4/17/02 tornadic and nontornadic supercell environments over Kansas, Oklahoma, and Iowa  (several different supercells, including "elevated" environments)

On the evening of 4/17/02, severe thunderstorms erupted from western Oklahoma to central Kansas and western Iowa with several supercells that lasted well into the nighttime hours.  This page presents parameter fields and RUC-2 model analysis profiles associated with several different areas and supercells, showing estimated shear-CAPE (0-1 km EHI), deep layer vertical shear, and several low-level thermodynamic parameters.  The environment varied considerably across several states, as will be seen by evaluating parameter characteristics together.  An objective table with comparison of environments, parameters, and tornado reports will be found at the bottom of this page.

At 00 UTC 4/18/02, thunderstorms were continuing to develop rapidly over the central plains ahead of a 500 mb short wave in moderate to strong southwest flow aloft and east of a dryline that extended through western Kansas into the eastern Texas panhandle:
<visible satellite 00 UTC     < Eta 500 mb analysis 00 UTC   < 01 UTC surface map

Supercell 1 environment (south-central KS)
By 01 UTC, severe storms were occurring over south central Kansas and western Oklahoma, and by 0122 UTC a radar-based tornado warning was issued for a rotating storm just northeast of Medicine Lodge.  The RUC-2 analysis profile for Medicine Lodge (P28) indicated strong shear and CAPE, but the shallow moist layer suggested LCL heights and cloud bases that would be relatively high.   A deep layer of near-surface convective inhibition (CIN) was also present.
< ICT radar base refl. 0122 UTC                   < Medicine Lodge KS RUC-2 analysis sounding 01 UTC (from FSL RUC-2 web site)

The large CIN (> 200 J/kg) and also high LFC heights (> 2500 m) suggested that the storms in south central and central Kansas were "elevated", occurring above a strong low-level stable layer.  This, along with the relatively high LCL heights and cloud bases suggested that, at least early on, the threat in Kansas would be from hail and maybe some localized strong winds.  It is interesting that none of the storms in Kansas, including several supercells, produced any tornadoes during the evening.

Estimated parameter fields from the Storm Prediction Center (SPC) mesoscale analysis page showed large low-level shear and CAPE combinations over a broad area from the western half of Oklahoma across the eastern half of Kansas.  Deep layer shear was strongest across Oklahoma and southern Kansas, and also across Nebraska into Iowa:
< 02 UTC 0-1 km EHI   < 02 UTC BL-6 km shear

Looking at low-level thermodynamic fields, mean-layer LCL heights and associated cloud bases were lowest over western Oklahoma, but mean-layer CIN was large over most of the area. The smallest CIN (around 125 J/kg) and lowest LFC heights (< 2000-2200 m) were also over western Oklahoma, co-located with strongest values of the significant tornado parameter guideline (Thompson, 2002):  
< 02 UTC LCL height   < 02 UTC CIN and Significant Tornado Parameter   < 02 UTC LFC height

Supercell 2 environment (west-central OK)
The co-location over western Oklahoma of shear-CAPE (EHI) and deep shear along with deeper low-level moisture and resulting lower LCL and LFC heights with weaker CIN (although still relatively large) suggested that this might be an area where tornado threat would be greatest.  Near 03 UTC, the southernmost supercell in western Oklahoma produced an F1 tornado in Beckham county.  Notice that the RUC-2 analysis profile for Hobart (HBR, southeast of the tornadic supercell) had roughly half the CIN of the earlier Medicine Lodge profile, with lower LCL and LFC heights and comparable EHI and deep shear:.
< VNX radar base refl. 0257 UTC   < Hobart OK RUC-2 analysis sounding 02 UTC (from NIU "Storm Machine" web site)

Supercell 3 environment (western IA)
Meanwhile in western Iowa, a single large supercell with strong rotation on radar near a warm front was prompting several tornado warnings in strong low-level shear (0-1 km SRH > 400 m2/s2) at the nose/edge of increasing shear-CAPE combinations and strong flow aloft (see 02 UTC SPC graphics above):
< OAX radar base refl. 0200 UTC

However, the SPC low-level thermodynamic fields from 02 UTC showed high LCLs (> 1500 m), large CIN (> 200 J./kg), and very high LFCs (> 3000 m), suggesting a strongly "elevated" environment unlikely to produce significant tornadoes.  A brief F0 tornado with no damage was reported in Harrison county, and a couple other brief weak tornado reports later came from Carroll and Boone counties, but by far the main severe feature of this storm was 2" hail.  The 03 UTC RUC-2 analysis profile for Boone (BNW) showed a very "elevated" environment ahead of the storm in central Iowa, with the only parcels producing significant CAPE coming from around 800-850 mb, well above the boundary layer:
< DMX radar base refl. 0418 UTC    <  Boone IA RUC-2 analysis sounding 04 UTC (from FSL RUC-2 web site)

Supercell 4 environment (central KS)
During roughly the same time frame, storms continued to move northeastward in Kansas.  One supercell southeast of Salina prompted a tornado warning based on radar shortly before 04 UTC.  However, as with other storms in Kansas, the environment continued to appear "elevated".  Although the RUC-2 analysis profile for Hillsboro (HBRK, south of the storm) at 03 UTC continued to show very large low-level shear and 0-1 km EHI (> 5.0), CIN for a mean layer parcel was also very large (> 200 J/kg), indicating a deep low-level stable layer still over central Kansas that was unlikely to support significant tornadoes.  As noted earlier, no tornadoes were confirmed with this or any other storms in Kansas.
< VNX radar base refl. 0353 UTC     < Hillsboro KS RUC-2 analysis sounding 03 UTC (from FSL RUC-2 web site)

Supercell 5 environment (NW Oklahoma, actually a continuation of supercell 2)
Back in Oklahoma, the 04 UTC surface map showed a deepening moisture axis on strong south winds through western Oklahoma, with surface dew points in the low to mid 70s (F) to the Kansas border near Alva.  Surface dew point depressions of only 2 to 5 deg (F) along this axis suggested low cloud bases and large low-level humidity favorable for tornadoes.  The 05 UTC SPC mesoanlysis graphics showed continuing shear-CAPE combinations that were very strong (EHI 5.0-6.0) along the same axis, and deep layer shear that was maximized near the Kansas border, even though values appeared to have weakened some since 02 UTC:
< 04 UTC surface map  < 05 UTC 0-1 km EHI  < 05 UTC BL-6 km shear

Low-level thermodynamic parameters at the same time showed a minimum area of mean-layer CIN (< 100 J/kg) along this moisture axis over western Oklahoma (more favorable values than at 02 UTC, light blue) and a corresponding axis of lower LFC heights (< 2200-2400 m) co-located with strong values of the significant tornado parameter guideline: 
< 05 UTC CIN and Significant Tornado Parameter   < 05 UTC LFC height 

Even though only one tornado had occurred prior to 04 UTC in Oklahoma, the juxtaposition of the strong shear-CAPE combinations with favorable low-level thermodynamic parameters  suggested that supercell tornadoes were still possible, even late at night in this situation at a diurnally unfavorable time (for the plains).  By after 04 UTC, storms over western Oklahoma had generally decreased in coverage except for a broken cluster over northwest Oklahoma that included a supercell moving northeast through Dewey and Major counties.  This strong cell, a continuation of the supercell that earlier had produced the tornado in Beckham county, was moving through the axis of favorable parameters noted above.  It produced an F3 tornado in Dewey county shortly after 0428 UTC, and an F2 tornado with a 30+ mile track in Major and Woods county between 05 and 06 UTC.  
       
^ VNX radar base refl. 0428 UTC   ^ ICT  radar base refl. 0534 UTC and 0559 UTC

The RUC-2 analysis profile for Enid (END) at 05 UTC, southeast of the tornadic supercell, showed a deeper and more humid low-level moist layer than indicated further north over central Kansas (see earlier Hillsboro RUC-2 profile), with less than half the CIN.  The low-level shear and 0-1 km EHI were quite large (> 6.0) with lower LCL heights indicated as well: 
< Enid OK RUC-2 analysis sounding 05 UTC (from FSL RUC-2 web site)

Additionally, comparing the model analysis dew point to observed dew points over northwest Oklahoma, it appears that the 05 UTC RUC-2 depiction under-analyzed surface dew points in the Enid-Alva area by 2 to 3 degrees F.  In this case, that would increase 0-1 km EHI to near 8.0 (!), and further decrease CIN by 20-40 J/kg, lowering LCL and LFC heights by around 200-300 m for an even more favorable low-level thermodynamic environment.  This supercell dissipated as it crossed the border into Kansas, apparently moving into an area of somewhat shallower moisture and larger low-level stability.

Here is a map of SPC log reports for 4/17/02, showing that, apart from the brief/weak tornadoes in Iowa, the only  tornadoes (including the F2 and F3) occurred over western and northwest Oklahoma where the most favorable juxtaposition of shear-CAPE and low-level thermodynamic factors was located:
<SPC log reports for 4/17/02

Summary
A summary of the five supercell environments examined here is shown below, using the following parameter ranges to categorize relative strengths and weaknesses:
                                                                                                                            

Supercell environments evaluated above from nearby RUC-2 analysis profiles from the evening of 4/17/02:

Notice that the environments where one or both EHI and BL-6 km shear were "strong" while low-level thermodynamic parameters were at least "ok" or "marginal", and no parameters fell into the "poor" category, turned out to be most favorable for tornadoes, with F1,F2, and F3 tornadoes reported in those two cases.  Also notice that all environments had at least"marginal" to mainly "ok" and "strong" amounts of EHI (shear-CAPE combinations) and BL-6 km shear (deep layer shear).  The large CIN (> 150 J/kg) and high LFC heights (> 2500 m) along with LCL heights lower than around 1000 m appeared to be the main difference between the nontornadic or brief/weak tornadic environments and those that produced more signficant tornadoes.  This case study seems to highlight the importance of assessing low-level thermodynamic factors in conjuction with accepted evaluations of CAPE and shear. 

Some additional comments:
1)  The RUC-2 profiles used in this case study were not compared or modified with actual surface observations, which would give a forecaster a better idea of actual real time conditions.
2)  Mean-layer CIN of 150 J/kg or larger and LFC heights of 2500 m or greater seem to work as useful discriminators for defining environments so "elevated" as to be unlikely to produce significant tornadoes when accepted shear-CAPE parameters are strong over a broad area.
3)  The stronger tornadoes in this case occurred with CIN environments that were still relatively large (e.g., 80-130 J/kg).  Recent experience is showing that this is not unusual in the plains states when shear-CAPE combinations are also strong (e.g., 0-1 km EHI 3.0-4.0 and greater) and deep layer shear is at least adequate.  Further east (e.g., east of the Mississippi River), significant tornadoes tend to be limieted to smaller CIN and lower LFC heights. 
4)  The "elevated" Iowa case in this study was able to support several brief tornadoes despite being strongly elevated.  Notice in that case that 0-1 SRH was very large( > 400 m2/s2), and that deep layer shear was also large (> 50kts).    Elevated environments with such large shear may be able to generate weak tornadoes through a deep low-level stable layer.

Parameter fields like those from the SPC mesoanalysis page shown earlier, and an "attributes" table concerning environment parameters like the one above, may be useful for forecasters in evaluating short-term and current environments to weigh relative strengths and weaknesses for  likelihood of significant supercell tornadoes.  Certainly, evaluating environments regarding tornado potential is not a "yes"/"no" proposition, but some objective measures like those above may be helpful in the warning decision process.

- Jon Davies  7/1/02

back to 2002 cases page