遥感订正作物种植结构数据对提高灌区SWAT模型精度的影响  

Effects of correcting crop planting structure data to improve simulation accuracy of SWAT model in irrigation district based on remote sensing

作　　者：王维刚 史海滨[1,2] 李仙岳 郑倩[1,2] 张文聪 孙亚楠[1,2] Wang Weigang;Shi Haibin;Li Xianyue;Zheng Qian;Zhang Wencong;Sun Yanan(College of Water Conservancy and Civil Engineering,Inner Mongolia Agricultural University,Hohhot 010018,China;High Efficiency Water-saving Technology and Equipment and Soil and Water Environment Effect in Engineering Research Center of Inner Mongolia Autonomous Region,Hohhot 010018,China)

机构地区：[1]内蒙古农业大学水利与土木建筑工程学院,呼和浩特010018 [2]高效节水技术装备与水土环境效应内蒙古自治区工程研究中心,呼和浩特010018

出　　处：《农业工程学报》2020年第17期158-166,共9页Transactions of the Chinese Society of Agricultural Engineering

基　　金：国家自然科学基金项目(51539005、51769024);内蒙古水利科技重大专项(213-03-99-303002-NSK2017-M1)。

摘　　要：为确保灌区水文过程与营养物流失过程模拟更接近于真实过程,进一步提高模拟精度,该研究综合考虑作物种植结构空间位置的准确性与作物种植结构数据的精度2个因素,利用GF-116 m遥感影像对耕地作物进行分类提取,并对土地利用类型图进行修正,从而分析比较作物种植结构空间位置的订正与作物种植结构数据精度的提高分别对SWAT(Soil and Water Assessment Tool)模型模拟精度的影响。结果表明:作物种植结构空间位置的订正或作物种植结构数据精度的提高均可提高径流和硝态氮模拟效率。经作物种植结构空间位置的订正和数据精度的提高可使得模型在径流模拟中,率定期和验证期决定系数R2分别达到了0.76和0.82,效率系数分别达到了0.69和0.79,相对误差分别降低至3.50%和-0.30%;在硝态氮模拟中,率定期和验证期决定系数R2分别达到了0.70和0.63,效率系数分别达到了0.55和0.53,相对误差分别降低至10.06%和6.42%。综合订正作物种植结构空间位置和提高作物种植结构数据精度可有效提高SWAT模型在灌区的模拟精度。This study synthetically considered the spatial position accuracy and data precision for crop planting structure in order to ensure that the hydrological and nutrient loss processes were more veritably simulated and the simulation accuracy was further improved.The classification and extraction of field crops were conducted and the land use map was modified using Normalized Difference Vegetation Index(NDVI)threshold method and Support Vector Machine(SVM)method based on GF-116 m WFV4 medium resolution remote sensing images in Hetao Irrigation District.The effect of the corrected spatial position and the improved accuracy of crop planting structure on the simulation accuracy of SWAT(Soil and Water Assessment Tool)model were evaluated using the modified land use map.The results showed that the classification of crops based on GF-116 m WFV4 remote sensing images agreed with the actual spatial distribution of crops in Hetao Irrigation District,with an overall accuracy of 89.61%,a mapping accuracy of over 88%,a user accuracy of over 88%and a Kappa coefficient of 0.86.The parameters with high level of sensitivity to the simulation of runoff and nitrate nitrogen were quite stable in the irrigation district.The simulation accuracy in terms of runoff was significantly affected by groundwater delay coefficient(GW_DELAY),groundwater evaporation coefficient(GW_REVAP),base flow alpha factor(ALPHA_BF),and soil evaporation compensation factor(ESCO).In addition,the simulation accuracy of nitrate nitrogen was markedly affected by nitrogen concentration in rainfall(RCN),the nitrate percolation coefficient(NPERCO),and the denitrification exponential rate coefficient(CDN).The corrected spatial position accuracy and data precision of crop planting structure effectively improved the accuracy of simulated values for runoff and nitrate nitrogen.In the calibration period(2009-2014),the R2 for simulated runoff and nitrate nitrogen were improved to 0.76 and 0.70 from 0.63 and 0.62,respectively by correcting crop pattern locations.The efficie

关 键 词：遥感 作物 SWAT模型 GF-1 种植结构 河套灌区 模拟精度 

分 类 号：X53[环境科学与工程—环境工程] S127[农业科学—农业基础科学]