Rare EF2 tornado in Oregon 12/14/10

Yesterday (Tuesday 12/14/10) saw an EF2 intensity tornado strike just southeast of Salem, Oregon shortly before noon PST (2000 UTC) at Aumsville (see photo above), with 2 injuries reported. It's been 17 years (1993) since a tornado that strong occurred in Oregon, and it was only the 4th tornado of F2/EF2 or stronger intensity to occur in the state since the early 1970's. Such events are so rare that they are typically unforecastable, as they usually occur in settings with small CAPE (<> 40 deg F) from western Washington into west central Oregon, and some spotty sun's heating on satellite. Although SPC mesoanalysis graphics showed total CAPE amounts not really registering (less than 250 J/kg, not shown), some low-level CAPE (SBCAPE below 3 km AGL) was indicated over the same area where dew points were maximized (see red areas in 4th graphic above at 1900 UTC). The small tornadic cell formed rapidly around 1910 UTC and began producing a five-mile track tornado around 1945 UTC (see radar inset on 4th graphic).

As is often the case in small CAPE settings, RUC soundings tended to under-estimate the environmental CAPE in this setting, showing surface temperatures around 45 deg F and surface dew points around 40 deg F. However, actual surface observations at Salem at 1900 UTC (not shown) before the tornadic cell formed were 47 F (temperature) and 42 F (dew point). When plugged into the RUC analysis sounding for Salem at 1900 UTC (last graphic above), these updated surface obs combined with cold air not far aloft (-10 deg C at 700 mb !) boosted total CAPE from around 100 J/kg to over 300 J/kg, with the "fattest" CAPE located in the 950-850 mb layer, very close to the ground. Even though total CAPE was unimpressive, this "packing" of CAPE down low in the sounding environment would likely help promote rapid vertical stretching in a local updraft, and given the accompanying sizable low-level and deep-layer wind shear, was able to support a tornadic storm as reported above. But picking out these subtle ingredients in a rapidly evolving environment is more than difficult, and everything apparently had to come together just right to generate a tornado in this case. Another similar setting might generate only storms with small hail. Adding to these problems, rotation in small cells (like this case) more than 25-35 miles from radar is usually not detectable for advance warning purposes.

Although it's not hard to go back and see contributing factors after a climatologically rare small CAPE event like this, detecting in real time these subtle ingredients that might make for a mesoscale "accident" in a location where tornadoes don't often occur is next to impossible.

- Jon Davies 12/15/10

Data collecting inside tornadoes - Will it improve warning lead times?

The Discovery television show "Storm Chasers" concluded another exciting season this past week. A repeating theme on the show is getting armored vehicles or probes into tornadoes (which makes for exciting storm chase footage) on the premise that this will help improve tornado warning lead times and save lives. But, in reality, that's really doubtful. Chuck Doswell's blog here discusses this issue very well. Measurements inside tornadoes may be useful toward understanding dynamics inside ongoing tornadoes (a valid goal), but will tell us next to nothing about which environments and storms will produce potentially deadly tornadoes. Chuck also points out that warnings for tornadoes most likely to cause deaths are already pretty good. Looking at four tornado days featured this season on "Storm Chasers" that each generated at least one violent EF4 tornado helps drive this point home rather well.

4/24/10 Yazoo City MS EF4 tornado (see photo by Dick McGowan above) - This tornado caused several deaths and was on the ground for 100+ miles in a setting that was well-recognized by NWS meteorologists as having high potential for tornadoes well in advance (for example, see experimental conditional sig tor probabiliities on 1st map graphic above). The specific warning for Yazoo City was issued approximately 30 minutes before the tornado hit town, and warnings for the same tornadic storm farther west had been ongoing for an hour. Can we really improve on those good lead times?

5/10/10 Tornadoes in central OK, including EF4 tornadoes just southeast of Oklahoma City - Again, the potential for this day was outlooked well in advance, with the coming together of environment ingredients quite evident (see 2nd graphic above). Tornado warnings in central Oklahoma were generally issued 30 minutes in advance of damage and deaths.

5/22/10 EF4 tornado near Bowdle in north-central ND - This setting was also quite evident in advance (see 3rd map graphic above), where a large tornado hit mainly in open country. A tornado warning for Bowdle was issued nearly 30 minutes in advance of the tornadic storm passing just northwest of town, and tornado warnings farther west had been ongoing for about an hour.

6/17/10 Tornadoes in MN, including the EF4 Wadena MN tornado - Another setting where tornado environment ingredients were very evident (see last graphic above) in advance, and most warnings were issued 30 minutes before sizable towns were struck.

Here are some representative values of individual bulk environmental ingredients associated with the above cases:

4//24/10 MLCAPE 1950 J/kg 0-1k SRH 670 m2/s2 0-6km shear 70 kts
5/10/10 MLCAPE 3340 J/kg 0-1k SRH 610 m2/s2 0-6km shear 62 kts
5/22/10 MLCAPE 3930 J/kg 0-1k SRH 345 m2/s2 0-6km shear 52 kts
6/17/10 MLCAPE 2525 J/kg 0-1k SRH 270 m2/s2 0-6km shear 54 kts

Any competent NWS meteorologist on days with environment ingredients like those above would be primed and ready to issue warnings immediately at the first indication on radar or credible information from spotters. The next level of improvement regarding saving lives will likely involve sociology to find ways to get people to hear and pay attention to those warnings already being issued.

The Discovery show is about 95% thrills and entertainment, and maybe 5% real science. But it makes interesting watching for legions of viewers, and I do commend the show producers this season for balancing the adrenaline quotient with some sober reality by showing the aftermath of tornadoes, including damage and injuries.

- Jon Davies 12/5/10

November funnels near Bakersfield CA

My friend composer/conductor William Stromberg (who also storm chases with his wife Anna Bonn) sent me the cool photos above taken by his brother Robert Stromberg (production designer and Oscar winner earlier this year for the film Avatar) yesterday morning while driving on Interstate 5 near Bakersfield, California (about 90 miles north-northwest of Los Angeles). These came out of some very low-topped convective showers associated with a potent upper storm system but limited instability, so I thought the setting was worth a look.

The 500mb forecast during the day on 11/8/10 (see 3rd image above) showed a strong upper trough (heavy black dashed line) moving across California and Nevada. Although a surface cool front (not shown) associated with this trough had already moved through the area, there was just enough moisture (surface dew points in the upper 40s F) and cold air aloft for a little sun-driven surface heating to pop some small showers (see satellite image above) in the Bakersfield vicinity, aided by forcing from the upper system.

A RUC model analysis sounding estimate for Bakersfield (BFL, see last image above) at around 16 UTC (8 am PST, near the time of the funnels), indicated a small pocket of CAPE/instability (possibly under-estimated by the RUC model) at around 3000 ft above ground, which is very near the surface. Apparently, enhanced stretching with the CAPE pocket so close to the ground, a steep lapse rate below cloud base, and the strength of the upper system combined to generate these "cold air" funnels within the updraft columns of 1 or 2 of the low-topped showers. As you can see, you don't always need alot of instability to get interesting things to happen with some fall and winter systems moving across southern and central California.

I didn't see any funnel or tornado reports online with the National Weather Service, so it may be that Robert was the only one to photograph these funnels. Many thanks to Robert and William for sharing these photos of an event that would otherwise have probably gone undocumented!

- Jon Davies 11/8/10

Late October EF2 tornado at Rice TX

There's been no shortage of interesting tornado events the past couple months, with a tropical system producing tornadoes in the Dallas TX area and a bow echo type tornado in the Queens/Brooklyn area of New York City in September, and damaging tornadoes in the high country of central Arizona in early October. I've been busy, so no case study posts on those events. Last Sunday saw a photogenic October EF2 tornado at Rice TX, southeast of Dallas and just north of Corsicana TX (see storm chaser Bridget Geaughan's photo above). Video shown on CNN from a range of about 100 yds from the tornado was quite impressive. Although there's nothing terribly unusual about an October tornado in Texas, I thought I'd at least post some brief material about the setting to help make up for lack of posts in recent weeks.

Probably the most interesting aspect of Sunday's afternoon setting is how quickly the low-level shear increased during the afternoon in response to the strong mid-level trough (see 2nd image above at 500 mb) moving west to east across Texas that helped fire up the storms ahead of a pre-frontal surface trough and wind shift (see surface map, 3rd graphic above). The SPC mesoanalysis showed 0-1 km storm-relative helicity (SRH, a measure of low-level wind shear, see 4th graphic above) at 2100 UTC / 4 pm CDT to be in the 100-150 m2/s2 range in the Rice TX area, values somewhat marginal for support of supercell tornadoes. However, by 2300 UTC / 6 pm CDT, about the time of the tornado, SRH values had doubled (> 250 m2/s2, same graphic) in response to the approaching mid-level trough, increasing the combination of SRH and MLCAPE (around 3000 J/kg) that could help generate low-level mesocyclones with stronger supercells (energy-helicity index near 3.0 and above by 2300 UTC, same graphic).

The RUC analysis profile at Corsicana TX, a few miles to the south of Rice at 2300 UTC during the tornado (see last graphic above), suggested strong combinations of parameters for supporting supercell tornadoes, even with a surface wind slightly west of due south. So a strong tornado occurring with a discrete supercell in this environment was not a surprise; SPC had a tornado watch issued well in advance, and a tornado warning was issued by NWS Ft. Worth based on radar roughly 15 minutes prior to the tornado.

- Jon Davies 10/26/10

Wind damage (and mesovortex) near my home 8/20/10

On Friday 8/20/10 while working at home (north of Kansas City), an intense squall line came through and caused an west to east swath of wind damage about two miles south of where I lived. The photos above show sections of corn crops flattened and some tree damage in that swath, damage I wasn't aware of until the next day when my wife Shawna noticed it while driving to Kansas City. There's really nothing unusual about this, except it seems the swath was associated with a forward "notch" in the line on radar, suggestive of a surface circulation with strong winds about 1-2 miles across called a mesovortex. The Kansas City NWS office at Pleasant Hill MO issued a tornado warning (based on this feature) a few minutes after it passed to the south of my house, and wind reports of 70-80 mph winds were verified in Holt MO, about 15 miles to my east. It's not too often a feature like this comes close to your home, so I thought I'd post a little documentation here.

NWS issued severe thunderstorm warnings for Clinton/Platte/Clay counties of MO well before the squall line passed my house. My photos above show views to the south and west as the line approached my home, with a bowing dark cloud shelf to my south and a ragged chaotic cloud base to my west just north of the "notch" on radar (I wasn't aware of the radar feature or damage until later). The forward flank "notch" can be seen on the radar base reflectivity images above (arrows), as well as the low-level base velocity image (circle). It was difficult to pick out a rotational "couplet" with this feature on storm-relative radar velocity images from the Pleasant Hill NWS radar (not shown). However, according to Evan Bookbinder at the NWS office, the closer and more favorably located Terminal Doppler Radar near MCI (KC International Airport, not shown) indicated a pronounced couplet with 80-90 mph winds toward the east on its south side as the feature approached Holt MO, prompting the tornado warning around 1905 UTC. The wind damage south of my home (a swath of 70-80 mph winds) clearly lined up with this radar feature before the tornado warning as it tracked eastward.

A surface map is also shown above, showing the squall line location after 1800 UTC, and surface temps in the 90s F ahead of the line with dew points in the mid-upper 70s F. The early afternoon environment over northwest Missouri (not shown) was quite unstable, with CAPE values near 4000 J/kg, around 30-35 knots of deep layer shear (0-6 km AGL), but little low-level shear (storm-relative helicity less than 80-90 m2/s2). So, the CAPE/shear setting was supportive of strong organized storms (possible supercells) with potential for wind damage, but suggested little support for much in the way of tornadoes. Forward flank squall line circulations in such settings can certainly produce notable wind damage, and even occasional weak/brief tornadoes of non-supercell origin.

- Jon Davies 8/22/10

Rare deadly tornado in Montana - 7/26/10

Killer tornadoes are rare in Montana due to the sparse population. Prior to 2010, only 2 deadly tornadoes had been recorded in the state, with the last tornado deaths occurring way back in July of 1983. But yesterday (Monday 7/26/10), sadly, 2 people were killed and another seriously injured when a large tornado struck a ranch in the northeast corner of Montana west of the town of Reserve (see photo and radar images above). This tornadic supercell moved east-southeast from Canada, deviating rightward under westerly flow aloft.

The surface map at early evening (above) showed a surface low over northeast Montana with easterly upslope winds along and just north of a stationary front where surface temperatures (80s F) did not cool too rapidly across the WNW-ESE boundary, and dew points feeding the storm were in the mid to upper 60s F. RUC model soundings had a little trouble depicting the moisture depth in northeast Montana near this boundary (moisture too dry/too shallow just above the ground, not shown) and wind profiles (hodographs a little "flat" in shape, not shown). However, the SPC mesoanalysis effective-layer sig tor parameter (STP, see graphic above) still suggested a "favorable" environment not too far to the east over extreme northwest North Dakota, likely feeding into northeast Montana on the upslope winds near and north of the stationary front.

A RUC analysis sounding at early evening near Reserve MT (last image above) had to be modified for moisture depth and the easterly wind pattern in order to be reasoanbly represnetative of the environment. This modified profile showed excellent deep layer shear (60 kts), good MLCAPE (2800-3000 J/kg), and workable low-level storm-relative helicity (SRH, > 150 m2/s2). Even though the low-level hodograph/wind profile wasn't large or notably strong, there was a "kink" in the profile and enough low-level shear to support strong tornadoes along and near the boundary. It appears this tornadic storm was able to parallel the boundary, taking advantage of the enhanced low-level wind environment there, along with better moisture just north of the stationary front without too much temperature contrast across it. Tornado warnings from the Glasgow NWS blanketed the area traversed by the supercell well in advance of it.

- Jon Davies 7/27/10

After-dark EF4 tornado in NW Iowa 6/25/10 - a look at the environment

Violent (EF4-EF5) tornadoes are rather rare after dark in the Plains. Last Friday's tornadoes during 03-05 UTC 6/26/10 after dark in northwest Iowa (see video image above) included an EF4 intensity tornado near Sibley, Iowa, which certainly makes the environment worth examining.

The tornadoes occurred just ahead of a cool front moving slowly southeastward (not shown) and an approaching upper trough (not shown), associated with the west end cell of a line sagging southward over northern Iowa (see SPC mosaic radar overlays above). The 03 UTC SPC mesoanalysis of 0-1 km energy-helicity index (EHI; see 2nd image above), which combines MLCAPE and 0-1 km SRH, showed very large values (near 10!) ahead of this west-end supercell over northwest Iowa, suggesting strong potential for low-level mesocyclones and possible tornadoes with discrete or tail-end supercells. The SPC mesoanalysis of low-level CAPE (0-3 km MLCAPE, see 3rd graphic above) at the same time also showed a notably surface-based setting over northwest Iowa, with low-level MLCAPE values near 100 J/kg. From a 2009 NWA electronic journal paper by Davies and Fischer, environments with large CAPE and SRH that are also relatively surface-based offer important support for tornadoes after dark, but aren't that common in the Plains due to nighttime cooling beneath typical warm layers advected eastward from the desert southwest. This latter issue did not appear to be a problem last Friday evening over northwest Iowa.

The 03 UTC RUC analysis sounding at Sheldon, Iowa (just south of the tornadic tail-end supercell; see 4th graphic above) confirmed large MLCAPE and SRH (near 4000 J/kg and 500 m2s2) and sizable deep layer shear (> 40 kts), along with a relatively surface-based setting (0-3 km MLCAPE > 100 J/kg, and MLCIN around -50 J/kg), all excellent supporting ingredients for tornadoes with supercells in that environment. In contrast, 110 miles to the east-southeast with another tornado-warned embedded supercell near Clarion, Iowa, the RUC sounding (see last graphic above) showed _no_ low-level CAPE, and MLCIN was quite large (-250 to -300 J/kg) with not nearly as much total MLCAPE above the CIN layer (only around 1600 J/kg),. This suggested somewhat "elevated" nighttime storms with considerably less tornado potential, and no tornadoes were reported in this area farther east.

NWS Sioux Falls has a page detailing their survey of the EF4 tornado near Sibley, Iowa, north of Sheldon.

- Jon Davies 6/30/10

Matt Hughes

This has been a sad couple weeks with Wichita chaser Matt Hughes (of Discovery Channel's Storm Chasers show) hospitalized in serious condition, and then passing away this past Wednesday. I am so very sorry for his family's loss.

I haven't known Matt for very long, but it is clear he was a bright and enthusiastic person, with much passion, and also very sensitive.

With the deadly Yazoo City MS and Oklahoma City OK tornadoes this year, and now Matt's death, I hope these events will serve as reminders to weather enthusiasts that there are more important things in life besides storm chasing, and to keep all that in its proper perspective.

Matt, we will remember you fondly...

- Jon Davies 5-28-10

Brief "chaser convergence" thoughts, and Missouri cold core tornado setting 5/20/10

Wednesday's massive chaser convergence and careless driving in central Oklahoma has been a huge topic of conversation the past couple days. Shawna and I actually decided not to chase that day because of the expected chaser hordes and narrow severe focus in central Oklahoma, and we both had work to do. Sometimes one asks, "If there's already hundreds of people out there shooting the same storm, what's to learn and what's the point of being out there adding to the dangerous traffic jams?" Wednesday was just such a day. Things don't look to improve in similar future situations, unless someone gets killed in traffic and it is well documented. Shawna and I are getting increasingly picky about going out on "big" days with a small focus drawing hundreds of chasers and weather yahoos to the same spot.

Anyway, stepping away from the situation in Oklahoma, Wednesday 5/19 brought some "cold core" action to the Garden City/Dodge City area of southwest Kansas in the form of "landspouts" along a surface boundary close to the 500 mb low, photographed nicely by Mike Umscheid. And Thursday 5/20 brought a few "cold core" supercell tornado reports to the Sedalia area in Missouri east-southeast of Kansas City (see photo above). I had family visiting in town, so no storm chasing for me or Shawna. But with my interest in such settings, the Thursday setup is worth documenting briefly. Thankfully, as with most "cold core" setups, none of the tornadoes were strong.

The 2nd graphic above shows visible satellite with overlaid surface and 500 mb features at 2145 UTC. Notice the boundary intersection in west-central MO, similar to the "cold core" tornado composite shown in this paper. This can be a favored area of severe focus (enhanced low-level shear and forcing) for tornado development when a midlevel low (strong cold air aloft) is nearby to the west (northeast KS on 5/20) and there is some instability. In this case, the N-S boundary wasn't a dryline, but a subtle wind shift intersecting the warm front southwest of Sedalia at late afternoon, notable on satellite along with some clearing for sun's heating. The 3rd graphic shows selected images from the SPC mesoanalysis at 2200 UTC , including the midlevel low, total CAPE (near 1000 J/kg over west-central MO, but certainly not as impressive as down in Texas), low-level CAPE (a significant maximum near the boundary intersection), and SRH (enhanced northeast of the warm front, but probably under-represented closer to the front in west-central MO). The presence of the 500 mb low in northeast KS and the low-level CAPE maximum over west-central MO near the boundary intersection were probably the biggest "heads-up" flags seen here for tornado potential in what otherwise looked like a fairly benign environment.

The 4th graphic above is a radar image at about 2340 UTC when the tornado in the photo above (likely EF0) was occuring near Sedalia (see arrow for cell location). The RUC analysis at 2300 UTC (last graphic above), as with many "cold core" type settings, showed most of the CAPE below 500 mb, suggesting sizable low-level stretching. The model-derived estimated hodograph (same graphic) suggested good clockwise curvature to the wind profile and more low-level shear than shown on the SPC mesoanalysis during the afternoon. So, while not an easily forecast event, the presence of these ingredients suggested taking rotational signatures on the nearby Pleasant Hill MO radar and spotter reports very seriously. Indeed, the Kansas City area NWS office did a good job jumping on the situation with tornado warnings and statements as the situation developed and evolved prior to the Sedalia tornado, which is what situational awareness is all about.

- Jon Davies 5/21/10

Monster Medford OK meso & tornadoes 5/10/10

My wife Shawna, her son Zach, and I intercepted the large supercell near Medford OK last Monday (5/10/10). As a rule, I don't chase storms going 50+ mph, but we decided to try it, given the high-end potential tornado environment (see significant tornado probability mesoanalysis graphic for 20z, 2nd image above, an experimental product that Bill Togstad in Minneapolis and I have been working on). We were prepared for very fast storm motions, probably giving us only a few minutes of intercept time.

There were many chasers on the Wakita-Medford storm, but we managed to avoid most of the crowds (and all the road "drama" south of Wakita) by pulling off on a county road 5 miles north of Medford and letting the dangerous storm come toward us. As it emerged from the haze and broken low clouds near Medford around 3:50 pm, the mesocyclone at first appeared to be rain-wrapped, but then we could begin to see features behind the rain curtains, including a narrow tornado on the north side of the meso (see 3rd image above, please excuse the poor video quality). We were using WxWorx due to some internet problems, and Threat Net was indicating that the meso would pass just to our northwest. Not so! It quickly became apparent that the storm was moving more rightward/eastward, and that we needed to move south FAST, out of the path.

It was unnerving to see a meso that large coming toward us at nearly 60 mph (a mile a minute!), and the pressure drop with the fast movement was causing our ears to "pop", the first time I remember experiencing this with an approaching mesocyclone. As we blasted south, the first tornado dissipated, and a larger one formed near the center of the meso, with rain and dust curtains moving around it. In fact, thoughout much of the meso, what appeared to be wispy "spin ups" and brief condensation columns were also occurring, causing us to wonder if much of the mesocyclone might actually be a very wide lower-end "tornado" a mile or more across, with more intense vortices moving around inside it (see 4th image above). There's certainly no clear-cut answer to that. Regardless, we needed to get out of the way.

As the larger cone tornado within the meso sped toward us (see 1st image above), we cleared the path of the meso and RFD winds to the south, and pulled over just north of Medford. Looking back to our north, the tornado(es) were no longer visible, as rain and hazy sunlight obscured our view (chasers farther east reported seeing only a rain-wrapped meso at this point). Looking to our west, we noticed an anti-cyclonic wall cloud and circulation on the trailing gust front (a feature not that uncommon), whch produced a very brief spin up beneath a small condensation funnel aloft (not shown). With the incredibly fast storm motion, our chase was over in only minutes, and Shawna's son Zach had videotaped his first tornado(es).

The last graphic above is a schematic and approximated path of this large fast-moving mesocyclone (several miles across), one of the largest and fastest I've ever seen. Despite the damage path from near Wakita OK to east of Arkansas City KS, it was great to hear that there were no deaths or serious injuries from this dangerous supercell.

For those on Facebook, additional video captures are under Shawna Davies' photo folder titled "5-10-10 Medford OK meso /tornadoes".

- Jon Davies 5/15/10

Dick McGowan's photos of 4/6/10 Wakeeney KS tornadoes

Dick McGowan (www.tornadolive.com) was kind enough to send me a few images from his Wakeeney KS chase yesterday, to compliment my earlier post about the meteorological setting for these unusual tornadoes (surface temperature mid 60s F, dew points mid-upper 40s F).

The first image shows the first tornado north of the interstate, with a condensation funnel well aloft and dust circulation at the ground. Dick says this tornado formed rapidly from a bowl-like high-based mesocyclone circulation and lasted maybe 2-3 minutes. The last image shows this tornado dissipating (right) with a 2nd tornado at left in what appears to be an occluded area. Dick says this 2nd tornado lasted about 3 minutes, and even hit a small farm structure, lifting pieces of debris into the air.

Looking back at the Hays RUC profile in my previous post, the high cloud bases and steep low-level lapse rate suggest to me that these tornadoes may have been a "hybrid" type of event, combining supercell and non-supercell processes, with significant heating at the surface (possibly upper 60s F) and very cold temperatures not far aloft (0 deg C around 700 mb). Dick says that he had the visual impression of updrafts that were shrinking horizontally and stretching very rapidly in the vertical as the tornadoes occurred. At any rate, these weak tornadoes occurred in an odd location relative to the surface pattern, away from the deeper moisture, and were the only ones I know of that were photographed over the past 4 days of severe weather in the Plains.

Dick, thanks for sharing these images.

- Jon Davies 4/7/10

Update 4/8/10 AM - I looked back at the surface analysis I did in the previous post, comparing it to Dodge City radar base reflectivity just before 2100 UTC on 4/6/10, and found I missed a boundary (Dick McGowan also mentioned this). I've drawn this in on the updated surface map (with radar) above, suggesting that the Wakeeney tornadoes occurred at or near a pretty decent boundary intersection, which may have added some nice vorticity ("spin").

Surprise tornadoes near Wakeeney KS 4/6/10?

Although not in yesterday's NWS storm reports for 4/6/10, storm chaser Dick McGowan (on his way to Oklahoma from Denver) happened to be in the right place at the right time, apparently photographing a couple tornadoes near Wakeeney, Kansas, just north of I-70. I haven't seen any pictures posted yet. but this was well northwest of the true warm sector, and in the vicinity of a secondary front moving across northwest Kansas. This wasn't a forecastable event, but it is worth looking at because it was certainly unusual.

The first graphic above is a SPC mesoanalysis estimation of 0-3 km MLCAPE and surface vorticity at 2100 UTC (4 pm CDT), with radar reflectivity overlaid. Notice the pronounced low-level CAPE maximum in the Wakeeney/Hays KS area. By itself, this doesn't mean much, but given its location far away from the warm sector over eastern Kansas, it suggests an area of cold air aloft combined with surface heating. The surface analysis just before 2100 UTC (2nd graphic above) shows an east-west frontal boundary in the area, with surface temperatures in the mid 60s F along this front (70s F further south) and dew points in the mid-upper 40s F. The RUC midlevel map analyses at 2100 UTC (3rd graphic above) definitely show cold air coming in aloft at 700 mb (0 deg C over northwest Kansas) and a closed 500 mb low over northern Colorado. So there was warmth at the surface, cold air aloft at 700 mb, and a frontal zone with just enough moisture in the local area. One probably wouldn't call this a real "cold-core" setting given the odd surface pattern and the distance from the midlevel low, but the cold air aloft with this system certainly played a role.

Visible satellitie and radar (4th graphic above) show the location of a small cell (possibly a mini-supercell) near Wakeeney, moving east, with plenty of clear skies (heating) to its south and east. This is near the time the tornadoes were reported. A RUC analysis sounding at Hays Kansas at 2100 UTC (east of Wakeeney, last graphic above), modified for more heating than the raw RUC file indicated, yields 200-300 J/kg of CAPE, all bunched below 500 mb, with decent wind shear, fairly steep low-level lapse rates, and small CIN. This is another one of those small CAPE soundings I've discussed recently, and the type of environment that would likely generate rapid low-level stretching within local updrafts. Given all the above ingredients, it's not too much of a "stretch" to see how the setting might support some weak tornadoes with the isolated cell over northwest Kansas. But it was certainly not an event that could really be anticipated.

What an oddball case! And as I write this, more cold air aloft from the same system is now generating a small late pm severe cell over north Kansas City, just behind a secondary surface cold front.

- Jon Davies 4/7/10