新疆北部棉花播种期热量指数预测研究

应用前期74类环流指数因子,考虑了统计模型中因子与预测量之间相关系数的不稳定性,用滑动集合回归方法建立了北疆棉区播种-出苗期间热量指数的集合回归预测模型。石河子、炮台、精河、博乐站1961-2006年的回归序列与实况序列的相关系数分别为0.804 0,0.756 2,0.850 3,0.825 2。4站1996-2005年的试报结果检验表明:集合回归预测模型的效果明显优于一级回归模型,而4站中又以博乐站的效果最优。应用前期月环流指数因子,经滑动相关建立的超级集合模型,对北疆棉区播种-出苗期间的热量指数有一定的预测作用,该方法在农业气象灾害预测等领域有推广价值。

Study on the Prediction of Heat Index during Cotton Seedtime in North Xinjiang

In this study,an integrated regression prediction model of heat index during the period from cotton seedtime to emergence season in 4 cotton-growing regions,i.e.the Shihezi,Paotai,Jinghe and Bole regions located in north Xinjiang,is developed using 74 factors of general circulations in previous one month based on considering the instability of the correlation coefficients between the factors in the statistical model and the predicted values and applying the smoothed integrated regression.The results show that the correlation coefficients between the predicted and observed values of heat index in these 4 cotton-growing regions during the period from 1961 to 2006 are 0.804 0,0.756 2,0.850 3 and 0.825 2 respectively.The tested results of experimental results show that the integrated regression prediction model is better than the regression models in predicting the heat index during the period from cotton seedtime to emergence season in the 4 cotton-growing regions,especially in Bole.The results also reveal that the integrated regression prediction model is applicable in predicting the heat index during the period from cotton seedtime to emergence season in the cotton-growing regions in north Xinjiang.