by Tim Marshall, Haag Engineering Co., Dallas, Texas


On May 3, 1999, a violent tornado traveled through the southern and eastern suburbs of Oklahoma City. The next morning, the author was asked to lead one of three damage survey teams sponsored by the Institute for Disaster Research at Texas Tech University in Lubbock, Texas. The teams were given five tasks: 1) to map out the damage path and assign F-scale ratings to the damaged residences, 2) to document the performance of housing, 3) to interview witnesses, 4) to document projectiles, and 5) to determine if there were any above or below ground shelters within the damage path and to assess their performance. Both aerial and ground surveys were conducted. The damage path was divided between the teams for the survey. A report summarizing the findings of all three teams was published by Gardner et. al. (2000).

This paper will present some of the findings from the authorís team which surveyed the damage at Moore, Oklahoma where the most severe damage (F-5) occurred. Various modes of house construction were found including various methods to secure wall bottom plates to the concrete slab foundations. Thousands of projectiles were launched by the tornadic winds ranging from conventional wooden boards to unusual projectiles such as a metal sewer pipe and a metal folding chair. The author returned to the area three months after the disaster and noted several deficiencies in the construction of new housing.


The tornado that went through Moore, Oklahoma initially began just northeast of Chickasha, Oklahoma located about 70km southwest of Oklahoma City. The tornado traveled northeastward paralleling I-44 and had a continuous damage path which at times reached 1.5km wide. The tornado continued through the city of Moore and into portions of Del City and Midwest City before lifting (Fig 1).

As the tornado entered the west side of Moore, the width of the damage path narrowed to about 400m wide. The tornado became quite intense inflicting F-4 and F-5 damage to a number of homes in the Country Place subdivision located just south of 134th Street (Fig 2). These wooden-framed homes were built on concrete slab foundations. Most of the wall bottom plates were attached with tapered concrete nails driven into the concrete slab foundation at various intervals.

The tornado damage path remained about 400m wide through the Briarhollow subdivision. These were older tract homes with attached two-car garages. The author found a mixture of homes that were either nailed, strapped and nailed, or anchor bolted to their concrete slab foundations. The Westmoore High School was located on the north side of the damage path and sustained some damage to the roof and metal cladding. An awards ceremony was being held at the time the tornado struck and hundreds of people were in attendance. The people were evacuated to interior portions of the school and all survived, however, their vehicles in the parking lot were tossed around. Some of the automobiles ended up in the houses south of the High School.

As the tornado crossed Western Ave., it struck the Emerald Springs Apartments. Some of the two-story wooden structures were reduced to one-story or less. Several injuries and fatalities were reported here. A metal church building was completely destroyed. The tornado continued through the Greenleaf subdivision reducing many wooden-framed townhomes to rubble. Numerous townhomes sustained F-4 damage and one home sustained F-5 damage. The tornado completely destroyed a two-story metal church building, then crossed Santa Fe Avenue entering the heart of city. The damage path extended diagonally from NW 12th to NW 19th Street through smaller wooden-homes with attached one-car garages inflicting up to F-5 damage.

The tornado traveled right over Kelly Elementary School. The steel frame buckled and most of the exterior walls collapsed. The tornado continued through another subdivision inflicting up to F-4 damage before crossing I-35 at Shields Blvd. A fatality and several injuries occurred when people tried to seek shelter under the bridge there. However, the bridge was a concrete slab (no steel girders) and offered no protection. The tornado crossed I-35 and struck an apartment complex and Best Western Hotel. Roofs were removed and some of the two-story buildings were reduced by one-level. The tornado then traveled through the Ridgewood subdivision causing up to F-4 damage before traveling out of town over rural areas. The tornado turned more northerly at I-240 and continued through portions of Del City and Midwest City before dissipating. In all, the teams rated F-1 damage to 284 homes, F-2 damage to 405 homes, F-3 damage to 558 homes, F-4 damage to 317 homes, and F-5 damage to 17 homes.


Houses surveyed were conventionally constructed. Wooden bottom plates were either nailed, strapped, or bolted to their concrete foundations with the majority being nailed. Nails were 5cm long, concrete type, with tapered shanks, and penetrated into the concrete foundation about 1.3cm. In many instances, the nailed connection failed. The wooden bottom plate either pulled through the nail leaving the nail in the concrete foundation or the wooden bottom plate and nail were removed. Stronger straps and anchor bolts held the wooden bottom plates in place. In these cases, failure occurred where the wall studs were attached to the bottom plate. Most wall studs were straight-nailed (two nails) to the wooden bottom plate and toe-nailing was observed in a few instances. The straight and toe-nailed connections were weaker than the straps or anchor bolts. Thus, the nailed connections failed first. Similar observations were made by the Building Performance Assessment Team (1999).

Typical house failure initiated as the garage doors buckled inward and windows broke allowing the wind to enter and pressurize the buildings. Roofs were then uplifted and walls removed. Houses with attached garages facing the wind sustained more damage. Similar observations were made by Marshall (1982). As expected, none of the houses were built to withstand an F-5 tornado. Failure of most homes initiated at wind speeds closer to code values which are 40 m/s (3-sec gust) for the Oklahoma City area. It should be noted that such building codes represent the minimum requirements.


The survey team assigned F-scale numbers to each damaged house (Fujita, 1971). The F-scale is a subjective, visual interpretation of damage ranging from 0 to 5 based on increasing severity of damage to a structure. Homes with minor roof damage or toppled chimneys were rated F-0. Homes which lost their roof covering or had attached garages that were destroyed were rated F-1. Homes which lost their roof structures were rated F-2. Homes which lost their roofs and most exterior walls were rated F-3. Well-constructed homes that were leveled with a pile of debris remaining were rated F-4, and homes that were completely removed from their foundations and no debris remained were rated F-5.

Wind speed ranges associated with the F-scale have been known to be inaccurate especially in the higher F-scale numbers. The wind speeds were originally derived when Fujita divided the gap between the top wind velocity in the Beaufort-scale and Mach 1 into 12 equal increments. Minor et. al (1993) suggested that the wind velocity ranges be lowered for F-2 or greater damage based on engineered assessments (See Table 1).

Other shortcomings in using the F-scale wind speed ranges are noted. The F-scale assumes each structure is homogeneously constructed. Thus, a weak building may fail when an adjacent building that is better built remains. Residences are usually non-engineered, so there will be greater uncertainty in an F-scale rating. Minor et. al (1993) indicated an accuracy of plus or minus one F-scale rating for buildings rated F-3 or above especially if inherent strengths or weaknesses in buildings are not considered.

There are also problems with rating tornado intensity based on the degree of damage to a structure. A house may fail at a lower wind velocity than the maximum wind speed of the tornado or the house may not be in the direct path of the tornado and sustain minimal damage in a strong tornado. Thus, the F-scale rating of damage to a structure does not necessarily reflect the strength or intensity of the tornado. Also, the translational speed of the tornado may affect the severity of residential damage as noted by Phan and Simiu (1998) in the Jarrell, Texas tornado in 1997. Wind speeds of longer duration generally causes more severe damage to residences.


The Moore, Oklahoma tornado generated thousands of projectiles. These projectiles impacted and/or went through houses and vehicles. Most of the projectiles were broken wooden pieces from houses and trees. The largest projectile found was a 3.7m diameter, 4.3m tall steel oil tank which tumbled end-over-end for 276m just west of the Newcastle bridge. The tank was not anchored. Several vehicles traveled up to 300m and ended up in ravines or culverts. Parts of mobile homes (i.e. steel beams up to 11m long) traveled similar distances.

The most unusual projectiles were: 1) the leg of a steel lawn chair which penetrated a solid wooden post that measured 13cm x 13cm, 2) a 2.4m long wooden board which entered a hole in the roof of a house and penetrated a refrigerator freezer, 3) a 2.4m long wooden fence post which went through a window and lodged in an interior wall, 4) a 1.8m long steel pipe, weighing about 280kg, which went through the front door and came to rest in an interior hallway. Also, a bathtub was moved 100m in which two people were seeking shelter.


The author revisited the disaster area three months after the tornado to monitor the rebuilding efforts. A total of 40 buildings were examined in Moore and southern Oklahoma City to see if building practices have changed in the wake of the tornado. Out of the 40 new homes inspected, five homes had bottom plates bolted to their foundations, six homes had bottom plates strapped to their foundations, and 29 had bottom plates nailed to their foundations. Tapered concrete nails were spaced from 30 to 130cm apart. Nails had a maximum of 1.3cm penetration into the concrete. In general, we found that the new construction was no better in quality than before the tornado. However, six of the 40 new homes did have concrete safe rooms constructed on their foundations. Information on constructing safe rooms has been published by the Federal Emergency Management Agency (1999) or can be obtained on their web site at http://www.fema.gov


The tornado disaster at Moore, Oklahoma provided an opportunity to learn about building response. As expected, there was significant variability in house construction and performance. Houses with F-4 and F-5 damage provided little shelter to the occupants. Portions of houses need to be "hardened" to provide occupant safety. This is especially true for schools. A number of projectiles were documented including a few never before seen items.


The author would like to thank Dr. Ernest Kiesling and Dr. Kishor Mehta at Texas Tech University for being invited to serve as a member of the disaster team.


Building Performance Assessment Report, 1999: Midwest Tornadoes of May 3, 1999, Federal Emergency Management Agency, Denton, Tx, 216pp.

Federal Emergency Management Agency, 1999: Taking Shelter from the Storm: Building a Safe Room Inside Your House, 28pp.

Fujita, T.T., 1971: Proposed characterization of tornadoes and hurricanes by area and intensity, SMRP Research Report 91, University of Chicago, Chicago, IL, 15pp.

Gardner, A. et. al., 2000: The Tornadoes of Oklahoma City of May 3, 1999, Wind Science and Engineering Center, Texas Tech University, Lubbock, Tx, 38pp.

Marshall, T.P., and J.R. McDonald, 1982: An Engineering Analysis of the Grand Island Tornadoes, 12th Conf. on Severe Local Storms, San Antonio, Tx, p. 293-296.

Minor, J.E., J.R. McDonald, and K.C. Mehta, 1993: The Tornado: An Engineering-Oriented Perspective, NOAA Technical Memorandum, NWS SR-147, 196pp.

Phan, L.T., and E. Simiu, 1998: The Fujita Tornado Intensity Scale: A Critique Based on Observations of the Jarrell Tornado of May 27, 1997, NIST Technical Note 1426.



(after Minor et. al 1993)

Original* Adjusted**

Wind Speed Wind Speed

(mph) (mph)

F0 40-72 40-75

F1 73-112 75-110

F2 113-157 110-140

F3 158-206 140-170

F4 207-260 170-200

F5 261-318 200-240

*after Fujita (1971)

**after Minor (1993)