Wednesday, May 28, 2008

Deadly nighttime tornado in eastern Pratt Co. Kansas on 5/23/08

It seems like there have been an unusual amount of nighttime tornadoes in Kansas the past couple years, with the 2007 Greensburg event being the most notable. Last Friday evening (5/23/06) was yet another nighttime tornado that killed a married couple in their car shortly before 11:00 pm CDT on Highway 54 in Pratt County, Kansas east of Cairo (60 miles west of Wichita), not far from where I grew up. The map above, with radar inset, shows the approximate location of the tornado.

It is very unfortunate that the couple was driving from Colorado to visit family in the Wichita area, and had pulled off the road to wait out the storm. Unwittingly, in the the dark and rain and hail, they had parked directly in the path of a large tornado that blew their car at least 200 yards off the road, killing both. This event reiterates the importance of having a weather radio available in your car during tornado watches, and knowing your location to avoid driving into a potentially tornadic storm.

Shawna and I looked at the damage a couple days later while in the area. Pictures that Shawna took (above) show how mangled the vehicle was, and how far off the road this car was blown. In the first photo (looking northeast), the remains of the car can be seen with a large "scrape" in the foreground where the vehicle bounced. In the second image (looking southwest), Highway 54 can be seen in the distance behind the crumpled car, along with tree damage caused by this half-mile wide tornado. These photos are a reminder that vehicles are clearly a deadly place to be in a strong tornado. My informal survey suggested at least EF3 intensity, given tree and building damage in the area.

The local environment that evening was excellent for tornadoes after dark. The Medicine Lodge RUC profile (above) located south of the supercell at 10:00 pm CDT showed an extremely favorable for tornadoes, with CAPE > 3000, SRH > 500 (!), and EHI > 10.0 (!!!). In the last image above, the large 1/2 mile wide tornado can barely be seen, backlit by lightning, in a video grab courtesy of Brandon Ivey and The Storm Report ( Thanks, Brandon.

Shawna and I have thought about this couple and their final moments a great deal the past few days. It makes us very sad.

Sunday, May 11, 2008

Deadly tornado setting on 5/10/08 in northeast OK & southwest MO

Tornadoes in northeast Oklahoma and southwest Missouri late afternoon on Saturday May 10 killed at least 18 people. This deadly event was the result of a classic tornadic supercell setting with a strong weather system and storms moving rapidly (around 40 m.p.h). People in the town of Picher OK (see radar above) had around 15 minutes warning from NWS before a large wedge tornado (see Chris Wilburn's image, also above) plowed into town; sadly, 6 people were killed in Picher. The rapid storm movement and hazy conditions may have caught some people unaware. This was a very different and much more dangerous situation than the 5/8/08 western KS landspout (see my previous post) with a storm chaser running up to it, a point not properly highlighted or emphasized in chaser TV interviews following that event. I elected not to chase on Saturday due to the fast storm motion in poor visibility terrain that was more than 3 hours from my place north of Kansas City, and instead monitored the situation from home.

Shown above at 22z (5 p.m. CDT) are SPC mesoanalysis graphics of storm-relative helicity (SRH, low-level wind shear that can be tilted into storms for low-level rotation) and CAPE below 3 km (suggesting instability that is strongly surface-based). These suggested that the newly-formed Picher supercell (located at circled "s") with a storm-free area to its south was moving toward an area of strongly enhanced low-level wind shear within a strongly surface-based environment.

Also shown above at 22z is a graphic of 0-1 km EHI (energy-helicity index, combinations of SRH and instability), which, like the prior graphic, also confirmed potential for strong low-level rotation in the supercell moving across extreme southeast KS into northeast OK and southwest MO. The RUC analysis profile for Grove OK (located between Tulsa and Joplin) at late afternoon also showed a large clockwise wind profile (hodograph shown at upper left of EHI graphic), indicating large low-level wind shear in a surface-based environment with large CAPE as well as deep layer shear. So it came as no surprise that this supercell (see earlier radar) proceded to produce tornadoes over the next 2 hours on its trek into southwest MO.

Morning forecasts from both the NAM/WRF model (shown in last graphic above) and the RUC (not shown) did a reasonable job of indicating the supercell tornado potential for late afternoon. These showed a deepening upper trough moving through the central plains with midlevel wind flow of 50-75 kts spreading out (diverging to induce upward motion) over southeast KS/southwest MO/eastern OK/northwest AR, as well as strong CAPE/low-level shear combinations over the same area in bright colors south of a surface warm front (not shown). Large low-level CAPE (a surface-based unstable environment) was also indicated on the forecasts, which did a good job highlighting the strong tornado potential. Unfortunately, that potential was realized, resulting in more deaths than with the Greensburg tornado last year.

Jon Davies - 5/11/08

Thursday, May 8, 2008

Western Kansas landspout tornado setting 5/8/08

My friend photographer Jim Reed got some good images of a landspout (nonsupercell) tornado in Western Kansas between Tribune and Leoti near the town of Selkirk on 5/8/08 (see one of Jim's pretty shots above, along with a couple other local images courtesy of KAKE-TV). Thankfully, the tornado stayed in open country with little or no damage. The setting this day affords a good opportunity to look a typical high plains landspout setting.

Above is a surface map at about 2 p.m. CDT (1900z) showing a surface low near the KS/CO border, and a dryline boundary north-south through the low where some surface convergence was taking place. Notice, too, the surface heating axis that pointed into this boundary area northeast of the surface low, with temperatures mid-upper 70s F impinging on dew points in the low-mid 50s F east of the dryline. This type of setting with heat axis intruding on a moisture axis suggests potential for rapid storm development.

Satellite images shown above at 1:45 p.m. (1845z) and 2:15 p.m. CDT (1915z) show sunny skies and strong surface heating over western Kansas helping to explode towering cumulus along the boundary northeast of the surface low into a thunderstorm within 30 minutes or less (very rapid development). Shortly after the image at 2:15 p.m., the landspout tornado formed and stayed on the ground for about 20 minutes.

A graphic from SPC's mesoanalysis at 2 p.m. CDT (1900z, also shown above) depicts surface vorticity ("spin") and low-level CAPE northeast of the surface low in the area where the storm was forming. This graphic suggests that good vertical "spin" was available along the boundary where the storm formed, as well as a surface-baced environment with strong heating (notice I also drew in the heat axis in red from the earlier surface map). These ingredients are typical building blocks for landspout tornadoes: steep near-surface lapse rates helping enhance low-level stretching with developing updrafts directly over a surface boundary that has some good "spin" within a surface-based environment.

A modified analysis profile from the NAM/WRF model an hour or so before the storm formed is also shown above. This profile, modified temperature-wise in the lowest 1-2 km based on observed surface temperatures near the KS/CO border in the mid 70s F, suggests an excellent environment for supporting landspout tornadoes, if a storm or storms were to develop directly over a windshift boundary, as happened in this case. Notice that there was good instability (CAPE), little if any convective inhibition (CIN), and that the lapse rate in the lowest 2 km was very steep (around 9 deg C per km), similar to soundings associated with dust devils in elevated terrain on sunny days. This would allow for rapidly rising air parcels beneath the storm updraft on the boundary, providing focused stretching of vertical spin for a possible nonsupercell tornado.

A somewhat similar situation occurred in northwest Iowa on 5/2/08 with some of the same ingredients when a dusty EF-2 tornado occurred near Rock Valley. A big difference that day was the presence of strong low-level wind shear (absent on 5/8/08 in western KS at the time of the landspouts) that helped the Iowa tornado become a long track supercell tornado, in essence a kind of "hybrid" event.

Jon Davies - 5/9/08

Saturday, May 3, 2008

Bow echo tornadoes in Kansas City early a.m. on 5/2/08

The most damage of any of the storms that occurred last Thursday evening into early Friday morning (5/1/08-5/2/08) happened with a bow echo/derecho that moved through the Kansas City area around 2 a.m. CDT. After about a hour of sleep, Shawna and I awoke near Trimble to the loud roar of near-hurricane force winds and canon-like booms of thunder. Our TV and internet immediately went out, and our grill and patio furniture went flying off into the darkness.

Damage occurred scattered throughout the Kansas City area. Shawna's folks in North Kansas City had a large tree fall close to their house, damaging their neighbor's garage and car. A very interesting part of this event was a mesovortex that produced bow echo tornadoes near the apex of the bowing squall line in the Gladstone-Liberty area, causing a concentrated path of damage.

Bow echo tornadoes (also called mesovortex tornadoes or QLCS or quasi-linear convective system tornadoes) can occur within squall lines that have a bowing segment where air aloft is intruding and descending prominently from behind into the rear of the squall line (called a rear inflow jet or RIJ). Shallow circulations called mesovortices sometime form along the forward part of the squall line in front of this rear inflow jet and near or just north of the apex of the bow. These circulations can tighten up and produce tornadoes. Such tornadoes are typically short-lived and different from supercell tornadoes, but they are still potentially damaging. The first diagram above is a schematic showing how mesovortices and tornadoes from bow echoes can evolve.

Also above is an informal damage path survey above that I did on 5/2/08 in the Gladstone- Liberty area, showing a mesovortex that produced at least 3 tornadoes in sequence that I could find along an 8-mile narrow track of scattered wind damage. These tornadoes are difficult to pick out on radar, and if too much emphasis is given to detecting them, it can detract from the wider straight-line damaging wind threat across the bow echo in general. Unfortunately, these bow echo tornadoes on 5/2/08 were intense (EF-2 and EF-3 intensity, see this NWS link) and did significant damage to many homes. But fortunately, there were no serious injuries.

A close look at archived radar images suggests that the mesovortex formed rapidly near Gladstone and moved northeast over the next 10 minutes. At 0650z (1:50 a.m. CDT), a prominent rear inflow jet was evident and a subtle forward inflow notch was forming at the front and just north of the bow apex (locations marked on the 0650z reflectivity image above), suggesting the possible presence of a mesovortex. Subsequent images shown above at 0656z (1:56 a.m. CDT) and 0700z (2:00 a.m. CDT) showed this feature becoming more prominent and tracking northeast along the corridor where the mesovortex and tornado damage occurred from my informal survey. On radar velocity images (not shown), this feature was initially difficult to pick out as it formed near Gladstone, but then became somewhat more noticable as it passed northwest of Liberty and matured. Though not marked on the radar images above, another inflow notch suggesting a mesovortex (without a rear inflow jet behind it) could be seen along the bow farther southeast passing near the Independence area. This mesovortex was apparently nontornadic.

A similar mesovortex tornado occurred earlier west of Lawrence, KS just after 1 a.m. CDT from the same bow echo in an earlier stage, and another mesovortex tornado occurred later in southern Missouri (around 6 a.m. CDT) from a different bow echo segment that caused some damage in the town of Ava, southeast of Springfield. In both cases (not shown), the tornadoes were associated with a subtle inflow notch at the front of the squall line near or just north of the bow apex. Again, it should be emphasized that bow echo/mesovortex tornadoes are difficult to detect on radar (they are small and form quickly), and are generally impossible to see at all when located at distances more than 40-50 miles from the radar. This makes specific local tornado warnings by NWS meteorologists difficult to do, which is something the public needs to know and understand.

Jon Davies - 5/3/08

Southeast KS supercell chase 5/1/08

Shawna Helt and I storm chased in southeast Kansas Thursday evening before coming back to the Kansas City area ahead of the damaging bow echo that hit town with embedded tornadoes (see a later post that I'll get to after this one). With all the weather in the KC area that evening , including a supercell mesocyclone that passed very near our house NW of Trimble, we could have saved alot of gas had we just stayed put. :-)

Anyway, our chase in southeast Kansas in the Fredonia-Chanute area yielded some nice views of supercells and mesocylones, and one possible brief tornado. The first photo above shows a supercell with inflow bands and some lowerings viewed to the southwest from just east of Fredonia, KS around 7:30 p.m. CDT. The 2nd two images (one full, one zoomed) show a lowering and possible brief funnel or tornado in the distance from another supercell that had formed behind the earlier one. These last two images were a westward view from south of Altoona, KS shortly after 8:30 p.m. CDT. It was good to get out in the field on a real chase, but a little frustrating with the storms going on closer to KC and home.

Thursday was an interesting day. Why did the storm west of Tulsa in Oklahoma produce tornadoes, while those in southeast Kansas struggled to do the same? It's hard to say. Post analysis suggests that low-level shear and helicity may have been stronger farther south, but that's not clear. Also, for whatever reason, in the Fredonia area, new cells would form behind each supercell we watched and then appear to interfere with progressive structure and maturation of the older storm. Many questions. There was also a prolific "cold core" tornado event in extreme northwest Iowa east of the surface low and cold core midlevel low with many online photos of impressive tornadoes there.

After Shawna and I returned to the Kansas City area that night, the most damaging event of the day was about to happen. See my next post.

Jon Davies - 5/2/08