关键词: 冬小麦返青期, 线性回归, 灰色系统模型GM（1, 1）, 河北省

Abstract: Using observed data from 11 agrometeorological stations in Hebei province from 1981 to 2018, the forecast models of returning green stage of winter wheat were constructed in Hebei province based on the methods of linear regression and grey system model GM （1, 1）, and the performance of the simulation effect of forecast models was evaluated by historical data. The results showed that the linear trend simulation curve of the returning green stage of winter wheat in each station was insignificance. The grey system model G （1,1） is used to simulate the winter wheat returning to green period in each station, and the simulation accuracy of 10 stations out of 11 stations is above the second level. By analyzing the correlation between the accumulation temperature sequences and the returning green stage of winter wheat, it was determined that the negative accumulation temperature from October 1 to February 15 of next year was the key influencing factor of the returning green stage of winter wheat. The linear regression equations of the negative accumulation temperature of key period and integrated sequence of the returning green stage of winter wheat of all stations reached a very significant level. The probability of an absolute error of less than one week（7 days） of single station linear trend mode, grey system model GM （1,1）,and the linear regression model for all stations with negative accumulation temperature as the dependent variable were 73%, 94%, and 80%. The prediction model of the returning green stage of winter wheat can be constructed based on the grey system model GM （1, 1） under the condition that the change of data is relatively stable. When there is mutation fluctuation in data, the prediction model can be constructed by linear equation at single station. Under the background of climate change,the prediction of winter wheat greenback period by negative accumulated temperature in critical period is suitable for regional analysis based on the linear regression equation of the comprehensive data series of various stations.

Key words: the returning green stage of winter wheat, linear regression, grey system model GM （1, 1）, Hebei province