东北地区春播期土壤水分预测方法研究

根据东北五个土壤湿度分区春播期土壤相对湿度与气象因子、环流因子相关关系,甄选土壤相对湿度主要影响因子,构建多元线性回归模型以预测4、5月土壤相对湿度。结果表明:(1)五个分区土壤相对湿度与封冻雨、降水量呈正相关,与透雨日期、气温及日照时数呈负相关,且4月份土壤相对湿度与封冻雨及透雨日期的相关性较5月更显著;(2)各分区春播期土壤相对湿度与多项环流因子关系较密切,相关系数绝对值大于0.4且通过信度为0.05的显著性检验,并且这些影响因子主要集中在前一年11、12月;(3)基于关键气象因子和环流指数构建的4、5月土壤相对湿度多元线性回归模型均通过F显著性检验,复相关系数达0.5以上,且模型预测准确率基本在80%以上,精度较高。

Research on forecasting method for soil moisture during spring sowing period in Northeast Area of China

According to the related relationship between soil relative moisture with climate factors and circular factors in spring sowing period of five soil moisture regions in Northeast Area, selected the major impact factor for soil relative moisture, constructed the multiple linear regression models to predict the soil relative moisture during April and May. The results indicated that: (1) The soil relative moisture with freezing rain and precipitation were shown positive correlation in five regions, while negative correlation with the date of soaking rain, temperature and sunshine hours. Moreover the relativity of soil relative moisture with freezing rain and the date of soaking rain in April were more remarkable than in May. (2) The soil relative moisture in spring sowing period for each region was rather close relation with multiple circular factors, the absolute values of correlation coefficients was more than 0.4 and passed the significant test at 0.05 level, also these impact factors were major concentrated in November and December of the previous year. (3) Based on key climate factors and circular index to construct the multiple linear regression models for the soil relative moisture during April and May were total passed the F test and the multiple correlation coefficients were larger than 0.5, also the forecasting precision of model was basically more than 80% with rather high accurac y.