Hydrology and Earth System Sciences | Hanasaki et al. 
Humans abstract water from various sources to sustain their livelihood and society. Some global hydrological models (GHMs) include explicit schemes of human water abstraction, but the representation and performance of these schemes remain limited. We substantially enhanced the water abstraction schemes of the H08 GHM. This enabled us to estimate water abstraction from six major water sources, namely, river flow regulated by global reservoirs (i.e., reservoirs regulating the flow of the world’s major rivers), aqueduct water transfer, local reservoirs, seawater desalination, renewable groundwater, and nonrenewable groundwater. In its standard setup, the model covers the whole globe at a spatial resolution of 0.5° × 0.5°, and the calculation interval is one day. All the interactions were simulated in a single computer program and the water balance was always strictly closed at any place and time during the simulation period. A global hydrological simulation was conducted to validate the performance of the model for the period of 1979–2013. The simulated water fluxes for water abstraction were validated against those reported in earlier publications, and showed a reasonable agreement at the global and country level. The simulated monthly river discharge and terrestrial water storage (TWS) for six of the world’s most significantly human-affected river basins were compared with river gauging observations and a satellite product of the Gravity Retrieval and Climate Experiment (GRACE) mission. It showed that the simulation including the newly added schemes outperformed the simulation without them. The results indicated that, in 2000, of the 3628 km3 yr−1 global freshwater requirement, 2839 km3 yr−1 was taken from surface water and 789 km3 yr−1 from groundwater. Streamflow, aqueduct water transfer, local reservoirs, and seawater desalination accounted for 1786, 199, 106, and 1.8 km3 yr−1 of the surface water, respectively. The remaining 747 km3 yr−1 freshwater requirement was unmet, or surface water was not available when and where it was needed in our simulation. Renewable and nonrenewable groundwater accounted for 607 and 182 km3 yr−1 of the groundwater total, respectively. Each source differed in its renewability, economic costs for development, and environmental consequences of usage. The model is useful for performing global water resource assessments by considering the aspects of sustainability, economy, and environment.
Full text can be found here.