Predicting agroecosystem responses to identify appropriate water-saving management in arid irrigated regions with shallow groundwater: Realization on a regional scale

Scenario analysis is the basis of developing rational water management practices (WMPs) for watersheds. How to predict future hydrological responses on a regional-scale is still a challenge for modeling work in irrigated watersheds with shallow groundwater environments. Therefore, this paper presents an efficient realization of predicting regional agroecosystem responses and searching for appropriate WMPs, through using a water balance-based, semi-distributed hydrological model (SWAT-AG). The scenario case study is carried out in the Jiyuan Irrigation System located in the Hetao of upper Yellow River basin, based on the calibrated and validated modeling work in our previous companion paper. Eight scenarios of water-saving practices (WSPs) are proposed, with consideration for reducing irrigation depth and controlling initial groundwater depth. Then the coupled responses of agroecosystem processes to various WSPs are predicted for the case study region in 2012 and 2013, mainly related to the groundwater depth, root zone soil water and salinity, and crop yield/natural vegetation biomass. Based on the analysis for proposed scenarios, the 100% of present irrigation depth combined with increasing initial GWD by 50 cm are recommended as appropriate WSPs for dry years, and the 80% of present irrigation depth combined with increasing initial GWD by 100 cm are recommended for wet years, in order to maintain good environmental conditions for both crops and natural vegetation. In addition, results show that SWAT-AG could overcome the scale/function limitations of traditional soil/crop models and also avoid computational issues of numerical models. We further point out that the scenarios in reality will be more complicated and comprehensive in time and space, and thus the predictions should be updated accordingly. Overall, this case study fully presents the feasibility and practicality of using the SWAT-AG model to realize the scenario response analysis and water management decision-making on a region scale for irrigated watersheds with shallow groundwater environments.