Evaluation of the Blitzortung lightning locating system

Blitzortung is a non-commercial lightning detection system intended for recreational use only. In this thesis the performance of the Blitzortung system is evaluated in order to see to which extent it can be compared to a high performance commercial lightning detection system for use in Sweden. Also a graphical tool for visualising lightning strikes detected by Blitzortung on maps containing an electric grid with the related infrastructure is proposed. The evaluation is accomplished by comparing lightning data registered by Blitzortung with lightning data from the national lightning locating system provided by the Swedish meteorological institute. By analysing lightning data from the lightning high-season in Sweden between the year 2012-2016, and by conducting a more in depth evaluation of the system based on data from 2016, the Blitzortung system performance is evaluated. Results for calculated flash density and flash rate density suggests that the Blitzortung system is improving over the study time, which is probably due to the fact that the number of sensors in Sweden has increased from a single station in 2012 to 30 stations in 2016. The performance evaluation show a relative detection efficiency of 34 % for strokes correlated between the systems in relation to the total strokes detected by SMHI. For strokes detected south of latitude 61 in Sweden the relative DE (detection efficiency) is 44.7 %, and for CG(cloud-toground) strokes it is 43.0 %. If only CG-strokes south of latitude 61 is considered the corresponding value is 55.1 %. The performance of the Blitzortung-system seems to be depending on the number of sensors in the region of interest. The relative positional error between the networks south of latitude 61 was shown to have a smaller median error of 1750 m and an upper quartile on 3350 m, compared to north of latitude 61 where the median error was 2780 m and the upper quartile 5900 m. The results indicates a systematic directional error for correlated strokes between the systems. It is also shown that the included parameters for Blitzortung strokes could not be used in any meaningful way for predicting positional errors between the networks