3种山核桃属植物叶面积预估模型构建  

Construction of Leaf Area Prediction Model for Three Carya Species

作　　者：陈柱伯 李乾华 吕勇[1] 魏海林[1,3] CHEN Zhu-bo;LI Qian-hua;LYU Yong;WEI Hai-in(College of Forestry,Central South University of Forestry and Technology,Changsha 410004,PRC;Daoxian County ForestryBureau,Daoxian 425399,PRC;Hunan Academy of Forestry Sciences,Changsha 410018,PRC)

机构地区：[1]中南林业科技大学林学院,湖南长沙410004 [2]永州市道县林业局,湖南道县425399 [3]湖南省林业科学院,湖南长沙410018

出　　处：《湖南农业科学》2023年第1期80-88,共9页Hunan Agricultural Sciences

基　　金：湖南省杰青培养科研项目(XLK202108-4)。

摘　　要：以山核桃、湖南山核桃和美国山核桃3种经济价值高的山核桃属植物为研究对象,通过方差分析对比3个树种叶片的形态差异,利用Pearson相关分析筛选最优自变量参数,采用线性回归法与最小二乘法分别构建3个树种及整体的小叶面积(LA)和复叶面积(CLA)与叶形态学指标间的预估模型,并通过决定系数(R^(2))、均方根误差(RMSE)、赤池信息量准则(AIC)和预测精度(FA)评估模型的优劣,最后检验预估模型的适用性。结果表明:3树种间叶片形态参数差异显著,尤其LA变异系数最大,达75.08%;Pearson相关分析显示,3树种均以小叶长宽积(P_(1))和复叶长宽积(P_(2))与相应叶面积的相关系数为最高;预估山核桃、湖南山核桃、美国山核桃和整体的LA最优模型分别为LA=0.7125 P_(1)^(0.9816)、LA=0.6346 P_(1)+1.3891、LA=0.6277 P_(1)^(1.0012)和LA=0.6466 P_(1)+0.1341;预估CLA的最优模型分别为CLA=0.2931 P_(2)+9.1523、CLA=0.3851 P_(2)-33.95、CLA=0.316 P_(2)+16.697和CLA=0.2903 P_(2)+28.53。构建的3树种LA和CLA最优模型平均FA分别为95.85%和92.67%,且模型检验的平均FA分别为96.04%和95.29%,表明构建的预估模型是可靠的,可满足供试树种叶片面积预估的要求。Carya plants are of odd pinnate compound leaves with a large number of leaflets and irregular curvature,which increases the difficulty and error of leaf area measurement.In this study,Carya cathayensis,Carya hunanensis and Carya illinoinensis with high economic values were taken as the research objects to build a fast,convenient and accurate model to estimate the leaf area of Carya.The differences of morphological characteristics among the three tree species were compared by variance analysis;Pearson correlation analysis was used to screen the optimal independent variable parameters;linear regression method and least square method were used to build prediction models of leaflet area(LA)and compound leaf area(CLA)of three tree species and the overall based on leaf morphology indexes;finally,based on the determination coefficient(R^(2)),root mean square error(RMSE),Akaike information criterion(AIC)and forecast accuracy(FA),the accuracy and applicability of the prediction models were evaluated.The results showed that the difference of leaf morphological parameters among the three tree species was significant,especially the variation coefficient of LA was the largest,reaching 75.08%;Pearson correlation analysis showed that the correlation coefficients between LA and the product of length and width of leaflets(P_(1)),and between CLA and the product of length and width of compound leaves(P_(2))were the highest among the three tree species.For predicting Carya cathayensis,Carya hunanensis,Caryaillinoinensis and the overall,the optimal LA models were LA=0.7125 P_(1)^(0.9816),LA=0.6346 P_(1)+1.3891,LA=0.6277P_(1)^(1.0012)and LA=0.6466 P_(1)+0.1341,respectively;the optimal CLA models were CLA=0.2931 P_(2)+9.1523,CLA=0.3851 P_(2)-33.95,CLA=0.316 P_(2)+16.697 and CLA=0.2903 P_(2)+28.53,respectively.The average FA of the optimal LA and CLA models of the three tree species constructed was 95.85%and 92.67%,respectively,and the average FA of the model test was 96.04%and 95.29%,respectively,indicating that the prediction model co

关 键 词：山核桃属 叶面积 预估模型 长宽积 小叶 复叶 

分 类 号：S792.99[农业科学—林木遗传育种]