CHAPTER 7 
7 Preliminary 
Conclusions 

The preliminary conclusions presented in this report are based on the BPAT's 
observations, an evaluation of relevant codes and regulations, meetings with 
State and local officials, and other interested parties such as organizations 
representing builders and contractors. The conclusions of this report are 
intended to assist communities, businesses, and individuals and to provide 
technical guidance for personal and property protection. 

7.1 RESIDENTIAL PROPERTY PROTECTION 
The BPAT observed considerable damage to single-family housing, multifamily housing, and manufactured and modular housing. Failures observed 
resulted from windborne debris and high winds that often produced forces on 
buildings not designed to withstand such forces. Failures, in some cases, also 
were observed that were due to improper construction techniques and poor 
selection of construction materials. Damage, in some situations, could have 
been reduced or avoided if newer building codes and engineering standards 
that provided better guidance for high wind events had been adopted, 
followed, and enforced. 

The majority of residential construction in Oklahoma and Kansas is currently 
required to be designed per the 1995 CABO One and Two Family Dwelling 
Code. Although local municipalities have adopted some amendments to this 
code, it does not incorporate wind speed design parameters used by the 
newer 1997 UBC and 1996 NBC codes. Furthermore, engineering standards 
such as ASCE 7-95 and 7-98 provide better structural and non-structural 
design guidance for wind loads than these newer codes. Although designing 
for tornadic wind events is not specifically addressed in any of these newer 
codes or standards, constructing residential homes to these codes and 
standards would improve the strength of the built environment. Building to 
these codes and standards would have led to reduced or minimized damage 
in areas that were affected by the inflow winds of all tornadoes and reduced 
the damage to residential construction impacted by the vortices of moderate 
tornadoes. 

7.1.1 Single-and Multi-family Homes 
The BPAT observed many single-family residential buildings that were in 
the direct path of tornado vortices or in the inflow areas that received 

BUILDING PERFORMANCE ASSESSMENT: OKLAHOMA AND KANSAS TORNADOES


7-1 

CHAPTER 7 
PRELIMINARY REPORT


structural damage. This damage was typically a result of the lack of 
redundancy in the structural system to resist wind-induced uplift loads, wind-
induced lateral loads, or increased loads on the building due to internal 
pressurization and a breach of the building envelope. It is crucial to establish 
a continuous load path in order to provide improved resistance to wind forces 
and windborne debris. 

In residential areas affected by the vortices of violent and severe tornadoes, it 
is difficult to economically construct a home that is tornado-resistant. 
However, improved construction and implementation of construction 
techniques that are used in other high wind-prone regions of the country may 
have significantly reduced the property damage caused by moderate tornado 
vortices and inflow winds of severe and violent tornadoes. 

7.1.1.1 Load Path and Structural Systems 
Foundations in the single- and multi-family homes performed adequately 
during the tornadoes in both Oklahoma and Kansas. The deficiency or failure 
mode of the load path at this point was the connection of the structural 
systems to the foundation. Wood framing relied on the connection of the sill 
plate or floor framing to the foundation wall or slab to maintain the load path. 
Straps, anchor bolts, epoxy set anchors, and nails were the most common 
fasteners. When properly used, the straps, anchor bolts, and epoxy set bolts 
maintained the connection of sill plate and floor framing to the foundations 
for most wind conditions. However, numerous instances of anchor bolts 
without nuts or misaligned anchor bolts at the sill plate and floor framing 
resulted in the house lifting off the foundation. Nailing of the sill plate to the 
foundation was adequate only in the areas that incurred minimal damage 
from inflow winds along the periphery of the tornado paths. 

Wall framing in single-and multi-family houses failed at the sill plate (sole 
plate) to stud connection under all wind conditions. This was the most 
common failure observed by the BPAT in wall framing. Revisions in the 
normal way of constructing wall framing are necessary if these weak links 
are to be addressed. A positive method of connecting the studs to the sill 
plate that can resist reasonable uplift forces is a necessity for providing a 
continuous load path. 

Wood framed walls also saw failures at the double top plate connection with 
the wall and the roof systems. Attention must be given to ensure a positive 
connection is provided for the uplift load transfer from the double top plate to 
the wall below. Straps or other connectors that would ensure a continuous 
load path to resist uplift loads were not observed at this location. Nails were 
the primary fasteners at this connection. Failures were observed between the 
studs and the top plate and between the two top plates. Typically, when this 
connection failed, no continuous structural sheathing was observed to help 
with this load transfer. Full length wood structural panels, from the top plates 
to the sill plate or floor framing, could act well as the uplift load transferring 

7-2 
FEDERAL EMERGENCY MANAGEMENT AGENCY



PRELIMINARY REPORT PRELIMINARY CONCLUSIONS 


CHAPTER 7 
mechanism. The sheathing or other means of transferring the force must be 
connected to the double top plate by sufficient fasteners such as those 
required to attach sheathing as contained in the model building codes. 

The primary shear wall failure observed was that of garage return end walls 
that frame the garage door. The narrow walls where failure was observed 
possessed an aspect ratio (height to width ratio) that generally was less than 
recommended by industry or allowed by model building codes. The current 
building codes, which contain industry recommendations that are intended to 
provide a narrower shear wall, but yet be capable of resisting the design wind 
loads, should be followed. 

Although most of the roof framing configurations observed did not include a 
sufficient connection of the rafter to the ceiling joist, at least one of the 
model building codes does require such connection. In those cases where the 
ceiling joists existed and were parallel and adjacent to the roof rafters, 
additional resistance would have been provided if roof framing was 
connected to the ceiling joists. For the cases where the roof framing and 
ceiling joists were not parallel or adjacent, an insufficient number of 
observations were made to be able to draw any conclusions. 

Roof geometry was observed to affect building performance in two 
significant ways. First, the roof geometry affected both the local and overall 
wind loads acting on the roof. Second, the roof geometry affected the overall 
strength of the roof system based on its framing configuration (e.g., hip 
versus gable framing). In general, for residential (low rise) structures, the 
more complex shaped roofs experienced lower local loads (e.g., component 
and cladding loads such as individual roof deck sheathing loads) than did 
more simple roof geometry. The lower localized loads usually are the result 
of interference between the complex roof shape and the wind flow, which 
hinders the development of large negative pressure regions on the roof 
surfaces. 

In general, for flat, gable, and hip roof geometry, the largest localized loads 
occurred near the corners, the gable ends, and the edges of the roof ridge, 
respectively. However, the largest localized loads for gable roofs are 
noticeably higher than those for hip roofs. The net uplift forces acting on 
roofs are less dependent on roof geometry than are the localized component 
and cladding loads that are often the governing design loads when 
engineering standards such as ASCE 7-95 are used. As the total roof area 
increases, the contribution of intense localized loads to a total roof failure is 
reduced significantly. Although a localized load may fail a single piece of 
roof sheathing, it certainly will not cause the entire roof to fail. Such 
localized roof failures often allow rainfall to enter the structure causing 
significant collateral damage to the building interior and furnishings. When 
the roof fails as a single entity, it is the overall combination of all wind loads 
that will cause this failure. 

BUILDING PERFORMANCE ASSESSMENT: OKLAHOMA AND KANSAS TORNADOES 


7-3 

CHAPTER 7 
PRELIMINARY REPORT


The effect of roof shape on the performance of residential structures in high 
winds varies with the size of the component being considered (e.g., roof 
covering, roof sheathing, single truss, entire roof, etc.), the wind directions 
producing the high winds, and the method and quality of construction. 
However, hip roof systems are a stronger system because of the method by 
which they are constructed. 

7.1.1.2 Increased Load Caused by Breach of Envelope 
BPAT inspections of wind-induced damage to residential buildings indicate 
that internal pressurization is a major contributor to poor building 
performance under moderate to severe wind loading conditions. Field 
observations provide strong evidence of partial and total roof and exterior 
wall failures that may have been initiated due to breaches in the building 
envelope leading to internal pressurization and significant load increases. 
Roof and wall coverings, garage doors, entry doors, and windows that are 
exposed to severe or violent tornado vortex winds are not expected to 
survive. However, on the periphery of severe and violent tornado tracks and 
in moderate tornadoes where the wind speeds may be near or below design 
wind speed conditions, the performance of these elements was less than 
expected. If non-structural envelope elements are suitably designed and 
tested to meet the wind loads derived newer standards, such as ASCE 7-95 
and ASCE 7-98, and are appropriately installed, much of the damage on the 
periphery of severe and violent tornado tracks and in the track of the vortex 
of moderate tornadoes would be significantly reduced. An exception to this 
would be missile-induced damage. 

For residential buildings, a significant contributor to catastrophic failures due 
to internal pressurization appeared to be the failure of single skin, non-
insulated, and non-reinforced double width garage doors. Breaches of 
windows and entry doors might also cause significant damage to the 
residential building through internal pressurization. However, if wind speed 
and direction do not produce high local loads elsewhere on the building, the 
effect might not be as dramatic as that associated with a larger breach such as 
a residential garage door. Preliminary investigations determined that most 
garage doors are not rated or tested for wind pressures that may be calculated 
from the design wind speeds indicated in the currently enforced 1995 CABO 
Dwelling Code. Although this code does not address designing garage doors 
and other architectural finishes for the wind speeds prescribed in the code, if 
these doors had been designed for the design wind speed and associated wind 
pressures, damage in the inflow areas of the moderate and severe tornadoes 
might have been significantly reduced. 

The observed wind performance of T-lock asphalt shingles was not 
significantly better than that of three-tab or laminated strip asphalt and 
composition shingles. Wind-induced damage to T-lock shingles was 
observed on roofs that were likely exposed to wind speeds that were in the 

7-4 
FEDERAL EMERGENCY MANAGEMENT AGENCY



PRELIMINARY REPORT PRELIMINARY CONCLUSIONS 


CHAPTER 7 
range of current code prescribed design conditions (i.e., 70 mph, fastest mile 
or 90 mph 3 second peak gust). 

Vinyl siding offers very limited resistance to low-energy missiles. The vinyl 
siding investigated also offered limited wind load resistance. Although the 
nailing patterns were erratic and the distance between nails was relatively 
large, it is difficult to envision that the investigated products had sufficient 
strength to meet the wind loads derived from new codes and standards such 
as ASCE 7-95 and ASCE 7-98. 

7.1.1.3 Masonry 
The BPAT observed extensive brick veneer loss in homes of all ages, 
indicating inadequate composite action caused by a failure of the brick ties. 
Masonry veneer and framed walls are generally assumed to provide some 
level of composite action to resist dynamic high wind forces, even though 
this is not considered explicitly in design. However, to act as a composite 
section, the connection between the veneer and backup wall (normally 
galvanized steel brick ties) needs to be maintained. Extensive degree of brick 
veneer loss in homes of all ages indicates inadequate composite action 
caused by a failure of the brick ties or due to the flexibility (relative to 
stiffness of the wall) of brick ties as installed. Some walls appeared 
undamaged, but could be deflected with hand pressure. 

Many of the failures observed stemmed from brick tie bond failure. In a 
majority of cases of masonry veneer loss, either corrugated or scalloped-edge 
galvanized steel brick ties remained attached to wall studs with one 6d 
common nail (withdrawal load = +/- 30 lb times a safety factor of 4 or 5), 
when a non-structural foam sheathing was used. The bond between mortar 
and brick tie was not sufficient to even exceed the withdrawal capacity of the 
tie nail. Therefore, there was inadequate bond between mortar and brick tie 
to resist the high wind forces experienced. The 1995 CABO One and Two 
Family Dwelling Code specifies that the maximum horizontal spacing of 
brick ties is 24-in on center, and each tie shall support not more than 3.25 sq. 
ft of wall area. At the code-required spacing to support 3.25 sq. ft, the 
maximum wind suction pressure on the veneer prior to failure could not have 
exceeded 37 psf, unless the rigid brick facing failed prior to the deflection 
required to allow the brick tie to develop its full capacity. 

There were a few instances of nail pull-out at brick ties fastened to wall 
studs. Therefore, in these cases the wind suction pressure exceeded the 
withdrawal strength of the one nail holding the brick tie. Causes of failure 
could be insufficient nail length or diameter, low withdrawal resistance, or 
ties having too high a tributary area. 

The 1995 CABO One and Two Family Dwelling Code also requires that if 
sheet metal is used, it shall not be less than No. 22 U.S. gauge by 7/8-in 
corrugated. There were many instances of brick ties spaced at greater 

BUILDING PERFORMANCE ASSESSMENT: OKLAHOMA AND KANSAS TORNADOES 


7-5 

CHAPTER 7 
PRELIMINARY REPORT


distances. The most common form of tie was a 7/8-in wide galvanized steel 
strip with a ¼-in deep scalloped edge on each side (center strip was 3/8-in 
wide), with very minor corrugation (less than 0.5 mm). There was notable 
absence of code compliance in what could be considered a random sample of 
homes impacted by the tornadoes. 

Because failures of masonry veneer were found at homes from less than 1 
year old to over 20 years old, mortar bonding strength did not seem to vary 
with age. Most brick used in the Oklahoma City area seemed to be of high 
density and low porosity, which could have affected bonding of mortar. 
There were several instances of loose brick on the ground with no mortar 
attached or only attached to one side. Mortar bond strength was inadequate to 
bond bricks together and to bond mortar to brick ties to resist negative 
(suction) wind pressures experienced. Some possible causes could be from a 
weak mortar mix, a too dry mortar, or use of low porosity brick. 

There were several instances where an air space between veneer and plastic 
foam insulation sheathing was 1.5-in or more, which reduced embedment 
length of brick ties in mortar joints to 1-in or less. Some model building 
codes specify 1-in maximum air space or grouted space, and 1.5-in minimum 
embedment of brick tie into mortar. 

7.1.2 Manufactured Housing 
The design and construction of manufactured housing has been governed 
since 1976 by Federal preemptive standards that are enforced by the U.S 
Department of Housing and Urban Development (HUD) under Federal 
Regulation and through a Monitoring and Enforcement Contractor, the 
National Conference of States on Building Codes and Standards (NCSBCS). 
Recently, the HUD Standard has been placed under a consensus process 
administered by National Fire Protection Association (NFPA). 

Wind resistance standards for manufactured housing differ from and are less 
than model building code provisions and standards for conventional site-built 
and modular or panelized construction. Minimum wind pressures for design 
of all homes located outside of the hurricane coastline are 15 psf for 
horizontal wind loads and 9 psf for net uplift load (equivalent to about a 65 
mph fastest mile wind speed, less than the 70 mph fastest mile wind speed 
specified in the CABO One and Two Family Dwelling Code, and lass than 
the 80 mph fastest mile wind speed specified in the 1997 UBC for this area 
of the country). Explicit engineering or test-based performance provisions 
require a minimum safety factor of 1.5 relative to these nominal design loads. 
However, simplified nominal design wind loads and the required safety 
factors do not consider the rare but significant overload that may occur due to 
inflow winds of violent and severe tornadoes or direct strike by the vortex of 
any tornado. Nominal loads are primarily associated with the level or risk 
that is associated with extreme thunderstorm winds. Also, the affects of 

7-6 
FEDERAL EMERGENCY MANAGEMENT AGENCY



PRELIMINARY REPORT PRELIMINARY CONCLUSIONS 


CHAPTER 7 
exposure (not necessarily a factor for tornadic events) are not considered in 
nominal design wind loads. 

Installation and setup of manufactured housing, including foundations, 
ground anchors, and strapping or cables, are enforced by State and local 
officials. The Federal Standards only address the design of the overall 
anchoring and tie-down systems and require that they be designed by a 
qualified professional. 

In general, manufactured housing did not resist wind forces as well as 
conventional site-built detached single-family dwellings for inflow winds of 
violent and severe tornadoes and vortex winds from all tornadoes. This was 
primarily because of inadequate fastening of roof systems to wall systems 
and inadequate resistance to uplift and overturning provided by anchorage 
and tie-downs. An exception to this was the observed improved performance 
of newer manufactured homes that had been installed on permanent 
foundations. 

7.1.2.1 Foundations 
Permanent foundations performed better in resisting lateral wind loads than 
did ungrouted and unreinforced CMU piers having wood leveling shims 
under the chassis beams. However, the BPAT observed that connections of 
chassis and perimeter joists to permanent foundations were inadequate to 
resist the moderate wind uplift and overturning forces generated at the 
periphery of most tornado track investigated. It is difficult to make positive 
connections and then to inspect once the units are erected. In addition, local 
building officials the BPAT interviewed did not seem to be aware of 
manufacturer’s installation or set-up instructions with specific connection 
requirements for permanent foundations. 

7.1.2.2 Anchors 
Depths and locations of helical ground anchors and soil conditions varied 
considerably from site to site. Ground anchors pulled out of the soil because 
of inadequate depth, or steel anchor shafts bent over from lateral wind forces, 
thus leading to failure of the superstructure. Some ground anchors were 
installed at an angle with the base under the home, leading to bending of the 
shaft from lateral wind forces. Thus, deformation of the anchor and strapping 
arrangement could allow significant movement (vertically and horizontally) 
prior to developing substantial resistance to wind loads. Most ground anchors 
did not appear to comply with requirements of the Federal Manufactured 
Home Construction and Safety Standards (MHCSS), which states the 
following: 

“Sec. 3280.306(f) Anchoring equipment shall be capable of resisting 
an allowable working load equal to or exceeding 3,150 pounds and 
shall be capable of withstanding a 50 percent overload (4,725 pounds 

BUILDING PERFORMANCE ASSESSMENT: OKLAHOMA AND KANSAS TORNADOES 


7-7 

CHAPTER 7 
PRELIMINARY REPORT


total) without failure of either the anchoring equipment or the 
attachment point on the manufactured home.” 

In 1994, the standard was revised to add Sec. 3280.306(b)(2) For anchoring 
systems, the instructions (provided by the manufacturer) shall indicate: 

“(ii) That anchors should be certified by a professional 
engineer, architect, . . . as to their resistance, based on the maximum 
angle of diagonal tie and/or vertical tie loading . . . and angle of 
anchor installation, and type of soil in which the anchor is to be 
installed; (iv) That ground anchors should be installed to their full 
depth, and stabilizer plates should be installed to provide added 
resistance to overturning or sliding forces.” 

7.1.2.3 Strapping 
Galvanized steel strapping in several instances failed in tension from wind 
uplift and overturning forces, or became loose when the home moved 
laterally from wind forces. In addition, connections of strapping to chassis 
beams often came loose and were on the ground, and there was no positive 
bolted or welded connection. The apparently premature failure of these ties 
was related to the number of ties, location of first ties from end of chassis, 
and tensile strength or ductility of steel. Several of the following provisions 
of the Federal MHCSS appeared to not be consistently complied with, 
possibly leading to failure: 

“Sec. 3280.306(c)(1) The minimum number of ties required per side 
shall be as required to resist the design loads . . . 

(2) Ties shall be evenly spaced as practicable along the length of the 
manufactured home with not more than 8 feet open-end spacing on 
each end.” (This provision was revised in 1994 to require not more 
than 2 feet open-end spacing on each end). 
The Material Specification for “Strapping, Steel, and Seals, with 
Notice #1 and Amendment #2, only Type 1, Finish B, Grade 1 of the 
plating/coating sections are applicable,” was Federal Spec. FS QQS-781H-1974 with 1977 amendments. (This was revised in 1994 to 
“Standard Specification for Strapping, Flat Steel and Seals – ASTM 
D3953-91”). 

7.1.2.4 Superstructure 
Generally, newer manufactured housing units, particularly multi-wide units 
on permanent foundations, resisted straight-line inflow wind forces better 
than older single-wide units. Newer units are generally constructed of more 
conventional wall and roof framing, and connections between roof systems 
and walls, and walls to floors, provide load paths to transmit wind uplift, 
lateral, and overturning forces to the foundations. Internal shear walls, and 

7-8 
FEDERAL EMERGENCY MANAGEMENT AGENCY



PRELIMINARY REPORT PRELIMINARY CONCLUSIONS 


CHAPTER 7 
bolted or steel strapped floors and roofs of multiple units at marriage walls 
provide a stiffer three-dimensional structure. However, attention does need to 
be paid to uplift straps from roofs to walls and walls to floors, and to bolting 
of units to permanent foundations, similar to conventional site-built home 
construction in tornado-prone areas. 

7.2 NON-RESIDENTIAL PROPERTY PROTECTION 
Visual observations indicated that non-residential structures were, with few 
exceptions, as vulnerable to damage as conventionally built residential 
construction. Many non-residential buildings received structural damage as a 
result of lack of redundancy in the load path to resist wind-induced uplift 
loads. Observed damage, however, was not as complete or devastating for 
non-residential buildings that were exposed to similar vortex winds of violent 
and severe tornadoes as that observed in residential construction. This was 
primarily due to the engineering that is required by model building codes for 
non-residential buildings and that is not required for residential buildings. 

Non-residential construction in Oklahoma is currently required to be 
designed per 1996 NBC and non-residential construction in Kansas is 
designed per the 1994 and 1997 UBC, depending upon local jurisdiction. 
Although local municipalities have adopted some amendments, these 
amendments were not significant relative to the structural issues discussed in 
this report. For current construction, these model building codes provide 
guidance for loads other than gravity loads. However, engineering standards 
such as ASCE 7-95 and 7-98 provide better structural and non-structural 
design guidance for wind loads than these newer codes. Although designing 
for tornadic wind events is not specifically addressed in any of these newer 
model building codes or standards, constructing non-residential buildings to 
these codes and standards would improve the strength of the built 
environment. Building to ASCE 7-95 or 7-98 would have led to reduced or 
minimized damage in areas that were affected by the inflow winds of all 
tornadoes and reduced the damage observed where moderate tornadoes 
impacted non-residential construction. 

7.2.1 Load Path 
Although non-residential construction is currently designed to specifically 
consider some wind load resistance, a lack of attention to uplift and lateral 
loads resulted in failure to provide a continuous load path and greatly 
increased damage to the buildings. In many cases, structural damage would 
have been negligible if adequate uplift resistance had been provided to steel 
roof joists and metal roof deck systems. Additional resistance to uplift could 
have significantly reduced damage to engineered construction on the 
periphery of severe and violent tornadoes or a moderate tornado track. 

Construction with materials such as URM block and brick that is capable of 
carrying gravity loads, but unable to carry uplift loads, will continue to lead 

BUILDING PERFORMANCE ASSESSMENT: OKLAHOMA AND KANSAS TORNADOES 


7-9 

CHAPTER 7 
PRELIMINARY REPORT


to wall and roof failures during moderately high wind events. Better attention 
to the design of and selection of materials for connections throughout the 
structural system will also minimize and reduce the number of failures that 
are currently observed in non-residential construction after moderately high 
wind events such as along the periphery of severe and violent tornadoes or in 
the tracks of moderate tornadoes. 

After roof decking and other parts of the structure were blown loose by the 
wind, these pieces became windborne missiles that created additional 
damage to nearby structures. Greater attention to attachment of perimeter 
wood nailers, copings and metal edge flashings, and perimeter attachment of 
metal roofing panels will enhance performance of roof coverings and reduce 
the debris on the periphery of severe and violent tornadoes and in the tracks 
of moderate tornadoes. 

7.2.2 Increased Load Caused by Breach of Envelope 
The BPAT observed that the failure of commercial overhead doors, 
depending on their location, may initiate or contribute to major failures of 
primary structural systems. Observations suggest that overhead doors failing 
near building corners may significantly contribute to catastrophic failures of 
exterior walls and roof systems. This is particularly true for pre-engineered 
metal (light-steel frame) buildings that typically have little redundancy in 
their load transfer paths. For buildings that have several interior rooms or 
partitions, the propagation of internal pressures may be hindered and 
collateral damage to exterior walls minimized. 

Breach of the building envelope was observed to result in extensive collateral 
damage to non-residential buildings. Garage doors and large windows were 
particularly vulnerable. All garage and overhead doors should have adequate 
strength to resist wind loads derived from the latest engineering standards, 
such as ASCE 7-95 or 7-98, that provide design guidance for non-structural 
elements such as garage doors and windows. Also, owners of buildings that 
use EIFS for exterior walls should be advised by the building designer that, 
although the wall has the appearance of concrete, it offers insignificant 
resistance to high wind pressures and windborne missiles unless the EIFS is 
over concrete or reinforced CMU. 

A large number of missiles, which had been generated from roofs on 
essential facilities (e.g., hospitals), and buildings such as schools were 
observed. Aggregate and paver surfacing should not be used for roofs 
because they can be picked up by winds and cause significant damage to 
architectural finishes, windows, and doors. 

Protection of windows from wind pressures and windborne debris was not 
extensively investigated by the BPAT. However, it is important to consider 
protecting glass in essential facilities. Laminated glass, like shutter protection 
systems, can offer substantial protection from modest-energy windborne 

7-10 
FEDERAL EMERGENCY MANAGEMENT AGENCY



PRELIMINARY REPORT PRELIMINARY CONCLUSIONS 


CHAPTER 7 
missiles. Laminated glass has the potential to offer significant occupant 
protection along the periphery of severe tornado tracks and in the tracks of 
moderate tornadoes and is a permanent protection device that does not need 
warning time to be installed, which can be a problem with many storm 
shutter systems. 

7.3 PERSONAL PROTECTION AND SHELTERING 
The best way to reduce loss of life and minimize personal injury during any 
tornadic event is to take refuge in a specifically designed tornado shelter. 
Although improved overall construction may reduce damage to buildings and 
contribute to safer buildings, an engineered shelter is the only means of 
providing individuals near absolute protection from severe and violent 
tornadoes. 

7.3.1 Residential Shelters 
The residential shelters observed by the BPAT included above-ground in-
residence shelters and storm cellars. Although the above-ground in-residence 
shelters provided safety for the occupants, no direct missile strikes were 
recorded on the shelter doors that the BPAT was able to locate and visit. The 
doors observed were light gauge hollow metal with a single deadbolt locking 
device, which is less than the 14 gauge hollow metal door held by three 
hinges and three deadbolts, as required in FEMA 320: Taking Shelter From 
the Storm: Building a Safe Room Inside Your House. 

Assuming proper construction and location outside flood-prone areas, storm 
cellars offered safety during severe wind events. Observed problems with 
storm cellars included lightweight doors and hardware, poor maintenance, 
and unprotected ventilators. Storm cellars are typically not fully 
waterproofed and therefore can be damp, musty environments with poor 
ventilation. Ventilators were not constructed of heavy gauge steel or 
protected by heavy gauge shrouds or saddles that would have prevented their 
removal by debris or extreme winds during a tornado, allowing the 
subsequent entrance of free falling missiles and debris through the remaining 
openings in the shelter roof. 

7.3.2 Group Shelters 
The BPAT observed group shelters at a manufactured housing rental 
development and at a plastics manufacturing plant in Haysville, Kansas. A 
rental development of manufactured homes provided shelters at a rate of one 
shelter per four homes. Shelters were located in close proximity to the homes 
and were quickly accessible by the occupants, but none of these shelters were 
easily accessible to persons with disabilities. All group shelters were below 
or partially below ground and required access by stairs. 


BUILDING PERFORMANCE ASSESSMENT: OKLAHOMA AND KANSAS TORNADOES 


7-11 

CHAPTER 7 
PRELIMINARY REPORT


The group shelter at the plastics manufacturing plant functions daily as a 
conference room and lunchroom. On May 3, 1999, it performed its third 
function as a tornado shelter. Although the building housing the shelter was 
not significantly damaged (one area suffered roof damage), other buildings 
that are part of the plant complex suffered substantial damage. The workers 
at the plant when the tornadoes struck and who were able to utilize the 
shelter were uninjured. 

7.3.3 Community Shelters 
The BPAT observed two community shelters that were utilized during the 
May 3 storm. One shelter was located in a manufactured housing park in 
Wichita, Kansas. The second shelter was located in Mid West City at the Mid 
West City High School gymnasium. Both were partially below-ground 
shelters and suffered from problems of moisture infiltration, mustiness, poor 
ventilation, and poor exterior doors and hardware. Other concerns common 
to community shelters include travel time required to access the shelter, 
accessing the shelter when the shelter is locked, accessibility for persons with 
disabilities (ADA compliance), and rules for gaining admittance. 

7.3.4 Other Places of Refuge 
Not all buildings, residential or non-residential, have designated tornado 
shelters or staffs with tornado plans for implementation during an event. 
Subsequently, in buildings without designated shelters or places of refuge, 
occupants are left on their own to identify places of refuge appropriate in a 
tornado event. The observations of the Oklahoma City and Kansas tornadoes, 
as well as other tornado events, indicate that small interior rooms within 
buildings often survive when the other portions of the building are destroyed. 
Rooms such as closets beneath staircases, small bathrooms, or other small 
rooms are the preferred place of refuge when no hardened shelter is provided 
in the building. 

Basements can also offer another alternative place of refuge. However, the 
observed basements demonstrated vulnerability from windborne missiles 
through windows, window wells, and through the above wood floor/ceiling 
structure. Although not observed in this storm event, previous observations 
have shown unreinforced basement walls collapsed as the result of the 
floor/ceiling diaphragm displacement by the winds of the tornadoes. 

The BPAT visited public use facilities during the field investigations to 
determine how these facilities addressed tornado threats that affect the users 
of the facilities. The team interviewed staff at schools, day-care centers, 
nursing homes, and churches, and found that not all public use facilities had a 
formalized tornado emergency refuge plan. Additionally, not all public 
facilities had a NOAA weather radio in continuous operation to monitor 
storm events that may lead to a tornado. When tornado plans were 
implemented by a facility, these plans were often not conspicuously posted 

7-12 
FEDERAL EMERGENCY MANAGEMENT AGENCY



PRELIMINARY REPORT PRELIMINARY CONCLUSIONS 


CHAPTER 7 
and the plans were not always exercised as drills so building occupants could 
become familiar with the plan. It is unclear whether all plans allow sufficient 
time for the building occupant type (e.g., children, elderly, etc.) and if the 
shelter had adequate capacity for the quantity of building occupants and 
others who may attempt to seek shelter in the planned place of refuge. 

BUILDING PERFORMANCE ASSESSMENT: OKLAHOMA AND KANSAS TORNADOES 


7-13