FTA Workshop on Travel Forecasting for New Starts 1March 2009

Before-and-After Studies – Examples

Session 12 Interstate MAX LRT in Portland

Joe Recker, TriMet Jennifer John, John Parker Consulting

Newark-Elizabeth Rail Link Bill Woodford, AECOM

Charlotte South Corridor LRT Bill Woodford, AECOM

FTA Workshop on Travel Forecasting for New Starts 2March 2009

FTA Workshop on Travel Forecasting for New Starts 2March 2009

Before-and-After Study for Interstate MAX LRT in Portland

Key Findings and Lessons Learned

Joe ReckerTriMet

Jennifer JohnJohn Parker Consulting

Presented by Ken CervenkaFTA

FTA Workshop on Travel Forecasting for New Starts 3March 2009

Project Description Planning began in 1980’s

Bi-state south/north corridor Project salvaged from failed bond

measure for entire corridor 5.8-mile extension (10 new

stations) north from existing 33-mile east-west system Bus service replaced with light rail Traffic lanes replaced with

exclusive transit lanes

FTA Workshop on Travel Forecasting for New Starts 4March 2009

FTA Workshop on Travel Forecasting for New Starts 5March 2009

Project Purpose

Provide a faster/more reliable connection to largest employment center in region - downtown Serve commuter market

Majority of trips in/out of the downtown Intercept trips crossing the river through park and ride facilities

and bus transfer points at the northernmost terminus of the route

Other goals included: Maximize effectiveness and efficiency of transit system in older,

urban neighborhood Support land use goals Accommodate future growth

FTA Workshop on Travel Forecasting for New Starts 6March 2009

Weekday Project Ridership

2005 forecast (completed in 2000) – 13,900 Opened May 2004 for revenue service

Spring 2005 – 11,830 Spring 2007 – 12,900 Spring 2008 – 13,800 Summer 2008– 14,700

Aggregate ridership forecast pretty good Insights are gained from delving deeper into

the forecasts at a more disaggregate level

FTA Workshop on Travel Forecasting for New Starts 7March 2009

Interstate Max Corridor Average Weekday Ridership

Actual (2005) Compared to 2005 Model Forecast

0

5,000

10,000

15,000

Yel

low

Line

Li

ne 5

Li

ne 4

Li

ne 6

Li

ne 8

Li

ne 4

0

Li

ne 8

5

Rid

ers

hip

Model

Actual2005Actual2008

Weekday Corridor Ridership

FTA Workshop on Travel Forecasting for New Starts 8March 2009

Transit Rider Travel Patterns

FTA Workshop on Travel Forecasting for New Starts 9March 2009

2005 Trip Purpose Comparison Interstate MAX Only

Model vs. Actual53%

30%

4%

6%

40%

27%

11%

9%

12%

8%0%

10%

20%

30%

40%

50%

60%

HBW HBO NHBW NHBNW SCHOOL

Model

Actual

Distribution by Trip Purpose

FTA Workshop on Travel Forecasting for New Starts 10March 2009

Predicted – Roughly 2% annual growth thru 2025 Actual – Roughly 2% contraction of employment

between planning and project opening (national recession)

Opening year interpolated from horizon year forecast (not a 2005 land use allocation process)

Economic cycles impact land use projections more in the short term than long term

Land Use

2005 Projected

2005 Actual

Absolute Difference

Households 10,587 9,217 -1,370

Employment 104,022 94,851 -9,171

FTA Workshop on Travel Forecasting for New Starts 11March 2009

Service Levels

Predicted Replace bus with rail Reallocate bus service to rest of N/NE Portland

Actual Modest service changes Generally met expectations

Issues Line coding checks Horizon vs. opening year service planning

FTA Workshop on Travel Forecasting for New Starts 12March 2009

Understand software Emme2 multipath

assignment

Mode of access Park-and-ride

assumptions Walk connections Transfer rates

Modeling Inputs

FTA Workshop on Travel Forecasting for New Starts 13March 2009

Overall Ridership By StationInterstate MAX 2005

Model vs. ActualAverage Weekday

Station Ons/Offs

9,12

0

3,57

0

1,92

0

900 1,

370 2,

600

1,45

0

2,74

0

830

900

2,45

0

9,64

7

2,91

6

626 92

2 1,25

3

1,81

9

1,06

2

3,19

9

872

646

636

0

2,000

4,000

6,000

8,000

10,000

12,000

Portla

nd C

BD

Rose

Quarte

r TC

Albina

/Miss

issipp

i

Ove

rlook

Presc

ott

Killing

swor

th

Portla

nd B

lvd.

Lom

bard

Kento

n

Delta

Par

k/Van

port

Expo

Cente

r

Stations

On

s/O

ffs

Model

Actual

FTA Workshop on Travel Forecasting for New Starts 14March 2009

Predicted: heavy transfer activity to MAX trunk line

Actual: long walk at transfer point with 2-3 signalized crossings – fewer transfers

30% fewer transfers and 18% less station activity than forecast

Rose Quarter Transit Center

Interstate MAX

Blue/RedLines

FTA Workshop on Travel Forecasting for New Starts 15March 2009

Predicted: walk access to dense employment/ medical center TAZ

Actual: indirect, uphill walk connection to that TAZ on narrow sidewalk underneath freeway

67% less station activity than forecasts

Albina/Mississippi Station

FTA Workshop on Travel Forecasting for New Starts 16March 2009

Predicted: typical urban neighborhood station with mix of walk and transfer access

Household and employment projections not realized

Standard coding conventions utilized in model, did not reflect actual access to the station

75% actual walk access vs. 54% forecasted

Killingsworth Station

½ Mile Buffer

TAZs not connected to station

FTA Workshop on Travel Forecasting for New Starts 17March 2009

Predicted: main access station for Vancouver residents via park-and-ride and Line 5 bus transfer

Actual: State DOT didn’t allow bus/car access from freeway expected in model

Park-and-ride and buses routed to different station

Transfers occurred at Lombard – 17% more station activity there

Park-and-ride underutilized

Expo Center Station

Anticipated vs. Actual Access Plan

FTA Workshop on Travel Forecasting for New Starts 18March 2009

Overlook: Actual land use and access modes close to forecast – 2% higher station activity than forecasted

Kenton and Lombard: Unanticipated transfers led to higher actual station activity – 5% and 17% over forecast, respectively

Delta Park/Vanport: Recent direct access to C-Tran transit market – 2008 actual station activity 11% over forecast

Other Stations

FTA Workshop on Travel Forecasting for New Starts 19March 2009

Line 5 vs. Interstate MAXMAX riders more likely to… Walk further Drive to transit Need fewer transfers to

complete trip Own a car Have higher incomes Make more occasional trips

Number of Blocks Walked

2004 Line 5

2005 Interstate

MAX% Point Change

0 - 1 35% 25% -10%2 - 3 36% 32% -4%4 - 5 18% 19% 0%6 - 7 5% 9% 5%8 - 9 2% 5% 4%10+ 5% 10% 5%Total 100% 100%

Number of Blocks Walked at Origin Interstate Corridor O/D SurveyLine 5 and Interstate MAX Only

Before-versus-After Observations

FTA Workshop on Travel Forecasting for New Starts 20March 2009

Key Takeaways Portland Metro has a good model that has been calibrated with

LRT experience in place Use available tools (GIS, Google Earth, existing

service/operational data, along with understanding of software) for realistic coding conventions in networks

These before and after studies help us understand ridership and aid in forecast assumptions (“unofficial” park-and-riders)

Economic cycles unpredictable – attention to detail in land use forecasts is important

While in the past, opening year was not main focus of modeling/New Starts in terms of reporting process, it is very important