Todd E. Holsten
National Weather Service Forecast Office
North Webster, Indiana



Eric A. Helgeson
National Weather Service Forecast Office
Rapid City, South Dakota



An outbreak of severe weather occurred across Eastern Iowa during the late evening and early morning hours of May 9-10, 1996. A mesoscale convective system developed during the evening hours and evolved into a severe bow echo squall line that tracked for 90 miles across the Davenport, Iowa (KDVN) county warning area. As the bow echo moved across Muscatine, Cedar, Scott, and Clinton Counties, a 1-3 mile wide path of intense damaging wind was observed (Figure 1). $3.3 million of damage was reported (Storm Data, May 1996).


Figure 1
Figure 1. May 10, 1996 severe weather reports.


The strongest winds (estimated at 45+ m s-1) caused extensive damage to farmsteads in southeast Cedar, central Scott and southeast Clinton as well as parts of neighboring counties. Most of the farmsteads in these areas lost nearly all of their outbuildings with many homes suffering moderate (F1) structural damage. An F1 tornado caused extensive damage to several homes and a cemetery in Clinton, Iowa.

Calculations estimating the strength and depth of the cold pool in relationship to the ambient shear will be shown which will offer insight into why this storm was so severe. The baroclinic generation of horizontal vorticity on the leading edge of the cold pool contributed to rapid and intense low level mesocyclone development of three distinct cyclonic circulations and the eventual creation of a long-lived intense downburst. Radar data from the KDVN WSR-88D illustrates the sequence of events that led to the downburst and brief F1 tornado. Both Archive level IV and level II data in conjunction with the WSR-88D Algorithm Testing and Display System (WATADS) data were used.


At 0000 UTC 10 May 1996 broad southwest upper flow dominated the central United States with a weak mid-level shortwave at 500 mb noted over Michigan and a stronger system over Eastern Nebraska (Figure 2). A 65 m s-1 250 mb jet streak extended from southeast Canada to the northern Plains. Upper level divergence was enhanced under the right entrance region of the jet streak. Under this enhanced upper level divergence, the lower troposphere responded with a 16 m s-1 low level jet from Texas to northern Missouri focusing 850 mb moisture convergence across the southeast half of Iowa (Figure 3). At 0000 UTC convection increased across Missouri and Iowa.


Figure 2
Figure 2. 0000 UTC 10 May 1996 500 mb heights (dm), vorticity (s-5 ), and wind barbs (m s-1).



Figure 3
Figure 3. 0000 UTC 10 May 1996 250 mb wind barbs and isotachs (m s-1), 850 mb wind barbs and isotachs (m s-1), and 850 mb moisture convergence.


The 0000 UTC 10 May 1996 KDVN sounding was not conducive to surface based convection, but was unstable with respect to parcels starting ascent around 850 mb. 1182 m2 s-2 of convective available potential energy (CAPE) is noted at 0000 UTC with a slight increase to 1334 m2 s-2 modified for conditions at 0700 UTC (Figure 4).


Figure 4
Figure 4. 0700 UTC 10 May 1996 modified KDVN sounding




A. Storm Trends

Figure 5 shows the position of the developing bow echo in the tall echo phase (Fujita 1978) at 0616 UTC. Archive II reflectivity data (not shown) indicated a storm top of 14.0 km. As this particular storm cell collapsed, it produced an intense downburst near Washington, Iowa causing part of the convective line to bow. The recorded wind gust was 41 m s-1. New cells are generated when the horizontal vorticity generated by the buoyancy gradient at the edge of the cold pool is matched by the opposing horizontal vorticity inherent in the ambient low-level vertical wind shear (Rotunno 1988). Strengthening convection was noted to the north and northeast of the storm collapse.


Figure 5
Figure 5. 0616 UTC 0.5° KDVN base reflectivity. The white arrow denotes the location of a bowing line segment and the initiation point of Storm A.


Equation 1


Figure 6a
Figure 6a. 0636 UTC KDVN cross section of storm relative velocity data normal to the squall line. Note the deep layer shear ahead (right) of the deep convergence zone associated with the leading edge of the cold pool. Figure 11 shows the cross section path as a white line.



Figure 6b
Figure 6b. 0636 UTC KDVN cross section of base velocity through the mesocyclone as seen in Figure 11. Note the descending rear inflow jet present from left to right.


Equation 2


Figure 7
Figure 7. 0636 UTC KDVN cross section of base reflectivity through the squall line. The severe storm is located at the far left. Note the more upright orientation and extent of this storm's reflectivity compared to the other non-severe thunderstorms in the figure.


Three main cyclonic circulations were observed through the life time of this severe bow echo (Figure 8) and coincided with the observed wind damage paths. As the bowing segment moved and accelerated east-northeast, the leading edge of the cold pool generated numerous transient low level circulations which were best seen on 0.13 nm resolution velocity data via WATADS. Each transient circulation is not shown (Figure 8), but on average they originated 1.6 - 4.8 km south of the three main cyclonic circulations that developed along the apex of the bow. The latter two cyclonic circulations (Storms B and C) which were observed to develop, began as one of these transient vortices. In addition, several of the observed or reported tornadoes were traced back to approximately the same time these transient vortices were observed (via radar data) to merge with one of the three main storm circulations (Figure 8).


Figure 8. Map denoting the three main mesocyclonic circulations (large X's) that were observed within the bowing segment of this squall line.


Several low reflectivity notches were noted to the rear of the bow echo across extreme western Scott County, Iowa at 0646 UTC (not shown), which would suggest a rear inflow jet was present. However there remains some question on whether low reflectivity notches to the rear of a squall line in themselves denote a rear inflow jet (Lemon 1999). However, upon further examination of the velocity data, especially the velocity cross section shown in Figure 6b, a rear inflow jet is indeed quite apparent. This led to an acceleration of the southern end of the convective line with several transient vortices observed to develop along the apex of the bowing segment similar to a storm described by Burgess and Smull (1990). The strongest of these vortices (eventually the Storm B cyclonic circulation) went on to become the next dominant and most damaging circulation.

Vr shear [(Vin+Vout)/2] diagrams were constructed for each main circulation and are shown (Figures 9a-c). Storm A (Figure 9a), which was embedded within the line, developed it's initial circulation at 0601 UTC from the ground up, and remained intact until it dissipated as Storm B initiated at 0621 UTC. As Storm B, which initially had developed as a transient vortex along the apex of the bow, migrated from the apex of the bow into the best inflow region of the system, its circulation rapidly increased, reaching a maximum of 28 m s-1 of shear at 0706 UTC (Figure 9b).

Figure 10 shows the storm relative velocity data at 0646 UTC, which indicates a very pronounced cyclonic circulation in the center of the image. Much of the bow echo's damage is attributed to Storm B's downburst, which was particularly intense. This intensity was likely due to its strong pressure perturbation as the low-level circulation rapidly spun-up (Figure 9b). This circulation went on to evolve into the classic comma echo phase ( Fujita 1978) of the system as seen in the reflectivity data (Figure 11) by 0716 UTC.


Figure 9a
Figure 9a. Storm A rotational velocity (Vr, knots) time-height cross section.


Figure 9b
Figure 9b. Storm B rotational velocity (Vr, knots) time-height cross section.


Figure 9c
Figure 9c. Storm C rotational velocity (Vr, knots) time-height cross section.



Figure 10
Figure 10. 0646 UTC KDVN 0.5° storm relative velocity. Note the pronounced mesocyclone in the center of the image. Solid white line indicates orientation of the cross sections used in Figures 6a and 6b.



Figure 11
Figure 11. 0716 UTC KDVN 0.5° base reflectivity. Note the classic comma echo pattern just northeast of the radar site.


Lastly, Storm C developed near a triple point as the comma echo became occluded at 0711 UTC. Radar data showed Storm C's circulation originated as a transient vortex similar to what was observed with Storm B across eastern Scott County Iowa. This vortex rapidly migrated northward into the triple point where the leading edge of the cold pool outflow and the inflow into the "comma head" intersected and rapidly intensified (Figure 9c). Stretching of this low-level vortex by the inflow updraft likely led to the F1 tornado in Clinton, Iowa. As Circulation B slowly weakened, Circulation C became dominant and continued to produce wind damage across northwest Illinois over the next half hour.

All three cyclonic circulations spun-up from low levels to mid levels. Storm B did have weak rotation around 4 km AGL at 0621 UTC, but at 0631 UTC the low level circulation rapidly spun-up near the ground and then was vertically stretched to 4 km AGL. It is important to note that both Storm B and Storm C began as transient vortices that were generated along the spreading cold pool at the apex of the bow echo. Circulation B evolved into the comma head, while Circulation C produced a tornado before it became the last dominant circulation. The bow echo finally weakened across northern Illinois later that morning. In addition to those discussed herein, an additional ten transient vortices were observed to develop along the apex of the bow echo and migrate into the cyclonic inflow side of the bow echo.


Even though previous convection during the evening of May 9, 1996 stabilized the boundary layer, strong low level theta-e advection on the nose of a 25 m s-1 850 mb low level jet combined with increasing deep layer shear under the favorable right rear quad of a 60 m s-1 250 mb jet max , combined to produce elevated convection. Parcels lifted from the top of the inversion present within the boundary layer were unstable, as depicted by the KDVN modified sounding.

Subsequent evolution of the convective line and development of a bow echo failed to produce any reported wind damage (except the Washington, Iowa report) outside the observed radar tracks of the three dominant low level cyclonic circulations. Use of the Vr shear function to diagnose these intensifying low-level circulations proved very useful. As discussed by Klemp and Rotunno (1982) and also shown in numerical simulations (Wakimoto et al. 1996), development of the occlusion downdraft associated with the strong low-level rotation within the developing mesocyclone exerts a downward-directed pressure gradient and development of an occlusion downdraft. This is in addition to the rear flank downdraft, which is driven primarily by precipitation evaporation within the downdraft. These two downdrafts often merge to produce a continuous area of downward motion and subsequently damaging downburst winds at the surface, even with a stable boundary layer in place.


A continuous path of severe to extreme wind damage was observed during a National Weather Service storm damage survey the following day. The damage path coincided with the individual tracks of the low-level circulations. Numerous farms and homes from southeast Cedar county, across Scott county and into southeast Clinton County sustained moderate to severe wind damage.

It is likely that the development of the cold pool in balance with the ambient low-level wind shear provided deeper low level forced ascent and helped to intensify both the initial squall line and subsequent evolution into a severe bow echo (Weisman 1993). Numerous transient vortices were observed to form along the leading edge of the cold pool. These vortices were observed to migrate along the leading edge of the bow echo into the cyclonic inflow region of the developing comma head. Updraft stretching of three of these vortices likely contributed to the rapid development of the three main storm circulations that were observed with this bow echo and the observed Clinton, Iowa tornado touchdown. Radar operators can focus on these low level circulation trends in real-time using time-height cross sections of base data as well as the Vr shear function. Rapid low-level spin-ups even within a stable planetary boundary layer can produce large pressure perturbations that then allow damaging winds to penetrate to the surface and cause severe wind damage.


We wish to thank Ray Wolf for his initial encouragement and support on this project. We also thank Ron Przybylinski and Julie Adolphson for their helpful reviews and suggestions.


Burgess, D.W., and B.F. Smull, 1990: Doppler Radar Observations of a Bow Echo Associated with a Long-track Severe Windstorm. Preprints, 16th Conf. On Severe Local Storms, Kanaaskis Park, Alberta, Canada, AMS (Boston), 203-208.

COMET, 1998: Mesoscale Convective Systems: Bow Echoes and Squall Lines. (28 Feb 1999).

DOC, NOAA, 1996: Storm Data, 38, No. 5, 65-66.

Fujita, T.T., 1978: Manual of Downburst Identification for Project NIMROD. SMRP Research Paper #156, Univ. of Chicago, Chicago, Illinois.

Hart, J.A., W. Korotky, and G. Jackson, 1998: The Sharp Workstation v 1.80. A Skew-T/Hodograph Analysis Research Program for the IBM and Compatible P.C. DOC/NOAA/National Weather Service Office Charleston, WV, 30 pp.

Lemon, L.R., 1999: Personal Communication.

Rotunno, R., and J.B. Klemp, 1982: The Influence of the Shear-Induced Pressure Gradient on Thunderstorm Motion. Mon. Wea. Rev., 110, 136-151.

Wakimoto, R.M., W.C. Lee, H.B. Bluestein, C.H. Liu, and P.H. Hildebrand, 1996: ELDORA Observations During VORTEX 95. Bull. Amer. Meteor. Soc., 77, 1465-1481.

Weisman, M.L., J.B. Klemp, and R. Rotunno, 1988: Structure and Evolution of Numerically Simulated Squall Lines. J. Atmos. Sci., 45, 1990-2013.

____________, 1990: The Numerical Simulation of Bow Echoes. Preprints, 16th Conf. On Severe Local Storms, Kanaaskis Park, Alberta, Canada, AMS (Boston), 428-433.

____________, 1993: The Genesis of Severe, Long-lived Bow Echoes. J. Atmos. Sci.,50, 645-670. is the U.S. government's official web portal to all federal, state and local government web resources and services.