Geographical adaptability for optimizing the recommendation of soybean cultivars in the Brazilian Cerrado

Abstract

Yield multi-location trials associated to geostatistical techniques with environmental covariables can provide a better understanding of G x E interactions and, consequently, adaptation limits of soybean cultivars. Thus, the main objective of this study is understanding the environmental covariables effects on soybean adaptation, as well as predicting the adaptation of soybean under environmental variations and then recommend each soybean cultivar to favorable environments aiming maximize the average yield. The trials were carried out in randomized block design (RBD) with three replicates over three years, in 28 locations. Thirty-two genotypes (commercial and pre-commercial) representing different maturity groups (7.5-8.5) were evaluated in each trial were covering the Edaphoclimatic Region (REC) 401, 402 and 403. The covariables adopted as environmental descriptors were accumulated rainfall, minimum temperature, mean temperature, maximum temperature, photoperiod, relative humidity, soil clay content, soil water avaibility and altitude. After fitting means through Mixed Linear Model, the Regression-Kriging procedure was applied to spacialize the grain yield using environmental covariables as predictors. The covariables explained 32.54% of the GxE interaction, being the soil water avaibility the most important to the adaptation of soybean cultivars, contributing with 7.80%. Yield maps of each cultivar were obtained and, hence, the yield maximization map based on cultivar recommendation was elaborated.

Figure 1

Maximum percentage contribution of significant environmental effects followed by their means, represented on the left axis. The right axis refers to the representativeness of the significant effects observed for the eight soybean genotypes evaluated in Brazil's Soybean Macroregion 4.

Figure 2

Graphical representation of the productive adaptability (kg ha⁻¹) of eight soybean genotypes, evaluated in Brazil's Soybean Macroregion 4, modeled using climatic and geographical environmental variables. The adaptation of the genotypes was classified using a light (low adaptation) to dark (high adaptation) color gradient; for more details, see appendices B and C.

Figure 3

Graphical representation of the adaptation of the winning soybean genotypes in the state of Mato Grosso ("who-won-where" map). Percentage values refer to the proportion of area in which the genotype outperformed the others in the target region.

Figure 4

Geographical distribution of field trial sites (green dots) for the 2019/2020 to 2021/2022 harvests. The orange highlights indicate soybean production, in millions of tons, by municipality throughout the State of Mato Grosso.