Abstract:

To scientifically assess the losses caused by continuous rain disasters during the maturity stage of tobacco leaves in Xuancheng, this study was conducted based on daily meteorological data, tobacco yield data, and phenological development data from 7 counties (cities, districts) in Xuancheng City from 2005 to 2024. Using statistical analysis, correlation analysis, and the K-Means clustering algorithm, we systematically carried out research on the screening of disaster-causing factors, determination of critical thresholds, construction of intensity indices, and classification of disaster grades for continuous rain disasters. The results showed that total precipitation (R_R) and duration of rainy days (D_R) during continuous rain were the core disaster-causing factors for continuous rain disasters during the maturity stage of tobacco leaves in Xuancheng. Both factors were significantly positively correlated with yield loss rate (ΔY) (P<0.05), among which R_R during the middle leaf maturity stage was extremely significantly correlated with ΔY (P<0.01). Based on the weight coefficient method, continuous rain intensity index models were constructed for three developmental stages: lower leaves, middle leaves, and upper leaves. Combined with K-Means clustering, the disasters were divided into four grades: slight, moderate, severe, and extremely severe, with the corresponding intensity index thresholds, yield loss rate intervals, and occurrence probabilities clarified. Verification results of measured data showed that the overall accuracy of the evaluation system with a grade difference ≤1 reached 88%, with an accuracy of 86% for slight and moderate disasters, and 89% for severe and extremely severe disasters. In terms of spatial applicability, the average accuracy was 90% in the northern region and 86% in the southern region. The continuous rain disaster loss evaluation system for tobacco leaves during the maturity stage established in this study fills the research gap in the Xuancheng tobacco-growing area and realizes precise stage-based and grade-based disaster evaluation. It can provide scientific technical support and decision-making basis for the monitoring and early warning, risk prevention and control, and optimization of tobacco planting layout under continuous rain disasters in Xuancheng tobacco-growing area.

Key words: Xuancheng, tobacco maturity stage, continuous rain disaster, disaster-inducing factors, intensity index, loss assessment system