Keir Opie
National Center for Transportation and Industrial Productivity (NCTIP)
New Jersey Institute of Technology (NJIT)
Paramics North American User Group Meeting
New Brunswick, NJ
July 14, 2008
Cape May County Hurricane Evacuation And Elevation Study
Project Sponsor & Client
Sponsored by:
• NJDOT Bureau of Research
• South Jersey Transportation Planning Organization
Clients:
• New Jersey State Police, Office of Emergency Management
• NJDOT, Office of Emergency Management
• South Jersey Transportation Planning Organization
• Cape May County
Project Objectives
Elevation Surveys:
• Conduct accurate GPS elevation surveys of county roadways that will feed into the NJ 47 / 347 Corridor
• Determine category of hurricane that will cause roadway inundation by storm surge
Evacuation Simulations:
• Determine total evacuation times for the corridor
• Different lane reversal treatments
• Different evacuation demand and evacuee response scenarios
Background – Cape May County
Southern peninsula of New Jersey
Off-season population of 100,000
Peak-season daytime population of 1 million or more
4 Roads in and out (5 total lanes per direction)
Background – Cape May County
Very low-lying terrain
A hurricane strike on the area could cause storm surges that would inundate much of the county
A major hurricane would force a mandatory evacuation of the entire county
USACE SLOSH Flood Modeling
Background – Cape May County
Density of Different Types of Housing
No FloodingCategory 1Category 2Category 3Category 4
Evacuation – Study Area
NJ Route 47 / 347 highlighted in yellow
Simulation study area shown in green
Pink region is the secondary study area (external zones)
Evacuation – Scenario Constants
Constant in all scenarios:
• Number of different types of housing units from household inventory in USACE HES, plus 2005 campground survey by NJIT
• Number of housing units vulnerable to inundation from different level of hurricanes from the USACE HES
• No day-trippers included in evacuation demands
Evacuation – Scenarios: Traffic Operations (3)
Normal operations (No reversal)
• Police control overrides signal controls
Reversal as currently planned (NJ 83 – NJ 55)
• Follows existing NJ 47 / 347 Reverse Lane Plan
Extended Reversal (US 9 to NJ 55)
• New reversal section assumed to operate similar to current reversal plan
Note: New Parameter in Extension Study
Evacuation – Scenarios: Area Population (2)
Peak Season
• 100% Permanent Residents
• 100% Occupancy of Seasonal / Tourist
Off-peak Season
• 100% Permanent Residents
• 50% Occupancy of Seasonal / Tourist
In both cases:
Household Inventory from USACE HES used to determine type and number of housing units in various inundation zones
No day-trippers included in evacuation demands
Evacuation – Scenarios: Vehicle Demands (2)
Census-based Vehicle Demands
• Data from Census 2000 used to estimate evacuating vehicles per housing unit
• Permanent and Seasonal Housing Units: 1.54 veh/hh
• Hotel / Motel / Campground: 1.0 veh/hh
Higher Estimate of Vehicle Demands
• Higher number of vehicles per household is assumed
• Permanent and Seasonal Housing Units: 3.0 veh/hh
• Hotel / Motel / Campground: 1.0 veh/hh
Note: New Parameter in Extension Study
Evacuation – Scenarios: Routing Assumptions (2)
Balanced NJ 47/347 and Other Corridors Routing
• Assumed balance of traffic between NJ 47/347 corridor and other corridors (US 9 / Parkway, NJ 50 / NJ 49)
• Manual assignment of traffic from each evacuation district to available evacuation corridors
Heavy NJ 47/347 Corridor Routings
• All households south of NJ 83 forced into NJ 47/347 corridor
• All households north of NJ 83 use other corridors
Note: New Parameter in Extension Study
Evacuation – Scenarios: Hurricane Intensity (2)
Category 1 Hurricane
• Evacuation of category 1 inundation areas
• Additional volunteer evacuees from other inundation areas
Category 2 (and up)
• Full scale evacuation of county
• Evacuation of category 4 inundation area
Evacuation – Scenarios: Behavior Response (3)
Hourly vehicle rate of departing evacuees calculated from behavioral response curves (Sigmoid or S-Curves)
• Fast Response Rate (approx. loading time 12 hours)
• Medium Response Rate (approx. loading time 18 hours)
• Slow Response Rate (approx. loading time 24 hours)
Evacuation – Analysis Scenarios
Parameters
Traffic Operations 3
Area Population x 2
Vehicle Demands x 2
Routing Assumptions x 2
Hurricane Intensity x 2
Behavior Response x 3
Total Scenarios 144
Evacuation – Scenarios for Paramics Files
Scenario Parameters
• Traffic Operations Network Structure
• Area Population
• Vehicle Demands
• Routing Assumptions
• Hurricane Intensity
• Behavior Response Profile
Demands File
Evacuation – Total Vehicle Demands
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
180,000
200,000
220,000
240,000
260,000
280,000
300,000
Cat. 1 Cat. 1 Cat. 2+ Cat. 2+ Cat. 1 Cat. 1 Cat. 2+ Cat. 2+ Cat. 1 Cat. 1 Cat. 2+ Cat. 2+ Cat. 1 Cat. 1 Cat. 2+ Cat. 2+
Off-Peak
Peak Off-Peak
Peak Off-Peak
Peak Off-Peak
Peak Off-Peak
Peak Off-Peak
Peak Off-Peak
Peak Off-Peak
Peak
BalancedBalancedBalancedBalanced Max47/347
Max47/347
Max47/347
Max47/347
BalancedBalancedBalancedBalanced Max47/347
Max47/347
Max47/347
Max47/347
Census Census Census Census Census Census Census Census High High High High High High High High
Other Routes
Rt 47/347
Census Based Vehicles, Balanced Routing
Census Based Vehicles, Heavy NJ 47 Routing
Higher Estimate of Vehicles, Balanced Routing
Higher Estimate of Vehicles, Heavy NJ 47 Routing
Paramics Network
450 miles of roadway
• Just under half of county’s roadways
• Centered along a 30-mile stretch of the NJ 47/347 corridor
Approximation of Police Officer override of signals
Multi-day Simulation in Paramics
Some evacuations would take more than 24 hours
One simulation run for each 24-hr period
Data to be passed to the next day
• Queued traffic on the roadways
• Snapshots taken at end of first sim period (23:59:59)
• Queued traffic in zones
• Releases Counts file counts blockages and releases (calculate queue)
Multi-day Simulation in Paramics
Next Day Demands
• Scheduled releases
• Matrix 1, hourly profile releases
• Snapshot loads vehicles en-route
• Loaded at time 0:00
• In-zone queue traffic
• Matrix 2, profile releases all in first 10 minutes of sim period
Run for 24 hours sim period
• Process next 24-hour period (if needed)
Multi-day Simulation in Paramics
Multiple Iterations
• Day 1: Processor runs 5 iterations of each assignment
• Day 2: Modeller run (one iteration for each of day 1 iterations)
• Day 3: Modeller run (if needed)
• Day 3: Modeller run (if needed)
Many Runs
• 144 Scenarios needed nearly 1500 individual 24-hour long simulations
Scenarios: Evacuation Times
Total Evacuation Times (in hours):
Ranges from 16 to 89 hours
Scenarios: Lane Reversal Savings
Shortening of Evacuation Times vs. No Reversal (in hours):
Effects of Lane Reversals - Example
Hours after Call for Evacuation
Cumulative Percentage of Vehicle evacuated via NJ 47/347 Corridor
Study Conclusions
Total evacuation time ranges from 16 to 89 hours (from time of call to evacuate)
Current lane reversal has very limited benefits
• Most traffic assumed to enter NJ 47/347 corridor south of NJ 83
• Current lane reversal plan will not help this bottleneck
Extending the current lane reversal plan further south will help remove this bottleneck
In order to best utilize the existing plan, some traffic must be diverted from points south of NJ 83 to use US 9/GSP to NJ 83 to access the contraflow section
Modeling Conclusions
Paramics can be used to simulate multi-day events
• Some data loss is seen moving between days
• Delays of vehicles unreleased from zones
Beneficial aspects of simulation in evacuation assessment
• Detailed vehicle flow modeling in contraflow lanes
• Detailed queue analysis for extensive queues
Addition of a multi-day profile would make multi-day simulation much easier
Thanks & Contact Information
Keir OpieManager, Simulation & ModelingNCTIP / IITCNew Jersey Institute of TechnologyUniversity HeightsNewark, NJ 07102
[email protected]. 973.596.6454www.transportation.njit.edu