Near-Ground Wind Profiles of Tornadic and Nontornadic Environments in the United States and Europe from ERA5 Reanalyses

Abstract

The near-ground wind profile exhibits significant control over the organization, intensity, and steadiness of low-level updrafts and mesocyclones in severe thunderstorms, and thus their probability of being associated with tornadogenesis. The present work builds upon recent improvements in supercell tornado forecasting by examining the possibility that storm-relative helicity (SRH) integrated over progressively shallower layers has increased skill in differentiating between significantly tornadic and nontornadic severe thunderstorms. For a population of severe thunderstorms in the United States and Europe, sounding-derived parameters are computed from the ERA5 reanalysis, which has significantly enhanced vertical resolution compared to prior analyses. The ERA5 is shown to represent United States convective environments similarly to the Storm Prediction Center’s mesoscale surface objective analysis, but its greater number of vertical levels in the lower troposphere permits calculations to be performed over shallower layers. In the ERA5, progressively shallower layers of SRH provide greater discrimination between nontornadic and significantly tornadic thunderstorms in both the United States and Europe. In the United States, the 0–100 m AGL layer has the highest forecast skill of any SRH layer tested, although gains are comparatively modest for layers shallower than 0–500 m AGL. In Europe, the benefit from using shallower layers of SRH is even greater; the lower tropospheric SRH is by far the most skillful ingredient there, far exceeding related composite parameters like the significant tornado parameter (which has negligible skill in Europe).

Reconstruction of Violent Tornado Environments in Europe: High‐Resolution Dynamical Downscaling of ERA5

“…However, violent tornado hodographs presented in this study (especially for WRF; Figure 3a) seem to be more comparable to those presented in Coffer et al. (2020) and Nixon and Allen (2021).…”

Machine learning classification of significant tornadoes and hail in the U.S. using ERA5 proximity soundings

“…Low-level SRH ranked in the top 5 for both models, but with different choices of integration bounds (10-500 m for RFC, and 0.01-1 km for LOGIT). A majority (7 of the top 10) of predictors for the RFC model were noted as being kinematic, which agrees with previous research examining the best environmental discriminators for tornadoes of various magnitudes (e.g., Brooks et al 2003a;Rasmussen 2003;Markowski et al 2003;Thompson et al 2003;Nowotarski and Jensen 2013;Hampshire et al 2018;Coffer et al 2019;King and Kennedy 2019;Coffer et al 2020). LOGIT had fewer kinematic variables (5) in the top 10.…”

“…First, vertical profiles were uniformly assessed across the entire study domain and temporal record. Previous studies have examined SCS environments as a function of space and time, noting that a variety of parameters used in this study displayed geographic and/or seasonal relevance (e.g., Davies and Johns 1993;Johns et al 1993;Brooks et al 2003b;Gensini and Ashley 2011;Thompson et al 2012;Sherburn and Parker 2014;Sherburn et al 2016;Coffer et al 2019;Gensini and Bravo de Guenni 2019;Taszarek et al 2020). Seasonal and geographic variability was not examined in this study, however, as performing this for the current study design would lead to significant reduction in sample sizes associated with SIG events, both seasonally and geographically.…”

“…Effective-layer calculations were applied to the supercell composite parameter (SCP) and significant tornado parameter (STP), improving their diagnostic discriminatory skill. Recent studies have indicated that further improvements in the skill of STP as a statistical discriminator can be achieved by using shallower layers for SRH integration bounds (e.g., 0-500 or 0-100 m; Coffer et al 2019Coffer et al , 2020, and, generally, stronger ground-relative winds and more rightwarddeviant storm motions contribute to more favorable conditions for tornadoes (Coniglio and Parker 2020). RUC-2 data were also used to stratify large hail events, where significant class overlap (severe and significant severe classes) was noted for thermodynamic variables (Johnson and Sugden 2014).…”

Modelling deep convective activity using lightning clusters and machine learning

“…Considering the shortcomings of radiosonde sounding data, reanalysis datasets are more and more widely used (e.g., Allen and Karoly, 2014; Brooks, 2013; Coffer et al ., 2020; Diffenbaugh et al ., 2013; Gensini and Brooks, 2018; Li et al ., 2020; Li and Chavas, 2021; Tang et al ., 2019; Taszarek et al ., 2021a, 2021b Taszarek et al ., 2018; Zhou et al ., 2021) due to their continuity in space and time. ERA5 reanalysis has high resolution, and can provide atmospheric homogeneity records with long time series (Czernecki et al ., 2019; Rodríguez and Bech, 2020; Taszarek et al 2021a, 2021b; Ukkonen and Mäkelä, 2019).…”