SWAT-MODSIM-PSO optimization of multi-crop planning in the Karkheh River Basin, Iran, under the impacts of climate change
Majid Fereidoon,Manfred Koch
DOI: https://doi.org/10.1016/j.scitotenv.2018.02.234
2018-07-15
Abstract:Agriculture is one of the environmental/economic sectors that may adversely be affected by climate change, especially, in already nowadays water-scarce regions, like the Middle East. One way to cope with future changes in absolute as well as seasonal (irrigation) water amounts can be the adaptation of the agricultural crop pattern in a region, i.e. by planting crops which still provide high yields and so economic benefits to farmers under such varying climate conditions. To do this properly, the whole cascade starting from climate change, effects on hydrology and surface water availability, subsequent effects on crop yield, agricultural areas available, and, finally, economic value of a multi-crop cultivation pattern must be known. To that avail, a complex coupled simulation-optimization tool SWAT-LINGO-MODSIM-PSO (SLMP) has been developed here and used to find the future optimum cultivation area of crops for the maximization of the economic benefits in five irrigation-fed agricultural plains in the south of the Karkheh River Basin (KRB) southwest Iran. Starting with the SWAT distributed hydrological model, the KR-streamflow as well as the inflow into the Karkheh-reservoir, as the major storage of irrigation water, is calibrated and validated, based on 1985-2004 observed discharge data. In the subsequent step, the SWAT-predicted streamflow is fed into the MODSIM river basin Decision Support System to simulate and optimize the water allocation between different water users (agricultural, environmental, municipal and industrial) under standard operating policy (SOP) rules. The final step is the maximization of the economic benefit in the five agricultural plains through constrained PSO (particle swarm optimization) by adjusting the cultivation areas (decision variables) of different crops (wheat, barley, maize and "others"), taking into account their specific prizes and optimal crop yields under water deficiency, with the latter computed in the LINGO-sub-optimization module embedded in the SLMP-tool. For the optimization of the agricultural benefits in the KRB in the near future (2038-2060), quantile-mapping (QM) bias-corrected downscaled predictors for daily precipitation and temperatures of the HadGEM2-ES GCM-model under RCP4.5- and RCP8.5-emission scenarios are used as climate drivers in the streamflow- and crop yield simulations of the SWAT-model, leading to corresponding changes in the final outcome (economic benefit) of the SLMP-tool. In fact, whereas for the historical period (1985-2004) a total annual benefit of 94.2 million US$ for all multi-crop areas in KRB is computed, there is a decrease to 88.3 million US$ and 72.1 million US$ for RCP4.5 and RCP8.5, respectively, in the near future (2038-2060) prediction period. In fact, this future income decrease is due to a substantial shift from cultivation areas devoted nowadays to high-price wheat and barley in the winter season to low-price maize-covered areas in the future summers, owing to a future seasonal change of SWAT-predicted irrigation water available, i.e. less in the winter and more in the summer.