Parsivel激光雨滴谱仪观测降水中异常数据的判别及处理

The detection and elimination of abnormal data for the precipitation observed by Parsivel precipitation particle spectrometer

2009年利用Parsivel激光雨滴谱仪连续观测太湖地区自然降水,发现对流云降水采样的粒子谱数据常常包含有少数异常高、低数据。从Parsivel激光雨滴谱仪的测量原理出发,分析了产生测量误差的原因。然后依据3σ准则(莱以特准则)对采集的一次对流云降水粒子谱数据进行分析,对其是否存在粗大误差进行判别和消除。结果表明:异常数据即为仪器测量的粗大误差,消除误差后的数据在表征降雨量、雷达反射率因子等降水特征量时更加接近于仪器自身输出值。同样地将此方法用来分析层状云降水,有利于消减仪器测量中粒子重叠所造成误差。

Observing continuously the precipitation of Tai Lake region by a laser optical precipitation particle spectrometer in 2009, we find that there are a few of abnormal data in the cumulonimbus precipi- tation data. According to working principle of the Parsivel precipitation particle spectrometer, the paper analyses and discusses the origins of abnormal data. Then, the 3σ criterion is used to detect whether er-ror exists in the cumulonimbus precipitation data and tried to eliminate. The result shows that the abnormal data resulted from the gross error of instrument measure, and the processed data is better compared to the output of the instrument in evaluating precipitation physical parameters, such as rainfall, radar reflec- tivity factor. The same method is used in analyzing the stratiform precipitation, which is in favor of de- creasing the measure error resulting from raindrop superposition.