Figure 1. Box and whiskers plot of mlCAPE. The top and bottom of each shaded box denotes the 75 th and 25 th percentiles, respectively, and the heavy horizontal line is the median value. The heavy vertical lines extend upward to the 90 th percentile, and downward to the 10 th percentile. Groups of supercells are labeled, with sample size in parentheses. P12.3 An Assessment of Supercell and Tornado Forecast Parameters with RUC-2 Model Close Proximity Soundings Richard L. Thompson * , Roger Edwards, and John A. Hart Storm Prediction Center Norman, OK 1. INTRODUCTION In a preliminary investigation, Edwards and Thompson (2000; hereafter ET00) examined multiple sounding parameters related to supercell and tornado potential with a sample of 188 close proximity soundings derived from RUC-2 model hourly analyses. Building upon that initial work, we have expanded our sample to include 548 close proximity soundings associated with supercells, and a smaller set (75) of close proximity soundings for discrete non-supercell storms. A more detailed discussion regarding our sounding collection methodology can be found in Thompson et al (2002, this volume; hereafter T02). Numerous studies have examined proximity sounding characteristics for severe, tornadic, and supercell thunderstorms. Several of the more recent studies, namely Rasmussen and Blanchard (1998; hereafter RB98) and Craven et al (2002; hereafter C02) have collected samples with thousands of individual soundings. While such large samples are desirable, these studies relied on relatively coarse proximity criteria in time and space. The hourly RUC-2 model analyses allowed relatively strict temporal resolution (within 30 minutes) in our proximity soundings, at roughly the spacing of the surface observing network (40 km horizontal resolution). To keep the sample size reasonably large, we considered the majority of discrete radar-identified supercells across the contiguous United States during a 27 month period from April 1999 through June 2001. An adaptation of the original SHARP sounding analysis software (Hart and Korotky 1991) was utilized to calculate all variables. 2. THERMODYNAMIC PARAMETERS From the box and whiskers plot of lowest 100 mb mean parcel (ml) CAPE shown in Fig. 1, it is clear that larger mlCAPE values tend to be associated with significant tornadic supercells, and lesser CAPE with nontornadic supercells and non-supercell storms. The significant tornadic and nontornadic supercells are offset by almost one quartile from the 25 th to 75 th percentiles. These results are similar to the findings of RB98, ET00, and C02. Several recent studies (namely RB98, ET00, and C02) have identified lifting condensation level (LCL) height as an important discriminator between tornadic and nontornadic supercells. Our RUC-2 proximity sounding sample reaffirms these findings, with a substantial offset (more than one quartile) between significant tornadic and nontornadic supercells (Fig. 2). The lower LCL heights of the significant tornadic storms supports the hypothesis of Markowski et al. (2000) that increased low-level relative humidity may contribute to increased buoyancy in the rear flank downdraft, and an increased probability of tornadoes. The substantially higher LCL heights for the non- supercell storms may be misleading in that our sample is probably not representative of the full spectrum of storms. All soundings with a surface pressure greater than 999 mb were excluded because of a truncation error in our sounding analysis software. Thus, most thunderstorms from the moist coastal (low elevation) areas were not included in this sample. * Corresponding author address: Richard L. Thompson, Storm Prediction Center, Norman, OK. E-mail: thompson@spc.noaa.gov