Early on the morning of September 7, 1996, a cluster of thunderstorms that formed the previous night about 100 miles east of Brownsville, TX, suddenly collapsed in a uniform manner (Fig 1). The outflow from this collapse spread in a circular swath and eventually affected 400 miles of the Gulf of Mexico (Gulf) Coast from Brownsville to Lake Charles, LA.
The synoptic situation over the United States on the morning of September 7, 1996, was dominated by the remnants of Hurricane Fran, which was centered near Pittsburgh. The 500 mb pattern over the southwestern states was weak and westerly with shortwave "ripples" moving through the flow. High pressure was centered over northern Mexico with more ridging off the southeastern coast of the U.S. and into the eastern Gulf. The major surface feature west of the Mississippi River was a low pressure system over Kansas with an associated cold front stretching across the Texas Panhandle and central New Mexico.
1200 UTC CONDITIONS
Upper air analysis of the 500 mb level at 1200 UTC on September 7th indicated a shortwave trough approaching the Texas Coast with the axis from Shreveport to Laredo. Temperature analysis at 500 mb indicated a pool of -6 C over eastern Texas with -8 C over the western Gulf (Fig 2). The analysis at 700 mb also confirmed an approaching trough and a temperature area of 7 C over the western Gulf. High pressure dominated the eastern Gulf at the surface and at 850 mb, with a trough line along the South Texas coast. The wind field was weak and southerly at all levels in advance of the shortwave. At Corpus Christi, precipitable water was 2.04 inches, the cap was a mere 0.2 with CAPE of 3300 j/kg. The relative humidity layer from the surface to 650 mb was 88%.
ETA MODEL FORECAST
A look at the 7 September 0000 UTC run of the Eta model revealed a moist and unstable corridor over the northwestern Gulf with an area of convective precipitation forecasted near and east of Brownsville. The model advanced the shortwave to the coast with a wind direction alteration at forecast hour 1200 for the 500 mb level, and diminished adequate upward vertical velocity by hour 1500 (Fig 3). 1200 UTC temperatures were -6 C over most of the Gulf and -5 C inland at 500 mb, and 8 C over the western Gulf at 700 mb. Also forecast were an 850 mb 342 K theta-e ridge over the central and western Gulf at 0000 and 1200 UTC, a 2.00 inch precipitable water contour over the northwestern Gulf, and CAPE of 1000 j/kg near the coast with a maximum of 3000 j/kg located in the central Gulf. Equivalent potential temperature cross sections showed convective instability over the zone and a deep pool of moist air (Fig. 4).
The cluster of storms measured about 22,500 mi with cloud top temperatures ranging from -50 C to about -65 C as revealed by infrared satellite imagery late in the event (Fig. 1). This mesoscale convective system (MCS) was centered about 120 miles east-northeast of Brownsville. The MCS began to collapse with the passage of the shortwave, which coincided with the rising sun. The wind at 500 and 700 mb veered to the north and the atmosphere warmed enough to cap the cluster and suppress convection. The 0000 UTC September 8, 1996, 500 mb analysis showed the -8 C cold pocket had disappeared with the -6 C contour aligned along the coast. The 500 mb -5 C isotherm had built over central Texas and northern Mexico and probably extended into the Gulf (Fig. 5). The 700 mb temperatures show the same pattern with the 7 C area over the western Gulf warming to 9 C and 10 C at Corpus Christi and into Mexico.
By 1815 UTC, circular outflow was enhancing the seabreeze and initiating scattered showers and thunderstorms along the Gulf Coast between Brownsville and Galveston. Late afternoon satellite pictures revealed the remnants of the outflow still in a circle with a diameter stretching from San Antonio to about 400 miles into the west central Gulf (Fig 6). The showers and storms around the perimeter of the outflow had become isolated by 2200 UTC with the most significant storms staying near the coast, and the strongest near and west of Brownsville (Fig 7). This convection was in an area of convergence from the southerly wind field and the outflow, and was supported by a lingering trough along the Texas Coast. This trough is evident as a corridor of "cool" airand by height analysis (Fig 5). The outflow boundary made very little southward progression through the day, as the surface south wind inhibited such movement. The symmetry of the outflow in the other directions is attributed to the weakness of the wind field.
Exclusive of marine interests, the significance of this event is probably best appreciated when viewed in relation to the previous days limited penetration of the seabreeze front. The seabreeze front frequently acts as the focusing source, and sometimes as the "disperser" for showers and thunderstorms over South Texas.
Another lesson for South Texas forecasters is the necessity of checking temperatures above the boundary layer. Instability indices and moisture may be sufficient, but temperatures of -5 or -6 C at 500 mb are generally too great to support showers and thunderstorms, as was evident by the lack of shower and thunderstorm development along the leading edge of the boundary as it progressed inland during the late afternoon.
Finally, the outflow boundary from this MCS probably remained across south central Texas and provided the focus for showers and thunderstorms with heavy rain that prompted some flash flooding the next morning in San Antonio and in counties to the southeast of San Antonio. Although no deaths were attributed to the flash flooding, some rescues were necessary in the San Antonio area. Conversely, the showers and storms that formed the morning of the 8th were likely routed that afternoon by a strong seabreeze front.
The authors would like to thank Mr. Robert Blaha for his guidance in obtaining satellite photographs. Thanks also to Austin/San Antonio SOO Jim Ward for his review and suggestions.