Abstract: In this paper, a new data processing method based on a kind of improved "PauTa" criterion is put forward for the treatment of real time precipitation data of regional automatic weather stations, to solve the problems occurred in related data applications. First, sample space is collected by Euclidean distance clustering, then the sample mean and the residuals are calculated, after that the definition for a suspicious degree is given, to calculate real suspicious degree of various sites. In the end, actual data in Zhejiang Province are used for parameter analysis, which proves the validity and practicability of the new method.

Key words: automatic station, quality control, real time information, ", PauTa", rule