Approbation of automatic allocation of thunderstorm foci in the south of Western Siberia according to WWLLN data according to the FRIS-Tax clustering algorithm

Abstract

The article shows the ability of the FRiS-Tax algorithm to separate lightning discharge clusters and combine them into classes. Lightning discharge clusters are a model of an individual lightning cell. Classes (groups) of clusters of lightning discharges that are close in space and in time of action can be associated with a model of a multi-cell thunderstorm or a thunderstorm focus. Clustering of lightning discharges is carried out using normalization values specifying the linear dimensions of a thunderstorm cell (ρ*) and its lifetime (τ*). The experimental sample consisted of 3934 lightning discharges registered by WWLLN in the south of Western Siberia, August 7-8, 2017. The results of lightning discharge clustering are compared for two parameterization conditions: ρ1*=50 km, τ1*=30 min and ρ2*=15 km, τ2*=20 min For the first set of parameterization conditions, the resulting lightning discharge clusters represent zones of possible activity of a thunderstorm cell for one to two hours. Classes formed from these clusters describe the behavior of long lasting and vast thunderstorm foci. The clusters obtained for the second set of parameterization conditions better correspond to the linear sizes and lifetime of individual thunderstorm cells. An adjustment of the clustering algorithm is required to combine such clusters into classes corresponding to a multi-cell thunderstorm.