Global patterns in water flux partitioning: Irrigated and rainfed agriculture drives asymmetrical flux to vegetation over runoff

The partitioning of precipitation water input on land between green (evapotranspiration) and blue (runoff) water fluxes distributes the annually renewable freshwater resource among sectors and ecosystems. The patterns and main drivers of this partitioning are not fully understood around the global land area. We decipher the worldwide patterns and key determinants of this water flux partitioning and investigate its predictability based on a global machine learning model. Available data for 3,614 hydrological catchments and model application to the global land area agree in showing mostly larger green than blue water flux. Possible expansion/intensification of irrigated and/or rainfed agriculture to feed a growing human population, along with climate warming, will tend to increase this flux partitioning asymmetry, jeopardizing blue water security. The developed machine learning model presents a promising predictive tool for future blue and green water availability under various forthcoming climate and land-use change scenarios around the world.