Future Changes in Global Runoff and Runoff Coefficient From CMIP6 Multi‐Model Simulation Under SSP1‐2.6 and SSP5‐8.5 Scenarios

Abstract This paper assesses the performances of runoff (Ro) and runoff coefficient ( α , the ratio of runoff to precipitation) simulations from 23 models during the historical period and then projects their future changes under the two emission scenarios (SSP1‐2.6 and SSP5‐8.5) in the Coupled Model Intercomparison Project. Compared with the UNH/GRDC Ro dataset (0.82 mm day −1 ), the multi‐model median (MME) Ro of 1995–2014 produces a comparable global mean magnitude (0.80 mm day −1 ), displays a similar spatial distribution of mean Ro, and also well captures the seasonal cycles at both global and basin scales. The global mean Ro of MME is projected to be increased by 0.01–0.02 mm day −1 (SSP1‐2.6) and 0.02–0.10 mm day −1 (SSP5‐8.5) during the twenty‐first century. Regional hotspots for strong increasing Ro appear across most areas of northern high latitudes, Africa, and southeastern Asia, with high inter‐model consistency. The global mean α is projected to be slightly decreased (−0.17 to −0.63%) except for the long‐term under the SSP5‐8.5 (0.26%). Although signs of changes in Ro vary with the river basins, periods, and scenarios, α in more than half (7 out of 12) river basins are projected to decrease. The uneven distributions of projected Ro changes over global land areas are related to the response of multiple hydroclimatic variables to the global warming. Given regions with inconstancy change signs of the projected precipitation, we speculate that changes in Ro are affected by more complicated hydroclimatic processes that warrant further investigations with physical‐based approaches.