Abstract:

This study aims to select a prediction method of rice chilling damage in different regions of Heilongjiang Province and to provide references for making food production and adjusting crop-planting structure. 11 agro-meteorological observation stations were used as the test objects, Heilongjiang Province was divided into three areas, east, west and south area, by using temperature records, 74 types of atmospheric circulation characteristics and the datum of rice phenology stage from 1971 to 2016. The stepwise regression function and the GM (1, 1) forecasting model and mean regression function of heat index were established respectively in every area. Heat index was dynamically forecasted, and then the results were analyzed. The results showed that the three prediction models all passed the residue test. The average regression-calculating accuracy of these models was above 95% from 1971 to 2010, they had little difference; the average forecast accuracy was 85%-99% from 2011 to 2016 and the average forecast accuracy of the GM(1, 1) forecasting model was 97%-99%, which was better than that of the stepwise regression function (91%-97%) and the mean regression function (85%-95%).

Key words: rice, heat index, prediction model