A statistical quality control approach of ground-based weather radar data

Ground-based weather radars are usually suffers from surrounding terrain, and thus the radar data are contaminated by ground clutter and beam blockage. Therefore, quality control on ground-based radar data must be carried out before the data can be further used.

Conventional methods are rather difficult to entirely eliminate ground clutter and beam blockage for long-term massive data due to the complexity and diversity of radar echoes. In this study, a new quality control method is proposed based on statistical features of radar echoes. With the assumption that precipitation distribution within the radar detection range is of climate homogeneity, radar echoes have features: 1) the distribution of ground clutter and beam blockage is relatively fixed while the distribution of precipitation is relatively even; 2) the reflectivity of ground clutter is very large while the reflectivity of beam blockage is very small, and the reflectivity of precipitation falls between the two extremes. Consequently, the ratio of the times that the reflectivity exceeds a given threshold can be calculated, and then a fuzzy clustering technique is used to identify areas where ground clutter and beam blockage frequently appear. Based on the identification results, quality control can be implemented by removing the contaminated echoes appear in these areas.

A comparative analysis of quality control results on Nanjing weather radar data with the new and traditional methods suggests that the new method can effectively identify ground clutter and beam blockage to improve the radar data quality.