DAMAGE ANALYSIS OF THE MESQUITE, TEXAS TORNADOby Timothy P. Marshall
(Presented at the 8th Conference on Wind Engineering, Seattle, Washington)
Investigation of buildings damaged from the Mesquite, Texas tornado revealed several interesting features in the debris pattern. Most of the debris was scattered northward with little rotation. Tornado characteristics which effect the debris pattern will be discussed.
Some residences in the tornados' path were heavily damaged while adjacent homes remained unscathed. It will be shown how the variability in construction practices effected building performance. Attention to details like the installation of anchor bolts or metal straps helped explain large variations in the damage zone. Performance of residences was directly related to the type and amount of anchorage used and how well they were installed. Four primary connection failures were observed: wall/foundation, wall stud/bottom plate, roof joist/top plate, and rafter/top plate. Suggestions are presented to help minimize structural damage in high winds.
An unusual winter tornado struck Mesquite, Texas, on December 13, 1984. The tornado moved northward 7.5 miles through heavy residential areas causing a maximum of F-3 intensity damage. Nearly 600 homes sustained damage. This author conducted a ground survey after the storm passed for the purpose of determining the path of the tornado and to rate the intensity of damage according to the F-scale (Fujita, 1971). In addition, failure origin was documented for hardest hit residences, metal and masonry buildings.
TORNADO DAMAGE PATH
Tornado damage was first apparent just south of Lake June Rd. (see Figure 1). Bent television antennas, overturned carports, missing roof shingles and broken tree limbs were observed. As the storm continued northward, damage intensity increased. After removing roofs from a dozen homes and downing a 138 KV power line, the tornado crossed a large field and struck another group of homes. One house was leveled while several others lost their roofs and exterior walls. Three steel-frame warehouses were damaged, while one was totally destroyed. North of Bruton Rd., two churches and a grocery store, constructed of block masonry walls, collapsed. The tornado weakened and dissipated as it started to enter the Garland community.
General Debris Pattern
Along the tornado path, debris was scattered in a north direction with the exception of roof debris from an apartment complex at Peach Tree and Bruton roads which was scattered eastward. This indicated that the center line of the tornado was probably on the extreme west side of the damage path. The unidirectional pattern of the damage appeared to be the result of the fast translational speed which averaged 50 m.p.h. in the north direction. From the extent of damage observed, the rotational speed did not appear much greater than the translational speed.
Metcalf, (1978) showed numerically that a straight-line damage pattern could occur in tornadoes when the rotational and translational velocities are equaled for certain critical speeds based on the resistance of a structure and tornado inflow angle (Figure 2). In that case, the model shows that the center line of a northward moving tornado could occur on the extreme west side of the damage path.
On the other hand, Minor et al., (1977) have shown that rotation in the debris pattern could occur in straight-line winds. Objects near ground level can be transported around buildings giving the impression of rotation. Thus, it is important for the damage investigator to study windstorm debris carefully, keeping in mind the various differences in the aerodynamics and resistance of structures which can affect the debris pattern.
RESIDENTIAL BUILDING PERFORMANCE
Investigation of damaged buildings in the Mesquite tornado revealed four types of connection failures; a) wall/foundation, b) wall stud/bottom plate, c) roof joist/top plate, or d) rafter/top plate connections.
Inadequate wall/foundation connections usually means a loss of large portions of the structure above ground level. A wood-framed residence not anchored properly to the foundation was completely destroyed (Figure 3). Note how smooth the concrete slab surface is where the perimeter wall once stood. There was little evidence of anchor bolts or concrete nails in the foundation.
The same type of failure was observed in a non-loadbearing masonry wall (see Figure 4). The 25 foot high wall was not reinforced with steel, and hollow block cells were left ungrouted. This wall was essentially free standing as it was not well secured to the intersecting walls, roof or the foundation. Luckily, no one was in or near the building when the wall collapsed. It was interesting to note that the rest of the building did not sustain wind damage, indicating wind velocities enveloping the building were relatively low.
Wall stud/bottom plate failure was observed at the First Assembly Church (Figure 5). Examination of the wall/foundation connection revealed that anchor bolts, which secured the bottom plate to the foundation, had remained in place. However, wall failure occurred between the wall stud and bottom plate where the nailed connection was pulled. Two straight nails had attached the wall stud to the bottom plate. Lateral forces due to wind effects apparently pivoted walls about the base until collapse occurred.
Inadequate roof/wall connections usually meant loss of large roof sections. Most homes examined had minimal roof/wall anchorage and apparently were designed to remain on the structure by gravity force. Light framed gables appeared most susceptible to be damaged by wind effects (Figure 6). Roof losses most likely occurred from inadequate roof joist/top plate connections.
STEPS TO MINIMIZE WALL AND ROOF DAMAGE
Along the damage path, there was a wide variation in construction practice. Some of this could be attributed to the types of materials used in housing construction. However, it was particularly bothersome to find that newly constructed residences were not much better in terms of the amount and extent of anchorage than older buildings. In each case, structural failure evolved from the weakest connection. Thus, gradations in the tornado damage appeared more readily explained by the variations in housing construction rather than a sudden change in wind velocity.
There are several publications available to builders today which illustrate how to correctly install anchors and braces to resist wind forces. A brochure from the Southern Forest Products Association, (1974), is one such publication which shows how wind resistant connections should be installed during construction or in existing structures.
Wall/floor failures can be minimized by the placement of metal bolts in the concrete slab to secure the bottom plate (Figure 7). Additional support to the wall stud/bottom plate can be obtained with a galvanized plate nailed at the connection. This will help resist overturning of the wall.
Avoidance of roof/wall failures involves the placement of metal clips or straps between rafters, roof joists, and top plate of the wall (Figure 8). Metal clips are readily available especially in some coastal areas that are hurricane prone. Adding these types of connections will help transfer wind loads to other structural members and to the foundation.
PROBLEMS WITH USING THE F-SCALE
The F-scale is frequently used by the scientific community for assessing the intensity of damage to structures. F-scale ratings are assigned to a structure according to one of six damage intensity classifications ranging from F-0 to F-5 with increasing amounts of damage. The F-scale is a subjective, visual interpretation of the damage.
One problem with the F-scale is that it treats each building as being constructed the same when, actually, vast differences in structural strength occurs. This problem is magnified in that tornado intensity is rated according to the greatest amount of building damage. Thus, a single "poorly" constructed home which is heavily damaged by a tornado can lead to an over-estimate of the actual tornados' strength. Factors which can influence the F-scale rating are listed in Table 1.
Such was the problem with the Mesquite tornado in that a single residence was leveled. Would this then be classified as an F-5 tornado? Since the residence was not anchored to it's foundation, it had little resistance to the wind. Considering the inherent weakness of the structure a rating of F-2 or 3 would seem more appropriate. For this reason, damage investigators should use caution when assigning F-scale intensity.
Investigation of structural damage in the Mesquite, Texas tornado revealed that gradations in damage were primarily attributed to variations in construction, rather than sudden changes in wind velocity. Residences, metal buildings and masonry structures had common failures originating at the weakest connections. In each case, the origin of structural failure could be traced back to wall/foundation, wall stud/bottom plate, roof joist/top plate, or rafter/top plate connections. It has been shown how the addition of anchors or braces would help minimize wind damage.
The fast translation of the tornado caused the debris pattern to look straight-line. Care must be taken when investigating windstorm damage to understand all the possibilities of how damage occurred. Problems with the F-scale must be realized in order to obtain a more realistic view of the damage intensity.
I would like to thank the board of directors at Haag Engineering Co., for their support and continuing encouragement to complete this paper.
Fujita, T.T., 1971: Proposed Characterization of Tornadoes by Area and Intensity, SMRP Research Report #91, University of Chicago Press, 15 pp.
Metcalf, D., 1978: Simulated Tornado Wind Fields and Damage Patterns, M.S. Thesis in Geosiences, Texas Tech University, 67 pp.
Minor, J., McDonald, J., and Mehta, K., 1977: The Tornado: An Engineering-Oriented Perspective, NSSL Techincal Memorandum, #82, 195 pp.
Southern Forest Products Association, 1974: How to Build Storm Resistant Structures, P.O. Box 52468, New Orlean, LA 70152, 10 pp.
TABLE 1. FACTORS WHICH CAN INFLUENCE THE F-SCALE
Load and Resistance Properties
a. Type of construction material.
b. Degree of engineering attention.
a. Traditional building methods applied.
b. Building code utilized.
c. Extent and type of anchoraging used.
d. Orientation of eaves and/or garage.
e. Type of roof geometry.
f. Extent and type of diagonal bracing.
a. Type of survey- ground, aerial, both.
b. Documentation of missile events.
c. Shielding from adjacent buildings