Developing Statewide Evacuation Model

1
Developing Statewide Evacuation Model

• Chi Ping Lam, Houston-Galveston Area Council
• Chris Van Slyke, Houston-Galveston Area Council
• Heng Wang, Houston-Galveston Area Council

2
Outlines

• Background
• Phases for statewide evacuation model
• Re-generate Real World Scenario
• Detect Network and Demand coding issues through
  normal daily run
• Evacuation results
• Sensitivity for Different Evacuaton Scenarios
• Next Steps

3
Background
4
Motivation

• In September 2005, Hurricane Rita landed east of
  Houston
• Over 1 million people attempted to evacuate from
  the eight county region
• Severe congestion as a results

5
Retreat!

• Evacuation routes became parking lots.
• Some people spent more than 18 hours on the
  evacuation routes
• Fatal accidents, abandoned cars, and other safety
  issues

6
(No Transcript)
7
(No Transcript)
8
In response

• H-GAC coordinated with various governmental
  agencies to develop a hurricane evacuation plan.
• H-GAC was asked to develop a tool for evacuation
  planning an evacuation model

9
Phases for statewide evacuation model
10
Phases

• Phase 1 Develop evacuation model on our
  8-county MPO network
• To model how well the transportation system could
  move evacuee outside our region
• 90 completed
• Phase 2 Expand to statewide network
• Model impacts from outside the MPO region
• Provide a more complete evacuation experience
• Early stage

11
Limitation on Phase 1 Model

• Around 90 of Rita evacuation trips travel
  outside of MPO region. The queues extended far
  away our region.
• The external stations are treated as
  destinations in phase 1 trip distribution.
  Those external stations are not real destination.
  
• Some known bottlenecks are outside the MPO
  network, and the traffic queued back to the MPO
  network. The Phase 1 model could not model the
  effect of the bottlenecks well.
• Some evacuation policies, like contraflow lanes,
  extend to and impact area outside the MPO
  networks. Phase 1 model could not model their
  full impact

12
(No Transcript)
13
Outside Region
Inside Region
14
Goals of Phase 2 Model

• Generate a complete evacuation trip tables
• Model impact of bottlenecks outside the MPO
  region
• Model policies outside the MPO region
• Measures congestion outside the MPO region
• Provide a complete picture of evacuation
  experience, such as total travel time

15
Phase Two Processes

• Get a copy of statewide model (in TransCAD)
• Convert the trip tables and the network from
  TransCAD format to Cube Voyager format
• Merge the statewide and regional networks and
  trip tables
• Manually coding the missing bottleneck
• Develop statewide evacuation trip tables
• Test Run
• Model Performance Improvement
• Calibration and Validation

16
Progress on DevelopingStatewide EvacuationTrip
Table
17
Houston TranStar Rita Evacuation Survey

• Solicited participation on website
• Participants responded to questions online
• 6,570 respondents
• 6,286 usable household responses
• 3,886 households evacuated by car or truck

18
Phase 1 Evacuation Trip Tables

• Six-day event modeled
• Cross-classification variables
• 6 geographical districts
• 5 household size groups
• Production models
• Probability of evacuating
• Vehicle trips/evacuation household
• Trip purpose split
• Simple attraction models
• Non-resident trip models

19
Six Districts

• Districts 1-3 are the three mandatory evacuation
  districts
• Districts 4-6 are other part of the MPO regions,
  defined by its distance to the coastline

20
(No Transcript)
21
(No Transcript)
22
Rita Evacuation Generation Results
23
External Station Evacuation Attractions

• Distributed attractions to other urban areas
  based on their population and relative
  accessibility
• Allocated results to external stations

24
Distribution Model For External Station
Attractions

• Similar to traditional external-local models
  using a gravity model
• Primary difference is that the external stations
  are treated the attractions
• Somewhat relaxed version of the normal
  external-local friction factors used

25
From Region-wide to state-wide trip tables

• Obviously, the external stations of the MPO
  network are not true destinations but rather a
  gate to outside MPO destinations.
• Trip generation and distributions should use real
  destinations
• First task is to geo-coding the survey to
  statewide level

26
Geo-coding the OD

• 4,092 records are inputted for geo-coding
• ArcGIS automatically geo-code the destinations if
  the destinations are cities inside Texas.
• Most coding errors are mis-spelling which could
  be corrected
• Only 0.4 of records are without sufficient
  information to identify the destinations

27
Where are people going to?
Top 10 Destinations (by Transtar
Survey) Austin 9.7 Dallas
7.9 San Antonio 7.1 Houston
5.3 Louisiana 2.8 College Station
2.4 Conroe 1.6 Fort Worth
1.6 Waco 1.6 Livingston
1.5 Hundreds of other destinations!!
28
Summary of Survey Findings

• Hundreds of destinations!
• 91.5 evacuation trips are in-state
• 16.9 evacuee change their destinations
• Most evacuee visit their friends or families
• Austin, Dallas, and San Antonio, the three
  largest adjacent metropolitan area, are the top
  three destinations
• Other major destinations concentrated on US-59,
  I-45, and I-34 corridors.
• 82 of planned in-state evacuation trips are
  within 4.5 hours (free flow time) from Houston

29
Factors of Determining Attractions

• Population is the most dominant factor
• Over 80 of evacuees visit family or friends
• Hotel and Shelters
• Coastal area are not very popular
• Evacuated or closed
• There is strong evidence that most evacuees
  disfavor long trips

30
Survey Trips normalized by county population
Big cities are not necessary most attractive
other than number of people Hill countries are
somehow very attractive A few outliners
31
Develop Trip Production Model for Rita Scenario

• Our model will re-generate the Rita scenarios
• For trip produced inside our MPO region, use the
  same production model from Phase 1 model
• Use the same trip table of Phase 1 model for
  internal-internal trips.
• Very few information regarding trip produced
  outside our MPO region.

32
Develop Trip Attractions for Rita Scenario

• The considered factors
• Population
• Accessibility or Distance
• Distance from the coast or coastal area indicator
• Potential bias factor
• Initial analysis suggests linear regression model
  is not a good fit as population rules over other
  factors
• Trip normalized by population may be a better
  variable than population

33
Progress on PreparingNetwork
34
Statewide Network

• Import Texas Statewide Model (SAM) Network
• Include air, marine, freight rails
• Less roadway details inside our MPO region than
  our MPO network
• Adequate for modeling major traffic flow
• Does not support every details in bottleneck,
  like ramps in direct interchange or traffic light
  in a small town
• Have to add some details to the statewide network

35
State Network
36
Merging Statewide and MPO Networks

• The statewide network does not provide enough
  access points or local road capacity to load
  evacuation traffic to evacuation routes inside
  Houston metropolitan area
• For Mesoscopic assignment, this could mean
  evacuation traffic get stuck in local streets,
  execrates local congestion while underestimates
  congestion on evacuation routes. Trial run of
  using a simplified network in Phase 1 model
  supports this logic.
• Therefore, MPO network is required to load the
  evacuation traffic
• Only auto links are merged. All non-auto links
  are deleted.

37
(No Transcript)
38
Handle Inconsistencies between the two networks

• The external stations of the regional model
  connects to the statewide network very well (only
  1 or 2 minor stations without a match)
• The Statewide model and region model have
  different number and definitions of facility
  types and area types
• Therefore, for the same road inside MPO region,
  its capacities for the statewide model and
  regional model are different
• Use regional model setting inside MPO region

39
Bottleneck at US 290 _at_ US-36

• Major bottleneck outside MPO region
• Cause by 1 lane direct interchange not coded in
  the statewide network

40
Other network Modifications

• Adding Contraflow lanes in the network
• May put toll identification to the network

41
Addressing Model Running Time
42
Slow Test Run

• Perform a test run using regular trip tables only
• The mesoscopic assignment complete, and only take
  96 hours to complete!
• Need to reduce running time for more efficient
  calibration and validation process

43
Option1Aggregate Zone System

• SAM models has 4800 TAZ
• Aggregate zone. The aggregated zones cross
  county boundary (zone is always within one
  county)
• The number of zone could be reduced to between
  1000-1500 zones.
• Reduce path-building time
• Avoid assign short trips as those trips are now
  intra-zonal trips
• This should reduced the running time to less than
  2 days

44
Option 2 Sub-area

• Based on the survey, most of the evacuee did not
  travel west of the Hill Country area
• West Texas and the Pan-handle areas could be
  removed from the model
• Affected cities includes El Paso, Amarillo,
  Midland, Lubbock

45
Option 3 Reduce Regular-Day Traffic

• One goal of model is to measure how evacuation
  traffic impacts regular traffic flow in
  destinations
• The static statewide model shows serious
  congestion inside metropolitan area, maybe
  because lacks of details
• Perhaps reduce regular day traffic inside urban
  area to compensate lacks of roads

46
Summary

• The statewide evacuation model is the next phase
  of regional evacuation model.
• Provide more complete pictures of evacuation
  experience and the impact of evacuation
  strategies.
• Need to develop evacuation trip attraction model
• Need to add crucial details to the statewide
  network
• Need to reduce running time