Regional Modelling of Water Stress; Irrigation water requirement meets water availability in the Oum Er Rbia basin

Abstract

Current large scale hydrological models have little value for local water resource management, even when applied regionally on downscaled data. They lack the representation of the water-transfers, allocation strategies and productivity that shape the impacts of future climate change and increased water use. This study seeks ways to improve this representation, and tests three combinations of the large scale model PCR-GLOBWB and the crop model AquaCrop, applied to a semiarid Moroccan basin. In this region 90 percent of the water resources is required for irrigated food-production, and future scarcity is projected. The first two multi-model approaches extract water available to agriculture and express allocation to Tadla, the largest irrigated perimeter. Under historical application from 1979 to 2012, this results in a water stress index with monthly detail, and in productivity series for the dominant crops. For 2020 to 2050, RCP4.5 and RCP8.5 project an increased meteorological dryness. But only for the latter scenario does the water availability subsequently decrease, which limits crop growth severely after 2033. For the former scenario productivity increases because of CO2-fertilization. With the expressed allocation and inter-crop prioritization does this study find that production can be optimized for wheat. And that these valuable expressions for local impacts are currently lacking in the irrigation routines of large scale models. The third approach redirected AquaCrop calculations to the irrigation routine of PCR-GLOBWB to express hydrological impacts of agriculture and showed a successful application of three indicators that reflect different stages of drought.