AbstractElectric field values measured inside thunderclouds have consistently been reported to be up to an order of magnitude lower than the value required for the conventional electrical breakdown of air. This result has made it difficult to explain how lightning frequently occurs in thunderclouds. A few different theories have been offered to explain the lightning initiation process, one of them being the theory of lightning initiation from hydrometeors. According to this theory, lightning can be initiated from electrical discharges originating around thundercloud water or ice particles in the measured thundercloud electric field. These particles, called hydrometeors, are believed to cause significant enhancement of the thundercloud electric field in their vicinity and then initiate streamers that are the precursor discharges for the hot lightning leader channel. Previously, Liu et al. (2012a) reported streamer formation from a model hydrometeor in an electric field value of half of the conventional breakdown threshold (Ek) for air. In this paper, we present modeling results for streamer formation in electric fields as low as one third of the breakdown threshold. According to our results, initiation of stable streamers from thundercloud hydrometeors in a 0.3Ek electric field is possible, only if enhanced ambient ionization levels (e.g., the ionization created by corona discharges around the same or other nearby hydrometeors) are present ahead of the streamer. The magnitude and distribution of this ambient density may be a determining factor on whether the streamer branches, recovers after the prebranching stage, or continues propagating stably. We investigate the streamer branching behavior and characteristics and test a theory that has recently been proposed to explain this phenomenon. We find that the geometry of the streamer head plays an important role in the streamer branching phenomena. The fast radial movement of the maximum streamer head curvature, combined with the slow reduction of the maximum curvature value, eventually leads the streamer head to branching. Finally, we compare our modeling results with laboratory experiments and realistic thundercloud conditions and discuss the implications of this study to lightning initiation and other lightning‐related phenomena.
