基于SWAT模型的干旱绿洲灌区径流模拟研究

Research on Runoff Simulation of Arid Oasis Irrigation Area Based on SWAT Model

为研究干旱绿洲灌区的径流变化规律,采用1955—2017年的中国大气同化数据集(CMADS)、传统水文气象站数据、土地利用数据、土壤数据、数字高程数据,构建了基于SWAT的焉耆盆地分布式水文模型。并应用SUFI-2(Sequential Uncertainty Fitting)算法分析参数的敏感性,使用SWAT-CUP(SWAT Calibration Uncertainty Procedures)软件对影响径流模拟结果的几个重要参数进率定,利用1991—2017年的实测径流量进行模型验证。结果表明:新疆焉耆盆地长时间尺度下径流的模拟值与实测值吻合良好,模型率定期和验证期的纳什系数NES均在0.82以上,皮尔逊相关系数R^(2)分别为0.86和0.90,相对误差R_(e)小于10,由此表明该SWAT模型在焉耆盆地的适应性良好。基于此,模拟了RCP4.5和RCP8.5情景下焉耆盆地从现状年到2100年的径流变化,结果表明开都河流域的降水量变化对焉耆盆地的径流量影响显著,即在降水量增加20%,温度升高2℃时,径流量较现状年增加了26.53%。

In order to study the runoff changes in the arid oasis irrigation area,this paper constructs a distributed hydrological model of Yanqi Basin based on SWAT model with the Chinese meteorological assimilation driving datasets(CMADS)from 1955 to 2017,traditional hydrometeorological station data,land use data,soil data,digital elevation data,analyzes the sensitivity of parameters by SUFI-2(Sequential Uncertainty Fitting)algorithm,calibrates several important parameters influencing runoff simulation results through SWAT-CUP(SWAT Calibration Uncertainty Procedures)software,and verifies the model by the measured runoff from 1991 to 2017.The results show that:the simulated runoff of Yanqi Basin in Xinjiang for a long time scales coincides with the measured runoff.During the calibration period and verification period,the Nash coefficient(NES)is above 0.82,the Pearson correlation coefficient(R^(2))is 0.86 and 0.90 respectively,and the relative error(R_(e))is less than 10.Therefore,SWAT model is well applicable to the runoff simulation of Yanqi Basin.Based on this,runoff changes from 2016 to 2100 in Yanqi Basin under the scenarios of RCP4.5 and RCP8.5 are simulated.It can be found that precipitation changes in Kaidu River basin have a significant impact on runoff in Yanqi Basin,that is,when precipitation increases by 20%and temperature increases by 2℃,the runoff increases by 26.53%compared with the current year.