Mesoscale Aspects of the 05 June 1997 Severe Weather Event

Overview:
On 5 June, 1997, NWSFO DMX issued 10 warnings of which 7 were verified. There were at least 2 unconfirmed tornadoes in Butler county causing considerable damage to a farm shed near the town of Dumont. There were numerous reports of hail, with one report up to 1.75 inch. A news report the next day revealed how a whole field with corn and soybeans near Newton was totally flattened by golf-ball sized hail. Damaging winds were not widespread, but there was a report of 50 mph in Blackhawk County. A TV station reported that a "microburst" occurred near Merle Hay Road and Hickman Road in Urbandale where large tree limbs were downed. However, the DMX office never confirmed damage caused by winds above severe criteria. The observing network in Des Moines recorded gusts of 35 to 40 kts at DMX. Reports of hail lessened as the storms moved into south central Iowa. See Figure 1 for a graphical plot of all severe weather reports for this event. The initial severe weather occurred at a time of day not climatologically favorable for severe weather in Iowa. The first report of severe weather occurred at 1804 UTC and the last one occurred at 2245 UTC. This event illustrates the importance of using mesoscale data to assess the storm environment and the potential for different types of severe weather.
	Figure 1. Severe weather reports from 1200 to 2300 UTC 05 June 1997.

Mesoscale/Synoptic scale Interactions: Doswell (1987) stresses how the synoptic scale and mesoscale interact in convective situations. The 05 June case illustrates how this interaction took place. Continuous and thorough mesoanalysis can play a key role in the warning process. This case will also document this role. Synoptic Environment: At 1200 UTC a 500 mb north-northwest flow regime was present over the upper Midwest. A potent and compact circulation was observed on the water vapor imagery and in the 500 mb analysis over the arrowhead region of Minnesota. The 1200 UTC model runs including the ETA and RUC placed this feature over southeast Minnesota by 0000 UTC. Figure 2 shows a PCGRIDDS display of the 12 hour ETA forecast position. Note the 12 hr forecast cold pool at 500 mb of less than 19 C over southeastern MN. Modified 1200 UTC soundings at MPX and DVN (Figure 3) based on forecast maximum surface temperatures and expected 500 mb temperatures of -16 to -19 deg C supported lifted indices of -4 to -6 and CAPES ranging from 1500 to 2500 J/kg. The hodograph from DVN (Figure 4) at 1200 UTC suggested tornadoes associated with any supercells would not be anticipated given that the shear was so minimal. The Bulk Richardson Number (BRN) forecast also favored multi-cell development with values >=50 over a large part of north central and northeast Iowa (Figure 5). Given the presence of convection, large hail was the anticipated severe weather type due to the cold temperatures aloft and unusually low Wet-Bulb Zero heights of 8000-9000 ft, especially close to the cold pool center. Another favorable factor for convective formation was a 850 mb thermal ridge at 1200 UTC extending from western South Dakota/Nebraska eastward across the Iowa/Minnesota border (Figure 6). Moisture was quite shallow with 850 mb dewpoints of >8 C located across western South Dakota southward through the central Plains and along portions of the mid-Missouri River Valley. Deep convection would have to rely upon pooling of moisture along the cold front. This is one reason that mesoscale analysis is essential. The analysis reveals key features and aspects that were not noted on the 1200 UTC model forecasts.

Figure 2. ETA 12hr forecast 500mb winds and temperatures. Figure 3. Modified 1200 UTC Davenport, IA sounding. Figure 4. 1200 UTC DVN Hodograph.

Figure 5. ETA 12hr forecast of Bulk Richardson Number.	Figure 6. 850mb 1200 UTC analysis.

Mesoscale Environment: As stated earlier, convective development occurred early in the day around 1700 UTC. The first storms developed along the cold front in south central Minnesota and quickly became large hail producers. However, as the storms moved southeast into the DMX CWA, no large hail was reported until the storms reached Butler county between 1800 and 1900 UTC. The earlier development time of the storms in southern MN can be attributed to a weak cap in place. It can also be attributed to enhanced moisture convergence near the surface low / cold front intersection. This was not anticipated looking at the 12 hr ETA forecast valid for 0000 UTC 6 June 1997. In Figure 7, note that the 12 hr forecast 1000 mb winds were not showing convergence or much of a cyclonic circulation center along the front at 0000 UTC. The 2 meter (AGL) dew point temperature forecast, however, did reflect a pooling of moisture in central Iowa. Observing the 1800 UTC surface map in Figure 8, there is a weak surface low (1008 mb) that has developed between Mason City and Waterloo. Careful examination of the analysis indicated some 3-hour pressure falls of -1 to -2.4 mb hr-1 were taking place ahead of this low which would eventually move southwest to just west of Des Moines by 0000 UTC. Also notice the triple point at the juxtaposition of the low, cold front, and surface trough which extends into southwest Iowa. What do you think has happened here? This is not a convective outflow boundary or a surface reflection of a developing synoptic scale low. The GOES sounder information was helpful in determining atmospheric instability. Figure 9 shows the effects of the cold pool aloft on the resulting CAPE values. By 1800 UTC, CAPE values approaching 1000 J/kg were analyzed across portions of southern MN and northern IA where storms were starting to reach severe limits.

Figure 7. ETA 12hr forecast 1000mb winds. Figure 8. 1800 UTC analyzed surface chart.

Radar operators began to monitor a storm with weak rotation moving south into Butler County (See the 1825 UTC WSR-88D radar images, Figures 10a and 10b.) Several other cells had also developed along the frontal band. The Butler County storm was positioned near the triple point. Other smaller cells had developed along the surface trough near the RDA. The first of two public reports of a tornado was received at 1823 UTC. The tornado touchdown was south of Dumont in Butler County in association with the storm located near the triple point. Although WSFO DMX was initially unsure of the tornado report, the environment had become much more favorable for supercell and incipient tornado development. The storm motion determined by the 88D for the Butler county storm was 335/24. Manual estimates were closer to 350 or 360 degrees, mostly due to the back-building nature of the storms to the west. The other cells exhibited movements of 280-300 degrees. Any deviant storm motion should always alert the radar meteorologist to evolving conditions in a storm. In this case, it had evolved into a rotating storm with the potential for tornadoes. Remember that the shear pattern earlier based on the ETA forecast 850 mb/500 mb winds favored multicellular storms. Looking at the 1900 UTC Iowa mesonet plot (Figure 11), note how the surface winds have backed to a more southerly direction on the southern and eastern periphery of the meso-low center.

Figure 9. 1800 UTC CAPE data from GOES.

Figure 10a. 1825 UTC DMX Base Reflectivity.

Figure 10b. 1825 UTC DMX Storm Relative Mean Radial Velocity Map.

Figure 11. 1900 UTC Surface mesonet plot.

In addition, the vertical wind profile as depicted from the 1904 UTC WSR-88D KDMX Velocity Azimuth Display Wind Profile (VWP) product showed the backing low-level winds now producing some significant directional wind shear in the lowest 10,000 feet. (Figure 12) There are three ways to assess mesoscale changes in the vertical wind profile: Modify a sounding using model forecast data or observed data. Use wind profiler data (from the VWP product on the WSR-88D or from a profiler in the Profiler Demonstration Network) and then modify the hodograph for the observed storm motion. Use a RCM program (Hinson, 1995) which utilizes output from the radar coded messages (including the VAD Wind Profile data). By using the WSR-88D derived storm motion of 335/24, the 0 - 3 km storm-relative helicity (as calcuated by the RCM program) (Figure 13), has increased to 111 m2/s2. By modifying the storm motion to a more northerly direction, this number would increase even more. By 2000 UTC, the storm over Butler County had moved south into Grundy County, still coincident with the triple point. Note that the 60F dewpoints lie near the triple point in central Iowa (Figure 14). Convergence was maximizing in this area. The strongest storms maintained themselves on their western flank. This can likely be attributed to the thermal ridge which was positioned to the west. Note the temperatures at Carroll, Denison and Audubon (sw quadrant of the meso-low) in the middle 80s to around 90F. The storm- relative inflow originating from this warmer tongue of air was helping to increase and maintain updraft strength. The Interaction: Why did this meso-low develop? Again, there was no MCS which nearly always creates a mesoscale pressure system of it's own. It is possible that the upper level jet structure had a lot to do with the development of this low. Recall that ageostrophic divergence aloft results in a compensating convergence in the lower levels of the atmosphere. This was indeed the case as a strong 80 kt, 250 mb jet dove southward into Iowa during the afternoon of the 5th. The 12 hour ETA forecast (Figure 15) valid at 0000 UTC 07 June shows this feature and the resulting ageostrophic divergent center within the left front quadrant of this cyclonically curved jet max. The maximum was forecast to be above where the surface low developed. This synoptic feature enhanced the mesoscale features in the convective environment by increasing low-level convergence which resulted in lowered surface pressures (and increasing vertical motion). Resulting surface pressure falls enhanced low-level storm-relative flow into updrafts of convective storms, (which in turn, increased generation of horizontal vorticity and resultant storm-relative helicity). The pooling of low-level moisture continued to increase local areas of instability which maintained updraft strength as the storms moved south-southeastward. Conclusions: This case represents a good example of utilizing mesoscale data (such as surface mesonet and WSR-88D data) to modify expectations of convective storm type. Originally, the anticipated severe storm type of the day was thought to be hail , with forecast shear parameters such as Bulk Richardson Number favoring multi-cellular storms. However, due to changes brought about by increased vertical motion, formation of a meso-low, and differential thermal (cold) advection increasing instability over the low-to mid layers of the atmosphere, the potential for tornadic storms increased. It is very important for forecasters to always examine any changes in the low- level wind fields (increasing convergence, backing winds, etc. ) and be alert for strong convective development along boundaries. These mesoscale changes can be monitored effectively and will improve warning decisions. During a potential severe weather event it is important to: Always utilize a severe weather meso-analyst (SWAN) before and during severe weather. Carefully analyze potential vertical wind structure by modifying soundings, using the VWP product, RCM decoder, or wind profiler data. Note the thermodynamic and moisture parameters (i.e. pooling of moisture and strong ageostrophic divergence) which may indicate mesoscale development that the model may not be resolving. For example, in this case, the strong geostrophic divergence is a clue to increased surface low development.

Figure 12. KMDX 1904 UTC VAD Wind Profile. Figure 13. Radar Coded Message. Figure 14. 2000 UTC Surface mesonet plot. Figure 15. ETA 250mb heights, winds and divergence of ageostrophic winds.

Radar Data

• Large Version (3.3MB, 640x480, continuous loop.)
• Small Version (724KB, 320x240, continuous loop.)

References:

Radar Coded Message Decoder Program, 1995. Larry J. Hinson, WSFO Omaha, NE. Version 1.82.

Acknowlegements: