Understanding the Basin-Wide Impact of Agricultural Irrigation on the Water Cycle in Dry Inland Areas: An Integrated Modeling Approach

Thursday, 18 December 2014: 9:45 AM
Yi Zheng1, Yong Tian1, Bin Wu1, Xin Wu1, Jinguo Li1, Chunmiao Zheng1,2 and Jie Liu1, (1)Peking University, Beijing, China, (2)University of Alabama, Tuscaloosa, AL, United States
Irrigation is indispensable to the agriculture in dry inland river basins, and may significantly alter the regional water cycle which is often featured by complicated surface water-groundwater (SW-GW) interaction. Although field-scale processes of irrigation water are clearly understood, the hydrological impact of irrigation at a large basin scale has not well examined. It poses a challenge to the water resources management. Our studies presented a modeling approach to address this challenge. GSFLOW, an integrated SW-GW model, was improved to represent irrigation practices in the model simulation, and applied to the HeiHe River Basin (HRB), the second largest inland river basin in China. The mid-stream area of HRB is a semi-arid region with extensive oasis agriculture, typical of western China. A series of studies have been conducted. First, the model was applied to fuse the limited hydrological observations in the area and generate a coherent understanding of the regional water cycle. Spatially and temporally detailed information on the impact of irrigation was then achieved. Second, an innovative optimization approach, coupling the Support Vector Machine (SVM) and SCE-UA algorithms, was proposed for the complex model. Allocation of surface water and groundwater for the irrigation was optimized under different management scenarios. The optimization results not only helped understand the hydrological processes, but provided insights into the water resources management. Third, the Probabilistic Collocation Method (PCM) was implemented to systematically address the modeling uncertainty. Spatial variation and temporal dynamics of the uncertainty associated with the irrigation practice in this area were revealed. The uncertainty results shed light on further data collection and model improvement. Overall, our study demonstrated the applicability and significant value of the integrated modeling approach in understanding the basin-wide impact of irrigation on the water cycle in dry inland areas.