天气现象自动化观测资料可用性分析与应用探讨

通过对北京地区天气现象组网试验资料的可用性分析以及在精细化预报业务中的应用,发现自动观测天气现象资料连续时长多于20 min以上时,数据可信度较高,对预报员有较高的参考价值。自动观测系统对雨、雾、霾现象的识别结果正确率总体较好,其高频次的探测数据对低频次的人工观测起到补充作用。但在天气现象识别算法、质控、雷暴和降水性质识别、现象识别连续性等方面还存在不足,需要进一步改进。

Analysis and Discussion On Auto-observation of Weather Phenomena

Abstract: Based on the analysis on the data from the automatic weather phenomena observing net and the application on refined forecasting, it is found that the antomatic observation data, considering the temperal span binned as 20 minutes, could have a high reference value for the forecasters. The accuracy rate of the identified weather phenomena which is complementary to manual observation, like rain, fog and haze, is totally high. Also the results show that the defects on the identified algorithm, the identification of thunder and precipitation phase, the continuity still exist, and should be further improved.