Terrestrial water storage (TWS) strongly modulates the hydrological cycle and is a key determinant of water availability and an indicator of drought. While historical TWS variations have been studied, future changes in TWS and the linkages to droughts remain unexamined. Here, using ensemble hydrological simulations, we show that climate change could reduce TWS in many regions, especially in the southern hemisphere. A strong inter-ensemble agreement indicates high confidence in the projected changes that are driven primarily by climate forcing, rather than land-water management activities. Declines in TWS translate to increase in future droughts. By the late-21 st century global land area and population in extreme-to-exceptional TWS drought could more than double, each increasing from 3% during 1976-2005 to 7% and 8%, respectively. Our findings highlight the importance of climate change mitigation to avoid adverse impacts on TWS and related droughts, and the need for adaptation to improve water resource management. TWS-the sum of continental water stored in canopies, snow and ice, rivers, lakes and 51 reservoirs, wetlands, soil, and groundwater-is a critical component of the global water and energy budget. It plays key roles in determining water resource availability 1 and modulating water flux interactions among various Earth system components 2 . Further, observed changes in TWS are inherently linked to droughts 2-6 , floods 7 , and global sea level change [8][9][10][11] . Despite such importance, global TWS remains less studied relative to hydrological fluxes (e.g., river discharge, evapotranspiration, and groundwater flow) owing to the lack of large-scale observations and challenges in explicitly resolving all TWS components in hydrological modeling 12 . This generally holds true for historical analyses; crucially, no study has to date examined the potential impacts of future climate change on global TWS. Recent modeling advancements 13 have improved the representation of TWS in global hydrological models 14,15 (GHMs) and land surface models 12 (LSMs). The Gravity Recovery and Climate Experiment (GRACE) satellite mission provided added opportunities to improve and validate TWS simulations in these models. GRACE TWS data and model simulations, often in combination, have been used for wide ranging applications including the assessment of water resources and impacts of human activities on the water cycle 14,16 , quantifying aquifer depletion 12,14,[17][18][19] , monitoring drought [3][4][5][6]20 , and assessing flood potential 7 . These studies have advanced the understanding of global TWS systems that are continually changing under natural hydro-climatic variability and accelerating human land-water management activities, but the 70 focus has been on historical variabilities in TWS. Further, future projections from general 71 circulation models (GCMs) have been used to quantify climate change impacts on hydrological 72 fluxes [21][22][23] and storages, but the projections of storages are limited to a subset of T...

Anthropogenic climate change is expected to affect global river flow. Here, we analyze time series of low, mean, and high river flows from 7250 observatories around the world covering the years 1971 to 2010. We identify spatially complex trend patterns, where some regions are drying and others are wetting consistently across low, mean, and high flows. Trends computed from state-of-the-art model simulations are consistent with the observations only if radiative forcing that accounts for anthropogenic climate change is considered. Simulated effects of water and land management do not suffice to reproduce the observed trend pattern. Thus, the analysis provides clear evidence for the role of externally forced climate change as a causal driver of recent trends in mean and extreme river flow at the global scale.

No abstract

No abstract

designed the analysis. S.L. processed the raw ISIMIP2b simulation output to generate the exposed land fraction data with help from T.G., D.N.B., J.V., and M.B.. K.E. generated the tropical cyclone tracks. F.Z. and S.W. ran the CaMa-Flood simulations and flood processing chain for generating the flood maps. W.T. performed all further analyses and wrote the manuscript with help from J.