Climate-adaptive optimization of rice sowing windows for yield stability: A case study of Anhui, China

Adjusting rice sowing dates in response to climate change is an effective strategy for sustainable production, as it improves the climatic resource use efficiency and reduces the weather-related risks. We developed a climate adaptive sowing optimization model that integrates rice phenology simulation, estimation of climate-constrained productivity, and multi-hazard stress correction for drought, flooding, heat, and cold damage. Using daily meteorological records from 77 stations across Anhui Province, China, during 1971–2024, we evaluated 20 sowing date scenarios for single season rice production. The model evaluation showed that incorporating phenological development and correction functions for seed setting rate (against drought, heat, and flood), substantially improved the correlation between simulated climate constrained productivity and observed yields, increasing the coefficient of determination (R 2 ) from 0.310 to 0.538. The results indicated that both early and late sowing reduced climate constrained productivity and increased inter-annual variability. Optimal sowing windows for most regions occurred from early to mid-May, while flowering under stable yield scenarios occurred around late July in Northern Anhui and mid-August onward in Central and Southern Anhui regions, thereby reducing exposure to critical heat stress periods. Climate constrained productivity under optimal stable-yield scenarios exhibited a significant declining trend, highlighting increasing climate pressure on rice production. Among climatic hazards, drought caused the maximum yield losses, followed by heat stress, floods, and cold damage. These findings provide a practical framework for optimizing rice sowing windows and strengthening climate resilience in regional rice production systems.