| description abstract | he impact of climate change on rice productivity in China remains highly uncertain because of uncertainties from climate change scenarios, parameterizations of biophysical processes, and extreme temperature stress in crop models. Here, the Model to Capture the Crop?Weather Relationship over a Large Area (MCWLA)-Rice crop model was developed by parameterizing the process-based general crop model MCWLA for rice crop. Bayesian probability inversion and a Markov chain Monte Carlo technique were then applied to MCWLA-Rice to analyze uncertainties in parameter estimations and to optimize parameters. Ensemble hindcasts showed that MCWLA-Rice could capture the interannual variability of the detrended historical yield series fairly well, especially over a large area. A superensemble-based probabilistic projection system (SuperEPPS) coupled to MCWLA-Rice was developed and applied to project the probabilistic changes of rice productivity and water use in eastern China under scenarios of future climate change. Results showed that across most cells in the study region, relative to 1961?90 levels, the rice yield would change on average by 7.5%?17.5% (from ?10.4% to 3.0%), 0.0%?25.0% (from ?26.7% to 2.1%), and from ?10.0% to 25.0% (from ?39.2% to ?6.4%) during the 2020s, 2050s, and 2080s, respectively, in response to climate change, with (without) consideration of CO2 fertilization effects. The rice photosynthesis rate, biomass, and yield would increase as a result of increases in mean temperature, solar radiation, and CO2 concentration, although the rice development rate could accelerate particularly after the heading stage. Meanwhile, the risk of high-temperature stress on rice productivity would also increase notably with climate change. The effects of extreme temperature stress on rice productivity were explicitly parameterized and addressed in the study. | |
