基于Logistic模型的南疆无膜滴灌棉花产量模型构建与评价

Construction and Evaluation of Cotton Yield Model Based on Logistic Model for Filmless Drip Irrigation in Southern Xinjiang

为了阐明灌溉对无膜滴灌棉花生长动态和产量形成的影响,在2019-2020年大田试验基础上,采用Logistic模型定量分析不同灌溉处理对无膜滴灌棉花茎粗、株高和叶面积指数动态发育的影响,建立棉花生长模型,分析棉花生长动态变化过程与产量及其构成因素的回归关系,构建棉花产量模型。结果表明,无膜滴灌棉花茎粗、株高及叶面积指数实测值与模拟值标准化的均方根误差(nRMSE)均小于20%,协同指数(d)与决定系数(R2)均大于0.91,Logistic模型可用于精确描述不同灌溉处理下无膜滴灌棉花茎粗、株高及叶面积指数的动态过程;无膜滴灌棉花产量与最大茎粗、最大株高、最大叶面积指数和耗水的相关性均呈现极显著水平;棉花产量模型预测值和实测值的nRMSE、d与R2分别为5.11%、0.88和0.62,产量模型具有较高的预测精度。

In order to elucidate the effects of irrigation on growth and yield formation of cotton with filmless drip irrigation,with a two-year field experiments from 2019 to 2020,the Logistic model was adopted to quantify the impactions of different irrigation treatments on dynamics of stem diameter,plant height,and leaf area index of drip-irrigated cotton without film mulching.The growth models were constructed for analyzing the regression relationship between the dynamics of cotton growth and yield and its components,and the cotton yield model also was constructed.The results indicated that the normalized root-mean-square errors(nRMSE)of the measured and simulated values of cotton stem diameter,plant height and leaf area index were all less than 20%,and synergy index(d)and coefficient of determination(R 2)were higher than 0.91,suggesting that the Logistic model could accurately describe the growth dynamics of cotton stem diameter,plant height,and leaf area index under different drip irrigation treatments without film mulching.There were significant correlations between yield and maximum value of stem diameter,plant height,leaf area index and water consumption of drip-irrigated cotton without film mulching.The nRMSE,d and R 2 of the predicted and measured values of the cotton yield model were 5.11%,0.88 and 0.62,respectively,and the yield model had high prediction accuracy.