Modelling the severity-yield relationship for the late-season disease complex in soybean small-plot trials

A meta-analytical approach was employed to assess the relationship between late-season disease (LSD) severity and soybean yield across 39 field trials conducted from 2020/21 to 2023/24 in nine Brazilian states. Pearson's correlation coefficient (r) and Fisher's transformation (Z) were calculated to examine the association between disease severity and yield. Additionally, random- and mixed-effects meta-analyses were performed to estimate the intercept (beta(0)) and slope (beta(1)) of the linear regression model. Disease pressure, yield class, region, altitude and year were included as categorical moderators. The damage coefficient (in percentage) was calculated from beta(0) and beta(1) values, while the economic damage threshold (EDT) was derived using the estimated damage coefficient, attainable yield and fungicide efficacy values. A strong negative correlation between LSD severity and yield was observed (Z = -1.03, r = -0.77). The population average estimates for beta(0) and beta(1) were 4589.0 kg/ha and -22.6 kg/ha per unit increase in severity, respectively (p < 0.001). While none of the categorical moderators significantly affected the slope (p > 0.19), disease pressure, yield class and altitude significantly affected the intercept (p < 0.014). The overall relative damage coefficient was estimated at -0.49% (95% CI = -0.61 to -0.38), with EDT values ranging from 0.41% to 4.33% (mean = 1.57%). In conclusion, a significant negative linear relationship between soybean yield and LSD severity was confirmed, particularly for Cercospora leaf blight and Septoria brown spot. However, these relationships should be validated further before being extrapolated to larger field scales.