不同网格加密方法在河流回补地下水模拟中的对比.

groundwater level near the river is a key to accurately assessing the effect of recharge. Based on unstructured grid refinement method，four different grid generation methods (two structural grids and two locally refined unstructured grids) were set，and different boundary conditions and subdivision of two unstructured grids were used to numerically simulate the process of groundwater recharge by river. Through a comparison of the final flow field, water levelvariations before and after recharge under different distances with river, groundwater balance and operation eficiency of model, the advantages and disadvantages of different grid refinementmethods in the linear recharge simulation were demonstrated. The results show that two locally refined unstructured grid models and fine structured grid model are more accurate than coarse structured grid model, resulting in more accurately description of the changes in groundwater levels caused by river recharge； two locally refined unstructured grid models have a shorter running tme than fine structured grid model, and run more efficiently without significantly reducing the accuracy of the simulation; Q-tree grid model has some advantages over nested grid model in long-term simulation, and the simulation accuracy is slightly higher than nested grid model in the later period of river recharge； moreover, Q-tree grid model has the advantage of multilevel refinement with fewer grids and shorter running tme than nested grid model. In conclusion, the Q-tree grid refinement method not only guarantees emulation accuracy, but also reduces running tme to a large extent, which indicates that the Q-tree grid refinement method should be recommended ｏｒ ｓｉｍｕｌａｔｉｏｎｓｏｆｇｒｏｕｎｄｗａｔｅｒｒｅｃｈａｒｇｅｂｙｒｉｖｅｒ． [ABSTRACT FROM AUTHOR]