In the wake of major hurricanes suchas Katrina that recently devastatedmuch of the Gulf Coast area nearNew Orleans, roofing professionalsare often asked to investigate thecause of roof wind damages. A key to

many damage investigations is approximat-ing the maximum gust wind velocity experi-enced at roof level. Although wind velocitydata are usually available from the NationalWeather Service (NWS), theNational Oceanic and Atmo-spheric Administration(NOAA), and other sources,the data are usually for ter-rain exposures, heightsabove ground, and averagingtimes different than what isneeded. This article discuss-es how to adjust wind veloci-ty data for site-specific ter-rain exposures, heights, andaveraging times.

Wind Warranty ClaimTo illustrate how these

adjustments work, we willuse the following hypotheti-cal example scenario:

On August 29, 2005,a 2-year-old, built-uproof covering on a 45-foot tall building in amostly suburban andwooded area south-east of New Orleans,

Louisiana, lifted near one cornerand peeled back (see Figure 1). Thebuilding owner submitted a warran-ty claim, but it was denied because,according to the manufacturer, windgusts in the greater New Orleansarea exceeded those covered by thewarranty. The warranty excludescoverage for damages experiencedby wind gusts greater than 100

mph at the roof. The building ownerhires you to provide an approxima-tion of the maximum wind gustvelocity experienced at roof level onhis building. No nearby wind veloci-ty data are available, but the site iswithin the geographic region coveredby wind swath data published byNOAA shortly after HurricaneKatrina made landfall.

Figure 1: A built-up roof covering lifted and peeled back near one corner.

O C T O B E R 2 0 0 5 I N T E R F A C E 4 1

NOAA H*WindFigure 2 shows an excerpt from wind

swath data published by NOAA shortly afterHurricane Katrina made landfall on August29, 2005. The graph shows the estimatedmaximum sustained winds in miles perhour along the hurricane path at a height of33' (10 meters) in open terrain (exposure C).Sustained winds are defined as the maxi-mum velocity averaged over a 1-minuteinterval. See the NOAA Hurricane ResearchDivision, Surface Wind Analysis website(www.aoml.noaa.gov/hrd/data_sub/wind.html) for information about how the datawere gathered and reduced to graphicalform.

Because many roofs are not positionedat a height of 33' in open terrain, andbecause current wind design guides (andmany wind warranties) are based on three-second gusts, not one-minute winds, theNOAA wind swath data often need to be

adjusted for height, exposure, and/or aver-aging time before they can be used.

Approximate Sustained WindsApproximating the maximum sustained

winds experienced at a building site is sim-ply a matter of interpolation between thewind isotach lines on the NOAA H*Windmap. Assuming the example building site islocated at the asterisks on Figure 2 yields amaximum sustained wind velocity of 85mph.

Adjust For Averaging TimeThe H*Wind data are wind velocities

averaged over one minute. Wind velocitiesreferred to in most building codes and mostroof manufacturers warranties are windgusts. Wind gusts are typically wind veloci-ty data averaged over three seconds. TheDurst Curve (Durst 1960 and ASCE 7-02)can be used to adjust one-minute windvelocities to equivalent wind gust velocities.

The Durst Curve (Figure 3) shows howwind velocities averaged over t secondscompare to wind velocities from the samewindstorm averaged over one hour (3600seconds). The ratio is 1.0 for an averagingtime of one hour and something higher forshorter averaging times.

For our example, the Durst Curve indi-cates a 103 mph 3-second gust wind veloc-ity is considered equivalent to an 85 mph 1-minute wind velocity. The math for this isfound in Formula 1.

Adjust for HeightWinds in a wind stream are assumed to

be moving at a constant velocity above acertain height (i.e., the gradient height, Zg).Winds in this same wind stream but belowthe gradient height are moving at a slowervelocity. They are slowed by frictionbetween the wind stream and the ground.Figure 4 illustrates how the rate of windstream slowing increases as the terrain getsrougher [source: Texas Tech University].

The wind velocities discussed in ourexample so far are applicable to a roofheight of 33' (10 meters). However, while thewind is blowing at 103 mph at a height of33', it is blowing faster at greater heights.

The Power Law can be used to adjustwind velocity data from one height to anoth-er. The Power Law equation shows the rela-tionship between the wind velocity at agiven height (Vz) and the gradient velocity(Vg). (See Formula 2.)

Figure 5 provides numeric values forgradient heights (Zg) and the alpha expo-nents for different terrain exposure cate-gories [Davenport, 1960].

For our example, the Power Law is usedin a two-step manner to indicate that a 107-mph gust measured at a height of 45' isconsidered equivalent to a 103-mph gustmeasured at a height of 33'. The math forthis is found in Formula 3.

Adjust for TerrainThe wind velocities discussed so far are

applicable to open terrain (Exposure C), aterrain similar to that surrounding mostairports. Since the rate at which a windstream is slowed by friction depends on

Figure 3: The Durst Curve [Durst 1960 and ASCE 7-02] shows how wind velocitiesaveraged over t seconds compare to wind velocities from the same wind storm averagedover 1-hour (3600 seconds).

Formula 2

Figure 2: Excerpt from Hurricane Katrinawind swath data published by NOAA.

4 2 I N T E R F A C E O C T O B E R 2 0 0 5

Formula 1

ground roughness, wind velocities need tobe adjusted if the building is in a terrainexposure other than C. See the commentaryof ASCE 7, Minimum Design Loads forBuildings and Other Structures fordescriptions and photo examples of differ-ent terrain exposures.

The gradient velocity of a given windstream is the same no matter what the ter-rain exposure is. The Power Law can be usedto adjust wind velocities from one exposurecondition to another. Referring again toFigure 4, one could say terrain-adjustedwind velocities are approximated by slidingup to the gradient velocity along one terraincurve and sliding down another.

The suburban and wooded terrainsurrounding our example building is classi-fied as terrain Exposure B. Using the previ-ous approximation that the wind stream atthis site has a gradient wind velocity of 146mph, the Power Law (with exposure B val-ues), indicates the wind gust velocity at 45'in Exposure B would be 92 mph, which isless than the 107 mph approximated forExposure C. The math for this final step isshown below in Formula 4.

Since the final approximated wind gustvelocity of 92 mph at roof level in our exam-ple is less than 100 mph, the buildingowner thanks you and submits your infor-mation, along with his warranty claim forreconsideration.

Be CarefulWarranties are legal documents and

require legal assistance to interpret. Roofingprofessionals are encour-aged to limit the informationthey offer as part of warran-ty claims to technical con-siderations.

The procedures dis-cussed provide approxima-tions of site-specific windvelocities based on general-ized wind field type data.Wind velocities at specificsites can vary significantlyfrom these types of general-ized data. Investigators areencouraged to supplementand to corroborate NWS andNOAA data with wind veloci-ty data obtained from othersources. More sophisticatedanalytical procedures, in-cluding wind tunnel model-ing and/or site instrumenta-tion, are recommendedwhen data beyond approxi-mations are desired.

This article focuses on the velocity of thewind stream as it approaches a building. Itdoes not address how wind velocities areaffected as the wind stream is diverted up,over, and around a building and it does notaddress how wind streams are affected byupstream terrain features such as escarp-ments and valleys. Readers are directed toASCE 7 for more information.

This article also does not address howroof wind damages can start and progress,yet these considerations are important interms of what may or may not be covered byroof warranties.

SummaryA key to many wind damage investiga-

tions is approximating the maximum gustwind velocity at roof level. Wind velocitydata available to investigators, however, areoften for heights, terrain exposures, or aver-aging times different than what is needed.For example, the NOAA H*Wind data repre-sents one-minute sustained wind velocitiesat a height of 33' in open terrain (ExposureC). The Durst Curve and the Power Law canbe used to adjust NOAA data, as well asdata from other sources, for desired averag-ing times, roof heights, and terrain expo-sures.

AcknowledgmentsThe author wishes to thank Jim

McDonald, PhD, PE, with McDonald, Mehta& Yin for useful comments and suggestionsoffered during preparation of this article.

Figure 4: Illustration of how the rate of wind streamslowing increases as the terrain gets rougher [source:Texas Tech University].

O C T O B E R 2 0 0 5 I N T E R F A C E 4 3

Formula 4

Figure 5: Numeric values for gradient heights (Zg) and the alphaexponents for different terrain exposure categories [Davenport,1960].

Formula 3

ReferencesDavenport, A.G., A Rationale for the

Determination of Design WindVelocities, Proceedings ASCEStructural Division, 86: 39-66, 1960.

Durst, C. S., Wind Speed Over ShortPeriods of Time, MeteorologicalMagazine, 89, 181-187, 1960.

SEI/ASCE 7-02, Minimum DesignLoads For Buildings and OtherStructures, ASCE/SEI, 298, 2003.

Philip Dregger, PE, RRC, FRCI

Philip Dregger, PE, RRC, FRCI, is president of Technical RoofServices, Inc., Concord, CA. He is a past region director ofRCI, a former faculty member of RIEI, and serves as RCIsrepresentative to RICOWI. Mr. Dregger has designed roof andwaterproofing systems to meet some unusual and challengingrequirements of clients, including the University of California,Pacific Bell (now SBC Communications), Kaiser Hospitals,Mervyns, and Disney. He has investigated numerous roof andwaterproofing problem conditions, including damages sus-tained from major hurricanes and earthquakes. Dregger is the author of several articleson roof technology, including The Role of Air Retarders Deserves Closer Scrutiny, ref-erenced in the National Wind Design Standard ANSI/SPRI/RP-4.