Sunday, March 29, 2009

Unexpected tornadoes in North Carolina on 3/27/09, while snowstorm raged in the Plains

It was a wild week of weather across the country, with flooding in North Dakota and a blizzard in Kansas (the town of Pratt where I grew up got 28 inches of snow!). Early in the week, there were tornadoes and injuries in Nebraska on Monday 3/23/09, and tornadoes again in Mississippi early Thursday morning with 20 injuries and many homes destroyed in the town of Magee. However, the event that really caught my attention was the tornadoes in North Carolina on Friday 3/27/09. It was unexpected (no severe outlook, watches, or even warnings early on), and involved several tornado reports over a 2 hour period with one tornado rated EF2 on the south and east side of Fayetteville, North Carolina (see photos above).

Unexpected events are always useful to look at to see if there are clues from which we can learn as forecasters. Looking back at Friday's case, there was a strong short wave and wind max aloft forecast to eject northeastward across the Carolinas (see 700 mb forecast map above), out ahead of the large blizzard-producing trough over the central plains. At the surface, there was a significant axis of dew points pointing northward into the Carolinas ahead of the shortwave (see circled area on the SPC dew point analysis, 3rd image above, at early to mid afternoon). On the same map, I've drawn in the estimated position of the freezing line at 700 mb. Notice how this colder air aloft was overlying the dew point axis over North Carolina. This would likely contribute to more significant instability in that area than one might detect on available total CAPE analyses (only around 250 J/kg, not shown).

In the 4th image above, I've drawn in the surface front on the 20 UTC SPC surface wind and pressure analysis, and also included the 20 UTC 850 mb map and 20 UTC analysis of estimated low-level CAPE below 3 km AGL. Notice how there was a low-level jet of 30-40 kts at 850 mb overrunning the surface front over northern South Carolina and southeast Nouth Carolina, ahead of the short wave aloft. This would provide lift and focus for storms near the warm front/stationary front, and probably increase the wind shear. The low-level CAPE map also indicates that there was a maximum of CAPE close to the ground in this same area, suggesting significant CAPE that might be missed when looking at relatively small values on corresponding total CAPE analyses.

The last image above shows a small supercell storm (see arrow) approaching Fayetteville (FAY) on radar at 2024 UTC and 2057 UTC, and a NAM/WRF model forecast profile for Fayetteville near the same time. Notice how the CAPE on the profile was bunched down low, with the fattest CAPE located near 700 mb (3 km above ground). A more typical tornado sounding in the Plains associated with supercells would have the CAPE distributed through a much deeper layer, with the fattest CAPE area located up around 400 mb (near 6-7 km above ground), more than twice as high as the 3/27/09 Fayetteville profile. Even though the low-level shear and storm-relative helicity (SRH) on this profile were not impressive (< 100 m2/s2, a value reinforced by RUC model profiles and SPC estimates, not shown), the rapid upward acceleration of updraft parcels due to CAPE residing so close to the ground may have facilitated tilting and stretching of the available low-level SRH near and north of the front. This could be a key issue regarding tornado potential in this case that involved tornadoes from very low-topped supercells.

Events and settings like this appear to reinforce the importance of detecting areas where CAPE is located atypically low in the atmosphere. This is particularly true when total CAPE looks marginal (say, < 500 J/kg), yet there are other favorable features present such as a short wave aloft, surface boundary, and surface dew point axis in place. Thankfully, tornadoes in most such events are not that strong (usually EF2 or less in intensity).

- Jon Davies 3/29/09


Adam L said...

awesome write up, proof that you dont need 1000+ CAPE to get things going. Thats why I dont get why so many chasers bad mouth a setup that doesnt have CAPE screaming off the charts.

Jon Davies said...

Thanks for the comment, Adam. These small CAPE tornado settings are certainly very interesting.

I didn't get around to responding to your post last week, but it was great to meet you and fellow northeast Illinois chasers at the DuPage Seminar a couple weeks ago. You guys keep up the good work!

chumpson said...

Nice write up, Jon. One of the most difficult aspects of these events is that they almost never seem to occur in isolation. Other severe threat areas serve as distractions (e.g., the complex scenario that same day from FL to TX/OK, including the blizzard), and we simply didn't recognize the threat in SC/NC until it was too late.

Forecast routines can certainly be modified to place the appropriate emphasis on CAPE and its vertical distribution. But, as you know, it's not as simple as saying "you don't need much CAPE to get tornadoes". That statement is true *in the right circumstances*. Weak instability typically works best in moist environments with boundaries and lots of vorticity sources, without excessive deep layer vertical shear.

We'll see if we can catch the next one...

Jon Davies said...

Good comments, Rich.

I agree... a really tough situation to pick out. I missed it when looking at severe potential over the southeast U.S. on Friday morning. The moisture axis, surface boundary, and rather strong short wave aloft might have bemen a tip off to check out the area more closely. But I agree that it's hard to notice such setups when so much else is going on, and standard parameters like total CAPE, EHI, and STP don't highlight them.

Good to see you in Denver last month, and thanks for the post!

Unknown said...

I was the guy that took the pic for WRAL. I have over 2:00 of video of it on Hi-Def if you ever what to see it. Just let me know.

torn80cj said...

This reminds me a lot of the F4 tornado that hit the Raleigh area on Nov 28, 1988. There was no outlook for severe weather, then this storm exploded into a supercell and produced a tornado that would track a path about 82 miles long. At it's worst damage area, it was rated F4 on the Fujita damage scale. I remember 4 people were killed in that twister, which was actually an amazingly low number. It happened overnight. I remember telling my brother, (Mark my words, I guarantee you somewhere tonight there will be a tornado in N.C.). I just didn't think it would be one of that magnitude. The storm was like a time lapsed video, just constant lightning striking without a chance of total darkness in the night sky. It was a bad twister. I live in Oklahoma now and chase these storms.