中国农业科学2023，Vol.56Issue(22)：4403-4416,14.DOI:10.3864/j.issn.0578-1752.2023.22.004
基于综合指标协同优化的冬小麦植株水分含量预测
Prediction of Water Content of Winter Wheat Plant Based on Comprehensive Index Synergetic Optimization
摘要
Abstract
[Objective]To find a more comprehensive and accurate method to monitor the water deficit and to provide a theoretical basis for drought relief of winter wheat,the present study was conducted to construct an inversion model of plant water content(PWC)at different growth stages based on three comprehensive indexes,namely,canopy temperature,morphology and physiology indexes of winter wheat.[Method]The winter wheat was studied by setting up three water treatments(water deficit treatment W1:35 mm,water deficit treatment W2:48 mm,and control treatment W3:68 mm)and two wheat varieties(general drought resistant variety Luomai 22 and weak drought resistant variety Zhoumai 27).Canopy temperature parameters(canopy temperature standard deviation(CTSD)and crop water stress index(CWSI)),morphological indicators(plant height,stem diameter,aboveground biomass,and leaf aera index(LAI))and physiological indicators(stomatal conductance,transpiration rate,and photosynthetic rate)of winter wheat were obtained at jointing,booting,and filling stages,respectively.Comprehensive temperature parameter indicators(CTPI),comprehensive growth indicators(CGI)and comprehensive physiological indicators(CPI)based on the average weight principle were constructed.The correlation between PWC and comprehensive indicators was analyzed,and multiple linear regression(MLR),partial least squares recurrence(PLSR)and support vector machine(SVM)methods were used to construct the PWC inversion model based on comprehensive indicators according to the growth period.[Result]The canopy temperature parameters,morphology and physiological indexes of winter wheat at different growth stages showed significant differences between water deficit treatments(W1,W2)and control treatment(W3)(P＜0.05).Comprehensive indicators(CTPI,CGI and CPI)at booting and filling stages have a significant correlation with PWC,with correlation coefficients(r)of-0.70(-0.78),0.84(0.80)and 0.83(0.76),respectively.Using MLR,PLSR and SVM methods,the PWC inversion prediction model based on comprehensive indicators(CTPI,CGI and CPI)has high prediction accuracy,among which the PWC model built by SVM is the best,R2cal(R2val),RMSEcal(RMSEval),and nRMSEcal(nRMSEval)were 0.878(0.815),2.06%(2.37%),and 3.10%(3.33%),respectively.[Conclusion]The SVM-PWC model based on the comprehensive indicators CTPI,CGI and CPI can well predict the water deficit of winter wheat at different growth stages,and provide theoretical basis for drought prevention and drought resistance of winter wheat in the Huang-Huai-Hai Plain.关键词
冬小麦/水分亏缺/综合指标/植株含水量/支持向量机Key words
winter wheat/water deficit/comprehensive index/plant water content(PWC)/support vector machine(SVM)引用本文复制引用
高晨凯,温鹏飞,刘水苗,李煜铭,吴鹏年,王艳丽,刘长硕,乔毅博,关小康,王同朝..基于综合指标协同优化的冬小麦植株水分含量预测[J].中国农业科学,2023,56(22):4403-4416,14.基金项目
国家重点研发计划(2021YFD1700900)、河南省高等学校重点科研项目计划(23A210017)、河南省重点研发与推广专项(232102110298) （2021YFD1700900）
