新疆巴州农区气温的多尺度变化及其预测

利用Morlet复数小波变换和最大熵谱分析了巴州农区各个季节气温的多时间尺度变化特征，利用均生函数预测模型对巴州地区气温进行预测试验。结果表明：巴州农区冬季、秋季气温年代际变化不明显，年际变化明显；春、夏季气温的年代际变化明显；四季气温均存在显著的周期变化。对2000～2002年预测试验表明，均生函数预测模型对巴州地区气温有较好的预测能力。

Multiple- Time- Scale Change and Prediction of Agricultural District Temperature in Bazhou,Xinjiang

Based on the Morelet complex number wavelet transformation and max-entropy spectrum analysis, analyzed the multiple-time-scale change of temperature in every season in the agricultural district of Bazhou, then conducted prediction test with catch function predict model. The results were shown:(1)the interdecade change of temperature in fall and winter was not obvious, but interannual change was obvious, the interdecade change of temperature in spring and summer was obvious. The temperature exist remarkable periodic variation in four seasons.(2)the catch function model had better capacity to predict the temperature in the agricultural district of Bazhou by the prediction test from 2000 to 2002.