基于逐步多元回归的眉山市晚熟柑橘干旱预测模型

Drought Prediction Model of Late Maturing Citrus in Meishan City Based on Stepwise Multiple Regression

摘要: 干旱影响晚熟柑橘产量和品质，研究应用2023年5月~2034年3月土壤含水量数据和气象数据构建了基于逐步回归的晚熟柑橘干旱预测模型，4层土壤含水量模型相关系数R² > 0.93，模型验证结果得到未来6天内的误差率均小于5%，表明模型能较好地应用于预测未来土壤含水量动态变化，可为晚熟柑橘精准灌溉提供决策依据。

Abstract: Drought affects the yield and quality of late maturing citrus. The soil water content data and meteorological data from May 2023 to March 2034 were used in this study to build a drought prediction model of late maturing citrus based on stepwise regression. The correlation coefficient R² of the four-layer soil water content model was > 0.93, and the model verification results showed that the error rate in the next 6 days was less than 5%. The results showed that the model could be used to predict the dynamic change of soil water content in the future, and could provide decision-making basis for precision irrigation of late maturing citrus.