Milton-Madison Bi-State Travel Demand Model Rob Bostrom Planning Application Conference Houston, Texas May 19, 2009.

Milton-Madison Bi-State Travel Demand Model

Rob Bostrom Planning Application Conference

Houston, TexasMay 19, 2009

Presentation Overview• Milton-Madison Project Overview• Bi-state model development• Model Results• Corridor study traffic forecasts

Milton-Madison Project Overview• Location – existing US 421 bridge over the

Ohio River at Milton, KY and Madison, IN• Project Goals

– $5.3 M project to look at bridge replacement and rehabilitation options

– Possible funding using tolls– Examination of I-71 to I-74 corridor necessitated the

creation of a bi-state model using the Kentucky and Indiana statewide models

Bridge Location

Bridge Characteristics• Built in 1929, rehabilitated in 1997 (deck

replacement, structural steel repairs and concrete patching)

• 3,181 feet long, steel truss w/ 10-ft lanes (substandard)

• Current ADT is 10,700• Truck percentage is 4%• Temporary 15-ton weight limit

Needs & Deficiencies Report

• Extensive data collection– OD survey– Historical data review– Blanket counts including turning movement

counts around area

Bi-State Model • Purpose• Source data• Network development• Zone system development• Trip table development• Traffic assignment• Forecasting procedures

Bi-state model development• Purpose:

– Use for bridge replacement alternatives (in vicinity of Madison)

– Use to determine if new bridge would create demand in I-71 to I-74 corridor

• Model data sources– KY statewide travel demand model and IN statewide travel

demand model– Field data including O-D surveys and traffic counts– KY CIMS and IN Commodity flow database for commodity

flows

KYSTM Background

• Base Year = 2007• Future Year = 2030• 4870 TAZs, including

3651 in-state TAZs• 71640 Network links

including 55662 in-state links

• Developed in 2003 by WSA

• Revalidated in 2007 by KYTC

INSTM Background

• Base Year = 2000 • Future Year = 2030 • 4720 TAZs, including

4579 in-state TAZs • 45468 Network links

including 32304 in-state links

• Developed in 2003 by BLA

Combining Networks

Combining Networks

Combining Networks• Converting INSTM network attributes to KY format

TAZ Development• Necessary to update INSTM base year from 2000

to 2007.– The key data needed was population, households and

employment– Data was interpolated between the base year and the 2030

future year– The county and state control totals were compared to Woods

and Poole 2007 data as a reasonableness check– The resulting data for Jefferson County (the Indiana county in

which the bridge is located)

Jefferson County SE Data

2000 2007 2030Population 31,665.00 33,133.00 37,182.00Employment 17,938.00 13,585.00 19,604.00

Consolidated TAZs

TAZs

• Also had to change the zone numbering scheme as follows:– Original numbering

scheme:– New numbering

scheme:

Trip table development• KYSTM generates trip tables for the following

purposes– Short distance Private Occupancy Vehicles (POV) – HBW, HBO, NHB– Long distance Private Occupancy Vehicles – Business, tourist, other

long distance– Trucks – long distance trucks, local trucks

• INSTM generates trip tables for the following purposes

– Short distance Private Occupancy Vehicles (POV) – HBW, HBO, NHB– Long distance Private Occupancy Vehicles– Trucks

Trip table disaggregation• KY long distance trips need to be disaggregated for

POV, LD and Truck trips• Factors were developed to transform the KY TAZs to

the INSTM TAZs. An example of the factors for a single KYSTM zone is shown below:

External Special Generator Trip Table Disaggregation

• KYSTM only validated within KY

• External trips (outside KY) were treated as special generators

• Used INSTM to distribute special generators, especially 600013 (Ext Sta on US 421 bridge)

Final Bi-state model triptable• Short distance Private Occupancy Vehicles (POV)

– KYSTM Home Based Work (HBW)– KYSTM Home Based Other (HBO)– KYSTM Non Home Based (NHB)– INSTM Home Based Work (HBW)– INSTM Home Based Other (HBO)– INSTM Non Home Based (NHB)

• Long distance Private Occupancy Vehicles– KYSTM Business– KYSTM Tourist– KYSTM Other long distance– INSTM Long distance

• Trucks– KYSTM Trucks– INSTM Trucks

Assignment

• AON Trucks Use an All or Nothing assignment method to Assign KYSTM Trucks and INSTM Trucks

together

• Preload Trucks

• UE for POV use a user equilibrium method to

assign all private occupancy vehicles together

Validation

• Visual review of the number of lanes and network connectivity

• Shortest path traffic assignments to ensure that no breaks in the network existed

• Review of the output travel times between traffic analysis zones, to ensure that there were no unreasonable travel times caused by network coding errors

Validation• Validated only for Jefferson County (containing Madison)• Used RMSE and maximum percent deviation (NCHRP

255) comparisons• Bridge assignment was 10,200 compared to 10,300

actual count

Volume Group Observations Average Count % RMSE

< 2000 23 638 147% 2000 - 4000 14 2,723 25% 4000 - 6000 11 4,701 37% 6000 - 8000 5 7,108 29%

>8000 8 10,776 21% All 61 3,709 38%

Actual counts vs. modeled volumes

O-D Comparison

• WSA performed OD data collection in Madison• Performed select link analysis to compare• Model seems reasonable although there are no standards for OD

comparison

Location Survey % of Bridge Traffic using Route

Model % of Bridge Traffic using Route

KY56 West of Bridge and KY 7 12.2% 16.5% KY 7 Downtown 9.4% 15.5% KY 7 North of KY 62 5.2% 3.9% KY56 East of Bridge 4.5% 8.7% US421/KY62 26.3% 25.2%

Future Model

• Both IN DOT & KYTC provided their E+C networks

• Used 2030 as the forecast year• Looked at No-Build conditions• Future growth shown below:

2007 2030 Change Jefferson County, IN 78,600 87,400 11% Trimble County, KY 13,400 18,300 37% Kentucky and Indiana 24,472,200 29,243,800 19%

Model Results

• Milton-Madison bridge– 2030 volume = 12,900 (25% increase)– Select link analysis for trips using bridge shown below

Distribution of Bridge Trips

Forecasted volumes: POV & Truck

Forecasted 2007-2030 Change

Corridor Study Analysis

• Purpose of corridor study – to see if combination of new Ohio River Bridge and improved corridor results in substantial new traffic

• The alternatives analyzed were:– No Build– Remove the bridge– I-71 to I-74 on three different corridors crossing the

Ohio River in the vicinity of Madison & Milton

Corridor Study Map

Corridor Study Results

• Additional trips– Total volume: 5,000 to 6,000– Trucks: 1,000 to 1,500

2030 Bridge VolumesLocation Do-Nothing West Alt. Middle Alt. East Alt.

Total Trucks Total Trucks Total Trucks Total TrucksExisting Bridge 12900 700 11400 500 9600 200 9200 400New Bridge 8000 1100 8400 1300 10000 2100Total 12900 700 19400 1600 18000 1500 19200 2500

Corridor PMs• The usual Performance Measures (PMs) were produced by the

model:– 2030 daily vehicle miles traveled (VMT)

– 2030 daily vehicle hours traveled (VHT)

• The PMs were compared to the No Build alternative

Corridor Performance Measures

VMT Change VHT ChangeDo Nothing 341,384,644 NA 7,744,820 NAEast Alt. 341,351,371 33,273 7,740,745 4,075

Middle Alt. 341,448,805 -64,161 7,740,339 4,481West Alt. 341,418,899 -34,255 7,736,112 8,708No Bridge 341,832,841 -448,197 7,754,376 -9,556

Conclusions

• Need for new tool Milton-Madison bridge project a major investment study needing refined modeling tools

• Project solution was a Bi-state model This tool, created from the KYSTM and INSTM was relatively easy to create and validate

• Successful implementation The Bi-state model results provided valuable information to decision-makers for future bridge volumes and for possible corridor options

Thank You

• Questions?

• Rob Bostrom, [email protected]