关键词: 马铃薯晚疫病, 最大熵模型MaxEnt, CARAH模型, 气候风险

Abstract:

The hourly average temperature and average relative humidity data of 260 meteorological stations in Chongqing from February to June in 2019−2023 were used to simulate the geographic distribution of spring potato late blight infection risk using the CARAH late blight model. The accuracy of the simulation was tested by using the late blight infection data of 26 monitoring stations in Wuxi county, Chongqing in 2022. Based on the geographic distribution of spring potato late blight infection risk simulated, the maximum entropy model was constructed using the climate grid data of the monthly average temperature, maximum temperature, minimum temperature, water vapor pressure and precipitation from February to June in 1970−2000 to analyze the climate impact factors of spring potato late blight in Chongqing, and to evaluate the climate risk of spring potato late blight, providing a reference for the prediction and scientific prevention of the disease. The results showed that simulations of late blight infection based on hourly weather data had high accuracy, with a false positive rate of 12.5%, false negative rate of 18.5% and TS score of 0.73. The mean area under curve (AUC) of the receiver operating characteristic (ROC) was above 0.9, indicating higher accuracy of the simulation results. Precipitation was the dominant climate factor affecting the risk distribution of late blight of spring potato in Chongqing, while relative humidity and temperature were important climate factors. Climate variables at the seedling stage and bud flowering stage had a great impact on the distribution of late blight risk. The low risk area of late blight of each maturity(early/late) and susceptibility (resistant/susceptible) combination of spring potato was less than or close to 10000km², with an average area proportion of 10.2%. The medium risk area and high risk area were both more than 30000km², with an average area proportion of 43.7% and 46.1%, respectively. The climate risk of spring potato late blight showed a spatial distribution characteristic of "high in the middle and low in the periphery" in Chongqing. The high risk area was concentrated in the parallel valley area in eastern Sichuan, the medium risk area was mainly distributed in Daba mountain area in northeast Chongqing, Wuling mountain area in southeast Chongqing, and the hilly area in central Sichuan in west Chongqing, and the low risk area was scattered in the fringe of Chongqing. Spring potato production in Chongqing area faces a high climate risk of late blight, with significant spatial differentiation characteristics. It should be addressed through reasonable production layout and improved cultivation techniques.

Key words: Potato late blight, MaxEnt model, CARAH model, Climate risk