New Severe Storms Conference paper online

I won't be able to go to the 24th Conference on Severe Local Storms in Savannah, Georgia later this month. But I have submitted a paper for the online preprint, titled "Three Strong Tornadoes in 2008 associated with Boundary Intersections and Narrow Instability Axes near 700-mb Lows". The final version is now online HERE.

The paper focuses on 3 strong tornado events (see photos above) that were difficult to forecast in 2008:

- May 1 in northwest Iowa (F2 tornado near Rock Valley)
- May 22 in Colorado (the F3 Windsor tornado, with 1 death)
- June 6 in north-central Minnesota (F2 and F3 tornadoes near Park Rapids)

These had some features and ingredients that were similar to so-called "cold-core" events, but probably wouldn't be considered as such using a rigid definition. While the paper is not anything "earth-shattering", I hope some people find the case studies useful.

- Jon Davies (updated 10/22/08)

High Plains presentation now online

Andy Fischer of NOAA/AWC Kansas City and I submitted a paper to the 12th Annual High Plains Conference (Hays, Kansas Sept. 4-5), "Significant Nighttime Tornadoes in 2008 Associated with Relatively Stable Low-level Conditions". I could not attend, so Andy presented the paper. Andy's presentation is now online in PDF format here.

The presentation focused on the setting and environment with the Beloit-Jewell-Belleville tornadic supercell in northern Kansas on 29 May 2008, and the Salina-Chapman-Manhattan supercell in central/northern Kansas on 11 June 2008. Both storms were associated with unusually large CIN for such intense tornadoes. MLCIN from lowest 100-mb mixed-layer lifted parcels was large for both events, probably between -120 and -170 J/kg, depending on the computer model used. This seems very large for significant tornadoes based on my database study from a 2004 paper in Weather and Forecasting.

What we found was that 0-1 km storm-relative helicity (SRH) was also unusually large (500-800 m2/s2) for these two supercell tornado events, particularly on 29 May. When combined with moderate total CAPE (at least 2000 J/kg) and strong deep shear (at least 55-60 kts), it appears that the environments for these events supported and enhanced intense mesocyclones to the extent that they were able to overcome the stable near-ground layer to generate tornadoes. Andy's presentation suggests that these combined ingredients (unusually large SRH, strong deep shear, at least moderate CAPE) can definitely support tornadoes in large CIN warm sector environments, and should be noted carefully by meteorologists, even when CIN suggests that an environment is not strongly surface-based. I'm hoping to post some addiitonal material on these events in the near future.

- Jon Davies 10/5/08

Tornadoes on 9/12/08 near Kansas City

A somewhat rare September tornado episode occured on Friday 9/12/08 in the Kansas City area. While visually impressive (see photos above near Eudora & Desoto, Kansas), the tornadoes were weak, with damage only in the EF0-EF1 range (the National Weather Service has officially rated the tornadoes near DeSoto, Kansas and Sedalia, Missouri EF0 in intensity).

Shawna and I had a birthday celebration planned in north KC for her son Zach on Friday evening, so we missed alot of the excitement. When picking up Zach from school in Plattsburg, Missouri at mid-afternoon, Shawna and I noticed SE to NW bands of low dark clouds with rapid motion, suggesting strong low-level wind shear. With rain all day and a cool surface air mass north of a boundary to our south (see surface map above), I didn't give the setting much thought until Shawna saw what looked like a wispy funnel that lasted a few seconds under a cloud band in the distance. I looked at some maps online, and began to wonder a bit what might happen. SPC issued a tornado watch around 4:30 pm with tornado reports coming in to the southwest of Kansas City near the boundary, but we were committed to our celebration. Watching TV later at Shawna's parent's house, we saw how large the tornado near DeSoto was, and heard about reports of damage. Thankfully, the tornado warned cells that went over the KC metro area did not produce any damaging tornadoes.

The DeSoto, Kansas tornado was a classic case of a supercell crossing a boundary where wind shear increased rapidly, with easterly surface winds. The surface map above shows the stationary front in the Topeka-Olathe area, and the SPC maps above show the large increase in low-level wind shear (storm-relative helicity or SRH) along and north of the front, with CAPE near and south of the front. The best combination of CAPE-SRH was right along the front just south and southwest of the KC metro area. Deep-layer shear through 6 km above ground (not shown) was also on the order of 40 kts along and north of the front, suggesting good support for tornadic supercells. On radar images above, notice how the DeSoto cell (indicated by white arrow) crossed the boundary moving northeastward, which is about the time it produced a tornado. As this cell moved deeper into the cool surface air north of the front, it stopped producing tornadoes.

The RUC analysis profile at Olathe (OJC, also shown above) is interesting, showing a large veering wind profile in the lowest 1-2 km, and around 900 J/kg of CAPE. While the instability wasn't super-impressive, notice that it was bunched low in the profile, with the "fattest" CAPE roughly 13,000 ft above ground. Most warm-season tornadic thunderstorms average larger CAPE, with the "fattest" CAPE up around 24,000-30,000 ft above ground. With the CAPE in the Olathe profile "squeezed" down fairly low, that suggests more rapid movement of air accelerating upward, like a hot air balloon encountering cold air aloft and moving upward faster. This could help with upward stretching, and combined with the wind shear, might be another factor to help increase potential for a tornado in a "smaller CAPE" environment.

- Jon Davies 9-14-08

Denver area landspout setup on 8-24-08

With the Democrats tooling up for their convention this past weekend in Denver, weather tried to steal the show for a bit with a long-lived landpout tornado (see picture sequence courtesy of KUSA above) just southeast of the Denver Metro area.

The setting was fairly typical for landspout tornadoes (a form of non-mesocyclone or non-supercell tornado) near Denver, with a boundary NE-SW across the metro area (see surface map above). Storms formed along this boundary just south of Denver between 4:00 pm and 5:00 pm MDT. Then another sharper southeastward-moving boundary moved off the foothills to the west (see radar image above), and collided with the storms over the southern metro area. These colliding boundaries and the storm updrafts stretching the vorticity ("spin") along them produced the long-lived landspout tornado over northeast Elbert County between 5:30 PM and 6:00 pm MDT, as well as a couple other briefer landspouts elsewhere along the line.

A sharp windshift boundary (or colliding boundaries in this case) is the most important component in landspout cases, providing the vertical "spin" to be stretched into a tornado. But the environment contributes as well, with strong surface heating and rapid temperature change (steep lapse rates) in the lowest 2 km or so, helping to accelerate air parcels upward in low-levels to augment stretching by storm updrafts along the boundary or boundaries (see Davies and Caruso 2005 at http://www.nwas.org/ej/cardav/). Such environments are similar to those that encourage dust devils on hot days.

Above, a RUC analysis profile in the southeast Denver metro area shows these very steep low-level lapse rates, and also the deep well-mixed moisture layer, even though cloud bases were high (MLLCL around 2500 m AGL). I used to think that some low-level CAPE (e.g., CAPE below 3 km AGL) was needed in these events, but I've seen several Colorado events in recent years where this just isn't the case. It seems that a deep well-mixed moist layer (to reduce parcel entrainment and drying) with little if any CIN (to let parcels rise rapidly) is what is really needed, along with the steep low-level lapse rates, and of course, the boundary or boundaries. These ingredients thrown together in the same area increase the chance of a mesoscale "accident" that might result in a landspout tornado. With the local topography around the Denver area that can help set up boundaries on days with steep lapse rates, such events are relatively common there.

- Jon Davies 8-26-08

Niagara Falls trip & photos

This isn't directly related to weather, but Shawna and I went to Niagara Falls last weekend for a very pleasant trip. The sky was picturesque with lots of bubbly cumulus clouds against crisp blue, courtesy of an upper trough with cold air aloft moving through. With that as background, the falls were truly beautiful!

Above are some photos. The first is the American side of the falls, with the city of Niagara Falls, New York, visible. But I think the Canadian side of the falls is more magnificent, in the remaining shots. Shawna, as seen in the second image above, agrees. Enjoy!

- Jon Davies 8/17/08

6/11/08 Little Sioux-Moorhead, Iowa tornado path revisited - solving a "mystery"

The Little Sioux campground EF3 tornado was one of the more publicized tornadoes in 2008 because of 4 boy scouts killed when it destroyed a bunk house where they had been drilled to take shelter (see my earlier post in June). Near the end of the 14 mile track (see first map shown above), this tornado struck storm chasers Kelly Martinson (also known as Kory Hartman) and Kenny Allen as it crossed Highway 183 a couple miles southwest of Moorhead, Iowa, yielding the most impressive up-close example I've ever seen of what it is like to encounter a rain-wrapped tornado. Contrary to some speculation by other chasers online, Kelly and Kenny did not intend to get that close, as the tornado was embedded in rain and the Omaha radar went down near the time of their encounter, making it difficult to judge where the tornado was as they waited for the mesocyclone to approach from the southwest. Also above are images of the same rain-wrapped supercell near Moorhead (observed by storm chasers from Iowa State), and later northwest of Dunlap (from the viewpoint of Shawna and myself).

The ground survey by NWS Omaha the day after the tornado suggested that the tornado did not cross highway 183 southwest of Moorhead, but Kelly and Kenny's video and their personal survey information indicated that it did (see the second map above, a Google maps closeup of the path end showing the discrepancy between the 2 survey tracks). To solve this mystery, Shawna and I revisited the area southwest of Moorhead on a trip to Iowa about a month after the tornado. Using a reference copy of Kelly and Kenny's video provided by Doug Kiesling (see video grab insets on second map above), Shawna and I located where the tornado crossed the highway (see third map above, my own survey of the tail end of the tornado track). From fence damage just west of highway 183 and south of 314th street (see photo above), it was clear that the most intense part of the tornado (50-100 yards wide) did indeed cross the highway where Kelly and Kenny's map indicated. We also found the two broken road signs that are seen being snapped off in Kelly and Kenny's video. Without the video the day after the tornado, it would have been easy to miss this damage, as the fence is located below and some distance west of highway 183 in a treeless area, and the broken signs were hidden in tall grass. Tree damage along and north of 314th street (what was shown as the center of the tornado path in the official survey) appears to have been a result of the north edge of the broader tornado circulation, or possibly inflow winds coming into the tornado from the north or northeast (see my close up map above). This example goes to show that surveying damage is difficult, and not always an exact science, particularly in rural areas.

Also, important to note: The intensity of the low-to-the-ground fence damage lined up with highway 183 strongly suggests that, had Kelly and Kenny been located another 50 yards southwest on highway 183, their vehicle would most likely have been rolled off the highway down a steep embankment, possibly resulting in injuries (the swirling condensation marking the most intense part of the tornado can be seen passing just south of their vehicle on the video). So, they are very lucky... as is the occupant of another vehicle (tail lights seen on the video) a few hundred yards farther down the highway (the most intense part of the tornado passed between the two vehicles). This is a graphic illustration of how dangerous it is to chase high-precipitation supercells with rain-wrapped tornadoes, as well as the potential problems of relying entirely on radar information when you can't see very well.

- Jon Davies 8/3/08

Nontornadic supercell in northeast Kansas - 7/24/08

It's been awhile since I've posted... I've been enjoying (and getting adjusted to) married life!

Last Thursday evening (7/24/08), Shawna and I storm chased in northeast Kansas during what appeared to be a fairly promising tornado setup for July. SPC even issued a tornado watch extending southeastward across the Kansas City area (see the insets on surface map graphic above). But, apart from a supercell along the Nebraska/Kansas border that rapidly gusted out and moved southeast into a line (see our photo images above), not much happened, and there were certainly no tornadoes. Why?

One can only speculate. Here's some thoughts...

Thunderstorm outflow from morning storms over Missouri had reinforced a stationary front/boundary during the day over northeast Kansas that was quite evident at late afternoon and early evening (see the surface map above). Even though wind shear was strong with easterly winds north of this boundary, temperatures were 10-15 degrees F colder on the cool side of the boundary in extreme northeast Kansas and southeast Nebraska where the original supercell developed between 5 and 6 pm CDT. Radar images above show this storm at 6 and 7 pm CDT, respectively, and its location is also marked with a circled "S" on the surface map above.

At Falls CIty, Nebraska (FNB), located just east of where the supercell formed, a RUC analysis profile at 6 pm CDT (shown above) indicated a good wind profile for tornadoes (a clockwise looping hodograph) and good CAPE (instability). BUT, the low-levels were rather stable (blue area on the RUC profile) beneath the CAPE area (red area on the RUC profile). Because tornadoes are a surface-based phenomenon involving near-surface air parcels, this low-level stable area with the cooler surface air may have been a negative factor, particularly because the supercell stayed north of the boundary.

I've noticed (as have other researchers) that when supercells form and _stay_ on the cool side of a significant boundary, they are less likely to produce tornadoes than if they _cross_ the boundary from warm to cool, or develop and stay within the warm sector (if there is significant warm sector wind shear, too). It may be notable that a couple storms that produced surprise tornadoes in north-central Missouri on 7/21/08 (no, we weren't there) occurred on the warm side of a prominent boundary. So the location of the supercell relative to the cool side of the boundary may be one clue in this case as to why no tornadoes happened. In addition, the cooler, heavier surface air on the cool side of the boundary may have made it easier for the storm's outflow to gust out, undercutting the mesocyclone and setting up a more linear configuration. But that's also speculation.

Anyway, it made for an interesting storm chase not too far from home (N of Kansas City).

- Jon Davies 7/26/08