1 JULY 1997 DERECHO - A RADAR PERSPECTIVE

Eric D. Howieson
NOAA/National Weather Service Tulsa, OK

Gregory A. Tipton*
NOAA/National Weather Service Minneapolis, MN

1. Introduction

During the afternoon and evening of 1 July 1997 an isolated thunderstorm, exhibiting High Precipitation (HP) supercell characteristics, and a squall line merged and became a large bow echo as it moved from west to east across southern Minnesota. Numerous severe weather reports fell in a geographical area that easily met the derecho criteria established by Johns and Hirt (1987). Severe weather reported included downburst winds estimated at 54 m s^-1 (120 mph), 12 tornadoes ranging from F0 to F3, hailstones with diameters approaching 3 cm (2 in.), and numerous occurrences of flash flooding (NCDC 1997). The only element, from the above list, often not associated with the conceptual model of a squall line/bow echo, was the number of tornadoes. The tornadoes will be the focus of this paper and presentation.

Of the 12 tornadoes, 3 developed from a merger between the HP supercell and squall line, 6 from the leading edge/apex of the derecho/bow, and 3 from the bookend vortex. Examples of the 12 tornadoes will be discussed in detail in this manuscript.

2. Tornadogenesis

Detecting nonsupercell tornadoes was a challenge for this particular event for the following reasons. First, tornadogenesis occurred with the non-bookend segment of the squall line/bow echo. Second, tornadogenesis did not take place aloft (0.3 to 1.8 km or 6,000 to 15,000 ft) then descend to the ground [hereafter mode I (Trapp and Davies-Jones 1997)]. Instead, tornadogenesis took place in the lower levels (1.8 to 4.6 km or 1,000 to 6,000 ft) then ascended to mid levels [hereafter mode II (Trapp and Davies-Jones 1997)]. A preliminary study by Trapp and Mitchell (1995) found, that of 16 tornadoes, 50% exhibited

mode II characteristics. One characteristic of mode II tornadogenesis within squall lines is a strong circulation develops in the lower levels, on average, within five minutes of tornado touchdown (Trapp et al. 1998). However, strong circulations for mode I tornadogenesis may appear 30 to 45 minutes before tornado touchdown (Burgess et al. 1993). This case provided an excellent opportunity to investigate reflectivity and velocity data for certain signatures that could aid the WSR-88D operator in detecting mode II tornadogenesis associated with the leading edge of a squall line or apex of a bow echo.

2.1 Pre-Bow Echo Tornadoes

An excellent example of mode I tornadogenesis took place along the leading edge of the squall line near Willmar, MN or about 120 km (65 nm) west of the WSR-88D in Chanhassen, MN (KMPX) between 2232 and 2247 UTC. Tornadogenesis in this case seemed to coincide with the squall line beginning to bow. The 0.5 storm relative motion (SRM) product (Fig. 1) showed convergence, reflecting the proximity of a Front Inflow Jet (FIJ) and a Rear Inflow Jet (RIJ) (both jets referenced to the actual thunderstorm), over Willmar.

As the squall line began to bow, it caught up to and merged with the HP supercell over Wright County Minnesota, about 40 to 75 km (~20 to 40 nm) northwest of KMPX. This merger resulted in three tornadoes, an F1, F2, and F3. The F3 tornado occurred near the town of Buffalo or 43 km (23 nm) northwest of KMPX at 2343 UTC. The 0.5 reflectivity product (Fig. 2) indicated a well defined front inflow notch, with at least a 19 km (10 nm) Bounded Weak Echo Region (BWER), adjacent to the merger location.

The 0.5 SRM product (Fig. 2) identified strong convergence (reflecting the presence of the strong front and rear inflow jets) at a location coincident with the aforementioned front inflow notch. The signatures associated with these three tornadoes were similar in nature to those discussed for the Willmar tornadoes, however, the F3 tornado was more difficult to identify due to the merging of the squall line and HP supercell. Interrogation of the circulation on the 2343 UTC 0.5 SRM product showed 51 m s^-1 (99 kts) of gate-to-gate shear at 0.9 km (3,000 ft). The circulation associated with this F3 tornado was detectable for more than an hour with the HP supercell. The circulation developed into a cyclonic convergence signature between 2307 and 2313 UTC with convergence values in excess of 85 kts for three consecutive volume scans beginning at 2313 UTC. At 2328 UTC, the convergence signature evolved into a cyclonic circulation which first interacted with the strong updraft at 2333 UTC, 10 minutes before tornado touchdown.

2.2 Bow Echo Tornadoes

Between 0013 and 0034 UTC, the thunderstorm complex rapidly evolved into a strong derecho with a well defined bookend vortex on the northern end and an anticyclonically rotating supercell on the southern end. During this transition, the RIJ surged into the of the northern portion of the line, as evident by a large rear inflow notch in the reflectivity imagery (not shown), leading to the development of the first bookend vortex (Fig. 3).

Another example of leading edge tornadogenesis began at 0044 UTC when a circulation became evident on the 0.5 and 1.5 SRM along the leading edge of the bow. The circulation was also coincident with an inflection along the leading edge of the line in the vicinity of a reflectivity notch (Fig. 4).

By 0059 UTC, the circulation appeared to weaken at 0.5 until 0114 UTC when the circulation appeared again. The circulation was evident during the entire period at the 1.5 elevation. At 0115 UTC, an F1 tornado occurred near Forest Lake, Minnesota, located 65 km (35 nm) northeast of KMPX. In addition to the circulations in the velocity data, a reflectivity notch formed on the leading edge of the bow in the vicinity of the circulation from 0044 UTC to 0114 UTC when a well defined hook echo appeared with the circulation. This leading edge reflectivity notch and developing circulation were very similar to the radar signatures associated with the St. Francis and Monticello tornadoes. All three developed on the leading edge of the bow and were intensified by the interaction of the RIJ and the LLJ out in front of the system. The strong updrafts along the leading edge then helped to stretch the circulations to the surface. Przybylinski (1995) noted similar transient tornadoes linked to leading edge shear regions. The St. Francis tornado differed from the others in that tornadogenesis was aided by the cyclonic rotation of the bookend vortex instead of solely by the interaction of the vortex and a strong updraft.

Tornadogenesis along the leading edge or apex of squall lines or bowing systems has been documented by Pryzbylinski (1995) and DeWald et al. (1998). DeWald et al. (1998) found that out of six identifiable circulations along the bowing portion of the line, four produced tornadoes. The circulations were also observed in the lowest levels before increasing in depth and strength. Tornadogenesis in this case appears to be no different than the examples given by the authors mentioned above and similar to the refined model of nonsupercell tornadogenesis presented by Lee and Wilhelmson (1997) (hereafter LW97). LW97 simulated nonsupercell tornadogenesis along a weak outflow boundary and then refined the previous working model developed by Wakimoto and Wilson (1989), Brady and Szoke (1989), and Roberts and Wilson (1995). The main difference between this event and LW97 is that of the environment. The common environment for nonsupercell tornadogenesis, according to LW97, is 1) presence of mesoscale surface boundary with significant across-front shear, 2) misocyclone circulations along boundary, 3) rapidly growing cumulus congestus or young cumulonimbus along a boundary, and 4) weak mid and upper-tropospheric winds.

In this case the leading edge of the squall line/bow echo acted as the surface boundary where misocyclone circulations existed along with rapidly growing updrafts. The major difference between LW97 and this case is that the mid and upper tropospheric winds were moderately strong (25 m s^-1 or 50 kts at 500 mb and 30 m s^-1 or 60 kts at 250 mb). Based on this case and the work of LW97 and their predecessors a conceptual model has been created (Fig. 5), which identifies the apex of the bow echo just north of the rear inflow jet, as being a favored location for mode II tornadogenesis. On a smaller scale, detection of the alignment of the velocity couplet with an updraft (e.g. WER, BWER, 0.5 front inflow notches, etc.) near or just north of the apex of the bow echo seemed to be successful for this event.

3. Conclusion

Detecting tornadoes using the WSR-88D is difficult under the best of circumstances, even when they develop from supercells (i.e. mode I). This detection becomes more difficult when dealing with tornadogenesis associated with a squall line/derecho (i.e. mode II). The conceptual model from this case will hopefully help the WSR-88D operator anticipate mode II tornadogenesis just north of the apex of a bow echo or a bowing segment of a squall line where a velocity couplet coincides with a WER or BWER.

4. Acknowledgments

The authors would like to thank Dan Effertz and Jeff Trapp for their assistance with this paper, and Rich Naistat for relentless reviewing and providing helpful suggestions on how to improve this manuscript.

5. References

Brady, R. H., and E. J. Szoke, 1989: A case study of non-mesocyclone tornado development in northeast Colorado: Similarities to waterspout formation. Mon. Wea. Rev., 117, 843-856.

Burgess, D. W., R. J. Donaldson, Jr., and P. R. Desrochers, 1993: Tornado detection and warning by radar. The Tornado: Its Structure, Dynamics, Prediction, and Hazards. Geophys. Monogr., 79, C. Church, D. Burgess, C. Doswell, and R. Davies-Jones, Amer. Geophys. Union, 203-222.

DeWald, V. L., T. W. Funk, J. D. Kirkpatrick, and Y. Lin, 1998: The 18 May 1995 squall line over south central Kentucky: An examination of complex storm reflectivity trends and multiple mesocyclone development. Preprints, 16^th Conf. On Wea. And Forecasting, Phoenix, AZ, Amer. Meteor. Soc., 148-151.