Modelagem da evapotranspiração e predições do rendimento de soja sob condição irrigada e de sequeiro

Abstract

Understanding the impacts of reduced soil water availability on soybean yields is paramount in rainfed and irrigation agriculture, especially from the perspective of climate change. Combining field experiments with water balance models can better explain the effect of different agroecological conditions on yield and support management decisions that help mitigate the impact of water deficit by adjusting cropping factors, especially irrigation management. Improving irrigation management implies improving estimates of crop evapotranspiration (ETc), which also involves a more accurate prediction of crop coefficients (Kc) throughout the cycle. An approach recently developed by Allen and Pereira (2009) (A&P) estimates single (Kc) and basal (Kcb) crop coefficients from observations of the fraction of ground covered (fc) and plant height (h) since these parameters represent the physical basis of the vegetation, such as quantity, type, stomatal resistance, and conductance. In this study, this approach was assessed, along with the SIMdualKc soil water balance model that was calibrated and validated using field observations of soil water content to determine soybean water use and the actual crop evapotranspiration from four relative maturity groups (MG’s), under irrigated and rainfed conditions. Field experiments were carried out during the 2018/19 and 2019/20 growing seasons in Rio Grande do Sul, Brazil. The observed data allowed estimating the actual Kc and Kcb for each segment of the FAO Kc-curve using the A&P approach and estimating and partitioning the actual ETc in soil evaporation (Es) and crop transpiration (Tc). The SIMDualKc model was able to simulate the dynamics of the available soil water (ASW), with a coefficient of determination (R2 ) varying from 0.85 to 0.98 and with a regression coefficient (bo) ranging from 0.97 to 1.03, along the crop cycle of the four MG’s, for both irrigation management and growing seasons. The estimation errors were low, with the root mean square error (RMSE) ranging from 3.5 to 4.3% of the total available water (TAW). The actual Kc and Kcb estimated with both approaches were compared, showing the perfect accuracy of the A&P approach to improve soybean irrigation scheduling. The rainfed treatments showed an average reduction in productivity of 8% in 2018/19 and 14.8% in 2019/20. Comparing observed and estimated soybean grain yield by the linear regression model resulted in a b0 value between 1.0 and 1.02, indicating that the estimated yield corresponded well with the observed one, showing a good association. The highest Ky values were obtained for soybean cultivars with MG 5 and 5.5, demonstrating that special attention should be given to water management in cultivars with lower MG.