关键词: 渠井结合灌区, 随机-确定模型, 地下水, 动态监测, 优化配置, 陕西

Abstract: Situated in Shaanxi Province of northwest China, the Jinghui Canal Irrigated Area is a typical area irrigated with both pumped groundwater and canal-diverted surface water. The Jinghui Canal Irrigated Area belongs to a continental semiarid climate region with high temperature and relatively plentiful rainfall in summer, and it is cold and arid with high evaporative capacity in winter. In fact, groundwater has been the main water source for most water demands in this area, including agriculture, industry and livelihood. However, water consumption has been dramatically increased with rapid industrial and agricultural development, population growth and sharp reduction of surface runoff inflow, and groundwater is pumped in big volume in most areas. The overdraft of groundwater has caused many problems. On account of irrational groundwater utilization, a continuous drawdown of groundwater level occurs and results in a series of environmental geological problems, such as the cone of depression, subsidence of land and deterioration of water quality. Therefore, it is significant to study how to develop and utilize groundwater resources rationally and how to redistribute irrigating water in an optimized way in the irrigated area. At present, the most urgent challenges for the local government are how to perform the rational redistribution of irrigating water resources and realize the sustainable development of this irrigated area so as to solve the relevant water issues caused by groundwater overexploitation. In recent decades, the researchers at home and abroad have adopted large amounts of methods to study the optimized redistribution of water resources in various irrigated areas in China. However, most of the current studies focus on the numerical solution of optimized redistribution per se, as a result, some significant facts are ignored, such as the impacts of mechanism conditions including geology, hydrogeology and environment in research areas. In order to employ the research results in practice, the Visual Modflow software is used as a platform, and a groundwater stochastic-determination model (GSDM) is developed based on time-series model and radial basis function neural network model (RBF-ANN) to make an intensive analysis of water pattern on a basis of prediction model of groundwater regime for Jinghui Canal Irrigated Area.The results provide a useful and powerful basis for the sustainable development and utilization of groundwater and optimized redistribution of water resources in irrigated areas.

Key words: canal-well irrigated area, GSDM, groundwater, dynamic monitoring, water optimized redistribution, Shaanxi Province