【摘要】 为了满足本地实际业务和服务需求,利用0.125°×0.125°欧洲中期天气预报中心ECMWF模式逐3 h最低气温预报产品和1 km×1 km逐时气温格点实况场资料,使用三种Kalman滤波方法订正贺兰山东麓144 h内逐3 h最低气温和霜冻1 km×1 km的格点预报,并对比检验订正前后四种产品的预报效果,检验结果表明:经过Kalman滤波方法订正后,贺兰山东麓最低气温预报误差和霜冻预报准确率及稳定性明显提高,三种Kalman滤波方法的平均绝对误差分别较ECMWF降低了0.76℃、1.04℃和1.12℃,霜冻预报准确率提高了3.2%、3.9%和5.3%;三种Kalman滤波方法对于1 400 m以上的高海拔山区订正效果显著,误差降低了2～5℃,霜冻预报准确率提高了10%～30%,且海拔越高,订正效果越明显,而对于1 200 m以下的低海拔平原地区,由于ECMWF模式预报效果较好,订正能力有限;相较而言,最优集成Kalman方法的订正效果最好。更多还原

【Abstract】 To meet the local actual meteorological service demand, based on the ECMWF 3-hourly minimum air temperature(0.125°×0.125°) and hourly air temperature grid data(1 km×1 km), the gridded weather forecasting correction of 3-hourly minimum air air temperature and frost(1 km×1 k m) within 144 hours in the Eastern Region of Helan Mountain were investigated with three methods of the Kalman filter. The results showed that minimum air temperature prediction error reduced dramatically and frost forecast accuracy and stability improved significantly after using the correction of the Kalman filter. The mean absolute error of three Kalman filtering methods were 0.76℃, 1.04℃ and 1.12℃ lower than ECMWF, frost forecast accuracy enhanced 3.2%, 3.9% and 5.3% respectively. Three methods of Kalman filter had noticeable correction efficiency in the high altitude mountain area above 1 400 meters, the error was reduced by 2～5℃, frost forecast accuracy was increased by 10%～30%. And the higher the altitude, the more obvious the correction efficiency was. However, the correction capability was limited in the low altitude plateau area below 1200 meters due to the better forecasting effectiveness of ECMWF. Comparatively speaking, the correction of optimal integrated method Kalman filter worked best.更多还原