Flood disasters pose one of the greatest threats to humanity. Effectively addressing this challenge requires improving the accuracy of flood simulation. Taking Xunhe watershed in Shandong Province as the study area, t...
详细信息
Flood disasters pose one of the greatest threats to humanity. Effectively addressing this challenge requires improving the accuracy of flood simulation. Taking Xunhe watershed in Shandong Province as the study area, the Random Forest model was utilized to classify historical flood events within the watershed based on rainfall conditions, such as varying rainfall durations, intensities, and total precipitations. Multiple sets of hydrological model parameters were established to conduct flood classification simulation, reducing the error caused by using a single parameter set for the entire watershed. The results indicate that the Random Forest model can be applied to flood classification simulation in Xunhe watershed. Compared to unclassified simulations, the method proposed in this study leads to an improvement in the Nash coefficient by 0.06 to 0.14, a reduction in the relative error of peak discharge by 3% to 11.24% and a reduction in the relative error of flood volume by 1.46% to 9.44%. The flood classification simulation method proposed in this study has certain applicability in reducing flood simulation errors under different rainfall scenarios and improving accuracy in the watershed, providing new insights for flood control and disaster reduction efforts.
暂无评论