbon tax by industry, agriculture, and rises over time to limit atmospheric CO2 concentra- residential users often have no relation to each other. tions to a prescribed stabilization level. The different In the desert state of Arizona in the United States, for types of policies lead to dramatically different out- example, water prices vary from $27/acre-foot for comes for deployment of bioenergy, for emissions agriculture to $3,200/acre-foot for urban uses, stemming from changes in land use, and consequently Olmstead (2013) finds.7 While some of the variation greenhouse gas emissions and climate change. The can be explained by the difference in nature and FFICT case is characterized by very high deployment quality of the product being delivered, most of it is a of bioenergy and emissions associated with changes in function of institutions that do not allocate water land use, which lead to greater emissions and climate based on economic criteria. change than the UCT case. Hejazi et al. (2014b) report ongoing research in which There are also differences in water demand under the water withdrawal in each of the 235 river basins is two regimes. The results depend on both the tax tracked in the GCAM model previously discussed. regime (UCT and FFICT) as well as the mitigation tar- They find that with no change in pricing, withdrawals get chosen. The targets considered are scenarios will double over this century (from around 2,600 A2  and B1 (resulting in best guess temperature ­billion m3/year in 2000 to around 5,000 billion m3/ increases by 2095 of 3.4°C and 1.8°C respectively). year in 2100). By contrast, with a price to balance sup- With the UCT regime, populations under severe ply and demand, the increase is limited to between stress decline by 2.0–2.4 percent by 2095 (depending 3,100 to 4,000 billion m3/year, depending on what on the mitigation target chosen). Under the FFICCT considered accessible. amount of the stream flow is ­ regime, they increase by 0.2 percent under scenario Thus even if only a part of water use is allocated A2 and by even more (5.6 percent) under scenario B1. based on a price that applies across users to bring This unusual result arises because under scenario B1 supply and demand into balance, many of the prob- demands for water when implementing the FFICT lems of scarcity related to climate change and socio- regime are particularly high. economic factors scarcity will be resolved. State of Knowledge on Climate Change, Water, and Economics 9 Pricing is one method of conserving water use and driver of increased efficiency is the price of water increasing efficiency in its allocation, but other mea- relative to other inputs. Another factor is techno- sures of a more technological nature should also make logical change that improves efficiency. The latter, a contribution. Hertel and Liu (2015) notes the follow- however, is not enough to ensure a reduction in ing possibilities, in particular: water use. As Hertel and Liu (2015) note, farmers tend to irrigate larger areas and increase irrigation a. Evaporation from water storage. Such evaporation ­ echnology. intensity after adopting a more efficient t accounts for a significant fraction in some regions. Hertel cites the example of the Upper Rio Grande For example, reservoir evaporation in the semi- Basin in the southwestern United States, an area arid U.S. state of Texas amounts to about 61 per- studied by Ward and Pulido-Velazquez (2008). cent of total agricultural irrigation use during the They find that adoption of the more efficient drip year 2010 (Wurbs and Ayala 2014). In Australia, irrigation system through a subsidy for upstream Craig (2005) estimated this loss to be about 40 per- irrigators reduces return flow, leading to larger cent of the total storage volume. In many develop- depletion of downstream water. The overall with- ing countries, storage rates are low because of drawal from the basin turns out to be larger than inadequate water infrastructure. In Pakistan for before the subsidy for drip irrigation was provided. percent of average annual flows example, only 9  ­ The message is that technological efficiency needs are stored, while the world average is 40 percent. to go hand in hand with water pricing to have a full As a consequence, the country has limited oppor- effect on water conservation. tunity to conserve flood waters, to be able to release water during periods of low river flow, such c. Productivity of water. Unlike irrigation efficiency, as Rabi (winter) (GOP and UNEP 2013). This evapo- which measures the share of the diverted water rative loss could increase by about 15 percent by finally applied to plants, water productivity refers 2080, as surface temperatures rise in the face of to gaining more output per drop of water. This can climate change (Helfer, Lemckert, and Zhang be achieved either by raising yields (increasing the 2012). Increasing total usable water storage by numerator) or by reducing non-beneficial con- reducing this type of loss depends on the adoption sumptive water use (decreasing the denominator). of evaporation suppression technology, which is Increasing water productivity places emphasis driven by the marginal value product of the water on  agricultural practices. For example, limiting to be saved. non-beneficial evaporative loss could boost yields from 1 to 3 metric tons/ha; limiting deep percola- b. Irrigation efficiency. This refers to reliable and pre- tion of rainfall could further boost yields by another cise delivery of water to plants. One definition is 2 tons/ha (Hertel and Liu, 2015). Other measures the ratio of crop water requirement to irrigation that contribute to raising yields relative to evapo- water withdrawal. Average world irrigation effi- transpiration are better pest and disease control ciency was around 50 percent in 2005-2007, and adopting drought-tolerant cultivars. according to the United Nations Food and Agricultural Organization (FAO). In other words, The literature indicates considerable room to about half of the water withdrawal is “lost” improve water productivity. A number of recent between the source and the destination. Among all findings confirm a nonlinear relationship between the regions, Sub-Saharan Africa has the lowest irri- water productivity and yields. At a lower level of gation efficiency, averaging about half of global yield, even a small gain in yield can significantly efficiency (Alexandratos and Bruinsma 2012). A key increase water productivity. But when the farm 10 State of Knowledge on Climate Change, Water, and Economics moves to a higher level of yield, water use tends to throughout the world (see rise in direct proportion to output, providing much box  1  for a brief description Changes in precipitation and less incentive for farmers to save water. The of the model). The OECD proj- river flows, as well as adverse threshold for this nonlinearity is around 3 metric ect ­ contains one of the most climate-related shocks, will tons/ha. Most small-scale farmers in developing robust  assessment of agricul- have a detrimental impact countries operate below this threshold, suggesting tural impacts from climate on agricultural production, particularly in some of the that they could significantly increase water effi- change published to date. poorest and driest parts of ciency (Hertel 2015). The model finds that climate the world. change affects crop yields het- Threats to Agricultural Production erogeneously in different world regions. Further, the A recent model, ENV-Linkages, was developed by effects are not the same for different crops. Figure 2 the  Organisation of Economic Co-operation and illustrates changes in crop yields for paddy rice and Development (OECD) to estimate the impact of climate wheat at the global level in 2050. Falls in yields of change on 35 economic sectors and 25 regions paddy rice by 2050 are strongest in  tropical areas, Box 1. Modelling the Economic Impacts of Climate Change Estimates of the economic costs of climate change are generally conducted using Integrated Assessment Models (IAMs) with long-term perspectives, to the end of this century and beyond.a Most of these studies have a stylized, aggregated representation of the economy focusing on projections of climate change impacts over time. They often include highly aggregated integrated structures, in which climate change impacts in different sectors are aggregated and used to re-evaluate welfare in the presence of climate change. An IAM projection is presented in detail in the section on the Uncertainty of Cost Projections. A smaller strand of literature uses computable general equilibrium (CGE) models to examine the economic implications of climate change impacts in specific sectors, often using a comparative static approach.b Because CGE models have a more disaggregated structure, they need more information to determine annual equilibria, and to run them forward, linking annual changes for more than 40 to 50 years, becomes very difficult. On the other hand, they can track the impacts of climate in a more detailed way than IAMs, which rely on reduced form functions linking impacts to temperature. Recent work at the Organisation for Economic Co-operation and Development (OECD 2015a) has attempted to address these issues by combining a CGE model to investigate the economic impacts of climate change to 2060 with an IAM model (AD-RICE) to look at impacts beyond that. Because their results are similar to a number of other models for the two periods, it is instructive to discuss them in some detail. The OECD CGE model (ENV-Linkages) contains 35 economic sectors and 25 regions. It models trade flows as well as capital accumulation using capital vintages, in which technological advances trickle down only slowly over time to affect existing capital stocks. The model estimates the impacts of changes in different box continues next page State of Knowledge on Climate Change, Water, and Economics 11 Box 1. continued inputs (including water) as a result of climate change using a production function that represents the activity of a specific industry or group of industries in the basic structure of the model. Climate impacts have the potential to directly affect sectors’ use of labor, capital, intermediate inputs, and resources. They also affect the productivity of inputs to production. Adverse climate-related shocks to the economy therefore increase the need for more inputs to generate a given level of output. Compared to Integrated Assessment Models in which climate damages are subtracted as a total from GDP, the production function approach can also explain how the composition of GDP is affected over time by climate change: what sectors are most affected and what changes in production factors contribute the most to overall changes in GDP. a.  See, for example, Nordhaus (1994, 2007, 2010); Tol (2005); Stern (2007); Agrawala et al. (2011). b.  See, for example, Bosello, Roson, and Tol (2006); Bosello, Eboli, and Pierfederici (2012). Figure 2. Impacts of Climate Change on Crop Yields by 2050 Percent change in yields in 2050 relative to current climate 40 30 20 10 0 –10 –20 –30 –40 –50 –60 ASEAN 9 Other OECD South Africa Chile Korea China Caspian region Other Europe Brazil Canada Mexico Middle East USA India Aus. & New Z. Other Lat.Am. Other OECD EU Non-OECD EU Other Asia Other Africa EU large 4 Russia Indonesia Japan North Africa OECD America OECD OECD Rest of Europe & Asia Latin Middle South and South- Sub Europe Pacific America East & East Asia Saharan North Africa Africa Rice Other grains Sugar cane and -beet Plant fibres Wheat Fruits and vegetables Oil seeds Source: OECD 2015a. 12 State of Knowledge on Climate Change, Water, and Economics including Central American and Mexico, Sub- total loss of GDP due to changes in agriculture is around Saharan African countries, some parts of the Middle 0.7 percent of GDP in 2060, but with notable difference East, and a large part of South and Southeast Asian between countries and regions, as shown in figure 3. countries. Some regions experience large increases There are small gains for two countries (Russia and in paddy rice yields. In particular, the highest gains Chile), and modest losses for all the other countries are estimated for the southern parts of Latin America and regions, with the largest impacts in “Other Africa,” (particularly Chile), in large parts of Africa (includ- India, North Africa, and the Caspian region. The coun- ing Morocco, South Africa and other Sub-Saharan tries and regions with positive impacts are also the African countries), and in parts of Eastern Europe ones with the expected improvements in river water and continental Asia. Such heterogeneity in impacts flow (Chile and Russia), while some of those with neg- suggests that climate change will cause major alter- ative impacts are also the ones that will be water ations in trade patterns in widely traded commodi- stressed (India and the Middle East). On the other ties such as rice. hand, some of the countries and regions expected to be water stressed, such as China, have only a modest loss Changes in yields of wheat by 2050 are somehow of agricultural output. less  differentiated, and most regions are negatively affected. The most severe negative impacts take place These predictions are of course subject to considerable in Mexico, Western and Eastern Africa, some Southern uncertainty for several reasons, especially stemming African countries such as Namibia and Lesotho, the from the assumed level of climate sensitivity as well as Middle East, South and Southeast Asia, and some the model that predicts impacts due to changes in Western European countries, such as Belgium, the water stress. Nevertheless, the range of damage esti- Netherlands, and Germany. While these are the most mates when taking account of these factors remains affected regions, negative impacts are widely spread relatively small up to 2060; the upper bound is at most and are also present in most of Europe, continental three times the lower bound. This level of uncertainty Asia, and North America. In a few regions, wheat yields increases, however, further out in time, when emis- are positively affected by climate change, particularly sions scenarios also start to play a bigger role. The those with cold climates such as Canada, Russia, and impact of longer time periods is discussed at the end of the Scandinavian countries, most of Central America, this section. Argentina, some countries in Eastern Europe and These estimates do not cover all the factors affecting continental Asia, and a few African countries. ­ agriculture. As noted, an important factor is the The estimated effects on GDP in 2060 are calculated as effect  of higher carbon dioxide (CO2) concentrations a percentage of “no damage” or baseline GDP in that on crop growth—the so-called CO2 fertilization effect. year, which, of course, also needs to be estimated. The Including such an effect, which has been highly latest set of baseline growth estimates prepared by the debated, would lead to higher agricultural productiv- community working on these issues (O’Neill et al. ity, especially for wheat and soybeans, although less 2012) covers a range of growth rates by region, but they so for maize. Rosenzweig et al. (2013) find “approxi- are all positive and imply a world that is considerably mately ±10 ­ percent yield change” by the end of the more affluent, with a minimum annual growth rate of twenty-first century from CO2 effects across a range of 2.2 percent and a maximum of 4 percent between 2010 models and climate scenarios, but also note that there and 2060. The OECD study reviewed here took an is wide variation between models and that “crop intermediate value of 2.8 percent for world growth model parameterization of CO2 effects remains a cru- during this period. With such a baseline growth, the cial area of research.” State of Knowledge on Climate Change, Water, and Economics 13 Figure 3. Changes in GDP Due to Climate Change in 2060 by Region and Impact Category Canada OECD America Chile Mexico USA EU large 4 OECD Pacific OECD Europe Other OECD EU Other OECD Aus. & New Z. Japan Korea China Europe and Asia Non-OECD EU Rest of Russia Caspian region Other Europe Brazil Africa America Latin Other Lat. Am. Middle East & North Middle East North Africa ASEAN 9 Southeast Asia South and Indonesia India Other Asia South Africa Saharan Africa Sub- Other Africa –5.0 –4.0 –3.0 –2.0 –1.0 0 1.0 2.0 Agriculture Ecosystems Health Coastal zones Extreme precipitation events Energy & Tourism demand Source: OECD 2015a. Note: ASEAN = Association of Southeast Asian Nations; Aus. & New Z = Australia and New Zealand; EU = European Union; Lat. Am. = Latin America; OECD = Organisation for Economic Co-operation and Development. 14 State of Knowledge on Climate Change, Water, and Economics Livestock is also likely to be considerably impacted Impacts are then assessed in terms of both physical by climate change, but is not covered in this analysis. losses (sq. km of land lost) and economic costs (value It is an important part of the agricultural sector for of land lost and adaptation costs). both OECD and non-OECD countries. Although the The regions that are most affected by sea level rise are effects of climate change on livestock are much less those in South and Southeast Asia, with the highest exhaustively explored than crop production, the larg- impacts in India and “Other Asia.” The projected est part of the literature finds negative effects of cli- land  and capital losses—expressed as a percentage mate change, not least through heat and water stress of  total regional land area in 2060 with respect to on animal growth, animal health, and the commodi- the  year  2000—are, respectively, -0.63 percent for ties they produce, such as dairy (IPCC 2014). There is, India and ‑0.86 percent for the Other Developing Asia. however, a lack of studies with a global coverage of Other countries in the region are also affected, but the impacts of climate change on livestock produc- to  a smaller extent. Some impacts are also felt in tion (IPCC 2014). North America, with Canada, Mexico, and to a lesser Similarly, information on the effects of warming and extent, the United States being affected. Canada has other climatic drivers on aquaculture is limited. highest loss in land (and capital) in this region the ­ Pickering et al. (2011) conclude that climate change will percent in 2060 with respect to 2000). Smaller (-0.47 ­ likely be beneficial for freshwater aquaculture, except impacts occur in the Middle East (-0.35 percent) and in in coastal zones. No comprehensive economic study Europe, where the highest impacts are felt in the on the impacts of climate change on changes in aqua- aggregate non-OECD Europe region (-0.37 percent), culture productivity currently exists. which includes Israel, Norway, and Turkey. Other world regions, such as Africa, South America, and con- tinental Europe, are on balance hardly affected by sea Coastal Land Losses level rise. Estimates of loss of Coastal land losses due to sea level rise are included in GDP due to coastal zones are the economic modelling of the OECD study as changes figure 3. shown in ­ Global sea-level rise threatens in the availability of land, as well as losses to physical to flood costal lands in many This estimate of the impacts of capital. Because information on capital losses is not regions of the world, destroying climate change on coastal readily available, changes in land and capital stock are natural capital, and potentially zones leaves out the loss of a approximated by assuming that changes in capital ser- swallowing up entire island number of nonmarket goods nations in the Pacific Ocean. vices match land losses, measured as a percentage and services that these zones change from baseline. provide. Although scenic and recreational benefits are Estimates of coastal land lost to sea level rise are partly captured in the values of coastal properties, not based  on the DIVA model outputs (Vafeidis et al. all such benefits are capitalized in this way. OECD 2008). DIVA is an engineering model designed to (2015a) also notes that sea level rise might also lead to address the vulnerability of coastal areas to sea level the loss of entire nation states and their distinctive cul- rise and other ocean- and river-related events, such tures. Low-lying island states such as the Maldives, as storm surges, changes in river morphology, and Kiribati, Palau, the Seychelles, and Tuvalu are particu- altered tidal regimes. It is based on a world database 8 larly at risk of being completely flooded. Apart from of natural system and socioeconomic factors for some case studies, evidence on the magnitude of these world coastal areas, reported with spatial details. impacts in economic terms is very limited. State of Knowledge on Climate Change, Water, and Economics 15 Shifts in Fisheries Economic valuations of other water-related ecosys- tem services, such as wetlands, mangroves, coral A major water-related impact reefs, and rivers and lakes (OECD 2015a), can be con- As ocean waters warm, fishery of climate change is on the fish- ducted by using a modified willingness to pay (WTP) ecosystems adapt and change. eries sector, where there are approach (Bosello, Eboli, and Pierfederici 2012). The This will lead to global changes some estimates of changes in WTP to avoid a given loss in ecosystems is used to in the fisheries sector, with some catch potential. This is mod- approximate the lost value in case these habitats are warmer regions, such as North elled as a change in the natural not protected. This is, for instance, the methodology Africa and Indonesia seeing resource stock available to fish- output decline, and some colder applied in Stanford University’s MERGE model ery sectors, which affects the regions like Russia and North (Manne, Mendelsohn, and Richels 1995).9 In this output of that sector. The input America benefiting. approach, the monetized ecosystem losses related to data uses results from Cheung temperature increase above preindustrial lev- a 2.5°C ­ et al. (2010), which estimate els is found to be approximately equal to 2 percent of maximum catch potential as dependent upon primary GDP when per capita income is above $40,000. This production and  distribution. It  considers 1,066 spe- calibration to 2 percent GDP loss for 2.5°C tempera- ­nvertebrates. Future pro- cies of exploited fish and i ture increase represents the U.S. Environmental jected changes in species ­distribution are simulated by Protection Agency (EPA) expenditure on environmen- using a model (Cheung et  al. 2008, 2010) that starts tal protection in 1995. The strong implicit assump- with identifying species’ preference for environmental tions are that what is actually paid is reasonably close conditions and then links them to the expected carry- to the WTP, and roughly sufficient to preserve ecosys- ing capacity. The model assumes that carrying capac- tems and their services in a world with moderately ity varies positively with habitat suitability of each increasing temperatures. Such an approach is also spatial cell. Finally, the related change in total catch only partial; it does not pick up other losses of ecosys- potential is determined by aggregating spatially and tem services, such as those arising from the extinc- across species. tion of certain species. The input data for the fisheries sector is the percentage change in fish catch with respect to the year 2000. By 2060, the countries with the highest WTP for all The  most negatively affected regions by 2060 are these ecosystem services are mainly the largest econ- North Africa (−27 percent) and Indonesia (−26 ­percent). omies, including Canada, Japan, Korea, Mexico, Some European countries, the Middle East, Chile, and South Africa, and the United States, as well as many several countries in Southeast Asia have impacts rang- European countries. These countries are willing to ing from −10 percent to −15 percent. Smaller negative pay around 1 percent of GDP to protect these impacts also take place in China, Korea, Brazil, and ­ ecosystem services. The WTP is smaller (between other Latin American countries, Mexico, and some percent and 0.7 percent of GDP) in much of Latin 0.3 ­ European countries. In some countries, fish catches America, China, Russia, the Middle East, and in OECD increase. The highest increases will occur in Russia EU regions. Other regions have very small WTP; the (+25 percent) and in the five major European econo- smallest is in the group of Sub-Saharan African mies (+23 percent). Small positive impacts are seen in countries. This distribution of WTP values is not ­ the United States, Canada, Oceania, and the Caspian surprising because there is a strong connection ­ region. Other world regions (India, other developing between the willingness to pay for ecosystem ser- countries in Asia, South Africa, and the rest of Africa) vices and average income in certain countries. Hence, are mostly unaffected. it is natural that the WTP for ecosystem services 16 State of Knowledge on Climate Change, Water, and Economics is higher in high-income countries such as the United conditions not related to climate change. However, States or Canada, while being much lower in other these damages are projected to double again by the areas of the world, such as continental Africa. end of the century due to climate change: that is, they are estimated to increase by a factor of 4. Most addi- The total value of these losses in terms of GDP is shown tional climate-­ induced damages are predicted to take in figure 3 (which includes losses from both fisheries place in North America, East Asia, and the Caribbean– and other ecosystem services). They are not large in Central American region, where the United States, any country or region, but are notable in Australia and Japan, and China will be most affected. New Zealand, the ASEAN 9, Canada, China, the largest four EU states, Indonesia, Korea and the United States These estimates do not include damages from and the. extreme events in the form of illness, disease and pre- mature mortality. So far, there is no robust database to Nevertheless, these estimates are only lower bounds, capture the impacts of climate change on river and as they do not take into account non-use values of eco- other floods. As OECD (2015a) observes, flood risk mod- systems and potential biodiversity loss, which could els exist, which calculate indicators such as area at risk be a consequence of climate change. The assessment of flooding and population at risk of flooding, but these has also not accounted for possible changes in forest do not easily translate into economic costs. A groups of areas and the associated changes in services from the models, including GLOFRIS (Ward et al. 2013, 2014), related ecosystems. can be used to compare projections of future flood risks with and without climate change (Winsemius and Ward Increasing Flood Damages 2015) and thus establish the additional damages due to climate change. Using the There are many types of extreme events and they framework of Ward et al. (2013), affect the economy in different ways. Given the uncer- Although the economic costs the OECD (2015a) has estimated tainties involved in predicting the frequency and dam- of tropical storms attributable the excess urban flood damages ages caused by these events, however, and the to climate change are from climate change. difficulties in attributing such events to climate relatively small, flood damages change, the available data on how the economy will be Map 2 shows the projected are expected to increase significantly in some regions, affected are still scarce. The assessment by Mendelsohn urban climate damages from with global urban damage et al. (2012) has provided some quantitative assess- floods for 2080 under the RCP estimates potentially reaching ment and projections on damages from hurricanes. 8.5 scenario, for data aggre- $1.8 trillion by 2080. The authors stress that the regional damages are quite gated to the country level.10 climate model that is used to project sensitive to the ­ The two countries with by far future climate ­ conditions. They find overall that cli- the largest projected urban flood damages are India mate change is predicted to increase the frequency of and China. The main driver for this is the huge high-intensity storms in selected ocean basins; the increase in the urban assets that are exposed in these predicted extent depends on the climate models used. countries. The scale of flood risks is so large in these In value terms, Mendelsohn et al. (2012) find the cur- countries that the additional damages from climate rent global damage from tropical cyclones to capital change are also huge. Bangladesh is also high in the stocks to be around $26 billion, which is equivalent to ranking of most affected countries, but in this case 0.04 percent of global GDP. That is expected to roughly the role of climate change is substantially larger. The double by the end of the century under current climate opposite is true for Indonesia, Russia, Thailand, and conditions, due to changes in socioeconomic the main Nile countries, where flood risks are State of Knowledge on Climate Change, Water, and Economics 17 Map 2. Urban Climate Change Damages from Floods by 2080 IBRD 42868 | APRIL 2017 Below −1 From −1 to 0 From 0 to 1 From 1 to 10 From 10 to 100 Above 100 No data Source: OECD 2015a, based on Winsemius and Ward 2015. currently relatively high, but the additional damages can only be considered as indicative. As table 2 from climate change are projected to be negative. For shows, there are significant differences depending OECD countries, the climate-induced urban flood on which climate models are adopted to make these damages are limited to less than $50 billion a year by projections. For instance, only the HadGEM model 2080. That does not mean that total urban flood dam- projections imply a reduction in urban flood dam- ages— either climate-induced or not—are much ages in Indonesia; the other models all predict smaller than in non-OECD regions. For example, the increased damages for this country. For the OECD annual damages by 2080 amount to $170 billion in the region, the largest uncertainty is in the projections United States, $58  billion in Mexico, $20 billion in for Mexico. Nonetheless, there are also some consis- Germany, and $17 billion in the Netherlands. But the tent patterns across the models, including the fact climate-induced component of these damages is sub- that the largest climate-induced urban flood dam- stantially smaller than for many non-OECD ages are in Asia in general and in India in particular, countries. and that the flood damages in Russia decrease due to climate change. Using climate scenarios from the Given the importance of the projected regional pre- HadGEM model, global annual urban flood damages cipitation patterns for these simulations, and the are projected to amount to between $0.7 and $1.8 large uncertainties surrounding them, these results trillion in 2080. 18 State of Knowledge on Climate Change, Water, and Economics Table 2. Urban Flood Damages by Region and Model ($US billion, 2005 PPP exchange rates) HadGEM GFDL IPSL MIROC NorESM 2010 2030 2080 2080 OECD America Canada 0.0 0.5 0.0 1.5 −3.6 1.9 0.8 Chile 0.0 0.3 2.0 −3.4 −3.1 −3.4 −0.4 Mexico 0.0 −0.5 0.7 66.3 −49.7 −15.6 −29.9 USA 0.0 2.3 19.4 10.2 16.6 5.4 3.5 OECD Europe EU large 4 0.0 1.9 11.2 0.8 3.9 4.8 2.2 Other OECD EU 0.0 1.6 8.8 1.6 5.8 4.6 2.6 Other OECD 0.0 −0.2 1.5 −6.2 −5.2 −4.5 0.1 OECD Pacific Aus. & New Z. 0.0 −0.3 1.3 −4.2 1.4 1.2 0.4 Japan 0.0 0.6 3.4 2.6 1.2 0.9 1.5 Korea 0.0 0.2 0.9 1.1 0.4 2.0 0.7 OECD 0.0 6.2 49.2 70.3 −32.1 −2.8 −18.5 Rest of Europe and Asia China 0.0 48.0 427.9 343.0 88.8 102.5 184.4 Non-OECD EU 0.0 −0.8 −3.6 −1.7 −2.3 4.4 0.7 Russia 0.0 −5.4 −32.6 −7.6 −7.8 −4.7 −44.8 Caspian region 0.0 1.9 17.6 2.6 −4.6 2.9 −6.4 Other Europe 0.0 −2.6 −13.5 −7.9 −6.8 2.1 −12.6 Latin America Brazil 0.0 0.9 12.6 6.7 98.1 −15.1 −40.3 Other Lat. Am. 0.0 −0.7 15.2 −10.5 10.5 −16.6 −26.9 Middle East & North Africa Middle East 0.0 −0.3 39.8 −60.9 −32.2 −34.2 9.4 North Africa 0.0 −2.5 −44.9 128.0 243.2 47.2 25.0 South and Southeast Asia ASEAN 9 0.0 −0.7 65.1 185.2 139.1 57.9 196.1 Indonesia 0.0 −2.7 −29.0 5.2 152.8 11.2 38.4 India 0.0 51.5 1,094.9 432.7 718.3 362.2 207.8 Other Asia 0.0 2.4 184.0 153.9 148.9 117.8 114.1 Sub-Saharan Africa South Africa 0.0 0.1 3.3 4.8 2.2 −2.0 −1.4 Other Africa 0.0 3.1 59.3 85.6 178.4 225.1 76.1 World 0.0 98.4 1,845.3 1,329.4 1,694.3 857.9 701.2 Source: OECD 2015a, based on Winsemius and Ward 2015. Note: HadGEM, GFDC, IPSL, MIROC, and NorESM are specific climate models that are used to project precipitation and temperature patterns (see Winsemius and Ward 2015 for more details). PPP = purchasing power parity. ASEAN = Association of Southeast Asian Nations; Aus. & New Z = Australia and New Zealand; EU = European Union; Lat. Am. = Latin America; OECD = Organisation for Economic Co-operation and Development. State of Knowledge on Climate Change, Water, and Economics 19 It is also important to note that these calculations pro- The costs of extreme events in the form of deaths, vide an estimate of potential damages, without any illnesses, and injuries from floods and sea level rise ­ adaptive behavior to deal with increased flood risks. due to climate change are not available on a systematic Hence, the numbers presented here should be inter- basis. According to the EMDAT disaster database,11 preted as an upper bound of the costs that will occur there have been an annual average of 16,117 deaths when such adaptation is taken into account. On the other worldwide between 2000 and 2014 from floods and hand, urban damages are only one element of flood dam- storms. In addition, an average of 75.4 million people ages, and the local disruption effects are excluded here. have been affected every year. Most, but not all, of These are likely to have severe consequences for local these impacts have been in developing countries. In communities, even if their economic effect may be rela- the future, with climate change such losses can be tively small (compared to the ­ damages from hurricanes). expected to increase, but reliable estimates are not available. Hinkel et al. (2014) estimate the number of people who may be displaced during this century due Increases in Water-borne Diseases and Water-related Deaths and Injuries to sea level rise under different modelling assump- tions, but do not estimate deaths and injuries. There The health impacts of climate change are generally are also estimates of people at risk from flooding for divided into those arising from climate-related dis- some regions,12 but they also do not estimate deaths or eases, effects of heat stress, and the effects of extreme injuries. Even if estimates are available, converting events. Effects that can be attributed to heat stress these into monetary terms would require valuing loss and climate-related diseases are not particularly of life, which is controversial. related and they are not discussed ­ water-­ further, except for diarrhea. For  water-borne diseases, esti- mates of the costs are generally measured in terms of Increase in the Cost and Decrease in Reliability of Power Supply the additional expenditures needed to treat the cases aris- Water plays an important role in the supply of energy, Climate change will have water- ing from climate change. Ebi not only in the generation of hydropower but also related impacts on health, (2008) estimated the costs of in  providing an input into the generation of thermal with increases in water-borne specific interventions for and  nuclear power. A number of studies have high- diseases and deaths and injuries treatment of additional cases lighted the increased cost of reduced water availability caused by flooding. of malaria, diarrhea, and mal- from climate change for supplying electrical energy. nutrition that are expected to Under higher temperatures, the efficiency, output, and occur between 2000 and 2030 because of climate reliability of thermal power plants is expected to suffer change. Her projections show increases of 5 percent as a consequence of reduced water volume and higher in malaria disease, 3 percent in diarrhea, and 10 per- water temperature—two factors that are crucial for cent in malnutrition. Ebi’s projections for  diarrhea cooling of most of these plants (alternative processes, are a little lower than those of Kolstand and Johansson such as dry cooling, typically consume more electricity (2011), who project an increase of 8–11 percent in the and require higher investment costs). Climate change risk of diarrhea in the tropics and subtropics due to could raise the costs of power plants in areas where cli- climate change using the A1B scenario (see descrip- matic factors increase water scarcity. China could be tion above in footnote 1). Ebi estimates the additional particularly affected by this development, given that annual costs of treating diarrhea in 2030 to be much of the existing and planned coal power capacity between $1.7 and $9.0 billion. is located in regions with high risks of water stress. 20 State of Knowledge on Climate Change, Water, and Economics Cost increases in India are expected to be smaller, prices are projected on a mean annual basis for most given that Indian coal mines, power stations, and European countries (except for Sweden and Norway), industrial demand are located mostly in areas with with the strongest increases for 2031–60 for Romania lower risks of water scarcity (IEA 2015; WRI 2014). (31–32 percent), Bulgaria (21–23 percent), and Slovenia A  case study by Hurd et al. (2004) has assessed the (12–15 percent), where limitations in water availability likely welfare costs of climate change impacts on water mainly affect power plants with low production costs. use in electric power generation in the United States, The implications of water scarcity will need to be projecting losses of about $622 million per year up to taken into account more fully when choosing energy 2100 due to changes in cooling water for combustion in options for mitigating CO2. The World Bank’s Energy coal, natural gas, and other thermal power stations. Sector Management Assistance Program (Ebinger and The study assumed warming of +2.5°C above prein- Vergara, 2011) produced figure 4, which shows both dustrial levels and a drop of 10 percent in monthly the water intensity (M3/kWh) and CO2 (kg/kWh) inten- average precipitation. Water shortage can also nega- sity of different electricity options. Several low-carbon tively affect the operation of hydropower plants (IPCC options such as nuclear, solar, thermal, and geother- 2014). mal have a high water intensity, which may determine There is also an issue with hydropower, where low where they can be located. flow rates will create difficulties in maintaining the current and proposed levels of generation. In Europe, for example, recent warm, dry summers have shown Stress on Municipal Water Supplies and Quality the vulnerability of the European power sector to low water availability and high river temperatures. Climate In addition to negative eco- change is likely to affect electricity supply, in terms of nomic consequences for A changing hydrological both water availability for hydropower generation water-intensive economic environment will also put stress and  cooling water usage for thermoelectric power activities, reduced water avail- on municipal water supplies production. Van Vliet, Vogel, and Rubbelke (2013) ­ ability from climate change will and affect water quality at the estimate the impacts of climate change and changes in ­ likely also be felt by house- household level. water availability and water temperature on European holds and municipalities through impacts on the avail- electricity production and prices. Using simulations of ability and quality of drinking water (OECD 2013, daily river flows and water temperatures under future 2015a; IPCC 2014). However, economic values of these climate in power production models from 2031 to impacts are not available. 2060, they show declines in both thermoelectric and hydropower generating potential for most parts A recent study by Henderson et al. (2015) attempted to of  Europe, except for the most northern countries. estimate the economic impacts of climate change on Gross hydropower potential of Europe is estimated to water resources in the United States, covering several decrease on average by 4–5 percent for 2031–60 (SRES other types of water use beyond irrigation and cooling. B1–A2) relative to 1971–2000, with decreases of around The study suggests annual damages of approximately 16–20 percent in Bulgaria and 15–21 percent in Spain. $2.1 billion by 2050 and $4.2 billion by 2100 without new Based on changes in power production potentials, they ­ olicies. The largest impacts are pro- climate change p assess the cost-optimal use of power plants for each jected to affect non-consumptive activities, such as European country by taking electricity import and hydropower and environmental flows. Agriculture and export constraints into account. Higher wholesale other consumptive uses will be impacted by climate State of Knowledge on Climate Change, Water, and Economics 21 Figure 4. Water and Carbon Intensity of Different Fuel Cycles 1.2 Evaporation recapture Dry condenser Hybrid 1.0 Cooling Open loop Coal: Closed loop Blowdown recycling 0.8 IGCC Carbon intensity (kg/kWh) Blowdown 0.6 recycling Hybrid Dry condenser Cooling Open loop 0.4 Evaporation Inlet cooling recapture High temperature 0.2 Solar Thermal: Geothermal Nuclear: Closed loop Closed loop Hydro- Photovoltaic Wind electric Dry condenser Open loop 0 Gen IV 1.E-06 1.E-05 1.E-04 Dry condenser 1.E-03 Pond 1.E-02 Water footprint (m3/kWh) Nonrenewable sources Coal Natural gas PV Solar thermal Renewable sources Nuclear Geothermal Wind Hydroelectric Low current availability High current availability Source: Ebinger and Vergara 2011. change less negatively. Similarly, Strzepek et al. (2014) The Uncertainty of Cost Projections suggest negative welfare consequences for the United The expected damages discussed in the preceding sec- Sates in the order of $6.5 to $15 billion by the end of the tions are small relative to the expected GDP in 2060; century in their assessment of they amount to about 1.5 percent of GDP in that year. the impacts of climate change on That figure, however, is misleading in a number of water supply, management, and Although climate and economic respects. First, as can be seen from figure 2, there are use of water resources. For the models universally agree that significant variations between regions, with Northern damages from the effects of year 2050, results are more Europe and North America having much lower dam- climate change on water will be ambiguous, with one scenario ages and South Asia and Sub-Saharan Africa having costly, the precise magnitude suggesting positive effects on much higher damages. of the costs are uncertain and welfare from climate change and projections should be treated others suggesting negative two ­ Second, there is considerable uncertainty in the esti- with caution. effects. mates. If the upper bound turns out to be right, the 22 State of Knowledge on Climate Change, Water, and Economics figures could be two to three times higher. A key source ­ is thorough their direct impact on growth. These are of uncertainty is the equilibrium climate sensitivity examined in the next section. (ECS): the amount by which temperature will increase The damages or costs of with a doubling of GHG concentrations. inaction in the face of climate ­ When estimating climate Third, the process for making the estimates is strongly change will not stop in 2060. change damages beyond 2060, driven by the underlying growth in the economy, The modelling indicates an uncertainty increases, but it is which is assumed to be around 2.8 percent per year. All increase in the rate of damage likely that the rate of damage the models that make projections to 2050 and beyond after that date, at the same will accelerate. assume some global growth rates of at least 2 percent time as an increase in the level per year, which decreases the importance of more of uncertainty in the projections and underlying data. ­climate-related sectors such as agriculture in the struc- This applies to all sources of damages, including those ture of the economy. Furthermore, scarcity of water in related to the water sector. Figure 5 shows the annual the coming decades may make these projections infea- damages from 2010 to 2100 relative to a no-damage sible, as discussed. baseline for all damages (not just from water-related sectors). The main estimate increases from around Fourth, the CGE model assumes relatively easy substi- 2 percent of GDP in 2060 to around 6.8 percent by the tution between factors. When there is a shock and an end of the century. These projections are based on the input such as land or water is reduced, the model AD-DICE model, which is one of the IAMs with esti- assumes that the input can be replaced with other fac- mates in the middle of the range generated by such tors such as capital, and any displaced labor can be models.13 In the long term, the equilibrium climate absorbed by other sectors of the economy in a painless sensitivity (ECS) has even more impact on the esti- fashion. In practice, such substitution will take time mated damages. Within the range of values of this and will involve transition costs that are not accounted parameter, the corresponding range of damages for. Some ongoing work indicates that allowing for by 2100 is between 2 percent and 10 percent of GDP. imperfect substitution will raise the cost of climate With a wider interval for ECS that is plausible, dam- change, but the adjustment does not appear to be huge. ages can vary from 1 percent to 15 percent of GDP, with Fifth, the scenarios modelled in the CGE part of the a corresponding range of temperature increase under analysis are based on relatively small increases in tem- inaction of 2.4 to 5.5 degrees Celsius (OECD 2015a). perature (around 2 degrees Celsius). If higher increases The figure also shows the extent to which damages occur, then the assumptions of the model in terms of beyond 2060 are committed as a result of emissions climate impacts may not hold, and there is no experi- produced by 2060. If emissions are assumed to stop ence on which to base the sectoral changes. These completely in 2060, damages will continue to 2100 as higher increases are discussed in the next section. a result of the inertia in the system. This is especially true for sea level rise damages, which respond very Lastly, the estimates of losses are incomplete. They do slowly to a change in emissions. Consequently, dam- not include all impacts, nor  do they value losses of ages that were around 2 percent of GDP in 2060 will life. All these factors would  raise damage estimates rise to around 3  percent by 2100. Thus, the case for from climate change through a number of pathways, action on emissions now is partly to forestall damages of which water is an important one. in the future. Alternative ways to look at the link between factors As noted, these long-term estimates are uncertain such as water and the economy under climate change because of the unknown ECS, but also because the State of Knowledge on Climate Change, Water, and Economics 23 Figure 5. Climate Change Impacts in the Very Long Run 0 –2.0 –4.0 Percent –6.0 –8.0 –10.0 –12.0 20 60 00 30 40 50 70 80 90 10 35 45 55 65 75 85 95 15 25 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 21 Likely uncertainty range - Full damages Likely uncertainty range - Committed by 2060 Central projection - Full damages Central projection - Committed by 2060 Source: AD-DICE model. damage functions linking temperature and precipita- in potential d ­ amages by 2100 that increases from the tion to damages are crude and not well determined figure of 6.8 percent previously discussed to as much (up to 2060, the use of such a function is avoided in as 18 percent of GDP. the reported figures by using direct links between A main message from the longer-term analysis there- impacts and sector-level production functions).14 As fore is that the costs of inaction in terms of economic noted, continuing the projections to 2100 with no damages rise substantially beyond 2060, as the world action may result in very high increases in tempera- approaches 2100. At the same time, so does the uncer- ture. There is no basis for estimating damages for such tainty, making the case of action now much more one scenarios. An alternative to the “standard” function of applying the precautionary principle. The second used in the AD-DICE model is one proposed by message is that action now will affect the future path Weitzman (2013), in which damages at high tempera- of damages to a significant extent because of the iner- ture increases are modelled using higher power terms tia in the system. This also makes a case for reducing in the damage function. The result is an increase emissions significantly in the period to 2060. 24 State of Knowledge on Climate Change, Water, and Economics Other Models Another interesting study that  looks at the historic Several other types of models An alternative approach to looking at the nexus of link between economic perfor- exist for predicting climate water, climate change, and the economy is through the mance and growth is Hsiang change and water’s effect effects of climate change on water and thereby on the and Jina (2014). Using meteoro- on future economic growth, growth of the economy. There is an emerging literature logical data, they reconstruct which provide alternative and that takes the view that an important pathway for cli- every country’s exposure to the sometimes different predictions. matic impacts is through the damage they do to the universe of tropical cyclones capital stock and to the capacity of the economy to from 1950 to 2008 and estimate the causal effect such increase total factor productivity and thereby growth. events have on long-term growth. Comparing each This literature is in part motivated by empirical studies country’s growth rate to itself in the years immediately that show how historic climate-related events have in before and after exposure, the data reject the hypothe- fact reduced growth in some economies. Dell, Jones, sis that disasters stimulate growth or that short-term and Olken (2012) were the first to investigate the effects losses disappear following migrations or transfers of of climate variability (temperature and precipitation), wealth. Instead, they find “robust evidence that on GDP per capita, agricultural value added, industrial national incomes decline, relative to their pre-disaster value added, and investment. The study also assesses trend, and do not recover within twenty years.” The the lagged effects of climate variability for periods of conclusion holds both for developed and developing 1, 5, and 10 years. It finds that for Sub-Saharan Africa, countries. They note: “Income losses arise from a small a one-degree increase in temperature has been statisti- but persistent suppression of annual growth rates cally significantly associated with a 1.8 percent decline spread across the fifteen years following disaster, [gen- in the growth rate. In poor countries, as a group, the erate] large and significant cumulative effects: a 90th effect of a one-degree increase is to reduce growth by percentile event reduces per capita incomes by 1.4 percent. percent two decades later, effectively undoing 7.4  ­ 3.7  years of average development.” Based on these While Dell, Jones, and Olken (2012) do not find a signif- results, they estimate that projections of future cyclone icant link between precipitation and GDP growth, activity would result in a discounted present value cost Brown et al. (2013) do. They apply a similar regression that is about $10 trillion larger than previous estimates. model, allow for temporal and spatial variation in pre- cipitation, and use a more sophisticated index for pre- These studies are valuable in pointing to a pathway by cipitation variability. The analysis for a panel dataset which climate change could have bigger impacts and of 133 countries showed that an increase of 1 percent need to be taken seriously. Yet there are some difficul- in  the area of a country exposed to a drought for a ties with them. First, the econometric estimates can be given period results in a decline in GDP per capita of criticized on a number of grounds, such as: using linear 2.7 percent, and an increase of 1 percent in the area of functions where nonlinearity is more likely to be the a country exposed to a flood is associated with a reduc- case; not taking account of year-to-year variability of tion in GDP per capita of 1.8 percent. More recently, temperature and rainfall (in the case of Dell, Jones, and Moore and Diaz (2015) take the results from Dell, Jones, Olken 2012); and not allowing for individual country and Olken (2012) and use them to project growth rates effects. Second, in terms of economic structure, they with different mitigation polices, concluding that do not indicate how the causality works. This leaves growth could be significantly reduced with a business- open the possibility that the common observations of as-usual policy. climate variability and GDP changes are the result of State of Knowledge on Climate Change, Water, and Economics 25 some other factors such as a structural shift to lower place at two different levels: competition for water growth in Sub-Saharan Africa countries during a period within river basins and competition for land within when an increase in temperatures has been observed. agro-ecological zones (AEZs). This design significantly More work is needed to confirm these results and to improves the adaptability of the model. For example, understand the causality. the irrigated and rain-fed production functions operate independently from one another. That means irrigated One way to understand the growth impacts of climate crop production can be completely removed from a cer- change is to link the changes caused by climate not tain part of the country if water supply for irrigation only to flows of goods and services but also to the cap- falls short. Moreover, in GTAP-BIO-W, intersections ital stock and the productivity of that stock. The latter between different river basins and agro-ecological can come through changes in the rate of technological zones are featured by different technologies (produc- change of the economy, measured in terms of total fac- tion functions) that reflect water availability, growing tor productivity (TFP). It is possible to incorporate the conditions, and soil quality peculiar to that area. capital effects by allocating part of the damages esti- mated in the previous section to the capital stock. Because a computable general equilibrium (CGE) An  estimate for the share to be allocated is around model requires detailed specification of individual percent (Dietz and Stern 2014). Alternatively, it can 30 ­ sectors of the economy, it is not possible, for reasons be assumed that this 30 percent of damages affects not explained earlier, to provide long-term projections. the capital stock but the growth rate of TFP. If either Thus the exercise is limited in time—in this case, to path is followed, the level of damages does increase, 2050. In that time frame, the study investigates the but not dramatically. By 2060, expected damages rise consequences of the expected growth in GDP and pop- about 1 percentage point under the assumption that ulation and how economic output by sector may be residual damages are allocated to TFP growth. If they affected by constraints on water supply, including are allocated to the capital stock, the total damage goes those resulting from climate change. up by only 0.2–0.3 percent. By 2100, the differences are more pronounced, with the effects via TFP growth Given the strong baseline projected growth in GDP and damages increasing by about 2 percentage points and population in the region, and the water scarcity that with the effects operating via the capital stock increase has already been noted, it is not surprising that such by about 0.5 percent. scarcity could compromise the baseline projections. The modelling shows that climate change adds to the Another way to look at the growth issue is to model it effects of the expected scarcity of water. With the two explicitly in a CGE model. Some work is ongoing in that taken together, losses of GDP by 2050 relative to the direction. Taheripour et al. (forthcoming) have devel- baseline are now 5.2 percent in Bangladesh; 1.8 percent oped a version of the GTAP model with detailed model- in India; 0.8 percent in Nepal; 5.6 percent in Pakistan; ling of biofuel supply as well as water demand and 0.6 percent in Sri Lanka; and 0.5 percent in Rest of supply through the GTAB-BIO-W Model, and applied it South Asia. If, however, water for agriculture were to to the South Asia region (Bangladesh, India, Pakistan, be available (that is, the non-climatic uncertainty were Sri Lanka and the rest of South Asia).15 This model to be removed), the effects of climate change on GDP retains the multilevel constant elasticity of substitution would be very small; they would be reduced by more (CES) structure and irrigated/rain-fed crop production than an order of magnitude. of the existing GTAP-W model, but overcomes some of its shortcomings. The most marked difference is that The main implication of water scarcity then is one of GTAP-BIO-W permits competition for resources to take addressing the scarcity problem, through improvements 26 State of Knowledge on Climate Change, Water, and Economics in water and land productivity as much as possible. In additional expenditure on a particular type of adap- general, water efficiency use in irrigation is relatively low tation. These levels will depend of course on the in this region. It can be improved substantially, but this level of baseline damages and will vary over time. requires additional investments and changes in water Bosello, Carraro, and De Cian et al. (2013) have sum- allocation rules. marized the outlays on proactive adaptation, reac- tive adaption, and innovation activity for a situation in which concentrations of CO2 double.16 The figures Reducing Future Damage through are  based on the literature. In the case of water, Investment in Adaptation Technology they consist of additional The damages assessed to 2060 and beyond make a costs for agriculture— largely Investments in adaptation strong case for action on adaptation; up to 2060 these for irrigation, for  providing technology and capital can actions are more or less independent of the mitigation water for other uses in areas significantly reduce future policies undertaken, but beyond that date, the mitiga- where scarcity is expected to impacts from water damages tion and adaptation policies are interrelated. The increase, for dealing with induced by climate change. higher the level of mitigation, the lower the damages in flood risks in river systems, the future and the less is needed in the way of adapta- and for expenditure on coastal protection. Their tion expenditure to reduce these damages. The two data are given in table 3, in both billions of U.S. dol- become substitutes for each other. At a more specific lars and as a percent of GDP. level, however, adaptation and mitigation can work The table indicates about 60 percent of all adaptation against each other, particularly over water use. expenditures as allocated for water-related impacts, Adaptation to changing hydrological regimes and water with the greatest amounts going to the Middle East scarcity, for example, takes place through increasing and North Africa (MENA), East Asia, and South Asia, reuse of wastewater and the associated treatment, for agriculture and other vulnerable areas; and to through deep-well pumping, and possibly large-scale Latin and Central America and the Caribbean (LACA), desalination. These adaptation measures increase East Asia, and Western Europe for coastal protection. energy use in the water sector, leading to increased As a percent of estimated GDP in the year of calibra- emissions and mitigation costs (Klein et al. 2007). tion (2050), the costs are modest, ranging from 1.5 Adaptation costs are made up of proactive (or anticipa- percent (MENA) to 0.2 percent (Canada, Japan, and tory) adaptation (taken in anticipation of expected New Zealand, CAJANZ). In absolute terms, they damages often by the public sector), reactive adapta- amount to $613 billion, which is a much greater tion (take after the impacts have occurred so as to mini- amount than the current finance for adaptation—esti- mize their consequences), and innovative activity mated at around $26 billion to $32 billion by Buchner (undertaken to make adaptation responses more effec- et al. (2014)—but recall that the figures in the table are tive) (Bosello, Carraro, and De Cian 2013). The anticipa- for 2060. tory and reactive adaptations are also referred to in the Although the table provides some of the best available literature as stock and flow forms. The former often estimates, it should be viewed as only a rough guide to involve an investment in capital, while the latter consist the likely adaptation needs. The underlying studies are of sector-related periodic expenditures (OECD 2015a). quite crude; they are not based on a detailed bot- Estimates of the amounts of adaptation expenditure tom-up assessment, and for some categories, such as should be determined by the point at which the early warning systems, the figures appear to be merely marginal damages reduced are equal to the a placeholder. Furthermore, it is not clear what is State of Knowledge on Climate Change, Water, and Economics 27 Table 3. Annual Adaptation Costs in Response to a Doubling of CO2 Concentrations in 2060 Water Water in Early in other Coastal Cooling Disease R&D for Total as % agriculture warning Settlements Total vulnerable protection expenditure treatments adaptation of GDP (Irrigation) systems markets USA 5.0 2.1 5.0 3.6 31.3 1.1 2.9 2.9 53.9 0.1 W. Europe 7.8 3.3 5.0 5.0 63.3 −0.7 2.4 2.4 88.5 0.2 E. Europe 12.3 5.3 5.0 0.3 2.4 −0.1 0.0 0.0 25.2 0.7 KOSAU 0.1 0.1 5.0 1.8 3.7 1.9 0.3 0.3 13.2 0.5 CAJANZ 2.7 1.1 5.0 2.9 23.1 3.0 1.7 1.7 41.2 0.2 TE 16.9 7.2 5.0 1.7 2.0 0.1 0.1 0.1 33.1 0.5 MENA 79.1 33.9 5.0 1.2 3.2 2.1 0.1 0.1 124.7 1.5 SSA 16.1 6.9 5.0 2.7 4.0 0.5 0.0 0.0 35.2 0.9 SASIA 28.4 12.2 5.0 1.3 12.8 1.1 0.0 0.0 60.8 0.6 CHINA 12.5 5.4 5.0 1.3 9.7 0.3 0.2 0.2 34.6 0.3 EASIA 31.2 13.4 5.0 4.3 6.0 4.7 0.0 0.0 64.6 0.9 LACA 7.2 3.1 5.0 7.7 15.0 5.7 0.1 0.1 43.9 0.2 Total 219.3 94.0 60.0 33.8 176.5 19.7 7.8 7.9 619.0 0.4% As % 35.4% 15.2% 9.7% 5.5% 28.5% 3.2% 1.3% 1.3% 100.0% Source: Adapted from Bosello, Carraro, and De Cian (2013). Note: CAJANZ = Canada, Japan, New Zealand; EASIA = East Asia; KOSAU = Korea, South Africa, Australia; LACA = Latin and Central America and the Caribbean; MENA = Middle East and North Africa; SASIA = South Asia; SSA = Sub-Saharan Africa; TE = transition economies. assumed about future changes in the pattern of use of it. The marginal costs of reducing damages through existing resources such as water. Earlier, it was adaptation increase as the percentage rises. The esti- observed that major inefficiencies in water use exist, mates in table 3 are based approximately on the level and that removing these would increase water avail- at which the marginal reduction in damages is equal to ability significantly, without further investment in irri- the marginal cost. Table 4 shows the reduction in dam- gation. The adaptation options included in most age that is estimated to be achieved from the expendi- analysis tend to focus on engineering alternatives and tures given in table 3. give less weight to changes in behavior and current practices of water use, as well as on other softer alter- For the water sector, damage reductions from the natives involving the use of ecosystems and the like. adaptation expenditures range from 36 percent (water For these reasons, estimates of adaptation cost could in agriculture) to 60 percent (water in other vulnerable be on the high side, although this needs further markets). Reductions are lower from early warning investigation. systems for extreme events. Variations by region show higher levels of damage reduction in developed regions The study by Bosello, Carraro, and De Cian (2013) relative to developing ones. (which uses the adaptation calibration in table 3 in the framework of the AD-Witch model),17 as well as other An alternative way to look at the amount of damages similar studies, note that the amount of damage reduced is to divide adaptation into stock and flow. reduced by adaptation is never 100 percent or close to Stock adaptation refers to measures that require 28 State of Knowledge on Climate Change, Water, and Economics Table 4. Percentage of Damages Reduced as a Result of Adaptation Costs Presented in Table 3 Water in agriculture Water in other Early warning Coastal Disease Weighted Settlements (Irrigation) vulnerable markets systems protection treatments total USA 48.0 80.0 10.0 75.0 40.0 90.0 25.0 W. Europe 43.0 80.0 10.0 54.0 40.0 90.0 20.0 E. Europe 43.0 80.0 10.0 63.0 40.0 60.0 34.0 KOSAU 27.0 80.0 10.0 62.0 40.0 81.0 24.0 CAJANZ 38.0 80.0 10.0 37.0 40.0 69.0 25.0 TE 38.0 40.0 10.0 37.0 40.0 70.0 20.0 MENA 33.0 40.0 10.0 55.0 40.0 60.0 38.0 SSA 23.0 40.0 0.1 30.0 40.0 20.0 21.0 SASIA 33.0 40.0 0.1 47.0 40.0 35.0 19.0 CHINA 33.0 40.0 10.0 76.0 40.0 40.0 22.0 EASIA 33.0 40.0 1.0 25.0 40.0 40.0 19.0 LACA 38.0 40.0 0.1 46.0 40.0 90.0 38.0 Average 35.8 60.0 6.8 50.6 40.0 62.1 25.4 Source: Adapted from Bosello, Carraro, and De Cian (2013). Additional material from the report was provided by the authors. Note: CAJANZ = Canada, Japan, New Zealand; EASIA = East Asia; KOSAU = Korea, South Africa, Australia; LACA = Latin and Central America and the Caribbean; MENA = Middle East and North Africa; SASIA = South Asia; SSA = Sub-Saharan Africa; TE = Transition Economies. investments beforehand to build adaptation capital. rising over time, with the share of reduction attributed This adaptation stock reduces the damages of climate to flow adaptation also increasing. In 2015, stock adap- change in the future. Flow adaptation refers to adapta- tation accounts for more than half the reduction in tion measures that do not require investments before- damages, but by 2100 it accounts for only around hand but where benefits are reaped almost 14  percent of the reduction. Total damages reduced instantaneously. Government involvement can facili- rise from around 27 percent in 2015 to 70 percent in tate the efficient application of this adaptation (for 2100. example, by overcoming knowledge barriers), but is The benefits of adaptation, therefore, are to reduce not necessary for its implementation. As a crude damages significantly by the end of the century. approximation, OECD (2015a) assumes stock adapta- Equally valuable, the adaptation also reduces the risks tion to be driven by the public sector and the market of climate change. Even though the benefits of adapta- and would hence require government coordination for tion are in themselves also uncertain, the range of its successful implementation. Flow adaptation is damage uncertainty is considerably lower with adapta- assumed to be private and market driven. Using a cali- tion than it is without it. bration similar to that in Bosello Carraro and De Cian (2013) but in the framework of the AD-DICE model, This discussion demonstrates the shortcomings in the they estimate the optimal level of each type of adapta- analysis of adaptation in the context of global climate tion over time. If adaptation is undertaken to minimize policy. The appropriate level of adaptation as calcu- total damage costs, it takes the form shown in figure 6. lated in the Integrated Assessment Models depend on The figure shows the percentage of damages avoided macro-level functions linking damage reduction by State of Knowledge on Climate Change, Water, and Economics 29 Figure 6. Percentage of Damages Reduced By Adaptation (No Mitigation) 100 90 80 70 60 Percent 50 40 30 20 10 0 20 60 00 30 40 50 70 80 90 10 25 35 45 55 65 15 75 85 95 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 21 Likely uncertainty range - Flow adaptation Central projection - Flow adaptation Full adaptation Source: OECD 2015a AD-DICE model. sector to adaptation expenditure. The basis of these about the options chosen). As tables 3 and assumptions and this analysis needs to be strength- i ndicate, adaptation expenditures are modest as 4 ­ ened, with more account taken of soft measures and p ercentage of GDP, while they reduce damages to a­ improvements in efficiency of resource use. This is a significant extent; a 70 percent reduction of dam- particularly the case for water. Even without climate ages in 2100 would amount to a reduction in GDP change, water scarcity is likely to be a problem in many loss of around 6–7 percent. Globally, therefore, parts of the world, and action will be needed to address a daptation expenditures amounting to around ­ that. More work is needed to determine the detailed p ercent of GDP would help avoid damages that 0.4 ­ policies and measures that could reduce damages— would represent a GDP loss of around 6–7 percent in especially in the areas of public (stock) adaptation— 2100. Inexpensive strategies with high benefit-cost and how these can complement actions in the area of ratios could be considered in priority. private (flow) adaptation. Nevertheless, the analysis shows how important Conclusions and Next Steps adaptation is in reducing future damages, at a cost that is considerably lower than the reduction in Water is a key channel for climate change to affect the damages (notwithstanding all the qualifications economy and society. The current state of knowledge 30 State of Knowledge on Climate Change, Water, and Economics on climate change and water points to predominantly there is considerable room for such actions to reduce negative effects. Studies indicate variations in precipi- water demand, especially in developing countries. To tation and run-off, but with many regions facing nega- be effective, however, they will in some cases need to tive effects, on balance. Changes in river flow favor be combined with water pricing. some regions and reduce flow rates in others, with pos- In-depth estimates of damages from climate change sible negative impacts on groundwater and on water related to water have been made to 2060 using a com- quality. The models also predict a shrinking of most putable general equilibrium model, and to 2100 using glaciers and increases in frequency and intensity of integrated assessment models that are less able to cap- floods and droughts, but with notable regional differ- ture links between climate and economic output. The ences. In terms of marine areas, predictions of sea level 2060 estimates indicate that the impacts from water rise and storm surge increases are made with some supply changes or changes in water-related extreme confidence, with variations in space and time. The events and marine flows add up to about 1.5 percent of global catch potential of fisheries is likely to increase in GDP in 2060 in the absence of mitigation or ­adaptation. high latitude regions and drop in the tropics. A major This average figure, however, may be misleading for a exacerbating factor is an increased frequency of number of reasons: there is a large uncertainty range, extreme events (flood and droughts) in several parts of with big differences between regions; a number of the world (though not all). impacts are not covered; the estimates depend to some In all these projections, there is a considerable element extent on achieving a strong underlying growth in the of uncertainty, which is reflected in statements of dif- economy; shifts in the structure of the economy may ferent levels of confidence in the data and in the likeli- not be as easy as is assumed in the modelling; and hood of the events. Nevertheless, strong economic modelling of changes for extreme temperature changes impacts can be expected from the changes in water is not based on any real experience. Taking these fac- availability, through the role of water in coastal areas tors into account makes damage estimates larger, with and through floods and droughts at the local level. increased regional variations and higher uncertainty. Importantly, these effects will occur on top of a Estimates to 2100 of potential damages in economic water scarcity situation that already prevails in many terms are even more uncertain, but there are strong parts of the world. Studies indicate an increase in reasons to believe they will be greater as a percent- demand for water due to climate change, which will be age of GDP—perhaps around 10 percent globally, and overlaid on a background of increasing scarcity result- possibly even higher. The long-term projections are ing from growing demand and inefficient allocation of particularly sensitive to the assumed value for equi- scarce water. The impact of climate change on water librium climate sensitivity (ECS)—the amount by scarcity is present but small in general, compared to which ­ temperature will increase with a doubling of the impact of the socioeconomic factors. Changes in concentrations. In addition, damages depend GHG ­ efficiency of water use could make a big contribution more on the emissions scenario that is realized. to water problems, including those caused by climate Lastly, the future path for damages depends to a sig- change. Pricing is one method of conserving water use nificant extent on actions taken now: the less that is and increasing efficiency in its allocation, but techno- done to reduce emissions by 2060, the greater is the logical measures should also make a contribution. estimated damage in 2100. These include measures to reduce evaporation from water storage, increases in irrigation efficiency, and Alternative approaches of linking climate impacts to increases in the productivity of water. Research shows the economy work through their effects on growth, State of Knowledge on Climate Change, Water, and Economics 31 rather than output. There is some empirical evidence and declines thereafter, and the rapid introduction of new and more efficient technologies with a balance between fossil and non-fossil in support of this path, but the results are not firmly energy. In Scenario B2, the emphasis is on local solutions to eco- established and it is difficult to see the causal path- nomic, social, and environmental sustainability. It is a world with ways. Nevertheless, some attempts have been made to continuously increasing global population at a rate lower than the A scenarios, intermediate levels of economic development, and less estimate damages through their impacts on the capital rapid and more diverse technological change than in the A1 sto- stock. They indicate an increase in damage relative to rylines. The scenario is also more oriented toward environmental protection and social equity. See https://www.ipcc.ch/publications_ the computable general equilibrium (CGE) model and_data/ar4/wg1/en/spmsspm-projections-of.html. approach, but not a large one. Further work is needed 2. Scenario B1 represents a world with significant reduction in emis- in this area. sions and a mean temperature increase by the end of the century of 1.8°C. In Scenario A2, emissions rise more or less as at current Adaptation can make a major contribution to reducing rates, population increases are greater, and mean temperature damages from climate change for all mitigation scenar- increases by  3.4°C by the end of the century. See https://www​ ios, and more so when mitigation is absent or limited. .ipcc.ch​/­p ublications_and_data/ar4/wg1/en/spmsspm-projec- tions-of.html. Adaptation will require both private and public actions. 3. The RCPs are pathways for radiative forcing developed by the Public action may need to be at least as large as private research community to model different changes in climate. RCP2.6 action initially, but by 2100 the main focus will be on is  the most optimistic with CO2 emissions staying at today’s level private action. If undertaken optimally, at a cost of less until 2020, then declining and becoming negative in 2100. RCP8.5 is pessimistic scenario, with CO2 emissions rising to three times a  ­ than 0.5 percent of GDP, adaptation could remove up present levels by 2100. See https://www.sei-international.org​ ­ to around 70 percent of damages by the end of the cen- mediamanager/documents/A-guide-to-RCPs.pdf. /­ Under RCP2.6 tury, at a cost that would leave net damages consider- temperature increases by 2081-2100 are likely to be in the range of 1°C, while under RCP8.5 they are likely to be in the range of 3.7°C. ably reduced. But adaptation options need further See Table SPM-2 in: http://www.climatechange2013.org/images​ analysis to include more of the softer options, such as /­report​/­WG1AR5_SPM_FINAL.pdf. those involving ecosystems, and approaches that 4. http://www.un.org/waterforlifedecade/scarcity.shtml. incorporate increased efficiency in the use of scarce 5. The Global Change Assessment Model (GCAM) is an integrated assess- water, among other resources. ment tool for exploring consequences and responses to global change developed primarily by the Joint Global Change Research Institute. For In terms of next steps, work is needed on how eco- more information, visit http://www.globalchange.umd.edu/gcam/. nomic growth in the future could be affected by the 6. In figure 1 the mitigation scenarios are shown in terms of radiative effects of climate change on water and on water-­related forcing, measured in Watts per square meter. The Scenario A2 corre- extreme events. In addition, a better understanding of sponds to a radiative forcing of 7.7W/m2, Scenario B2 corresponds to a radiative forcing of 5.5W/m2 and scenario B1 to a radiative forcing how increases in the efficiency of water use could of 4.2W/m2. Recall that the scenarios correspond to an expected affect the water-energy-economic nexus under climate temperature increase by 2095 of 3.4°C (A2), 5.5°C (B2) and 1.8°C (B1). The case of radiative forcing of 8.8W/m2 represents Business as change is needed. Most models of climate change Usual. assume a more-or-less constant level of efficiency in 7. One acre foot is equal to 1,233.5 cubic meters. water use: if this can be changed, the predictions of losses could be significantly reduced. Finally, a better 8. For more information about the DIVA model, visit http://www.diva​ -model.net/. estimate of the likely reduction of damages from adap- 9. MERGE stands for a Model for Evaluating the Regional and Global tation is needed, based on a detailed bottom-up assess- Effects of GHG Reduction Policies. For more information, visit http:// ment rather than a top-down one. web.stanford.edu/group/MERGE/. 10. See note 3 for descriptions of the scenarios. Notes 11. The  Emergency Events Database (EM-DAT) is maintained by the 1. Scenario A1B represents a world with describes a future world of very Centre for Research on the Epidemiology of Disasters (CRED). For more rapid economic growth, global population that peaks in mid-century .html. information, visit http://emdat.be/advanced_search/index​ 32 State of Knowledge on Climate Change, Water, and Economics 12. For estimates for Europe, see Feyen et al. 2015. Bosello, F., F. Eboli, and R. Pierfederici. 2012. “Assessing the Economic Impacts of Climate Change. An Updated CGE Point of View.” Working 13. The AD-DICE and its sister model AD-RICE are integrated economic Paper 2012.02, Fondazione ENI Enrico Mattei (FEEM), Milan. and geophysical model of the economics of climate change devel- oped at Yale University. For more information, visit http://www.econ​ Bosello, F., R. Roson, and R. S. J. Tol. 2006. “Economy-wide Estimates .yale.edu/~nordhaus/homepage/dicemodels.htm. of the Implications of Climate Change: Human Health.” Ecological Economics 58: 579–91. 14. Pindyck (2012) makes the telling comment that these are “arbitrary functions made up to describe how GDP goes down when tempera- Brown, C., R. Meeks, Y. Ghile, and K. Hunu. 2013. “Is Water Security ture goes up.” Necessary? An Empirical Analysis of the Effects of Climate Hazards on National-level Economic Growth.” Philosophical Transactions of the 15. The GTAP is named after the Global Trade Analysis Project at Purdue Royal Society 371: 1–19. University. The standard GTAP model is a multiregion, multisector computable general equilibrium model. The GTAP-W model is Buchner, B., M. Stadelmann, J. Wilkinson, F. Mazza, A. Rosenberg, and expands the GTAP model to include more details on the demand and D. Abramskiehn. et al. 2014. “Global Landscape of Climate Finance supply of water and The GTAP-BIO-W model takes the GTAP-W fur- /wp​ 2014.” Climate Policy Initiative. http://climatepolicyinitiative.org​ ther to include land use for energy and its implications for water -content​/uploads/2014/11/The-Global-Landscape-of-Climate​-Finance​ demand. For more information, visit https://www.gtap.agecon​ -2014.pdf. .­purdue.edu/models/current.asp. Calzadilla, A. 2010. “Water, Agriculture and Climate Change: A Global 16. Further details of the adaptation modelling, and a comparison across Computable General Equilibrium Analysis.” PhD thesis, University of different integrated assessment models is available in Agrawala et al. Hamburg. (2011). Calzadilla, A., K. Rehdanz, R. Betts, P. Falloon, A. Wiltshire, and 17. The WITCH (World Induced Technical Change Hybrid model) is a R.  J.  S  Tol. 2013. “Climate Change Impacts on Global Agriculture.” modelling tool developed within the Mitigation, Innovation and Climatic Change 120 (1-2): 357–74. Transformation Pathways research programme of the Fondazione Cheung, W. W. L., V. W. Y. Lam, J. L. Sarmiento, K. Kearney, R. Watson, ENI Enrico Mattei (FEEM), Milan. For more information, visit http:// D. Zeller, and D. Pauly. 2010. “Large-Scale Redistribution of Maximum www.witchmodel.org/. Fisheries Catch Potential in the Global Ocean under Climate Change.” Global Change Biology 16: 24–35. References Dell, M., B. F. Jones, and B. A. Olken. 2012. “Temperature Shocks and Economic Growth: Evidence from the Last Half Century.” American Agrawala, S., F. Bosello, C. Carraro, K. de Bruin, E. De Cian, R. Dellink, and Economic Journal: Macroeconomics 4 (3): 66–95. E. Lanzi. 2011. “Plan or React? Analysis of Adaptation Costs and Benefits Using Integrated Assessment Models.” Climate Change Economics 2 (3): Dietz, S. and N. Stern (2014), “Endogenous growth, convexity of dam- 175–208. ages and climate risk: how Nordhaus’ framework supports deep cuts in carbon emissions”, The Economic Journal, forthcoming; available as Alexandratos, N., and J. Bruinsma. 2012. “World Agriculture towards Grantham Research Institute on Climate Change and the Environment 2030/2050 (the 2012 Revision).” ESA Working Paper 12-03, Agricultural Working Paper 159, London. Development Economics Division, Food and Agriculture Organization of the United Nations. Döll, P. 2002. “Impact of Climate Change and Variability on Irrigation Requirements: A Global Perspective.” Climate Change 54 (3): 269–93. Arndt, D. S., M. O. Baringer, and M. R. Johnson, eds. 2010. “State of the Climate in 2009.” Bulletin of the American Meteorological Society 91 (7): Ebi, K. L. 2008. “Adaptation Costs for Climate Change-related Cases S1–S224. of Diarrheal Disease, Malnutrition, and Malaria in 2030.” Global Health 4 (1): 9. Bates, B. C., Z. W. Kundzewicz, S. Wu, and J. P. Palutikof, eds. 2008. Climate Change and Water. Intergovernmental Panel on Climate Change Ebinger, J., and W. Vergara. 2011. Climate Impacts on Energy Systems: Key (IPCC) Technical Report VI. Geneva: IPCC Secretariat. Issues for Energy Sector Adaptation. Washington, DC: World Bank. Bolch, T., T. Pieczonka, and D.I. Benn. 2011 “Multi-decadal mass loss Feyen, L., R. Dankers, K. Bódis, P. Salamon, and J. I. Barredo. 2010. of  glaciers in the Everest area (Nepal Himalaya) derived from stereo “Climate Warming and Future Flood Risk in Europe.” Working Report, imagery.” The Cryosphere, 5, 349–358. Joint Research Centre, European Commission, Ispra, Italy. Bosello, F., C. Carraro, and E. De Cian. 2010. “Climate Policy and the Fischer, G., F. N. Tubiello, H. van Velthuizen, and D. A. Wiberg. 2007. Optimal Balance between Mitigation, Adaptation and Unavoided “Climate Change Impacts on Irrigation Water Requirements: Effects of Damage.” Climate Change Economics 1: 71–92. Mitigation, 1990–2080.” Technological Forecasting and Social Change, Greenhouse Gases-Integrated Assessment, 74 (7): 1083–1107. ————. 2013. “Climate-Change Adaptation.” In Global Problems, Smart Solutions: Costs and Benefits, edited by B. Lomborg, 225–59. Cambridge GOP and UNEP (Government of Pakistan and United Nations University Press. Environment Programme). 2013. Government of Pakistan and the United State of Knowledge on Climate Change, Water, and Economics 33 Nations Environment Programme (2013). The Environment and Climate of the Intergovernmental Panel on Climate Change, edited by C. B. Field, Change Outlook for Pakistan. Retrieved from http://www.mocc.gov.pk​ V. Barros, T. F. Stocker, D. Qin, D. J. Dokken, K. L. Ebi, M. D. Mastrandrea, /­gop/index.php?q=aHR0cDovLzE5Mi4xNjguNzAuMTM2L21vY2xjL2Zyb​ K. J. Mach, G.-K. Plattner, S. K. Allen, M. Tignor, and P. M. Midgley. URldGFpbHMuYXNweD9pZD00JmFtcDtvcHQ9cHVibGljYXRpb25z. Cambridge, U.K. and New York: Cambridge University Press. Helfer, F., C. J. Lemckert, and H. Zhang. 2012. “Influence of Bubble ————. 2013. Climate Change 2013: The Physical Science Basis. Plumes on Evaporation from Non-stratified Waters.” Journal of Hydrology Working Group I Contribution to the Fifth Assessment Report of the 438-439: 84–96. Intergovernmental Panel on Climate Change. Edited by T.F. Stocker, D. Qin, G-K Plattner, M.M.B. Tignor, S.K. Allen, J. Boschung, A. Nauels, Hejazi, M. I., J. Edmonds, L. Clarke, P. Kyle, E. Davies, V. Chaturvedi, Y. Xia, V. Bex, P.M. Midgley. Cambridge, U.K. and New York: Cambridge M. Wise, P. Patel, J. Eom, and K. Calvin. 2014a. “Integrated Assessment University Press. of Global Water Scarcity over the 21st Century under Multiple Climate Change Mitigation Policies.” Hydrology and Earth System Sciences. 18: ————. 2014. Climate Change 2014: Impacts and Vulnerability Part A: 2859–83. Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Hejazi, M. I., L. Clarke, J. Edmonds, E. Davies, M. Huang, S. Kim, P. Kyle, Cambridge, U.K. and New York: Cambridge University Press. R. Leung, H. Li, L. Liu, P. Patel, J. Rice, T. Tesfa, N. Voisin, Marshall Wise, Tris West and Yuyu Zhou. 2014b. “The Water System in GCAM: Key Jiménez Cisneros, B. E., T. Oki, N. W. Arnell, G. Benito, J. G. Cogley, Developments and Future Directions.” Joint Global Change Research P. Döll, T. Jiang, and S. S. Mwakalila. 2014. “Freshwater Resources.” Institute (JGCRI) College Park, MD. Available at: http://www.globalchange​ In Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part .umd.edu/data/annual-meetings/2014/Hejazi-GTSP14_Hejazi_et_al.pdf. A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Hejazi, M. I., J. Edmonds, L. Clarke, P. Kyle, E. Davies, V. Chaturvedi, Change, edited by C.  B.  Field, V. R. Barros, D. J. Dokken, K. J. Mach, M. Wise, P. Patel, J. Eom, K. Calvin, R. Moss, and S. Kim. 2013. “Long- M. D. Mastrandrea, T.  E. Bilir, M. Chatterjee, K. L. Ebi, Y. O. Estrada, term Global Water Projections Using Six Socioeconomic Scenarios in an R. C. Genova, B. Girma, E. S. Kissel, A. N. Levy, S. MacCracken, P. R. Integrated Assessment Modelling Framework.” Technological Forecasting Mastrandrea, and L. L. White, 229–69. Cambridge, U.K. and New York: and Social Change 81 (January): 205–26. Cambridge University Press. Henderson, J., C. Rodgers, R. Jones, J. Smith, K. Strzepek, and Klein, R. J. T., S. Huq, F. Denton, T. E. Downing, R. G. Richels, J.  Martinich. 2015. “Economic Impacts of Climate Change on Water J.  B.  Robinson, and F. L. Toth. 2007. “Inter-relationships between Resources in the Coterminous United States.” Mitigation and Adaptation Adaptation and Mitigation.” In Climate Change 2007: Impacts, Strategies for Global Change 20 (1): 135–57. Adaptation and Vulnerability. Contribution of Working Group II to the Hertel, T. W., and J. Liu. 2015. Implications of Water Scarcity for Economic Fourth Assessment Report of the Intergovernmental Panel on Climate Growth. OECD Environment Working Paper 109, OECD Publishing, Paris. Change, edited by M. L. Parry, O. F. Canziani, J. P. Palutikof, P. J. van der Linden, and C. E. Hanson, 745–77. Cambridge, U.K.: Cambridge Hinkel, J., D. Lincke, A. T. Vafeidis, M. Pettette, R. J. Nicholls, R. S. J. Tol, University Press. B. Marzeion, X. Fettweis, C. Ionescu, and A. Levermann. 2014. “Coastal Flood Damage and Adaptation Costs under 21st Century Sea Level Rise.” Kolstad, E., and K. A. Johansson. 2011. “Uncertainties Associated with PNAS 111 (9): 3292–97. Quantifying Climate Change Impacts on Human Health: A Case of Diarrhea.” Environmental Health Perspectives 119 (3): 299–305. Hsiang, S. M., and A. S. Jina. 2014. “The Causal Effect of Environmental Catastrophe on Long-Run Economic Growth: Evidence from 6,700 Manne, A., R. Mendelsohn, and R. Richels. 1995. “MERGE: A Model Cyclones.” NBER Working Paper 20352, National Bureau of Economic for Evaluating Regional and Global Effects of GHG Reduction Policies.” Research, Cambridge, MA. Energy Policy 23 (1): 17–34. Hurd. B.H. et al. (2004), “Climatic change and U.S. water resources: Mendelsohn, R., K. Emanuel, S. Chonabayashi, and L. Bakkensen. 2012. From modeled watershed impacts to national estimates”, Journal of the “The Impact of Climate Change on Global Tropical Cyclone Damage.” American Water Resources Association, Vol. 40 (1), pp. 129–148. Nature Climate Change 2: 205–09. IEA (International Energy Agency). 2015. World Energy Outlook Special Moore, F. C., and D. B. Diaz. 2015. “Temperature Impacts on Report on Energy and Climate Change. Paris: International Energy Agency. Economic Growth Warrant Stringent Mitigation Policy.” Nature Climate Change 5: 127–31. IPCC (Intergovernmental Panel on Climate Change). 2001. Climate Change 2001: Impacts, Adaptation and Vulnerability edited by J.J. McCarthy, O.F. Nordhaus, W. D. 1994. Managing the Global Commons: The Economics of Canziani, N.A. Leary, D.J. Dokken, K.S. White. Contribution of Working the Greenhouse Effect. Cambridge, MA: MIT Press. Group II to the Third Assessment Report of the  Intergovernmental Panel on Climate Change. Cambridge, U.K. and New York: Cambridge ————. 2007. A Question of Balance. New Haven, CT: Yale University Press. University Press. ————. 2010. “Economic Aspects of Global Warming in a Post-Copenhagen ————. 2012. Managing the Risks of Extreme Events and Disasters to Advance Environment.” Proceedings of the National Academy of Sciences 107 (26): Climate Change Adaptation. A Special Report of Working Groups I and II 11721–26. 34 State of Knowledge on Climate Change, Water, and Economics O’Neill, B. C., T. R. Carter, K. L. Ebi, J. Edmonds, S. Hallegatte, E. Kemp- Water Scarcity: Growing Risks for Agricultural-based Economies in Benedict, E. Kriegler, L. Mearns, R. Moss, K. Riahi, B. van Ruijven, and South Asia.” In Handbook of Sustainable Development in Asia, edited by D. van Vuuren. 2012. “Meeting Report of the Workshop on the Nature S. Hsu et al. Routledge. and Use of New Socioeconomic Pathways for Climate Change Research,” Tol, R. S. J. 2005. “Emission Abatement versus Development as Strategies Boulder, CO, November 2–4 (final version). http://www.isp.ucar.edu​ to Reduce Vulnerability to Climate Change: An Application of FUND.” socio-economic-pathways accessed 14/11/2014. /­ Environment and Development Economics 10 (5): 615–29. OECD (Organisation for Economic Co-operation and Development). 2012. ————. 2009. “The Economic Effects of Climate Change.” Journal of OECD Environmental Outlook to 2050: The Consequences of Inaction. Paris: Economic Perspectives 23 (2, Spring): 29–51. OECD Publishing. doi: http://dx.doi.org/10.1787/9789264122246-en. Vafeidis, A. T., R. J. Nicholls, L. McFadden, R. S. J. Tol, J. Hinkel, ————. 2013. Water and Climate Change Adaptation: Policies to Navigate T. Spencer, P. S. Grashoff, G. Boot, and R. J. T. Klein. 2008. “A New Global Uncharted Waters. OECD Studies on Water. Paris: OECD Publishing. doi: Coastal Database for Impact and Vulnerability Analysis to Sea Level http://dx.doi.org/10.1787/9789264200449-en. Rise.” Journal of Coastal Research 24 (4): 917–24. ————. 2014. OECD Economic Outlook, Vol. 2014/1. Paris: OECD Publishing. Van Vliet, M. T. H, S. Vogel, and D. Rubbelke. 2013. “Water Constraints doi: http://dx.doi.org/10.1787/eco_outlook-v2014-1-en. on European Power Supply under Climate Change: Impacts on Electricity ————. 2015a. The Economic Consequences of Climate Change. Paris: Circle Prices.” Environmental Research Letters 8 (3). Project, OECD Environment Directorate, OECD. Ward, F., and M. Pulido-Velazquez. 2008. “Water Conservation in ————. 2015b. Water Resources Allocation: Sharing Risks and Opportunities. Irrigation Can Increase Water Use. ” Proceedings of the National Academy OECD Studies on Water. Paris: OECD Publishing. doi: http://dx.doi​ of Sciences 105 (47): 18215-20. .org/10.1787/9789264229631-en. Ward, P.J. et al. (2013), “Assessing Flood Risk at the Global Scale: Model Olmstead, S. M., 2013. “Climate Change Adaptation and Water Set up, Results, and Sensitivity. ” Environmental Research Letters, Vol. 8, Resource  Management: A Review of the Literature.” Energy Economics 044019. 46 (November): 500–09. Ward, P.J. et al. (2014), “Strong Influence of El Niño Southern Oscillation Pickering, T. D.,B. Ponia, C. A. Hair, P. C.Southgate, E. Poloczanska, on Flood Risk around the World. ” Proceedings of the National Academy of L. D. Patrona, A. Teitelbaum, C. V. Mohan, M. J. Phillips, J. D. Bell, and Sciences 111(44): 15659-64. S. De Silva. 2011. “Vulnerability of Aquaculture in the Tropical Pacific Weitzman, M.L. (2013), “Tail-Hedge Discounting and The Social Cost of to Climate Change.” In Vulnerability of Tropical Pacific Fisheries and Carbon. ” Journal of Economic Literature 50 (3): 873–82. Aquaculture to Climate Change, edited by J. D. Bell, J. E. Johnson, and A. J. Hobday, 647–731. Noumea, New Caledonia: Secretariat of the Pacific Winsemius, H. C., and P. J. Ward. 2015. “Projections of Future Urban Community. Damages from Floods.” Personal communication to the Organisation of Economic Co-operation and Development (OECD). Available on Pindyck, R. S. 2012. “The Climate Policy Dilemma.” NBER Working Paper request from the Environment Department, OECD. 18205, National Bureau of Economic Research, Cambridge, MA. Wong, P. P., I. J. Losada, J.-P. Gattuso, J. Hinkel, A. Khattabi, Rosegrant, M. W., and the IMPACT Development Team. 2012. K. L. McInnes, Y. Saito, and A. Sallenger. 2014. “Coastal Systems “International Model for Policy Analysis of Agricultural Commodities and  Low-lying Areas.” In Climate Change 2014: Impacts, and Trade (IMPACT): Model Description.” International Food Policy Adaptation,  and Vulnerability. Part A: Global and Sectoral Aspects. Research Institute (IFPRI), Washington, DC. www.ifpri.org/sites/default​ Contribution of Working Group II to the Fifth Assessment Report of /files/publications/impactwater2012.pdf. the Intergovernmental Panel on Climate Change, edited by C. B. Field, Rosenzweig, C., J. Elliott, D. Deryng, A. C. Ruane, C. Müller, A. Arneth, V.R. Barros, D. J. Dokken, K. J. Mach, M. D. Mastrandrea, T. E. Bilir, K. J. Boote, C. Folberth, M. Glotter, N. Khabarov, K. Neumann, F. Piontek, M.  Chatterjee, K. L. Ebi, Y. O.  Estrada, R. C. Genova, B. Girma, T. A. M. Pugh, E. Schmid, E. Stehfest, H. Yang, and J. W. Jones. 2013. E. S. Kissel, A. N. Levy, S. MacCracken, P.  R.  Mastrandrea, and “Assessing Agricultural Risks of Climate Change in the 21st Century in a L.  L. White, 361–409. Cambridge, U.K. and New  York: Cambridge Global Gridded Crop Model Intercomparison.” Proceedings of the National University Press. Academy of Sciences (PNAS) 111 (9): 3268–73. WRI (World Resources Institute). 2014. “Aqueduct Water Risk Atlas.” Stern, N. 2007. The Economics of Climate Change: The Stern Review. World Resources Institute. www.wri.org/our-work/project/aqueduct​ Cambridge, U.K. and New York: Cambridge University Press. /­aqueduct-atlas. Strzepek, K. et al. (2014). “Benefits of Greenhouse Gas Mitigation on the Wurbs,  R. A., and R. A. Ayala. 2104. “Reservoir Evaporation in Texas, Supply, Management, and Use of Water Resources in the United States,” USA.” Journal of Hydrology 510: 1–9. Climatic Change, November 2014. Yang, D., S. Kanae, T. Oki, T. Koike, and K. Musiake. 2003. “Global Taheripour, F. T.W., B. Hertel, S. Narayanan Gopalakrishnan, A. Sahin, Potential Soil Erosion with Reference to Land Use and Climate Changes.” A.  Markandya, and B. K. Mitra. Forthcoming. “Climate Change and Hydrological Processes 17 (14): 2913–28. State of Knowledge on Climate Change, Water, and Economics 35 SKU W16008