Bridging the gap between observed and simulated supercell cold pool characteristics

A number of recent investigations of near storm environment (NSE) characteristics have identified specific favorable environmental conditions for supercell storms which can also to some extent discriminate the likelihood of significant tornadoes (e.g. Thompson et al., 2003, 2007; Davies, 2004; Bunkers et al., 2006a,b). Among those conditions identified for mature storms is the presence of modest convective inhibition, which as Bunkers et al. (2006b) noted aids in storm isolation and longevity leading to enhanced hazard potential. However, there may be a further aspect of convective inhibition (CINH) that may favor significant tornado potential when this CINH owes largely to the presence of a capping inversion. Notably, surface observation based studies (Markowski et al., 2002; Shabbott and Markowski, 2006; Grzych et al., 2007) have found limited baroclinity along gust fronts of significantly tornadic storms. This finding contrasts with numerical simulations of supercell storms which appear to invariably develop strong cold pools (large gust front baroclinity) for both tornadic and non-tornadic storms (Markowski, 2002; Davies-Jones, 2009). Traditionally, a numerical simulation of a supercell would often require prior modification of the raw NSE sounding to remove features such as capping inversions so that a warm bubble can initiate convection (e.g., Klemp and Wilhelmson, 1978). Recently, Snyder and Zhang (2003), Dowell et al. (2004) and others have developed methods to initialize storm-scale ensembles and to use the ensemble Kalman filter (EnKF) to retrieve atmospheric state estimates roughly consistent with observed storms. A benefit of these methods is that it is possible to simulate storms with raw soundings, including soundings with significant capping inversions, to define the background state. Using the case study presented by Romine et al. (2008) and the Dowell et al. (2009) EnKF data-assimilation system, we will evaluate the influence of the capping inversion (or lack thereof) on the retrieved cold pool. A preliminary result is that removing the capping inversion from the raw sounding leads to increased size and intensity of the retrieved cold pool and poorer kinematic retrieval.