A fall “second season” severe weather event occurred over portions of the Austin-San Antonio county warning area during the evening hours on Monday, October 24. Multiple reports of wind damage were received along and ahead of an advancing squall line over portions of Hill Country and the Interstate 35 corridor. A storm survey would later conclude that a brief tornado touched down to the immediate south and southwest of Jarrell as well. This review takes a closer look at the event and the meteorology behind it.
A link to these severe weather reports can be found here.
Multiple large-scale features would influence our weather through the day on October 24. To our northwest, an upper level trough was emerging from the New Mexico Rocky Mountains into the Texas Panhandle. An intensifying jet streak swung through the base of the trough, with surface cyclogenesis becoming apparent under its left exit region during the early afternoon hours. An infant surface low would reach the Red River Valley by sunset. A cold front branched out to the southwest of the newly developed cyclone, surging east and southeast into the Edwards Plateau and Big Bend by early evening (figure 1). To the southeast, the remnants of the former Hurricane Roslyn drifted out of Mexico and into South Texas through the morning hours, helping to trigger light showers and thundershowers from the I-35 corridor northeast into East Texas and the Piney Woods. This activity would vacate from the region by the mid afternoon hours, leaving an east to west-oriented outflow boundary in its wake. The feature served as a demarcation between warming afternoon surface temperatures to its south and a rain-chilled air mass to its north. All of these features were evident in satellite imagery through the day on October 24 (figures 2 and 3).
Figure 1: Surface analyses at 15 (10:00 AM CDT), 18 (1:00 PM CDT), and 21Z (4:00 PM CDT) 24 October, as well as 00Z (7:00 PM CDT October 24) October 25. Images obtained from the Weather Prediction Center surface analysis archive.
Figure 2: Loop of GOES-16 Channel 9 (mid-level water vapor) imagery between 06 and 22Z (1:00 AM and 5:00 PM CDT) October 24. The approaching upper level trough and jet streak, along with the remnants of the former Hurricane Roslyn, are marked accordingly. Imagery obtained from the Regional and Mesoscale Meteorology Branch (RAMMB).
Figure 3: GOES-16 visible satellite imagery between 19 and 22Z (2:00-5:00 PM CDT) 24 October. Animation time step is in ten minute increments. Outflow boundary and approaching cold front are annotated accordingly. Imagery obtained from the Regional and Mesoscale Meteorology Branch (RAMMB).
With mid and upper level moisture from Roslyn vacating to the northeast, our eyes turned further to the west and the advancing cold front. Most noticeable over areas to the northwest of our CWA, storms had been ongoing ahead of the boundary through the middle and late afternoon hours. This prompted the issuance of Severe Thunderstorm Watch #553 over portions of the Dallas/Fort Worth and San Angelo CWAs, with Burnet and Llano Counties also being included in the Austin-San Antonio service area (figure 4). Despite the activity ongoing to the northwest, conditions remained quiet over South-Central Texas through the dinner hour. This would change moving deeper into the evening, as the swiftly-moving cold front began to advance upon the outflow boundary situated across northern portions of the region. Two well-defined regions of showers and thunderstorms began to mature following 6:30 PM—one of which was located ahead of the front along the outflow boundary over portions of Burnet, Lampasas, Llano, and San Saba Counties. The other, located further to the west and extending along a line from Mason County to southwest of the DFW metroplex, was positioned along and just ahead of the quickly moving cold front (figure 5). The activity appeared to be triggered by a combination of convergence along the outflow boundary and front, combined with an inflowing Gulf air mass having dew points in the mid and upper 60s (figure 6). The leading batch of convection, exhibiting supercellular characteristics at times, would track along the Llano-San Saba and Burnet-Lampasas County lines, prompting issuance of the first severe thunderstorm warning at 7:01 PM CDT.
Figure 4: Counties that were included in Severe Thunderstorm Watch 553 from the Storm Prediction Center (SPC). Image obtained from the SPC convective watch archive.
Figure 5: Still image of Granger (KGRK) 0.5 degree tilt reflectivity at 7:13 PM CDT. Two distinct regions of convection—one rooted to the outflow boundary and one connected to the cold front—were evident at this time. Data obtained from the Amazon Web Services (AWS) cloud archive. GR2Analyst was implemented for visualization.
Figure 6: Crude hand analysis performed at 23Z (6:00 PM CDT) by an EWX mesoanalyst. Yellow and green shading indicates lines of constant dew point (isodrosotherms), with darker shades of green implying higher values. EWX CWA is outlined in black. A narrow corridor of elevated moisture transport into Hill Country was evident at this time as the cold front was beginning to encroach from the west. Surface data obtained from SPC.
With discrete storms rooted to the outflow boundary, and an estimated 200-300 m2/s2 (figure 7) of effective storm relative helicity in the ambient environment, near term tornado concerns began to grow across the northeastern CWA. This was addressed by the Storm Prediction Center in mesoscale discussion #1893, with a tornado watch eventually being coordinated for areas to the east of I-35 and north of I-10 (figures 8 and 9). Cells connected to the outflow boundary continued to track across portions of northern Burnet County through the beginning of the period and would go on to produce half dollar to golf ball-sized hail in Bell County in the Dallas/Fort Worth CWA. Between 7:45 and 8:30 PM CDT, the trailing region of frontal convection began to catch the leading cells and clusters along the outflow boundary. Storms associated with the cold front raced east-northeast across portions of Blanco, Burnet, Gillespie, and Llano Counties through 8:00 PM, producing straight-line wind damage to trees and power poles. This prompted the extension of pre-existing severe thunderstorm warnings downstream into northwestern Williamson County at 8:10 PM CDT. The same storms would ultimately track into and merge with the quasi-supercellular structure initiated by the outflow boundary in southwestern Bell County. By 8:30 PM CDT, a single band of convection was evident from Killeen to just south of Fredericksburg (figure 10).
Figure 7: Effective storm relative helicity (blue contours) and effective inflow bases (red shading) (reference Thompson et. al. 2007 for further information) at 00Z 25 October (7:00 PM CDT 24 October). As a rough rule of thumb, values reaching and exceeding 150 m2s-2 suggest near-term tornadic potential. Image obtained from the SPC mesoanalysis archive.
Figure 8: Companion graphic of SPC mesoscale discussion (MCD) 1893, which included portions of Bastrop, Lee, Travis, and Williamson Counties. Image obtained from the SPC mesoscale discussion archive.
Figure 9: Counties that were included in Tornado Watch 554 from the Storm Prediction Center. Image obtained from the SPC convective watch archive.
Figure 10: KGRK 0.5 degree reflectivity loop between 7:16 and 8:30 PM CDT. Note the consolidation of prefrontal convection along the outflow boundary and storms immediately along the cold front into a single, broken line through the period. Data acquired from the AWS radar archive. Visualization done in GR2Analyst.
The newly-consolidated line of thunderstorms continued to push east into portions of western Williamson County between 8:30 and 8:45 PM CDT. Of particular interest during this period was the convective line’s gust front, which presented itself as a fine line extending from Florence to Liberty Hill to just west of Lakeway. Fine lines are narrow bands of low radar reflectivity returns that appear when winds surge out ahead of thunderstorms. These gusty winds pick up dust, leaves, bugs, and other particles, which register as weak power returns on the radar display. We can also use a dual-pol product called correlation coefficient to track these features, as non-meteorological scatterers show up in dark blue coloration. This is in stark contrast to rain drops and hail stones, which will typically present themselves in green, yellow, or red coloration. With this in mind, it became apparent that most of the convection ongoing across Blanco, Burnet, and Travis Counties was behind the gust front. When this occurs, we know that the storms have been undercut by cold, outflowing air, which acts to cut their updrafts off from the surface. This eliminates the storm as a potential tornado producer. Conversely, storms over western portions of Williamson County during this timeframe were snugly collocated with the gust front, indicating that they had not been undercut and continued to remain rooted to the surface (figure 11) This area of the line thus demanded thorough observation for potential spin up tornadoes, which was further reinforced by the extreme values of low level storm relative helicity being estimated by the Granger (GRK) radar (figure 12). Evident in the 8:44 PM CDT scan from GRK, such a spin up appeared imminent following a roll up and attendant “kink” in the convective line just to the southwest of Jarrell (figure 13). A tornado warning was issued at 8:46 PM CDT, just as a tornado was touching down four miles southwest of town.
Figure 11: Animation of KGRK 0.5 degree reflectivity (left) and correlation coefficient (right) between 8:30 and 8:39 PM CDT. Based on the respective locations of the outflow boundary and convection, it was apparent that storms were elevated over portions of Blanco and Burnet Counties through the period. Further north, storms were nearly surface-based across western Williamson County. Data acquired from the AWS radar cloud archive. Visualization done in GR2Analyst.
Figure 12: Vertical wind profile (VWP) sampled from KGRK at 8:49 PM CDT. Note the very high to extreme values of storm-relative helicity (SRH) in the 0-500 meter, 0-3 kilometer, and 0-6 kilometer layers. Such numbers are very favorable for near-term tornado potential. Image obtained from a database maintained by Autumn Sky.
Figure 13: 8:44 PM CDT 0.5 degree reflectivity scan from GRK. Tornadogenesis would occur two minutes following this scan to the southwest of Jarrell.
Tracking just over four miles to the southeast between 8:46 and 8:53 PM CDT (figure 14), the first signs of tornado damage in Williamson County were to the west of CR 234 and north of CR 239. It was in this portion of the county that survey crews encountered snapped trees, as well as structural damage to a barn and a partially-constructed two story house. Upon crossing the intersection of the two county roads, the tornado turned east while concurrently strengthening and widening, imparting high-end EF0 damage to trees and homes immediately to the west of I-35. Further intensification occurred as the tornado crossed I-35 to the north of Ronald Reagan Boulevard, where a tractor trailer was flipped and nine power poles were downed. EF1 damage was noted to the east of the freeway, where a home suffered roof and tree damage. A small cargo trailer residing on the same property was tossed approximately one hundred yards downwind. These observations were consistent with sustained wind speeds of 100 MPH. The tornado would lift to the east Little Road shortly thereafter. Additional, non-tornadic wind damage would be observed by crews on the south side of Jarrell, where numerous fences were toppled and several buildings had doors blown out.
Figure 14: Social media graphic prepared by EWX staff once the preliminary damage survey of the Jarrell tornado had been completed.
Additional tornado warnings would fortunately not be needed through the remainder of the event as convection would remain elevated behind the undercutting cold front and outflows. Several additional wind damage reports would filter in from portions of the I-35 corridor, including a measured 60 MPH gust near San Marcos at approximately 9:42 PM CDT. This would prompt the issuance of several severe thunderstorm warnings along and east of I-35 through 10:30 PM CDT. Additional thunderstorm development was observed between 10:30 and 11:00 PM CDT to the south and east of San Antonio as the cold front and remnant outflow impinged upon a progressively more humid air mass in the Coastal Plains. The final severe thunderstorm warning of the event was issued for portions of Atascosa, DeWitt, Karnes, and Wilson Counties at 11:22 PM CDT as the southern batch of storms reached peak intensity. Like their counterparts further to the north, these storms would get undercut by outflow and the advancing cold front following the midnight hour.
Alluded to many times over in this review, the speed of the approaching cold front played a critical role in the outcomes observed during the October 24 squall line. Storms forming along the boundary struggled to remain surface-based due to its rapid forward propagation across the area. This limited tornado potential, which could have been much greater per model and observational estimates had storms remained rooted to the surface. Said potential was realized in just one location, and occurred following the consolidation of along-front and prefrontal convection over Hill Country. Damaging winds were the otherwise predominant hazard in the Austin-San Antonio service area. Given these details, continuous maintenance and diagnosis of the near-storm environment was necessary during this event. Key messages from these analyses were conveyed publicly through a relatively new “mesoscale update” (figure 15) heading in the EWX Area Forecast Discussion. We plan to continue leveraging this messaging medium in future events.
Figure 15: Mesoscale AFD update issued at 7:36 PM CDT providing further messaging as the event was unfolding. Excerpt obtained from the IEM COW NWS text product archive.
Follow us on Twitter
Follow us on Facebook
Follow us on YouTube
EWX RSS Feed