Hurricane Allen formed on August 1st off the west coast of Africa and became a classic Cape Verde hurricane as it ravaged its way toward the west-northwest on an unwavering track toward the western Gulf of Mexico. The storm passed near the Yucatan Peninsula on August 7th and made landfall on the lower Texas coast north of Brownsville (near Port Mansfield) about midnight on the night of August 9th. The storm continued moving to the west-northwest to near Laredo around 12pm on August 10th, while gradually weakening. Allen was a Category 5 storm 215 miles east-southeast of Brownsville but had weakened to Category 3 at the time of landfall.
From August 9th through 6am on the 10th, nine tornadoes occurred, all in counties close to the coast, and mainly from Corpus Christi southward. Four of those tornadoes produced damage rated F2, with estimates of damage running into the $2 million range. These events were consistent with the landfall of a major hurricane, but I doubt that anyone could have anticipated what would occur later on the next day, August 10th.
An outbreak of tornadoes occurred across south-central Texas during the day on August 10th, including tornadoes in Williamson, Travis, Caldwell, Hays, and Bexar counties. Tornado expert Dr. T.T. Fujita surveyed the damage in and near San Marcos, where a tornado caused extensive damage, including a campground of trailer homes, a nursing home and an apartment complex. He found a damage path that was 6.75 miles long with evidence of multiple vortices and some F3 damage. Another strong tornado (damage rated F2) caused $250 million in damage at the Austin Municipal Airport about the same time. At least 16 of Allen’s 29 tornadoes occurred in Central Texas, in an area that was in both the right-front quadrant (relative to storm motion) and the northeast quadrant (relative to Allen’s position), at a distance of some 135 to 215 miles from the storm center. Several of the tornadoes were long-tracked and strong, producing property damage estimated at well over $250 million. Thirty people were injured, some seriously, but fortunately, there were no deaths.
Examination of upper air data from 12UTC on the morning of the 10th revealed a pronounced dry intrusion that appears to have been entrained into Allen’s circulation from the east and southeast. Best depicted at the 700 mb level (Fig. 3), the intrusion rotated westward over a period of 24–36 hours, producing a very sharp north–south gradient of relative humidity above the Austin–San Antonio area during the daytime of August 10th. The intense midlevel gradient (from very dry in north central Texas to very moist in southern Texas) coincides with the tornado outbreak area. Figures 5a and 5b (below) offer a comparison of the morning and evening soundings at Victoria, TX and clearly reflect the impact of the mid-level dry intrusion.
In a research paper that I authored (published in the A.M.S. journal Weather and Forecasting in 2004), I reviewed all tropical cyclones that made landfall in the United States from 1960 to 1999 in which twenty or more tornadoes occurred on the day of, or the day after, the storm’s landfall. Of the thirteen cases in that study, eleven were found to have clear evidence of a mid-level dry intrusion, typically in the right-front quadrant relative to storm motion.
From my research, as well as that of others, it appears that the mid-level drying can substantially alter the thermodynamic structure of the tropical cyclone environment, with substantial enhancement of CAPE and surface-based instability. The drier mid-level air erodes the thick cloud blanket generated by the tropical cyclone, allowing some incoming solar radiation to penetrate into the lower troposphere, creating enhanced instability which leads to stronger updrafts. When that occurs within the peripheral wind field in the right-front quadrant (relative to storm motion) of a tropical cyclone, conditions can become favorable for the development of stronger tornadoes. If you are interested in delving deeper into this area of tornadogenesis, an excellent place to start is an article by Roger Edwards, one of the lead forecasters at the N.W.S. Storm Prediction Center. His article is online in the Electronic Journal of Severe Storm Meteorology (EJSSM), at this link: http://ejssm.org/ojs/index.php/ejssm/article/view/97. Roger worked at the National Hurricane Center before moving to the Storm Prediction Center. He is one of the most knowledgeable people in the field of tornadoes associated with tropical cyclones.
Lon Curtis Bio: Although educated as an attorney, and serving 24 years as a Texas prosecutor, Lon Curtis pursued his passion for weather as a serious avocation before realizing his dream of becoming an on-air meteorologist at KWTX-TV (Waco) in 1997. That opportunity came about, in part, because of his role in storm spotting during a tornado outbreak in May of 1997, including the tornado at Jarrell that killed 27 people. Lon is a member of the American Meteorology Society and the National Weather Association, and holds the AMS Certified Broadcast Meteorologist certification (on inactive status since his retirement from broadcasting in 2011). Lon authored a peer-reviewed research article published in the AMS journal Weather and Forecasting (2004), on mid-level dry air intrusions as a factor in tornado outbreaks associated with landfalling tropical cyclones. Since retiring, he has continued his research interest in severe convective storms, and is building a database of all tornadoes in Texas since 1880 that have caused one or more deaths. He also maintains a personal website on weather conditions and hazards affecting Central Texas.