Close Window

Saint Louis University: "Case studies of tornadoes associated with tropical cyclones in the southern United States based on conventional and WSR-88D data"

Final Report

1.1 Objectives

The primary objective has been, to compare the commonly known climatologically inferred key factors responsible for tropical cyclone (TC) tornadoes with those occurring in Earl (1998); using both conventional and WSR-88D radar data. Forecasting the typically shortlived and shallow TC tornadoes is a matter of concern to National Weather Service Forecasters (Edwards 1999).

The commonly known key factors are:

In the second instance of tornadoes near Charleston, South Carolina; the observed CAPE lied within climatology but the observed SREH exceeded significantly the climatology. A nearly saturated atmosphere produced only small values of CAPE, and because storm cells moved against the low-level winds near Charleston, South Carolina SREH value was high. In the above, CAPE was calculated using the conventional radiosonde data, while SREH was calculated using the conventional wind data or the Velocity-Azimuthy-Display (VAD) data

In earlier studies, the climatologically preferred quadrant for tornado occurrence was the right forward quadrant one. In our case with Earl, this quadrant also produced tornadoes. For a track of Earl and the related tornadoes, see Edwards et al. (2000).

1.2 Accomplishments

An accomplishment of this research was the verification through the Earl radar data, that in addition to CAPE and SREH low-level baroclinicity was responsible for tornadoes. There were two separate volumes of low-level airmasses forming boundaries with the tropical cyclone circulation in the Tampa and Charleston areas, differing in temperature from the low-level tropical circulation by a degree or two Celsius. Consequently, mesoscale baroclinicity and vertical shears were produced. Doppler radar detected tornadoes in the vicinity of these boundaries. Identifying such air mass boundaries and showing that tornadoes occurred at those boundaries through the WSR-88D radar loops is regarded as an accomplishment of this research. Prior research (Vescio et al. 1996) has indicated such a possibility, though none has shown such occurrences through animated observed Doppler radar.

A second accomplishment has been the finding, that the actual storm relative helicities (SREH) were somewhat less than the default SREHs. The default SREHs were derived on the premise that a mesocyclone moves 30 degrees to the right of the mean (0-6 km) ambient wind with a speed equal to 75% of the mean wind. If a sufficient amount of data on the observed and default storm motions accumulate, the current default motions can be revised (for TC studies) so that realistic estimates of SREH can be made routinely.

A third accomplishment has been a clear documentation of a downburst in a TC. Previous tropical studies have confirmed mostly mesocyclones and tornadoes, but not downburst. Stewart and Robbins (2000) mentioned a downburst with a land falling hurricane Irene (1999). However, no structure of the downburst was presented. Our downburst near Charleston, South Carolina was about 8 km in horizontal extent and nearly 3 km in the vertical. Because of the 60-km distance to the WSR-88D, we could detect the downburst only from a base height of 700 meters. The air constituting the downburst had a westerly component and was sandwiched between two slabs of air with easterly components. Since the low-level flow in the vicinity of downburst has an easterly component, it is likely that the downburst air had originated from the middle levels, e.g., 700 mb. The Eta model 700-mb wind for that time supported this hypothesis. It was possible to estimate the duration of the downburst also. For example, the radar imagery was available at 5-minute intervals. The downburst was detected only on one Plan Position Indicator (PPI) and not on the one preceding or following it. Thus, the downburst lasted less than 10 minutes.

A fourth accomplishment has been a compilation of the mean characteristics of the mesocyclones that bore tornadoes at some stage or other. Table 1 contains the data from the Tampa area while Table 2 from the Charleston area.


Showing the mean characteristics of mesocyclones gathered by the Tampa Bay WSR-88D Doppler radar. Sample size 15 with A (elevation in degree), B (range in km), C (echo top in km), D (maximum reflectivity in dBZ), E (height of max reflectivity in km), F (rotational velocity in m s -1), and G (diameter in km).

0.5 69.6 10.8 57.0 4.8 13.8 5.6
1.5 69.6 10.8 57.0 4.8 15.0 5.4


Showing the mean characteristics of mesocyclones gathered by the Charleston WSR-88D Doppler radar. Sample size 15 with A (elevation in degree), B (range in km), C (echo top in km), D (maximum reflectivity in dBZ), E (height of max reflectivity in km), F (rotational velocity in m s -1), and G (diameter in km).

0.5 53.6 12.0 59.7 6.3 17.3 5.9
1.5 53.6 12.0 59.7 6.3 20.5 6.0

Tables 1 and 2 suggest that the rotational velocity is more intense at a higher elevation in both locations. Also, the mesocyclones in the Charleston area are deeper and stronger than those in the Tampa area. One explanation is that the Charleston area is closer by about 200 km to the TC center than the Tampa area. Besides, the large-scale ascent obtained from the Eta model and the low-level vertical shears are more intense in the Charleston region than the corresponding ones in the Tampa region.

A fifth accomplishment has been the recognition that most of the tornadoes associated with Earl had been with supercells that are combined and nonisolated. We found two tornadoes that were generated from isolated supercells in Beaufort and Colleton Counties, and in Bufford County near Charleston, South Carolina.

A sixth accomplishment has been the ability to attract undergraduates to do TC tornado research. Four undergraduate students took a total of eleven (11) hours research credit toward this problem. They identified hook echo, mesocyclones, and a bounded weak echo region with great interest.

1.3 Experiences Gained

It is advisable to provide a month to two months for acquiring radar data on 8 mm tapes.
A WATAD (WSR-88D Algorithm Testing and Display System, NSSL 2000) user should be aware that the T) C related shallow mesocyclones and the accompanying tornadoes may be missed by the tornado algorithm, because of their short duration or shallowness. The investigator should examine in detail the imagery for mesocyclones, tornadoes, and downbursts.
While using the WATADS (version 9.0 or later) reliability dictates the Vertically Integrated Liquid Water (VIL) estimates be based on the storm cell rather than on the grid volume.

1.4 Software Developed

To better our estimates of SREH, we calculated the observed mesocyclone movement based on radar imagery. We plotted the coordinates (latitude and longitude) of the mesocyclone as a function of time and calculated the storm motion. These storm velocities were given to our collaborator, Roger Edwards, who calculated the observed storm relative helicities. The default values were calculated using the Hart and Korotky (1991) algorithm.


The SPC forecaster, Roger Edwards, visited Saint Louis University, September 9-12, 1999 and talked to Professor G. V. Rao about the Eta model products relating to the TC Earl (1998). He mentioned that although a tropical cyclone consists of a homogeneous sample of air, its circulation after land falling, invariably interacts either with the topographical features; or with the existing or traveling (middle latitude) weather systems. In some instances, dry air in the middle levels intrudes into the right forward quadrant of the TC causing convective instability and tornadoes. In other instances, the TC interacts with an existing low-level baroclinic zone (Vescio et al. 1996) and produces tornadoes at the boundary.

In the spring semester, Professor Rao and Joshua Scheck, a graduate student, and four other undergraduate students examined the radar data pertaining to Earl (1998) in detail. Tornadic mesocyclones were identified and their trajectories plotted. The storm motions were computed. Several kinematic quantities based on radar data were derived. Some of these were shown in Tables 1 and 2. From these data other quantities such as the normal component of vorticity, horizontal shear, and lifetimes of mesocyclones, etc., were estimated. Besides being vitally important to the National Weather Service forecasters, they are of basic significance to the scientists who wish to simulate or develop theories of the TC tornadic mesocyclones.

Throughout this phase of intense investigation in the spring semester, the university researchers and the SPC forecaster were in constant contact with each other. While preparing a preprint article (see Rao et al. 2000) for presentation at the Hurricane Conference, there was a regular exchange of ideas and preprint drafts. This preprint article is enclosed.

Frequent transfers of maps, radar PPI imageries, and tornadic mesocyclone trajectories and their velocities of propagation took place in May and June; while a preprint article (Edwards et al. 2000) was prepared for presentation at the Severe Local Storms Conference. This preprint article is enclosed.

Professor G. V. Rao and Joshua Scheck visited the SPC during June 25-29, 2000. They continued their discussions with the SPC forecaster, Roger Edwards, and the Director of SPC, Joseph Schaefer. Professor Rao presented a seminar at the SPC and NSSL on June 27, 2000. Valuable discussions took place after the seminar with the SPC and NSSL scientists on topics such as: VIL estimates, vortex couplets, and downbursts. G. V. Rao and Josh Scheck saw how hand analysis of the wind field takes place in the SPC Forecast Center and the various steps leading to the issuance of Severe Weather Watches in real time. Josh performed a hand analysis of a temperature field in real time while watching the different weather alerts.


3.1 Department Presentation

The SPC forecaster, Roger Edwards, presented a talk entitled, May 3, 1999 Oklahoma City Tornado, at Saint Louis University, Department of Earth and Atmospheric Sciences on September 10, 1999.

3.2 University Presentation

Joshua Scheck participated in the Annual Graduate Student Research Day on April 7, 2000 and presented a poster entitled, Radar Case Studies of Tornadoes Associated with Earl (1998). There were about fifty (50) similar posters by other graduate students.

3.3 Regional Presentation

Professor Rao presented a talk entitled, Case Studies of Tornadoes Associated with Earl (1998), at the Annual Meeting of the Missouri Academy of Science in Columbia, Missouri on April 15, 2000.

3.4 National Presentation

Professor Rao presented a talk on, Case Studies of Tornadoes Associated with Tropical Cyclones Based on Conventional and WSR-88D Data, at the 24th Conference on Hurricanes and Tropical Meteorology, May 29-June 2, 2000 in Fort Lauderdale, Florida.

3.5 Storm Prediction Center and National Severe Storm Laboratory Presentation

Professor Rao presented a talk on Some Studies of Tornadoes Associated with TC Earl (1998), on June 27, 2000.

3.6 National Presentation

The SPC forecaster, Roger Edwards, will present a talk on, Examination of Tornadic Supercells in Tropical Cyclone Earl (1998) using Conventional and WSR-88D Data Suites, at the 20th Conference on Severe and Local Storms, September 11 - 15, 2000 in Orlando, Florida.

3.7 Publication

A publication encompassing the results of Earl (1998) and Frances (1998) is being prepared by G.V. Rao, R. Edwards, and J. Scheck for submission either to Monthly Weather Review or Weather and Forecasting in December 2000.


4.1 Benefits to the University

Professor Rao had profited immensely with this project. He gained knowledge specifically in the areas of rainband formation and growth of convection in the outskirts of a tropical cyclone leading to its movement. His knowledge of hurricane convection has improved as a result of examining the radar imagery of mesocyclones; and radar loops of the life cycle of mesocyclones. He wishes to compare the tornadic mesocyclonic CAPE and SREH values obtained in this project with those computed in numerical models (such as in the NOAA-GFDL, Princeton model) dealing with severe convection in tropical cyclones. His concepts in radar meteorology have become clearer after examining the imagery of weak echo region, hook echo, and vortex couplet, etc. Professor Rao teaches regularly Tropical Meteorology and Convection and wishes to add some of this research material to those courses.

The graduate students of Saint Louis University enjoyed the presentation of the SPC meteorologist, Roger Edwards. They were treated to a firsthand account of the deadly May 3, 1999 Oklahoma City tornado. Joshua Scheck enjoyed his visit to SPC and NSSL, and understood the rigor of real-time forecasting.

4.2 Benefits to SPC

Because of the frontal interaction observed with TC Earl in Florida, awareness and understanding have improved regarding the role of subtle baroclinic zones in the production of tornadoes in land falling TCs. Our findings indicate that known midlatitude roles of mesoscale baroclinic boundaries in tornado production (e.g., Markowski et al. 1998) may be analogous to interactions of TC supercells with pre-existing fronts in the vicinity of TC landfalls. We can, therefore, narrow the spatial focus of a tornado watch in such a situation to the favorable corridor along and leeward of the frontal location, in the climatologically favored quadrant of the cyclone. Smaller, more precise watches reduce the dead area in a watch (where no events occur), thereby providing a better public forecast service.
Our findings regarding the use of default versus observed storm motions in evaluating SREH, reinforce the operational necessity of incorporating observed cell motion into any storm-scale nowcast (e.g., NWS warnings and nowcasts) and mesoscale forecast (e.g., SPC mesoscale discussion or watch using cell motion).
Affirming the findings of earlier studies (e.g., Spratt et al. 1997), our observations on the longevity and trajectories of outerband supercells may lead to more effective SPC mesoscale discussions, describing the most probable strike areas and duration of lifetime for offshore TC supercells, further narrowing the greatest tornado threat within a watch. This should best benefit the mesoscale forecaster, writing discussion updates within a watch and preparing watch clearance lines that denote the end of a severe threat.
The potential usefulness of vertical motion fields in the ETA model, combined with real-time analysis of boundaries in the imminent landfall region, should help the SPC outlook forecaster to narrow the spatial extent of tornado threat in such situations.


Davies-Jones, R. P., D. W. Burgess, and M. Foster, 1990: Test of helicity as a tornado forecast parameter. Preprints, Conf. on Severe Local Storms, Kananaskis Park, AB, Canada, Amer. Meteo. Soc., 588-592.

Edwards, R., 1999: Storm Prediction Center forecast support for land falling tropical cyclones. Preprints, 23rd Conf. on Hurricanes and Tropical Meteorology, Dallas, Amer. Meteor. Soc., 53-56.

G. V. Rao, and J. Scheck, 2000: Examination of tornadic supercells in Tropical Cyclone Earl (1998) using conventional and WSR-88D suites. Preprints, 20th Conf on Severe Local Storms, Orlando, FL, Amer. Met. Soc.

Hart, J. A., and W. C. Korotky, 1991: The SHARP Workstation VI-50, a skew-T hodograph analysis and research program for the IBM compatible PC. NOAA Users Manual, 62 pp. [Available from NWS Forecast Office, Charleston, WV].

Markowski, P. M., E. N. Rasmussen, and J. M. Straka, 1998: The occurrence of tornadoes in supercells interacting with boundaries during VORTEX-95. Wea. Forecasting, 13, 852-859.

McCaul, E. W., Jr., 1991: Buoyancy and shear characteristics of hurricane-tornado environments. Mon. Wea Rev., 119,1954-1978.

NSSL, 2000: WATADS (WSR-88D Algorithm Testing and Display System) 2000: Reference guide for version 10.2. [Available from Storm Scale Applications Division, National Severe Storms Laboratory, 1313 Halley Circle, Norman, OK 73069.]

Rao, G. V., R. Edwards, and J. Scheck, 2000: Case studies of tornadoes associated with tropical cyclones based on conventional and WSR-88D data. Preprints, 24th Conf. on Hurricanes and Tropical Meteorology, Ft. Lauderdale, Fl, Amer. Meteor. Soc.

Spratt, S. M., D. W. Sharp, P. Welsh, A. C. Sandrik, F. Alsheimer, and C. Paxton, 1997: A WSR-88D assessment of tropical cyclone outer rainband tornadoes. Wea. Forecasting, 12,479-501.

Stewart, S. R., and C. C. Robbins, 2000: Mesocyclone induced downburst associated with the landfall of Hurricane Irene (1999) over South Florida. Preprints, 24th Conf on Hurricanes and Tropical Meteorology, Ft Lauderdale, IL, Amer, Met. Soc., 304-305.

Vescio, M. D., S. J. Weiss, and F. P. Ostby, 1996: Tornadoes associated with Tropical Storm Beryl. Natl. Wea. Dig., 21,2-10.