典型喀斯特流域介质结构的降雨-径流机制研究——以黔中地区为例  

Study on the Rainfall and Runoff Mechanism of Media Structure in Typical Karst Basin-Taking the Central Guizhou Region as An Example

作　　者：杨树平 贺中华[1,2,3] 王茂强 潘杉[1] 陈莉会 谭红梅 顾小林 许明金 Yang Shuping;He Zhonghua;Wang Maoqiang;Pan Shan;Chen Lihui;Tan Hongmei;Gu Xiaolin;Xu Mingjin(School of Geography and Environmental Science,Guizhou Normal University,Guiyang 550001,China;National Karst Rock Desertification Prevention Engineering Technology Research Center,Guizhou Normal University,Guiyang 550001,China;Key Laboratory of Remote Sensing Application for Mountain Resources and Environment,Guizhou Province,Guiyang 550001,China;Bureau of Hydrology and Water Resources,Guizhou Province,Guiyang 550000,China)

机构地区：[1]贵州师范大学地理与环境科学学院,贵阳550001 [2]贵州师范大学国家喀斯特石漠化防治工程技术研究中心,贵阳550001 [3]贵州省山地资源与环境遥感应用重点实验室,贵阳550001 [4]贵州省水文水资源局,贵阳550000

出　　处：《水土保持研究》2024年第3期49-60,共12页Research of Soil and Water Conservation

基　　金：贵州省水利厅自然科学基金项目(KT202237);贵州省自然科学基金项目(黔科合基础-ZK[2023]重点028);国家自然科学基金项目(u1612441,41471032)。

摘　　要：[目的]探究喀斯特流域介质结构的降雨—径流机制,为合理利用黔中地区水资源及防洪防旱等方面提供理论参考。[方法]选黔中20个典型流域为研究样区,分析流域降雨、径流时空分布特征,利用支持向量回归(SVR)开展多尺度降雨—径流模拟,采用Spearman方法分析流域介质结构的降雨—径流机制。[结果](1)黔中1979—2021年降雨呈缓慢上升趋势(5.02 mm/a),径流呈缓慢下降趋势(-0.17亿m^(3)/a),两者在空间上总体分布不均匀;(2)通过径向基核函数(RBF)训练的SVR相对准确,模拟精度:秋季>冬季>春季>全年>夏季;(3)单因素单因子影响:非喀斯特(0.764)>低活性强酸土(0.566)>峰林地貌型(0.459)>草丛(0.456)>连续性白云岩喀斯特(-0.435)>栽培植被(-0.426);(4)单因素耦合影响:土层覆盖结构(0.714)>地貌组合结构(-0.529)>岩石组成结构(-0.446)>植被分布结构(0.323),多因素耦合影响:234(-0.702)>1234(-0.679)>23(0.607)>34(-0.604)>24(-0.547)>134(-0.544)>13(0.507)。[结论]SVR适用于黔中地区降雨—径流过程的模拟,流域介质结构对降雨—径流影响较大。[Objective]The aims of this study are to investigate the rainfall-runoff mechanism in the medium structure of karst basin,and to provide theoretical references for the rational utilization of water resources and flood and drought prevention in the Central Guizhou region.[Methods]20 typical watersheds in the central Guizhou were selected as the study sample areas,the spatial and temporal distribution characteristics of rainfall and runoff in the watersheds were analyzed,multi-scale rainfall-runoff simulation by using Support Vector Regression(SVR)was carried out,and the rainfall-runoff mechanism of the watershed media structure was analyzed by using Spearman′s method.[Results](1)Rainfall in the Central Guizhou from 1979 to 2021 showed a slowly increasing trend(5.02 mm/a)and runoff showed a slowly decreasing trend(-0.017 billion m^(3)/a),and both of them unevenly distributed in the space in general.(2)SVR trained by radial basis kernel function(RBF)was relatively accurate,and simulation accuracy decreased in the order:fall>winter>spring>year-round>summer.(3)The single-factor single factor effects decreased in the order:non-karst(0.764)>low activity strong acidic soil(0.566)>peak forest geomorphic type(0.459)>grassland(0.456)>continuity dolomite karst(-0.435)>cultivated vegetation(-0.426).(4)Single-factor coupled effects decreased in the order:soil cover structure(0.714)>geomorphic combination structure(-0.529)>rock composition structure(-0.446)>vegetation distribution structure(0.323),and multi-factor coupling influence decreased in the order:234(-0.702)>1234(-0.679)>23(0.607)>34(-0.604)>24(-0.547)>134(-0.544)>13(0.507).[Conclusion]SVR is suitable for the simulation of rainfall-runoff process in the central Guizhou region,and the structure of watershed media has a large influence on rainfall-runoff.

关 键 词：流域介质结构 支持向量回归(SVR) 降雨—径流机制 

分 类 号：P429[天文地球—大气科学及气象学]