摘要
采用2007~2010年宝象河流域邻近的昆明、呈贡、太华山3个国家气象站逐日降水量及宝象河干海子水文站同期逐日径流量数据,构建支持向量机(SVM)模型,分析流域内降雨与径流的关系,进而对验证期2011~2013年的年径流量进行模拟对比,并结合纳什效率系数(NSE)和相对误差(RE)对全年、雨季、旱季3个时段的模型逐日拟合结果进行评估,评价模型的模拟精度。结果显示:(1)率定相关参数后建立的SVM模型能较好模拟宝象河径流的变化过程,SVM模型的NSE和RE分别达到了0.0309和24.50%。(2)在汛期,雨量较大的年份,SVM模型模拟精度相对更高,雨季时SVM模型的纳什效率系数达到0.2031,相对误差为9.38%;验证期内,年径流量相对最大的2012年NSE相对最大(0.055),RE相对最小(24.74%)。
Using the daily precipitation of the three national meteorological stations Kunming Station,Chenggong Station,and Taihuashan Station adjacent to the Baoxiang River Basin from 2007 to 2010,and the daily runoff data of the Ganhaizi Hydrological Station of the Baoxiang River during the same period,a Support Vector Machine(SVM)model was constructed and analyzed the relationship between rainfall and runoff in the basin.Then,the annual flow data during the verification period from 2011 to 2013 was simulated and compared,and combined with the Nash efficiency coefficient(NSE)and relative error(RE),the daily fitting results of the model in the three periods of year,rainy season and dry season were assessed,and the simulation accuracy of the model was evaluated.Results were as follows:(1)The SVM model established after calibrating the relevant parameters could better simulate the change process of the Baoxiang River runoff.The NSE and RE of the SVM model reached 0.0309 and 24.50%,respectively.(2)In the flood season,the year with large rainfall,the SVM model simulation accuracy was relatively higher.During the rainy season,the Nash efficiency coefficient of the SVM model reached 0.2031,and the relative error was 9.38%.During the verification period,the annual runoff was the largest in 2012,the NSE was the largest(0.055),and the RE was the smallest(24.74%).
作者
王瑞芳
姜玥玮
易琦
WANG Rui-fang;JIANG Yue-wei;YI Qi(School of Earth Sciences,Yunnan University,Kunming 650500,Yunnan,China)
出处
《云南地理环境研究》
2020年第5期1-8,共8页
Yunnan Geographic Environment Research
基金
国家自然科学基金项目“滇池流域不同下垫面降雨产流特征及产水量模拟研究”(41761109)
国家级大学生创新训练项目“变化环境下南充市雾、霾污染态势诊断解析及其归因”(201810638055).
关键词
支持向量机
降雨—径流过程
宝象河流域
Support Vector Machine model(SVM)
rainfall-runoff process
Baoxiang River Basin
