April 10 2005 tornadoes in NW Kansas associated with a 500mb cold core low system        by Jon Davies

Early spring 2005 has been the season of the 500 mb cold core low, with most tornadoes in the plains so far (prior to mid April) associated with such systems.   March 21 (n-c OK/s KS) and April 10 (nw KS) both produced several tornadoes from low-topped supercells with low cloud bases.  March 30 (n IA/se MN) and April 5 (sw KS) also produced tornadoes associated with 500 mb closed lows, but from storms with much higher cloud bases that were less typical of such systems and that appeared nonsupercell in nature. 

This study will look at the April 10 case in Kansas that involved several supercells and my first real storm chase (with Jim Reed) of 2005.  (Just click on any yellow-bordered images to see a larger version.)  The photos below show a tornado from one of the lesser-photographed low-topped supercells that day that passed just west of Wakeeney.  Note the sharp tilt angle of the updraft in the first picture (looking ENE from near Collyer KS), resulting from the large deep layer shear that day (50-60 kts through 6 km, not shown).  Low-topped tornadic supercells can occur in environments with weak deep layer shear (such as March 21 in n-c OK, < 30 kts) or strong deep layer shear (such as April 10, > 50 kts), but the days with stronger shear like April 10 will tend to produce longer lived and possibly stronger tornadoes.

Jared Guyer at SPC and I worked on a study for the recent AMS Severe Local Storms conference that looked at tornado-producing 500 mb closed cold core lows.   We found that most such systems had surface dew points at least in the low-mid 50s F within 200 miles of the 500 mb low center, and a surface"focus" area near a boundary intersection at the northwest "tip" of the warm sector east or northeast of the surface low, as shown in the idealized composite below.  Tornadoes (usually from low-topped supercells) were most likely near and just east of this surface "focus" area, even if total CAPE did not appear large (300-900 J/kg) and composite tornado parameters such as EHI (energy-helicity index) and STP (significant tornado parameter) were not suggestive of supercell tornadoes.  As will be seen in graphics below, the April 10 case matched the basics of this composite well, though both the 500mb low and surface low were southwest of the surface "focus" area, with an occluded front extending back southwestward from this focus to the surface low.

The morning 500 mb forecasts below from the Eta/NAM model showed the closed 500 mb low moving slowly into the central high plains from the Rockies during the day, with a strong vorticity max ("X") moving northeastward into western KS.   From experience, if there is adequate moisture and a good surface "focus" in the vicinity, tornadoes with 500 mb cold core lows often occur just ahead of compact areas of associated midlevel vorticity.

The surface forecasts below (from the RUC model) also showed a surface low with the midlevel system, moving northeastward through western KS during the day, and suggested a surface" focus" area in the wind pattern somewhere in western KS northeast of the surface low with southeasterly and easterly winds converging into northeasterly winds.

Using Earl Barker's site, a look at forecast parameter fields for mid afternoon from the RUC such as surface-based total CAPE and surface-based 0-3 km CAPE suggested a narrow but enhanced area of instability northeast of the surface low in the surface "focus" area suggested by the RUC surface forecast maps.  The northwest extent of this forecast instability would likely be enhanced by any surface heating and the proximity of cold air aloft associated with the 500 mb low.  Surface-based instability fields often work best in 500 mb cold core events because instability tends to be located lower to the ground with the cold air aloft and lower tropopause nearby, and if a warm front is involved, low-level conditions may be near saturated near the front.  In many such cases, mean lowest 100 mb mixed layer lifted parcels "mix out" the relevant instability and do not capture representative CAPE values.

Surface heating is an important factor in tornado events associated with 500 mb cold core lows because the events are diurnally driven.  The RUC forecasts of surface temperature and MLLCL heights for mid afternoon below suggested an axis of strong surface heating and steep low-level lapse rates pointing into the surface "focus" area discussed above, the result of a synoptic-scale dry intrusion extending around the southeast and east side of the 500 mb low, usually present in most organized 500 mb cold core low tornado events.

Moving from forecasts to "real-time", a satellite photo at mid afternoon (below), and a detailed surface map (below, courtesy of Al Pietrycha) clearly showed a well-defined dry intrusion (sunny skies and surface dew points in the 20s F) pointing into the surface "focus" area (in this case a "quadruple point" joining 4 air masses) where surface dew points were in the mid-upper 50s on southeasterly winds in the warm moist sector.  A photo I shot (below)  in southern Gove County at about 2035 UTC looking NE from within the dry intrusion shows what may have been the boundary intersection of the surface warm front and the Pacific front acting as the dryline/dry intrusion... notice the abrupt change in cloud base/LCL heights under the low-topped storms reflecting the rapid transition from colder air to warmer air.   Cloud bases to my immediate north at this time were nearly on the ground on the cold side of the occluded front, in marked contrast to cloud bases further east along the leading edge of the dry intrusion.  As an aside, my chase partner Jim Reed and I encountered an impressive hail accumulation swath from one of these storms near the occluded front north of Gove KS that covered the ground 1-2 inches deep (see photo 1 and photo 2), plunging us into a winterscape of white, with areas of hail fog and slick roads making for difficult driving.

SPC mesoanalysis graphics at 21 UTC showed an axis of steep low-level lapse rates (0-3 km, indicated in the images below by red dots) pointing toward the boundary intersection "focus" area noted above, intersecting the northwest "tip" of the warm sector and axis of surface-based instability (total CAPE and low-level CAPE, below).   In the 500mb cold core low tornado events I've examined, this axis of steep lapse rates is almost always present , and points toward the general area where tornadoes are most likely just ahead of the advancing dry intrusion if adequate moisture is present.  This seems to be a valuable tool in suggesting an area of enhanced and relevant localized buoyancy in 500 mb cold core cases, and it is best for forecasters to pay more attention to the orientation of this lapse rate axis and where it is pointing, rather than specific values of lapse rate or CAPE.  Notice also the low LFC heights (below) near the NW "tip" of the instability axis, suggesting that instability near the surface "focus" area (yellow ellipse) was located close to the ground, possibly enhancing potential for vertical stretching in low-levels, moreso than total CAPE values (larger further southeast) would suggest.   Often the instability axis in 500 mb cold core events may look too narrow to provide an adequate window of opportunity for storms to produce tornadoes, but the subtle yet intensified thermodynamic environment at  the northwest extent of this instability axis under the colder air aloft may accelerate tornado development processes in some way, particularly in well-organized surface and synoptic settings such as this one.

The 500 mb isotherms superimposed on the surface-based CAPE field (below) also suggest that the coldest air aloft overran the instability axis mainly on its northwest extent, probably enhancing local buoyancy and potential for vertical stretching in the localized area near the surface "focus" (northwest Kansas) compared to areas more distant to the southeast (e.g., south central Kansas).  This seems to be an important thermodynamic factor in events with tornadoes that are associated with 500 mb cold core lows.  In comparing surface-based with mean parcel CAPE (also shown below), notice how MLCAPE more than halves the instability, a typical problem in using lowest 100 mb mean parcels in 500 mb cold core events.

The tornado near Wakeeney that I photographed below (looking northeast from near Collyer KS) appeared to occur with a cell at the northwest tip of the warm sector in under the cold air at 700 mb and 500 mb.  This was immediately east of of the warm front/Pacific front intersection discussed earlier, as Jim and I drove east out of the cold air north and northwest of the occluded front.  As the Wakeeney storm (see radar image a few minutes after the tornado photos) moved north and crossed over into the cold surface air, it appeared to become rapidly elevated and ceased producing tornadoes.  The next storm to the southeast near Ogallah also produced a widely-photographed tornado until it moved north out of the warm sector into colder air (see web sites of Mike Umscheid, Melanie Metz, and Eric Nguyen, among others).

It is important to remember that with most supercell tornado events occurring near 500 mb cold core lows, parameters used for forecasting "typical" supercell tornadoes will be less than impressive, as shown in the SPC graphics below on April 10.  Low-level storm relative helicity (0-1 km SRH) appeared adequate but not impressive (largest values were northwestward in the cold surface air), and composite parameters involving SRH/CAPE combinations appeared insignificant where tornadoes occurred during the hour following 21 UTC.  This is largely because total CAPE values are relatively small in 500 mb cold core tornado events, probably masking the enhanced low-level thermodynamic environment at the northwest extent of the narrow warm moist sector.  It may also be that SRH/CAPE environments in such cases are so localized as to not be handled well by model forecasts (the SPC mesoanalysis graphics are based on 1 hour RUC forecasts).

As the synoptic scale dry intrusion and Pacific front advanced east and northeastward, overtaking the warm front, the surface "focus" boundary intersection and northwest "tip" of the warm moist sector slid eastward along and near I-70 into the Hays and Russell area where tornadoes occurred from 2 supercells around 23 UTC.  The Russell tornadic storm that produced a multiple-vortex tornado is visible as a prominent overshooting top on the satellite photo below.  Other storms developed southward along the Pacific front/dryline into south central Kansas, but were not tornadic (see SPC storm reports graphic below) and took on a more linear orientation (see satellite photo).  Whether storms further southeast along the Pacific front/dryline and further from the cold air aloft in 500 mb cold core low cases become tornadic may depend largely on the orientation and intensity of the midlevel flow.   In the SPC graphic below showing 850 mb (low-level) and 500 mb (midlevel) winds, notice that the flow was very unidirectional in south central Kansas as opposed to northwest Kansas, due to the orientation of the surface system and the shape and wind pattern around the midlevel closed low.  In contrast on March 21 in Oklahoma, supercells became tornadic in eastern Oklahoma southeastward along the dry intrusion away from the 500 mb low, possibly due to the smaller/rounder shape of the 500 mb low that day which resulted in midlevel winds crossing perpendicular to the Pacific front/dryline further to the southeast of the 500 mb low, rather than parallel to it as in the April 10 case over south central Kansas.

In summary, the April 10 case involved a well-organized 500 mb cold core low system that matched the essential elements from Davies and Guyer (2004) for tornadoes occurring near 500 mb cold core lows.  Apart from the specific orientation of the surface and midlevel lows, it matched a composite (below) that I've been using in presentations over the past few years that hints at important ingredients:

        1)  A surface"focus" area or boundary intersection at the northwest extent of an axis of adequate moisture (low-mid 50s F surface dewpoints or greater) within roughly
                        200 miles of a closed 500 mb cold core low center (warm moist air under very cold air aloft to provide instability and stretching in an area of potentially enhanced horizontal                           and/or vertical vorticity) 
        2)  A surface dry intrusion and heating axis of steep low-level lapse rates that points into the surface focus area and moist axis (generation of locally enhanced low-level CAPE).
        3)  Development of storms in and close to the surface focus area, where tornadoes are most likely. (Note:   "mainstream" supercell parameters typically will do little to highlight                          the tornado threat in this area.) 

Some other comments and impressions:

-  When the dry intrusion is prominent and involves very dry air as in the April 10 case, this may somehow help storms to be more prolific in producing tornadoes, and may also be one reason that tornadoes associated with 500 mb cold core lows appear to be more common in the plains adjacent to a source of dry air from elevated terrain to the west and southwest.
-  For whatever reason (possibly mixing out and lessening of the dry air intrusion as it works eastward on following days), cold core midlevel lows generally don't seem to be good tornado producers on subsequent days as the system moves east (e.g., April 11 in northeast KS following April 10 above).
-  Cold core 500 mb lows that are positive-tilted and/or are  moving southeastward appear to be least likely to produce tornadoes, possibly because these characteristics result in a less favorable orientation of the midlevel low and cold air aloft relative to the surface moisture axis. 

Hopefully this case study will be useful to others in summarizing and using ingredients from operational products that can help highlight potential for tornadoes that may occur near 500 mb closed cold core lows.  The potential for tornadoes in such cases can sometimes be overlooked because moisture and thermodynamic factors may at first glance appear limited.

Special thanks to Jared Guyer (SPC) and Al Pietrycha (NWS GLD) for helping to provide graphics for this case study, and NWS Dodge City and Tyler Ummel for providing an image.

-  Jon Davies  4/17/05