基于物理相似法的山丘区流域水文分类及模型参数移用方法

Hydrological Classification and Model Parameter Transfer of Basins in Eastern Hilly Region Based on Physical Similarity Method

以我国东部64个山丘区流域为研究对象,提取35个流域地理特征(降雨、地形地貌、土壤和地质)作为相似指标衡量流域间的水文相似性,结合K-means聚类和参数区域化技术识别相似流域群和移用模型参数。结果显示,该64个山丘区流域依据降水、地形等特征可划分成11类相似的流域群。在具有水文相似的流域群内进行水文模型参数移用试验,水文过程模拟的确定性系数中值集中在乙级、丙级预报精度,且洪量相对误差不超过实测值的20%,能有效重现无资料流域的径流过程。该方法对无资料山丘区的水文预报具有重要借鉴意义,有助于山丘区洪水灾害防治及水资源管理。

64 basins in the eastern hilly region were selected as the research objects,and 35 geographical features of the basin(rainfall,topography,and soil and geology)were extracted as the similarity indexes to measure the hydrological similarity among basins.The K-means clustering and parameter regionalization were combined to identify similar basin groups and transfer model parameters.The results show that the 64 basins can be divided into 11 similar groups according to the geographical features(e.g.precipitation and topography).In the basin groups with hydrological similarity,the hydrological model parameter transplantation experiment is carried out,and the median certain coefficients of hydrological process simulation are concentrated in Serie B and Serie C forecast accuracy,and the relative error of flood volume is not more than 20%of the measured value,which can effectively reproduce the runoff process in the ungauged basins.This method can be used as a reference for hydrological prediction of ungauged basins in hilly regions,and is helpful for mountain flood prevention and control and water resources management.