SR/SSD 98-30

7-1-98

Technical Attachment

A Comparison of VIL Density and Wet-Bulb-Zero Height

Associated with Large Hail over North and Central Georgia

Patricia A. Hart and Kent D. Frantz

NWSFO Peachtree City, Georgia

Introduction

Recent studies in Oklahoma and Tennessee have examined the relationship between Vertically Integrated Liquid (VIL) and storm echo top (ET). This study was primarily done in an effort to verify the conclusions of the previous studies (Amburn and Wolf 1996, Troutman and Rose 1997). The airmass over Georgia is heavily influenced by moisture from the Gulf of Mexico and it was hypothesized that the resulting VIL Density, which was shown to be a reliable indicator of large hail in Oklahoma and Tennessee, would also be successful in predicting hail size from severe thunderstorms in Georgia.

Another well known predictor of large hail has been the wet-bulb-zero (WBZ) height (Air Weather Service, Miller 1972). A wet-bulb-zero height in the range of 7,000 to 11,000 ft AGL is favorable for large hail reaching the ground. It was also hypothesized that a relationship between WBZ height and VIL Density would result in an increase of a meteorologist's confidence in determining hail size from severe thunderstorms over Georgia.

Only large hail events (hail diameter 3/4 in or larger) were studied across north and central Georgia during 1995-1996 in the NWSFO Peachtree City (Atlanta) County Warning Area (CWA). Thunderstorms that produced hail smaller than 3/4 in, or no hail at all, were not examined due to the magnitude of data that would result. The VIL Density was computed and compared with the actual wet-bulb-zero height for 165 large hail events. The results show a strong relationship in determining thunderstorm potential for producing large hail.

Vertically Integrated Liquid, VIL Density and Wet-Bulb-Zero Height

The VIL is a function of reflectivity, and converts reflectivity into an equivalent liquid water content value based on studies of drop-size distribution and empirical studies of reflectivity factor and liquid water content (Amburn and Wolf 1996, Troutman and Rose 1997). The general equation for VIL as used with the WSR-88D is:

VIL = E 3.44 x 10^-6 [(Z_i + Z_i+1)/2]^4/7 D h

which has units of (kg/m²). Z_i and Z_i+1 are radar reflectivity at the bottom and top of the layer Dh whose thickness is in meters.

VIL Density is simply the VIL divided by the echo top (in meters), multiplied by 1000 g/kg to express the result in units of g/m³. The importance of the VIL Density is its use in quickly identifying thunderstorms with high reflectivities relative to their height. Such thunderstorms often contain hail cores, and as the VIL Density increases, the hail core tends to be deeper and more intense, resulting in larger hail (Amburn and Wolf 1996, Troutman and Rose 1997).

Wet-bulb temperature is the temperature an air parcel would have if cooled adiabatically to saturation at constant pressure by evaporation of water into the air, all latent heat being supplied by the parcel (Glossary of Meteorology 1959). Wet-bulb-zero height is the height above the surface in which the air parcel wet-bulb temperature becomes 0C. Midwest studies of thunderstorms disclosed that over 90% of reported surface hail occurred where the WBZ height was between 5,000 and 12,000 ft. In those cases in which large hail was reported, the WBZ heights were clustered around an average of 9,000 ft above the surface. When WBZ heights were above 11,000 ft and below 7,000 ft, the frequency and size of hail diminished rapidly (Miller 1972).

Methodology

The WSR-88D radars in Peachtree City (KFFC) and at Robins AFB, Georgia (KJGX) were used for this study due to the large geographical area monitored by the NWSFO. There are 93 counties in the Peachtree City CWA, which covers most of north and central Georgia. The data analyzed for this study were from confirmed large hail events using the usual storm verification data. During the years 1995-1996, 165 severe hail events occurred within the CWA. VILs and echo tops were collected for each of these events. Also, the wet-bulb-zero height was recorded from the Peachtree City 1200 UTC sounding for the day of the event, utilizing the SHARP program (Hart, et al. 1991).

The highest echo top selected was not always located on the same pixel as the maximum VIL pixel, as done in the Troutman and Rose (1997) study. After conferring with Steve Amburn, Science and Operations Officer at NWSO Tulsa, the highest echo top pixel used was located either on the same maximum VIL pixel or within one pixel in any direction. This allowed for possible tilt in the thunderstorm due to upper level wind shear, which was clearly evident in many of the 165 cases examined. The echo top and VIL values used were as precise as possible. For example, if the maximum VIL or echo top detected for the volume scan was determined to be the severe thunderstorm event being analyzed, then those values were used. If the precise VIL or echo top values could not be determined and a range was indicated, then an average was used. A VIL in the range of 50-54 kg/m² was recorded as 52 kg/m², and an echo top in the range of 45,000 to 49,000 ft was recorded as 47,000 ft.

While all 24 months of 1995-1996 were analyzed, it was interesting to note that the large hail events occurred from January through the beginning of September in each year. No severe hail events were reported during October, November or December. However, this is not to imply that large hail may not occur from October through December; more likely it is an indication of the limited data studied. The WSR-88D data used were from both Volume Coverage Patterns (VCP) 11 and 21. While VCP 21 samples five fewer elevation slices than VCP 11, gaps in the scan strategy occur mainly within 50 nm of the radar site. Beyond 50 nm the elevation slices are the same below 30,000 ft, but not above (Federal Meteorological Handbook, No. 11, Part C 1991).

Results

Amburn and Wolf (1996) and Troutman and Rose (1997) have demonstrated the relationship between VIL and echo top could be useful in determining the potential of a thunderstorm for producing large hail. We hypothesized the VIL Density would also be successful in Georgia. However, it was also hypothesized the corresponding WBZ height would be a very useful parameter that could be compared with the VIL Density to obtain hail size. The results of this study in fact show a strong relationship between VIL Density and hail size. Also, VIL Density in combination with the WBZ height can give the meteorologist a high level of confidence in predicting large hail.

Figure 1 shows a scatter plot of VIL vs. echo top for the 165 severe hail events studied. Calculated VIL Densities of 3.0, 3.5, 4.0 and 4.5 g/m³ are shown by straight lines. These VIL Density lines were chosen to correspond with the graph (Fig. 2) derived by Amburn and Wolf (1996), to facilitate a comparison of Tulsa and Georgia data. Only 9% (15 of 165) of the severe hail events occurred with a VIL Density value less than 3.0 g/m³. Very few of these 15 cases were larger than 3/4 in hail. Another 12% (19 of 165) of the severe hail events occurred between the VIL Density values of 3.0 to 3.5 g/m³. The remaining 79% (131 of 165) of the large hail cases studied had a VIL Density of 3.5 g/m³ or larger. See Table 1 for specific numbers of events, by category. A VIL and echo top falling to the right of the 3.5 line would indicate severe hail likely due to a large VIL Density. (Although it must be kept in mind that we did not analyze VIL Density for cases in which large hail was not reported.) Our Georgia data are similar to results derived from the Oklahoma study, indicating a strong positive correlation between VIL and echo top in cases of large hail.

As indicated by the earlier studies, VIL Density when compared with hail size provides useful information for the meteorologist. A scatter plot of hail size vs. VIL Density is shown in Fig. 3. For the 165 large hail cases studied, 79% had a VIL Density of 3.5 g/m³ or larger. The Amburn and Wolf (1996) study indicated a VIL Density of 3.5 g/m³ as correctly identifying 91% of the severe hail cases in Oklahoma, while Troutman and Rose (1997) also found 79% of their cases in middle Tennessee to produce large hail when the VIL Density was at least 3.5 g/m³. Figure 4 shows the best-fit straight line for the plotted data. Although the figure shows a slight upward trend, the greater the VIL Density the more likely the hail size will increase, the data in this study appears to be inconclusive. Of the 83 severe hail events that had a VIL Density value of 4.0 g/m³ or greater, only 46% (38 of 83) produced one inch hail or larger (see Data Table). It is hoped a more definite trend will be indicated after data from 1997 and 1998 are included in a future expanded study.

Figure 5 shows a third comparison between VIL Density and WBZ height. It is notable that 67% (111 of 165) of the large hail cases studied had WBZ heights between 7000 and 11,000 ft. This would likely confirm the Miller (1972) Midwest study is appropriate for indicating the potential for large hail in Georgia as well. It is rare for hail of 2 in or greater to occur in Georgia. However, of the 4 cases that did occur out of the 165 cases examined, 75% (3 or 4) were found in this prime WBZ height zone. The largest severe hail normally reported in Georgia is the size of golfballs (1.75 in). Of these cases, 63% (19 of 30) were in the prime WBZ height zone as well. The comparison between VIL Density and WBZ height shows a slightly higher concentration of plots, 52% (85 of 175) in this prime WBZ height zone with VIL Densities of 3.5 to 5.0 g/m³. The best-fit straight line for the plotted data seems to show a slight inverse relationship trend in which hail producing thunderstorms developing in lower WBZ height environments tend to have larger VIL Densities. However, all four of the highest VIL Density plots lie above the best-fit straight line. One explanation could be greater weight was given to the plots below the WBZ height of 7000 ft. This will be explained further.

Figure 6 is a scatter plot of WBZ height compared with echo top for the 165 cases and it also shows the best-fit straight line for the plotted data. The line shows a trend as the echo top increases so does the WBZ height increase. This would seem to be typical of thunderstorm development. However, there were many cases where the WBZ height remained low (6,500 to 8,500 ft) and there are a few reasons for this. Shallow or low top thunderstorms occur during a variety of atmospheric conditions. Some of these include, but are not limited to, a low seasonal tropopause, weak or non-divergence aloft, weak vertical motion, marginally unstable environment, or climatologically colder atmosphere during the late fall through early spring seasons. All of these would tend to suppress or cap thunderstorm growth (Djuric 1994).

The colder atmosphere that climatologically occurs from December through March over the United States would affect the WBZ height the most by lowering it toward the surface until low-level heat, convergence and moisture could stabilize the downward trend. Thunderstorms developing in this kind of environment would tend to have low echo tops (18,000 to 30,000 ft) and a lower WBZ height (6,500 to 8,500 ft). The VIL value in low echo top thunderstorms can become high as abundant low-level moisture spreads upward uniformly in the thunderstorm and forms highly reflective hailstones above the WBZ height. As shown in Figs. 1 and 2, a high VIL compared with a low echo top will produce a large VIL Density.

Another parameter that would affect the WBZ height is dryer air at the mid-levels (7,000 to 11,000 ft). This would lower the wet-bulb temperature as the dew-point depression increases. This 700 mb dry intrusion has been an indicator for the classic conceptual model for severe thunderstorm forecasting. This appears to be an essential ingredient for any significant outbreak of severe thunderstorms (Miller 1972). Under this condition large hail would be likely as the WBZ height lowers.

While low echo top thunderstorms during the cooler seasons and/or 700 mb dry intrusion can explain a lower WBZ height, Fig. 6 still shows a trend; as the WBZ height decreases the VIL Density increases. Is this a coincidence of the limited cases studied, radar scan strategy, or is this a definite trend? While there seems to be an association with VIL Density and WBZ height in determining hail size, further research will be needed.

Conclusions

While researching this study there were a few inconsistencies noted that need to be examined. The first was the 34 cases in which the VIL Density ranged from 2.2 to 3.4 g/m³. Hail size in 76% of these cases was 3/4 in, but there were three cases of golfball-size hail or larger. (See Table 1.) A possible reason for this is that the VIL is computed by integrating reflectivity for each vertical column of pixels. For slow moving, vertical thunderstorms VIL calculations should be quite accurate. However, for fast moving or strongly tilted thunderstorms, VIL calculations will not be as accurate (Amburn and Wolf 1996).

Echo tops can also vary significantly with changes in radar range, even though the thunderstorm may not vary in height. The WSR-88D uses a finite number of elevation angles in the scan strategy and this tends to truncate the actual echo top for the thunderstorm (Federal Meteorological Handbook 11, Part C). Another reason these thunderstorms could produce large hail, even though the VIL Density was less than 3.5 g/m³, is that 79% of the 34 cases occurred with the WBZ height in the prime WBZ height zone. Once again, the meteorologist needs to know the WBZ height for the event. The lower the WBZ height, the greater the potential for large hail to occur, even though the VIL Density of 3.5 g/m³ has not been reached.

This study, as in previous studies, continues to show a strong relationship between VIL and echo top in thunderstorms which produce large hail. An association between hail size and VIL Density appears to exist. Also, when a VIL Density of at least 3.5 g/m³ is compared with the WBZ height for the day of the event, and is found within the prime WBZ height zone, then it is likely severe hail will occur. Determining the VIL Density for individual thunderstorms requires some extra time from the meteorologist, but these relationships can be used successfully in issuing severe thunderstorm warnings.

Acknowledgments

The authors would like to thank Gary Beeley (SOO) and Carlos Garza (MIC), NWSFO Peachtree City, and Steve Amburn (SOO), NWSO Tulsa, for their helpful comments and suggestions while proofreading this study.

References

Amburn, S., and P. Wolf, 1996: VIL Density as a Hail Indicator. 18th Conference on Severe Local Storms. San Francisco, CA, American Meteorological Society, 581-585.

Troutman, T. and M. Rose, 1997: An Examination of VIL and Echo Top Associated with Large Hail in Middle Tennessee. Technical Attachment, SR/SSD 97-15, 4-1-97.

Miller, R., 1972: Notes on Analysis and Severe-Storm Forecasting Procedures of the Air Force Global Weather Central. Air Weather Service Technical Report 200 (Rev), United States Air Force, Chapters 5 and 7.

Djuric, D., 1994: Weather Analysis. Chapter 12. Prentice-Hall, Inc. Englewood Cliffs, New Jersey.

Hart, J.A., J. Korotky and G. Jackson, 1991: The Sharp Workstation V (5/97a) Skew-T/Hodograph Analysis and Research Program for the IBM and compatible PC. Users Manual. NWSFO Charleston, WV, Appendix IV-14.

Federal Meteorological Handbook 11, 1991: Doppler Radar Meteorological Observations, Part C. Office of the Federal Coordinator for Meteorological Services and Supporting Research, U.S. Department of Commerce/NOAA, Washington, D. C.

Glossary of Meteorology, American Meteorological Society, Boston, Massachusetts, 1959.