Dr' Jawhar Bouabid

1
Customizing HAZUS at Grid Cell Level

• Dr. Jawhar Bouabid
• Durham Technologies, Inc
• Advanced Training
• June 24, 2002

2
HAZUS Main CD
Demographic Data
County Boundaries
Census Tract Boundaries
Dam Inventory
Schools, hospitals ER
HazMat Sites
Square Footage Data
Hwy, Rlwy, Port Airports
Water, Oil, Gas, Power, Comm
3
HAZUS Installed on Hard Drive
Classifications and Other Non-editable Items
Data Processed by the Building Import Tool
User-created Study Regions
Heading Files
HAZUS Engine
Template Files
4
Objectives

• Create a HAZUS compatible study region at a much
  finer resolution than a census tract
• Improve the accuracy of results
• Enable a much better microzonation capabilities
  for the hazard and for the building inventory
• Generate more uniform and elegant maps
• Provide the flexibility for more accurate land
  use planning in the undeveloped areas

5
Pros and Cons

• Advantages
• High resolution
• Consistent with soil liquefaction data
• No Large Census Tract / Census Block issues
• Independent of tract boundaries
• Easy to update with 2000 Census Data
• Most accurate results
• Disadvantages
• Computation intensive
• Minor round-off errors of the input data during
  the transformation from census block to

6
Example Comparison of Boundaries

• Grid- based data

• Census Tract-based

7
Data Layers Affected

• Only aggregated data is affected
• Better done at source level (i.e., HAZUS CD)
• Less files to worry about
• Global change transferred to all study regions
  created with the enhanced county
• Boundary file (mttr file under CTRACK\)
• Demographic Data (pophsng file under CENSPOP\)
• Square footage data (mtsqft file under MSH\)
• Probabilistic ground motions file (tr file under
  MSH\)

8
Major Considerations in the Approach

• Obtaining the more accurate data at the much
  finer resolution
• Creating a unique identification scheme for the
  grid cells consistent with the census tracts /
  census blocks
• Maintaining the same name convention and data
  structure requirements
• Striking a balance between the size of the grid
  cells and the needs of the project , the data
  gaps, and the computation limitations of HAZUS
• Developing rational models for complementing the
  data gaps

9
Overall Data Transformation Methodology

• Three possible cases to worry about
• Case 1 Grid cell contains entirely the census
  block
• Case 2 Grid cell is entirely within the census
  block
• Case 3 Grid cell intersects with the census
  block
• Original data is then split into the three
  subgroups according to the three cases above
• In case 1, census block-based data was aggregated
  to the grid cell level
• In cases 2 and 3, block-based data was
  distributed among the grid cells proportional to
  the total kilometers of streets in each grid cell
  or in the intersected (partial) portion of the
  cell
• The processed data from the three subgroups is
  then grouped back

10
Example Cases 2 and 3
Case 2
Case 3
11
Assumptions Made where Gaps Existed

• Daytime population is assumed to be about 35 of
  the total population (based on correlation
  analysis)
• Nighttime population is assumed to be about 95
  of the total population (based on correlation
  analysis)
• Commuting population based on average bridge
  daily traffic and on an assumed distribution of
  the traffic during a 24-hour period
• Commercial population based on number of
  commercial employees in LC county distributed in
  proportion to the square footage in each cell
• Industrial population based on number of
  industrial employees in LC county distributed in
  proportion to the square footage in each cell
• Educational square footage was based schools in
  each cell and an assumed square footage per school