Hotspots for social and ecological impacts from freshwater stress and storage loss

The global co-occurrence of freshwater stress and freshwater storage trends

We mapped freshwater stress and trends in freshwater storage at the basin scale and analyzed the co-occurrence of these phenomena (Fig. 1).

Fig. 1: Global co-occurrence of freshwater stress and storage trends.

a Freshwater stress, derived from freshwater withdrawal and streamflow datasets (see “Methods” section). b Freshwater storage trend per basin. c Combinations of freshwater stress and storage trend per basin, which together derive basin freshwater status (shown in Fig. 2b). Values overlaying the legend indicate the number of basins satisfying each set of conditions. For categorical plotting purposes only, ±3 mm year−1 is used as the threshold denoting a clear directional storage trend, based on the error level of the underlying observations25. dg The exposure of social-ecological activity to freshwater stress and storage trends. Each plot represents storage trends as the x-axis coordinate, and log-transformed freshwater stress as the y-axis coordinate with the size of each circle based on the basin’s value respective to each plotting dimension.

Full size image

Freshwater stress represents the state of demand-driven water scarcity15 and is defined as the ratio of freshwater withdrawal to streamflow (Fig. 1a). Trends in freshwater storage, conversely, represent the evolution of total storage, defined as the vertical sum of groundwater, soil moisture, surface water, and snow water equivalent storages (Fig. 1b). Freshwater stress and storage are linked, as freshwater storage becomes a required source of water during periods when demands exceed supply. As climate change intensifies hydrological extremes globally, the strategic importance of the world’s largest store of liquid freshwater, groundwater, will only continue to increase24. Though studies have focussed on global assessments of freshwater stress13,14,15 and trends in freshwater storage9, no study to date has mapped these two variables against one another. Doing so provides important context to differentiate basins of equal freshwater stress, as drying trends are likely to exacerbate challenges derived from freshwater stress, while wetting trends may yield offsetting effects. However, as freshwater stress calculations do not differentiate between withdrawals sourced from streamflow or storage, the two variables are not necessarily independent.

We found that 201 (42%) of the 478 currently stressed basins (withdrawal/streamflow > 0.10) are simultaneously losing freshwater storage (Fig. 1c). These basins are located in south and southwestern USA, northeastern Brazil, central Argentina, Algeria, and concentrate throughout the Middle East, the Caucasus, northern India, and northern China. Predominantly, these regions are agriculturally significant and heavily irrigated9, with the exception of a few basins in South America whose trends are likely the product of natural variability9. Conversely, 98 (21%) of the currently stressed basins are gaining freshwater storage. The storage trends in these basins have largely been attributed to natural variability with the exception of central India, whose trends are partially attributed to groundwater recovery following groundwater policy change9. The remaining 179 stressed basins have freshwater storage trends that are smaller than can be definitively interpreted from the satellites monitoring these trends25. This skew towards negative storage trends (i.e., drying) in the world’s water-stressed basins dissipates and even reverses in the non-stressed basins, where drying and wetting trends are found in 23% and 32% of the 726 non-stressed basins, respectively. While previous work has shown that the world’s dry regions are becoming drier while the wet regions are becoming wetter26, this work reveals that the stressed regions of the world are becoming drier while the non-stressed regions of the world have no clear overall trend in freshwater storage.

The encompassed human population, food crop production, gross domestic product (GDP), biodiversity, and wetlands enumerate the potential social-ecological impacts from the current state of global freshwater stress and storage trends. Around 2.2 billion people, 27% of global food crop production, and 28% of global GDP live, grow, and situate in freshwater stressed basins that are drying (Fig. 1d–f). These totals represent an upper limit as not all social and ecological activity within these basins will be affected by freshwater stress and storage loss, which will depend on local levels of adaptive capacity and ecological sensitivity22 (our focus in the subsequent sections). Conversely, 1.2 billion people, 24% of global food crop production, and 19% of global GDP are found in stressed basins that are wetting. We find less taxonomic biodiversity in the freshwater stressed and drying basins, and greater biodiversity in unstressed and wetting basins. Roughly the same number of wetlands of international importance are found in stressed and drying basins as in stressed and wetting basins. While these totals represent the magnitude of potentially affected biodiversity and wetlands, taxonomic biodiversity is only one of many critical facets of biodiversity27, and freshwater stress and storage trends are but two of many variables impacting global biodiversity28. Thus, we urge caution in interpreting the role of freshwater stress and storage in driving differences in these biodiversity distributions.

The most vulnerable populations to freshwater stress and storage loss

To better characterize social vulnerability, freshwater stress and storage loss must be placed in the context of social adaptability. We mapped and analyzed the co-occurrence of freshwater stress and storage trends with an existing global dataset of social adaptive capacity23 summarized at the basin scale (Fig. 2). Social adaptive capacity (Fig. 2a), or adaptability, represents “the ability of the system to respond to disturbances”29 and is derived based on input indicators of governance, economic strength, and human development. This consideration of social adaptability enables more representative estimates of social, agricultural, and economic activity that are vulnerable to the co-occurrence of freshwater stress and storage loss. To consider freshwater stress and storage loss together, we developed the basin freshwater status indicator (Box 1) where higher values indicate co-occurring freshwater stress and storage loss (Fig. 2b, see “Methods” section).

Fig. 2: The relationship between basin freshwater status and social adaptive capacity.

a Social adaptive capacity, or adaptability, per basin. b Basin freshwater status, representing the combination of freshwater stress and storage trend per basin (see “Methods” section). c Combinations of basin freshwater status and social adaptability. Values overlaying the legend indicate the number of basins satisfying each set of conditions. dg The exposure of social-ecological activity to basin freshwater status (x-axis coordinate) and social adaptive capacity (y-axis coordinate), with the size of each circle scaled based on the basin’s value respective to each plotting dimension. These distributions are summarized below each plot. P notation represents the percentile distribution.

Full size image

We found 73 basins to possess low levels of social adaptability and severe basin freshwater status (Fig. 2c). These basins concentrate in Northern, and Eastern Africa, the Arabian Peninsula, and Western, Central, and Southern Asia; although vulnerable basins are also found in northeast Brazil, Southern Africa, and northern China. These basins encompass approximately 1.2 billion people, 12% of global food crop production, and 6% of global GDP (Fig. 2d–f). Conversely, 119 and 49 basins are found to have similarly severe basin freshwater status yet have moderate or high levels of social adaptability, respectively. These basins are located in southwestern USA and Mexico, Chile and Argentina, the Arabian Peninsula, regions surrounding the Caspian Sea, western Australia, and the North China Plain.

These differences in social adaptability across basins with severe freshwater status (i.e., co-occurring freshwater stress and storage loss) raise important economic considerations. First, greater social adaptability likely coincides with greater technological and economic capacity to pursue development. This development may consume greater volumes of freshwater and drive basins towards greater levels of freshwater stress or storage loss, while simultaneously increasing institutional and technical capacity to cope with limited water resources. Furthermore, freshwater stress and storage loss are not certain to induce negative economic impacts on basins, and can lead to positive impacts if a region is able to leverage its comparative advantages (e.g., irrigation efficiency) among other stressed regions30. Second, the divergent economic situation facing basins with severe freshwater status is particularly evident on a per-capita basis. In severe freshwater status, low adaptability basins, there resides 17% of the global population yet only 6% of global GDP. Conversely, in severe freshwater status basins with moderate-and-greater social adaptability, there resides 14% of the global population and an outsized 18% of global GDP (Fig. 2d, f). It is thus paramount that global initiatives prioritize and link economic inequality with freshwater goals. One such example is Sustainable Development Goal (SDG) 6.4 (“reduce the number of people suffering from water scarcity”), which we argue should increasingly be linked to targets of SDG 10 (“reduce inequality within and among countries”).

Hotspot basins found on all continents

We mapped the global gradient in social-ecological vulnerability to freshwater stress and storage loss at the basin scale and, from this, identified those with the greatest vulnerability as hotspot basins (Fig. 3). Hotspot mapping has been a successful endeavor within the field of conservation biogeography31,32, and many global hydrology studies have identified regions of exceptional water scarcity and security challenges e.g.,13,14,15,17,18,19. Here, we seek to combine and apply these concepts in an integrated global social-ecological vulnerability context. As a useful reference, biodiversity hotspots aim to “maximize the number of species “saved” given available resources” by asking “where are places rich in species and under threat?”33. For comparison, the aim of our hotspot mapping is to ‘minimize the social and ecological impacts of freshwater stress and storage loss given available resources’ by asking “what basins with sensitive ecosystems and limited social adaptive capacity are exposed to freshwater stress and storage loss?”

Fig. 3: Hotspot basins for social and ecological impacts from freshwater stress and storage loss.

ad Social-ecological vulnerability results. a Hotspot basins of social-ecological vulnerability to freshwater stress and storage loss. b Vulnerability classification, based on the product of basin freshwater status and social-ecological sensitivity to freshwater stress and storage loss (see “Methods” section). c Histograms of the global distribution of vulnerability classes by basin count and surface area. d Summarized social-ecological activity within transitional and hotspot basins. e Ecological vulnerability results, presented as vulnerability classes. f Social vulnerability results, presented as vulnerability classes. Vulnerability classes for e and f are derived using the same methods as shown for social-ecological vulnerability in b.

Full size image

We conceptualize vulnerability as the product of (i) ecological sensitivity, (ii) social adaptive capacity, and (iii) basin freshwater status. To represent ecological sensitivity, we derived an indicator using data products from two global ecohydrological studies that assess broad ecosystem sensitivity to freshwater storage and use (see “Methods” section). To represent social adaptability, we utilized the same adaptive capacity dataset as used in the previous section (Fig. 2a). To classify the derived global vulnerability results into hotspot basins, we implemented a simple classification algorithm developed for heavy-tailed distributions34, which appropriately describes the global vulnerability distribution.

The most vulnerable basins are constrained to regions confronting co-occurring freshwater stress and storage loss. When considering social and ecological vulnerability individually (Fig. 3e, f), we find spatial variation between ecological vulnerability (Fig. 3e) and social vulnerability (Fig. 3f). For instance, several basins in affluent nations with sensitive ecosystems reveal high ecological vulnerability but low social vulnerability (southwestern USA; western Australia). Conversely, several basins in Eastern Africa and northeastern India possess high social vulnerability but low to moderate ecological vulnerability. While these differences are notable and could impact regional strategies, it remains essential in most, if not all, regions that social and ecological vulnerabilities be confronted simultaneously4. For this purpose, we combined ecological sensitivity and adaptive capacity indicators into a combined social-ecological sensitivity indicator (see “Methods” section) to map combined social-ecological vulnerability (Fig. 3a).

We identify 168 basins, representing 14% of all basins and 11% of the global land area considered in our study, as vulnerability hotspots (Fig. 3a–c). These hotspot basins consist of basins receiving “high” and “very high” vulnerability scores through our classification procedure. Of the 168 basins, 78 (6% of all basins) are classified in the most-severe “very high” vulnerability class, while 90 (7% of all basins) are classified in the “high” vulnerability class. We also identified 232 basins (19% of all basins) as “transitional” basins, which are not classified alongside basins with null vulnerability yet also do not possess extreme values within the global vulnerability distribution. The 78 hotspot basins with “very high” vulnerability represent the multiple epicenters for potential social and ecosystem impacts from freshwater stress and storage loss. These basins are found in Argentina, northeastern Brazil, the American southwest, Mexico, Northern, Eastern, and Southern Africa, the Middle East and Arabian Peninsula, the Caucasus, West Asia, northern India and Pakistan, Southeastern Asia, and northern China.

A total of over 1.5 billion people, 17% of global food crop production, and 13% of global GDP are found within hotspot basins (Fig. 3d). Of these, ~300 million people, 4% of global food crop production, and 4% of global GDP situate within the 78 “very high” vulnerability basins. Consistent with the relationship between biodiversity and basin freshwater status, we find the most vulnerable basins to be less taxonomically biodiverse than less vulnerable basins. While it is possible that these lower biodiversity levels may have eroded due to freshwater stress and storage loss, a proper investigation is outside the scope of this study and would require a wider array of pressures to be considered. The hotspot basins encompass 157 wetlands of international importance, which we highlight to prioritize their conservation in these vulnerable environments (Supplementary Table 2).

While the degree of social-ecological activity within hotspot basins is substantial, the global proportion of each dimension found in hotspot basins is roughly proportional to the fraction of basins within each vulnerability class. Thus, as the hotspot basins do not contribute disproportionately to global totals of social-ecological activity, we find it important to restate and clarify the motivating purpose of this hotspot mapping. The hotspot basins do not identify the greatest contributors to global social-ecological activity that face severe freshwater challenges. Rather, the hotspot basins are those with sensitive ecosystems and adaptability-limited societies exposed to the co-occurrence of freshwater stress and storage loss, and thus are the basins most likely to suffer social and ecological harms due to these freshwater conditions.

The identification of hotspot basins shows high levels of consistency across two uncertainty analyses and a sensitivity analysis focused on the impacts of subjective methodological decisions (Supplementary Section 4). We consider individually the impacts of (i) uniform over-estimation and under-estimation of each data input (spatially uniform uncertainty) and (ii) heterogeneous uncertainty in each data input (spatially variable uncertainty) on our hotspot basin results. Performing 10,000 realizations for each uncertainty analysis, we find that 98% of the identified transitional and hotspot basins are identified as at least transitional basins in over 50% the realizations considering spatially uniform uncertainty, and 96% when considering spatially variable uncertainty (Supplementary Figs. 8 and 9). The subjectivity-focused sensitivity analysis considered 24 alternative methodological configurations, and revealed that our identified transitional and hotspot basins are consistently identified across the majority of configurations (Supplementary Fig. 10).

Implementation of integrated water resources management is inconsistent across hotspot basins

We compared national implementation levels of integrated water resources management (IWRM) with our global vulnerability results (Fig. 4). For IWRM implementation data, we rely on the IWRM Data Portal35 which tracks progress on SDG 6.5.1 (“IWRM implementation at the national scale”).

Fig. 4: Integrated water resources management in hotspot basins.

a Map of IWRM implementation overlaid by hotspot basin results. b Scatterplot of individual basin values of social-ecological vulnerability (x-axis) and IWRM implementation (y-axis). Transboundary basins are represented by concentric red circles, with the number of circles representing the number of nations present within each basin. See text for interpretation of labels 1, 2, and 3.

Full size image

IWRM is defined as “a process which promotes the co-ordinated development and management of water, land and related resources, in order to maximize the resultant economic and social welfare in an equitable manner without compromising the sustainability of vital ecosystems”36, while the SDG framework notes that IWRM implementation “supports all Goals across the 2030 Agenda”37. Thus, as the IWRM paradigm seeks to guide management of water resources to minimize trade-offs between human well-being, ecological health, and water resources sustainability, assessing implementation levels of IWRM against our vulnerability results provides insight regarding the performance of IWRM globally while simultaneously emphasizing the broad sustainability implications within hotspot basins.

Globally, we find no direct relationship between vulnerability and IWRM implementation at the basin scale. There is thus a wide range of IWRM implementation across all levels of social-ecological vulnerability to freshwater stress and storage loss, and there is no indication that IWRM implementation levels are greatest where they are most needed. This finding likely derives from variations in proactive versus reactive governance and management approaches to freshwater challenges across the globe. As our analysis is conducted at a snapshot in time (input data align to ~2015), we can only generate hypotheses about the performance of IWRM globally. For example, basins with high levels of IWRM implementation and low vulnerability (label 1 in Fig. 4b) have either proactively implemented IWRM, have effectively reduced their vulnerability through IWRM implementation, or simply benefit from a favorable intersection of regional climate and economy.

Alternatively, basins with low levels of IWRM and low vulnerability can be categorized as non-proactive in their IWRM implementation (label 2 in Fig. 4b). We place particular emphasis here on basins with low levels of IWRM where vulnerability is high (label 3 in Fig. 4b), which we argue should be the priority basins and regions of SDG 6.5-focused initiatives. Identified nations with low levels of IWRM implementation and very high vulnerability include Afghanistan, Algeria, Argentina, Egypt, India, Iraq, Kazakhstan, Mexico, Somalia, Ukraine, Uzbekistan, and Yemen. As one-third (36%) of all hotspot basins are transboundary (Fig. 4b), improving basin-level IWRM implementation will require multilateralism and hydro-diplomacy and cannot be left to individual nations acting alone. Furthermore, we observe a lower level of IWRM implementation across hotspot basins that are transboundary versus non-transboundary hotspot basins (mean basin IWRM Data Portal score = 50 vs. 56), suggesting greater multilateralism and cooperation are needed in transboundary basins.

Source: Resources -

Complex marine microbial communities partition metabolism of scarce resources over the diel cycle

Deep learning increases the availability of organism photographs taken by citizens in citizen science programs