基于水箱模型的会仙岩溶湿地睦洞河小流域径流模拟研究.

Influenced by the distribution of karst landforms and karst development, the process of rainfall runoff in karst regions is complicated. Runoff simulation can help identify the main factors influencing runoff in karst areas. In this study, based on the karst landform characteristics of the small watershed of the Mudong River in Huixian Karst Wetland in the Lijiang River basin, parallel staggered time model of double tanks was established and an artificial neural network was built by using Microsoft Excel for parameter optimization. The relative error, the Nash-Sutcliffe efficiency coefficient and the correlation coefficient were used to evaluate the simulation effects during calibration and validation based on the measured discharges in the small watersheds. The mean annual relative error of the regularization rate was 7.5%, the Nash-Sutcliffe coefficient was 0.63 and the correlation coefficient was 0.72. During the validation period, the mean annual relative error was 10.5%, the Nash-Sutcliffe coefficient was 0.56 and the correlation coefficient was 0.63, indicating that the model was able to achieve good simulation results in the study area. Through the parameter sensitivity analysis of the runoff simulation of the Tank model in the small watershed of Mudong River, the results show that the karst geomorphological characteristics such as surface infiltration performance, the existence of karst fast flow and slow flow, and karst base flow mainly affect the runoff distribution and production flow in the karst small watershed. [ABSTRACT FROM AUTHOR]