Unit 5: Conducting Realistic BenefitCost Analyses

1
Unit 5 Conducting Realistic Benefit-Cost
Analyses
2
Limitations of the FD Module

• Three major limitations/shortcomings
• Seismic hazard data are old, outdated and
  inaccurate.
• Seismic damage functions are old, outdated and
  inaccurate.
• Casualty estimates are systematically low by a
  substantial amount.

3
Seismic Hazard Data

• Seismic hazard data in FD module are based on a
  very limited number of data points from early
  1990s national seismic hazard maps.
• Much better, state-of-the-art consensus
  nationwide data from the USGS now exist.
• USGS website has 3 data points per location, by
  lat/long and by zip code.
• Supporting documentation has full seismic hazard
  curve data - much better than only 3 data points

4
Seismic Hazard Curves

• Seismic hazard curves give the annual
  probabilities of exceeding various levels of
  ground shaking (expressed in PGA)
• For Washington State, we have compiled an Excel
  File that has the USGS seismic hazard data for
  the entire state.
• Seismic hazard for a project site requires only
• Zip Code and Soil Type

5
Seismic Hazard Curve
6
Soil/Rock Types (IBC 2000)

• Class A hard rock
• Class B rock
• Class C very dense soil and soft rock
• Class D stiff soil
• Class E soft soil
• Class F very soft soil
• NOTE See technical definitions handout from
  IBC 2000

7
Seismic Hazard Data by Zip Code for Washington
State

• 1. Enter Zip Code and soil type for project
  site.
• 2. If soil type is unknown, enter Class D,
  firm soil, as a default estimate.
• 3. The Excel template returns annual
  probabilities of ground motions in the 7 bins of
  PGA used in the FD module

8
Seismic Hazard Data by Zip Code for Washington
State

• Demonstrate
• Excel spreadsheet with Seismic Hazard Data for
  Washington State
• File Hazard_WA Beta Version.xls

9
WA Seismic Hazard Data

• 1. USGS data are for rock sites.
  Soil adjustments were made based on IBC-2000
  soil/rock factors
• 2. Adjusted for magnitude/duration effects
  (longitude dependent)
• 3. Interpolated for PGA bins in FEMA FD module.
• 4. Automatic lookup tables for 700 WA
  Zip Codes

10
Building Seismic Damage Functions

• 1. Data in FD module are from mid-1980s (ATC-13
  seismic damage estimates for California).
• These estimates are too high at low PGA levels
  and often too low at high PGA levels
• 2. Best current data are based on Fragility
  Curves (e.g., HAZUS typical values).

11
Fragility Curves

• 1. Fragility curves are lognormal probability
  distributions that express the probability that a
  building will be in defined damage states at a
  given PGA level.
• 2. Best current building seismic damage
  function data are based on Fragility Curves
  (e.g., HAZUS typical values).
• 3. Curves expressed as median PGA values for
  each damage state, plus lognormal dispersion
  (uncertainty) parameter beta

12
Fragility Curve Example
13
Fragility Curve Math
14
Fragility Curves

• Demonstrate
• Fragility Curve Calculator Version 1.0
• Enter median PGA values (and betas) for slight,
  moderate, extensive, and complete damage states
• Output building damages for each PGA bin
• Includes HAZUS fragility curves for 36 types of
  buildings for 4 levels of seismic design

15
Fragility Curves

• 1. The best structural mitigation projects are
  not for typical buildings, but rather for
  buildings that are much worse seismically than
  typical buildings.
• 2. WARNING the extent of improved performance
  (after mitigation) varies enormously, depending
  on type of building, type of retrofit, robustness
  (completeness) of the retrofit.
• 3. Close consultation with a structural engineer
  specializing in seismic vulnerability is strongly
  recommended

16
Suggestions

• 1. NEVER use the SDFs in the FD module.
• 2. Use HAZUS fragility curves as default or
  typical values.
• 3. Adjust HAZUS curves for building specific
  characteristics whenever possible.
• 4. ALWAYS work in close consultation a
  structural engineer experienced in making
  fragility curve evaluations of buildings seismic
  performance before and after mitigation.

17
Suggestions (Continued)

• 5. NEVER use building SDFs including those based
  on fragility curves for non-structural components
  or infrastructure components - they simply do not
  apply.
• Doing so will produce completely meaningless
  results.
• 6. Mitigation projects that differ in scope will
  have very different improvements in seismic
  performance. Such differences may require
  building specific fragility curves for meaningful
  results.

18
Casualty Estimates

• 1. Life safety is often the driving force for
  seismic mitigation projects.
• 2. Most casualties in earthquakes arise when
  buildings fail - partial or full collapse.
• 3. Benefits of reducing casualties can be a
  large component of total benefits - indeed, often
  the largest benefit.
• 4. Casualty rate estimates (default) in FD
  module are lousy.

19
FD Module Casualties

• 1. Default values based on ATC-13 building
  damage functions - outdated.
• 2. Casualty estimates based on average percent
  building damage at each PGA .
• 3. BUT, most casualties arise from partial or
  full collapse, not from average percent damage.
• 4. Need to estimate probability of collapse at
  each PGA level - cant do from ATC-13 seismic
  damage functions.

20
Fragility Curve Based Casualty Rate Estimates

• 1. Fragility curve calculator (Excel)
  automatically calculates casualty rate estimates
  for deaths, major injuries, and minor injuries
• 2. Casualty rates, especially deaths, depend
  largely on the probability of the building being
  in the complete damage state.
• 3. Fragility curve based casualty rates are
  calculated both before and after mitigation

21
After Mitigation Casualty Rate Estimates

• 1. FD Module default assumption is that the
  casualty rates for minor injuries, major injures,
  and deaths are reduced after mitigation
  (life-safety structural retrofit) by factors of
  10, 100, and 1000, respectively.
• 2. After mitigation, fragility curve calculator
  shows these FEMA default rates, based on the
  above assumption, as well as the fragility curve
  based estimates.
• 3. Which to use is a question of FEMA policy.

22
Suggestions

• 1. NEVER use the casualty rates in the FD
  module.
• 2. Use fragility curve based estimates instead
• 3. Adjust casualty estimates for building
  specific characteristics whenever possible.
• 4. ALWAYS work in close consultation a
  structural engineer experienced in making
  fragility curve evaluations of buildings seismic
  performance before and after mitigation.

23
Suggestions (Continued)

• 5. NEVER use building casualty rates, including
  those based on fragility curves for
  non-structural components or infrastructure
  components - they simply do not apply.
• Doing so will produce completely meaningless
  results.
• 6. After mitigation, be sure that the reduction
  in casualties estimated is commensurate with the
  actual project design and characteristics

24
Unit Summary

• 1. Three major shortcomings of the seismic FD
  module can be easily overcome by
• a) using the Washington State seismic hazard
  data,
• b) using fragility curve based seismic damage
  functions for buildings, and
• c) using fragility curve based casualty rate
  estimates for buildings.

25
Unit Summary (Continued)

• 2. Site specific hazard curves and building
  specific fragility curves are REQUIRED whenever
  buildings or projects are not typical and are
  ALWAYS recommended.
• 3. Evaluating the seismic performance of
  buildings, before and after mitigation, requires
  a great deal of highly specialized engineering
  judgement.
• 4. Working in close consultation with a very
  experienced structural engineer is always highly
  recommended.