Real-time risk assessment of rice chilling injury based on disaster processes in Northeast China

Although global warming has reduced the frequency of large-scale severe chilling injury in Northeast China, such risks have not been eliminated. Regional and periodic chilling injury events still occur frequently and are even showing signs of increasing intensity. To enhance the capability of chilling injury identification and early warning, this study constructed a cumulative intensity index based on the daily accumulation process. Meteorological data from Northeast China (1991–2020) were used to analyze the spatiotemporal differentiation characteristics of rice chilling injuries. It also integrates sun-induced chlorophyll fluorescence (SIF) to assess crop adaptability. Furthermore, random forest (RF), the DSSAT crop model, and the analytic hierarchy process (AHP) were employed to incorporate a natural disaster risk index and develop a real-time risk assessment model for the rice chilling injury. The model was validated using three chilling events from 2002, 2017, and 2020. Results showed that during the tillering and booting stages, SIF is significantly negatively correlated with the cumulative intensity index, indicating that SIF had strong potential as an early remote sensing indicator for chilling injury. Through retrospective real-time risk assessments of three representative historical chilling injury events, we found that high-latitude rice growing areas are more vulnerable to chilling stress. For instance, during the 2002 chilling event, the proportion of moderate-risk areas rose to 49 %, while high-risk areas reached 17 %. This spatial distribution pattern reflects historical disaster trends and supports farmers and agricultural agencies in assessing chilling injury risk, enabling timely preventive measures and contributing to national food security.