不同灌水处理下CROPGRO-Tomato模型全局敏感性和不确定性分析  

Sensitivity and Uncertainty Analysis for CROPGRO-Tomato Model at Different Irrigation Levels

作　　者：李波[1] 李长信 姚名泽[1] 魏新光[1] 包正荣 孙翔龙 LI Bo;LI Chang-xin;YAO Ming-ze;WEI Xin-guang;BAO Zheng-rong;SUN Xiang-long(College of Water Conservancy,Shenyang Agricultural University,Shenyang 110161,China)

机构地区：[1]沈阳农业大学水利学院,沈阳110161

出　　处：《沈阳农业大学学报》2020年第2期153-161,共9页Journal of Shenyang Agricultural University

基　　金：中国博士后科学基金项目(2019M661128);辽宁省自然基金计划项目(2019-BS-211)。

摘　　要：作物品种参数和土壤参数是作物生长过程模拟和产量预测过程中的基本输入参数,输出变量对输入参数的敏感性分析对提高作物模型的模拟精度具有重要作用。依据辽宁地区的温室番茄田间试验资料,进行了不同灌水水平下CROPGRO-Tomato模型的全局敏感性分析。在4个灌水处理(50%θf、60%θf、70%θf和80%θf为灌水下限,θf为田间持水率)条件下,利用SimLab软件,采用Morris法和EFAST法分析了CROPGRO-Tomato模型的5个输出变量(番茄的开花期、成熟期、产量、地上干物质量和最大叶面积指数)对作物品种参数和土壤参数的全局敏感性,并对两种方法的分析结果进行了比较,同时对EFAST法分析结果进行了不确定性分析。结果表明:两种方法分析得到的开花期天数、成熟期天数、地上干物质量和最大叶面积指数的分析结果相关性较高;产量方面,两种方法得到的XFRT(每日分配给果实的干物质量的最大比例)、EMFL(出苗到第一朵花出现的光热时间)、LFMAX(最适条件下叶片最大光和速率)等作物品种参数的敏感度排序基本一致。其中最敏感参数为XFRT,在EFAST法和Morris法中其值分别为0.505和1305;两种方法得到开花期、成熟期、地上干物质量和叶面积指数等输出变量的最敏感参数均为EMFL。其中EFAST法得到的上述各参数依次为0.999,0.615,0.463,0.690;Morris法得到的各参数依次为19.9,18.2,2450,5.13;不确定性分析结果表明:当土壤水分在70%θf和80%θf条件下,CROPGRO-Tomato模型对番茄生育期划分和地上干物质量的模拟精度比在50%θf和60%θf水分条件下平均高7.25%。研究结果为利用DSSAT-CROPGRO-Tomato模型模拟辽宁地区温室番茄生长过程中参数的优化和适用性的提高提供理论依据。Crop models need to simulate crop growth process or predict yield based on a large number of variety parameters and soil parameters.Sensitivity analysis is helpful to adjust parameters and verify applicability of GROPGRO tomato model in greenhouse environment.By screening sensitive parameters,the operation time of the model was shorten and the accuracy of simulation was improved.Based on the greenhouse tomato experiments in Shenyang region of Liaoning province,we analyzed the sensitivity of five outputs(days of initial flowering and maturing,yield,ground dry biomass and maximum leaf area index of tomatoes)of the CROPGRO-Tomato model to the variety and soil parameters with the methods of Morris and Extended Fourier Amplitude Sensitivity Test(EFAST)at four lower irrigation limits of 50%θf,60%θf,70%θf and 80%θf respectively(θf is field water capacity),and compared the results drawn from the 2 methods.The results show that there is a high correlation for the 2 methods in the days of flowering,maturing,ground dry biomass and maximum leaf area index,and a same sensitivity order for XFRT,EMFL and LFMAX in terms of yield.Therein XFRT is the most sensitive parameter with 1305 for Morris method and 0.505 for EFAST,and EMFL is the most sensitive parameter in the days of flowering,maturing,ground dry biomass and maximum leaf area index,including 0.999,0.615,0.463 and 0.690 for EFAST and 19.9,18.2,2450 and 5.13 for Morris.The results of uncertainty analysis indicated that the accuracy of division of tomato growth period and the simulation of ground dry biomass was 7.25%higher for the model with field water capacity being 70%θf and 80%θf than that being 50%θf and 60%θf.Above all,this study provides a theoretical basis for the simulation of parameters optimization and applicability improvement of greenhouse tomato growth process s in Liaoning with the DSSAT-CROPGRO-Tomato model.

关 键 词：温室番茄 灌水水平 CROPGRO-Tomato模型 全局敏感性分析 不确定性分析 

分 类 号：S641.2[农业科学—蔬菜学]