Effects of delayed sowing on physiological traits and plant height dynamics in rice cultivar ‘Yefengzhan’: physiological response analysis and logistic-based empirical modeling of plant height

Investigating the impact of delayed sowing on rice is an important strategy for improving rice growth and yield. This study, based on sowing experiments conducted in Guangdong Province, China, from 2018 to 2024, compared delayed sowing treatments (delayed by 10 and 20 days) with a control group of conventional sowing dates. The study quantified the effects of delayed sowing on rice photosynthetic parameters and endogenous hormone content, and developed a logistic growth model for plant height based on the time extreme values from the green recovery to the milk stage. The results showed that delayed sowing significantly affected rice physiological characteristics, with the most significant decline in photosynthetic parameters observed during the heading stage, ranging from 9.15% to 22.56%. The indole-3-acetic acid (IAA) content increased by 4.98% during the green recovery stage at the second sowing date, but subsequently decreased by more than 14%. Abscisic acid (ABA) increased by less than 5% only during the jointing stage at the third sowing date, while gibberellin (GA) showed the least response to delayed sowing, with fluctuations remaining within 3%. Delayed sowing significantly shortened the rice growth cycle, reduced heat accumulation, and limited plant height, particularly at the heading and milk stages. Compensatory growth was observed at the green recovery and jointing stages, but delayed sowing led to a more than 10% increase in sterility at maturity and a 20% reduction in total yield. The constructed Logistic plant height prediction model effectively captured the regulatory mechanism of environmental and physiological factors on rice growth rate by extracting the typical distribution characteristics of plant height and growth period data and fitting the intrinsic growth rate (0.1–1.31 cm/day). The model validation results showed that the coefficient of determination (R²) for simulating plant height at different growth stages exceeded 0.5 (with R² as high as 0.82 at the jointing stage), and the root mean square error (RMSE) was controlled within 4 cm. This study provides a dynamic forecasting tool for sowing optimization, field management, and disaster early warning, with significant implications for precision agriculture and food security.