CLUSTER TORNADO OUTBREAK IN HOUSTON, TX
by Tim Marshall
Storm Track, January/February 1993
© Copyright 1993 Tim Marshall
On November 21, 1992, a concentrated and very unique series of tornadoes struck the Houston, Texas metropolitan area. Six tornadoes occurred within Harris County from four tornadic storms. At one time, there were three tornadoes on the ground simulatenously in the same county! The most damaging tornado (F4) developed in the Channelview suburb, on the east side of Houston. This tornado began as a thin rope and widened dramatically into a wedge shape (3/4 of a mile wide) within two miles of its inception. I cannot recall such a slow moving, cluster outbreak of tornadoes this far south and this late in the season.
Gene Rhoden and myself conducted a damage survey within days following the storms. The survey team involved identifying tornado tracks by aerial reconnaisance, conducting door-to-door ground surveys to document the severity of building damage, and meeting with Mr. Bill Read (MIC) ant others at the National Weather Service in Dickinson, Texas to gather Doppler Radar information in order to piece together a chronology of the tornado events.
A. STORM CHRONOLOGY
An unusually powerful storm system developed across central Texas on the morning of November 21, 1992. Cloudy skies prevailed across the Houston area keeping temperatures in the 60's (°F). However, the air was quite humid with dewpoint temperatures also in the 60's (°F). A cold front extended northeast-southwest from the a surface low pressure system centered near Victoria, TX. That morning, the National Weather Service had issued advisories stating that severe weather conditions were possible later that day.
Of particular significance was the wind velocity profile in the lower atmosphere over Harris county. During the morning of November 22, 1992, there were east winds near the ground surface (33 feet, or 10 meters) between 10 and 15 miles per hour (8 to 13 knots). At 1,000 feet above the ground, the wind was south@erly and at 58 mph (50 knots)!! Near 40,000 feet above the ground, the wind velocity was southwesterly at 92 miles per hour (80 knots)! "Holy wind-shear", I exclaimed. The clockwise turning of the wind direction with height (directional shear) no doubt helped the storms rotate. The increase in wind velocity with height (speed shear) tilted the storms so that their updrafts and downdrafts were separated helping sustain the storms and allowing the tornadoes to become wore visible. Several eyewitnesses told us that the rain was located several miles northeast of the tornado such that they could see the storm coming, and rain immediately followed tornado passage.
By mid-morning, the cold front had surged into southeast Texas and a line of thunderstorms (squall line) had developed along the front and began moving eastward towards Harris County. Around noon, several isolated thunderstorms formed ahead of the squall line within and near Harris County. Well of the tornadoes in Harris County were produced by the isolated storms that were ahead of the squall-line! Most of the storms were "classic" supercells except one storm made a transition to "HP" type. The storms were BO tightly packed together that it appeared the rear flank downdrafts (RFD's) of the leading storm actually enhanced and became part of the inflow on the trailing storm.
The first tornadic activity began in Fort Bend county (southwest of Harris County) around 1:30 pm. This storm (STORM A) produced several tornadoes in Fort Bend County and was responsible for the Pecan Grove tornado and Denver Harbor tornado. About the same time the storm was near Pecan Grove, another storm (STORM B) had formed to the north and produced a long-tracked, multi-vortex tornado east of Katy, Texas.
By 3 pm, two more storms had formed in Harris County. One storm near Pearland, Texas (STORM C), produced a tornado south of Hobby Airport and went on to produce the Channelview, Tx tornado. The other storm (STORM D) produce a tornado that cross Lake Houston. At this time, three storms had tornadoes on the ground simultaneously. The squall line passed through the area within one or two hours after the tornadic storms. Eventually, the squall line caught up to the isolated storms east of the Houston area.
B. SUMMARY OF TORNADO TRACKS
The six tornado paths were labelled one through six based on their time of occurrence. All tornado paths were surveyed except for the Lake Houston tornado. The severity of the tornado damage was rated for each tornado on the Fujita-scale from zero to five, five being most severe.
Tornado #1: Pecan Grove, TX. (2:20-2:40pm): This tornado formed near Richmond (SW of Howton) and moved northeastward through a residential area. Damage included removal of roof shingles, some decking, and destruction of garages. Naximum damage will be rated F1 on the Fujita scale. Total path length was unknown as the tornado mainly traversed open country.Tornado #2: Kelliwood area, Katy, TX. (2:20-3:10 pm): This tornado formed south of town and moved northeastward through several exclusive sub-divisions including Kelliwood Greens, Kelliwood Creek, Green Trails, and Windsor Park. Damage included the removal of roofs, some exterior walls, and destruction of garages. Some homes under construction had collapsed. Maximum damage will be rated F2 on the Fujita scale south of I-10. The tornado continued northeastward across I-10 and struck the West Houston Lakeside Airport destroying several metal hangars. The damage path continued for several miles through residential areas including the Bear Creek area removing roof shingles, some decking, and destroyed several garages. Roof damage occurred to several apartment complexes and to the Randalls store in northwest Houston. Maximum damage will be rated F1 on the Fujita scale north of I- 10. Total path length of this tornado was 19 miles and maximum path width was about .1 mile.
Tornado #3: Lake Houston Tornado (3:00-3:30 pm): This tornado formed just south of Lake Houston in a rural area and continued northeastward across the lake entering a subdivision on the north end. This damage path was not surveyed by us, but the National Weather Service indicated that damage to F3 was reported.
Tornado #4: Pearland, TX. (3:05-3:15 pm): This tornado began in open country northeast of town and proceeded northeastward through a subdivision on the eastern end of Knapp Road. Some roof shingles and cladding were removed on houses. The tornado crossed Fuqua Road before dissipating. Maximum damage will be rated F1 on the Fujita scale. Total path length was two miles and maximum path width was .1 mile.
Tornado #5: Denver Harbor, near downtown Houston, TX. (3:25-3:50 pm): This tornado formed from the same storm that produced the first tornado near Richmond. Damage began near Rice University and moved east-northeast through largely industrial and older residential areas. Many homes lost roofs and some exterior walls had collapsea on masonry buildings. Naximum damage will be rated F2 on the Fujita scale. Total path length was about 11 miles and maximum path width was about .1 mile.
Tornado #6: Channelview, TX (3:30-4:10 pm): This tornado formed from the same storm that produced the Pearland, TX tornado. The first damage was spotted just north of the Houston ship channel in a heavily wooded area. From that point, the tornado moved northeastward, widening rapidly and increasing in strength. The tornado crossed I-10 and entered several subdivisions. Damage ranged from missing roof shingles to complete destruction of residences. We found 14 homes with no interior walls left standing (F4), 88 residences with only a few interior walls left standing (F3), and 183 homes that lost roof structures. Maximum damage will be rated F4 on the Fujita scale. The tornado continued across a rural area, crossing the San Jacinto River and dissipating in the Diamondhead subdivision. Total path length was 12 miles and maximum path width was .6 mile.
C. PATTERNS OF DAMAGE
We found a number of similar failures in residential structures. It was the same old story: if you don't anchor it down, it blows away! A number of homes did not have hurricane straps (These are strips of metal which attach the rafters and ceiling joints to the top of the walls helping to hold down the roof). Many people tola me that hurricanes only affected coastal areas so the straps weren't needed as they lived 40 miles from the coast! Here are some of our observations:
1. Homes with attached garages that faced the wind sustained considerably more damage than those homes where garages faced the opposite direction. Frequently, the windward side of the garage door/rails were pushed inward. Positive internal wind pressures then aided in removing a portion of the garage roof and/or wall(s). In some instances, the roof over the garage collapsed when one or more perimeter walls toppled. Similar observations were made with large atriums in houses.
2. Homes with "L-shaped" plan views appeared to "catch" the wind and thus, incurred more damage than rectangular plan homes oriented parallel to the wind.
3. Homes with hip roofs outperformed homes with gable roofs. However, we noticed that many roofs had minimal attachment to the wall top plates and did not have hurricane straps. Typical roof failures included removal of roof decking, broken rafters, and separation of nailed connections at rafter/wall joints.
4. Homes sustained more damage if they were located on a corner at the end of a block, or adjacent to a prairie/vacant lot than if they were positioned in the middle of the block. Streets apparently acted like wind corridors such that the greater the wind fetch, the higher the wind velocity, and greater the damage.
5. Most homes had minimal anchorage to their concrete foundations. Usually, a concrete nail (without washer) extended through the bottom plates into the concrete foundation about every four lineal feet. Only in a few instances did we find the bottom plates bolted to the foundation. There were four types of wall/foundation failures as the walls rotated: 1) the nails pulled through the bottom plates leaving nails imbedded in the foundation, 2) the nails were removed from the foundation and remained attached to the bottom plate, 3) the nails were sheared off, and 4) the wall studs pulled away from the bottom plate leaving the bottom plate still anchored to the foundation.
6. Many homes had brick veneer failures. Close inspection revealed the following brick wall problems: 1) there were no brick ties, 2) brick ties were mounted on the wall but not secured in the brick, 3) brick ties were pulled from the wall when they were attached between studs or too small of nail was used. In rare instances were the ties actually torn. Many brick wall failures occurred on the leeward side of the homes where negative wind pressure occurred. Several windward walls bowed inward or were impacted by flying debris. In each instance, wall distress was visible as fresh cracks in the masonry or separations between freize boards and the walls.
7. Many small projectiles were generated by these tornadoes and usually involved 2 x 4's or other wooden timbers. Vehicles moved or rolled, few became airborne.