102490 IOP PUBLISHING ENVIRONMENTAL RESEARCH LETTERS Environ. Res. Lett. 8 (2013) 044014 (9pp) doi:10.1088/1748-9326/8/4/044014 Toward evaluating the effect of climate change on investments in the water resources sector: insights from the forecast and analysis of hydrological indicators in developing countries Kenneth Strzepek1 , Michael Jacobsen2 , Brent Boehlert3 and James Neumann3 1 Joint Program on the Science and Policy of Global Change, Massachusetts Institute of Technology, 77 Massachusetts Avenue, E19-411, Cambridge, MA 02139-4307, USA 2 World Bank, MC 6-235, 1818 H Street, NW Washington, DC 20433, USA 3 Industrial Economics, Incorporated, 2067 Massachusetts Avenue, Cambridge, MA 02140, USA E-mail: strzepek@mit.edu Received 12 August 2013 Accepted for publication 1 October 2013 Published 23 October 2013 Online at stacks.iop.org/ERL/8/044014 Abstract The World Bank has recently developed a method to evaluate the effects of climate change on six hydrological indicators across 8951 basins of the world. The indicators are designed for decision-makers and stakeholders to consider climate risk when planning water resources and related infrastructure investments. Analysis of these hydrological indicators shows that, on average, mean annual runoff will decline in southern Europe; most of Africa; and in southern North America and most of Central and South America. Mean reference crop water deficit, on the other hand, combines temperature and precipitation and is anticipated to increase in nearly all locations globally due to rising global temperatures, with the most dramatic increases projected to occur in southern Europe, southeastern Asia, and parts of South America. These results suggest overall guidance on which regions to focus water infrastructure solutions that could address future runoff flow uncertainty. Most important, we find that uncertainty in projections of mean annual runoff and high runoff events is higher in poorer countries, and increases over time. Uncertainty increases over time for all income categories, but basins in the lower and lower-middle income categories are forecast to experience dramatically higher increases in uncertainty relative to those in the upper-middle and upper income categories. The enhanced understanding of the uncertainty of climate projections for the water sector that this work provides strongly support the adoption of rigorous approaches to infrastructure design under uncertainty, as well as design that incorporates a high degree of flexibility, in response to both risk of damage and opportunity to exploit water supply ‘windfalls’ that might result, but would require smart infrastructure investments to manage to the greatest benefit. Keywords: climate change, water resources, infrastructure, economic development, investment Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 1748-9326/13/044014+09$33.00 1 c 2013 IOP Publishing Ltd Printed in the UK Environ. Res. Lett. 8 (2013) 044014 K Strzepek et al 1. Introduction a method to evaluate the effects of climate change on six hydrological indicators across 8951 basins of the world. The Major infrastructure investments in virtually any sector indicators are designed for decision-makers and stakeholders require rigorous economic/financial analysis to ensure that to consider climate risk when planning water resources and expected returns justify investment, and that key risks that related infrastructure—here we refer to risk as the product might jeopardize those returns are fully evaluated. Failure to of severity (the magnitude of change) and frequency (the consider changes in future climate impacts risks reliance on likelihood of change). These indicators reflect impacts of a faulty time series of future returns, but with most economic climate change (severity) on irrigation and drainage, large analysis incorporating discount rates on the order of 7–10%, water supply and urban wastewater treatment, small water decision making is heavily influenced by the net monetary supply and rural wastewater treatment, flood protection, and flows of the first two decades—too short to reflect most effects river basin management and multipurpose infrastructure. To of a changing climate. In this time perspective other variables fully understand climate change as a risk factor, however, we are much more important. However, a bigger issue, in are limited by an inability to attribute reliably the frequency particular in developing countries, is whether water resources (or probability) of alternative projections of climate change. infrastructure investments that look economically attractive The next best solution is to provide a representation of the today are consistent with the best long-term development breadth of future change across many plausible predictions of path. National governments and international financial in- future climate. To accomplish this goal, the analysis examines stitutions should consider, for example, whether a large relative changes from an historical baseline to three future multipurpose dam, with attendant irrigated agriculture, elec- periods for 56 GCM-SRES combinations available from the tricity dependent industry and related settlement patterns is IPCC Fourth Assessment (IPCC 2007), enabling users to sustainable in the face of long-term water challenges. Practical employ a risk-based approach to the effect of climate on examples include options to invest in irrigation infrastructure investment plans. As described here, the results provide in the Okavango basin in Botswana (World Bank 2010a); insights into key water resources challenges likely to arise options to invest in high value irrigated agricultural production in developing regions, including the prospect of much larger in parts of the Balkans, Central Asia, and the Southern variability in key hydrological indicators in the poor countries Caucasus (Sutton et al 2013); and proposed hydropower least able to manage those risks. investments in northern and western sub-basins of the Zambezi River basin in southern Africa (World Bank 2010b). 2. Methods In developed country contexts alternatives to large-scale Developing projections of hydrological indicators for 8951 infrastructure investments may be reasonable substitutes for world river basins under a wide range of possible future infrastructure (e.g., water efficiency, input substitution, and climate conditions presents challenges in characterizing other non-infrastructure related changes might be employed to baseline conditions (including the unit of analysis), projecting maintain service levels). Nonetheless, while such alternatives key climate variables, and developing hydrological indicators may also play a role in developing country contexts, the at the basin level. We review our methods for each of these general under-investment in large-scale infrastructure here three steps below. (see Foster and Brice˜ no-Garmendia 2010) suggests that long-lived infrastructure investments should continue to be 2.1. Characterizing baseline conditions proposed and thus require more rigorous analysis. The focus of this study is water resources planning and The best analyses of large-scale infrastructure include development at the regional and local level, and as such, the consideration of future climates and sensitivity analyses, but river basin was identified as the appropriate scale for this they are typically not tied to the specific, internally consistent analysis. A key challenge then is determining an appropriate scenarios of future precipitation and temperature changes that global definition for river basins. We rely on the USGS have been developed for climate change assessments (IPCC HydroSHEDS global basin definitions, based on a 1 km digital 2007), do not incorporate the full range of changes that could elevation model. We chose a combination of Level 3 and be associated with future climates and in particular do not Level 4 basins from HydroSHEDS, in an attempt to roughly adequately take into consideration the uncertainty with respect match basin size to the size of a typical GCM gridbox, in order to future climates which is indicated by the full suite of to ensure the results were not over-specified relative to the climate models and emission scenarios of IPCC (2007) (e.g., scale of GCM results. Nonetheless, the Level 3 and 4 basins Kuik et al 2008, Kirshen et al 2008, Ward et al 2010). It defined in this study vary significantly in size, ranging from is now clear that the wide range of potential future climate approximately 2500 km2 , which is similar to a grid cell of and hydrologic outcomes suggest the use of planning tools 0.5◦ × 0.5◦ , to more than 62 500 km2 , which is similar to a such as robust decision-making (Lempert and Groves 2010), grid cell of 2.5◦ × 2.5◦ . which focus on resilience to uncertain futures rather than For climate data, we rely on a 30-year historical baseline optimization in relation to predicted futures and on methods of (1961–1990), with the goal of projecting to future 30-year decision making for large-scale infrastructure that put a very periods centered on three future eras: the 2030s, 2050s, and high value on flexibility (De Neufville and Scholtes 2011). 2080s. Baseline precipitation and temperature data for the In response to this growing need to evaluate the climate 1961–1990 baseline was taken from the University of East resilience of proposed development paths and related infras- Anglia’s Climate Research Unit (CRU) TS 2.1 data set, which tructure investments, the World Bank has recently developed provides monthly data at a 0.5◦ × 0.5◦ resolution. 2 Environ. Res. Lett. 8 (2013) 044014 K Strzepek et al 2.2. Projecting key climate variables from a suite of natural runoff with six calibration parameters (Strzepek and GCM/SRES combinations McCluskey 2010). This class of model requires natural runoff Projecting changes in climate variables from GCM sim- data to calibrate the model over an historic period. ulations has often involved downscaling approaches, but While global databases of gauged flow are available (e.g., both statistical and dynamical downscaling have well-studied WMO 2012) there is no corresponding database of natural uncertainties (Kerr 2011), and the time and costs of these flows to use in assessing the performance of this procedure at computationally intensive approaches rarely allow the use global scale. McMahon et al (2007) are developing a global of more than a few GCMs. Our goal in this work is to natural flow database based on statistical characteristics of characterize the broadest possible range of ‘not implausible’ natural flow and recreating natural flows from gauged flow, climate futures, as defined by the currently available set of but this effort is limited in scope and not appropriate for our GCM-SRES combinations. The only practical approach for application. Hydrologists have taken an alternative approach a global analysis is to use projected changes in temperature using global gridded databases of climate time series and and precipitation for 56 GCM-SRES combinations at their using hydrologic models to simulate natural flows. The Global native resolutions. These native resolution changes were Runoff Data Centre (GRDC) has developed a composite mapped onto a 0.5◦ × 0.5◦ grid, and then combined runoff database that combines simulated water balance model with the corresponding 0.5◦ × 0.5◦ grid of CRU baseline runoff estimates with monitored river discharge (Fekete et al modeled data. Basin-scale aggregation was then achieved 2002). This data set consists of average monthly runoff values using GIS software to overlay basin boundaries with the for each cell at a 0.5◦ × 0.5◦ global land grid. 0.5◦ × 0.5◦ grids, and then aggregating cells based upon their We calibrated the model by minimizing the squared weighted area in each basin. This approach was designed to deviation between the 12 monthly GRDC runoff values and capture the range of potential climate change impacts at a the 12 monthly averaged CLIRUN-II model outputs from the higher resolution without downscaling the GCMs themselves, 10-year simulation period, which was chosen to best represent thereby achieving a balance between precision and accuracy. the decade used to generate the 12 months of GRDC runoff Note that the 56 climate projections represent the full data. The limitations of using a modeled ‘natural’ runoff range of available models for the B1, A1B, and A2 Special for calibration and having only monthly average values add Report on Emissions Scenario (SRES) scenarios evaluated uncertainty to the results. Other issues with the GRDC data in the Intergovernmental Panel on Climate Change (IPCC) that add to uncertainty in the analysis include: (1) there are Fourth Assessment Report (AR4). There are 17, 22, and 17 large areas (especially in dry regions) that do not have gauge GCM runs, respectively, available for the three emissions data, (2) the time period of available gauge data varies by scenarios, leaving a total of 56 GCM-SRES combinations. station, therefore the resulting monthly discharge regimes are These three SRES scenarios were chosen because they are not fully consistent, (3) the historical climate data used in the generally in the middle range of the marker SRES scenarios water balance model (WBM) of the GRDC data set is not the identified by the IPCC, and are the most commonly used same that was used in the CLIRUN-II model analysis, and emissions scenarios for impact and adaptation assessments. (4) the data set is only provided for 12 average monthly values, To compare across GCMs, we converted GCM modeled not for a full time series. Additional uncertainty also exists in baselines and projections into decadal average monthly the choice of CLIRUN and its model uncertainty. Based on changes relative to the model baseline by subtracting the multi-model assessments, Haddeland et al (2011) and Schewe modeled baseline from the projected values to produce et al (2013) report that differences between hydrological delta temperature and precipitation derived from the model results are also a major source of uncertainty. archived CMIP3 IPCC AR4 outputs. For each GCM-SRES CLIRUN-II produces a 30-year time series of monthly combination, these relative changes for the decades of the hydro-climatic variables that are used in calculation of 2030s, 2050s, and 2080s, were then coupled with the six hydrologic indicators4 : (1) mean annual runoff (MAR); 30 year CRU historical dataset to generate three 30 years (2) river basin yield; (3) annual high flow (q10), or 10% absolute monthly projections representative of potential future 4 MAR: the average annual runoff across 30 years. Basin yield : the maximum conditions in decades of the 2030s, 2050s, and 2080s. sustainable reservoir releases within a basin using derived storage yield curve and the reservoir storage available in each basin. Annual high flow (q10): the 2.3. Translate trends from climate models into hydrological annual runoff that is exceeded by 10% of years in a given period, also referred indicators to as the 10% exceedence flow. In a 10-year period, the q10 flow would be the second highest flow of the 10 available, which is exceeded only by the highest Basin-scale runoff is a key component of the six hydrological flow in that decade. Change in q10 is used as an indicator of flood risk. Annual indicators. To model changes in runoff, this study employed low flow (q90): the converse of annual high flow, this is the 90% exceedence CLIRUN-II: a hydro-climatic modeling framework with flow, or the annual runoff that is exceeded by 90% of years in a designated period. For a 10-year period, this would correspond to the second lowest components that model, PET, Snow Water Balance, and soil recorded flow. Change in annual low flow is used as an indicator of drought moisture. Potential evapo-transpiration (PET) is a necessary risk. Groundwater (baseflow): the sustained flow in a river basin resulting input into runoff modeling as well as irrigation water from groundwater runoff. This indicator is used as a proxy for groundwater requirements. CLIRUN-II uses the Modified Hargreaves availability. Reference crop water deficit: the crop water demand that exceeds available precipitation. Because it was not possible for this study to measure method (Allen et al 1998, Droogers and Allen 2002). biophysical crop water demand, PET was used to represent the water demands The runoff modeling component is a two-layer, one- of a typical perennial grassland over the typical growing season of crops in dimensional conceptual rainfall-runoff model that simulates the basin. 3 Environ. Res. Lett. 8 (2013) 044014 K Strzepek et al Figure 1. Per cent changes in A2 SRES scenario ensemble mean MAR (left) and reference crop water deficit (right), baseline to 2050s. exceedence flow; (4) annual low flow (q90), or the 90% 2050s across the GCMs run for the A2 SRES scenario—the exceedence flow; (5) baseflow or the sustained flow in a river A2 scenario was chosen for presentation because it was also basin resulting from groundwater runoff; and (6) reference used in the World Bank Economics of Climate Change study crop water deficit, which is the crop water demand less (World Bank 2009). Regionally, model results suggest that, available precipitation. on average, MAR will decline in southern Europe; most of As crop modeling and analysis of agricultural water Africa; and in southern North America and most of Central use at the global basin scale were well beyond the scope and South America. Asia, most of North America, and the of this work, we employ a simplified version of the water Pacific Islands are projected to experience increases in water deficit index approach (Woli et al 2008) to estimate reference availability. These general patterns hold for the q10 and q90 crop water deficit. For a given basin-specific growing season, indicators as well. Mean reference crop water deficit, on the this formulation reduces to the sum of monthly PET other hand, combines temperature and precipitation and is minus precipitation for those months in which PET exceeds anticipated to increase in nearly all locations globally due to precipitation. For a more detailed investigations of the impact rising global temperatures. The most dramatic increases in of climate change on irrigation water demand for a range of crop water deficit are projected to occur in southern Europe, GCMs, see Konzmann et al (2013). southeastern Asia, and parts of South America. As part of our evaluation of these mean results, we 3. Results compared our MAR projection to those from another recent analysis (Milly et al 2005). Figure 2 compares 2050s MAR The result of our analysis is a dataset that provides six projections of the current study to Milly et al (2005), each hydrological indicators for over 8000 basins worldwide, for using the same set of the GCMs under the A1B SRES up to 56 alternative climate futures. The methodology and scenario. Although the results differ in several locations such data set has been utilized by the World Bank in a number of as parts of South America and Australia, the general pattern is cases for example for a policy note on adaptation options in very similar globally. Botswana (World Bank 2010a), for a policy note on adaptation options for the Sava River basin and for a multi-sector investment opportunity analysis in the Zambezi River basin 3.1.1. Observation 1: hydrological indicators show a clear (World Bank 2010b). A dataset of this size could easily regional pattern that intensifies and grows less certain over overwhelm users, so the data also includes a user-friendly time. For each of the World Bank regions, figure 3 provides interface that allows for analysis at the country and regional boxplots of per cent changes in MAR from baseline to the level, with mapping products and statistical representations 2030s, 2050s, and 2080s across the 17 A2 GCMs. The of output, such as box and whisker diagrams. The full data World Bank region results are population-weighted averages set and interface can be accessed at the World Bank Climate of basin-level values, grouped into regions based on basin Knowledge Portal, by pointing on a map5 . In this section, centroids. The clear regional trends in MAR become more then, we first provide a summary overview of our global pronounced and less certain over time, illustrating the widely results, and then outline three observations from our analysis different challenges in water resources planning in different of the results. parts of the world. For example, planning for the projected increases in MAR and q10 in the Europe and Central Asia 3.1. Overview of GCM ensemble mean results region poses vastly different challenges for infrastructure development than planning for the anticipated reductions in Figure 1 provides an overview of the mean changes in MAR MAR and q90 in the Middle East and North Africa (MENA). and reference crop water deficit from the baseline to the Our data suggest that these differences are much greater at the 5 See catchment level. It is important to note, however, that in some for example, the following: http://sdwebx.worldbank. org/climateportal/index.cfm?page=country impacts water& regions, the direction of change in MAR become more certain ThisRegion=Africa&ThisCcode=KEN over time. For example, within the MENA region, changes in 4 Environ. Res. Lett. 8 (2013) 044014 K Strzepek et al Figure 2. Comparison of projected per cent changes in mean MAR, baseline to 2050s. Milly et al (2005) on the left, current study on the right. Maps show A1B SRES scenario and GCMs used by Milly et al (2005). Figure 3. Box and whisker diagrams across A2 GCMs showing per cent changes in MAR between baseline and three periods for the World Bank regions. Regional changes are averages of per cent changes in basin MAR weighted by basin populations. Key: AFR is Africa; EAP is East Asia and the Pacific; ECA is Europe and Central Asia; LAC is Latin America and the Caribbean; MNA is the Middle East and North Africa; and SAR is South Asia. MAR are both positive and negative in the 2030s, whereas by (at left), and boxplots of per cent changes in MAR from the 2080s, almost all models project a decrease. baseline to the 2080s across the A2 GCMs for countries within These results, as presented, suggest overall guidance each income category (at right). Income region boxplots are on regions in which to focus water infrastructure solutions population-weighted averages of basin-level values, grouped that could address future runoff flow uncertainty. The full spatially based on basin centroids. World Bank per capita dataset is much richer, however; the country and basin-level income categories include lower (<$1005); lower middle results provide insights at a finer geographic scale, but remain ($1006–$3975), upper middle ($3976–$12 275), and high consistent with the geographic scale of results from GCMs. (>$12 276). Although uncertainty increases over time for Nonetheless, the indicators do not support project-level all income categories, basins in the lower and lower-middle analyses. Concerns over whether a particular hydropower income categories are forecast to experience dramatically investment may face substantial reductions in future flow, for higher increases in uncertainty relative to those in the example, require a yet more detailed site-specific analysis that upper-middle and upper income categories. Strzepek and incorporates engineering considerations that could be adopted Schlosser (2010) find similar results for 2050 and the A2 to adapt to changes in flow. In addition, because our results GCMs when analyzing climate change impacts on the Climate indicate that the full range of available GCMs span a wide Moisture Index. range of hydrologic outcomes, they suggest that project-level analyses may require a new method of decision-making for Figure 5 displays the relationship between income and water infrastructure that puts a very high value on flexibility uncertainty in projected country-level MAR and q10. The (De Neufville and Scholtes 2011). figure plots per capita country income against the IQR of projected percentage changes across the 17 A2 GCMs for the 3.1.2. Observation 2: uncertainty in projections of MAR basins in that country (aggregated based on population; the and high runoff events is higher in poorer countries, and size of each marker corresponds to the population of each increases over time. Our results also suggest that country). All trends are statistically significant (p < 0.001), lower-income countries will face greater uncertainty in future and steepen over time. Note that both the larger and smaller hydrological conditions, particularly mean annual runoff population countries appear to follow these trends. This result and 10% exceedence flows (q10). Figure 4 displays the is not surprising, as precipitation is much more variable in low inter-quartile range for each country of per cent changes in income countries currently, but our work shows that trend will MAR from the baseline to the 2080s across the 17 A2 GCMs be exacerbated by climate change. 5 Environ. Res. Lett. 8 (2013) 044014 K Strzepek et al Figure 4. Inter-quartile range of per cent changes of MAR between baseline and the 2080s for each country (left); box and whisker diagram across A2 GCMs showing per cent changes in MAR between baseline and the 2080s for income categories (right). In boxplots, regional changes are averages of per cent changes in basin MAR weighted by basin populations. Key: income categories include lower (<$1005); lower middle ($1006–$3975), upper middle ($3976–$12 275), and high (>$12 276). Figure 5. Country GDP (2008, US$) versus inter-quartile ranges of per cent changes in MAR (top) and q10 (bottom) across A2 scenarios, changes are from baseline to three periods; marker sizes scaled by country population. Note: p < 0.001 for all trends. Although the observation 1 results indicates more water (e.g., construction standards, concrete investments) that are runoff in general, the Observation 2 results suggests more flexible (see De Neufville and Scholtes (2011)). uncertainty about the amount, and in particular for poorer countries, who are least prepared to manage uncertainty for 3.1.3. Observation 3: uncertainty in projections of reference reasons related to information, institutions, and infrastructure. crop water deficit is higher in wealthier countries, and First, poor countries have less knowledge about current increases over time. Interestingly, our analysis suggests and future climate. Second, poor countries seldom have the that while the uncertainty in MAR and high runoff events regulatory and institutional capacity (including the capacity increases with income, the opposite trend exists in projections for cross sectoral collaboration) to deal with uncertainty. of uncertainty in reference crop water deficit over time, (WMO 2013, Sivakumar et al 2011). Third, poorer countries as illustrated in figure 6. This trend appears to be more often (though with many notable exceptions in regard to pronounced than for uncertainty in MAR. While this may water storage infrastructure) have less water infrastructure, appear to be a contradiction, as both measures consider an investment which can serve as an effective response to temperature and precipitation forecasts, MAR is more uncertainty. In policy terms, then, this result underscores dependent on precipitation outcomes, while reference crop the need for an analytical approach to investment evaluation water deficit is more dependent on temperature for the that focuses on uncertainty (e.g. robust decision-making, PET component, and also exhibits a threshold effect (when see Lempert and Groves (2010)) and on practical solutions precipitation exceeds PET, deficit is 0). To the extent that 6 Environ. Res. Lett. 8 (2013) 044014 K Strzepek et al Figure 6. Country GDP (2008, US$) versus inter-quartile ranges of per cent changes in reference crop water deficit across A2 scenarios, changes are from baseline to three periods; marker sizes scaled by country population. Note: p < 0.001 for all trends. Figure 7. Relationship between reference crop water deficit in A2 forecast scenarios and current country-level percentage of agricultural land that is equipped for irrigation. higher income countries are in higher latitudes, then, what agriculture sector (where possible). All of those measures, appears to be at work is temperature outcomes exhibit higher however, will require good information and advance planning variability in higher latitudes, while precipitation outcomes to address. exhibit higher variability in lower latitudes. Additional work is underway to evaluate the robustness of this outcome. 4. Limitations We also conducted analyses of mean reference crop water deficit (rather than uncertainty) for the A2 scenario There are several key limitations to this analysis. First results versus income, and found no relationship between are the limitations of any hydrological study relying on our projections with either income or with per cent of climate change projections, namely (1) the assumptions, land area irrigated by country, suggesting that it is only model physics, and parameterization of the GCMs; and (2) the the uncertainty in projections which vary with income. The unpredictability of future development pathways and the result is potentially good news for poor agriculturally oriented resulting scenarios for emissions of greenhouse gases, land countries, and presents a challenge for the agriculture sector use changes, and other factors influencing climate change; and in wealthier countries, particularly in areas where adding (3) fundamental uncertainties in the impact of climate change traditional water storage infrastructure has proven difficult on the hydrologic cycle and water resources and the modeling owing to environmental concerns. hereof. We also examined the relationship between the projected In addition, there are several uncertainties which stem reference crop water deficit with climate change and the directly from using rainfall-runoff models in global climate percentage of agricultural land that is currently irrigated change studies. These lumped models tend to be relatively across countries—in this case we forecast that countries with simple, and often require a minimum amount of input in the highest current irrigation penetration also tend to face order to reduce both the uncertainty associated with inputs the highest increases in reference crop water deficit. This and the possibility of compounding errors. Their performance relationship is presented in figure 7 for the three future eras. also relies heavily upon the quality of the calibration process, The relationship is not as strong as for other results presented which is driven by the quality of the naturalized runoff here, but does suggest that areas currently equipped for inputs. Where the GRDC inputs are actually gauged flows, irrigation may face particular challenges related to increased CLIRUN-II is being calibrated to human influenced flow crop water demand. Some of those issues could be resolved rather than naturalized flow. Yet another issue is that because by altering crop choice, improving basin level and/or farm both the GRDC and CRU datasets tend to include too few level water use efficiency, or increasing allocations to the extreme events (runoff and weather, respectively), there is a 7 Environ. Res. Lett. 8 (2013) 044014 K Strzepek et al good chance that extreme events are under-represented in the for better hydrometeorological data—in particular in poor CLIRUN-II results. countries. The benefits of better data will be realized not In terms of input data, both the CRU and GRDC datasets only in the planning phases of these projects, but also in have additional uncertainties. Climatological station data is the operational phases. A better understanding of current not always available for every time and place, an issue that variability may be at least as important as improving the tends to be more common in developing countries where physics in the GCMs, particularly when it is made clear station coverage is often poor. When and where weather that current water infrastructure is poorly adapted to current records are not available, the CRU team uses an interpolation climate, let alone future climate risks. Improvements are method to fill in missing data. Interpolation accuracy is particularly needed in both precipitation monitoring and of particular concern in areas with significant variation in understanding of naturalized runoff flows. Second, as noted elevation, and the accuracy of the original station data, in above, a clear short-term need while data are enhanced and itself, is a source of notable uncertainty. GCMs improved is focus in the near term on better planning models and practices for managing ‘deep uncertainty’. Third, 5. Discussion efforts are needed to mainstream what we have learned into the policies, planning and practice of vulnerable countries and the international finance community. The results presented here are designed to provide a sense of Finally, our results in figures 4 and 5 in particular the value of hydro-indicators developed through this work; provide a new insight about the relationship between water the real value rests in the value of these indicators to inform resources, climate, and country-level income, which deserves project planning, using a consistent and broad set of results. further attention. There is already a substantial and growing Infrastructure project design will nonetheless continue to literature linking the temperature component of climate to require much more detailed hydrologic analyses. For example, country-level income (Acemoglu et al 2002, Dell et al climate change is expected to alter the seasonal pattern of 2009), and suggesting that the temperature component of precipitation, with the result that water can be in short supply climate may provide an indicator of future impacts of climate at exactly the time it is needed most, during the high power change (Horowitz 2009). Our work is prospective, concluding demand or agricultural growing season. Higher temperatures that lower-income countries that are least able to manage also lead to more rapid evaporation from reservoirs, already uncertainty in water availability are likely to face the greatest a major consumptive use of water in many basins, and challenges in this area as a result of climate change. The potentially more rapid evaporation from wetland areas such results also suggests a subtle but potentially powerful factor in as those that characterize some areas, such as the Kafue flats development research as well, that not only water availability region of the Zambezi River basin in southern Africa. These but the level of certainty in water availability may be a key finer scale project-level assessments require a greater spatial component of development success (Brown et al 2008), which and temporal scale than we can achieve with an indicators is deserving of further exploration. approach. At a minimum, the enhanced understanding of the uncertainty of climate projections for the water sector that References this work provides strongly support the adoption of rigorous Acemoglu D, Johnson S and Robinson J A 2002 Reversal of approaches to infrastructure design under uncertainty, as fortune: geography and institutions in the making of the well as design that incorporates a high degree of flexibility, modern world income distribution Q. J. Econ. 117 1231–94 in response to both risk of damage and opportunity Allen R G, Pereira L S, Raes D and Smith M 1998 Crop to exploit water supply ‘windfalls’ that might result, evapotranspiration—guidelines for computing crop water requirements FAO Irrigation and Drainage Paper 56 (Rome: but would likely require infrastructure to manage to the United Nations Food and Agriculture Organization) greatest benefit. In addition, it may make sense to not Brown C, Domeisen D, Meeks R, Hunu K and Yu W 2008 only consider changes to infrastructure investment levels, Exploring the Effect of Hydroclimate Variability on Economic project design, and project operating rules, but also to Growth in Sub-Saharan Africa: A Water Security Index (Report consider non-infrastructure alternatives that can be effective 2 of the Water and Growth Study) (Washington DC: World Bank) (www.africainfrastructure.org) ‘in-the-moment’ adaptations to changing climate. Despite De Neufville R and Scholtes S 2011 Flexibility in Engineering the well-established infrastructure investment gap in many Design (Cambridge, MA: MIT Press) developing countries (SOFRECO Consortium 2011, Vivid Dell M, Jones B F and Olken B A 2009 Temperature and income: Economics 2012, Foster and Brice˜ no-Garmendia 2010), non- reconciling new cross-sectional and panel estimates Am. Econ. infrastructure alternatives may in specific situations postpone Rev. 99 198–204 Droogers P and Allen R 2002 Estimating reference the need for some new climate-sensitive infrastructure. evapotranspiration under inaccurate data conditions Irrig. Certainly, climate change is not the only driver behind the Drain. Syst. 16 33–45 need for more rigorous evaluations—issues of governance, Fekete B, Vorosmarty C and Grabs W 2002 High-resolution fields institutional capacity, and the need for education and outreach of global runoff combining observed river discharge and to support wise use of infrastructure investments continue to simulated water balances Glob. Biogeochem. Cycles 16 15-1–10 be important as well. Foster V and Brice˜no-Garmendia C (ed) 2010 Africa’s The work also suggests a number of improvements Infrastructure: A Time for Transformation (Washington, DC: that could be made in future efforts. First, there is a need World Bank) 8 Environ. Res. Lett. 8 (2013) 044014 K Strzepek et al Haddeland I et al 2011 Multimodel estimate of the global terrestrial Planning and Coordination Agency, and the African water balance: setup and first results J. Hydrometeorol. Development Bank) 12 869–84 Strzepek K and McCluskey A 2010 Water and Climate: Modeling Horowitz J K 2009 The income–temperature relationship in a the Impact of Climate Change on Hydrology and Water cross-section of countries and its implications for predicting Availability (Washington, DC: The World Bank) (http:// the effects of global warming Environ. Resour. Econ. documents.worldbank.org/curated/en/2010/09/12815320/ 44 475–93 modeling-impact-climate-change-global-hydrology-water- IPCC (Intergovernmental Panel on Climate Change) 2007 Regional availability) climate projections Climate Change 2007: The Physical Strzepek K and Schlosser C 2010 Climate Change Scenarios and Science Basis. Contribution of Working Group I to the Fourth Climate Data. Development and Climate Change (Discussion Assessment Report of the Intergovernmental Panel on Climate Paper no. 9) (Washington, DC: The World Bank) (http:// Change ed S Solomon, D Qin, M Manning, Z Chen, documents.worldbank.org/curated/en/2010/10/12924239/ M Marquis, K B Averyt, M Tignor and H L Miller (Cambridge: climate-change-scenarios-climate-data) Cambridge University Press) pp 847–940 Sutton W, Srivastava J and Neumann J 2013 Looking Beyond the Kerr R 2011 Vital details of global warming are eluding forecasters Horizon: How Climate Change Impacts and Adaptation Science 334 173–4 Responses will Reshape Agriculture in Eastern Europe and Kirshen P, Watson C, Douglas E, Gontz A, Lee J and Tian Y 2008 Central Asia (Directions in Development Series) (Washington, Coastal flooding in the Northeastern United States due to DC: The World Bank) doi:10.1596/978-0-8213-9768-8 climate change Mitig. Adapt. Strateg. Glob. Change 13 437–51 Vivid Economics 2012 The Cost of Adaptation to Climate Change Konzmann M, Gerten D and Heinke J 2013 Climate impacts on in Africa (Report Prepared for the African Development Bank global irrigation requirements under 19 GCMs, simulated with Group) a vegetation and hydrology model Hydrol. Sci. J. 58 1–18 Ward P, Strzepek K, Pauw W, Brander L, Hughes G and Aerts J Kuik O, Buchner B, Catenacci M, Goria A, Karakaya E and Tol R 2010 Partial costs of global climate change adaptation for the supply of raw industrial and municipal water: a methodology 2008 Methodological aspects of recent climate change damage and application Environ. Res. Lett. 5 044011 cost studies Integr. Assess. 8 19–40 WMO (World Meteorological Organization) 2012 Global Runoff Lempert R J and Groves D G 2010 Identifying and evaluating robust Data Centre (www.bafg.de/nn 266918/GRDC/EN/Home/) adaptive policy responses to climate change for water WMO 2013 Overview of Current and Emerging Major management agencies in the American west Technol. Forecast. Partnerships/Regional Hydromet Development Projects Soc. Change 77 960–74 (Geneva: World Meteorological Organization) McMahon T A, Vogel R M, Peel M C and Pegram G G S 2007 IPM/(2013)/Doc. 4; ITEM 2.1 Global streamflows—part 1: characteristics of annual Woli P, Ingram K and Jones J 2008 Crop water stress as estimated streamflows J. Hydrol. 347 243–59 by water deficit index and some popular drought indices Conf. Milly P, Dunne K and Vecchia A 2005 Global pattern of trends in Proc.: Climate Information for Managing Risks streamflow and water availability in a changing climate Nature World Bank–EACC 2009 The costs to developing countries of 438 347–50 adapting to climate change: new methods and estimates The Schewe J et al 2013 Multi-model assessment of water scarcity under Global Report of the Economics of Adaptation to Climate climate change Proc. Natl Acad. Sci. at press Change Study (Washington, DC: The World Bank) (available Sivakumar M V K, Motha P, Wilhite D A and Qu J J (ed) 2011 online at www.worldbank.org/eacc) Towards a compendium on national drought World Bank 2010a Botswana Climate Variability and Change, policy—proceedings of an expert team meeting Proc. an Understanding the Risks (Policy Note to the Government of Expert Mtg. on the Preparation of a Compendium on National Botswana, November 2010) Drought Policy (July 2011) (Geneva: World Meteorological World Bank, Water Resources Management, Africa Region 2010b Organization) p 135 AGM-12; WAOB-2011 The Zambezi River Basin: A Multi-Sector Investment SOFRECO Consortium 2011 PIDA: Interconnecting, Integrating, Opportunities Analysis vol 1 (Washington, DC: The World and Transforming a Continent (Programme for Infrastructure Bank) pp 1–37 (http://siteresources.worldbank.org/ Development in Africa, a joint initiative of the African Union INTAFRICA/Resources/Zambezi MSIOA - Vol 1 - Commission, the New Partnership for Africa’s Development Summary Report.pdf, last accessed Jan 13) 9