基于Logistic回归的砀山春霜冻预测模型研究与应用  

Research and application of spring frost prediction model in Dangshan based on Logistic regression

作　　者：张欣然 ZHANG Xinran(Dangshan Meteorological Bureau,Dangshan 235300,China)

机构地区：[1]砀山县气象局,安徽砀山235300

出　　处：《安徽农学通报》2025年第4期108-112,共5页Anhui Agricultural Science Bulletin

摘　　要：本研究利用2011—2018年3—5月砀山国家基本气象观测站日最低草面温度、日最低气温、日最低0 cm地温、日平均相对湿度、日最小相对湿度以及日平均露点温度观测数据和春霜冻资料,使用Logistic回归分析建立春霜冻预测模型,并对预测模型进行计算和检验,利用MICAPS软件对模型进行应用。结果表明,采用Logistic回归方法得到的影响砀山春霜冻出现的解释因子为日最低气温、日最低0 cm地温以及日平均露点温度,将这3个因子作为变量建立预测模型。该模型的检验准确率和应用准确率均在90%以上。综上,本研究建立的春霜冻预测模型使用方便快捷、预测准确率高,可应用于实际生产。The observation data of daily minimum grass surface temperature,daily minimum temperature,daily minimum 0 cm ground temperature,daily average relative humidity,daily minimum relative humidity,and daily average dew point temperature and spring frost data in Dangshan National Basic Meteorological Observation Station from March to May 2011 to 2018 used to establish a spring frost prediction model by Logistic regression analysis.The model was calculated and tested,and the MICAPS software was used to apply the model.The results showed that the factors influencing the occurrence of spring frost in Dangshan were the daily minimum temperature,the daily minimum 0 cm ground temperature,and the daily average dew point temperature,which were obtained by Logistic regression method.The 3 factors were used as variables to establish the prediction model.The test accuracy and application accuracy of the model were both above 90%.In conclusion,the spring frost prediction model established in this study is convenient,fast,and has high prediction accuracy,which can be applied to actual production.

关 键 词：梨树 春霜冻 LOGISTIC回归 预测模型 

分 类 号：S425[农业科学—植物保护]