Spatiotemporal variation characteristics and forecast model construction of high-temperature heat damage intensity in rice

An evaluation index is crucial to quantitatively represent the intensity of high-temperature heat damage in rice. A medium and long-term forecasting model can be established to accurately forecast the trend of high-temperature heat damage, and then deploy heat-resistance measures in advance. In this research, spatiotemporal evolution was used to fully consider the cumulative effects of heat damage. The disaster indicators were combined to build an intensity index, and then objectively grade using the Fisher partition. According to medium and long-term weather forecasting, two types of forecasting factors were selected as the sea surface temperature (SST) and atmospheric circulation indices. The forecasting models were finally constructed and then validated for the index of high-temperature heat damage intensity using two-dimensional optimization. The results show that: 1) There was a significant increase in the annual frequency and intensity of high-temperature events and heat damage on the rice crops since 2010. Notably, the intensity of high temperature and heat damage reached the highest grade 4 in 2013 and 2022. While it was only in grade 3 in 2016, 2017, and even as far back as 1992. There was a significant inter-decadal difference in the heat damage intensity index. There was an increase in the intensity of rice high-temperature heat damage in the province each year since the 1990s. In spatial patterns, the areas with a cumulative value in the heat damage intensity index of ≥10 were gradually expanded from south to north, extending to the west of Huai Bei in the 2010s. A significant increase was found in the intensity of heat damage south of the Huai River, indicating the strongest in history. 2) The prediction models were constructed for two kinds of high-temperature heat damage intensity index in early July using Pacific SST and 88 atmospheric circulation indexes. At the same time, Jiangsu rice had just been transplanted, and there were still 1-2 months before the stage of rice joining and heading. The forecast lead was 1-2 months after the fitting and trial test. The forecast models generally simulated the interannual fluctuation of the intensity index. But there was some variation in the amplitude difference between the simulated and actual values. 3) The optimal correlation between the dependent and independent variables was carried out to optimize the forecasting factors after comparison, leading to effectively improv the simulation of the forecasting model. Among them, the best performance was achieved in the forecasting model of sea temperature factors for the high-temperature heat damage intensity index after optimization. The findings have important scientific significance and reference to defend against high-temperature disasters in agricultural meteorological services for food security.