Estimating the temperature sensitivity of rice (Oryza sativa L.) yield and its components in China using the CERES-Rice model

Abstract

The effects of temperature changes on rice (Oryza sativa L.) yield and its components have been widely documented. However, most existing studies are based on small-scale, short-term field experiments, with few assessing these effects on a large scale or over long periods. Here, the calibrated Crop Environment Resource Synthesis (CERES)-Rice model was used for numerical simulations over six climate regions in the major rice cultivation areas of China for the period of 1989–2018. The simulated results were used to estimate the temperature sensitivity of rice yield with a panel model in each climate region, and the yield sensitivity was then decomposed into the temperature sensitivity of three components: panicle number per unit area (Pan_no), filled grain number per panicle (Grain_no), and grain weight (Grainwt). Results indicated that rice yield exhibited negative temperature sensitivity across all climate regions, driven primarily by the temperature sensitivity of Grain_no in most regions. Additionally, Grainwt did not vary with temperature in all regions. Further analysis suggested that yield, Pan_no, and Grain_no were more sensitive to high temperature degree days (HDD) than to growing degree days (GDD). Under the warmer scenarios, HDD increase played a dominant role in the reduction of Grain_no, while the joint effect of GDD and HDD resulted in an increased Pan_no in most regions. However, the negative effect of temperature on Grain_no outweighed its positive effect on Pan_no, leading to a decline in yield. This study provides insight for understanding the temperature response of rice yield and its components and will be beneficial for developing targeted adaptations to ensure rice sustainable production under global warming.