| 多模式集合优选方案在淮河流域夏季降水预测中的应用 |
| |
| 引用本文： | 程智,段春锋,邓淑梅.多模式集合优选方案在淮河流域夏季降水预测中的应用[J].热带气象学报,2017,33(2):241-249. |
| |
| 作者姓名： | 程智 段春锋 邓淑梅 |
| |
| 作者单位： | 1.安徽省气候中心，安徽 合肥 230031 |
| |
| 基金项目： | 公益性行业（气象）科研专项GYHY201406021安徽省自然科学基金1408085MD73 |
| |
| 摘 要： | 基于国家气候中心提供的1981—2010年4种季节气候预测模式的资料，将两种互为补充的降尺度因子挑选方案应用于淮河流域夏季降水预测，利用距平符号一致率ASCR、等级评定PG、距平相关系数ACC方法，评定了每种模式及其所采用的两种降尺度方法对淮河流域夏季降水的预测效果，并采用了一种优选方案进行多模式集合。结果表明，从4种模式的降水预测效果来看，NCEP_CFSv2和TCC_CPS1模式的评分较高，NCC_CGCM1和ECMWF_SYSTEM4模式相对较低；采用2种基于最优子集回归的降尺度方法后，NCC_CGCM1、TCC_CPS1和ECMWF_SYSTEM4模式的降尺度方法相对于模式降水预测为正订正，NCEP_CFSv2模式为负订正；将模式和降尺度预测方案进行优选，其集合平均的评分不仅高于模式降水预测的集合平均，也优于降尺度方法的集合平均，该方法发挥了不同模式的区域性优势，改进了原始集合平均的效果，为提高多模式解释应用水平提供了一种参考性方案。
|
| 关 键 词： | 气候学 模式解释应用 优选法 夏季降水 淮河流域 |
| 收稿时间： | 2015-08-03 |
Application of Optimization Scheme of Multi-Model Ensemble in Prediction of the Huaihe River Basin Summer Precipitation |
| |
| Institution: | 1.Anhui Climate Centre, Hefei 230031, China2.Anhui Institute of Meteorological Sciences, Hefei 230031, China |
| |
| Abstract: | Based on the data of four seasonal climate models supplied by the National Climate Centre, two methods of choosing complementary downscaling factors are applied in predicting Huaihe River Basin summer precipitation, effects of every model and downscaling method predicting the precipitation are evaluated by comprehensively utilizing anomaly sign consistency rate, grading evaluation and anomaly correlation coefficient methods, and an optimization scheme is applied for ensemble. The results show that seen from effects of model precipitation prediction, the CFS and TCC models are better, while the NCC and EC models are relatively insufficient; and after downscaling by the two methods based on optimal subset regression, the downscaling results of the NCC, TCC and EC models have higher scores than their original precipitation prediction thanthe CFS model; by using optimal choice between model prediction and downscaling results, the effect of the final ensemble results isfound to be better than a simple ensemble of model prediction or that of model downscaling. This optimization scheme has the advantages of different regional modes and helps improvethe predicting effect of original model ensemble and can be a reference method to improve the technical level of multi-model interpretation and application. |
| |
| Keywords: | |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《热带气象学报》浏览原始摘要信息 |
| 点击此处可从《热带气象学报》下载免费的PDF全文 |
|
