Improving rice development and phenology prediction across contrasting climate zones of China

Modelling rice development rate and phenology is crucial for robust yield prediction. However significant uncertainties originating from the temperature response functions in rice models exist. In this study, both the Wang-Engel temperature response function and the Johnson temperature response function were coupled in the MCWLA-Rice model and evaluated at 10 agro-meteorological experiment stations located in contrasting climate zones across China. Results show that the MCWLA-Rice model with the Wang-Engel temperature response function is able to improve the predictions of heading and maturity date for double season rice at the stations in the middle and lower reaches of the Yangtze River. Mean improvements were 23.5% and 25.1%, respectively. The MCWLA-Rice model with the Johnson temperature response function is able to improve the predictions of heading date by 26.2% and maturity date by 22.9% on average at all stations. At the locations characterised by high temperature stress conditions, both the Wang-Engel and the Johnson temperature response functions are able to improve predictions of heading and maturity date, nevertheless the improvement is not as pronounced as for those locations not suffered from high temperature stress. These results demonstrate how rice phenology simulation can be improved, yet suggest that further improvements should be made to reproduce phenology in locations suffering from heat stress. This study highlights the value of environment-controlled experiments on the effects of exceeding threshold temperatures from several hours to several days on instantaneous crop development rate and improving crop phenology simulations.