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

The study addresses how co-occurring freshwater stress (high withdrawals relative to streamflow) and trends in freshwater storage (drying or wetting) impact human societies and ecosystems when considered together at a global, basin scale. While global analyses often evaluate water scarcity or broader water security, few jointly quantify social and ecological vulnerability to hydrologic change. The authors argue that understanding freshwater not only as impacted by society but also as a driver of social-ecological outcomes is essential to confronting global freshwater challenges. The purpose is to synthesize key global sociohydrological and ecohydrological datasets to: (1) assess the global co-occurrence of freshwater stress and freshwater storage trends at the basin scale; (2) analyze relationships between social adaptive capacity and ecological sensitivity with freshwater stress and storage trends; (3) derive a global gradient in social-ecological vulnerability and identify hotspot basins; and (4) evaluate integrated water resources management (IWRM) levels in hotspot basins. Analyses are conducted for 1204 large basins (median ~70,000 km²), using data aligned around 2015.

Two main bodies of prior work inform this study. First, freshwater scarcity studies typically compute ratios of water use to streamflow and/or streamflow per capita at basin scales, offering comparable global indicators of stress. However, they often apply uniform classification schemes that ignore spatial variability in social and ecological sensitivities and responses, and holistic derivatives (e.g., social water stress and water poverty indices) have mostly been evaluated nationally. Second, water security assessments integrate diverse indicators (physical, chemical, socioeconomic, institutional) using various aggregation schemes. While comprehensive, they can obscure the specific impacts of water scarcity because many combinations of indicators can yield similar overall security scores, and they often neglect interactions among elements of water security that shape system behavior. Social-ecological systems research emphasizes coupled human–natural feedbacks, yet global-scale assessments linking hydrological drivers to social and ecological vulnerability remain limited. The authors aim to combine strengths of scarcity and security approaches, integrating social-ecological concepts to better represent exposure, sensitivity, and adaptability to freshwater stress and storage change.

Study design and data: The analysis uses pre-existing, published, open datasets (12 geospatial datasets plus a basin scheme), harmonized to 0.5° resolution and summarized to HydroBASINS level 4 basins (n = 1204 after masking regions with poor coverage). Data are aligned to ~2015 where possible; freshwater storage trends come from satellite-based observations (2002–2016). Basin-level summaries use area-weighted averages for intensive variables and within-basin sums for extensive variables. Key indicators: - Freshwater stress (W/Q): Annual freshwater withdrawal divided by annual streamflow per basin; basins with W/Q ≥ 10% are stressed, ≥ 40% highly stressed. - Freshwater storage trend (ΔS): Year-over-year trend in total water storage (groundwater, soil moisture, surface water, canopy water, snow/ice water equivalents). Negative trends denote drying, positive denote wetting. Basin freshwater status (exposure): To integrate stress and storage trend into a single exposure indicator, each is normalized by 0.4Q (the common threshold for high stress), bounded to [-1,1] for storage trends (after flipping sign so drying is positive) and [0,1] for stress, and then averaged. Negative composite values (wetting offsetting stress) are floored at 0. In basins affected by large earthquakes compromising storage trend estimates, the indicator defaults to the normalized stress component. Social-ecological sensitivity: Ecological sensitivity is derived by combining two global ecohydrological datasets: (i) de Graaf et al. estimates of groundwater head decline at which environmental flows are transgressed (sensitivity of environmental flows to groundwater depletion where pumping occurs), and (ii) the water-specific component of Seddon et al.’s Vegetation Sensitivity Index (sensitivity of vegetation productivity to water anomalies). Each dataset is transformed to area-weighted percentiles at 0.5°, averaged within basins, then combined and normalized by the global maximum to yield [0,1]. Social adaptive capacity is taken from Varis et al. and inverted so higher values reflect lower adaptability (greater social sensitivity). Ecological sensitivity and inverted adaptive capacity are combined using a fuzzy sum operator to yield a composite social-ecological sensitivity in [0,1]. Vulnerability and hotspot classification: Vulnerability per basin is computed as the product of social-ecological sensitivity and basin freshwater status. Given heavy-tailed vulnerability distributions, the Head/Tail Breaks algorithm (applied with three iterations for consistency) classifies basins into four classes: low (non-hotspot), transitional (moderate), high (hotspot), and very high (hotspot). Uncertainty and sensitivity analyses: Two uncertainty analyses propagate spatially uniform and spatially variable uncertainties through 10,000 realizations each; robustness is assessed by the frequency of basins remaining at least transitional. A methodological sensitivity analysis tests 24 alternative configurations to assess the impact of subjective choices. Results show high consistency of identified transitional and hotspot basins across analyses. IWRM comparison: Basin-level vulnerability is compared with national IWRM implementation scores from the UNEP/UNEP-DHI IWRM Data Portal (SDG 6.5.1). Transboundary status is noted, and differences in IWRM levels between transboundary and non-transboundary hotspot basins are assessed.

- Co-occurrence of stress and drying: Of 478 stressed basins (W/Q > 0.10), 201 (42%) are simultaneously losing freshwater storage, concentrated in the southwestern and southern USA, northeastern Brazil, central Argentina, Algeria, the Middle East, the Caucasus, northern India, and northern China. Ninety-eight stressed basins (21%) are wetting; 179 have small, non-definitive trends. In non-stressed basins, drying and wetting occur in 23% and 32%, respectively. Thus, stressed regions are disproportionately drying, whereas non-stressed regions show no clear overall trend. - Exposed social-ecological activity: Approximately 2.2 billion people, 27% of global food crop production, and 28% of global GDP are in stressed and drying basins. Conversely, 1.2 billion people, 24% of food production, and 19% of GDP are in stressed and wetting basins. Taxonomic biodiversity is lower in stressed/drying basins and higher in unstressed/wetting basins; numbers of Ramsar wetlands are similar between stressed/drying and stressed/wetting categories. - Social adaptability versus freshwater status: Seventy-three basins exhibit both severe freshwater status and low social adaptability, concentrated in Northern/Eastern Africa, the Arabian Peninsula, and Western/Central/Southern Asia (also parts of NE Brazil, Southern Africa, northern China). These basins include ~1.2 billion people, 12% of global food crops, and 6% of GDP. In contrast, 119 and 49 basins with severe freshwater status have moderate or high social adaptability, respectively, in regions such as the American Southwest/Mexico, Chile/Argentina, around the Caspian Sea, western Australia, and the North China Plain. Economic disparities are stark: low-adaptability severe-status basins contain 17% of population but 6% of GDP; moderate-or-higher adaptability severe-status basins contain 14% of population but 18% of GDP. - Hotspot basins: 168 basins (14% of all; 11% of study land area) are classified as hotspots (90 high, 78 very high). Over 1.5 billion people, 17% of global food crop production, and 13% of global GDP are within hotspot basins; about 300 million people, 4% of food, and 4% of GDP are in the very high class. Hotspots occur across Argentina, NE Brazil, the American southwest, Mexico, Northern/Eastern/Southern Africa, the Middle East and Arabian Peninsula, the Caucasus, West Asia, northern India and Pakistan, SE Asia, and northern China. Hotspot basins encompass 157 Ramsar wetlands. - Social vs ecological vulnerability patterns: Affluent regions (e.g., southwestern USA, western Australia) can have high ecological but low social vulnerability; some regions (Eastern Africa, NE India) have high social but lower ecological vulnerability, highlighting the need for integrated strategies. - IWRM implementation: There is no global correlation between basin vulnerability and national IWRM implementation levels; high vulnerability is not associated with higher IWRM implementation. One-third (36%) of hotspot basins are transboundary; transboundary hotspots have lower mean IWRM scores than non-transboundary (50 vs 56), indicating a need for hydro-diplomacy and multilateral cooperation. Countries with low IWRM and very high vulnerability include Afghanistan, Algeria, Argentina, Egypt, India, Iraq, Kazakhstan, Mexico, Somalia, Ukraine, Uzbekistan, and Yemen. - Robustness: Across 10,000-run uncertainty analyses, 98% (uniform) and 96% (spatially variable) of identified transitional/hotspot basins are retained as at least transitional in over half of realizations; alternative methodological configurations also show consistent hotspot identification.

The findings reveal that existing water-stressed basins are disproportionately losing freshwater storage, compounding socio-ecological risks. By explicitly combining exposure (co-occurring stress and storage loss) with ecological sensitivity and social adaptive capacity, the study identifies where harms are most likely, rather than where absolute activity is greatest. This addresses the research objectives by (i) mapping co-occurrence of stress and storage trends, (ii) situating these within social and ecological context, and (iii) classifying a global gradient of vulnerability to prioritize basins. Significance and implications: Hotspot basins face elevated risks of environmental flow breaches, increased drought frequency, reduced ecosystem drought resilience, impacts to groundwater-dependent ecosystems, and altered land–atmosphere energy balance. Irrigation-driven depletion can propagate beyond basin boundaries by altering moisture recycling and precipitationsheds. Socially, shrinking water availability threatens domestic, industrial, and agricultural uses; declining water tables risk well failures and can exacerbate inequality as wealthier users drill deeper. Elevated vulnerability aligns with a substantial share of recent water conflicts (about 68% of conflicts since 2000 occur in transitional or hotspot basins), underscoring security implications. The analysis highlights economic inequality: low-adaptability severe-status basins contain a large population share but a small GDP share, suggesting that SDG 6.4 (reducing water scarcity) should be linked to SDG 10 (reducing inequality). The lack of alignment between IWRM implementation and vulnerability suggests inconsistent or reactive governance; improving IWRM, especially in transboundary hotspots, will require hydro-diplomacy, coordinated institutions, and integrated water resources management to balance human and ecosystem needs.

This study integrates freshwater stress, freshwater storage trends, ecological sensitivity, and social adaptive capacity to map a global gradient of social-ecological vulnerability and identify 168 hotspot basins for prioritization. It shows that stressed basins are disproportionately drying and that more than 1.5 billion people, 17% of global food crop production, and 13% of global GDP reside in hotspot basins. The results emphasize that vulnerability depends on both exposure and capacity/sensitivity, not solely on scarcity metrics, and reveal gaps between vulnerability and IWRM implementation, especially in transboundary contexts. Future directions include: (i) developing process-based, globally consistent ecological sensitivity indicators that integrate groundwater, surface water, and soil moisture processes; (ii) advancing sub-basin and sub-grid assessments of social adaptive capacity and ecological sensitivity; (iii) incorporating probabilistic exposure, thresholds, and early-warning indicators; (iv) extending analyses to other hydrological dimensions (seasonal/interannual storage variability, water quality) and indirect/non-local effects (e.g., virtual water/food trade); and (v) evaluating the effectiveness of IWRM and other governance interventions in reducing vulnerability over time.

- Ecological sensitivity indicator is constructed from two datasets (environmental flow sensitivity to groundwater decline and water-driven vegetation sensitivity) and does not capture the full spectrum of ecosystem processes/functions affected by freshwater changes. - Social adaptive capacity is represented by a general, reductionist proxy; it does not resolve sub-national heterogeneity or sector-specific adaptive responses. - Analyses are at large-basin scale (HydroBASINS level 4) with 0.5° inputs; sub-grid and within-basin variability are not addressed. - Freshwater stress is computed as within-basin W/Q without upstream–downstream allocation rules and without special arid-region treatment used in some other datasets. - Storage trends rely on satellite-derived total water storage with limited spatial resolution and potential contamination by large earthquakes; some basins default to stress-only exposure where trends are unreliable. - The approach evaluates the current state and trend deterministically; it does not incorporate probabilistic exposure or proximity-to-threshold scaling. - Only direct impacts of stress and storage loss are considered; indirect and teleconnected effects (e.g., virtual water trade) are not included due to data limitations. - Data are aligned to ~2015; temporal changes since then are not captured.