Key points are not available for this paper at this time.
This study quantifies, spatially explicitly and in a consistent modeling framework (Lund‐Potsdam‐Jena managed Land), the global consumption of both “blue” water (withdrawn for irrigation from rivers, lakes and aquifers) and “green” water (precipitation) by rainfed and irrigated agriculture and by nonagricultural terrestrial ecosystems. In addition, the individual effects of human‐induced land cover change and irrigation were quantified to assess the overall hydrological impact of global agriculture in the past century. The contributions to irrigation of nonrenewable (fossil groundwater) and nonlocal blue water (e.g., from diverted rivers) were derived from the difference between a simulation in which these resources were implicitly considered (IPOT) and a simulation in which they were neglected (ILIM). We found that global cropland consumed >7200 km 3 year −1 of green water in 1971–2000, representing 92% (ILIM) and 85% (IPOT), respectively, of total crop water consumption. Even on irrigated cropland, 35% (ILIM) and 20% (IPOT) of water consumption consisted of green water. An additional 8155 km 3 year −1 of green water was consumed on grazing land; a further ∼44,700 km 3 year −1 sustained the ecosystems. Blue water consumption predominated only in intensively irrigated regions and was estimated at 636 km 3 year −1 (ILIM) and 1364 km 3 year −1 (IPOT) globally, suggesting that presently almost half of the irrigation water stemmed from nonrenewable and nonlocal sources. Land cover conversion reduced global evapotranspiration by 2.8% and increased discharge by 5.0% (1764 km 3 year −1 ), whereas irrigation increased evapotranspiration by up to 1.9% and decreased discharge by 0.5% at least (IPOT, 1971–2000). The diverse water fluxes displayed considerable interannual and interdecadal variability due to climatic variations and the progressive increase of the global area under cultivation and irrigation.
Rost et al. (Mon,) studied this question.