Development and decay of the LPCR are controlled by the graupel movement in the atmosphere and by the changing temperature profile (Kuettner, 1950, Williams, 1989). When LPCR fully screens the detector site we observe a positive electric field; when LPCR decays the surface detectors are exposed to the charge of the main negative (MN) layer, and, if because this charge is sufficiently large TGE continued and reach large intensities. Sure, LPCR does not screen the MN all the time, there can be several transient cases as demonstrated by Nag and Rakov in their famous paper (Nag & Rakov, 2009). The graupel fall, coincides with the start of the NS electric field reversal, and with the TGE start, and the finish – with the TGE flux decay. Thus, we can conclude that EOSO starts with LPCR decay, afterward continued with the decay of the MN layer, and in the end – decay of the main positive layer. The graduate finishing of the storm coincides with the degradation of all 3 charged layers of a tripole. The same behavior of the NS electric field, with the graupel fall, is typical for the storms on the Tibetan plateau in China (Que et al.2005, Zhang et al., 2018). Typically, conditions at the ground during EOSOs in NM storms over Langmuir Lab would be rainfall and surface temperature there is usually 15-20 C^o. We can assume that the charge sitting on the graupel is reversed to positive when the graupel has grown big at temperatures larger than -10C, or even larger than -15C (Takanashi, 1978, BERDEKLIS and LIST, 2001). Thus, using adiabatic lapse rate (-9.8C per 1 km) we come to an estimate of the possible LPCR emerging height at 3 km or more. It is natural that in New Mexico storms between ground and LPCR there could exist several differently charged regions and the sequence of the NS electric field polarity reversals when these regions are ”landed” can be arbitrary. For Aragats EOSO the pattern is more-or-less stable because we select EOSO events when we observe TGEs, that occurred mostly at surface temperatures -2-+2C. Thus, we can estimate the charged reversal height 1-1.5 km above the ground where temperatures reach a freezing level of -10 - -15C. The strong electric field can be extended low, below 100 m when the falling positively charged graupel is “lowering” the plate of the “capacitor”. Thus, the vertical extent of the accelerating field is not stable and completely decays with graupel fall. On 24 May, the duration of TGE coinciding with graupel fall was 12 minutes, in agreement with 1-1.5 km LPCR vertical size, if we assume 0.1-0.2 km/min graupel fall speed.

On 25 May during a 1-hour long storm (the first category according to classification introduced in Chilingarian and Mkrtchyan, 2012, Fig. 5), no nearby lightning flashes were registered at all. TGE was lengthy, its duration was ≈18 minutes and particle flux continued both during positive and negative NS electric field, demonstrating that both main scenarios of TGE initiation (with and without emerging LPCR) can be rather smoothly continued. The NS electric field is an illustration of decay of a mature LPCR and turning from the second scenario of the RREA initiation (2 dipoles MN-MIRR and MN-LPCR are accelerated electrons) to the first one (only dipole MN-LPCR accelerates electrons, see Fig. 1 in Chilingarian et al., 2021a). Also, it is worth noticing an intense graupel fall during positive NS field, which is evidence of the decay of the LPCR that is “sitting” on graupels.

In Table 2 we summarize characteristics of TGEs registered on Aragats during storms at the end of May 2021.