Case study: November 27, 2005 tornadoes in Kansas with
cold core 500 mb low - by Jon Davies
It isn't unusual for tornado events to occur in November in the eastern U.S. (remember 11/10/02, 11/6/05 and 11/15/05 in the Ohio Valley and Southeast), but Kansas is another story. In the last 20 years (since 1985) there have been only two November tornado days (in 1988 and 1998) in the state. So, the tornadoes in east central Kansas between Marion and Manhattan on 27 November 2005 would have to be called a rare event, prompting the case study below (click on yellow-bordered images for larger versions). Thanks go to Darin Brunin and Dick McGowan (from the Lawrence-Olathe KS area) for sharing their images and chase information.
The GFS model five days in advance (below) was suggesting a closed 500
mb low would be somewhere in the southern plains on 11/27/05, with possibly enough time
for moist return flow from the Gulf for severe weather:
<GFS 120 hr 500 mb and
surface forecast for 6 p.m. CST Sunday 11/27/05
By Saturday night 11/26/05, it looked like the central plains might get
into the act with the surface low forecast to be further north in central Kansas on Sunday
evening, and the 500 mb low aloft not far away to the west or southwest:
<Eta/NAM 24 hr surface and 500 mb forecasts for 6 p.m. CST Sunday 11/27/05
The Eta/NAM was also forecasting some CAPE (around 500-600 J/kg over
eastern Kansas, not shown), and a surface dry surge punching northward into an axis of
low-level CAPE (CAPE below 3 km) just east of the surface low:
<Eta/NAM 24 hr surface dew point
and low-level CAPE forecasts
The panels above suggested a surface pattern in association with a cold
core 500 mb low that might produce tornadoes (see this
link), even though surface dew points would only be in the low to mid 50's (F) east of
the surface low:
Moving to the day of the event, on Sunday morning 11/27/05, the 9 hr RUC
model forecast suggested that the surface low and dry surge might not be as fast as
forecast the evening before, with an arc of precipitation at mid p.m. from the surface low
near Wichita southeastward into eastern Oklahoma. RUC forecast LCL heights
(approximating cloud bases) suggested that the dry punch east and southeast of the 500 mb
low would be nosing into an axis of surface-based total CAPE and low-level CAPE over
south-central or central Kansas at mid afternoon along the Pacific front (acting as a
dryline):
<RUC 9 hr surface/precip and LCL
forecasts for 3 p.m. CST Sunday 11/27/05
<RUC 9 hr sfc-based CAPE & 0-3 km CAPE forecasts for 3 p.m. CST Sunday
11/27/05
As it turned out, the RUC model was slow in its depiction of the
evolution of these features by mid p.m. on Sunday. Surface maps and satellite images
at late morning and early afternoon below show that the surface dry slot with sunshine
surged northward rapidly from Oklahoma to a position north of Wichita before 2
p.m. CST (20 UTC):
<visible satellite 11:30 a.m. CST and surface map 11:00 a.m. CST Sunday
11/27/05
<visible satellite 1:45 p.m. CST and surface map 1:00 p.m. CST Sunday 11/27/05
At the time of the 1:45 p.m. CST satellite image above, low-topped
storms had already formed in central Kansas ahead of the dry surge, and a tornado was
already on the ground in eastern Marion County, 60-70 miles northeast of Wichita:
<photo
by Dick McGowan, courtesy Darin Brunin, looking WSW
Several different small low-topped cells produced tornadoes over the
next 90-120 minutes ahead of the northward-advancing dryline boundary near the surface
warm front. Here are additional images of other tornadoes observed by chasers
Darin Brunin and Dick McGowan in northeast Marion County and southwest Morris County:
(See Scott Currens page for
additional tornado images a little further north.)
<photos by Dick McGowan, courtesy Darin Brunin, looking N, then E
Topeka NWS has a summary and map of tornado tracks in Morris, Dickinson, Geary, and Riley counties here. Thankfully, none of the tornadoes were stronger than F1 intensity, and no one was hurt, although damage did occur at Fort Riley. Tornado warnings from the Wichita and Topeka NWS offices were timely and well-placed.
On the 1 p.m. CST surface map earlier, notice the general location of
the warm front and the Pacific front/dryline, as well as the position of the heat axis
(red dots) just behind the Pacific front/dryline. The tornadic storms occurred not
far from this boundary intersection at the nose of the heat axis where the juxtaposition
of surface heating with the moisture axis probably maximized local CAPE values,
particularly with the cold air aloft (500mb temperatures around -20 C). See the SPC
graphics below at 2 p.m. CST (20 UTC), showing the axis of steepest low-level lapse rates
(from surface heating) pointing into the moist axis (low-level CAPE over central Kansas N
of Wichita) with cold 500 mb temperatures aloft:
^ 0-3 km lapse rate, 0-3 km CAPE, and 500 mb winds/temperatures from SPC
mesoanalysis at 20 UTC 11/27/05
Because total CAPE values over central Kansas were relatively small (250-750 J/kg), parameters such as the significant tornado parameter (STP, not shown) and energy-helicity index (EHI, not shown) did not indicate an increased risk for tornadoes over central or east-central Kansas on 11/2705, which is typical of most tornado events associated with 500 mb cold-core lows. This is just a reminder that CAPE amounts in 500 mb cold core situations near the surface low may look small, but these systems can do alot with that CAPE.
After 3:30 p.m. CST (2130 UTC) the tornado threat began to shift
southeastward as the warm moist sector over east -central Kansas became increasing
"pinched off": 
<surface map 3:00 p.m. CST Sunday 11/27/05
The area just east of the Pacific front/dryline in southeast Kansas became more favorable for tornadoes, with a damaging tornado reported near Erie, Kansas around 4 p.m. CST (see NWS Wichita link).
I was working on other things on Sunday 11/27/05, keeping only one eye
on the weather. Putting too much trust in the morning RUC model positioning of
features seen earlier, I was unprepared for how quickly the surface dry surge jumped past
Wichita going northward. By the time I left Wichita after 1:30 p.m. (a misjudgement
on my part), the storms entering Marion county were moving north at 40 mph, and I could
not catch them. Highway 77 at Florence (completely closed for construction) ended
the chase for me. Here is picture of low-topped tornadic storms heading rapidly away
toward I-70, shortly after 2:30 p.m. looking north from Florence:
<photo
by Jon Davies
And so... I missed out on a weather event one won't see in Kansas this late in the year very often :-( . The combination of a well-defined 500 mb cold core system and adequate Gulf moisture (50's F dew points) that was in place over Kansas since the evening before made for an interesting and unusual late-year severe weather episode.
- Jon Davies 12/4/05