Climate-responsive sowing strategies reduce reproductive heat losses in Eastern Indo-Gangetic wheat

CONTEXT Climate change is likely to exacerbate persistent yield gaps and food insecurity in the Eastern Indo-Gangetic Plains (EGP). Wheat, one of the EGP's staple crops, is particularly sensitive to rising temperatures and intensifying heat stress. Adjusting sowing timing offers a practical strategy to reduce heat exposure for resource-constrained smallholders. However, substantial variability in climate and cultivation practices precludes uniform sowing recommendations. To capture spatial thermal heterogeneity, we evaluated wheat responses to sowing timing across five agroclimatic zones (AZs) defined by growing degree days (GDD). OBJECTIVE The objectives of this study are to 1) characterize sowing-date responses under irrigated and rainfed conditions across different agroclimatic environments in the EGP wheatbelt; and 2) identify the effects of changes in sowing dates on wheat yield and related climate risks during different growth phases. METHODS The Decision Support System for Agrotechnology Transfer model (DSSAT-CERES-Wheat) was calibrated and evaluated using independent field datasets from different growing seasons across the EGP. The calibrated and validated model was then used to quantify the effects of altered sowing dates on irrigated and rainfed wheat yields across five AZs from 1984 to 2020. Simulations were conducted using 12 sowing scenarios at 5-day intervals from November 1 to December 26 to evaluate sowing-date responses for each AZ. We further examined how climate variability influences sowing-date responses across AZs by classifying growing-season precipitation into three categories: wet, normal, and dry. RESULTS AND CONCLUSIONS Our results show that rainfed yield declined progressively with late sowing, whereas irrigated systems exhibited clear climate-responsive optimal sowing windows. Within the simulated range, earlier sowing therefore generally favored rainfed wheat, and sowing beyond the irrigated optimal window exposed the reproductive stage to higher late-season temperatures. In the hottest zone, postponing sowing to late December reduced yields by up to 15%. Irrigation reduced yield losses associated with heat exposure, halving average yield losses relative to rainfed systems. For irrigated wheat, climate-responsive optimal sowing windows differed systematically across AZs: cooler and wetter conditions supported earlier and broader windows, whereas hotter and drier conditions were associated with later and narrower timing, reflecting greater constraints from reproductive heat stress. SIGNIFICANCE Our model-based results indicate that climate-responsive sowing strategies can reduce climate exposure and reproductive heat-related yield losses across the EGP, particularly in hotter and water-limited areas. However, translating these model-based responses into locally feasible sowing recommendations requires consideration of agroclimatic variation and production constraints.