基于SWAT模型的泾河流域月径流分布式模拟  

Research on the monthly runoff distributed simulation and its application in Jinghe River Basin based on SWAT model

作　　者：王尹萍 姜仁贵[1,2,3] 解建仓 赵勇[3] 吕喜玺[2] WANG Yinping;JIANG Rengui;XIE Jiancang;ZHAO Yong;Lü Xixi(State Key Laboratory of Eco-hydraulics in Northwest Arid Region of China,Xi’an University of Technology,Xi’an 710048,China;Department of Geography,National University of Singapore,117570,Singapore;State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin,China Institute of Water Resources and Hydropower Research,Beijing 100038,China)

机构地区：[1]西安理工大学省部共建西北旱区生态水利国家重点实验室,陕西西安710048 [2]新加坡国立大学地理系,新加坡117570 [3]中国水利水电科学研究院流域水循环模拟与调控国家重点实验室,北京100038

出　　处：《西安理工大学学报》2020年第2期135-144,158,共11页Journal of Xi'an University of Technology

基　　金：国家重点研发计划资助项目(2016YFC0401409);国家自然科学基金资助项目(51509201,51679188,51979221);陕西省自然科学基础研究计划资助项目(2018JM5031)。

摘　　要：受气候变化和人类活动变化的影响,流域水循环和基于物理机制的分布式水文模型已成为研究热点。本文采用泰森多边形计算泾河流域2008—2015年每月面降水量,对中国大气同化数据集(CMADS)和传统气象站降水量变化趋势及其相关关系进行分析。构建基于SWAT的泾河流域分布式水文模型,并采用CMADS驱动SWAT模型,通过对流域内张家山(二)水文站2008—2015年月径流的率定和验证,评价CMADS+SWAT模式在泾河流域的适用性,并采用SUFI-2算法对模型参数进行敏感性分析和模型不确定性分析。结果表明:泾河流域CMADS数据集和气象站监测数据相关性较高,可采用CMADS数据集驱动SWAT模型。所构建的泾河流域径流模拟模型中12个参数敏感性较高,包括平均坡度(HRU_SLP)、土壤表层到底层的深度(SOL_Z())、主河道河床有效水力传导度(CH_K2)等。采用确定性系数(R 2)、纳什效率系数(NSE)、百分比偏差(PBIAS)和均方根差与标准偏差的比值(RSR)四个指标对模型结果进行评价,率定期和验证期的R^2和NSE均达到0.8以上,RSR均小于0.7,且PBIAS均介于±25%,表明CMADS+SWAT模式在泾河流域的模拟结果较好。对于整个模拟期,验证期的模拟效果优于率定期。模型率定期和验证期的p-因子均大于0.7,r-因子均接近于1,不确定性较小。研究结果可为泾河流域水资源开发利用与管理提供参考,同时CMADS+SWAT模式能够为我国气象观测资料匮乏地区的水资源管理、土地利用和非点源污染研究等提供基础支撑。Affected by the changing environment of climate change and human activities,the watershed water cycle and distributed hydrological model based on physical mechanism have become research hotspots.The paper used Thiessen polyssen to calculate the monthly precipitation of Jinghe River from 2008 to 2015,analyzed the change trends and correlations of precipitation between the China Meteorological Assimilation Datasets(CMADS)and traditional meteorological stations,built the distributed hydrological model for Jinghe River Basin runoff simulation based on the Soil and Water Assessment Tool(SWAT),used CMADS to drive the SWAT model,and evaluated the applicability of CMADS and SWAT model by the calibration and verification of monthly runoff in Zhangjiashan hydrological station from 2008 to 2015.Further,the Sequential Uncertainty Fitting(SUFI-2)was employed to analyze the sensitivity of model parameters and its uncertainty.The results show that the CMADS data set in the Jinghe River has a high correlation with the monitoring data from meteorological stations,and therefore the CMADS data set can be used to drive the SWAT model.The 12 parameters for the runoff simulation model of Jinghe River Basin are highly sensitive,including average slope(HRU_SLP),soil depth from surface to bottom(SOL_Z()),effective hydraulic conductivity of main channel bed(CH_K2),etc.R 2,Nash-Sutchiffe efficiency(NSE),Percent bias(PBIAS)and RMSE-observations standard deviation ratio(RSR)were used to evaluate the simulation result by the SWAT model.As for the calibration and validation period,R^2 and NSE were both over 0.8,RSR were less than 0.7,and PBIAS were between-25% and 25%,indicating that the simulation result of CMADS and SWAT model in the Jinghe River Basin is good.For the whole simulation period,the simulation result of the verification period was better than that of the calibration period.Moreover,the p-factors of the calibration period and the validation period were both greater than 0.7,and the r-factors close to 1,indicating that the model u

关 键 词：SWAT模型 径流模拟 CMADS SUFI-2 泾河流域 

分 类 号：TV12[水利工程—水文学及水资源]