论文部分内容阅读
基于南方双季稻种植区708个气象站1961—2010年的逐日气象资料、双季稻低温灾害发生的气象行业标准和1960—2010年逐月74项大气环流特征量资料,采用因子膨化、相关性分析、逐步回归等方法,建立了针对不同风险和时空变化趋势的分区双季稻低温灾害历年第一次灾害发生等级预测模型。结果表明:高风险区(Ⅰ区)早稻春季低温灾害、晚粳稻寒露风、晚籼稻寒露风的预测模型平均外延预测基本一致准确率分别为100%、83.3%和83.3%;低风险且呈增加趋势区(Ⅱ区)早稻春季低温灾害、晚粳稻寒露风、晚籼稻寒露风的预测模型平均外延预测基本一致准确率分别为100%、83.3%和83.3%;低风险且呈减少趋势区(Ⅲ区)早稻春季低温灾害、晚粳稻寒露风、晚籼稻寒露风的预测模型平均外延预测基本一致准确率分别为83.3%、100%和83.3%;各预测区域各代表站预测模型的回代和预测等级误差基本在1个等级以内,具有较高的精度。
Based on daily meteorological data of 708 meteorological stations in the southern double cropping rice fields from 1961 to 2010, meteorological data of low temperature disaster occurrence in double cropping rice and 74 monthly atmospheric circulation characteristics from 1960 to 2010, This paper established a prediction model of the first disaster occurrence level for the double cropping rice in the past years with different risks and spatial and temporal variations. The results showed that the average accuracy of predictive models for predicting cold and dampness in late-season rice with cold-exposed or late-season indica rice was 100%, 83.3% and 83.3% respectively in high-risk areas (area Ⅰ), and low-risk and increasing In the trend area (Ⅱ), the average consistent predictive accuracy of the predictive models for cold morning dew of late japonica rice and cold dew of late indica rice was 100%, 83.3% and 83.3% respectively in low temperature disaster of early rice in spring; Area) The basic uniform accuracy of the predictive models for early rice in spring cold season, the cold dew in late japonica rice and the cold dew in late indica rice were 83.3%, 100% and 83.3%, respectively; and the regression and prediction of the prediction models of representative stations in each prediction area Level error is basically within 1 level, with higher accuracy.