Multi-scale analysis for automatic weather station data

Wavelet based method is presented in this paper to improve the quality of an observed series obtained with automatic weather station. The first part of this paper surveys the drawback of data series collected with automatic-weather-station. A method of multi-stage analysis is introduced in second session of this article. The third part of this article provides an application of denoising noisy data using wavelet method with different threshold. At the end of this paper, we propose an example to detect anomaly mixing in the date series. Being the experimental result that an outlier record of temperature series can be detected with the multi-scale technology, the propose way in this article guaranteed the quality control of automatic weather station data.