Article

Water competition between cities and agriculture driven by climate change and urban growth

  • Nature Sustainabilityvolume 1pages5158 (2018)
  • doi:10.1038/s41893-017-0006-8
  • Download Citation
Received:
Accepted:
Published:

Abstract

Urban water demand will increase by 80% by 2050, while climate change will alter the timing and distribution of water. Here we quantify the magnitude of these twin challenges to urban water security, combining a dataset of urban water sources of 482 of the world’s largest cities with estimates of future water demand, based on the Intergovernmental Panel on Climate Change (IPCC)’s Fifth Assessment scenarios, and predictions of future water availability, using the WaterGAP3 modelling framework. We project an urban surface-water deficit of 1,3866,764 million m³. More than 27% of cities studied, containing 233 million people, will have water demands that exceed surface-water availability. An additional 19% of cities, which are dependent on surface-water transfers, have a high potential for conflict between the urban and agricultural sectors, since both sectors cannot obtain their estimated future water demands. In 80% of these high-conflict watersheds, improvements in agricultural water-use efficiency could free up enough water for urban use. Investments in improving agricultural water use could thus serve as an important global change adaptation strategy.

Main

Cities around the world are markedly expanding in size, as global urban growth (that is, increasing urban population) leads to more than two billion additional urban residents by 20301. Today, approximately 54% of the global population (that is, 3.9 billion people1) lives in cities, a fraction that is likely to grow to between 60% and 92% by the end of the twentyfirst century, according to the scenarios from shared socioeconomic pathways2 (SSPs). Domestic water use almost quadrupled over the last 60 years due to increasing population, wealth and access to drinking-water infrastructure3, and there was an even higher increase in water use in cities4. This trend will continue, with domestic water use forecast under one scenario to increase by another 80% by 20305. A previous study6 has analysed a range of SSP scenarios and has estimated an increase of 50–250% in global domestic water withdrawals by 2050. This growth in demand comes just as cities are confronted with water-related problems caused by an increase in urban population, rising prosperity and changes in water supply, potentially leading to water shortage and groundwater over-abstraction7,8. Climate change poses an additional risk to metropolitan areas and will probably exacerbate rather than ameliorate the major water challenges in the future9,10,11. Climate change is expected to alter renewable freshwater resources in many regions of the world, causing change in spatiotemporal temperature and precipitation patterns and largely increasing evaporative demand12.

Global hydrological models (GHMs) are useful tools to simulate changes in the water cycle and to quantify the impacts of change in both water supply and demand for renewable freshwater availability and sectoral water use. Recent studies estimated that 3.5–4.4 billion people are expected to live under water scarcity in 2050, because of climate change and increasing water demand for human activities13,14. Similarly, water shortage has been identified as a major threat to urban water supply8,15. More than one billion urban residents may face water shortage in the future owing to urbanization and climate change11. A previous study16 has estimated the vulnerability of 70 large cities (of more than 750,000 inhabitants per city) to changes in future water demand under normal climate conditions, and found that some of these cities may face chronic water scarcity in the future.

However, previous global studies did not fully account for urban water infrastructure that transfers water across watershed boundaries. Accounting for these transfers in models is crucial to avoid a substantial overestimation of urban water stress7. Moreover, previous studies did not use an integrated approach when examining changes in climate as well as the broad spectrum of socio-economic changes envisioned in the SSP. Major socio-economic changes include urbanization, which leads to an increase in municipal water demand, as well as agriculture-sector changes, which will markedly increase water withdrawals in some areas.

The aim of our study is to assess urban water provision under climate change impacts and socio-economic changes in the future (2050s). We provide estimates of the urban surface-water deficit (that is, the supply gap) highlighting the competition between urban water provision, agricultural water demand and environmental flow requirements. Finally, implications for groundwater resource uses are addressed.

Our study focused on 482 cities containing 736 million people today. Climate change uncertainty has been addressed by using output from an ensemble of five general circulation models (GCMs) associated with RCP6.0 and socio-economic assumptions following a business-as-usual scenario (SSP2). The radiative concentration pathways (RCPs) are not associated with unique socio-economic scenarios17 and the selection of RCP6.0 and SSP2 is a suitable combination of the scenario matrix architecture18,19. We quantified the urban surface-water deficit as well as additional withdrawals from groundwater resources (urban groundwater footprint) in the future by testing two scenarios: that urban dwellers have the highest priority to be supplied with water (first priority), or that urban dwellers have the lowest priority after other water users, such as industry and agriculture (last priority)20. Furthermore, we considered the consequences of implementing environmental flow requirements as an additional constraint on water allocation. Then we identified which cities will be most affected in the future, determined the number of people and irrigated area not supplied, and assessed the possibility of water saving to reduce or prevent the impacts (Fig. 1). Because urban water transfers put pressure on water resources that are sometimes far away from the city’s location, we integrated detailed information on infrastructure and water management7 (that is, withdrawal points of cities) into the WaterGAP3 modelling framework, which comprises a global hydrology model21 and global water use models3,22 (see Methods and Supplementary Tables 1, 2).

Fig. 1: Schematic illustration of the methods adopted in this study
Fig. 1

.

Urban surface-water deficit and implications

If cities have first priority over other sectors, 16.1% of cities in our sample with surface-water withdrawal points experienced at least one month with a surface-water deficit within the baseline period 1971–2000 (hereafter referred to as the baseline). A surface-water deficit means that the available amount of surface water is less than the demand, and that cities must therefore depend on water storage to make it through such periods. By contrast, 38.9% of cities would be vulnerable to an urban surface-water deficit if cities have the last priority. Integrating environmental flow requirements (see Methods) into our analysis23,24 leads to a decrease in surface-water resources available for urban use and therefore to an increased number of cities vulnerable to water scarcity (first priority scenario, 36.3%; last priority scenario, 62.5%; see Supplementary Information and Supplementary Tables 3, 4).

Considering the effects of global warming and urbanization, the number of cities affected by surface-water deficit is likely to increase to 27.6% in 2050 (Fig. 2a). Almost 46.6% of cities may experience a surface-water deficit if environmental flow requirements also need to be fulfilled. In 41% of all river basins, the needs of agricultural water users will conflict with those of cities, as there is not enough water to supply both urban and agricultural needs. South Asia is a hotspot region where competition between urban and agricultural demand is generally highest. The United States shows a clear division between cities in the West that are vulnerable to a surface-water deficit and cities in the East that have little vulnerability.

Fig. 2: Global urban water security (n = 482).
Fig. 2

a, Urban surface-water deficit (n = 416). Subbasins are highlighted where monthly urban surface-water deficits occur in a 30-year period in the future (2041–2070). We distinguished subbasins with a urban surface-water deficit even when cities get the first priority for water (red), those basins where competing use of water by industry and agriculture would create a water deficit (last priority supply of urban water; blue), and basins in deficit after additionally accounting for environmental flow requirements (EFR, green). b, Urban groundwater footprint (n = 251). We show the trend in urban groundwater footprint in regional aquifers between the baseline (1971–2000) and the 2050s (2041–2070) and the main drivers of change. Groundwater withdrawal points will have more water vulnerability if marked in yellow (CC, climate change), orange (SE, socio-economic changes) and red (SE + CC, socio-economic and climate changes). Groundwater withdrawal points will have less water vulnerability if marked in green (climate change), light blue (socio-economic changes) and blue (socio-economic and climate changes). Black dots represent cities.

If a provision is made for environmental flows, many more river basins and the cities that source from them will experience water deficits. Globally, in our urban first priority scenario, about 14,000 km2 of irrigated area (in its extent of the year 2005) are at risk of crop water deficit in 2050 as a result of an increase in the urban water supply, a reduction in freshwater availability and simultaneous increase in agricultural demand. Similarly, setting additional water aside for environmental flow would further increase the irrigated area at risk of having a water deficit to 180,000 km2, which is about 6% of the current global area that is equipped for irrigation (see Methods, Supplementary Information and Supplementary Table 5).

Unsustainable urban groundwater provision

Increasing demand for urban water supply will also put additional pressure on groundwater resources. The quantification of a groundwater deficit (in units of volume, such as litres) was not possible owing to lack of complete global datasets of aquifer-specific storage. Storage is particularly important for groundwater, because the volume of water stored in aquifers sometimes greatly exceeds the flows of water into and out of the aquifer. Instead, we investigated urban groundwater stress by calculating the urban groundwater footprint (UGF)7,25 of regional aquifers26 providing cities with groundwater for the baseline period and the 2050s. The UGF is a unit-less ratio of groundwater use to estimate recharge, and is used to identify cities that are using groundwater in a way that will reduce aquifer storage.

Climate change and socio-economic developments, including urbanization, lead to an increasing urban groundwater footprint (Fig. 2b). As a result, the UGF increases in 238 cities with groundwater abstraction points (95%) and will more than double in 116 cities in 2050. Twelve cities might be regarded as groundwater-stressed (defined as UGF ≥1) where annual urban groundwater abstractions alone exceed groundwater recharge of regional aquifers. Although aquifer storages might be available to serve these cities in the future, a drop in groundwater levels may result in higher costs for abstraction. Furthermore, urban water supply of 61 cities depends solely on groundwater abstractions with 59 cities experiencing an increasing trend in UGF.

Future threats to urban water provision

Figure 3 presents the urban surface-water deficit under first priority and last priority assumptions as well as considering environmental flow requirements. The global urban surface-water deficit is about 1,386 million m³ in 2050 if urban water supply has first priority. This number corresponds to 3% of the baseline urban surface-water demand. Under last priority assumptions, 161 cities are likely to be vulnerable to urban water deficits, with a total surface-water deficit of 6,764 million m3 (that is, 15% of baseline demand). Under the last priority scenario, and with environmental flow requirements considered, the surface-water deficit will be significantly higher and increase up to 28,842 million m³ (that is, 62% of baseline demand) (Supplementary Table 6).

Fig. 3: Urban surface-water deficit by subregion for the baseline and 2050s.
Fig. 3

If cities get priority (first priority), the urban water deficit is the lowest. If cities come after other sectors (last priority), the urban surface-water deficit is higher. Finally, if in addition water is set aside for EFR, the urban surface-water deficit is highest. Median annual urban surface-water deficit is shown for the baseline (1971–2000) and 2050s (2041–2070) based on results taking into account climate forcing data of five GHMs (see also Supplementary Table 6). NWP, Northwest Pacific.

Regional differences in urban surface-water deficit are apparent for baseline conditions but are more pronounced in the future. For the baseline time period, the urban surface-water deficit is relatively minor in most UNEP GEO subregions, except North America. In 2050, the urban surface-water deficit is likely to exceed 400 million m³ in both South Asia and North America, even if urban water supply has first priority. Note that in these water deficit areas in the first priority assumption, any other water demand by another sector will therefore also by definition not be fulfilled. Other regions, such as southern Africa, South America as well as northwest Pacific and East Asia will become vulnerable to urban water supply due to rapid urbanization. Under the last priority scenario, the urban surface-water deficit is much larger, although with a corresponding decrease in water scarcity for other sectors. The marked difference between the first priority and last priority scenarios shows that integrated multi-sectoral management solutions will be required in many areas to adapt to water scarcity.

Under the assumptions of a business-as-usual scenario (SSP2) more than 440.5 million people will live in cities with a water deficit in the future even in the first priority scenario, 36.4% of the population of the cities that we studied. In the last priority scenario, 673.3 million urbanites will live in cities with a water deficit (55.6%). However, many more people live in the basins from which cities are withdrawing water, and are therefore likely to be affected by urban surface-water deficits. We estimated that 38.3 to 480.4 million others living in source watersheds will be affected by the cities’ transfer of water, showing that urban surface-water withdrawals decrease water security for others who live sometimes far from the city centre. Therefore, the total number of people (in our studied cities, plus in their source watersheds) affected by the urban surface-water deficit is much higher (up to 1,464.7 million if environmental requirements are considered) (Table 1).

Table 1 Number of people affected by surface-water decifit, in the cities that we studied (left), and in their source watersheds (right)

Solutions to reduce the urban surface-water deficit

Altogether 40% of cities (409 million people) are vulnerable to urban surface-water deficits under the last priority scenario, because of competition with water abstractions for irrigation. Urban water demand could be met in these basins by improving water-use efficiency of the agricultural sector as a soft-path measure to reduce irrigation water withdrawals14. Efficiency improvements can result from switching from less-efficient flood irrigation to more-efficient sprinkler or drip irrigation, reducing leakages in water infrastructure to the fields, changes in plant varieties, and simply better information about when and where to irrigate27. Assuming a moderate increase of 10% in irrigation water-use efficiency, which is about 0.3% per year over a 50-year time period, could reduce the urban surface-water deficit by about 2,618 million m³ and therefore help 78% of the vulnerable cities and their 236 million urbanites to overcome water deficits in the future (Fig. 4).

Fig. 4: Cities and subbasins where an improvement in irrigation water-use efficiency could help to overcome urban surface-water deficits in the future.
Fig. 4

The red colour indicates subbasins where the urban surface-water deficit exceeds the consumptive use of irrigation water and therefore, an efficiency improvement is not adequate to fully resolve the water deficit.

An efficiency improvement of 10% in agricultural crop irrigation will help to overcome the urban surface-water deficit in at least 93 subbasins without additional measures, whereas a further improvement up to 30% would be required in 14 subbasins. In 7 subbasins the improvement in irrigation efficiency will not be an adequate measure for reduction, because the urban surface-water deficit already exceeds the amount of water consumed by plants. Here, other measures should be considered, such as a reduction in irrigated land, wastewater reuse or improving agricultural water productivity14,28, but these measures are not addressed in this study. Nonetheless, it is necessary to note that there is room for efficiency improvements in the future that depends not only on current irrigation technologies but also on future investments, expected economic gains and the willingness to adopt practices.

In 2050, cities, such as Los Angeles, Jaipur and Dar es Salaam, are forecast to have the greatest surface-water deficits that will exceed, on average, 100 million m3 per year (Fig. 5). Adapting to these annual deficits corresponds to savings of between 30 and 150 litres per capita per day. The urban surface-water deficit of the top 20 cities is about 2,338 million m3 (35% of the total deficit).

Fig. 5: Top 20 cities under urban surface-water deficit affected by climate change and socio-economic development (including urbanization).
Fig. 5

Two scenarios were analysed: (1) urban surface-water supply gets first priority; or (2) cities receive the remaining water after water demand of other sectors has been fulfilled (last priority). Median urban surface-water deficit is shown; error bars represent the range of uncertainty resulting from the climate forcing data of the five GHMs. Only withdrawal points located in subbasins ≥10 grid cells (>800 km2) are considered.

Urban water transfers export water insecurity

Our implementation of inter-basin transfers into a global water modelling framework is a step forward toward quantifying urban water security. Our results suggest that cities are already vulnerable to water shortages today4,29 and one in six large cities is likely to be at risk of water deficits. Urban population growth will make solving this problem harder, as another 595 million people will be in the cities that were included in this study in 2050. This rapid growth will likely require large-scale maintenance, improvement and extension of the existing infrastructure.

Our findings help to quantify a statement from ref. 30 that indicates that interactions between cities and the countryside will become increasingly intertwined. Because of inter-basin transfers, many cities shift their water deficits to river basins far outside of the cities and remotely impact millions of people, economy and aquatic ecosystems. For every three urban residents living in a city that avoids water deficit via cross-basin transfer, there is one person affected remotely in the source basins (last priority assumption). Our results show that climate change and increased urban demand will lead to markedly greater competition between the agricultural and urban sectors for water, affecting a large fraction of the world’s river basins.

Adapting urban water supply to climate change impacts and increasing urbanization is a major challenge cities will have to face in the future. Technical and governance solutions to reduce or even prevent future urban surface-water deficits cannot be discussed in isolation31, but have to be integrated not only into water requirements of other sectors in the basins and aquifers where the water is abstracted but also in upstream areas. This is particularly important in case of cities that are likely to be at risk of deficits even under the first priority assumption. For cities that will have a deficit only under the last priority assumption, one possible solution is an increase in irrigation water-use efficiency to relieve conflict between urban and agricultural water demands. We found that an increase in irrigation water-use efficiency of 10% can serve as an adaptation option to overcome urban surface-water deficits for 236 million people.

Methods

We used the integrated global water modelling framework WaterGAP3 to simulate river discharge and sectoral water withdrawals and water consumption over the entire timeseries from 1971 to 2070. For all analyses, we considered the model results representing the baseline period (time period from 1971 to 2000) and future period (referred to as the 2050s, the time period 2041–2070). The components of the modelling framework, main input data and data analysis are briefly described in the following sections; for a more detailed description readers are referred to the indicated references.


Global hydrology model

The hydrological model simulates the characteristic macroscale behaviour of the terrestrial water cycle in order to estimate the renewable freshwater availability on a 5 × 5-arcmin spatial resolution21,32,33 (about 9 × 9 km at the equator). On the basis of timeseries of meteorological forcing data, the hydrological model calculated the daily water balance for each grid cell, taking into account physiographic characteristics such as soil type, vegetation, slope and aquifer type. Runoff generated on the grid cells was routed to the catchment outlet on the basis of a global drainage direction map34, taking into account the extent and hydrological influence of lakes, reservoirs, dams and wetlands35. The hydrological model was calibrated against observed mean annual discharge data by adjusting one parameter, that is, the runoff coefficient γ for all grid cells within each calibration basin36. This parameter corresponds to a linear relationship between soil saturation and runoff fraction. Observed discharge timeseries were provided by the Global Runoff Data Center and only a selection of stations that fulfilled the following criteria were considered in the calibration procedure: (1) an upstream area of at least 3,000 km2; (2) a timeseries of at least five (complete) years; and (3) an inter-station catchment area to the next upstream station of at least 5,000 km2. Finally, a total number of 2,446 stations, covering 51% of the global land area, excluding Antarctica and Greenland, were used for calibration. The calibration procedure and regional patterns of model performance of the global hydrology model have been described previously36.


Water-use models

Water withdrawal and consumption are calculated for the domestic, industrial (manufacturing, thermal electricity production) and agricultural (irrigation, livestock) sectors distinguishing between abstractions from groundwater and surface-water bodies37. Sectoral water use for the baseline period was calculated with WaterGAP3 driven by data input such as population, gross domestic product, gross value added, urbanization, thermal electricity production, actual irrigated area as percentage of area equipped for irrigation, and livestock numbers. These driving forces are statistical data originating from different sources such as the World Bank, United Nations Statistics Division, Food and Agriculture Organization of the United Nations and cover the time period 1971–2009 in this study (see also refs 3,22).

Water demand for households and small businesses is estimated per country by multiplying total population with a per-capita water-use intensity per country that accounts for structural change and technological improvements over time. The approach of structural change is expressed by a sigmoid curve and reflects the relationship between water-use intensity and income on a national or regional basis depending on data availability3. Besides structural changes that can either increase or decrease water-use intensitiy, the concept of technological change is applied to account for improvement in water-use efficiency. Water consumption is calculated by a consumptive-use coefficient method, that is, a water consumption to water withdrawal ratio, and derived from national statistics3 or data from ref. 38.

The manufacturing water-use model simulates the annual amount of water withdrawn and consumed for production and cooling processes, except water used for power generation. The model simulates the water withdrawals on a country scale by a specific structural water-use intensity, which is derived for the base year 2005, describing the ratio of water use over the manufacturing gross value added3. Technological improvements are considered through a technological change factor. Water consumption is estimated as the difference between water withdrawal and wastewater. Where no data are available, the fraction of consumptive water use is derived from neighbouring or economically comparable countries.

The amount of cooling water withdrawn and consumed for thermal electricity production is determined by multiplying the annual thermal electricity production with the respective water-use intensity of each power station3. Input data on location, type and size of power stations are based on the World Electric Power Plants Data Set39. The water use intensity is a function of the cooling system and the source of fuel of the power station. Four types of fuels (biomass and waste, nuclear, natural gas and oil, coal and petroleum) with three types of cooling systems (tower cooling, once-through cooling, ponds) are distinguished40.

Net and gross irrigation water requirements, which reflect an optimum supply of water to irrigated plants, are computed on the area equipped for irrigation based on the digital global map of irrigated areas around the year 200541,42 (Global Map of Irrigation Areas (GMIA) v.5). The model simulates cropping patterns, growing seasons, and net and gross irrigation requirements, distinguishing 21 crop types22. Irrigation water withdrawals are determined as the ratio of irrigation water consumption (that is, water evapotranspirated by crops) to irrigation project efficiency (country-specific)43,44. Because future land use changes, including the extent and location of irrigated areas and future irrigation practices, are not yet available for RCP–SSP scenario combinations, we kept the extent of irrigated areas and irrigation efficiencies at base year levels. Water withdrawals and consumption for livestock are computed by multiplying the number of animals per grid cell by the livestock-specific water-use intensity45,46.

While water demands for irrigation, livestock and thermoelectric power plants are computed on a 5-arcmin grid cell basis, countrywide estimates of water use in the manufacturing and domestic sectors are allocated to grid cells within the country based on the geo-referenced population density and urban population maps2,3.


Water transfers

In this study, we examined water transfers of 482 large cities, taking into account the location of withdrawal points from surface water (1,315 serving 416 cities), groundwater (313 serving 251 cities) and desalination plants (29 serving 20 cities)7. (See Supplementary Information, Supplementary Tables 1, 3). Information on where cities obtain water was taken from the City Water Map, which focused on mapping the water sources for the largest cities on earth, as well as primary cities in each country. See ref. 47 for more information on data collection.

The information on urban water transfers, such as location of withdrawal points and amount of water withdrawn, were implemented in WaterGAP3. Withdrawal points from surface water are located in 592 subbasins while the groundwater abstraction points are located in 138 regional aquifers. Because most cities have more than one withdrawal point, their water demands were proportionally allocated to the different withdrawal points depending on the reported volume. If this information was not available, water demands were equally divided among the respective withdrawal points. Technically, each withdrawal point was linked to its particular source of abstraction, that is, the river network, reservoir and groundwater aquifer, as well as to the city supplied by the withdrawal point. Current and future water demands were calculated with the domestic water-use model and downscaled according to the population grids and water transfers7,48,49. Whereas water withdrawals were transferred from the ‘source basin’ to the cities’ location, the consumption of water as well as return flows were allocated where urban inhabitants live.


Climate and socio-economic scenario data

For our study, we used the latest set of global change scenarios on climate and socio-economic developments to drive the integrated modelling framework and quantify the status of current and future water resources. Therefore we combined the 'business-as-usual’ shared socioeconomic pathway (SSP2)50 with a ‘higher emission’ representative concentration pathway (RCP)51 in the integrated modelling framework. GCM outputs are known to differ substantially from observations owing to biases52. To address climate uncertainty, the hydrological model was forced with bias-corrected climate projections from five GCMs53 as generated within the ISI-MIP framework54. The five GCMs were HadGEM2-ES, IPSL-CM5A-LR, MIROC-ESM-CHEM, GFDL-ESM2M and NorESM1-M and based on a radiative forcing of 6.0 W m−2 (RCP6.0) and cover the time period 1971–2099. RCP6.0 is characterized by a stabilization of radiative forcing in 2100 without peaking in prior years55. To quantify the impacts of climate change, we calculated changes over the future period 2041–2070 (representing the 2050s) relative to the baseline period 1971–2000 (baseline). Changes in water demand for the domestic, manufacturing and thermal electricity production sectors were driven by scenarios on population growth, urbanization, gross domestic product, gross value added, thermal electricity production and technological change rates for the time period 2010–2070 as developed within the Water Futures and Solutions initiative (WFaS)6. Future irrigation water withdrawals and water consumption were driven by climate variables from the climate projections mentioned above, while the irrigated area and water-use efficiencies were kept constant at current values56. Projections of future irrigated areas and irrigation water-use efficiencies under combined SSP–RCP scenarios are not yet available but under development in the WFaS framework.


Urban surface-water deficit

We used urban surface-water deficit as an indicator to assess the vulnerability of large cities to changes in surface-water availability caused by climate change and socio-economic developments. The indicator describes the difference between available water for urban supply and urban water demand and was calculated as follows:

  1. (1)

    Estimation of monthly water availability based on the climate scenarios from the five GCMs on a (sub)basin scale for the entire time period of 1971–2070.

  2. (2)

    Calculation of sectoral water withdrawals and consumption driven by socio-economic data (SSP2) and climate scenarios for the entire time period of 1971–2070 and allocated to (sub)basin scale.

  3. (3)

    Calculation of monthly urban surface-water deficit per (sub)basin by subtracting urban water demand from available surface-water resources considering upstream–downstream relationships between subbasins for the entire time period. Differentiation between first priority (that is, urban water supply was given the highest priority in the respective subbasins) and last priority (that is, urban water supply was given the lowest priority after the demand of other water users was satisfied) scenarios.

  4. (4)

    Calculation of long-term annual average urban surface-water deficit per (sub)basin, city, and regions for the baseline (1971–2000) and 2050s (2041–2070).

All analyses were performed at the scale of 143,653 individual (sub)basins globally57. The urban surface-water deficit was calculated for all basins and subbasins in which surface-water withdrawal points are located. The total number of (sub)basins is 592 with 498 (sub)basins larger than 800 km2 (that is, 10 grid cells).

In addition, environmental flow requirements were taken into account in our analysis. Rather than maintaining a minimum flow (for example, Q90) throughout the year, our approach is based on the natural flow paradigm58 and considers environmental flow requirements in each month of the 30-year timeseries. So far no generalizable relationships between flow alteration and ecological impact are available for large-scale assessments59. Thresholds gained from case studies are not often transferable across ecoregions as a high uncertainty still exists about how sensitivity to flow varies at different river types60. Therefore we follow the previously published approach23, which suggested that the protection of 80% of daily flows should be used as a ‘presumptive standard’ to maintain the ecological integrity. This approach was also applied in ref. 24 to calculate monthly blue water scarcity.

On the basis of the considerations described above, the irrigated area at risk of suffering from urban surface-water deficit was calculated. Furthermore, the number of population affected by surface-water deficit (including the population living in the ‘source basin’) was determined.

Irrigated area is located in 114 (sub)basins where urban surface-water deficit occurs under last priority assumptions. We evaluated the possibility of improving irrigation efficiencies in the source basins as an adaptation option to increase the water available for urban water supply and therefore to reduce the urban surface-water deficit. Project irrigation efficiencies typically range between 0.3 and 0.8, where 0.8 means that 80% of the water delivered to the crop is actually transpired. The upper range of 0.8 was assumed as the upper limit of improvement, that is, a reduction in irrigation water withdrawals was not possible if efficiency values exceeded 0.8. We calculated the improvement in irrigation water-use efficiency required to reduce the urban water deficit in the 2050s under the last priority assumption. First, future irrigation water withdrawals and consumption were simulated for all subbasins (>800 km2) containing surface-water withdrawal points. The target value to be achieved by a reduction of irrigation water withdrawals was set to the difference between the highest monthly mean urban surface-water deficit of the 30-year period and the corresponding irrigation water withdrawals in the subbasins. Then the improved irrigation efficiency was calculated to achieve the target value.


Urban groundwater footprint

The UGF was determined for 138 regional aquifers following the analysis described previously25. The UGF is defined as the ratio of aquifer-averaged annual urban water abstraction of groundwater to the groundwater recharge rate including artificial recharge from irrigation. The WaterGAP3 modelling framework was used to compute monthly groundwater recharge rates, including artificial drainage from irrigation return flows forced by the climate input from the five GCMs for the baseline and 2050s61,62. The areal extent of regional aquifers was taken from WHYMAP26 (World-wide Hydrogeological Mapping and Assessment Programme) and the cities’ groundwater withdrawal points were linked to the respective aquifers according to the information provided by the City Water Map database7. We calculated the fraction of water withdrawals and consumption from groundwater and surface-water bodies (rivers, lakes and reservoirs) separately for each sector37,61 and determined annual groundwater abstractions per city for the baseline and 2050s time periods. Finally, we calculated the UGF for each regional aquifer with city withdrawal points and identified cities that used groundwater resources in an unsustainable way (that is, UGF ≥ 1). In addition, the trend of the development in UGF was determined as well as the key drivers of change (climate change, socio-economic development or both). Whereas changes in groundwater recharge are impacted by climate change, changes in urban groundwater abstractions are driven by socio-economic developments and urbanization. Future changes in renewable groundwater resources and urban water abstractions are likely to lead to an increase or decrease in urban groundwater footprint.


Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    UNDP. World Urbanization Prospects: The 2014 Revision (United Nations Population Division, New York, 2014).

  2. 2.

    Jiang, L. & O’Neill, B. C. Global urbanization projections for the shared socioeconomic pathways. Glob. Environ. Change 42, 193–199 (2017).

  3. 3.

    Flörke, M. et al. Domestic and industrial water uses of the past 60 years as a mirror of socio-economic development: a global simulation study. Glob. Environ. Change 23, 144–156 (2013).

  4. 4.

    Richter, B. D. et al. Tapped out: how can cities secure their water future? Water Policy 15, 335–363 (2013).

  5. 5.

    Amarasinghe, U. & Smakhtin, V. Global Water Demand Projections: Past, Present, and Future (IWMI, Colombo, 2014).

  6. 6.

    Wada, Y. et al. Modeling global water use for the 21st century: the Water Futures and Solutions (WFaS) initiative and its approaches. Geosci. Model Dev. 9, 175–222 (2016).

  7. 7.

    McDonald, R. I. et al. Water on an urban planet: urbanization and the reach of urban water infrastructure. Glob. Environ. Change 27, 96–105 (2014).

  8. 8.

    Li, E., Endter-Wada, J. & Li, S. Characterizing and contextualizing the water challenges of megacities. J. Am. Water Resour. Assoc. 51, 589–613 (2015).

  9. 9.

    Hallegatte, S., Green, C., Nicholls, R. J. & Corfee-Morlot, J. Future flood losses in major coastal cities. Nat. Clim. Change 3, 802–806 (2013).

  10. 10.

    McDonald, R. I. et al. Global urban growth and the geography of water availability, quality, and delivery. AMBIO 40, 437–446 (2011).

  11. 11.

    Revi, A. et al. in Climate Change 2014: Impacts, Adaptation, and Vulnerability (eds Field, C. B. et al.) 535–612 (Cambridge Univ. Press, Cambridge, 2014).

  12. 12.

    Collins, M. et al. in Climate Change 2013: The Physical Science Basis (eds Stocker, T. F. et al.) 1062–1087 (Cambridge Univ. Press, Cambridge, 2013).

  13. 13.

    Hanasaki, N. et al. A global water scarcity assessment under shared socio-economic pathway — Part 2: water availability and scarcity. Hydrol. Earth Syst. Sci. 17, 2393–2413 (2013).

  14. 14.

    Wada, Y., Gleeson, T. & Esnault, L. Wedge approach to water stress. Nat. Geosci. 7, 615–617 (2014).

  15. 15.

    Engel, K., Jockiel, D., Kraljevic, A., Geiger, M. & Smith, K. Big Cities. Big Waters. Big Challenges. Water in an Urbanizing World (WWF, Berlin, 2011).

  16. 16.

    Padowski, J. C. & Gorelick, S. M. Corrigendum: global analysis of urban surface water supply and vulnerability (2014 Environ. Res. Lett. 9, 104004). Environ. Res. Lett. 9, 119501 (2014).

  17. 17.

    Moss, R. H. et al. The next generation of scenarios for climate change research and assessment. Nature 463, 747–756 (2010).

  18. 18.

    van Vuuren, D. P. et al. A new scenario framework for climate change research: scenario matrix architecture. Clim. Change 122, 373–386 (2014).

  19. 19.

    van Vuuren, D. P. & Carter, T. R. Climate and socio-economic scenarios for climate change research and assessment: reconciling the new with the old. Clim. Change 122, 415–429 (2014).

  20. 20.

    Wimmer, F. et al. Modelling the effects of cross-sectoral water allocation schemes in Europe. Clim. Change 128, 229–244 (2015).

  21. 21.

    Eisner, S. et al. An ensemble analysis of climate change impacts on streamflow seasonality across 11 large river basins. Clim. Change 141, 401–417 (2017).

  22. 22.

    aus der Beek, T. et al. Modelling historical and current irrigation water demand on the continental scale: Europe. Adv. Geosci. 27, 79–85 (2010).

  23. 23.

    Richter, B. D., Davis, M. M., Apse, C. & Konrad, C. A presumptive standard for environmental flow protection. River Res. Appl. 28, 1312–1321 (2012).

  24. 24.

    Hoekstra, A. Y. et al. Global monthly water scarcity: blue water footprints versus blue water availability. PLoS ONE 7, e32688 (2012).

  25. 25.

    Gleeson, T., Wada, Y., Bierkens, M. & van Beek, L. Water balance of global aquifers revealed by groundwater footprint. Nature 488, 197–200 (2012).

  26. 26.

    BGR/UNESCO. Groundwater Resources of the World 1:25,000,000 (BGR, 2008); https://www.whymap.org

  27. 27.

    Gleick, P. H. Global freshwater resources: soft-path solutions for the 21st century. Science 302, 1524–1528 (2003).

  28. 28.

    Molden, D. et al. Improving agricultural water productivity: between optimism and caution. Agric. Water Manage. 97, 528–535 (2010).

  29. 29.

    Zhao, X. et al. Burden shifting of water quantity and quality stress from megacity Shanghai. Water Resour. Res. 52, 6916–6927 (2016).

  30. 30.

    Cosgrove, W. J. & Loucks, D. P. Water management: current and future challenges and research directions. Water Resour. Res. 51, 4823–4839 (2015).

  31. 31.

    Kiparsky, M., Sedlack, D. L., Thompson, B. H. Jr. & Truffer, B. The innovation deficit in urban water: the need for an integrated perspective on institutions, organizations, and technology. Environ. Eng. Sci. 30, 395–408 (2013).

  32. 32.

    Schneider, C., Flörke, M., Eisner, S. & Voss, F. Large scale modelling of bankfull flow: an example for Europe. J. Hydrol. 408, 235–245 (2011).

  33. 33.

    Verzano, K. et al. Modeling variable river flow velocity on continental scale: current situation and climate change impacts in Europe. J. Hydrol. 424–425, 238–251 (2012).

  34. 34.

    Lehner, B., Verdin, K. & Jarvis, A. New global hydrography derived from spaceborne elevation data. Eos. Trans. 89, 93–94 (2008).

  35. 35.

    Schneider, C., Flörke, M., De Stefano, L. & Petersen-Perlman, J. D. Hydrological threats for riparian wetlands of international importance — a global quantitative and qualitative analysis. Hydrol. Earth Syst. Sci. 21, 2799–2815 (2017).

  36. 36.

    Eisner, S. Comprehensive Evaluation of the WaterGAP3 Model across Climatic, Physiographic, and Anthropogenic Gradients. PhD thesis, Univ. Kassel (2016).

  37. 37.

    Döll, P. et al. Impact of water withdrawals from groundwater and surface water on continental water storage variations. J. Geodyn. 59–60, 143–156 (2012).

  38. 38.

    Shiklomanov, I. in World Water Scenarios: Analyzing Global Water Resources and Use (ed. Rijsberman, F. R.) Ch. 12, 160–203 (Earthscan Publications, London, 2000).

  39. 39.

    Utility Data Institute. World Electric Power Plants Data Base (Platts Energy InfoStore, 2004); https://infostore.platts.com/infostore/samples/descmeth.pdf

  40. 40.

    Flörke, M., Bärlund, I. & Kynast, E. Will climate change affect the electricity production sector? A European study. J. Water Clim. Change 3, 44–54 (2012).

  41. 41.

    Siebert, S., Henrich, V., Frenken, K. & Burke, J. Update of the Global Map of Irrigation Areas to Version 5 (University of Bonn, Bonn; FAO, Rome; 2013).

  42. 42.

    Siebert, S. et al. A global data set of the extent of irrigated land from 1900 to 2005. Hydrol. Earth Syst. Sci. 19, 1521–1545 (2015).

  43. 43.

    Rohwer, J., Gerten, D. & Lucht, W. Development of Functional Types of Irrigation for Improved Global Crop Modelling Report 104 (Potsdam Institute for Climate Impact Research, Potsdam, 2007).

  44. 44.

    Rost, S. et al. Agricultural green and blue water consumption and its influence on the global water system. Water Resour. Res. 44, W09405 (2008).

  45. 45.

    Alcamo, J. et al. Development and testing of the WaterGAP 2 global model of water use and availability. Hydrolog. Sci. J. 48, 317–337 (2003).

  46. 46.

    FAOSTAT. Live Animals and Livestock Primary, 2014. http://faostat.fao.org/site/339/default.aspx (FAO, 2014).

  47. 47.

    McDonald, R. I., Weber, K. F., Padowski, J., Boucher, T. & Shemie, D. Estimating watershed degradation over the last century and its impact on water-treatment costs for the world’s large cities. Proc. Natl Acad. Sci. USA 113, 9117–9122 (2016).

  48. 48.

    Balk, D. in Global Mapping of Human Settlement: Experiences, Data Sets, and Prospects (eds. Gamba, P. & Herold, M.) 145–161 (Taylor and Francis, New York, 2009).

  49. 49.

    CIESIN, Columbia University, IFPRI, World Bank, CIAT. Global Rural–Urban Mapping Project (GRUMP): Urban Extents (Center for International Earth Science Information Network, 2004); http://sedac.ciesin.columbia.edu/data/collection/grump-v1

  50. 50.

    O’Neill, B. C. et al. A new scenario framework for climate change research: the concept of shared socioeconomic pathways. Clim. Change 122, 387–400 (2014).

  51. 51.

    van Vuuren, D. P. et al. The representative concentration pathways: an overview. Clim. Change 109, 5–31 (2011).

  52. 52.

    Piani, C. et al. Statistical bias correction of global simulated daily precipitation and temperature for the application of hydrological models. J. Hydrol. 395, 199–215 (2010).

  53. 53.

    Hempel, S., Frieler, K., Warszawski, L., Schewe, J. & Piontek, F. A. Trend-preserving bias correction — the ISI-MIP approach. Earth Syst. Dyn. 4, 219–236 (2013).

  54. 54.

    Warszawski, L. et al. The inter-sectoral impact model intercomparison project (ISI-MIP): project framework. Proc. Natl Acad. Sci. USA 111, 3228–3232 (2014).

  55. 55.

    Masui, T. et al. An emission pathway for stabilization at 6 Wm−2 radiative forcing. Clim. Change 109, 59–76 (2011).

  56. 56.

    Wada, Y. et al. Multimodel projections and uncertainties of irrigation water demand under climate change. Geophys. Res. Lett. 40, 4626–4632 (2013).

  57. 57.

    Brauman, K. A., Richter, B. D., Postel, S., Malsy, M. & Flörke, M. Water depletion: an improved metric for incorporating seasonal and dry-year water scarcity into water risk assessments. Elem. Sci. Anth. 4, 83 (2016).

  58. 58.

    Poff, N. L. et al. The natural flow regime. Bioscience 47, 769–784 (1997).

  59. 59.

    Poff, N. L. & Zimmerman, J. K. H. Ecological responses to altered flow regimes: a literature review to inform the science and management of environmental flows. Freshw. Biol. 55, 194–205 (2010).

  60. 60.

    Acreman, M. C. et al. The changing role of ecohydrological science in guiding environmental flows. Hydrol. Sci. J. 59, 433–450 (2014).

  61. 61.

    Müller Schmied, H. et al. Sensitivity of simulated global-scale freshwater fluxes and storages to input data, hydrological model structure, human water use and calibration. Hydrol. Earth Syst. Sci. 18, 3511–3538 (2014).

  62. 62.

    Portmann, F. T., Döll, P., Eisner, S. & Flörke, M. Impact of climate change on renewable groundwater resources: assessing the benefits of avoided greenhouse gas emissions using selected CMIP5 climate projections. Environ. Res. Lett. 8, 024023 (2013).

Download references

Author information

Affiliations

  1. Center for Environmental Systems Research, University of Kassel, Kassel, Germany

    • Martina Flörke
    •  & Christof Schneider
  2. Worldwide Office, The Nature Conservancy, Washington, DC, USA

    • Robert I. McDonald

Authors

  1. Search for Martina Flörke in:

  2. Search for Christof Schneider in:

  3. Search for Robert I. McDonald in:

Contributions

M.F. and R.I.M. designed the study, M.F. drafted the manuscript and performed the analyses. R.I.M. provided all information on basin transfers, C.S. implemented basin transfers into the model and prepared global gridded data of water availability and sectoral water uses. All authors discussed the results and contributed to the manuscript.

Competing interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to Martina Flörke.

Supplementary information

  1. Supplementary Information

    Supplementary Tables 1–6.