Prediction of Distribution of Dry Matter and Leaf Area of Faba Bean (Vicia faba) Using Nonlinear Regression Models

Abstract

Growth analysis is a valuable method for quantitatively investigating the growth and development of products. To analyze plant growth during the growing season, access to accurate and regular plant information is needed, which is obtained by measuring leaf surface and dry matter accumulation. The use of nonlinear regression models is expanding due to having parameters with physiological meaning in growth analysis. Of these models, there are beta, logistic, Gomperts, Richards, linear, cut and symmetric linear models. Therefore, this study was conducted on bean plant of the variety “Barakt” under factorial experiment in the form of basic randomized complete block design with four crop densities in four replications under rainfed conditions at the research farm of Gorgan University of Agricultural Sciences and Natural Resources in 2014–2015, located in the west of Gorgan, with a latitude of 37° and 45 min north and a longitude of 54° and 30 min east and an altitude of 120 m above sea level. In this study, the nonlinear beta and logistic regression models were fitted to leaf surface data, and beta, Gompertz and logistic models were fitted to bean dry weight. The AICc criterion analysis showed that the beta model had a better fit than the logistic model for leaf area. According to this model under various crop densities, LAI_max was between 2.30 and 5.30 g per square meter, t_m was from 131.90 to 144.20 days after planting, and t_e was between 158.7 and 163.50 days. Also, the analysis of the AICc criterion for dry matter accumulation showed that the beta model was better in fitting the dry matter accumulation than Gomperts and logistic models. According to this model, W_max varied between 1.725 and 1484.3 g per square meter, t_m between 138.30 and 146.40 days after planting, and t_e between 162.60 and 179.0 days in different densities.