Date of Award

January 2013

Document Type


Degree Name

Master of Science (MS)


Atmospheric Sciences

First Advisor

Matthew S. Gilmore


The first half of this two-part study explores two ways of producing composite environmental soundings (feature averaging versus height averaging; FA vs. HA), why those composites differ from one another, how the compositing technique itself affects the resulting thermodynamic and wind parameters, and which technique results in preserving features. This exploration was applied to three groups of supercell proximity soundings: low-precipitation (LP), classic (CL), and high precipitation (HP) and the HA analysis from the Rasmusssen and Straka (RS98) paper are reanalyzed in both the FA and HA framework. The second half of this study investigates how well previously reported LP, CL, and HP supercell radar behavior (Beatty et al. 2009) is reproduced in an idealized three-dimensional cloud model using both the original and composite soundings.

Reanalyzing the results from RS98 in both HA and FA frameworks, the LP group of soundings have a mean mixed-layer LCL (MLLCL) and mean MLLFC that are both significantly different (p < 0.05) than those from the other sounding groups. Also, the HP group of soundings has a mean MLLFC that is significantly different (p < 0.05) than the means from the other sounding groups. The HP sounding mean BL to 9 km shear and mean 4-10 km shear magnitude are significantly different (p < 0.05; RS98 found

p < 0.02) and the mean HP sounding 9-10 km storm relative wind is significantly different (p < 0.02) compared to the other sounding groups.

Wind parameters and thermodynamic parameters computed from surface-based parcels for both the FA and HA composite sounding lay within one standard deviation of the distribution mean for each sounding group and mixed-layer parcel parameters lay farther from the distribution mean. The FA soundings parameters are not consistently closer to distribution means despite features such as the capping inversion and low-level moisture being preserved better within the FA sounding. Using relative humidity for the LP and CL FA and HA soundings (and vapor pressure for the HP soundings) produces the largest CAPE and least CIN, although averaging water vapor mixing ratio is arguably the most accurate and appropriate.

From the dataset, 29 individual sounding cases were simulated--10 CL, 10 LP, and 9 HP supercells-- and only three storms in each class lasted at least 7200 seconds with an updraft helicity greater than 480 m2 s-2. Only two of these nine individual cases produced long lived supercells, one each from the LP and HP sounding classes, transitioned from a forward flank dominant to rear flank dominant maximum precipitation (following Beatty et al. 2009). The other seven cases maintained a forward flank dominant maximum precipitation. Compositing using only the three successful cases in each class only succeeded in producing long-lived supercells only for CL FA and HA composites and the HP HA composite. These cases produced forward flank dominant precipitation maximums, with no transition. Due to the lack of consistency in storm behavior within each class, it is concluded that cases should be simulated and studied individually, as compared to creating a composite sounding - particularly when studying environments with a very small sample size.