Properties of preliminary breakdown processes in Scandinavian lightning

Lightning flashes are usually preceded by preliminary breakdown processes (PBPs) before a stepped leader is initiated. These breakdown processes are not well understood. An early model, the so-called BIL model, has been called into question in later studies. However, we have found that the BIL model is quite successful in describing initial processes at least in high-latitude Scandinavian lightning. We present results from one summer of measurements in Finland, during which the vertical electric field was measured with a standard broadband plate antenna system. Lightning flash locations were provided by a lightning detection network and magnetic fields were measured with an experimental narrowband detection system. The relationship between the preliminary breakdown and the first return stroke (RS) is studied for 193 flashes at distances of 5–70 km. We can identify a preliminary breakdown in at least 90% of the flashes. The peak electric field of the RS is on average four times as intensive as the highest peak of the PBP. However, in 25% of the cases the PBP peak is more intensive. On the other hand, we show that this method of comparing intensities is physically arbitrary, since the PBP is continuous and the RS is impulsive. The narrowband measurement allows a physically consistent definition for intensities as the root-mean-square (RMS) sum of the most intense parts of signals. The PBP and RS are shown to have almost equal intensities at small distances. At larger distances, the PBP weakens more rapidly. This is suggested to be due to different propagation regimes, with the PBP signal changing from space-wave to ground-wave propagation with increasing distance, while the RS is predominantly ground wave at all distances. The result may have practical applications in narrowband detection of lightning. The BIL model suggests a characteristic signal in the narrowband signal, which could be used to identify the start of a lightning flash. The change in the RS–PBP ratio as a function of distance is statistically significant, but is too weak to significantly improve ranging methods.