Predicting maize yield under drought stress around flowering based on plant growth rate and canopy silking dynamics

Drought stress around flowering alters the temporal synchronization between male (anthesis) and female (silking) floral maturity, as well as dry matter accumulation, thereby posing a significant threat to maize production. However, accurately predicting maize yield components, particularly the kernel number under flowering drought stress remains challenging due to the intricate effects of drought on kernel setting processes. In this study, a four-year field experiment (2020-2023) was carried out in Jinzhou, Liaoning Province, Northeast China to investigate the impacts of drought stress around flowering on maize yield. Canopy silking dynamic, kernel number and kernel weight prediction model were developed to predict silking processes, kernel number and weight, and final yield outcomes. The results demonstrated that the canopy silking dynamic model effectively captured the silking progresses under various drought stress treatments ( R 2 =0.88-0.96, NRMSE=8%-27%). Based on estimated daily silking percentage, anthesis-silking interval, as well as kernel setting rate, kernel number per unit area under drought stress was predicted. By explicitly accounting for impaired silk emergence and development, this method achieved satisfactory kernel number prediction accuracy, with 89% for both the return test and prediction test. This study confirms that by incorporating crop growth conditions (represented as the population-level plant growth rate) around the flowering stage, maize yield under drought stress can be accurately predicted (NRMSE=14%), thereby providing a novel approach for precise yield prediction under drought stress.