WindBorne Debris Criteria for the Florida Panhandle

1
Wind-Borne Debris Criteria for the Florida
Panhandle

• June 19, 2006
• Florida Department of Community Affairs Draft
  Final Briefing

Prepared by Lawrence Twisdale, Peter Vickery,
Justin Chen, Chris Driscoll, Dhiraj Wadhera,
Jeff Sciaudone, William York Applied Research
Associates, Inc. Raleigh, NC Orlando, FL Contact
L.A. Twisdale 919-582-3336
Draft v 1.00
2
Outline

• 0. Pre-Project What Did We Know About Wind
  Borne Debris?
• 1. Project Overview
• 2. Wind Tunnel Tests
• 3. WBD Damage Observations and Model Validation
• 4. Selected Panhandle Houses and Subdivision
• 5. Simulate House Performance
• 6. Quantify Risks, Benefits and Costs
• 7. Conclusions
• Recommendations

3
0. Pre-Project What Did We Know About Wind
Borne Debris?
4
Pre-Project Preamble
5
Pre-Project Preamble (contd)
This project has focused on integrating hurricane
WBD data, hurricane models, damage and loss
models to evaluate the need for WBD protection in
the Florida Panhandle.

• 7. ASCEs WBD Criteria was Largely
  Judgment-Based
• 1996 vintage
• 120 mph region ignores terrain effects
• However, ASCE pressure loads considers terrain
• Includes 110 mph within 1 mile of coast
• No formal risk or benefit cost analysis was
  performed

6
1. Project Overview
7
Objective - Perform risk, benefit, and cost
assessment of hurricane wind-borne debris
protection options for the Florida Panhandle
8
Project Tasks

• Wind Loads. Perform wind tunnel tests for houses
  located in treed environments characteristic of
  the Florida Panhandle. Develop velocity profiles,
  turbulence intensity, and pressure load models
  for houses in treed environments.
  Compare/integrate with FCMP data.
• Model Representative Houses. Select
  representative Panhandle houses and develop
  computer models of these houses for analysis of
  wind borne debris protection effects.
• Update/Validate Wind Borne Debris Model. Update
  ARAs wind borne debris model to reflect new load
  models and impact resistance data. Perform
  validation comparisons on window/building
  performance with available field data on recent
  Florida Hurricanes.

9
Project Tasks (contd)

• 4. Simulate House Performance. Perform hurricane
  simulations of the representative houses located
  at various positions in the Panhandle. Evaluate
  building damage and loss with and without
  windborne debris protection.
• 5. Quantify Performance, Risks, Benefits, and
  Costs of Wind Borne Debris Protection. For each
  house in each Panhandle location, develop the
  detailed loss reduction data, costs, and risks.
  Summarize the results and present findings.

10
House Locations WBD Protection
ASCE Contours

• Each of the 6 house models will be placed on
• 110 mph contour
• 120 mph contour
• 130 mph contour
• Terrains modeled will include
• Open-Suburban (no trees)
• Suburban (no trees)
• Light Trees - Suburban
• Medium Trees -Suburban

• Alternative WBD Protection Options
• Option A No WBD Protection
• Option B Steel Panel Shutters with Accordian on
  2nd Floor
• Option C Plywood Shutters
• Option D Impact Resistant Glazing

11
Terrain Modeling is Critical
Beach
Town
Suburban
Forests
BarrierIsland
Suburban
Bay
Ocean
Z
V(z)
V(z)
V(z)
Displacement Height
X
Open to SuburbanTransition
Suburban Light and Medium Trees
Open

• Boundary layer transition will be affected by
  tall trees and influence windspeeds, loads, and
  WBD environment on one and two story residences

12
2. Wind Loads
13
Wind Loads

• Objectives
• Perform Wind Tunnel Tests to
• Determine reduction in wind speeds at eave
  heights due to trees
• Determine changes in pressure coefficients due to
  trees
• Incorporate results in wind borne debris modeling
  study
• Wind Tunnel Testing
• Seven Different Terrain Types
• Open, suburban, 5 tree variations
• Mean and turbulence intensity profiles for each
  case
• Pressure measurements on one and two story gable
  roof houses for each case two story hip roof for
  selected cases

14
Wind Tunnel Experiments
15
Tree Characteristics

• Two different tree densities
• 50 Tree
• Effective area (CdA 135 ft2)
• 75 Tree
• Effective area (CdA 240 ft2)
• Equal number of 50 and 75 trees in all cases
• Maximum of 400 between forest and model house

16
Modeled Terrain Characteristics

• Open Terrain
• Suburban Terrain
• Medium Density Trees
• 800 upstream 400 surrounding House
• 800 upstream 400 _at_ 50 density surrounding
  house
• 800 upstream 400 clear cut around house

• Light Density Trees (50 of Medium Density Trees)
• 800 upstream 400 around house
• 800 upstream 400 clear cut around house
• Some additional tests performed with surrounding
  buildings

MeasuredPressureTime-Series
17
Velocity Profiles(Shows how trees affect winds)
Large velocity reductions within first 40 ft of
ground
18
Peak Gust Wind Speeds Pressures (One Story)
Normalized Peak Gust
Normalized Kz
19
Pressure Reduction One Story
20
Pressure Reduction Two Story
21
FCMP Full Scale vs Wind Tunnel (Trees)

• Reasonable agreement, but note
• Building geometries are different
• Distance to neighboring buildings
• FCMP 5 apart
• Wind Tunnel 20 apart
• Peaks in worst case agree well
• Close neighbors expected to reduce peaks for some
  wind directions

22
Wind Load Summary Treed Terrain

• Trees significantly reduce wind speeds at eave
  heights of homes
• GCp values (referenced to gust speed at eave
  height) on roof increase with the existence of
  trees (10-40 increase)
• Reduction in wind speed somewhat offset by
  increase in GCp
• GCp values all greater than given in ASCE 7
• Typical pressure reductions associated with
  trees
• 30-40 for light tree case
• 50-60 for medium tree case
• Reductions in loads not as big on two story house
  compared to one story
• GCp and velocity profile data have been
  incorporated in the windborne debris damage and
  trajectory models.

23
Wind Load Summary Suburban Terrain

• Wind tunnel tests indicate pressure coefficients
  (for components and cladding) in ASCE 7 are too
  low for much of the roof.
• Positive pressures over much of the walls
  (windows) are underestimated using ASCE-7.
• Negative wall pressures underestimated over much
  of the walls (Zone 4). Overestimated in zone 5
  (edge zone).
• Pressure coefficients increase with increasing
  turbulence (even when normalized to the peak gust
  wind speed at eave height).
• Underestimate of pressure coefficients in ASCE-7
  previously identified by studies performed by
  Reinhold.
• More wind tunnel tests (different roof slopes and
  terrains) required to enable the new information
  to be incorporated into the next edition of
  ASCE-7 (or FBC)

24
Summary Light and Medium Tree Terrains
Tree Terrain Parameters

• We used two of the five tree terrains tested
  for the damage and loss modeling
• Light ? 34 trees/acre
• Medium ? 69 trees/acre
• Light trees ? subdivision with light tree buffer
  and clear-cut within subdivision
• Medium trees ? subdivision with medium tree
  buffer and some trees in subdivision
• Results valued for subdivision ? 400 ft

25
3. WBD Damage Observations and Model Validation
26
Wind-Borne Debris Model
Source (95E-5-20)

• Model was developed between 1995-1998
• Approach involves simulating hurricane winds,
  modeling building failures and debris sources
• Entire subdivisions are modeled
• Impact statistics are developed for use in single
  building analysis

Transport Summary
27
Wind-Borne Debris Simulation

• Full hurricane wind trace
• Geometric model of subdivision
• Building envelope modeled for each house
• Component failures
• Debris transport
• Impact energy and momentum

28
Model Validation

• Original Model Validation included
• Building performance for house components
• Component resistance based on test data where
  available
• Aerodynamic and transport models
• Windborne debris transport for wood missiles
• Current validation effort includes 6 cases
• Hurricane Bonnie (ARA)
• Hurricane Charley (Tile and Shingle, UF, ARA))
• Hurricane Andrew (2 Locations, NAHB)
• Hurricane Ivan
• WBD (UF)
• Tree Blowdown (ARA)

• Shingle transport calibration has become a key
  question of this effort

29
WBD Model Validation Steps
2. Site-Specific Hurricane Winds
1. Post-Storm Ground Survey
3. Simulated Trace
5. Simulated WBD and Building Damage
4. Model Subdivision
6. Compared Observed and Simulated Results
30
Hurricane Charley UF Survey
31
ARA Hurricane Charley Survey
Neighborhood Locations
32
ARA Hurricane Charley Survey

• 370 Houses
• Surveyed by Engineers
• Building Envelope Performance

33
Simulated WBD Environments

• Trajectories
• Impact points, energies
• Mean transport
• Max transport

Charley-B
Charley-C
34
Simulated WBD Environments
Andrew-F
Andrew-E
Ivan
35
Validation Results for Window Damage
36
Validation Summary

• Hurricane Bonnie demonstrated that
• Peak gust winds of 80 mph in open exposure
  produced window damage from WBD
• Shingles can be transported up to several hundred
  feet at these windspeeds
• Hurricane Charley indicated
• Significant WBD damage in both shingle and tile
  neighborhoods
• 20-30 of residences experience WBD breach of
  openings
• Hurricane Andrew Model underestimates 1 story
  damage, but compares well to 2 story subdivision
  observations
• Hurricane Ivan Little WBD damage noted in
  survey and reproduced by model
• Model agrees reasonably well, some
  underprediction of high damage states

37
Hurricane Ivan Aerial Photo Analysis(New
Information on Effects of Trees)
Ivan Peak Gust Wind Swath

• NASA Photographs near the coast
• Analyzed photos for
• Extent of tree canopy coverage
• Percent of trees blown down
• Observable damage to buildings (roofs)
• Compared with Citizens loss data for insured
  losses
• Compared to tree blowdown probability estimates

38
NASA Aerial Photo Locations
PenEast
PenWest
PB1
PenNAS
PB2
PB3
39
Example Tree Blowdown Counts

• PenWest Area

PB1 Area
40
Citizens Loss Data
Losses in Treed Terrain
Loss by Grid Location
Significant loss reduction in treed terrain!
41
Hurricane Ivan Summary
Aerial Photo Summary
Tree Blowdown Estimates (from HAZUS)
Blowdown reduces dramatically with increasing
tree density
Hurricane IvanWindspeeds
Good Agreement!

• Losses appear to be reduced in treed terrain
• Tree blowdown model agrees reasonably well with
  observations

42
4. Selected Panhandle Houses and Subdivision
43
Panhandle Houses

• Six new houses were selected for analysis
• CAD models of the building envelopes
• Strengths for components and cladding
• Building valuations and replacement costs
• Cost estimates for window protection options

44
Example Components

• House CAD models of building envelope were
  developed based on
• Plan Drawings
• Site Visit
• Photographs
• Sketches
• Window and door locations were maintained

45
CAD Models Bldgs 1-3
Building 2
Building 1
Building 3
46
CAD Models Bldgs 4-6
Building 4
Building 5
Building 6
47
Replacement Value Estimates

• Replacement value is used in the estimation of
  hurricane damage repair and reconstruction costs
• Replacement value has been estimated by
• Insurance Risk Services of Sanford, Florida
• ISO Home ValueTM
• Marshall Swift/Boeckh

• A benefit/cost sensitivity is included to reflect
  inflationary costs of repairs after hurricane
  catastrophe

48
Panhandle House Summary
49
Protection Option Costs

• House Upgrade Costs ()
• House Upgrade Cost ( of Base)

• Glazing Protection SF Cost Estimates

50
5. Simulate House Performance
51
Simulation Matrix
ASCE Contours

• 6 Houses
• 4 Terrains
• Open-Suburban
• Suburban
• Light Trees
• Medium Trees
• 3 Panhandle Locations
• 110 mph Contour
• 120 mph Contour
• 130 mph Contour

• 4 Glazing Protection Options
• None
• Steel Panels
• Plywood
• Impact Resistant Units
• Total 6 x 4 x 3 x 4 288

52
Two-Step Simulation Process
Step 1. WBD Environment
Step 2. Individual Building Performance
1. Simulate a Panhandle Subdivisionto produce
WBD environment for individual building analysis
2. Simulate Individual Buildings 200,000 yrs
of Hurricanesx 30 Building
Replications6,000,000 Building Responses Each

• WBD Impact
• 30 Hurricane Traces
• 6 Peak Gust Windspeeds
• 90, 110, 130, 150, 170, 190 mph

• Physical Damage
• Building Envelope
• Water Penetration
• Damage State

• WBD Parameters
• Number of impacts per SF wall
• Impact energy
• Impact momentum

• Economic Losses
• Repair and Reconstruction
• Contents
• Loss of Use

53
Panhandle Subdivision
Survey of Panhandle Region from Google
EarthProbability Distribution of Houses/Acre
Mixed Subdivision

• Density
• 3 Houses per acre
• Code
• ½ New Code, ½ Old Code
• Old Code Houses
• 44 6d roof deck nails
• 56 8d roof deck nails
• Roof Shape
• 28 Hip Roofs, 72 Gable Roofs
• Roof Cover
• 17 Tile, 83 Shingle
• 140 houses
• Number of Stories
• 50 1 Story, 50 2 Story

3 per acre is near median
Subdivision140 Houses104 Interior
WBD Environment
54
Hurricane Event Model

• Storms initiated in Atlantic, Caribbean and Gulf
  of Mexico
• Storm track and intensity modeled
• Central pressure modeled as a function of sea
  surface temperature
• Numerical Wind Field Model
• Terrain dependent velocity profiles and gust
  factors
• 200,000 years of hurricanes simulated
• Each Panhandle hurricane is used in the
  individual building performance simulations

55
Number of Missiles Produced in Subdivision
Suburban
Open-Suburban

• Heavily dependent on loads and, hence, terrain
• 300,000 Open-Suburban
• 200,000 Suburban
• 50,000 Light Trees
• 10,000 Medium Trees
• Number of missiles in Open-Suburban at 110 mph
  exceeds Light Trees at 130 and Medium Trees at
  170 mph
• Tree branch missiles considered

Medium Trees
Light Trees
56
Tree Branch Missiles
Tree Blowdown
Trees Standing
Branch Missiles Light Trees
V 130 mph Branch Impact Simulation
Cumulative Branch Missiles/acre
57
Panhandle WBD Impact Probabilities(House with
12.5 Glazed Area within Total Wall Area)
Open-Suburban
Suburban
Very High Risk
High Risk
Light Trees
Medium Trees
Very Low Risk
Low Risk
58
CDFs of Impact Energy of Different Wind Speed
Bins at Panhandle Subdivision in Suburban-Open
Terrain
Shingle Missile
Current WBD Test Standard 349 ft-lb

• Many wood and some tile impacts exceed test
  standard
• Shutters can fail
• Raises questions of shutter adequacy in very high
  windspeed locations

59
Individual Building Simulations
60
Simulation Approach for Individual Buildings
Modeled Building
CAD Model Building Frame Components and
Cladding Wind Resistive Features Building
Value Contents, ALE
Building

• Location
• Terrain

61
Example Building Performance Simulation (from
Model Outputs)
110 mph
75 mph
120 mph
143 mph
138 mph
160 mph
62
Building 1 Results Failure of at Least One
Glazed Opening 120 mph
Open-Suburban
Suburban
Light Trees
Medium Trees
63
Building 3 Results Failure of at Least One
Glazed Opening
Open-Suburban
Suburban
Light Trees
Medium Trees
64
Building 6 Results Failure of at Least One
Glazed Opening
Open-Suburban
Suburban
Light Trees
Medium Trees
65
Building 1 Results Component Reliabilities
66
Building 4 Results Component Reliabilities
67
Loss Reduction Example House 4

• Losses
• Building
• Contents
• ALE
• Limits
• Building 120
• Contents 50
• ALE 20

68
6. Quantify Risks, Benefits and Costs
69
Do Benefits Outweigh Costs?
70
Benefits and Costs

• Benefit-Cost Decisions
• Facilitates efficient allocation of societys
  resources
• Selection of optimal criteria from several
  alternatives
• Generally applied to specific projects,
  decisions, etc.
• Generally recommend alternative with largest net
  societal benefits
• Sensitivity analyses help assess how
  uncertainties affect results
• Benefits
• Reduction in losses due to protection of openings
• Considered as annualized losses
• AAL (No Opening Protection) AAL (Opening
  Protection)
• Depends on house and type of opening protection

• Costs
• Incremental cost of opening protection in Year 0
• Depends on house
• Depends on type of opening protection
• Benefit-Cost Ratio
• R gt 1 means that Benefits gt Cost
• PV Present Value
• NPV PV (Benefits) PV (Costs)

71
Benefit-Cost Time Line

• Time value of money is considered
• Provides a way to measure return on capital
  investment in opening protection
• Future benefits are discounted to present value
  to compare with initial investment

72
Benefit Cost Parameters - Sensitivity

• Minimal Benefit Case
• Heavily discounts future benefitsi 6 (real
  rate)
• Considers only building related losses
• Building
• Contents
• Loss of use
• Includes expected cost of shutter installation
  for panels and plywood options (including false
  alarms)
• 0.41 hurricanes/year gt 30 mph at PH coast
• 1 SF to put shutters up (except IRU)
• Neglects salvage value of opening protection
  investment
• Opening protection investment is not recovered in
  future

• Maximum Benefit Cost
• Real discount rate i 3
• Considers public costs of hurricanes
• The more damage, loss, and displaced homeowners,
  the greater the public costs (tax dollars)
• Factor of 2 applied to AAL
• Neglects shutter installation cost for
  approaching storm
• Includes salvage value of opening protection
• Recovered, but discounted to NPV
• Inflation effects of house repairs following
  hurricanes
• 30-40 increase in AAL estimated
• Can be considered to be included in PC multiplier

• Range of opening protection costs considered in
  both cases

73
Minimal and Maximum Benefit Cases

• Minimal Benefit
• Maximum Benefit

C0 Initial Cost Increment
i 6
c1
c2
c3
c4
c40
(Annual Shutter Installation Costs)
Benefits Average Annual Loss Reduction vs
Base
(Annualized Loss Reduction)
B1
B2
B3
B4
B40
C0 Initial Cost Increment
Bsalvage
i 3
Benefits Average Annual Loss Reduction vs
Base
(Annualized Loss Reduction)
B1
B2
B3
B4
B40
0
1
2
3
4
40 years
Time
74
AAL House 4

• Annualized Benefits AAL (No Protection AAL
  (Protection)

75
House 4 Net Present Value

• NPV PV (Benefits) PV (Costs)

Minimal Benefit
Maximum Benefit
OS Open-Suburban SS Suburban LT Light
Trees MT Medium Trees
76
House 4 Benefit Cost Ratios
Maximum Benefit
Minimal Benefit
OS Open-Suburban SS Suburban LT Light
Trees MT Medium Trees
77
Benefit Cost Results Template
1
2
3
1
2
3
1
2
3
Open- Suburban
4
5
6
4
5
6
4
5
6
House Number
Avg
Avg
Avg
3
1
2
3
1
2
3
1
2
Average of 6 Houses
Suburban
6
4
5
6
4
5
6
4
5
Avg
Avg
Avg
Terrain
1
2
3
1
2
3
1
2
3
Lt Trees
4
5
6
4
5
6
4
5
6
Benefit Cost lt 0.9
Avg
Avg
Avg
1
2
3
1
2
3
1
2
3
0.9 ? BC ? 1.1
Med Trees
4
5
6
4
5
6
4
5
6
BC gt 1.1
Avg
Avg
Avg
100
110
120
130
Windspeed (mph)
78
Visualizing Windspeed and Terrain Criteria
If Windspeed Dependent, results should slice
vertically
If Terrain Dependent, results should slice
horizontally
If Windspeed Terrain Dependent, results should
slice diagonally
Open- Suburban
Suburban
Terrain
Lt Trees
Med Trees
79
Steel Panel Shutters Minimum Benefit Parameters
BC gt 1.0 (40 yrs, I 6) Public Cost Multiplier
1, Salvage Value 0, Storm Installation Cost
Logical 1
80
Plywood Shutters Minimum Benefit Parameters
BC gt 1.0 (40 yrs, I 6) Public Cost Multiplier
1, Salvage Value 0, Storm Installation Cost
Logical 1
81
Impact Resistant Glazing Minimum Benefit
Parameters
BC gt 1.0 (40 yrs, I 6) Public Cost Multiplier
1, Salvage Value 0, Storm Installation Cost
Logical 1
82
Summary Minimum Benefit Parameters
Steel Panel Shutters BC gt 1.0 (40 yrs, I
6) Public Cost Multiplier 1, Salvage Value
0, Storm Installation Cost Logical 1
Plywood Shutters BC gt 1.0 (40 yrs, I 6) Public
Cost Multiplier 1, Salvage Value 0, Storm
Installation Cost Logical 1
Impact Resistant Glazing BC gt 1.0 (40 yrs, I
6) Public Cost Multiplier 1, Salvage Value
0, Storm Installation Cost Logical 1
83
Steel Panel Shutters Maximum Benefit Parameters
BC gt 1.0 (40 yrs, i 3) Public Cost Multiplier
2, Salvage Value 100, Storm Installation
Cost Logical 0
84
Plywood Shutters Maximum Benefit Parameters

• BC gt 1.0 (40 yrs, i 3)
• Public Cost Multiplier 2, Salvage Value 100,
  Storm Installation Cost Logical 0

85
Impact Resistant Glazing Maximum Benefit
Parameters
BC gt 1.0 (40 yrs, i 3) Public Cost Multiplier
2, Salvage Value 100, Storm Installation
Cost Logical 0
86
Summary Maximum Benefit Parameters

• Plywood Shutters
• BC gt 1.0 (40 yrs, i 3)
• Public Cost Multiplier 2, Salvage Value 100,
  Storm Installation Cost Logical 0

Steel Panel Shutters BC gt 1.0 (40 yrs, i
3) Public Cost Multiplier 2, Salvage Value
100, Storm Installation Cost Logical 0
Impact Resistant Glazing BC gt 1.0 (40 yrs, i
3) Public Cost Multiplier 2, Salvage Value
100, Storm Installation Cost Logical 0
87
Comparison of Risks

• Risk of glazed opening failure for houses in
  light trees without opening protection is about
  the same as shuttered houses in Open-Suburban
  Terrain
• Risk of glazed opening failure for houses in
  medium trees without opening protection is less
  than a shuttered house in Suburban Terrain

Similar Risks
Much Less Risk
88
Tree Fall Building Impact Risk
Probability of House Hit
89
WBD vs Tree Fall Risk on House

• Light Trees
• Medium Trees

WBD Risk
Tree Fall on House
Tree Fall Risk is same or higher than WBD Risk in
Light Trees
WBD Risk
Tree Fall on House
Tree Fall Risk is higher than WBD Risk in Medium
Trees
90
7. Conclusions and Recommendations
91
Conclusions

• WBD is a dominant risk to buildings in open and
  suburban terrains.
• Glazing WBD failure in Open Terrain have occurred
  in peak gust winds as low as 80 mph
• In treed terrain, no glazing failures were noted
  in the UF survey for Hurricane Ivan in 100-110
  mph
• In Hurricane Charley, the ARA survey of over 300
  houses indicated
• Similar roof cover loss for shingles and tiles
• 17-18 loss for old code, 8-9 for new code
• Tile neighborhoods experienced 33 window
  breakage for unprotected openings
• Shingle neighborhoods experience 24 window
  breakage for unprotected openings
• In Hurricane Andrew, over 90 of houses in the
  NAHB survey experienced broken windows from WBD
• Failed openings lead to internal pressures in the
  building (increasing chance for further failures
  due to increased loads) and water penetration in
  the building.
• Treed terrain dramatically reduce the loads on
  buildings and the low level windspeeds, thereby
  significantly reducing the WBD risk.

92
Conclusions (contd)

• Within the windspeed contours (110 to 130mph)
  investigated, terrain is more important than
  windspeed in determining the need for WBD
  protection.
• In medium treed terrain, the BC ratios are
  generally ltlt1.
• In light treed terrain, the results were mixed
  and dependent on the range of benefit cost
  parameters.
• The most beneficial solution for society is to
  implement a WBD criteria that considers both
  windspeed and terrain, much as the pressure load
  coefficients are terrain dependent.
• In light and medium tree terrains, tree fall risk
  on house seems to be higher than WBD risk.
  Cost-beneficial strengthening solutions should be
  investigated for tree fall protection.

93
Conclusions (contd)

• Key research qualifications in these results
  include
• Glass breakage and debris transport for shingle
  missiles
• Shingle debris transport validation
• Effects of tree blowdown on velocity profiles,
  loads
• Effects of tree blowdown on losses (overestimates
  effectiveness of shutters)
• Limited treed terrain test parameters more tests
  needed for
• Larger subdivision
• Fewer trees
• Smaller buffers
• Have only considered SF residential, and not
  commercial
• The results show that openings should be
  protected in
• Open-Suburban terrain
• The lowest winds (110 mph) considered produced
  average BCgt1
• Raises question of what 100 mph results would
  indicate for open-suburban
• Suburban (no trees) in the range 110-130 mph from
  public costs perspective
• Light Treed terrain in gt 130 mph, the results
  show generally beneficial BC ratios from public
  cost perspective

94
Conclusions (contd)

• There are many subdivisions that are much larger
  than those considered in the scaled wind tunnel
  tests
• A review of aerial photography of Panhandle
  showed that most subdivisions exceed our wind
  tunnel tested parameters with 400 ft depth of
  subdivision
• This raises major questions in implementing
  terrain-dependent WBD criteria
• More tests are needed to develop final criteria

200 ft
95
Recommendations

• Windspeed and terrain dependent WBD criteria
  should be implemented in Florida and nationally
• The Project Phase II should proceed to finalize
  windspeed and terrain parameters for building
  code implementation.
• Hence, we recommend a two-phased implementation
  approach
• Phase I 2007 Panhandle WBD Region
• Phase II 2008 Statewide Implementation of
  Windspeed/Terrain-Dependent WBD Criteria
• The Statewide Implementation will address all
  windspeeds and terrains to ensure maximum cost
  effective protection where needed
• We note that the recent NIST report on Hurricane
  Katrina recommends
• Evaluate the effects of shielded (e.g., wooded
  or wooded/suburban) exposures and their potential
  for reducing the wind loads on nearby residential
  structures and better explaining the variation in
  observed damage.

96
Phase I Panhandle WBD Region for 2007

• Our recommendation for the Phase I Panhandle
  Criteria (for 2007), based on a reasonable
  balance of benefits and costs
• Phase I. 2007 Panhandle Criteria - Adopt the 130
  mph contour as the WBD region in the Panhandle.
  This option would also include all areas within
  1500 feet of the inland Bays that are not within
  the 130 mph contour.

97
Phase I 2007 Criteria
Phase I. 2007 Panhandle Criteria Adopt the 130
mph contour as the WBD region in the Panhandle.
This option would also include all areas within
1500 feet of the inland Bays that are not within
the 130 mph contour.
Requires Protection

• Pros
• Includes 130 mph region that under the Maximum
  Benefit Parameters that often requires WBD
  protection for open-suburban, suburban, and light
  treed terrain.
• Includes 1500 feet of bays to reflect
  open-suburban terrain for winds less than 130
  mph. This is also consistent with analysis for
  both max- and min- benefit assumptions.
• Will insure protection of houses built in this
  windspeed region that would not qualify for treed
  terrain.
• Cons
• Includes region that likely includes many areas
  of medium trees and hence may unnecessarily add
  WBD protection costs with minimal benefits.
• Requires WBD protection in trees, but neglects
  tree fall risk, which may be higher. However, the
  tree fall risk will be addressed in Phase II to
  ensure balance of risks and cost- beneficial
  solutions

Minimum High Cost
Open-
Suburban
Suburban
Lt Trees
Med Trees
130
100
110
120
Maximum Low Cost
Open-
Suburban
Suburban
Lt Trees
Med Trees
100
110
120
130
98
Phase I 2007 Panhandle WBD Region
Phase I

• WBD Region
• ? 130 mph
• 1500 from bays

99
Phase II Research

• The research would address these limitations of
  the current study
• Additional Wind Tunnel testing to encompass
  larger subdivisions and fewer trees
• Shingle impact/transport testing
• Include additional damage validation cases from
  Florida hurricanes
• Include 100 mph locations to complete
  Benefit-Cost matrices
• Consider tree fall on residences
• Consider terrain transition due to tree-blowdown
• Final conclusions, recommendations, pros and cons
• Develop implementation language for code bodies
• Results would be applicable to all of Florida
• Research would lead to cost-effective and risk
  balanced building code implementation
• Parallel funding from outside the state would be
  sought to address national implementation by
  evaluating non-Florida locations
• The Phase II work is essential to ensure houses
  built in open-suburban and suburban terrains (no
  treed terrain) are not exposed to excessive
  risks.

100
Florida Windspeed Regions
101
At Risk?

• Study indicates 110-120 mph region should have
  opening protection for open-suburban and
  possibly, suburban terrain
• Phase II would address this issue with further
  experiments and analysis
• Phase II would also address 120-130 mph Panhandle
  as part of final terrain and windspeed dependent
  criteria
• Note that there are more new starts at risk in
  110-120 mph outside Panhandle than in gt120 mph in
  Panhandle

More new code starts at risk outside Panhandle
102
Phase II Should also Evaluate lt110 mph for
Open-Suburban and Suburban (No Trees)

• Should buildings in these locations in
  Open-Suburban terrains have opening protection?
• The risk of Open-Suburban at 110 mph gt risk of
  WBD damage in light trees at 120-130 mph