The Sioux City Downburst:
An Example of Mid-Altitude Radial Convergence and Bookend Vortices
Philip N. Schumacher
National Weather Service
Sioux Falls, South Dakota
1. Introduction
A downburst caused significant damage to portions of Sioux City, Iowa on 2 August 2001. The damage was primarily across the northern portion of the city - from North Sioux City and the Riverside area to the Plymouth County line. The worst damage generally ran from southwest to northeast across the city with a width of 3 to 4 miles. The worst damage extended for 3 or 4 miles and was 2 or 3 blocks wide. Wind damage continued to the east of Sioux City affecting the Hinton area as well as Kingsley Iowa. The strongest wind report was 98 mph in Sioux City with numerous 60-70 mph (with trees down) into eastern Woodbury and southeastern Plymouth County. As the storm moved to the east of Woodbury County, it took the form of a mature bow echo across Cherokee and Buena Vista County. The evolution of this type bow echo was similar that discussed by Przybylinski (WAF, June 1995) and Fujita (1978). The goal of this note is twofold. First, despite the fact this was not a classic event, this will show how the use of Mid-Altitude Radial Convergence (MARC) was useful in providing some lead time to the Sioux City winds. Second, to show a classic example of bookend vortices and the resultant subsidence seen in the rear of the storm as it moved across Cherokee County.
2. Use of Gate-to-Gate MARC
Przybylinski (Severe Local Storms Preprint Volume, 1999) shows how MARC can be used to provide up to 20 minutes of lead time for severe wind events. He notes that not all high winds events are accompanied by MARC signatures nor do all MARC signatures produce severe winds. However, identifying MARCs coupled with an understanding the mesoscale environment can help the warning forecaster to issue a severe thunderstorm warning for damaging winds. MARCs are calculated using a technique similar to storm top divergence. Look for near adjacent (separated by < 3 nm) pixels where there is convergence - usually near the front edge the storm (although behind the gust front). Add the absolute value of the inbound velocity to the absolute value of the outbound velocity. The total is the MARC value for the storm. For this note, I have calculated both the maximum radial convergence and the "gate-to-gate" convergence (adjacent pixels along the same radial). From Przybilinski's work, one looks for MARC values below 10 km (approximately 30000 ft) which have convergence greater than 25 m s-1 (50 kts). As with TVSs and mesocyclones, one needs to look for temporal continuity, vertical continuity, and location of the MARC with respect to the storm. If there is time, one can also determine if the maximum convergence is lowering with time and has temporal continuity. In general, the storm motion should be parallel to the radial you are calculating the convergence along. In this case, the storm motion was nearly perpendicular to the radial which led to more difficulty in calculating MARC.
Radar images will be followed from 1231 UTC to 1254 UTC. Examining the reflectivity and the storm relative motion (SRM) at 1231 UTC shows that an elevated reflectivity core was developing along the Missouri River west of Sioux City. The height of the 65 dBz core at 3.4 degrees (lower left) was 28 kft. Meanwhile the SRM data shows only weak convergence over northern Dakota County at 0.5 degrees. While the large scale convergence increases at 1.5 degrees (max inbound of 26 kts and max outbound of 45 kts separated by approximately 8 nm), gate-to-gate radial velocity convergence is still relatively low with the maximum value less than 20 kts. In fact, the 1.5 degree SRM is more indicative of a broad cyclonic convergent signature associated with a developing supercell. At this point, the concern would mainly center on large hail as the updraft is producing an elevated hail core.
At 1237 UTC, the reflectivity imagery shows an elevated core over the northwestern portion of Sioux City. This core has begun to lower as there is now 65 dBz at 0.5 degrees at the confluence of the Big Sioux and Missouri River. The threat of large hail remains large for the Sioux City metropolitan area. The SRM at the same time has evolved with convergence more evident at all levels. At 0.5 degrees, the cyclonic circulation is still dominant. The value of the maximum "gate-to-gate" (GTG) convergence is only 24 kts at the confluence of the Big Sioux and Missouri River. However, at 1.5 degrees, the convergence has increased markedly. Near North Sioux City, a GTG MARC value of 42 kts can be observed. Note that at the same time the broad convergence (taking the maximum inbound and maximum outbound) has actually decreased. It is the "near adjacent" values that are important in calculating MARC not maximum values separated by > 5 nm. Even the 3.4 degree SRM shows evidence of convergence west of North Sioux City with a GTG value of 32 kts. When calculating radial convergence for pixels 3 nm, the 1.5 degree MARC value is 52 kts and 3.4 degree value is 61 kts. The 0.5 degree MARC value is unchanged. It was at this point that Sioux City Airport recorded a wind gust of 53 kts. And at 1240 UTC, 70 mph winds were reported in North Sioux City South Dakota.
Moving to 1242 UTC, the reflectivity core continues to lower. The 3.4 degree reflectivity no longer shows greater than 65 dBz. These values are confined to 1.5 and 0.5 degrees over Sioux City. Of course, a collapsing reflectivity core can also be indicative of a developing downburst. But since most severe storms have collapsing reflectivity cores, this alone does not guarantee a large downburst. However, the SRM continued to show increasing mid-level convergence - especially at 0.5 and 1.5 degrees. The maximum value was at 1.5 degrees - where the MARC value increased to 61 kts. At 0.5 degrees, the MARC value was 42 kts while even 3.4 degrees was at 24 kts. When using pixels within 3 nm, the MARC values are the same. The value of 61 kts is above the critical value of 50 kts that Przyblinski has found. Of more importance, the GTG has temporal consistency for the last two scans and there is even some spatial consistency between 0.5 and 1.5 degrees. This increases the confidence that the MARC is indicative of a developing downburst. While wind gusts to 55 kts were reported at the Sioux City airport, this is also still 5 to 8 minutes before the most destructive winds hit the northern part of Sioux City.
Examining the 1248 UTC and 1254 UTC reflectivity show that the reflectivity core has collapsed. By 1254 UTC, there is not any reflectivity greater than 65 dBz and the highest reflectivity within the storm is at 0.5 degrees (in the Sioux City vicinity). Any large hail has likely fallen and the storm reports support this as the large hail was generally before 1250 UTC. However, the 0.5 degree reflectivity has begun to bow with a rear inflow notch evident at both 1248 UTC and 1254 UTC. This rear inflow notch is south of the location of the highest winds in this case. Also both the rear inflow notch and bowing is coincident with the onset of the strong winds and could not be used as a predictor. By 1248 UTC, the MARC signature is only evident at 1.5 degrees with GTG convergence of 52 kts observed over the northeast portion of the metropolitan area. At 3.4 degrees, there is 34 kts of convergence in approximately the same area. No significant GTG convergence is observed at 0.5 degrees. By 1254 UTC, the maximum GTG convergence is located at 1.5 degrees south of Hinton. But the value is only 24 kts. Over the same area there is divergence at 3.4 degrees and only weak convergence at 0.5 degrees. While wind damage was reported around 8 AM in the Hinton vicinity, the damage was not nearly as extensive as in Sioux City. The evolution of the MARC suggested not only the development of a downburst in the Sioux City area but also the weakening of the maximum wind speeds to the northeast of Sioux City. This, despite the fact the storm was showing evidence of bowing on the reflectivity.
3. Maturing of the bow echo
Skipping ahead to 1329 UTC, the storm has matured into a classic bow echo. In general, the highest reflectivities are in the lowest two elevation angles. The exception is the southern end where developing cells still result in high mid-level reflectivity. Some large hail was reported with the southern cell. Also, note that the bow echo tilts forward with height and that the 1.5 degree and 3.4 degree echoes show the bowing structure due to the rear inflow notch. This is indicative of the rear inflow jet air subsiding toward the surface. In the SRM, the northern end of the bow echo shows evidence of a cyclonic bookend vortex. Although somewhat ill-defined, there is also evidence of an anticyclonic vortex at 0.5 degrees and 1.5 degrees.
At 1335 UTC, the subsiding rear inflow jet is causing the development of rear inflow notch at 0.5 degrees and reflectivity values continue to decrease at 1.5 and 3.4 degrees in the rear of the storm. At the same time, both bookend vortices are evident at all 3 elevation angles (0.5, 1.5, and 3.4 degrees) of the SRM. At this time, it is most obvious at 3.4 degrees where cyclonic vortex is over northwestern Cherokee County and the anticyclonic vortex is on the Cherokee and Buena Vista County line. I have also placed 4 panel reflectivity and SRM figures every 6 minutes through 1404 UTC (1341 UTC reflectivity and SRM, 1347 UTC reflectivity and SRM, 1352 UTC reflectivity and SRM, 1358 UTC reflectivity and SRM, and 1404 UTC reflectivity and SRM) . Note that at 1358 UTC, the reflectivity 4-panel uses 0.5, 1.5, 2.4 and 3.4 degrees instead of 0.5, 2.4, 3.4, and 6.0 degrees. As the storm progresses, note how the rear inflow notch becomes increasing evident at 0.5 degrees as the rear inflow jet is subsiding toward the surface. In general, the bookend vortices also become more pronounced at all levels as the storm moves to the east. In fact, the rear inflow jet may progresses so far ahead of the vortices that it appears the updraft gets cut off and the northern half of the storm weaken. The southern half of the storm did show evidence of bowing as it moved through Buena Vista County between 1400 UTC and 1430 UTC.
With the development of bookend vortices, one would expect that severe winds would develop near the apex of the bow, especially as the rear inflow notch was developing across Cherokee County. Both vortices and rear inflow notch indicate an accelerating rear inflow jet subsiding toward the surface. This was certainly assumed during the event and several warnings were issued for strong winds associated with the bow echo. However, despite the classic structure of the bow echo, only one wind gust in excess of 50 kts was reported with this bow echo between 1330 UTC and 1404 UTC. It illustrates the importance of atmospheric stability in the development of wind events. Because of the lack of soundings it is impossible to determine how the low level stability changed from Sioux City to Storm Lake. One can only hypothesize that there were two factors which contributed to the difference in severe reports. First, in was probably less stable in the boundary layer over Sioux City and second the downdraft was more negatively buoyant over Sioux City than farther east. Obviously, neither hypothesis can be easily tested during the event. The best the warning forecast team can do is be knowledgeable about the general environment (is there dry air at mid-levels, how unstable is the lowest 5 to 10 kft) and aggressively seek reports from spotters on the ground when it appears that severe winds may be occurring or a bow echo is evolving. In my opinion, warning for damaging winds from sunset through mid-morning is one of the biggest challenges we have because it requires detailed knowledge of the atmospheric stability. Even an inversion less than 1 km deep can prevent severe winds from reaching the surface.
