Posted by Steve Keighton, WFO Blacksburg, VA
Often, High-Shear Low CAPE (HSLC) events have clear signals of a severe weather threat due to strong frontal scale forcing, very high low-level wind shear, and/or weak but obvious instability. Other times, these factors are much more subtle with severe potential considered more marginal, dependent on isolated weak updrafts or boundary layer features which may be conducive to some momentum reaching the surface. September 21 of this year was one of those marginal events, with a cold front moving through the Appalachians during the day, but lacking especially strong forcing, very limited instability, and mainly a broad zone of steady rain with only isolated heavier showers just ahead of the front. SPC outlooks issued that day included the general threat with only small chances of severe weather well to the east of the region…closer to the coast.
The image below shows the frontal analysis and radar mosaic valid at 2100 UTC (just before the time of the tornado in northwest NC), followed by images of SPC meso-analysis showing surface-based CAPE (about 250 J/kg near the region of the tornado), and 0-1km shear (with about 25 kts of shear in that layer).
The SHERBS3 parameter forecast from the NAM is shown below for a time just a couple of hours after the tornado touchdown, and while highlighting the region where the tornado occurred with relatively higher values, they are still below the threshold of 1.0. Keep in mind SHERBS3 uses a 700-500mb lapse rate which is likely to be moist adiabatic in the NAM at this time, and the shear in the 0-3km layer which may also have been weaker than the analyzed SPC 0-1km suggest that it actually was. An analysis SHERBS3 or SHERB is not available at this time.
SHERBS3 from the NAM 6hr forecast valid at 00Z.
Not shown, but of worthy note is that cloud bases were extremely low, indicated by LCL analyzed values of 500m or less, which suggests a higher likelihood of any tornadic circulations making it to the ground. Cloud bases at the nearby Mount Airy AWOS were actually reported at between 500-700ft (<200m) just before the time of the tornado.
Just before and at the time of the tornado, WSR-88D radar data from KFCX (about 70-80km away from the tornado and viewing almost 6,000 ft above the ground for the lowest elevation cut) showed some heavier showers but no obvious signs of appreciable updraft structure, yet there was some weak but broad shear with several of the heavier showers, and there was an evolution in the lowest slice of reflectivity data that suggested a very subtle acceleration of part of the echo and marginal “Broken-S” type of signature. Spectrum Width and NROT (from GR2Analyst) were also not especially high, nor were they showing signs of increasing in this region (in fact the signatures looked marginally better along other portions of this broken line of heavier showers just before and after this time. Also recall that no lightning was detected (including in-cloud lightning as observed in StreamerRT from, Earth Networks), and no thunder was reported that we know of.
Below is a 4-panel snapshot of the 0.5 deg slice of reflectivity, storm-relative motion, spectrum width, and NROT for the volume scan time just minutes before the tornado touchdown in King.
KFCX 0.5 deg Refl/SRM/NROT/SW (clockwise from upper-left), at 21:57 UTC, Sept 21, 2013.
Quick calculations of Vr shear were made from +/-15 min surrounding the tornado time, and compared with the Vr shear averages from all tornadic and non-tornadic circulations studied by Jason Davis for the CSTAR HSLC research (in this case specifically for the 50-100km range from closest radar data set), and these are plotted on that graph below. The values (plotted in blue) show that the low-level shear with this particular tornadic cell were well below most of the non-tornadic cases from the David database (at least below the lowest 25th percentile). The signature remained very broad at this height and distance from the radar, and indicated only very subtle strengthening, but mainly after the tornado touchdown.
It is worth noting that there were several other reports of wind damage across parts of NW NC, and south-central VA this afternoon/evening, but almost all were isolated trees down. In one case, just before the confirmed tornado and just a few miles to the southwest, there was some isolated structural damage but was determined to be most likely straight-line (these damage areas can be very hard to determine however, but the survey team did indicate in King it was much more obviously tornadic). One tornado warning was issued for a cell in south-central VA, but the circulation here was no stronger than over King, and no tornado was ever confirmed in VA).
One final interesting note is that this tornado occurred in almost the exact same location as one of the tornadoes examined in the late October 2010 back-to-back events that WFO Blacksburg contributed to the CSTAR HSLC research, and with the particular cell that caused the 2010 tornado in King, it too showed only very subtle shear signatures on radar was especially shallow. A web review of these Oct 2010 events can be found on the NWS Blacksburg web page under “Past Events”, and a brief review of this most recent event (Sept 21, 2013) will be there shortly.
In summary, many tornadoes in HSLC environments have been shown to be very challenging to detect given their shallow and fleeting nature, but in many cases where these parameters are even more marginal (especially the instability) and the updrafts especially shallow, it may be nearly impossible to observe any notable radar-based signatures suggesting a particular cell is more likely to produce a tornado than another cell, unless perhaps these cells are VERY close to a radar.