BROOKINGS COUNTY SOUTH DAKOTA TORNADOES
LARGE HAIL AND DAMAGING WINDS OVER SOUTHWEST MINNESOTA
JUNE 22, 1997
by
Ron Holmes
INTRODUCTION
The second major severe weather event for the Sioux Falls County Warning Area (CWA) struck late in the afternoon on Sunday, June 22, 1997. The event began just after 5:00 pm and produced a swath of severe weather from Brookings county South Dakota southeast into Cottonwood and Jackson counties in Minnesota by 9:00 pm. Two tornados were spotted in Brookings county, one by a National Weather Service employee, and this storm later produced 2.00" diameter hail when it moved southeast into Pipestone county. Another storm dropped softball size (4.50") hail in Lyon county, MN. which smashed in a car window and also broke windows in homes.
SYNOPTIC OVERVIEW
At 500 mb a broad ridge was located over the Northern Plains with a broad trough settled over the west coast. There were no apparent short waves moving through the area and temperature advection aloft was either negligible or slightly bordering on warm advection. At 700 mb a hot prod of dry air downsloping off the Rocky Mountains created a strong cap (shaded red) over much of South Dakota which inhibited convection from forming despite very warm, moist air at low levels. As a general rule in the Central Plains temperatures above +12 C at 700 mb often produce this strong cap and inhibit strong convection. However the edge of the cap was located over extreme eastern South Dakota into extreme southwestern Minnesota. The severe thunderstorms which later developed formed just along the edge of this cap to the east where 700 mb temperatures were cooler.
Lower in the atmosphere at 850 mb the leading edge of very warm air, which marked a warm frontal boundary, existed across extreme eastern South Dakota and a tongue of rich moisture with dewpoints greater than +15 C (shaded green) existed parallel to the warm front. The dewpoint at Aberdeen was a whopping +19 C. At the surface the warm front was located just south of Brookings and extended to the southwest corner of Minnesota at 2000 UTC (3:00 pm) with surface dewpoints greater then 70 F (shaded green) nosing up into southeast South Dakota.
FAVORABLE ENVIRONMENTAL HELICITY
Note in these upper air charts the strong southerly winds at 850 mb with the wind veering (becoming southwest then westerly) with height up through 500 mb. This produced strong directionally-varying wind shear and favorable storm relative helicity once storms formed. This can also be seen in the Neligh, Nebraska wind profiler which showed strong, low level southerly flow which decreased in strength but became more southwesterly at mid levels. The strong southerly winds at low levels advected abundant moisture northward up the isentropic surface north of the surface warm front which helped form the storms. Once they developed they propagated southeastward, feeding along the best moisture axis associated with the frontal boundary and to the right of the mean wind.
The upper air site nearest to where the storms first developed was at Aberdeen. The 0000 UTC (7:00 pm) sounding was quite unstable with an observed CAPE of 3923 J/kg, a Lifted Index of -10, and a 700-500 mb lapse rate of 8.5 C/km. The observed wind field also reflected the favorable shear profile that the Neligh profiler showed with good directional and speed shear. This resulted in a clockwise curved hodograph. Numerical modeling experiments by research meteorologists have shown that a strong updraft in an environment in which the shear vector turns with height (curved hodograph) is favorable for producing a differential pressure gradient directed from high pressure at low levels of the storm to low pressure in the mid levels of the storm. This dynamically forced pressure perturbation enhances the updraft on the storm relative inflow side of the storm. The storm then propagates to the right of the mean wind which often provides a better storm relative flow of warm, moist air. The mean 0-6 km wind for the observed profile was 213 at 24 knots and based on an estimated storm motion of 243 at 18 knots the storm relative helicity was 476 m2/s2. However due to dynamic processes associated with a strong updraft within a strong directionally sheared environment the storms continued to develop to the right of the mean wind towards the southeast. Using the observed motion of storms, 291 at 22 knots, the storm relative helicity increased to 631 m2/s2. In both the observed and modified hodograph the storm relative flow was from the southeast which continually supplied the storms with a fresh supply of moist, unstable air, however in the modified hodograph this flow was much stronger leading to a stronger updraft.
PHASE 1 - TORNADIC STORM ACROSS BROOKINGS COUNTY AND OBSERVATIONS OF SEPARATE HOOK ECHOS
The first storms developed south of Aberdeen and moved southeast across the northern border of our CWA into Brookings county. A 4-panel time lapse view of WSR-88D reflectivity shows the evolution of this storm shortly after it crossed the border. Time increases from panel 1 (21:59 UTC) (4:59 PM) to panel 4 (22:34 UTC) (5:34 PM) and shows how the storm developed a hook echo with a good inflow notch at 22:14 UTC (5:14 PM) (panel 3) which then wrapped around the storm by 22:34 UTC/5:34 PM (panel 4). A severe thunderstorm warning was issued for northern Brookings county at 5:02 PM and then a tornado warning was issued for eastern Brookings county at 5:26 pm valid until 5:55 pm. Around this time a tornado was reported 7 north of Brookings (at 22:30 UTC) (5:30 PM) and another tornado 2 north of Aurora (at 22:39 UTC) (5:39 PM).
A 4-panel time lapse of Storm Relative Velocity Map (SRM) matching the reflectivity 4-panel shows a strong mesocyclone in the low to mid levels of the storm (10,000 to 13,000 feet) with a rotational velocity >50 knots. The combination of good reflectivity structure and strong rotational shear were good visual clues to the radar operator that this storm was tornadic.
A 4-panel slice of the storm at 22:19 UTC (5:19 PM) details the structure of the storm just before it produced the tornado. In panel 1 you can see the early stage of the hook echo with a tight reflectivity gradient on the southeast side. When comparing panels 3 and 4 with panel 1 you can see higher reflectivity aloft over the reflectivity free area at low levels indicative of a strong weak echo region and updraft. In panel 4 the reflectivity aloft almost forms a BWER.
A cell trends product for this storm clearly shows the gradual increase in maximum reflectivity and cell based VIL as well as a dramatic increase in probability of severe hail from 22:09 UTC to 22:19 UTC just before the tornados formed.
One of the more interesting aspects of this event is the production of another, separate hook echo moving across Brookings county about an hour after the first one. Panels 1 and 3 show the reflectivity and SRM products at 23:34 UTC (6:34 PM) and panels 2 and 4 show the same products at 00:09 UTC (7:09 PM). A hook echo and inflow notch is clearly evident on the southeast side of the storm in panel 2. Cyclonic convergence can be seen in the SRM product in panel 3 with more pure cyclonic rotation in panel 4. We did not receive any severe weather reports with this second thunderstorm but the radar signatures clearly suggested it was most likely severe. In fact a severe thunderstorm warning was issued for eastern Brookings county at 7:43 PM.
PHASE 2 - PRODUCTION OF LARGE HAIL ACROSS SOUTHWEST MINNESOTA
As the event unfolded new storms developed over Lincoln and Lyon counties in Minnesota but these did not produce tornados. One explanation might be due to the fact that the vertical wind shear over southwest Minnesota was not as favorable for producing rotating updrafts as it was over eastern South Dakota. The Wood Lake profiler (located to the northeast of Marshall) revealed more of a unidirectional shear profile that was not as strong as the veering profile seen on the Aberdeen sounding. Although this information was not plotted on a hodograph it strongly suggests that the pre-storm environmental shear over southwest Minnesota would have resulted in a straight hodograph. This would favor splitting storms with both a left moving and right moving thunderstorm. Unfortunately, after 2300 UTC thunderstorms moved over the profiler site and contaminated the wind data resulting in the chaotic appearance of winds below 600 mb.
Lyon county Minnesota received the brunt of large hail with softball size (4.50") hail breaking a car window 13 west of Marshall at 6:20 pm. A 4-panel view of this storm at 6:04 PM (23:04 UTC) shows the highest reflectivities in the lower portion of the storm, a strong mesocyclone with rotational velocity >50 kts at 18,000 feet, and strong storm top divergence (>110 kts) . These were all good signals that the storm was severe. The composite reflectivity product identifies this storm as M3 and the table at the top indicated a 100% probability of hail with a maximum hail size of 4.00"...close to the 4.50" hail observed. Note also how the storm tracking algorithm shows this storm and the one over Lincoln county (to the west) moving to the southeast while all the other weaker storms are moving to the northeast. This again demonstrates the right moving nature of these storms. The cell trends product for this storm indicates this storm was severe long before it produced the softball size hail. Indeed, a severe thunderstorm warning was issued for Lyon county at 5:45 pm well before the first report of severe hail (at 6:05 pm). This warning was then extended at 6:30 pm until 7:00 pm.
PHASE 3 - STORMS PRODUCE SEVERE DOWNBURST WINDS
Toward the end of this severe weather event the storms evolved into a large cluster of cells which eventually organized into a bow echo and produced damaging straight-line winds. As the cluster of storms became larger in size rain-cooled air from the storms produced a meso-high over southwest Minnesota north of the warm frontal boundary. An outflow boundary along the leading, southern edge of this rain-cooled air interacted with the warm front to the south to produce an area of enhanced convergence across southern Minnesota into northwest Iowa.
A 4-panel time lapse of reflectivity shows how new storms formed in this area south of the outflow boundary and moved northward colliding with the cluster of storms moving across southwest Minnesota. Before the collision a bow echo was beginning to form over western Cottonwood county where winds gusted to 60 mph at 7:00 pm. However when the bow echo slammed into the northward moving storms the bow echo became much stronger and better defined. At this time (01:01 UTC) (8:01 PM - panel 3) 77 mph winds were measured by an automatic observing station in Windom. By 01:31 UTC (8:31 PM - panel 4) the bow echo was east of Jackson county but the leading edge extended southwest into the county. Wind gusts associated with this leading edge reached 60 mph in the town of Jackson.
