New methods of modeling water availability for agriculture under climate change: The US Cornbelt

This paper reports on new methods of linking climate change scenarios with hydrologic, agricultural an water planning models to study future water ava ilability for agriculture, an essential element of sustainability. The stud y is based on the integration of models of water supply and demand, and of crop growth and irrigation management. Consistent modeling assumptions, ava ilable databases, and scenario simulations are used to capture a range of p ossible future conditions. The linked models include WATBAL for water suppl y; CERES, SOYGRO, and CROPWAT for crop and irrigation modeling; and WEAP fo r water demand forecasting, planning and evaluation. These models are appli ed to the U.S. Cornbelt using forecasts of climate change, agricultural pro duction, population and GDP growth. Results suggest that, at least in the near term, the relative abundance of water for agriculture can be maintained under climate change conditions. Ho wever, increased water demands from urban growth, increases in reservoir ev aporation and increases in crop consumptive use must be accommodated by tim ely improvements in crop, irrigation and drainage technology, water managem ent, and institutions. These improvements are likely to require substantial resources and expertise. In the highly irrigated basins of the region, irr igation demand greatly exceeds industrial and municipal demands. When impro vements in irrigation efficiency are tested, these basins respond by reduci ng demand and lessening environmental stress with an improvement in system reliability, effects particularly evident under a high technology scenario. Rainfed lands in the Cornbelt are not forced to invest in irrigation, but there is some concern about increased water-logging during the spring and c onsequent required increased investment in agricultural drainage. One major water region in the Cornbelt also provides a useful caveat: change will no t necessarily be continuous and monotonic. Under one GCM scenario for the 2 010s, the region shows a significant decrease in system reliability, while the scenario for the 2020s shows an increase.