ABSTRACT
摘要
Target spot of soybean is strongly influenced by weather, but the major drivers of epidemics remain poorly understood. Using a dataset obtained in non-treated plots of 254 trials conducted across nine Brazilian states over 13 growing seasons, we applied a distributed lag nonlinear modeling to quantify delayed and nonlinear associations between weather and severity measured at the R6 growth stage. Daily and cumulative effects were evaluated across two epidemiologically-relevant pre‑assessment periods: an early-season period (41–85 days prior to R6), associated with epidemic onset, and a late-season period (0–40 days prior to R6), associated with epidemic development. Lag–response functions were incorporated into a beta mixed model with season as random effect. Maximum air temperature, accumulated precipitation, and vapour pressure deficit (VPD) exhibited period‑dependent associations with target spot severity. Increased levels of severity (>60%) were associated with temperatures above ~30 °C across periods. Precipitation showed linear cumulative effects during the late-season period, with increased severity (>70%) typically observed exceeding ~300 mm. In contrast, VPD displayed opposing period‑specific effects, with lower severity ( ~1.3 kPa during the later period. A historical assessment of spatial and temporal variation in climatic favorability further revealed consistent shifts across planting periods, indicating lower climatic favorability for outbreaks under delayed soybean planting. The results indicate that more severe outbreaks may arise from the combined influence of elevated temperature, sustained moisture availability, and high atmospheric drying demand during the two critical phases of epidemic development.