Travel Forecasting for New Starts

1
Travel Forecasting for New Starts

• The FTA Perspective
• September 27, 2004

2
Topics

• Why FTA cares about forecasts
• What FTA is doing about forecasts
• What FTA requires about forecasting
• What project sponsors should be doing

3
Why FTA cares about forecasts

• FTA responsibilities
• Accuracy record of forecasts
• External scrutiny

4
FTA Responsibilities

• Evaluation of New Starts proposals
• FTA Major Capital Investment Program
• Discretionary funding
• Annual recommendations to Congress based on
  mandated criteria

Strong ties to travel forecasting

5
FTA Responsibilities

• Count all of the benefits (and costs)
• Maintain a level playing field
• Ensure that promises can be kept
• Make solid cases for good projects

6
Accuracy of Forecasts

• FTA analysis of 19 latest New Starts
• Full Funding Grant Agreement
• Subsequent to 1990 Pickrell report
• Open to service
• Documented guideway ridership forecasts

7
Accuracy of Forecasts

• 2003 assessment
• Exceeded AA forecast 3 of 19
• 80-100 of AA forecast 3 of 19
• 70-80 of AA forecast 4 of 19
• 1990 assessment
• Exceeded AA forecast 0 of 10
• 80-100 of AA forecast 0 of 10
• 70-80 of AA forecast 1 of 10

8
Accuracy of Forecasts

• Conclusions
• Forecast accuracy is much better
• Risk of large errors still remains
• Enhanced quality control is crucial

9
External Scrutiny

• Annual
• Office of Management and Budget
• Congress
• General Accounting Office
• Special studies
• Office of the Inspector General
• General Accounting Office

10
What FTA is Doing about Forecasts

• User benefits
• Detailed reporting of forecasts
• Summit
• Research

11
User Benefits

• Transportation system user benefits
• User benefits are the changes in mobility for
  individual travelers that are caused by a project
  or policy change, measured as hours of travel
  time savings, and summed over all travelers.

12
User Benefits

• Changes in mobility
• Shorter transit times in-vehicle, walk, wait
• Fewer transfers
• Changes in unmeasured characteristics
• Relief of crush loading conditions
• (Shorter auto times due to lower congestion)
• Project-oriented growth new option in 2003

13
Detailed Reporting

• Reporting of trips and user benefits
• Totals across all socio-economic segments
• District-to-district summaries ? reports
• Row totals, column totals ? thematic maps
• Frequency distributions of per-trip benefits
• Results for individual socio-economic segments

14
Reporting Transportation Benefitsfor Individual
Travel Markets

Report 1-5 Total
User Benefits (hours) for the Build Alternative
All
Transit-Access Markets
Home-Based-Work
Production
Attraction District District 1 2
3 4 5 6 7 8 9 10 11 12
13 14 15 16 17 Total
-------------------------------------------------
--------------------------------------------------
------ 1 CBD 4 -1 4 0 5
0 0 6 15 1 0 9 0 0 0
0 0 41 2 Urban 194 86 67
0 39 0 0 73 281 8 0 220 0
15 0 0 0 984 3 N Suburb 135
50 37 0 21 0 0 10 39 2 0
54 0 3 0 0 0 351 4 N Rural
1 1 3 0 1 0 0 0 2
0 0 2 0 0 0 0 0 10 5
W Suburb 219 140 41 0 93 0 0
23 240 4 0 83 0 3 0 0 0
846 6 NW Suburb -13 2 7 0 0
0 0 1 10 0 0 8 0 0 0
0 0 15 7 NW Rural 42 18 13
0 5 0 0 2 4 0 0 8 0
0 0 0 0 93 8 S Suburb 150
86 14 0 17 0 0 130 63 7 0
72 0 1 0 0 0 540 9 SW Suburb
201 147 17 0 108 0 0 31 195
5 0 62 0 1 0 0 0 766 10
SE Suburb 18 12 3 0 4 0 0
3 7 0 0 14 0 0 0 0 0
62 11 SE Rural 2 4 2 0 1
0 0 1 3 1 0 8 0 0 0
0 0 22 12 E Suburb 832 467 88
0 111 0 0 97 191 25 0 909 0
20 0 0 0 2739 13 E Rural 0
3 3 0 1 0 0 1 3 0 0
9 0 0 0 0 0 20 14 NE Suburb
104 49 13 0 10 0 0 5 11
2 0 78 0 3 0 0 0 276 15
NE Rural -41 -16 -2 0 -3 0 0
-1 -1 0 0 -7 0 0 0 0 0
-72 16 External 835 345 123 0 79
0 0 37 95 7 0 138 0 8 0
0 0 1668 17 Other 2 1 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 4 -----------------------
--------------------------------------------------
-------------------------------- Total
2684 432 493 0 1158
0 0 0 0
1396 0 0 420
61 1665 55 0 8364
15
Reporting Transportation Benefitsfor Travel
Produced in Each Zone
16
Detailed Reporting

• Unhappy outcomes
• Previously unknown properties of models
• Problems with highway time savings
• Inconsistencies among models nationally
• Problems in definitions of the alternatives
• New opportunities
• Understanding and refinement of projects
• Making a better case for projects

17
Detailed Reporting

• Brief make the case write-up
• The case for your project as you and FTA can make
  it given the justification criteria in TEA-21
• Three pages supported by your forecasts
• Problem(s) that you are trying to address
• Causes of the problem(s)
• Specific ways the project addresses the
  problem(s)
• Reasons that the project is preferable to
  lower-cost options

18
Calculations in Summit

• User benefit calculations
• Embedded function
• Several (in-stream) runs per build alternative
• For each mode choice run (purpose time of day?)
• For summations across purposes, times of day
• User specifications
• Filenames
• Table titles

19
Calculations in Summit

• User benefit calculations
• Required inputs
• Special output file from base alternative
• FTA-standard output file from build alternative
• Zone?district equivalence file
• Outputs
• Report file district-to-district user benefits
  totals
• Output file district-to-district user benefits
  (binary)

20
Special Mode Choice Output Files for Summit User
Benefits
Prices/Quantities files from mode choice
application
HBW
HBO
NHB
TRIP GEN TRIP DIST
TRIP GEN TRIP DIST
TRIP GEN TRIP DIST
MODE CHOICE
HBW Ps/Qs
MODE CHOICE
HBO Ps/Qs
MODE CHOICE
NHB Ps/Qs
TIME OF DAY ASSIGNMENT
TIME OF DAY ASSIGNMENT
TIME OF DAY ASSIGNMENT
21
Summit Applications to Compute User Benefits
Summit applications Alternative versus Base
Alt HBW Ps/Qs
Alt HBO Ps/Qs
Alt NHB Ps/Qs
Summit
Summit
Summit
Base HBW Ps/Qs
Base HBO Ps/Qs
Base NHB Ps/Qs
User Bens D-D TEsum
User Bens D-D TEsum
User Bens D-D TEsum
Summit
District-District
Row/Col-Sums
22
Calculations in Summit

• Other features
• Analytical reporting of forecasts
• Row-sums and column sums ? GIS
• Selected rows and columns ? GIS
• Trip-length frequency distributions ? grapher
• Trip tables stratified by and user benefits
• Analytical summaries of trip tables
• Software interfaces TP TransCAD
• Emme/2 MinUTP
• TRANPLAN

23
Research

• Technical methods
• Reliable quantification of congestion relief
• Reasonable alternative-specific constants
• Synthesis of data on guideway ridership
• Approaches to quality control
• Others
• Guidance and requirements

24
What FTA Requiresabout Forecasting for New Starts

• Models that tell a coherent story
• Forecasts that can be explained
• A case for the project built upon insights
  obtained from the forecasts

25
Coherent Models

• Consistency with current good practice
• Level playing field
• Likelihood of promises kept
• Threats to coherency of models
• Naïve or less-than-rigorous calibration and
  validation
• Incorrect travel markets represented in
  person-trip tables
• Odd properties in mode choice models
• Inconsistencies between transit path-builder and
  mode choice
• Inaccurate network speeds for auto and bus travel

26
Model Calibration, Validation

• Does it tell a coherent story about behavior?
• Nesting structure and coefficients
• Constants and implied effect of unincluded
  attributes
• Does it reproduce current travel patterns?
• Any beginner can match totals by adjusting Ks
• Scrutiny of markets and patterns within the
  totals
• Does it predict rational responses to change?
• For changes inherent in New Starts projects
• For all model components

27
Travel Markets

• Trip productions traveler characteristics
• Production-attraction flows
• Characteristics of travelers
• Implications for mode choice
• Calibration
• Forecasting

28
Mode Choice

• Unusual coefficients
• Bizarre alternative-specific constants
• Non-Logit decision rules
• Problems in choice-set formation

29
Transit Path-Builder and Mode Choice Model

• Conformance between parameters in
• Transit path selection
• Mode choice utility expressions for transit
  choices
• Consequences of disagreement
• Better paths may look worse to mode choice
• Build alternatives may lose some trips and
  benefits
• Consistency crucial possible exceptions
• Bifurcation of 1st wait time?
• Treatment of transfers?

30
Network Speeds

• Highway
• Replication of current average travel times
• Comparability between alternatives
• Bus
• Relationship to auto speeds
• Replication of current average travel times
• Handling of dead highway links

31
Bottom Line

• Purpose of models
• Insights into problems, solutions, benefits
• Development of a solid case for a project
• Required performance by models
• Remain consistent with current good practice
• Provide coherent insights
• Support a coherent story about the project