Improvement and application of a distributed hydrological model in karst regions

In the karst peak forest area of southwest China, it is difficult to quantitatively characterize the process of groundwater recharge with the traditional hydrogeological models due to the spatial heterogeneity caused by the strong carbonate karst development and the influence of surface karst zone on rainfall infiltration. Based on the open-source SWAT (Soil and Water Assessment Tool) model and by improving the soil water and groundwater modules, the process of soil water entering groundwater through rapid flow and slow flow respectively by piston propulsion and by shortcut, is characterized in this paper. Taking the Lijiang River Basin in Guilin of Guangxi as an example, the authors establish a hydrological model in the basin with the cross distribution of carbonate and clastic rocks, and the model is calibrated and validated. The results show that compared with the original model, the distributed model has a better efficiency on simulation of regulation and storage function of the surface karst zone and dry season simulation in each hydrological station. The Nash coefficient of the Chaotian station increases from 0.78 to 0.92, and the determination coefficient increases from 0.85 to 0.92, obviously higher than that before the improvement. The improved model can be applied to the simulation of water cycle in karst basins with uneven distribution of clastic rocks and carbonate rocks.