Southern California Association of Governments (SCAG) Metropolitan Planning Organization (AMPO) Annual Conference. Prepared for

Transcription

1 Congestion Pricing Modeling and Results for Express Travel Choices Study Kazem Oryani and Cissy Kulakowski, CDM Smith Portland, Oregon, October 22 25, 2013 Prepared for Southern California Association of Governments (SCAG) 2013 Association of Metropolitan Planning Organization (AMPO) Annual Conference 1

2 Objective To estimate revenue potentials and network performance measures for range of pricing scenarios as an input for policy discussion and selection for pre implementation analysis. 2

3 Southern California Association of Governments (SCAG) Year 2010 Population 18 Million SCAG Region is Home to 49 Percent of California Population Year 2035 Population 22 Million Increase of 4 Million Population in 25 Years or 160,000 Person Per Year (Each Year, One Small City Added) Ventura Co. Los Angeles Co. Orange Co. San Bernardino Co. Riverside Co. Imperial Co. 3

4 Year 2035 Traffic System Vehicle trips 48.8 million Average speed (mph) 34.7 Average trip length (miles) Average trip time (min) 20.9 Freeways Vehicle trips 29.7 million Average speed (mph) 43.9 Average trip length h( (miles) 10.9 Average trip time (min)

5 Inter County Person Trip Flows Weekday Work 5

6 Inter County Person Trip Flows Weekday Non Work 6

7 Inter County Person Trip Flows Weekday Total 7

8 New Model Components Changes in Time of day Travel Due to Pricing Trip Suppression Due to Pricing Route Choice Due to Pricing 8

9 Model System Structure Original SCAG Model SCAG Model Enhanced For Pricing Analysis Trip Generation Trip Generation Trip Distribution Mode Choice Time of Day Destination Choice Mode Choice Enhanced Time of day Trip Suppression Enhancements By CDM Smith Enhancements By PB Trip Assignment (Route Choice) Enhanced Trip Assignment (Route Choice) 9

10 Tools and Databases Existing Time Period Model AM peak (6am 9am) MD (9am 3pm) PM peak (3pm 7pm) Night (7pm 6am) 4 periods Enhanced Model: 30 (½ hour periods) (6am 9pm) 1 night period (9pm 6am) 31 periods 10

11 Model Estimation Trip suppression / time of day changes based on Stated Preference Survey: More than 3,600 samples Coverage (Six county SCAG Region): Imperial, Los Angeles, Orange, Riverside, San Bernardino, and Ventura Counties. 11

12 Model Estimation Time of Day Model Estimation Based on More Than 16,000 SCAG Household Travel Surveys in 2001 Including More Than 88,000 Full Person Trip Records. 12

13 Year 2010 Stated Preference Survey Stated Preference Survey to Support Model Changes 3,600 survey record for all six SCAG counties Discrete choice model by trip purpose: work business trips, non work Time of day: peak, off peak Hourly Traffic 10,000 9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 $2.00 $3.00 $4.00 $3.00 $2.00 5:00 6:00 7:00 8:00 9:00 10:00 Hour 13

14 Hypothetical Reaction to Pricing for Range of Fees Percent Share 100% 90% 80% 70% 60% 50% 40% 4% 5% 7% 4% 5% 6% 14% 13% 13% 7% 8% 8% 7% 7% 8% 8% 10% 12% 7% 8% 9% 12% 11% 11% 8% 9% 9% 8% 8% 8% 15% 17% 9% 10% 10% 10% 9% 9% 9% 9% 20% 11% 9% 9% 9% 23% 12% 8% 10% 9% 30% 20% 64% 62% 60% 57% 54% 51% 48% 45% 42% 38% 10% 0% $1 $2 $3 $4 $5 $6 $7 $8 $9 $10 Area Pricing Fee Current Destination Peak Current Destination Shift Late Alternate Destination Current Destination Shift Early Current Destination HOV Transit 14

15 Variables Used in Model Estimation Used Multinomial Logit Formulation for Time of day Model Logit Based Toll Diversion Model for Trip Assignment Utilized Enhanced Model for Scenario Analysis 15

16 Variables Used in Model Estimation Departure Time Arrival Time Origin Zone Destination Zone Trip Purpose Mode Traveler s Household Size Traveler s Household Income The Number of Household Workers The Number of Household Vehicles Traveler s Age Traveler s Employment Industry Type 16

17 HBWD From Home Trip Time of Day Choice Model Summary Alternatives Shift Constant Delay Distance Delay Shift Delay Shift^2 Variables in Utility Functions Distance Shift Distance Shift^2 Inc_H Inc_M_H Inc_M_L HH_Size Age Drive Alone Pop_O AM1 AM2 AM3 AM4 AM5 AM (7.647) (8.770) (9.447) (9.624) (8.661) (8.217) (-5.342) (-3.600) (2.626) (1.579) (-1.303) (3.279) (-1.009) (8.559) (5.948) (2.787) ( ) (-2.697) (2.125) (-1.397) MD1 MD2 MD3 MD4 MD5 MD6 MD7 MD8 MD9 MD10 MD11 MD (7.085) (6.305) (4.935) (4.763) (4.476) (4.289) (3.753) (5.300) (4.997) (5.352) (5.287) (4.997) (Constrained) (-7.908) (-3.524) (4.842) (-7.054) (-3.437) (3.292) (-1.874) PM1 PM2 PM3 PM4 PM5 PM6 PM7 PM8 PM9 PM10 PM11 PM (5.975) (5.674) (4.883) (5.301) (4.623) (4.515) (3.783) (3.981) (3.302) (2.015) (1.543) ( ) ( ) (-1.476) (-2.048) (2.408) ( ) ( ) (2.425) ( ) NT (7.720) Note: Value in parentheses is the t-statistics. Observations: 7,368 Final Log Likelihood: -19,733 ρ2 w.r.t. 0:

18 Change in Tripmaking (Trip Suppression / Inducement) Peak Non work Trip Toll Difference 0 Travel Time Difference $0.00 $ % 3.8% +1.2% 2.6% +2.4% 1.5% +3.6% 0.3% +4.7% +0.9% $ % 6.5% 5.3% 4.1% 2.9% $ % 10.3% 9.1% 7.9% 6.7% $ % 14.1% 12.9% 11.7% 10.6% $ % 1% 17.9% 16.7% 15.6% 14.4% 4% (Negative = Suppression, Positive = Inducement) 18

19 Trip Suppression Model The models were developed from regression analysis on the responses to the trip suppression question in the survey developed by comparing the change in trips made against the change in utility of each trip. The generic trip regression equation is shown as: Tr m * cos t Cost after * LN( income / ) LN( d 1) Where: ΔTr is the percentage difference in the number of trips m is the regression coefficient LN(d+1) is the natural log of trip distance in miles plus 1 Β cost is the toll cost coefficient Cost after is the toll cost with pricing Income/λ is the median household income divided by λ 19

20 Base Case Performance Measures (000 s) Total System AM Peak Midday PM Peak Evening Night Total Vehicle Trips VMT VHT 9, ,398 4,163 18, ,119 4,513 14, ,554 6,367 3,040 36, ,542 65,166 1,214 48, ,308 17,003 Average Speed (mph) Average Trip Length (miles) Average Trip Time (min) All Freeways AM Peak Midday PM Peak Evening Night Total Vehicle Trips VMT VHT 6,126 62,476 1,740 9, ,845 1,839 8,616 90,831 2,832 2,069 21, ,180 44, , ,684 7,378 Average Speed (mph) 35.9 AverageTrip Lengthon Fwy. (miles) 10.2 Average Trip Time on Fwy. (min) All Other Roads AM Peak Midday PM Peak Evening Night Total VMT VHT 58,922 2,422 83,274 2,675 88,722 3,535 14, , ,623 9,626 Average Speed (mph) AM Peak Midday PM Peak Evening Night Total Vehicle Trips Crossing Downtown Cordon ,062 20

21 Scenarios Examined Base Case Regional Freeway System 21

22 Scenarios Examined (cont d) 1 Strategic Express Lanes Network: HOV3+Free 2 Strategic Express Lanes Network: HOV2+Free Strategic Express Lane Network 22

23 Scenarios Examined (cont d) 3 Full Express Lanes Network: 1 Lane 4 Full Express Lanes Network: 2 Lanes Full Express Lanes Network 23

24 Scenarios Examined (cont d) Downtown Los Angeles Cordon Area 5 Downtown Cordon Pricing All Trips 6 Downtown Cordon d Pricing i i Destination i i Only 7 Full Freeway Pricing 8 Region wide Region wide VMT Fees Flat Rate 9 Region wide VMT Fees Variable Rate 10 Combination I Region wide Variable VMT Plus Strategic Express Lanes Network Plus Downtown Cordon Pricing 11 Combination II Region Wide Variable VMT Plus Strategic Express Lanes Network 12 Combination III Region Wide Flat Rate VMT Plus Strategic Express Lanes Network 24

25 Primary Metrics for Comparison of Scenarios Total Vehicle Trips in the Model Total Vehicle Miles of Travel Total Vehicle Hours of Travel Average Speed Average Trip Length Average Ti Trip Time 25

26 Year 2035 VMT Estimate (000 s) and Percent Change From Base Peak All Other Peak All Freeway Off Peak All Other Off Peak All Freeway X% = Difference From Base Case Regionwide We eekday VMT (in thousands) 700, , , , , , ,000 0 Base Case 0% 0% 0% 0% 1% 2% 0% Scenario 1 Scenario 2 HOV3+Free HOV2+Free Strategic Express Lanes Network Scenario 3 1 Lane Scenario 4 2 Lane Full Express Lanes Network Scenario 5 Scenario 6 All Trips Destination Only Downtown LA Cordon Scenario 7 Freeway Facility Pricing 4% 2% 1% 1% 2% Scenario 8 Flat Rate Variable Rate Mileage Based User Fee Scenario 9 Scenario 10 Scenario 11 Scenario Combination Scenarios 26

27 Year 2035 VHT Estimate (000 s) and Percent Change From Base Peak All Other Peak All Freeway Off Peak All Other Off Peak All Freeway X% = Difference From Base Case Regionw wide Weekday VHT 20,000 18,000 16,000 14,000 12,000 10,000 8,000 6,000 4,000 0% 0% 1% 0% 6% 6% 7% 11% 11% 9% 10% 9% 2,000 0 Base Case Scenario 1 Scenario 2 HOV3+Free HOV2+Free Strategic Express Lanes Network Scenario 3 1 Lane Scenario 4 2 Lane Full Express Lanes Network Scenario 5 Scenario 6 All Trips Destination Only Downtown LA Cordon Scenario 7 Freeway Facility Pricing Scenario 8 Scenario 9 Flat Rate Variable Rate Mileage Based User Fee Scenario 10 Scenario 11 Scenario Combination Scenarios 27

28 Comparison of Annual Gross Toll Revenue Potential Year 2035 (Billions of 2011 Dollars) $12 $11 $10 Annual Reven nue $9 $8 $7 $6 $5 $4 $3 $2 $1 $0 Scenario 1 HOV3+Free Scenario 2 HOV2+Free Strategic Express Lanes Network Scenario 3 1 Lane Scenario 4 2 Lane Full Express Lanes Network Scenario 5 Scenario 6 All Trips Destination Only Downtown LA Cordon Scenario 7 Freeway Facility Pricing Scenario 8 Scenario 9 Flat Rate Variable Rate Mileage Based User Fee Scenario 10 Scenario 11 Scenario Combination Scenarios 28

29 Summary of Results Categories of Congestion Pricing Options Mileage Based User Fees Variable Rate CONGES STION IM MPACT medium high lo ow Strategic Express Lanes HOV 2+Free Strategic Express Lanes HOV 3+Free Downtown LA Cordon Pricing low medium high REVENUE POTENTIAL Combination 2 Combination 1 Not yet implemented in the U.S. 29

30 Summary of Results (cont d) Lessons Learned: Continuous and close coordination between the consulting team, client and system integrator during the project was vital for the success of the project. This included last minute updates of models, re integration of model components and model re runs for scenario consistency purposes. Future Work: Pre implementation Analysis 30

31 Acknowledgements Contribution ofannie Nam, Guoxiong Huang, and Warren Whiteaker of Southern California Association of Governments, Linda Bohlinger of HNTB Corporation, Edward Regan of CDM Smith, Thomas Adler, Mark Fowler of RSG, Jim Lam of Caliper Corporation are greatly appreciated. 31

32 Sources 1. Model Enhancement, Technical Memorandum of Time of day Model Development Express Travel Choices Study, by CDM Smith for Southern California Association of Governments, October Technical Memorandum, Summary of Modeling Results, Alternative Congestion Pricing Strategies, Express Travel Choices Study, by CDM Smith for Southern California Association of Governments, November Choices Perspectives on Southern California Traffic Congested Express Travel Choices Study. California Association of Governments, June