FR1 CCTA Travel Model Documentation COMPLETEtechnical details required in the model documentation...

Decennial Model Update CCTA Travel Model Documentation

finalreport

prepared for

Contra Costa Transportation Authority

prepared by

Cambridge Systematics, Inc.

with

Dowling Associates, Inc. Caliper Corporation

June 2003


Cambridge Systematics, Inc. i

Table of Contents

1.0 Introduction ............................................................................................................... 1-1 1.1 The Authority’s Modeling Needs.................................................................. 1-2 1.2 Model Update Objectives ............................................................................... 1-2 1.3 Model Framework ........................................................................................... 1-3

2.0 Validation Data ......................................................................................................... 2-1 2.1 Introduction...................................................................................................... 2-1 2.2 Validation Database Development ............................................................... 2-1 2.3 Database Architecture..................................................................................... 2-2 2.4 Transit Ridership Data.................................................................................... 2-4 2.5 Freeway Ramp Count Data ............................................................................ 2-7 2.6 Screenline Counts ............................................................................................ 2-7 2.7 Intersection Counts.......................................................................................... 2-11 2.8 Travel Speeds ................................................................................................... 2-13 2.9 HOV Lane Usage Data.................................................................................... 2-14

3.0 Highway Networks................................................................................................... 3-1 3.1 Introduction...................................................................................................... 3-1 3.2 Development of the Master Highway Network.......................................... 3-2

4.0 Transit Networks ...................................................................................................... 4-1 4.1 Introduction...................................................................................................... 4-1 4.2 Development of the Transit Master Networks............................................ 4-1

5.0 Zonal Data .................................................................................................................. 5-1 5.1 CCTA Master Zonal Database ....................................................................... 5-1

6.0 CCTA Model Development .................................................................................... 6-1 6.1 MTC Model System Overview ...................................................................... 6-1 6.2 Workers and Vehicles in Household Nested Choice Model ..................... 6-5 6.3 Trip Generation Models.................................................................................. 6-9 6.4 Trip Distribution Models................................................................................ 6-12 6.5 Mode Choice Models ...................................................................................... 6-16 6.6 Trip Assignment .............................................................................................. 6-28

7.0 Model Refinements .................................................................................................. 7-1 7.1 Introduction...................................................................................................... 7-1 7.2 School Trip Attractions Model....................................................................... 7-2 7.3 Special Generators ........................................................................................... 7-5


ii Cambridge Systematics, Inc.

Table of Contents (continued)

7.4 Park-and-Ride Trips ........................................................................................ 7-13 7.5 Peak Period and Daily Models ...................................................................... 7-15 7.6 Peak Hour Model............................................................................................. 7-17

8.0 Model Validation ...................................................................................................... 8-1 8.1 Systemwide Validation ................................................................................... 8-2 8.2 Highway Assignment Validation.................................................................. 8-2 8.3 Transit Assignments........................................................................................ 8-14 8.4 Link-Based Highway Validation Tests ......................................................... 8-14

9.0 Forecasts...................................................................................................................... 9-1 9.1 Introduction...................................................................................................... 9-1 9.2 Future Scenarios............................................................................................... 9-1 9.3 Land Use Data.................................................................................................. 9-2 9.4 Special Generator Data ................................................................................... 9-3 9.5 Supplementary Data Files .............................................................................. 9-5 9.6 Highway Network Data ................................................................................. 9-5 9.7 Transit Network Data ..................................................................................... 9-6 9.8 Future Intersection Geometry........................................................................ 9-7 9.9 Forecast Results................................................................................................ 9-7

10.0 Level of Service (LOS).............................................................................................. 10-1 10.1 Introduction...................................................................................................... 10-1 10.2 CCTALOS Validation...................................................................................... 10-4 10.3 Choose Input Files ........................................................................................... 10-4 10.4 Adjust Model Output...................................................................................... 10-5 10.5 Compute CCTALOS........................................................................................ 10-5 10.6 Compute ICU ................................................................................................... 10-8


Cambridge Systematics, Inc. iii

List of Tables

2.1 Data Types in Validation Database ......................................................................... 2-2

2.2 Linkage of Validation Database to CCTA Model Networks ............................... 2-3

2.3 Daily Boardings of Transit Operators ..................................................................... 2-5

2.4 RMSE Comparison of Caltrans Ramp Estimates (Caltrans Census) and Intersection Approach Volumes (Pang Ho, Cities) ............................................... 2-10

2.5 RMSE Comparison of Screenline Counts (MultiTrans, MTCI, Caltrans Census) Versus Intersection Approach Volumes (Pang Ho, Cities)................... 2-11

2.6 Mapping Turn Volumes into Approach and Departure Volumes ..................... 2-12

2.7 Caltrans Vehicle Occupancy Data ........................................................................... 2-15

2.8 Route 4 at Bailey, Westbound AM Peak Hour....................................................... 2-15

2.9 I-80, East of Hilltop, Westbound.............................................................................. 2-15

2.10 I-80, East of Hilltop, Eastbound ............................................................................... 2-16

2.11 I-680, at Stone Valley Road, Southbound ............................................................... 2-16

2.12 I-680, at Stone Valley Road, Northbound............................................................... 2-16

4.1 Number of Transit Routes by Location................................................................... 4-2

4.2 Transit Non-Motorized Access Links...................................................................... 4-3

4.3 Mode Table ................................................................................................................. 4-4

5.1 MTC Zone Data Files................................................................................................. 5-1

6.1 WHHAO Multinomial and Nested Choice Models – Model #9W ..................... 6-6

6.2 Detailed Definition of Variables used in BAYCAST Travel Demand Models (in Alphabetical Sort Order) ....................................................................... 6-7

6.3 Summary of BAYCAST Trip Generation Models.................................................. 6-10


iv Cambridge Systematics, Inc.

List of Tables (continued)

6.4 Final Friction Factors for BAYCAST Trip Distribution Models .......................... 6-13

6.5 Value of Time Estimates by Trip Purpose*............................................................. 6-18

6.6 Final Multinomial and Nested Home-Based Work Mode Choice Models – Multinomial Model #99W and Nested Model #97 .............................. 6-20

6.7 Final Nested Home-Based Shop/Other Mode Choice Model – Nested Model #73W – Nest 2 ................................................................................................ 6-21

6.8 Final Nested Home-Based Social/Recreation Mode Choice Model – Nested Model #35 ...................................................................................................... 6-22

6.9 Final Nested Non-Home-Based Mode Choice Model – Nested Model #14W-2 ............................................................................................................ 6-23

6.10 Final Home-Based School (Grade School) Mode Choice Model – Multinomial Logit Model #21W .............................................................................. 6-24

6.11 Final Nested Home-Based School (High School) Mode Choice Model – Nested Model #18W-3 ............................................................................... 6-25

6.12 Final Nested Home-Based School (College) Mode Choice Model – Nested Model #28W 2 ............................................................................................... 6-27

6.13 Terminal Times........................................................................................................... 6-31

6.14 Transit Travel Time Parameters............................................................................... 6-32

7.1 MTC School Trip Production and Attraction Equations ...................................... 7-3

7.2 School Enrollment by Super-District....................................................................... 7-3

7.3 Special Generators by Type and Size ...................................................................... 7-7

7.4 Special Generator Employment Estimates ............................................................. 7-9

7.5 Special Generator Trip Estimates............................................................................. 7-11

7.6 Assumptions for Special Generator Categories* ................................................... 7-12

7.7 Recommended Special Generators .......................................................................... 7-13


Cambridge Systematics, Inc. v


7.8 Drive Access Trips to Bart Stations inside the CCTA Study Area ...................... 7-15

7.9 Peaking Factors for AM and PM Peak Period ....................................................... 7-17

7.10 Comparison of Peak-Hour Assignments With and Without ODME ................. 7-19

8.1 Systemwide Validation Test ..................................................................................... 8-2

8.2 Validation of Highway Assignment by Facility Type and Time Period............ 8-3

8.3 Validation of Highway Assignment by Area Type and Time Period ................ 8-4

8.4 Validation of Highway Assignment by Screenline for A.M. Peak and P.M. Peak Hour................................................................................................... 8-10

8.5 Validation of Highway Assignment by Screenline for A.M. Peak Period, P.M. Peak Period, and Average Daily Traffic........................................... 8-12

8.6 Validation of Highway Assignment by Screenline for A.M. Peak Period, P.M. Peak Period, and Average Daily Traffic........................................... 8-13

8.7 Transit Assignment Validation Summary.............................................................. 8-15

8.8 Link-Based Validation Tests by Time Period......................................................... 8-16

9.1 Future Scenarios ......................................................................................................... 9-1

9.2 Future Schools in Dougherty Valley ....................................................................... 9-2

9.3 Special Generators ..................................................................................................... 9-3

9.4 Assumptions for Special Generator Categories..................................................... 9-4

9.5 Sources of Each Network Scenario .......................................................................... 9-6

9.6 AM Peak-Hour County-to-County Trip Table Forecasts ..................................... 9-8

9.7 PM Peak-Hour County-to-County Trip Table Forecasts...................................... 9-9

9.8 AM Peak-Period County-to-County Trip Table Forecasts................................... 9-10


vi Cambridge Systematics, Inc.


9.9 PM Peak-Period County-to-County Trip Table Forecasts.................................... 9-11

9.10 Off-Peak County-to-County Trip Table Forecasts................................................. 9-12

9.11 AM Peak Hour Screenline Results........................................................................... 9-13

9.12 PM Peak Hour Screenline Results ........................................................................... 9-14

9.13 AM Peak Period Screenline Results ........................................................................ 9-15

9.14 PM Peak Period Screenline Results ......................................................................... 9-16

9.15 Daily Screenline Results ............................................................................................ 9-17


Cambridge Systematics, Inc. vii

List of Figures

1.1 Regional Geographic Coverage in the CCTA Travel Model ............................... 1-4

1.2 Subarea Geographic Coverage in the CCTA Travel Model ................................. 1-5

2.1 Sample Interchange in TransCAD........................................................................... 2-13

6.1 Bay Area Travel Demand Model Forecasting System (BAYCAST).................... 6-2

6.2 WHHAO – Nested Choice Model #9W .................................................................. 6-9

6.3 Home-Based Work Mode Choice – Nested Model #97 ........................................ 6-17

6.4 Home-Based Shop/Other Mode Choice – Nested Model #2 .............................. 6-21

6.5 Home-Based Social/Recreation Mode Choice – Nested Model #35 .................. 6-23

6.6 Non-Home-Based Mode Choice – Nested Model #14W 2................................... 6-24

6.7 Home-Based School: Grade School Mode Choice – Model #21W .................... 6-25

6.8 Home-Based School: High School Mode Choice – Model #18W-Nest 3.......... 6-26

6.9 Home-Based School: CollegeMode Choice – Model #28W – Nest 2 ................. 6-27

6.10 Location of External Gateways ................................................................................ 6-30

6.11 Volume-Delay Functions by Facility Type ............................................................. 6-34

7.1 School Locations in Contra Costa ............................................................................ 7-4

7.2 Distribution of Special Generators .......................................................................... 7-8

7.3 Production-Attraction and Origin-Destination Trips ........................................... 7-16

8.1 East County Screenlines ............................................................................................ 8-5

8.2 Central County Screenlines ...................................................................................... 8-6

8.3 West County Screenlines........................................................................................... 8-7

8.4 Tri-Valley Screenlines................................................................................................ 8-8


viii Cambridge Systematics, Inc.

List of Figures (continued)

8.5 AM Peak Hour Screenline Validation..................................................................... 8-9

8.6 PM Peak Hour Screenline Validation...................................................................... 8-9

10.1 CCTA Interface........................................................................................................... 10-1

10.2 Turning Movement Diagram ................................................................................... 10-2

10.3 Data Diagram for CCTA Technical Procedures Implementation ....................... 10-3

10.4 Input Turning Movement Table .............................................................................. 10-7


Cambridge Systematics, Inc. 1-1

1.0 Introduction

The Decennial Model Update was a process to update current models being used for transportation planning applications in Contra Costa County. This process involved implementing the recommendations in the Modeling Evaluation Study, which addressed the planning and modeling needs of the Contra Costa Transportation Authority (CCTA) over the next 10 years. A brief summary of the needs and objectives of the study are presented in the next sections. In addition, an overview of the overall modeling framework is pre-sented to describe the geographic coverage, trip purposes, modes, and time periods used in the model. For the remainder of this report, the updated Contra Costa travel demand forecasting model will be referred to as the CCTA Travel Model.

The CCTA Travel Model was developed using a systems-up approach. This is a process for model development that develops and tests all model components before validation of the full model is complete. This approach used initially available or estimated input data for model development and Beta testing in Phase I. Final versions of input data were used for model calibration and validation and are referred to as Phase II. All results presented in this report are the product of Phase II.

This report is one of five reports written to document the work completed during the Decennial Model Update study:

1. Executive Summary;

2. CCTA Travel Model Documentation;

3. CCTA Travel Model User’s Guide;

4. CCTA Travel Model Technical Appendices; and

5. MTC Consistency Report.

The purpose of this report is to document the process of preparing the CCTA Travel Model and to provide results of the validation and forecasting model runs. The purpose of the executive summary is to provide a brief overview of the study and summaries of validation and forecasting results. The purpose of the user’s guide is to provide technical guidance on the use of the models, including documenting new software procedures developed as part of this study. The purpose of the technical appendices is to provide technical details required in the model documentation and user’s guide that are too volu-minous to be placed within these reports. The purpose of the MTC consistency report is to compare the CCTA travel model results with the MTC model results for each of the model components required by the MTC.


1-2 Cambridge Systematics, Inc.

1.1 The Authority’s Modeling Needs

In the early 1990s, the CCTA developed and has since been maintaining four subarea models and one Congestion Management Program (CMP) model. These models have been used extensively by the Authority, local jurisdictions, and private developers to gen-erate the 20-year traffic forecasts that are needed to address the transportation planning requirements of Measure C and the CMP. While the models are fully functional, it has become increasingly difficult and expensive to maintain all the models at an equal state of readiness. To address these needs, the Authority has updated their modeling capabilities to achieve the following goals:

• The modeling approach was restructured from four subarea models to one fine-grained countywide model based on recommendations from Cambridge Systematics as part of the Modeling Evaluation Study;

• The updated model contains approximately 1,500 zones in Contra Costa and the Tri-Valley region, and retains the 1,099-zone structure of the MTC’s model for the remainder of the bay area;

• The updated model also involved software conversion from EMME/2 to TransCAD®; and

• The Authority’s new countywide model is also consistent with MTC’s 2001 CMP Modeling Consistency Guidelines.

1.2 Model Update Objectives

To achieve the overall goals for the model update, a more specific set of objectives was developed as part of this model update:

• To develop a single, fine-grained, countywide model that addresses all of the Authority’s modeling requirements over the next 10 years. An added advantage of a single model over several subarea models is that adjustments made for an individual area would be included in all other applications of the model.

• To develop a model that meets both the Growth Management Programs (GMP) and CMP needs and allows the Authority to provide these results without having to maintain, store, and document five individual models for these purposes.

• To provide consistency for all planning applications in the County and eliminate ques-tions concerning inconsistent forecasts at the subarea borders.

• To use the travel demand forecasting software TransCAD to take advantages of its user-interface and GIS Integration.



• To review the assumptions for the input data and update them. For instance, the new Metropolitan Transportation Commission (MTC) models have additional require-ments for input land use data, such as school and college enrollment, which have not been standardized inputs to the subarea models.

• To include special generators to account for unique trip generation rates for hospitals, parks, or shopping centers.

• To update pricing and auto ownership assumptions to match the MTC Bay Area Travel Demand Model Forecasting System (BAYCAST).

• To review network attributes for reasonableness and update them.

All of the objectives have been achieved during the model update, and the documentation of the specific modeling features addressing these objectives is contained herein.

1.3 Model Framework

Geographic Coverage

The CCTA Travel Model was developed to include the nine-county region currently mod-eled by MTC (Marin, Napa, Sonoma, Solano, Contra Costa, Alameda, Santa Clara, San Mateo, and San Francisco). This area is presented in Figure 1.1 and represents the 1,099-zone system in use by the MTC in 2003. The geographic coverage is consistent with the MTC model in the areas outside the Contra Costa County study area and considerably more detailed for the area within the study area. Of the 1,099 zones in the MTC region, 921 of these are retained in areas outside the study area. The CCTA Travel Model study area is defined by the four subareas: West County, Central County, East County, and Tri-Valley. These subareas are presented in Figure 1.2. The CCTA study area contains 1,460 zones for a total of 2,581 in the full model.



Figure 1.1 Regional Geographic Coverage in the CCTA Travel Model



Figure 1.2 Subarea Geographic Coverage in the CCTA Travel Model



For purposes of clarification, the following items are defined for use in this project:

• CCTA study area – The CCTA 2000 Model Update study area includes Contra Costa County and the portion of Alameda County that is within the Tri-Valley subarea model.

• Region – This is equivalent to the MTC nine-county San Francisco Bay Area region. This is also the regional travel model study area.

• Externals – These are trips that are outside the nine-county region.

Trip Purposes

There are five primary trip purposes in the CCTA Travel Model. These are the same trip purposes defined by the MTC model for intraregional personal travel:

1. Home-based work (HBW),

2. Home-based shop and other (HBSH),

3. Home-based social and recreation (HBSR),

4. Home-based school (HBSK), and

5. Non-Home-Based (NHB).

Home-based school trips are further broken down into:

• Home-based school: Grade school (HBGS),

• Home-based school: High school (HBHS), and

• Home-based school: College (HBCol).

In addition to the intraregional person travel, there are additional sets of trip making included in the model, as follows:

• Trucks, and

• External trips (including internal-external and external-internal trips).

Modes

There are seven modes represented in the seven individual mode choice models (one for each of the seven personal trip purposes), but not all modes are represented in each model. These modes are defined by the MTC regional model. The home-based work model contains all seven modes:



1. Drive alone,

2. Shared ride 2 persons,

3. Shared ride 3+ persons,

4. Walk access to transit,

5. Drive access to transit,

6. Bike, and

7. Walk.

The home-based shop/other and home-based social/recreation trip purposes contain six modes, including all of the above, with walk and drive access to transit combined as a sin-gle transit mode. The non-home-based, home-based high school, and home-based college trip purposes have five modes, as follows: vehicle driver, vehicle passenger, transit, bike, and walk. The home-based grade school mode choice model has only four modes, as fol-lows: vehicle passenger, transit, bike, and walk.

Time Periods

There are three time periods in the CCTA Travel Model; these are a departure from the time periods currently being used in the MTC regional model. The three time periods are defined as follows:

1. a.m. peak period (6:00 a.m. to 10:00 a.m.);

2. p.m. peak period (3:00 p.m. to 7:00 p.m.); and

3. Off-peak periods (12:00 a.m. to 6:00 a.m., 10:00 a.m. to 3:00 p.m., and 7:00 p.m. to 12:00 a.m.).

The a.m. and p.m. peak periods were developed by evaluating the traffic counts in the CCTA study area over a 24-hour period.

In addition to the a.m. and p.m. peak periods, the CCTA Travel Model contains a.m. and p.m. peak-hour models that represent the highest hour volume at any location in the sys-tem. The peak hour can be any hour, based on 15-minute time increments, during the four-hour peak period.



2.0 Validation Data

This section documents the development and contents of the Validation Database for the Decennial Model Update. Information on the content of the validation database can be found in the CCTA Travel Model User’s Guide.

2.1 Introduction

The Validation Database was developed for the purpose of storing model validation information in a readily accessible form that is also linked to the CCTA TransCAD travel demand model.

This section first describes the steps involved in the development of the Validation Database. It then describes the architecture and content of the database. The following sections then describe the origin, quality control checks, and processing of the data contained in the database. These sections are organized into historic traffic counts, current traffic counts, high-occupancy vehicle (HOV) lane usage, travel speeds, and transit ridership.

The section concludes with how to access and modify the database, along with guidance on the implications of various highway and transit network editing processes on the database.

2.2 Validation Database Development

The Validation Database was developed in five steps. First, the overall database archi-tecture was designed. Then, the data was collected and reviewed/processed for use in the database. The processed data was then imported into the Validation Database. Next, the data was linked to the CCTA model master highway network in TransCAD. Finally, the database development was documented in a series of memoranda and this report.

Memoranda on the Validation Database were distributed to the Transportation Modeling Working Group (TMWG) on October 10, 2001; July 24, 2002; and October 28, 2002.



2.3 Database Architecture

The Validation Database consists of daily transit boarding data, daily and peak-period freeway ramp counts, peak-period intersection turn counts, 24-hour screenline counts, and peak-period freeway speed data (see Table 2.1).

Table 2.1 Data Types in Validation Database

Data Type Description 1. Transit boardings Daily boardings for all operators in Bay Area. Some more detailed data

available for the Alameda Contra-Costa Transit District (AC Transit) and the Bay Area Rapid Transit District (BART). Most data are from 1998, and some are from 1995.

2. Freeway ramp counts

California Department of Transportation (Caltrans) freeway on- and off-ramp counts (daily, a.m./p.m. peak hours and peak periods).

3. Intersection turn counts

Intersection turning movements and link approach and departure volumes for a.m. and p.m. peak hour. The date of the data varies by location between the years 2000 and 2002.

4. Screenline counts Screenline volumes for 15-minute increment, 1-hour increment, a.m. and p.m. peak hour, a.m. and p.m. peak period, and daily. The date of the data varies by location between the years 1998 and 2002.

5. Speed data Peak-period floating car mean speeds by freeway route for mixed-flow lanes only (for 1997)

Historic counts (counts from the previous CCTA model development effort) are stored in a separate unlinked spreadsheet comparing year 2000 screenline counts to the 1990 counts. The historical counts are not used in the model validation and are of primarily historical interest. They were, therefore, not included in the database itself. Historic intersection turning movement counts made in 1990 are not included in the database.

HOV lane counts for the year 2000 were gathered from Caltrans, but were not included in the Validation Database. The data was available only for one location on State Route 4, one location on I-80, and one location on I-680. For the I-80 freeway, the counts lumped two-person carpools in with single-occupant vehicles (SOVs). The data was so sparse and required a great deal of interpretation if it was to be used to validate the model. (The method that the MTC model uses to allocate HOVs to HOV lanes and mixed-flow lanes does not accurately reflect actual lane use.) Consequently, it was not included in the Validation Database, but is reported here (in a later subsection) for information only.



Linkage to CCTA Networks

The validation data is linked to the CCTA TransCAD highway and transit networks according to intersection identification (ID) number, link ID number, and transit operator. Table 2.2 shows these linkages to the highway and transit networks.

Table 2.2 Linkage of Validation Database to CCTA Model Networks

Data Type Data Table Linked? CCTA

Network How?

All Modes Boardings Yes Transit By Route Group 1. Transit Boardings All Operators Boardings Yes Transit By Operator

AC Transit Line Boardings

No N/A N/A

BART Station Boardings No N/A N/A

BART Line Boardings No N/A N/A

Alameda County Yes Highway By Directional Link 2. Freeway Ramp Counts

Contra Costa County Yes Highway By Directional Link

Intersection Approach Counts

Yes, by Macro Highway By Intersection Approach 3. Intersection Turn Counts Intersection Turn

Counts Yes Highway By Intersection

Screenline Counts Yes, by Macro Highway By Directional Link 4. Screenline Counts MTC Screenline Counts No N/A N/A

5. Speed Data MTC Speed Data Partially Highway By Node Start/Stop

Not all validation data in the Validation Database is linked to the CCTA highway and transit networks. This is either because the data was considered to be of secondary importance to model validation (superior data being available), or because of structural incompatibilities between the available data and the highway and transit networks.

If the resources required to establish the linkage were judged to be greater than the prob-able value of the linkage to model validation, then no linkage was developed. With or without linkage, the data is readily available in the database. The question was whether the calibration process is made sufficiently more efficient to warrant the investment and maintenance required to establish the linkage.

In the case of BART station-to-station data and the MTC travel time data, the point-to-point data was not easily compatible with link-by-link display on the highway and transit networks. This data was also considered to be of secondary importance to the model



validation effort. Thus, this data is contained in the Validation Database, but without a linkage to the highway or transit networks.

Similarly, the AC Transit boardings by route data are contained in the Validation Database, but are not linked to the transit network. This is because this detailed data was not available for the other major non-rail operators in Contra Costa. Thus, establishing the linkage for AC transit line data was not judged to be sufficiently useful to the validation process. The data is readily available in the database; it just cannot be displayed on the transit route layer at this time.

The MTC BAYCAST speed data from 1997 is partially linked to the CCTA highway net-work through identification of the network node numbers (Fields “A” and “B” in the data-set), where the travel time run started and ended. These fields were filled in for 875 of the 1,026 records in the speed dataset before it was decided that this information would be of marginal use to model development and calibration. (It was quite laborious looking up the appropriate node number for each start and stop point of the floating car runs, and associating the data with the appropriate links in between the points.) The age of the data and the difficulty of ensuring that TransCAD would select the same path (when creating skims) between starting and end points, as was actually run by the floating cars, sug-gested that project resources could be better spent on other endeavors.

The following sections describe the development of the validation data in more detail.

2.4 Transit Ridership Data

Regional and local operator daily boardings and transfers are stored in data tables with fields for route ID; route; operator; type of service (e.g., express, local); and boardings and transfers. This data is linked to individual TransCAD route systems by route ID. Table 2.3 gives details about the services and daily boardings of each operator.

Most operators keep only patronage data on routes. BART, through its data acquisition system (DAS), has automated origin/destination data by station and time of day. For AC Transit, the 1998 boarding/alighting survey, which covered every bus run on the system, may well be the best source of detailed counts by location and time of day.

Segment data are the most difficult to acquire, because few operators collect data on boardings and alightings.

The following are the major sources of transit patronage data for the transit database:

• BART Trips Between Line Segments (Lines F, K, C, R, D, M) 1998 data – Daily (Document: BART_segsum_RVAL98_10.xls);

• Regional Operator Weekday Boardings (BART, AC Transit, GGT, CCTA) Observed and Predicted, 1998 (Document: RVAL98 Transit Validation.xls);



Table 2.3 Daily Boardings of Transit Operators

Transit Operator Daily

Boardings Year Source AC Eastbay 180,063 1998 98 Operator Report AC Transbay 14,452 1998 98 Operator Report AirBART 500 1998 Analyst Estimate Amtrak 1,300 1998 Capitol: (1200) Adjusted from March 98 Transactions;

San Joaquin: (100) analyst estimate BART 275,742 1998 98 Operator report Benicia 500 1995 95 Operator report BEST 500 1998 Analyst Estimate BWS 1,000 1998 Analyst Estimate CalTrain 27,967 1998 98 Operator Report CCCTA 16,337 1998 98/99 from Aug./Sept. Transactions Emeryville 1,000 1998 Analyst Estimate Fairfield 3,000 1997 97 Operator Report Ferry – Eastbay 1,638 1998 98 Operator Report Ferry – Tiburon 1,389 1997 Dec 97 Transactions Ferry – Vallejo 950 1998 MTC, Dec. 98 Golden Gate Transit – Bus 32,118 1998 98 Operator Report Golden Gate Transit – Ferry 5,118 1998 98 Operator Report LAVTA 3,875 1998 MTC, Dec. 98 Muni-Bus 564,767 1998 97/98 Operator Report Muni-Metro 137,045 1998 97/98 Operator Report NVT/Vine 3,023 1996 96 Operator Report SamTrans 61,845 1998 MTC, Dec. 98 Santa Clara Valley Transit – Bus

149,050 1998 Interpolation from 97 & 99 Operator Reports

Santa Clara Valley Transit – LRT

23,400 1998 Interpolation from 97 & 99 Operator Reports

Shuttles 10,800 1998 98 Operator Reports Sonoma operators 11,463 1998 MTC, Dec. 98 Tri-Delta 6,200 1997 97 Operator Report Union City 1,700 1998 MTC, Dec. 98 Vacaville 514 1995 95 Operator Report Vallejo-Bus 9,500 1998 MTC, Dec. 98 WestCAT/30Z 1,633 1998 MTC, Dec. 98 Total 1,548,389

MTC = Statistical Summary of Bay Area Transit Operators, MTC, December 1998.



• All Transit Operator Boardings (BART, AC Transit, GGT, CCTA, LAVTA, TriDelta, WestCAT, Fairfield, Benicia, Amtrak) Observed (1990 and 1998) and Predicted (1998) (Document: RVAL98 Transit Validation.xls);

• All Transit Operator (BART, AC Transit, GGT, CCTA, LAVTA, TriDelta, WestCAT, Fairfield, Benicia, Amtrak) Predicted Boardings by Trip Purpose and Intrazonal vs. Interzonal, 1998 and Share of Predicted Boardings by Trip Purpose and Intrazonal vs. Interzonal, 1998 (Document: RVAL98 Transit Validation.xls);

• Transit Operator Weekday Boardings by Technology (Heavy Rail, Light Rail, Ferry, and Bus), Observed vs. Predicted, 1998 (Document: RVAL98 Transit Validation.xls);

• BART Ridership by Station, Observed and Predicted, 1998 (RVAL98 Transit Validation.xls);

• 1998 Transit Validation by Sub-Operator (AC Transit and GGT) Observed and Predicted Weekday Boardings (Document: RVAL98 Transit Validation.xls);

• Source of Data for 1998 Transit Validation by Operator (BART, CCTA, LAVTA, TriDelta, WestCAT, AC Transit, GGT, Fairfield, Benicia, Amtrak) (RVAL98 Transit Validation.xls);

• AC Transit Assignment Validation by Sub-Operator of Observed and Predicted Weekday Boardings 1998 (Document: RVAL98 Transit Validation.xls);

• GGT Assignment Validation by Sub-Operator of Observed and Predicted Weekday Boardings 1998;

• All Transit Operator Boardings (BART, AC Transit, GGT, CCTA, LAVTA, TriDelta, WestCAT, Fairfield, Benicia, Amtrak) Observed (1990 and 1998) and Predicted (1998) with source and Intrazonal vs. Interzonal prediction (Document: RVAL98 Transit Validation.xls);

• Regional Operator Boardings, Observed and Predicted for Large Operators 1998 (AC Transit, CCCTA, Golden Gate BART) (Document: RVAL98 Transit Validation.xls);

• Transit Boardings by Operator by Purpose – Predicted and Observed 1998 (BART, AC Transit, GGT, CCTA, LAVTA, TriDelta, WestCAT, Fairfield, Benicia, Amtrak) (RVAL98 Transit Validation.xls);

• Transit Boardings by Operator by Purpose – Shares (Benicia, GGT, BART, Amtrak) (Document: RVAL98 Transit Validation.xls);

• Compare TP+ Transit Assignment (Percentage) by Purpose 1998 (AC Transit, LAVTA, CCCTA, TriDelta, WestCAT, Fairfield, Benicia, GGT, BART, Amtrak) (Document: RVAL98 Transit Validation.xls);



• Bay Area Regional Trips (numbers and percentage share) by Trip Purpose and Travel Mode 1998 (Document: RVAL98 Transit Validation.xls);

• Trip-iz (?) by Operator (AC Transit, LAVTA, CCCTA, TriDelta, WestCAT, Fairfield, Benicia, GGT) (1998?) (Document: RVAL98_10);

• Daily Boardings by Line (AC Transit, LAVTA, CCCTA, TriDelta, WestCAT, Fairfield, Benicia, GGT) (1998?) (Document: RVAL98_10);

• Regional Transit Operator Boardings (BART, AC Transit, GGT, CCTA, LAVTA, TriDelta, WestCAT, Fairfield, Benicia, Amtrak) Observed and Predicted (1998) with source and Intrazonal vs. Interzonal prediction (Document: RVAL98_10); and

• BART Ridership for sample months July – December 2000 by station (BART_2000_SampleMonths.xls).

2.5 Freeway Ramp Count Data

Freeway ramp counts were obtained from Caltrans by CCTA staff. Caltrans counts free-way ramps once every three years according to the following cycles:

• The ramps on state routes: 12, 29, 35, 37, 116, 121, 128, 131, I-205, 221, I-280, I-580, and I-880 were counted during years 1990, 1993, 1996, 1999, and 2002;

• The ramps on state routes: 4, 13, 24, 61, 77, I-80, 82, 84, 92, 109, 112, 114, 123, 160, 185, 238, 242, 260, 262, I-680, and I-980 were counted during years 1991, 1994, 1997, 2000, and 2003; and

• The ramps on state routes: 1, 9, 17, 25, 85, 87, 101, 113, 130, 152, 156, 220, 237, I-380, I-505, I-780 were counted during years 1992, 1995, 1998, and 2001.

The count for the year nearest to the year 2000 was entered into the database. The count data was aggregated to a.m. peak hour, p.m. peak hour, a.m. peak period, p.m. peak period, and daily. The start time of the peak hours was determined from the counts to the nearest whole hour (e.g., 6:00 a.m., 7:00 a.m., etc.). The a.m. and p.m. peak periods are each three hours long (inconsistent with the four-hour length of peak periods used in the model).

2.6 Screenline Counts

Directional link traffic count data is stored in data tables with fields for location, date, day, jurisdiction, direction, 15-minute interval begin time, and actual count by direction and



total count. These data tables are uniquely related to the TransCAD “Master” geographic file line layer using a unique location code. For screenlines, this is a combination of screenline number, screenline location, and screenline direction. Then, as with all related data tables, the user can use TransCAD (or other database management software) to select and process records based on specific filters (e.g., by screenline, day = t-th, etc.); and data statistics can be compiled.

Directional link traffic count data, including freeway ramp volumes, have been collected and put into the database. Count data were stored for the following five time periods:

1. a.m. peak hour (highest hour at each location),

2. a.m. peak four hours (6:00 a.m. to 10:00 a.m.),

3. p.m. peak hour (highest hour at each location),

4. p.m. peak four hours (3:00 p.m. to 7:00 p.m.), and

5. Daily (24 hours).

This data has been uniquely related to the TransCAD network line layer using a code that is a combination of the screenline number, screenline location, and direction of flow. (This is important since TransCAD links are bi-directional and data must be coded in the AB and BA directions.) Then the user can use TransCAD to select and process records based on specific filters (e.g., by screenline and direction); and data statistics can be compiled. TransCAD has a built in dataview statistics function that creates a .DBF summary of the count, sum, minimum, maximum, mean, and standard deviation for each field in the dataview.

Count data included the following:

• 24-hour directional count data (some counts for more than one day) at up to 200 surface street locations broken down into 15-minute periods;

• 24-hour directional freeway mainline counts broken down into one-hour periods; and

• a.m. and p.m. peak three-hour period freeway ramp counts at 371 locations broken down into 15-minute periods.

MTC Screenline Volumes

MTC screenline counts are stored in the Validation Database table: Data_MTCscreenline. The table below explains the fields.



Field Name Field Description Units Data Type Field Size

SL_ID Screenline Number Text 255

Facility Location Text 255

AM_PkPer AM Peak Period Volumes Number Double

Source Source where data were obtained from Text 255

The MTC screenline data came from the 1998 Base Year Validation of Travel Demand Models for the San Francisco Bay Area (BAYCAST-90) Technical Summary (published in May 2001 by MTC). The data were obtained from Table 6.5 (A.M. Peak Period Traffic Assignment Validation at Key Locations, 1998). The MTC screenline data is not linked to the CCTA master highway network, because the screenlines fall generally outside of the CCTA study area. They are considered a secondary source of count data for validation. The CCTA screenline counts, located inside the CCTA area, are considered the primary source of count data for validation.

Freeway Mainline Volume Estimates

Freeway mainline volumes were estimated for the screenlines by Wilbur Smith Associates using Caltrans mainline counts (taken at nearby locations), and manually adjusted for ramp counts between the mainline count location and the screenline location.

The Caltrans Census Program contracts out for the collection of traffic count data at mainline control stations, as well as on- and off-ramps on a three-year schedule. This data is stored in data tables with fields for location type, control station number, county, route, milepost, leg, count year, etc. The data tables provided by Caltrans are “raw data only.”

Caltrans provides the disclaimer that all material is the direct product of field studies and machine processing, that mechanical errors may not have been corrected, that the raw data is not to be considered the results of a professional traffic engineer’s evaluation of validity or reliability, and that it is the user’s responsibility to determine the accuracy and usability of the data prior to using or release to others

Surface Street Machine Counts

Much of the traffic count data collected under Element III of the CCTA Decennial Model Update is stored in 15-minute increments. For validation of the model, the data must rep-resent an average year 2000 weekday for the 24-hour daily, a.m. and p.m. peak-period, and a.m. and p.m. peak-hour time periods.



This required a data reduction task that included the following:

• Review and evaluation of the data for outliers, etc.;

• Filtering for Tuesdays, Wednesdays, and Thursdays, if necessary;

• Interpolation of data, if necessary;

• Determination of the peak period and peak hours; and

• Averaging of counts.

There were locations where the intersection traffic count data were used to fill in missing directional link volumes. Where this was the case, the appropriate intersection count data was also converted to peak-period and peak-hour approach and departure volumes.

A major difficulty, which resulted in significantly increased labor, was the frequent dis-crepancies between the reported link count locations in the field and the location of the screenline it was intended to represent. In some cases, the screenline was relocated to match the location of the count. In other cases, the count was adjusted to represent the likely count at the true location of the screenline.

Comparison of Intersection and Link Counts

Where both the intersection count data (approach/departure volumes) and the directional link count data (ramp or screenline volumes) were available, they were compared to assess the variability in data sources and counts. As can be seen in the summaries of the resultant percent, root mean square errors (RMSE) are shown in Tables 2.4 and 2.5; the percentage RMSE varied from 30 to 44 percent of the mean count.

Table 2.4 RMSE Comparison of Caltrans Ramp Estimates (Caltrans Census) and Intersection Approach Volumes (Pang Ho, Cities)

Peak Hour

Caltrans Ramp

Estimates

Intersection Approach Volume Ratio Difference

Sum of Diff.

Squared AM 72,197 76,432 1.06 4,235 8,427,589 122 # of Counts 44% % RMSE

PM 79,434 81,429 1.03 1,995 7,968,817 122 # of Counts 39% % RMSE



Table 2.5 RMSE Comparison of Screenline Counts (MultiTrans, MTCI, Caltrans Census) Versus Intersection Approach Volumes (Pang Ho, Cities)

Bi-Directional

Peak Hour

Screenline Count

Volume from Intersection

Count Ratio Difference Sum of Diff.

Squared

AM 114,358 116,997 1.02 2,639 15,541,731 77 # of Counts 30% % RMSE

PM 129,045 134,986 1.05 5,941 20,323,269 77 # of Counts 31% % RMSE

2.7 Intersection Counts

Intersection traffic count data are stored in data tables with fields for location, date, day, 15-minute interval begin time, N/S street, E/W street, and each of the 12 individual turn movements. These data tables are uniquely related to the TransCAD “Master” geographic file node layer using an intersection number code. The user can then use TransCAD (or other database management software) to select and process records based on specific fil-ters (e.g., by location, day = tu-th, begin time = 7:00-8:00); and data statistics can be com-piled. TransCAD has a built-in dataview statistics function that creates a .DBF summary of the count, sum, minimum, maximum, mean, and standard deviation for each field in the dataview.

While ideally the traffic counts would all have been conducted in 2000 for validation of the Decennial Model Update year 2000 model run, the traffic count database is actually a collection of data from many sources and counts taken between the years 1998 and 2002. Based on a review of demographic trends in the last four years, it was decided not to growth factor any of the counts to the year 2000 model validation year.



Pang Ho Associates collected turn movement counts for about 360 intersections in Contra Costa County. Translating these into validation counts required the following two steps:

1. Aggregate turn movement counts into approach and departure volumes by direction; and

2. Map approach and departure directions at each intersection into links on the CCTA network.

Turn movements are recorded as northbound left, northbound through, northbound right, southbound left, etc. Aggregating these into approach volumes is fairly straightforward, as shown in Table 2.6.

Table 2.6 Mapping Turn Volumes into Approach and Departure Volumes

Departure Direction Approach direction North East South West

Approach Totals (Row Sums)

North – SBL SBT SBR East WBR – WBL WBT South NBT NBR – NBL West EBL EBT EBR – Departure totals (column sums)

NBL = Northbound left, etc.

Mapping approach directions to links is not always as straightforward, because link directions are based on the order in which they were originally digitized; and the direc-tions do not always fall on the cardinal points (north, east, south, and west). Because the nodes (and hence the link ends) are geocoded, it was possible to calculate the azimuths of the links. The process was as follows:

1. For each intersection, determine whether the azimuths for the majority of links at the intersection lie “close to” a cardinal direction (say, within 10 degrees), and whether all cardinal directions are covered. For these intersections, assign directions to links with a reasonably high degree of confidence.

2. For intersections that do not meet the above criteria, manually inspect the network and assign directions to the links.

The majority of intersections were dealt with by the first step, and only one-third or fewer required manual intervention.



Since TransCAD has the ability to code interchanges geographically, some intersections that the previous subarea models may have considered as a single node will actually be more than one node (see Figure 2.1 below). TransCAD identified six nodes on the north-south street crossing the freeway, while the field personnel will count this interchange as two intersections. Two of the six possible nodes will be selected for storage of the two intersection turn counts. Thus, in this situation, all turn movements are stored in the database at a single node, even when they represent turning movements occurring at several TransCAD nodes.

Figure 2.1 Sample Interchange in TransCAD

TransCAD Node

Signal

Freeway

2.8 Travel Speeds

Directional link travel speed data from the MTC model validation have been entered into the Validation Database for selected facilities. The data is stored in tables with fields for location, date, direction, distance, source, and time period. The speed data was obtained from the following sources:

• Table 6.6 of BAYCAST 90: A.M. Peak Period (6:30 to 8:30 a.m.) Traffic Assignments, Speed Validation for Key Segments, 1998;

• Table 6.7 of BAYCAST 90: A.M. Peak Period (6:30 to 8:30 a.m.) Traffic Assignment, Speed Validation by Detailed Segment, 1998;

• Table 6.8 of BAYCAST 90: Comparison of Predicted vs. Observed Arterial Speeds, 1998 in Hercules and Walnut Creek; and



• 2000 Monitoring of Compliance with Level-of-Service Standards: Freeway Segments Level of Service Results, including Average Speed from Plot (mph).

The Caltrans 1998 HICOMP Report found the following regularly congested freeways in Contra Costa (speeds regularly under 35 mph) during the a.m. peak period:

• I-80 from the Bay Bridge to the 24 and from I-880 to Martinez,

• Highway 4 between 160 and I-680,

• Western portion of the 24,

• I-680 south of Walnut Creek, and

• I-680 north of the 24.

For the p.m. peak period, the HICOMP report found the following regularly congested freeways:

• I-80 from the Bay Bridge to the 24 and from I-880 to Martinez,

• Highway 4 between 160 and I-680,

• 24 between I-680 and I-580,

• I-680 south of Walnut Creek, and

• I-680 north of the 24.

2.9 HOV Lane Usage Data

Caltrans counts of vehicle occupancy were available for three freeways in Contra Costa (see Table 2.7) for the following locations and times. The counts are contained in the Caltrans report, District 4 2000 HOV Report, December 2000.

The counts are shown below (see Tables 2.8 to 2.12).

Validate to Total Vehicles (HOV and SOV)

For remaining links, inspect HOV volumes for reasonableness (to check volumes and determine if HOV trips are being assigned correctly), as well as checking speeds to make sure they are equal to or faster than SOV lanes.



Table 2.7 Caltrans Vehicle Occupancy Data

State Route Location Direction Peak Period Peak Hour Route 4 Bailey Road Westbound 6:00-9:00 a.m. 6:00-7:00 a.m.

Interstate 80 East of Hilltop Drive Westbound 7:00-10:00 a.m. 7:00-8:00 a.m. 3:00-7:00 p.m. 5:00-6:00 p.m. Eastbound 7:00-10:00 a.m. 7:00-8:00 a.m. 3:00-7:00 p.m. 5:00-6:00 p.m.

Interstate 680 Stone Valley Road Southbound 6:00-9:00 a.m. 7:00-8:00 a.m. 3:00-6:00 p.m. 5:00-6:00 p.m. Northbound 6:00-9:00 a.m. 8:00-9:00 a.m. 3:00-6:00 p.m. 5:00-6:00 p.m.

Table 2.8 Route 4 at Bailey, Westbound AM Peak Hour

Vehicles HOV Lane Mixed-Flow Lanes

Motorcycles 8 13

Buses 0 4

Carpools (2+) 279 375

Vanpools 1 16

SOVs 21 3569

Total 309 3,977

Table 2.9 I-80, East of Hilltop, Westbound

AM Peak Hour PM Peak Hour

Vehicles HOV Lane Mixed-Flow

Lanes HOV Lane Mixed-Flow

Lanes Motorcycles 36 8 1 0 Buses 31 0 29 8 Carpools (3+)* 1,272 104 166 102 Vanpools 13 1 5 0 SOVs + HOV2 76 4,892 103 3,930 Total 1,428 5,005 304 4,040

*Includes eligible two-seater vehicles



Table 2.10 I-80, East of Hilltop, Eastbound




Lanes Motorcycles 2 0 47 11 Buses 24 9 24 2 Carpools (3+)* 91 152 810 97 Vanpools 9 4 37 4 SOVs + HOV2 41 3306 173 4899 Total 167 3,471 1,091 5,113

*Includes eligible two-seater vehicles.

Table 2.11 I-680, at Stone Valley Road, Southbound




Lanes Motorcycles 16 0 3 0 Buses 13 1 0 1 Carpools (2+) 765 139 363 0 Vanpools 28 0 6 0 SOVs 21 4,175 4 5,678 Total 843 4,315 376 5,679

Table 2.12 I-680, at Stone Valley Road, Northbound




Lanes Motorcycles 3 0 28 2 Buses 8 7 9 0 Carpools (2+) 467 149 1,308 127 Vanpools 10 1 7 7 SOVs 21 4,152 69 5,509 Total 509 4,309 1,421 5,645



3.0 Highway Networks

3.1 Introduction

This section describes the development and contents of the master highway network for the CCTA Decennial Model Update. Information on the content of the highway network can be found in the CCTA Travel Model User’s Guide.

The objective of the highway network task for the CCTA Model Update was to develop a comprehensive TransCAD highway and HOV geographic database file that has reason-able geographic shape (i.e., curvilinear streets, true interchange forms, etc.) and appropri-ate network attributes (lanes, functional classification, etc.) from which the following six network scenarios could be extracted:

1. Scenario #1 – Year 2000, existing conditions;

2. Scenario #2 – Regional Transportation Improvement Program (2000 RTIP) for year 2010;

3. Scenario #3 – 2000 RTIP, plus seven-year CMP Capital Improvement Program (CIP) (CCTA 2001 CMP Update), also for the year 2010;

4. Scenario #4 – Regional Transportation Plan (RTP) Track 1 (2001 RTP Update) for year 2020;

5. Scenario #5 – RTP Track #1 (same as Scenario #4), but for year 2025; and

6. Scenario #6 – RTP blueprint, plus selected projects from CCTA’s Comprehensive Transportation Project List (CTPL) for year 2025.

The 2000 Model Update highway network within the CCTA area (defined as those links and nodes within Contra Costa County and the Tri-Valley area) includes all of the road-ways in the detailed networks of the subarea models from the 1990 CCTA model set. All links within this area have been geographically shaped to overlay the CCTA Land Use Information System (LUIS) map, which, in turn, is based upon the Contra Costa County road centerline file.

Network links and attributes outside of the CCTA area are consistent with currently avail-able MTC networks. These networks were transformed to a coordinate system consistent with the LUIS; however, no additional geographic shaping was performed for links out-side of the CCTA study area.



3.2 Development of the Master Highway Network

Travel demand modelers have typically developed the base year highway network, and then copied it and edited the copy to create future networks. This meant that work on future highway networks could not start until all network edits associated with model calibration and validation had been completed. Also, as part of the forecasting process, the modeler invariably discovers errors in the future network that should also be fixed in the base network. Unfortunately, it is difficult to convey these corrections reliably to every copy of the network, so often forecasted volumes reflect differences in network copies, as well as actual highway improvements.

The “master” highway network concept overcomes these problems by placing all existing and future highway improvements in one master file. For example, a single link will have several fields indicating the number of lanes, each field appropriate to a particular future scenario. A future link may have zero lanes in the base year and non-zero lanes in one or more of the future years. The future networks are then generated from the master file, ensuring that all network edits are carried consistently through to all applicable future highway networks.

One difficulty working with master networks is visually checking the integrity of each future network that might be generated from the master network. Since the master net-work includes all existing and future links, it is hard to visually spot missing links. How-ever, this can be overcome by generating each future scenario network and proofing it.

The master highway network was created by combining the MTC years 2000, 2010, and 2025 highway networks outside of the CCTA area (defined as Contra Costa County, plus the eastern half of Alameda County that lies in the Tri-Valley) with years 2000 and 2010 highway networks from each of the CCTA subarea EMME/2 models.

Network Development Inside the CCTA Study Area

The portion of the master network that falls in the CCTA study area was developed through the following steps:

1. Select the base for network development and convert the base network to Contra Costa County Centerline File coordinate system. The Contra Costa County Centerline File coordinate system was selected as the geographic basis for the CCTA LUIS and, therefore, for the CCTA model highway network. It was determined that it would be easier to shape the existing MTC highway network and the CCTA EMME/2 Subarea model networks to the County Centerline File than it would be to add high-way network attributes to the County Centerline File. A coordinate transformation (basically a linear regression methodology) was then performed to convert the CCTA subarea networks, the MTC networks, and the Contra Costa and Alameda County Census TIGER files to the coordinate system and projection of the Contra Costa



County centerline file and the 2000 Model Update TAZ layer (State Plane NAD83, California Zone 3).

2. Code new centroids and centroid connectors in support of the refined Traffic Analysis Zone (TAZ) system inside of the CCTA area. The 2000 Model Update TAZ map and available TransCAD features were used to add centroid and centroid con-nectors for the refined zone system.

3. Relate the master highway network to the traffic count database. TransCAD’s azi-muth function was used to get compass bearings of roadway facilities. Link attribute data were added to relate the line layer with the screenline and transit route validation counts. Node attribute data were added to relate the node layer with the Validation Database intersection counts.

4. Identify highway network scenario improvements and code in master network. The Authority’s CTPL database was queried to obtain the project lists for the various future network scenarios. The project lists were then manually translated into changes in highway link attributes, which were then coded into the master highway network and identified by future network scenario.

5. Review highway network scenario improvements. The base year and future high-way networks were generated and plotted from the master network for review by the Transportation Model Working Group (TMWG).

Network Development Outside the CCTA Study Area

The MTC model networks (including geography and link attributes) were used outside of the CCTA Study Area. It was a fairly simple process to identify those links outside the CCTA area in each of the available MTC networks, and then to merge each of these indi-vidually with the internal CCTA area geographic file. However, since MTC’s coding con-ventions change from network to network, node numbers are reused, and a large number of improvements are coded outside of the CCTA area; it would have been a significant level of effort to merge the MTC networks together with each other. Therefore, the fol-lowing process was adopted:

1. Select the most comprehensive and best-defined MTC network scenario to which appropriate improvement projects will be added or subtracted to create other alterna-tive scenarios.

2. Track and address any issues of inconsistent and duplicative use of node numbering. Create a correspondence table of node renumbering between scenarios, so that transit networks can be updated after MTC networks are merged.

3. Renumber nodes as needed in each MTC network scenario.

4. Create attribute fields for all network scenarios and merge the MTC networks and attribute data into a single file.



5. Track and address any issues of inconsistent coding (e.g., HOV lanes).

6. Select merge nodes along the Contra Costa County and Tri-Valley boundary, and identify those links and nodes inside and outside the boundary. Add an appropriate internal/external link attribute to all links.

• Select the links and nodes outside the CCTA area, merge these with the CCTA area geographic file as developed above, and ensure connectivity at each merge node for each network scenario.



4.0 Transit Networks

4.1 Introduction

This section describes the transit network and the process used to develop it. In the CCTA model implementation of TransCAD, the transit network consists of two layers overlaid over the master highway network node and line layers: 1) route layer and 2) stop layer.

The transit access links (walk and drive) are stored in the master highway network:

• Walk access links are Mode 1,

• Drive access links are Mode 2,

• Transfer links are Mode 3 in the master highway network,

• Mode 4 links are auto access walk funnel links, and

• Mode 5 links are walk access walk funnel links.

A funnel link is the dummy link connecting the park-and-ride lot to the station or the walk point to the station.

4.2 Development of the Transit Master Networks

The master transit network was developed from the converted MTC years 2000 and 2025 transit networks. These networks had been converted by Caliper from TP+ to TransCAD.

Development of Route and Stop Layers

The master transit network was obtained from Caliper in September 2002. The following changes were made to the route system:

• The transit lines were examined in both the 1998 and the 2025 MTC models and for those lines which were exactly the same in both years, only one route with attributes for both the years was included. The other duplicate route was deleted.

• It was noticed that splitting some of the links in the underlying master highway net-work caused the transit routes to not reload properly. To avoid such errors, all the



routes inside the CCTA study area (Contra Costa County, plus the Tri-Valley portion of Alameda County) were first deleted. The highway layer was then edited to contain the Phase 2 zone centroids. The transit lines inside the study area were then added in appropriately on newly added local streets using maps obtained from the web sites of the transit operators.

• For the transit routes inside the CCTA study area, stops were added approximately near centroid connectors.

• All other route attributes (e.g., Headway, Headway2, Mode, Owner) for the transit lines inside the study area were brought in exactly the same as from the route systems in the MTC 1998 and 2025 models.

• Rail routes (BART and Amtrak) passing through the study area were shaped to look more realistic.

There are a total of 1,704 routes in the master route layer. The stop layer contains 6,047 stops and 154 park-and-ride lots for both 2000 and 2025. The transit network stop layer is presented in detail in Appendix E. A summary of the transit routes by location is pre-sented in Table 4.1.

Table 4.1 Number of Transit Routes by Location

Inside CCTA

Through CCTA

Outside CCTA Total

2000 138 60 896 1,094

2020/2025 138 68 907 1,113

Development of Transit Access and Transfer Links

The following steps went into the development of transit access and transfer links in the master network:

1. Walk Access Links – Outside the study area, all the walk access links in the MTC 1998 and MTC 2025 models were combined together in the form of a “master walk access network” and brought into the CCTA master network, along with all the walk access attributes. Inside the study area, walk access links were created from all centroids to all stops/walk aux rail nodes within 0.5 mile of them. For those zones that did not have any stop within 0.5 mile, walk access links were created to the nearest stop to them to a maximum distance of eight miles. Walking speed on the links is assumed as three miles per hour.



2. Drive Access Links – Drive access links inside and outside the study area were pro-vided for all the zones that had drive access in the MTC 1998 and MTC 2025 models. For example, inside the study area, if a zone with a drive access link was split into four new zones, all these four zones now have drive access links to the parking lot. The “time” field on the drive access links inside the study area was filled by skimming the network on the congested time.

3. Transfer Links – Outside the study area, all the transfer links were brought in exactly as they were. Inside the study area, transfer links were created from all stops to all stops within 0.5 mile to each other.

4. Auto Access Walk “funnel” Links – These are the links from a park-and-ride lot to the rail station. Funnel links inside and outside the study area were brought in exactly as they were in the MTC 1998 and MTC 2025 models.

5. Walk Access Walk “funnel” Links – These are the links from a walk aux rail node to the rail station. Funnel links inside and outside the study area were brought in exactly as they were in the MTC 1998 and MTC 2025 models.

The fields in the non-motorized links are presented in Table 4.2, and the modes used in the transit network are presented in Table 4.3.

Table 4.2 Transit Non-Motorized Access Links

Field Name Field Description

Mode Transit mode as described in the mode table

Mdistance Length on the access link (skimmed distance for drive and transfer links) (miles)

MAB_Time Time taken to walk/drive (3 miles per hour walking speed for walk links and skimmed congested time for drive access links) (minutes)

MBA_Time Time in the reverse direction (minutes)

AB_NT_Time MAB_Time*0.01 (hundredths of minutes)

BA_NT_Time MBA_Time*0.01 (hundredths of minutes)



Table 4.3 Mode Table

Mode Name Mode

ID Type Fare Type Headway Fare

Maximum Wait

Maximum Access

Highway 0 H 0 WACC 1 W 0 DACC 2 H 0 Transfer 3 W 0 DACC Walk Funnel 4 W 0 WACC Walk Funnel 5 W 0 BEST 10 T 0.10 15.00 5 Broadway_Shuttle 11 T 0.10 15.00 5 EMERY 12 T 0.01 15.00 5 STANFORD_Shuttle 13 T 0.01 15.00 5 CALTRAIN_Shuttle 14 T 0.01 15.00 5 SCVTA_Shuttle 15 T 0.01 15.00 5 Munimetro 20 T 0.80 15.00 20 Muni_CableCars 21 T 1.60 15.00 20 Muni_Richmond_Dist 22 T 0.80 15.00 20 Muni_Mission_Dist 23 T 0.80 15.00 20 MUNI_other 24 T 0.80 15.00 20 SAM_EXP 26 T 0.90 15.00 10 SAM_Coastal 27 T 0.90 15.00 10 SAM_Bayside 28 T 0.90 15.00 10 SAM_90 29 T 0.90 15.00 10 SCVTA_LRT 31 T 0.90 15.00 6 SCVTA_Local 32 T 0.90 15.00 6 SCVTA_Limited 33 T 0.90 15.00 6 SCVTA_EXP 34 T 1.40 15.00 6 DBX 35 T 0.80 15.00 6 AC_TransBay 37 T 1.00 15.00 6 AC_NORTH 38 T 1.00 15.00 6 AC_SOUTH 39 T 1.00 15.00 6 AC_EB_exp 40 T 1.00 15.00 6 WHEELS_Dublin 42 T 0.60 15.00 5 WHEELS_Pleasanton 43 T 0.60 15.00 5 WHEELS_Livermore 44 T 0.60 15.00 5 WHEELS_Intercity 45 T 0.60 15.00 5 UNIONCTY 47 T 0.60 15.00 5 AirBart 49 T 1.60 15.00 5 CCCTA_local 51 T 0.80 15.00 5 CCCTA_EXP 52 T 1.00 15.00 5



Table 4.3 Mode Table (continued)

Mode Name Mode

ID Type Fare Type Headway Fare

Maximum Wait

Maximum Access

TRIDELTA 54 T 0.60 15.00 5 WESTCAT 56 T 0.60 15.00 5 WESTCAT(ML30Z) 57 T 0.80 15.00 5 VALLEJO_Local 59 T 0.80 15.00 5 VALLEJO_Link 60 T 0.80 15.00 5 FAIRFIELD 62 T 0.60 15.00 5 FAIRFIELD(20ECL/WCL) 63 T 0.60 15.00 5 FAIRFIELD(40BEW/BEE) 64 T 0.60 15.00 5 Vacaville 66 T 0.80 15.00 5 Benicia 68 T 0.60 15.00 5 NVT 70 T 0.60 15.00 5 VINE 71 T 0.60 15.00 5 SONOMA_Local 73 T 0.60 15.00 5 SONOMA_Intercity 74 T 0.60 15.00 5 SANTA_ROSA 76 T 0.40 15.00 5 Petaluma 78 T 0.20 15.00 5 Ggt_SF 80 T 1.40 15.00 10 Ggt_SF_FerryFeeder 81 T 1.00 15.00 10 Ggt_FerryFeeder 82 T 1.00 15.00 10 Ggt_Marin_Sonoma 83 T 1.00 15.00 10 Ggt_Richmond 84 T 1.40 15.00 10 FERRY 90 T 5.00 1 FERRY(LARKS/LARKN) 91 T 10.00 1 FERRY(SSLTO/SSFW) 92 T 10.00 1 FERRY(TIBFB/TIBFW) 93 T 5.00 1 FERRY(VALFB) 94 T 10.00 1 BART 100 T 15.00 2 CALTRAIN 101 T 15.00 2 AMTRAK 102 T 15.00 2 AMTRAK(SJQ) 103 T 15.00 2



Development of Future Year Networks

The future year transit networks were developed based on the following assumptions:

• Ace Train exists in 2000.

• All transit lines in the 2010, 2020, and 2025 scenarios inside the study area have the same route, frequency, and run time as the 2000 routes, except for the following:

− Track 1 – New train station at Hercules on the Capitol Corridor line;

− Track 1 – Increased frequency of Amtrak Capitol Corridor line between Oakland and Sacramento from 99 minutes to 60 minutes; and

− Track 2 – East County Commuter rail on existing tracks from Brentwood to Tracy.

• All routes passing through the study area from outside are kept the same as in the MTC 1998 and MTC 2025 models.

• The following edits were not clear and need to be clarified:

− Track 1 – I-80: Expand Express Bus Service, Solano County to Richmond, Berkeley, and Oakland. Which transit line and increase frequency by how much?

− Track 1 – I-80: Purchase new Express Buses to provide increased express bus ser-vice within the I-80 corridor. Contra Costa portion only?

− Track 2 – East County Commuter rail on existing tracks from Brentwood to Tracy. Where to add stations and park-and-ride lots?

− Track 2 – Further expand service on Capitol Corridor consistent with Capitol Corridor Joint Powers Board Business Plan (?)



5.0 Zonal Data

The CCTA Travel Model uses four zonal data files from the MTC model as inputs, as shown in Table 5.1. To ease maintenance, updating, and inspection of the data, all four files were merged into a single, unified database for each analysis year. Prior to running the model, a pre-processing step creates each of the four ASCII files that are used as input into the numerous FORTRAN programs in the model.

Table 5.1 MTC Zone Data Files

Data File Description

ZMAST*.ASC Master zonal data file

ZHBSK*.ASC School enrollment data file

ZAGE*.ASC Population for age group categories

AZLOS*.DAT Auto zonal level of service file

*Indicates the analysis year.

These ASCII files include certain specific zonal characteristics that currently do not exist in the CCTA zonal database. The variables from each of the files identified in Table 5.1 are described in a data dictionary and presented in Appendix F. The data dictionary is com-prised of the field names, field type, description, and data source for Phases 1 and 2 and its corresponding MTC ASCII file. Many of the procedures developed to support the Phase 1 analysis were replaced with more local data in Phase 2. For future model updates, these data may be further reviewed and updated on a local jurisdiction basis. The remainder of this section describes procedures to derive the zonal data that was unavail-able at a local level.

5.1 CCTA Master Zonal Database

The CCTA master zonal database (CCTA_2000_Phase1.DBD) includes all the base eco-nomic conditions for the region. This file includes the following:

• Population,

• Number of households,



• Employment categories,

• Income classifications,

• School/college enrollment,

• Age categories, and

• Auto zonal level of service variables.

Appendix F shows all the data items that are needed for the CCTA master zonal database. The items highlighted in gray were not included in the Phase 1 LUIS database, and were generated from other data sources.

In order to implement Phase 1, testing of the new CCTA countywide model, the missing data was generated using the existing MTC zonal data. Because all CCTA zones nest within the MTC zones, it was possible to derive the values for each CCTA zone using the MTC data as a reference. For Phase 2, many of the data items were collected from local sources. If the data was from local sources, the derived values by analysis year serve as default values. Below describes how each value was derived using the MTC data as a reference.

• Total population – Computed total population to be the same ratio in each CCTA zone as household population in Phase 1; replaced by total population from the 2000 Census in Phase 2 as part of the LUIS update.

• Households in single and multi-family dwelling units – Computed the percentage of households in single and multi-family dwelling units to total households in each MTC zone and applied this percentage to total households in each CCTA zone Phase 1; replaced by households by type from the 2000 Census in Phase 2 as part of the LUIS update.

• Income quartiles – Used the MTC distribution and applied this to each CCTA zone and adjusted as necessary to the average household income in Phase 1; replaced by households by income quartile from the 2000 Census in Phase 2 as part of the LUIS update.

• Share of total population over age 62 – Computed the percentage of population over age 62 in each MTC zone and applied this percentage to total population in each CCTA in Phase 1; replaced by the share of population over age 62 from the 2000 Census in Phase 2 as part of the LUIS update.

• Agricultural, manufacturing, and wholesale employment categories – Computed the sum of agricultural, manufacturing, wholesale, and other employment to derive total other employment for each MTC zone; computed the percentage of each category to the total other employment in MTC and applied this percentage to other employment category in each CCTA zone in Phase 1; replaced by employment by type updated in the LUIS in Phase 2.



• School, college enrollment – In Phase 1, there was no valid method to derive these using the existing MTC or CCTA zonal data. In Phase 2, we compiled enrollment data for grade/middle school, high school, and college from the California Department of Education (CDE) web site. As part of this process, we requested and received a geographic information systems (GIS) file of all school locations from Caltrans.

• Calculation of age group profiles – Computed the percentage of the age group to total population for each MTC zone and applied this percentage to the total population for each corresponding CCTA zone in Phase 1; replaced this with age group profiles from the 2000 Census as part of the LUIS update.

• Auto zonal level of service variables – These variables have information pertaining to parking costs and terminal times for each zone in the CCTA model. The majority of these values are calculated as part of the modeling process; however, parking cost in the peak and off-peak time periods is a required input of this dataset. For Phase 1, the MTC data was used as the default values.



6.0 CCTA Model Development

The core modeling components of the CCTA Travel Model have been derived to match the procedures and parameters in the MTC regional travel model. During model calibra-tion, minor adjustments were made to the calibration files within Contra Costa County. A direct comparison of the model outputs for each model component is provided in a sepa-rate report, MTC Model Consistency. A series of flow charts to identify inputs, programs, and outputs throughout the modeling process is provided in the User’s Guide report. The remainder of this section provides an overview of the MTC BAYCAST derived from the technical summary report produced by MTC.1 Where applicable, calibration adjustments to the models are also documented. Additional references for the MTC BAYCAST model are provided in Appendix H.

6.1 MTC Model System Overview

BAYCAST is designed as an advanced state-of-the-practice trip-based travel forecasting system. It is designed to be tractable, sophisticated, and user-friendly.

As opposed to the typical “four-step” model, the BAYCAST modeling system includes the standard four steps of trip generation, trip distribution, mode choice, and trip assignment, as well as three extra main models: 1) workers in household, 2) auto ownership choice, and 3) time-of-day choice models (see Figure 6.1).

The following five principal trip purposes are defined for intraregional personal travel:

1. Home-based work (HBW),

2. Home-based shop/other (HBSH),

3. Home-based social/recreation (HBSR),

4. Home-based school (HBSK), and

5. Non-Home-based (NHB).

1 Travel Demand Models for the San Francisco Bay Area (BAYCAST-90) Technical Summary, Charles L.

Purvis, Planning Section, MTC, June 1997.



Figure 6.1 Bay Area Travel Demand Model Forecasting System (BAYCAST)

Workers in Household Choice(WHH = 0, 1, 2+)

Vehicles in Household Choice(VHH = 0, 1, 2+)

TripGeneration

TripDistribution

ModeChoice

Time-of-Day Choice(Peak/Off-Peak)

TripAssignment



Home-based school trips are further broken down into the following:

• Home-based school: Grade school (HBGS),

• Home-based school: High school (HBHS), and

• Home-based school: College (HBCol).

Trip-Based Versus Activity-Based Travel Demand Models

Stratifying trips by trip purpose is typical of trip-based travel demand model systems. This allows the analyst to subdivide the overall travel demand market into manageable and generally well-understood submarkets: the journey-to-work, the journey-to-school, other home-based travel, and non-home-based travel. Some of these trip purposes are restricted to certain groups (i.e., only workers take journeys-to-work and only students take journeys-to-school). Other trip purposes are considered household or family-based chores or activities, such as home-based shopping and home-based social/recreation trips. Non-home-based trips are taken by anyone.

The downside of trip-based models is the independence between these trip purposes in a typical zone-based (aggregate) forecasting system, such as BAYCAST. A good example of this problem is that a non-home-based trip, say, from work-to-lunch, is not linked with and has no “knowledge” or “memory” of the previous trip, say, from home-to-work, that the trip maker took (in terms of mode used, vehicle used and available, time of travel con-straints, etc.).

Alternatives to the standard trip-based modeling system are under development in other metropolitan areas (e.g., Stockholm; Portland, Oregon; Honolulu, Hawaii) to fully elimi-nate non-home-based trips and replace them with “half-tour” (e.g., home-to-work and work-to-home trip chains) or “full-tour” (home-to-home round trips) models.

(This short discussion on trip-based models with non-home-based trips and activity-based models that exclude non-home-based trips is intended as background to future model development activities planned at MTC.)

Market Segmentation

Market segmentation is a critical feature of advanced trip-based travel demand model systems. Market segmentation is a compromise between a fully disaggregate modeling system and a fully aggregate modeling system. In a fully disaggregate modeling system, the disaggregate demand models are applied at the disaggregate, individual level. Results are only summed at the end of the process for reporting purposes. In a fully aggregate modeling system, all persons and households within a travel analysis zone are assumed to be “average” with identical characteristics in terms of average household income, average household size, average vehicles per household, average workers in the household, average students in the household, etc.



Market segmentation is useful in adapting disaggregate demand models for use in an aggregate modeling system, such as BAYCAST. Models are applied to subgroups within travel analysis zones as opposed to no subgroups. This “no subgrouping option” is also known as “naive segmentation.”

Market segmentation is particularly useful in analyzing market captivity. For example, households without automobiles are highly unlikely to drive alone to work or to drive to a transit station. Another example is that households without workers are not going to take trips from home-to-work.

The market segments used in the BAYCAST model system application include the following:

• Household by workers in the household (WHH = 0, 1, 2+).

• Households by autos available in the household (AO = 0, 1, 2+).

• Households by household income quartile (Income = less than $25,000; $25,000 to $45,000; $45,000 to $75,000; and more than $75,000).

These three market segmentations are not used for all trip purposes and for all models (trip generation, trip distribution, and mode choice).

• Home-based work trips are market segmented by household income in the trip generation, trip attraction, trip distribution, and mode choice models; and also by auto ownership level in the mode choice model application. Non-working households are ineligible to take home-based work trips.

• Home-based shop trips are stratified by workers in household level and auto ownership level in the trip generation model; and by auto ownership level in the mode choice model. No segmentation is used in either trip attraction or trip distribution models for home-based shop.

• Home-based social/recreation trips are segmented by auto ownership level in the trip generation and mode choice models. No segmentation is used in either the home-based social/recreation trip attraction or trip distribution models.

• Home-based school trips are not market segmented by auto ownership level or workers in household level or income level in any of the home-based school models (generation, attraction, distribution, mode choice). The basic market segmentation for school trips is by level of school (grade school, high school, college) based on age of student (five to 13, 14 to 17, and 18 to 24).

• Non-home-based trips are also not market segmented. Early work on separating non-home-based trips by whether or not the non-home-based trip was a work-based trip concluded that the work-based versus non-work-based stratification was not an improvement.



6.2 Workers and Vehicles in Household Nested Choice Model

The MTC workers and vehicles in household model (WHHAO) is a nested logit choice model applied at the zone-of-residence level. The input to the WHHAO model application are number of households stratified by household income quartile level. Variables in this choice model include mean household income, mean household size, the share of households residing in multi-family dwelling units, the share of persons age 62 or older, and gross population density. Coefficients for the final nested choice model, Model #9W, are shown in Table 6.1. Detailed definition of variables in this and other models are included in Table 6.2.

Data on mean household income, mean household size, and gross population density is available from the Association of Bay Area Governments (ABAG) forecasts. Future year data on share of multi-family units and share of persons age 62 or more will be derived by MTC staff from the 1990 decennial Census data and ABAG county-level age forecasts.

The nested structure for the WHHAO model is shown in Figure 6.2. The upper level nest of this model splits households into households by workers in household level (0, 1, 2+ workers per household). The lower nest further splits these households by auto ownership level (0, 1, 2+ vehicles per household).

The output from this WHHAO model is the number of households by household income quartile (4), by workers in household level (3), by auto ownership level (3), or 36 different market segments per travel analysis zone.

A detailed example of the calculation of the “logsum” variables used in this WHHAO model is included in a memorandum by C. Purvis, dated October 3, 1995, which was included in Technical Memorandum Compilation, Volume III. The “logsum” is defined as the natural logarithm of the sum of the exponentiated utilities within the particular nest of interest.

Description of adjustments (calibration) to the utility constants and adjustments to the coefficients by market segment are included in the separate technical summary on calibration and aggregate validation of the BAYCAST model system. These calibration adjustments were not revised during the calibration of the CCTA model update.



Table 6.1 WHHAO Multinomial and Nested Choice Models – Model #9W

Model #9W (MNL)

Model #9W (Nested)

WH

H=0

A

O=0

WH

H=0

A

O=1

WH

H=0

A

O=2

+

WH

H=1

A

O=0

WH

H=1

A

O=1

WH

H=1

A

O=2

+

WH

H=2

+ A

O=0

WH

H=2

+ A

O=1

WH

H=2

+ A

O=2

+

Variable Beta t-stat Beta t-stat constant 1 0.9349 (1.4) 1.615 (1.4)

constant 2 2.33 (3.8) 3.084 (2.6) constant 3 0.6962 (1.1) 1.679 (1.4) constant 4 0.2607 (0.4) 1.586 (1.2) constant 5 1.719 (2.9) 3.284 (2.5) constant 6 -0.4014 (0.6) 1.237 (0.9) constant 7 -1.851 (2.0) -2.941 (2.8) constant 8 -0.3117 (0.5) -0.7834 (1.1) Income-Leg 1 4.633E-02 (2.4) 3.956E-02 (2.1) Income-Leg 1 0.1000 (4.2) 0.0888 (3.6) Income-Leg 1 0.1093 (4.3) 0.2853 (2.4) Income-Leg 1 0.1671 (8.1) 0.3433 (3.0) Income-Leg 1 0.2139 (8.2) 0.3907 (3.3) Income-Leg 1 0.1267 (3.0) 0.9325 (1.7) Income-Leg 1 0.1729 (5.9) 0.9719 (1.8) Income-Leg 1 0.2418 (8.8) 1.0320 (1.9) Income-Leg 2 1.213E-02 (0.8) 9.989E-03 (0.6) Income-Leg 2 2.564E-02 (1.6) 2.268E-02 (1.4) Income-Leg 2 1.122E-02 (0.7) 4.776E-02 (1.4) Income-Leg 2 2.310E-02 (1.6) 5.624E-02 (1.7) Income-Leg 2 4.604E-02 (3.1) 7.682E-02 (2.4) Income-Leg 2 3.035E-02 (1.5) 0.2699 (1.6) Income-Leg 2 4.740E-02 (3.1) 0.2866 (1.7) Income-Leg 2 6.685E-02 (4.5) 0.3048 (1.8) HH Size 0.4119 (5.4) 0.3311 (3.8) HH Size 0.5893 (9.7) 0.5986 (8.9) HH Size 0.4688 (7.0) 1.3790 (2.4) MFDU 0.5272 (2.8) 0.5662 (3.0)

MFDU -0.9346 (8.0) -1.0700 (8.8) SHPOP 62+ 3.4230 (16.8) 4.5390 (2.9)

SHPOP 62+ -2.5250 (7.3) -12.1900 (1.7) GPOPD-Leg 1 -0.03546 (1.0) -0.05354 (1.6) GPOPD-Leg 1 -0.05174 (1.4) -0.07401 (2.2) GPOPD-Leg 2 -0.04837 (3.5) -0.04987 (3.6) GPOPD-Leg 2 -0.10180 (6.4) -0.11170 (6.9) GPOPD-Leg 3 -2.378E-02 (4.0) -2.506E-02 (4.1) GPOPD-Leg 3 -2.380E-02 (2.7) -2.724E-02 (2.9) Theta - NWHH 0.7451 (3.0)

Theta - SWHH 0.4477 (2.7) Theta - MWHH 0.1968 (1.8) Log

Likelihood -2806.2 -2780.5

Source: BAYCAST Travel Demand Model Based on 1990 Bay Area Household Travel Survey, Single Day Sample.



Table 6.2 Detailed Definition of Variables used in BAYCAST Travel Demand Models (in Alphabetical Sort Order)

Variable Name Model(s) Definition AreaDeni NHBMC Area Density, zone of production, ((TOTPOP + 2.5 * TOTEMP)/(CIACRE

+ RESACRE)) Auto Distance HBWDT Door-to-Door Drive Alone Peak Period Distance, in miles Auto Distance^2

HBWDT Door-to-Door Drive Alone Peak Period Distance, in miles, squared

Berkeleyi HBCOLM Berkeley zones, zone of production (zones=718-722, 725-738, 741-747) Berkeleyj Multiple Berkeley zones, zone of attraction (zones=718-722, 725-738, 741-747) Bike Time HBCOLM Bicycle travel time, in minutes Bridge Dummy HBWDT Bridge crossing dummy variable (if Drive Alone Toll > 0.0, then

Dummy=1) CIACRE Multiple Commercial + Industrial Acres (ABAG Land Use) Constant Multiple Modal or Utility intercept. Corej Multiple Core zone of work (see Table A.2 for definition of CORE) Cost Multiple Trip Cost in 1990 cents (per person) CTFT HBWDT Congested Time less Free-Flow Time, in minutes GPOPD-Leg 1 WHHAO Gross Population Density (TOTPOP/TOTACRE), MIN(10.0,GPOPD) GPOPD-Leg 2 WHHAO Gross Population Density (TOTPOP/TOTACRE), MAX(0,MIN((GPOPD-

10.0),20.0)) GPOPD-Leg 3 WHHAO Gross Population Density (TOTPOP/TOTACRE), MAX(GPOPD-30.0) HH Income HBWDT Household Income in 1989 constant dollars (not divided by 1000) HH Size WHHAO Persons per Household (same as Pers/HH) HHINC Multiple Income in 1000s of 1989 dollars (same as Income) Income HBSRMC Household Income in 1000s of 1990 dollars. Income-Leg 1 Multiple Income in 1989 dollars. MIN(Income,25000) Income-Leg 2 Multiple Income in 1989 dollars. MAX(0,MIN(Income-25000),50000)) IVTT Multiple In-Vehicle Travel Time (IVTT) Ln Net ResDensI

HBCOLM Natural Log of Net Residential Density, zone of residence, Ln(TOTHH/RESACRE)

LnAreaDeni HBSHMC Natural Log of Area Density, Zone of Residence (see Table A.2) LnCost Multiple Natural Log of Trip Cost in 1990 cents (per person) LnEmpDi HBWMC Natural Log of Gross Employment Density, Zone of Residence,

Ln(TOTEMP/TOTACRE) LnEmpDj HBWMC Natural Log of Gross Employment Density, Zone of Work,

Ln(TOTEMP/TOTACRE) LnIncome HBSRMC Natural Log of Household Income in 1000s of 1990 dollars. LnPHH HBSHMC Natural Log of Household Size MFDU WHHAO Multi-Family Dwelling Unit Dummy Variable Multi-Wrkr/ HH

HBWMC Multiple Number of Workers in Household Dummy Variable

Net ResDensI HBHSM Net Residential Density, Zone of Residence, (TOTHH/RESACRE) No VHH HBWMC No Vehicle in Household Dummy Variable OVTT Multiple Out-of-Vehicle Travel Time



Table 6.2 Detailed Definition of Variables used in BAYCAST Travel Demand Models (in Alphabetical Sort Order) (continued)

Variable Name Model(s) Definition PaloAltoi HBCOLM Palo Alto zones, zone of production (zones=234-238, 245-248, 254-258) PaloAltoj Multiple Palo Alto zones, zone of attraction (zones=234-238, 245-248, 254-258) Pers/HH Multiple Persons per Household (same as PHH) PHH Multiple Persons per Household (same as Pers/HH) PHH^3 HBGSM Persons per Household, cubed (polynomial transformation) RESACRE Multiple Residential Acres (ABAG Land Use) Retail Industry HBWDT Retail Industry worker Dummy Variable Rurali HBGSM Rural zone of residence (see AREATYPE, Table A.2) SFOBB WB HBWDT Bay Bridge Westbound Dummy Variable (based on origin & destination

zones) SHPOP62+ WHHAO Share of Population Age 62+ Single VHH HBWMC Single Vehicle in Household Dummy Variable SR Dummy HBWDT Shared Ride 2+ choice, dummy variable Stanfordi HBCOLM Stanford zones, zone of production (zones=244, 249-252) Stanfordj Multiple Stanford zones, zone of attraction (zones=244, 249-252) Theta(Group) Multiple Nesting, or Scaling Parameter for Group submode choice nested logit Theta(Motor) Multiple Nesting, or Scaling Parameter for Motorized submode choice nested logit Theta(Transit) HBWMC Nesting, or Scaling Parameter for Transit submode choice nested logit Time (Total) HBSHMC Total Travel Time, in minutes TOTACRE Multiple Total Acres (ABAG Land Use) Veh/HH Multiple Vehicles Available per Household (same as VHH) VHH Multiple Vehicles Available per Household (same as Veh/HH) Wait Multiple Transit wait time, in minutes Walk Multiple Walk time, in minutes (may include walk-only utility walk time) Wrkr/HH HBWMC Workers in Household (same as WHH) ZVHH Multiple Zero-Vehicle Household Dummy Variable (same as Zero VHH) ZWHH Multiple Zero-Worker Household Dummy Variable (same as Zero WHH)



Figure 6.2 WHHAO – Nested Choice Model #9W

WHH = 2+(MWHH)

WHH = 1(SWHH)

WHH = 0(NWHH)

AO = 0 AO = 1 AO = 2+ AO = 0 AO = 2+AO = 1 AO = 2+ AO = 0 AO = 1

6.3 Trip Generation Models

Trip generation models include both trip production and trip attraction models. Production models are based on trips made by households, workers, or students at the home end of home-based trips. Attraction models are based on trips made at the non-home end of home-based trips. Trips as defined in these trip generation models include non-motorized trips (bicycle, walk), as well as motorized modes (auto, transit).

For non-home-based trips, the same production/attraction terminology can be applied, though non-home-based generation models are essentially trip origin (production) and trip destination (attraction) models.

With the exception of the home-based school trip generation models, all of the new trip generation models are multiple regression in form. The home-based shop trip generation model, in particular, is a hybrid of a cross-classification model (stratified by workers in household level) and a multiple regression model.

Coefficients and definition of variables for all trip generation and attraction models are included in Table 6.3.

The independent variables in these multiple regression trip generation models are either trip rates (e.g., work trips per employed person, home-based shop attractions per retail+service+other job) or trips (e.g., total home-based social/recreation attractions, total non-home-based productions).



Table 6.3 Summary of BAYCAST Trip Generation Models

Home-Based Work Trip Generation Generation: HBWG/EMPRES = 1.0525 + 1.632E-02 * HHINC - 2.190E-04 * HHINC^2

+ 8.50E-07 * HHINC^3 Attraction: HBWA/TOTEMP = 0.7782 + 0.5661 * WRKR/JOB10 - 0.1289 * WRKR/JOB^2

+ 0.00873 * WRKR/JOB10^3 - 0.03928 * GEMPG10 + 0.3396 * CORE Where: HHINC = Household Income in Thousands of 1989 Constant Dollars WRKR/JOB10 = Worker/Job Ratio Decile code GEMPDG10 = Gross Employment Density, of Work, Decile Code CORE = Regional Core Zones Dummy

Market Segmentation:

Household Income Quartile (Generation and Attraction)

Home-Based Shop/Other Trip Generation Generation: HBSHG/ZWHH = 0.3141 + 0.4709 * PHH + 0.4034 * VHH + 0.02052 * HHINC

- 0.000131 * HHINC^2 HBSHG/SWHH = -0.4419 + 0.7299 * PHH + 0.2279 * VHH + 0.005123 * HHINC HBSHG/MWHH = -0.4288 + 0.5921 * PHH + 0.09071 * VHH + 0.009143 * HHINC

- 6.054E-5 * HHINC^2 Attraction HBSHA/RSOEMP = 0.1363 - 0.04506 * LogNEMPD + 1.6169 * TOTHHRT1

+ 0.7365 * TOTHHRT2 + 2.9835 * RETEMPRT Where: PHH = Average Household Size (Persons Per Household) VHH = Average Vehicles per Household HHINC = Household Income in Thousands of 1989 Constant Dollars LogNEMPD = Natural Log of RSOEMP/Commercial/Industrial Acres TOTHHRT1 = Ratio of Total Households to RSOEMP, where ratio is less than 1.0 TOTHHRT2 = Ratio of Total Households to RSOEMP, where ratio is greater than 1.0 RETEMPRT = Ratio of Retail to RSO Employment RSOEMP = Retail + Service + Other Employment


Workers in Household (3) by Household Income Quartile by Auto Ownership Level (3) (Generation Only)

Non-Home-Based Trip Generation Generation: NHBG = 0.798 * OTHEMP + 2.984 * RETEMP + 0.916 * SEREMP + 0.707 * TOTHH Attraction: NHBA = 0.636 * OTHEMP + 3.194 * RETEMP + 0.730 * SEREMP + 0.803 * TOTHH

Where: OTHEMP = Other Employment RETEMP = Retail Employment SEREMP = Service Employment TOTHH = Total Households


None



Table 6.3 Summary of BAYCAST Trip Generation Models (continued)

Home-Based Social/Recreation Trip Generation Generation: HBSRG/HH = 0.4102 + 0.1176 * PHH + 0.002849 * HHINC - 0.4632 * WHHRATE

+ 0.1487 * VHH - 0.08118 * ZVHH - 0.1049 * ZWHH Attraction: HBSRA = 0.8674 * RETEMP + 0.1606 * SEREMP + 0.5216 * TOTHH

Where: PHH = Average Household Size (Persons Per Household) VHH = Average Vehicles per Household HHINC = Household Income in Thousands of 1989 Constant Dollars WHHRate = Share of Persons in Household who Work (EMPRES/HHPOP) ZVHH = Zero Vehicle Household Dummy ZWHH = Zero Worker Household Dummy RETEMP = Retail Employment SEREMP = Service Employment TOTHH = Total Households


Workers in Household (3) by Household Income Quartile by Auto Ownership Level (3) (Generation Only)

Home-Based School Trip Generation Generation: HBGSP = POP0513 * 0.923 * 1.314

HBHSP = POP1417 * 0.943 * 1.314 HBColP = POP1824 * <PCTENR_C> * 1.157

Attraction: HBGSA = GSENROLL * 1.314 HBHSA = HSENROLL * 1.314 HBColA = COLL_FTE * 1.157 Where: HBGSP, HBGSA = Home-Based Grade School Productions and Attractions HBHSP, HBHSA = Home-Based High School Productions and Attractions HBColP, HBColA = Home-Based College Productions and Attractions POP0513 = Number of Persons age 5-13 POP1417 = Number of Persons age 14-17 POP1824 = Number of Persons age 18-24 0.923, 0.943 = Percent of persons enrolled by age (1990 Census PUMS) 1.314, 1.157 = Trips per student (estimated from 1990 Survey) PCTENR_C = Percent of 18-24 year olds, enrolled in college, by County (PUMS) GSENROLL = Grade and Middle School Enrollment HSENROLL = High School Enrollment COLL_FTE = College Full Time Equivalent Enrollment


None



The home-based work and home-based school trip generation (production) models are applied to persons who are eligible to take either work or school trips, namely, workers or students. Given difficulty in estimating home-based school trip generation models, the final models are simple trip rate models: 1.314 trips per K-12 student and 1.157 trips per college student. The home-based grade school trip attraction model was modified from the original MTC version, which set attractions equal to productions (and was, therefore, using population to estimate trip attractions) to use grade and middle school enrollment. The trip attraction rate for this model was set equal to the trip attraction rate for home-based high school trip attractions and balanced to home-based grade school trip productions.

Adjustment (calibration) of these trip generation models is included in the separate technical summary on calibration and aggregate validation. This document includes the calibration constants, as well as a discussion on the trip rate “caps” and “floors” that are needed in model application. In terms of aggregate validation, trip generation results are compared at the MTC super-district and county level to census-based “observed home-based work trips,” or 1990 survey-based observed non-work trips. Different MTC trip calibration parameters by super-district were tested during the CCTA model calibration effort for super-districts within the CCTA study area, but the original MTC trip calibration parameters were retained in the final models.

6.4 Trip Distribution Models

Gravity models are the most common form of trip distribution models. Other forms include logit destination choice models (earlier Bay Area models) and intervening opportunities models (Chicago models). Fratar, or growth factor models, are also used for short-term extrapolation of base year trip tables. All of the new Bay Area trip distribution models are gravity in form.

The final set of friction factors used in the BAYCAST gravity trip distribution models are included in Table 6.4. Essentially, these are “lookup” tables to substitute friction values for travel time.

Travel time, as used in the BAYCAST gravity trip distribution model, is either a.m. peak period door-to-door drive alone travel time; or a blend of a.m. peak period and off-peak period door-to-door drive alone travel time.



Table 6.4 Final Friction Factors for BAYCAST Trip Distribution Models

Home-Based Work Travel Time IncQ1 IncQ2 IncQ3 IncQ4

Hbased Shop/ Oth

Hbased Soc/ Rec

Hbased Grade

Sch

Hbased High Sch

Hbased College

Non-Home Based

1 400,000 350,000 320,000 350,000 500,000 250,000 999,999 999,999 999,999 999,999 2 300,000 250,000 215,000 250,000 250,000 200,000 500,000 999,999 999,999 999,998 3 220,000 125,000 100,000 138,000 143,000 202,000 200,000 300,000 175,000 290,073 4 80,000 53,000 45,000 53,000 55,000 108,000 40,000 100,000 125,000 143,547 5 45,000 30,000 29,000 28,000 33,000 50,000 23,000 80,000 50,000 85,960 6 35,000 27,500 23,000 24,300 27,500 32,200 13,500 60,000 35,000 44,936 7 25,000 20,000 17,500 18,400 17,000 20,500 7,500 40,000 22,500 33,931 8 20,000 16,500 15,000 14,500 10,000 13,100 4,000 20,000 16,500 21,498 9 17,500 13,500 12,500 12,600 7,500 12,200 2,000 10,000 12,000 17,186 10 14,500 12,000 10,400 11,000 4,000 7,250 1,500 7,500 10,000 13,092 11 13,000 10,700 9,200 8,900 3,360 6,222 1,000 4,000 6,000 10,838 12 9,700 8,600 7,500 7,600 2,500 5,495 750 2,500 5,000 6,104 13 7,400 6,800 6,400 6,200 1,700 3,800 500 1,500 4,000 5,174 14 6,000 5,900 5,700 5,400 1,300 2,500 300 1,100 3,000 3,396 15 5,000 5,000 5,000 5,000 900 2200 200 950 2,000 2,453 16 4,800 4,700 4,600 4,200 800 1671 150 800 1,800 2,338 17 4,000 4,000 3,900 3,800 690 1341 125 675 1,600 1,642 18 3,700 3,500 3,600 3,500 475 1102 75 550 1,400 1,232 19 3,400 3,100 3,200 3,100 450 955 50 425 1,200 1,125 20 2,700 2,800 3,000 2,800 350 770 40 325 1,000 1,101 21 2,600 2,600 2,600 2,500 300 600 30 250 875 855 22 2,300 2,300 2,300 2,250 200 500 25 220 750 665 23 1,800 2,100 2,100 2,000 150 440 24 175 650 618 24 1,700 1,900 1,900 1,800 125 390 23 150 550 518 25 1,400 1,700 1,700 1,650 115 350 22 110 475 414 26 1,300 1,400 1,500 1,500 100 270 21 90 440 387 27 1,050 1,200 1,300 1,350 90 240 20 75 360 359 28 900 1,100 1,200 1,200 75 215 19 60 320 287 29 800 1,000 1,100 1,100 55 190 18 50 280 255 30 725 850 975 950 53 170 17 40 240 233 31 650 750 900 850 50 150 16 30 200 198 32 575 700 800 800 47 133 15 25 170 180 33 500 650 700 725 43 120 14 19 140 172 34 450 600 600 650 40 110 13 17 120 153 35 400 550 550 575 35 100 12 16 100 137 36 360 500 500 525 34 90 11 15 90 130 37 335 450 450 475 33 80 10 14 82 115 38 325 400 400 450 31 75 9 13 74 104 39 300 350 375 425 23 63 8 12 66 95 40 275 325 350 400 22 60 7 11 58 85



Table 6.4 Final Friction Factors for BAYCAST Trip Distribution Models (continued)


Hbased Shop/ Oth

Hbased Soc/ Rec

Hbased Grade

Sch

Hbased High Sch

Hbased College

Non-Home Based

41 240 300 325 350 21 57 6 10 50 77 42 220 260 295 325 20 54 5 9 43 71 43 200 230 275 300 17 52 4 8 38 63 44 180 210 250 275 15 50 3 7 33 59 45 160 195 225 250 13 47 2 6 28 54 46 140 185 200 225 10 44 2 5 25 48 47 130 175 185 200 9 41 2 4 23 43 48 120 170 170 190 8 38 2 3 21 41 49 110 155 160 170 7 39 2 2 20 38 50 100 140 150 160 6 36 2 1 19 36 51 95 125 140 150 6 33 1 – 18 34 52 90 110 130 140 6 30 1 – 17 32 53 86 100 120 130 6 28 1 – 16 30 54 82 90 110 120 5 26 1 – 15 28 55 78 85 100 110 5 24 1 – 14 27 56 74 80 95 100 5 26 1 – 13 25 57 70 77 90 90 4 23 1 – 12 24 58 66 74 85 85 4 20 1 – 12 22 59 62 71 80 82 4 18 1 – 11 21 60 58 68 75 77 4 17 1 – 11 21 61 54 63 70 72 4 16 1 – 10 21 62 50 58 65 67 3 15 1 – 10 20 63 46 53 60 63 3 14 1 – 9 19 64 42 48 56 59 3 13 1 – 9 18 65 38 43 51 56 3 12 1 – 8 17 66 34 38 47 52 3 11 1 – 8 16 67 31 33 43 48 2 10 1 – 7 16 68 28 30 39 44 2 9 1 – 7 15 69 26 28 35 41 2 8 – – 6 13 70 24 26 31 37 1 7 – – 6 12 71 22 24 28 33 1 6 – – 5 11 72 20 23 25 29 1 6 – – 5 10 73 18 22 22 25 1 6 – – 5 9 74 16 21 20 21 1 5 – – 4 8 75 14 20 19 18 1 5 – – 4 8 76 13 19 18 17 1 5 – – 4 7 77 12 18 16 16 1 5 – – 3 7 78 11 17 15 15 1 4 – – 3 6 79 10 15 13 14 1 4 – – 3 6



Table 6.4 Final Friction Factors for BAYCAST Trip Distribution Models (continued)


Hbased Shop/ Oth

Hbased Soc/ Rec

Hbased Grade

Sch

Hbased High Sch

Hbased College

Non-Home Based

80 9 13 12 13 1 4 – – 3 6 81 9 11 10 12 – 4 – – 3 6 82 9 10 9 11 – 4 – – 3 6 83 8 9 8 10 – 4 – – 2 5 84 8 8 8 9 – 4 – – 2 5 85 8 8 8 8 – 4 – – 2 4 86 7 7 8 8 – – – – 2 4 87 7 7 7 7 – – – – 2 4 88 7 7 7 7 – – – – 2 3 89 6 7 7 7 – – – – 1 3 90 6 6 6 6 – – – – 1 2 91 6 6 6 6 – – – – 1 2 92 5 6 6 6 – – – – 1 – 93 5 6 5 5 – – – – 1 – 94 4 5 5 5 – – – – 1 – 95 4 5 4 5 – – – – 1 – 96 3 4 4 4 – – – – 1 – 97 3 4 4 4 – – – – 1 – 98 2 4 4 3 – – – – 1 – 99 2 3 4 3 – – – – – – 100 2 3 3 3 – – – – – – 105 1 2 2 3 – – – – – – 110 1 2 2 2 – – – – – – 115 1 1 1 1 – – – – – – 120 1 1 1 1 – – – – – – 125 1 1 1 1 – – – – – –

In the case of home-based work and home-based school trips, only a.m. peak-period travel times are used. For home-based shop, home-based social/recreation and non-home-based trips, a “blended” travel time based on 32.4 percent peak and 67.6 percent off-peak travel time is used. These blending shares are based on time-of-day information by trip purpose from the 1990 MTC household travel survey, as follows:



HBSH, Time HBW HBSH + HBSK NHB HBW + HBSR, TOTAL Period HBSR HBSK NHB AM 1645741 656493 738907 325277 2384648 981770 3366418 Peak 36.9% 10.8% 44.2% 6.9% 38.9% 9.1% 19.9% PM 1345441 1516593 179723 1000769 1525164 2517362 4042526 Peak 30.2% 24.9% 10.8% 21.2% 24.9% 23.3% 23.9% Total 4461255 6096871 1670741 4715609 6131996 10812480 16944476 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% PEAK 2991182 2173086 918630 1326046 3909812 3499132 7408944 (AM+PM) 67.0% 35.6% 55.0% 28.1% 63.8% 32.4% 43.7% OFFPEAK 1470073 3923785 752111 3389563 2222184 7313348 9535532 33.0% 64.4% 45.0% 71.9% 36.2% 67.6% 56.3%

AM Peak Period = 6:00 – 9:00 a.m. (3 hours) PM Peak Period = 3:30 – 6:30 p.m. (3 hours) Off-Peak = 9:00 a.m. – 3:30 p.m.; 6:30 p.m.– 6:00 a.m. (18 hours)

The home-based work trip distribution model is actually four sets of friction factors applied to home-based work trip ends stratified by household income quartile level. Data from the 1990 Census-based “observed” home-based work trip tables were used in calibrating these friction factors.

In addition to friction factors, socioeconomic adjustment factors (k-factors) are used in calibrating and validating trip distribution models. These k-factors, along with model validation results that include average trip lengths, regional trip length frequency distributions, and modeled versus observed county-to-county and super-district-to-super-district trip tables, are included in the separate technical summary on calibration and aggregate validation. K-factors were further adjusted during the CCTA model calibration process to achieve validation targets for traffic counts by time period across the CCTA cordon screenline (all roads crossing the CCTA study area boundary).

6.5 Mode Choice Models

The standard form for mode choice models is the logit choice model. Logit models were introduced by researchers in the late 1960s and entered practice in the Bay Area and elsewhere in the early 1970s. Prior to logit models, the most common form of mode choice model was the “diversion curve” model used to split trips between auto and transit modes.



Various options in logit models are binomial logit (two alternatives); multinomial logit (multiple alternatives, typically 3+); sequential-nested logit (mechanically feeding the logsum from a lower-level logit choice model to an upper-level choice model); and the simultaneous-nested logit model (“full information” from the lower nest affecting the scaling, or nesting parameter to the upper nest).

Model development in the 1970s was limited to binomial, multinomial, and sequential-nested logit choice models. Simultaneous-nested logit procedures were developed in the late 1970s and made available in commercial software (e.g., ALOGIT, LIMDEP) in the late 1980s and early 1990s.

Of the seven mode choice models included in the BAYCAST model set, six are simultaneous-nested logit choice model (hereafter, “nested logit choice”) and one, the home-based grade school mode choice model, is multinomial logit. The overall structure of these seven mode choice models is shown in Figure 6.3. All of the detailed technical memoranda discussing mode choice model development are in the Technical Memorandum Compilation Volume VI.

Figure 6.3 Home-Based Work Mode Choice – Nested Model #97

DriveAlone

SharedRide 2

TransitWalk Acc WalkBicyleTransit

Auto AccShared

Ride

Motorized Theta = 0.9208 (t = 0.6).Transit Theta = 0.7194 (t = 2.2).

One key indicator in reviewing mode choice models is the “value of time” (Table 6.5). This value of time concept is useful in understanding tradeoffs between travel time (typically IVTT) and trip cost. The rule of thumb for work trips is that the value of time is 25 to 50 percent of the average wage rate for the area. Given an average wage rate of $20.82 per hour for the Bay Area, the expected work trip value of time ranges from $5.20 to $10.41 per hour. The final nested work trip mode choice model yields an average value of time of $9.65 per hour, or 46.4 percent of the average Bay Area wage rate.



Table 6.5 Value of Time Estimates by Trip Purpose*

Trip Purpose In-Vehicle Coefficient

Cost or LnCost

Coefficient Value

of Time

Share of Wage Rate Model Source

Home-Based Work -0.03326 -0.002067 $9.65 46.4% Nested Model #97

Home-Based Shop -0.05815 -0.2262 $6.58 31.6% Nested Model 73W-2

Home-Based Social/Rec -0.02745 -1.16 $0.79 3.8% Nested Model #35

Home-Based Grade School -0.05855 -1.93 $0.36 1.7% Model #21W

Home-Based High School -0.03228 -2.034 $0.23 1.1% Nested Model #18W-3

Home-Based College -0.02731 -0.692 $0.67 3.2% Nested Model #28W-2

Non-Home-Based -0.03232 -0.9862 $1.08 5.2% Nested Model #14W-2

*Based on motorized in-vehicle and cost coefficients (1990 constant dollars per hour). Notes: 1. In-Vehicle coefficient for the HBSH model is total travel time. 2. Cost coefficients for all non-work models are for natural log of trip cost. 3. Value of time for HBW trips is IVTT/COST * 0.60. 4. Value of time for HBSH trips is TTT/LnCost * 0.60 * 42.65. The 42.65 is average trip cost. 5. Value of time for HBSR trips is IVTT/LnCost * 0.60 * 55.33. The 55.33 is average trip cost. 6. Value of time for HBGS trips is IVTT/LnCOST * 0.60 * 19.57. The 19.57 is average trip cost. 7. Value of time for HBHS trips is IVTT/LnCost * 0.60 * 23.9. The 23.9 is average trip cost. 8. Value of time for HBCol trips is IVTT/LnCost * 0.60 * 28.1. The 28.1 is average trip cost. 9. Value of time for NHB trips is IVTT/LnCost * 0.60 * 54.92. The 54.92 is average trip cost. 10. The Bay Area average wage rate is $20.82 per hour.

The rules of thumb for non-work trip values of time are not as well agreed upon as the value of time for work trips. The general feeling is that non-work value of time should be some fraction of work trip value of time. The values of time for BAYCAST non-work mode choice models range from a high of $6.58 per hour for home-based shop trips (68 percent of the work trip value of time) to a low of $0.23 per hour for home-based high school trips (2.4 percent of the work trip value of time). All of these values of time for non-work trips are fairly reasonable and well within acceptable practice for analyzing value of time.

An important characteristic of most BAYCAST mode choice models (with the exception of the three home-based school mode choice models) is that both a.m. peak period and off-peak period travel times and trip costs are used in the model application. In previous versions of MTC model systems, home-based work trips were only sensitive to peak-period travel times and costs; and non-work trips were only sensitive to off-peak times and costs. This improvement in the model system means that mode choice for these trip purposes is sensitive to changes in both the peak and off-peak period, as opposed to just one or the other.

All mode choice models incorporate non-motorized alternatives: bicycle and walk-only. Travel times for bicycle and walk are based on a “non-motorized network” based on the



standard regional highway network, excluding freeway facilities where bicycles and pedestrians are not allowed (and including freeway facilities where bicycles and pedestrians are allowed!). Uniform speeds of three miles per hour for pedestrians and 12 miles per hour for bicyclists are used to convert non-motorized distance into travel time.

The home-based work mode choice model is the only two-level nested choice model in the BAYCAST model set. Trips are first split into motorized modes, bicycle, and walk-only modes. Motorized trips are then split into drive alone, shared ride 2, shared ride 3+, and transit. Lastly, transit trips are split into transit with walk access versus transit with auto access. Market segmentation into the HBW mode choice model is zone-to-zone trips by AO level (3) by household income quartile level (4). Where the auto ownership is zero, work trips are prohibited from taking the drive alone or transit-auto access modes. Coefficients for the final nested HBW mode choice model (Model #97) are shown in Table 6.6. Definitions for these variables are included in Table 6.2.

As is typical with mode choice models, the BAYCAST home-based work mode choice model include variables about tripmaker demographics (auto ownership, income, household size, workers in the household); trip characteristics (travel time and trip cost); and neighborhood characteristics (employment density; “dummy” variables to represent high bicycle commute shares in Stanford, Palo Alto, and Berkeley; and “dummy” variables for regional “core” zones in the San Francisco financial district).

Modal constants are estimated for six of the seven modal utilities in the home-based work mode choice model. These modal constants are then calibrated (adjusted) on a district-to-district interchange basis so that model predicted trips reasonably match observed trips by mode. These changes or “deltas” to the modal constants are included in the separate technical summary on calibration and aggregate validation. The modal constants were not modified during the CCTA model calibration effort.

The coefficients for the final home-based shop/other mode choice model (Model #73W-2) are shown in Table 6.7. Both the home-based shop and home-based social/recreation mode choice models include six alternatives (drive alone, shared ride 2, shared ride 3+, transit, bicycle, walk) and one nest (either motorized or group modes). The overall structure of these six mode models is shown in Figure 6.4. The nest for the home-based shop/other model splits motorized trips from bicycle and walk trips in the upper nest; and drive alone, shared ride 2, shared ride 3+, and transit in the lower nest. As with the home-based work model, trips, where the auto ownership level is zero, are prohibited from using drive alone or auto access to transit. The home-based shop mode choice model is the only model where a total travel time variable is used. All other models were successful in terms of separating IVTT from out-of-vehicle travel time (transit wait times, walk times).



Table 6.6 Final Multinomial and Nested Home-Based Work Mode Choice Models – Multinomial Model #99W and Nested Model #97

Utility Model #99W NestModel #97 DA SR2 SR3+ TR-A TR-W Bike Walk

Variable Name Coeff. T-Stat Coeff. T-Stat

Constant -6.305 (9.1) -9.234 (4.0) Constant -9.177 (12.8) -13.310 (4.1) Constant -9.524 (13.4) -13.780 (4.1) Constant -8.802 (12.4) -12.250 (4.6) Constant -7.164 (11.2) -10.380 (4.1) Constant -8.63 (13.6) -8.268 (12.4) LnEmpDi 0.3474 (2.4) 0.3243 (2.2) LnEmpDj 0.3834 (6.7) 0.5461 (3.3) Veh/HH 0.8841 (10.1) 1.2240 (4.5)

Veh/HH 0.6489 (7.2) 0.9023 (4.2) Veh/HH 0.6754 (7.1) 0.9357 (4.2) Single VHH 0.5871 (4.1) 0.8370 (2.9) Veh/HH 0.5109 (3.5) 0.5697 (3.1) No VHH 0.4731 (2.1) 0.5501 (1.4) Wrkr/HH -0.1781 (2.9) -0.2454 (2.3)

Multi-Wrkr/HH

0.6655 (4.6) -0.9297 (3.0)

Pers/HH -0.2202 (7.1) -0.3099 (3.6) IncomeLeg1 4.077E-05 (2.3) 5.878E-05 (2.0)

IncomeLeg1 3.388E-05 (1.8) 5.049E-05 (1.7) IVTT -0.02683 (5.3) -0.03326 (4.3)

Wait -0.04180 (4.0) -0.05233 (3.1) Walk -0.05776 (2.7) -0.09305 (2.2) Cost -0.001468 (3.2) -0.002067 (2.6)

Stanfordj 2.033 (2.9) 2.09 (3.0) PaloAltoj 1.626 (2.4) 1.584 (2.3) Berkeleyj 1.062 (1.6) 1.01 (1.5) Corej -0.7605 (3.7) -1.086 (2.7)

Corej 1.004 (3.5) 1.147 (3.3) LnWalkTime -2.174 (14.0) -2.137 (13.5) LnEmpDj 0.1418 (2.0) 0.1418 (2.1)

Theta (Transit)

0.7194 (2.2)

Theta (Motor)

0.9208 (0.6)

Value of Time (IVTT/Cost * .60) $10.97 $9.65 Ratio of Wait/IVTT 1.56 1.57 Ratio of Walk/IVTT 2.15 2.80

Note: Variable definitions are included in Table 6.2.



Table 6.7 Final Nested Home-Based Shop/Other Mode Choice Model – Nested Model #73W – Nest 2

Utility Nested Model #73W-2 DA SR2 SR3+ Trans Bike Walk Variable Name Coeff. T-Stat

Constant 0.5495 (0.7) Constant -0.3612 (0.5) Constant -2.4860 (3.4) Constant -1.7470 (2.4) Constant -3.9280 (13.5) LnPHH 0.6635 (7.8) LnPHH 2.2360 (17.9) Veh/HH -0.3352 (4.0) LnIncome 0.1952 (2.7)

LnIncome 0.1118 (1.6) Time (Total) -0.05815 (13.5) LnCost -0.2262 (1.4)

Corej 2.3750 (6.0) LnAreaDeni -0.4701 (3.8)

Stanfordj 2.488 (2.5) Berkeleyj 1.630 (3.0) PaloAltoj 1.377 (1.7) Zero WHH -0.2273 (2.0)

Zero VHH 3.2910 (10.8) Zero VHH 1.7350 (6.6) Theta (Motor) 0.4847 (4.9)

Value of Time (Time/Cost * .60 * 42.65) $6.58


Figure 6.4 Home-Based Shop/Other Mode Choice – Nested Model #2

DriveAlone

SharedRide 2 WalkBicyleShared

Ride 3+

Motorized Theta = 0.4847 (t = 4.9).

Transit



All non-work trip mode choice models use a natural logarithm transformation of trip cost. This was done since the straight or linear versions of trip cost yielded either unacceptable coefficients for cost or for the calculated value of time. Value of time is calculated using the average trip cost by trip purpose (see Table 6.5).

The coefficients for the final home-based social/recreation mode choice model (nested Model #35) are summarized in Table 6.8. The nest for the home-based social/recreation model is a “group nest.” The upper nest splits drive alone, group modes, bicycle and walk trips. The lower nest splits shared ride 2, shared ride 3+, and transit trips. The ratio of the out-of-vehicle to IVTT coefficients is 2.48 (-0.06806/-0.02745), which is consistent with a priori expectations. The value of time for home-based social/recreation trips at $0.78 per hour is on the low side, but is fairly reasonable relative to other trip purposes.

Table 6.8 Final Nested Home-Based Social/Recreation Mode Choice Model – Nested Model #35

Utility NestModel #35 DA SR2 SR3+ Trans Bike Walk Variable Name Coeff. T-Stat

Constant 1.295 (2.0) Constant -1.437 (2.2) Constant -2.486 (4.5) Constant 1.703 (1.6) Constant -3.149 (7.9) LnPHH 1.8340 (11.1) Veh/HH -0.7475 (3.6) LnIncome 0.2305 (2.5) Income -8.8820E-03 (1.7) IVTT -0.02745 (3.4) OVTT -0.06806 (11.9) LnCost -1.1600 (4.9)

Corej 0.9694 (1.7) LnAreaDeni 0.3217 (1.9) Stanfordj 2.2090 (2.9) Theta (Group) 0.6271 (3.2)

Value of Time (IVTT/Cost * .60 * 55.33) $0.78 Ratio of OVTT/IVTT 2.48




Figure 6.5 Home-Based Social/Recreation Mode Choice – Nested Model #35

DriveAlone

SharedRide 2 WalkBicyleShared

Ride 3+

Group Mode Theta = 0.6271 (t = 3.2).

Transit

The coefficients for the final nested non-home-based mode choice model (Model #14W-2) are shown in Table 6.9. This model includes five alternatives (driver, passenger, transit, bicycle, and walk) and one nest (motorized trips) as shown in Figure 6.6. The upper nest for the non-home-based mode choice model splits trips into motorized, bicycle, and walk modes. The lower nest splits motorized trips into vehicle driver, vehicle passenger, and transit modes. The ratio of the wait time to IVTT coefficients is a very respectable 2.42 (-0.07836/-0.03232). The ratio of the walk time to in-vehicle time coefficients is 2.35 (-0.07583/-0.03232). Value of time for non-home-based trips is a reasonable $1.08 per hour. Given that traditional non-home-based trips are not linked with the home characteristics of the trip maker, typical demographic variables, such as household income and household size, are excluded from this model.

Table 6.9 Final Nested Non-Home-Based Mode Choice Model – Nested Model #14W-2

Utility NestModel #14W-2 VD VP Trans Bike Walk

Variable Name Coeff. T-Stat

Constant 2.2330 (8.2) Constant 0.5104 (1.9) Constant 2.0540 (5.5) Constant -4.7690 (18.4) AreaDeni -5.277E-04 (2.7)

AreaDeni 4.173E-04 (1.8) IVTT -0.03232 (4.6)

Wait -0.07836 (6.1) Walk -0.07583 (19.5) LnCost -0.9862 (12.8) Theta (Motor) 0.9144 (1.0)

Value of Time (IVTT/Cost * .60 * 54.92) $1.08 Ratio of Wait/IVTT 2.42 Ratio of Walk/IVTT 2.35




Figure 6.6 Non-Home-Based Mode Choice – Nested Model #14W-2

VehicleDriver

VehiclePassenger WalkBicyle

Motorized Theta = 0.9144 (t = 1.0).

Transit

The coefficients for the final multinomial logit choice model for home-based grade school trips (Model #21W) are included in Table 6.10. This multinomial logit model has four alternatives: vehicle passenger, transit, bicycle, and walk (as shown in Figure 6.7). Grade school students are too young to drive to school, so the vehicle driver alternative is excluded in this model. The ratio of out-of-vehicle to IVTT coefficients is on the low side at 1.09 (-0.06384/-0.05855). The value of time for home-based grade school trips is also (reasonably) low at $0.36 per hour.

Table 6.10 Final Home-Based School (Grade School) Mode Choice Model – Multinomial Logit Model #21W

Utility Model #21W VP Trans Bike Walk Variable Name Coeff. T-Stat

Constant 2.6250 (5.3) Constant 7.3003 (7.4) Constant -3.1550 (9.3) PHH^3 0.004436 (5.4) Rurali 1.5440 (3.3) Income (000s) 0.009757 (3.3) IVTT -0.05855 (4.1) OVTT -0.06384 (10.7) LnCost -1.93000 (8.7)

Value of Time (IVTT/Cost * .60 * 19.57 ) $0.36 Ratio of OVTT/IVTT 1.09




Figure 6.7 Home-Based School: Grade School Mode Choice – Model #21W

VehicleDriver WalkBicyleTransit

The coefficients for the final nested home-based high school mode choice model (Model #18W-3) are included in Table 6.11. There are five alternatives in this model and the home-based college model: vehicle driver, vehicle passenger, transit, bicycle and walk. The upper nest in the home-based high school model splits trips into vehicle driver, “group modes,” bicycle and walk. The lower nest splits group modes into vehicle passenger and transit passenger modes. The ratio of out-of-vehicle to IVTT coefficients is also on the low side at 1.07 (-0.03463/-0.03228). The value of time is the lowest of all mode choice models at $0.23 per hour.

Table 6.11 Final Nested Home-Based School (High School) Mode Choice Model – Nested Model #18W-3

Utility Nested Model #18W-3 VD VP Trans Bike Walk Variable Name Coeff. T-Stat

Constant -0.6729 (1.0) Constant 0.1929 (0.2) Constant 2.9550 (2.8) Constant -3.5240 (5.5) Veh/HH 3.5580 (2.0)

Veh/HH 0.5994 (3.5) Pers/HH -1.5000 (1.6)

Net ResDensI 0.1442 (3.5) IVTT -0.03228 (1.7) OVTT -0.03463 (5.9) LnCost -2.0340 (5.6)

Theta (Group) 0.2583 (5.5) Value of Time (IVTT/Cost * .60 * 23.9) $0.23 Ratio of OVTT/IVTT 1.07




Figure 6.8 Home-Based School: High School Mode Choice – Model #18W-Nest 3

VehicleDriver


Group Mode Theta = 0.2583 (t = 5.5).

Transit

The final mode choice model, the home-based college mode choice model (Model #28W-2) is documented in Table 6.12. The upper level nest in this model splits motorized modes, bicycle, and walk trips. The lower level splits motorized trips into vehicle driver, vehicle passenger, and transit passenger modes. To represent the high bike-to-college share to Stanford and Berkeley, “dummy” variables are used to represent residential areas in Stanford, Berkeley, and Palo Alto. A separate bicycle time coefficient is estimated in the home-based college model; in comparison, all other models include bicycle travel time as IVTT. The out-of-vehicle to IVTT coefficient ratio is on the low side at 1.44 (-0.03923/-0.02731). Value of time is higher for college trips than for grade school or high school trips at $0.67 per hour.

The mode choice model applications are designed for preparing transit and auto person trip tables for trip assignment. Up to three transit trip tables are output per trip purpose (a.m. peak auto access, a.m. peak walk access, and midday walk access) for directly assigning transit trips to the appropriate transit path. Auto person trips need to be peak-hour factored using the home-to-work departure time model or peaking factors derived from household travel surveys. Auto person trips also have to be divided by appropriate vehicle occupancy levels to convert auto person trips into vehicle driver trips.

Certain travel modes, namely, vehicle passenger trips, bicycle, and walk trips, will not normally be assigned to networks. They will be used in conjunction with other evaluation programs to account for person miles of travel by these modes, but there will not be an ongoing need for assigning these particular trips.



Table 6.12 Final Nested Home-Based School (College) Mode Choice Model – Nested Model #28W-2

Utility Nested Model #28W-2 VD VP Trans Bike Walk Variable Name Coeff. T-Stat

Constant -1.4610 (0.9) Constant -5.5060 (3.4) Constant -1.4480 (0.7) Constant -3.3980 (4.7) Veh/HH 0.7718 (4.6) Pers/HH -0.2638 (3.0) Ln Net ResDensI -0.3973 (2.1)

Stanfordi 3.216 (3.1) PaloAltoi 2.668 (2.8) Berkeleyi 1.711 (2.5) Bike Time -0.07129 (2.6) Walk (Only) Time -0.09188 (6.2) IVTT -0.02731 (1.1) OVTT -0.03923 (2.0) LnCost -0.6920 (1.8) Theta (Motor) 0.5302 (2.6)

Value of Time (IVTT/Cost * .60 * 28.1) $0.67 Ratio of OVTT/IVTT 1.44


Figure 6.9 Home-Based School: CollegeMode Choice – Model #28W – Nest 2

VehicleDriver


Motorized Mode Theta = 0.5302 (t = 2.6).

Transit



6.6 Trip Assignment The trip assignment model is the last of the four primary model components identified as part of the four-step planning process. The trip assignment model estimates the volume on each link in the transportation system for both highway and transit modes. In addi-tion, the trip assignment model generates specific performance measures, such as the con-gested speed or travel time on a highway link or the boardings and alightings on a transit route. Trip assignment is performed separately for each mode (auto and transit) and time period (a.m. peak, off-peak, and p.m. peak periods; and a.m. and p.m. peak hours).

There are two primary objectives to using the trip assignment model. The first objective is to assign trip tables and produce measures of impedance for both trip distribution and mode choice models. The second objective is to assign the trip tables and produce vol-umes for auto and transit networks. These are described separately in the following sections.

Time Periods

The trip assignment model is applied separately for each of the five time periods, which supports the two primary objectives of producing impedance measures and producing volumes by mode. The following list described the various time periods and uses for gen-erating impedance measures and volumes:

• A.M. peak-period assignments are used to produce both impedance measures and volumes by mode. This is the only time period used for both purposes. The assign-ment is completed using three-hour trip tables and three-hour capacities to estimate volume and delay for the networks. This assignment is completed for both highway and transit modes.

• P.M. peak-period assignments are used to produce volumes by mode. The assign-ment is completed using three-hour trip tables and three-hour capacities to estimate volume and delay for the networks. This assignment is only for highway modes.

• Off-peak assignments are used to produce volumes by mode. The assignment is com-pleted using 18-hour trip tables (calculated as the daily trip table minus the a.m. and p.m. peak trip tables) and eight-hour capacities to estimate volume and delay for the networks. This eight-hour delay is designed to represent the volume-delay in the midday and early evening hours, which is when most of the off-peak trips are traveling. This assignment is only for highway modes.

• A.M. peak-hour assignments are used to produce volumes by mode. The assignment is completed using one-hour trip tables and one-hour capacities to estimate volume and delay for the networks. This assignment is only for highway modes.



• P.M. peak-hour assignments are used to produce volumes by mode. The assignment is completed using one-hour trip tables and one-hour capacities to estimate volume and delay for the networks. This assignment is only for highway modes.

The free-flow travel time is required as input to each of the time period assignments and, in the case of the daily assignment, is actual free-flow travel time. In the case of the a.m. peak, off-peak, and p.m. peak assignments, the initial free-flow travel time is calculated as the travel time resulting from the 1990 MTC model assignments and output as a function of the volume-delay on the daily network.

External Trips

External trips in the CCTA Travel Model are those that have origins or destinations out-side the nine-county MTC region. The MTC model predicts external travel at 21 gateways in and out of the nine Bay Area counties using statewide survey data. These external trips are incorporated into the MTC model as vehicle trips added to the highway assignment. These procedures will be maintained in the new countywide model. External trip tables for the CCTA Travel Model were created by disaggregating the MTC external trip tables to match the CCTA zone system. External trips were allocated to zones inside the study area based on the proportion of total trips from that zone. External trips into Contra Costa, Alameda, and Solano Counties were updated based on traffic counts at those loca-tions. Figure 6.9 shows the location of the 21 gateways.

Impedance Measures

Highway Travel Time and Parking Cost

Highway measures of impedance include travel time for each time period. Travel time is measured under peak conditions (represented by the a.m. peak-period assignments) and midday conditions (represented by the off-peak assignments). The parameters used to develop these assignments are described in the following section of this report.

Highway travel time is comprised of three components:

1. IVTT for each origin-destination zone pair;

2. Terminal time for each origin zone; and

3. Terminal time for each destination zone.

The IVTT are measured in minutes and estimated as a function of free-flow travel time and volume delay. Volume delay is determined as a function of the volume-to-capacity ratio for the time period being estimated. These functions are described in the Highway Assignment part of this section.



Figure 6.10 Location of External Gateways

Terminal times represent the time it takes to travel from one’s origin to one’s vehicle, and from one’s vehicle to one’s final destination. This would typically be higher in denser urban areas, where it is necessary to park further away from the final destination. There are two types of terminal times established by the MTC model: walk time and park time. These terminal times vary for peak and off-peak conditions. A summary of terminal times used in the CCTA model is provided in Table 6.13. Zonal access travel times are also established in the MTC regional model to replace the travel time on centroid connectors, but this feature is replaced in the CCTA travel model with actual distances and travel times from centroid connectors.



Parking cost is also input to the MTC model by zone for peak and off-peak conditions. These data have been used directly in the CCTA model update. A summary of these data are also presented in Table 6.13.

Table 6.13 Terminal Times

Average Minimum Maximum

P walk time 78.8 50 200

A walk time 151.7 100 500

P park time 22.6 0 100

A park time 78.8 50 200

Peak

A zone time 107.9 10 600

Park cost 8.1 0 541

P walk time 78.8 50 200

A walk time 118.8 100 400

P park time 22.6 0 100

Off-peak

A park time 78.8 50 200

Area type 3.5 0 5

Network-based models easily calculate the travel time between zones (interzonal time) as a function of the travel time required to transverse from one zone to another. Intrazonal travel times cannot be calculated in this manner, because the modeled trips do not use the roadway network and the time within a zone would be calculated as zero. As a result, intrazonal travel times are calculated as 50 percent of the time it takes to travel to the four nearest neighboring zones.

Transit Impedance

Transit impedance is measured in components of travel time and number of boardings. Transit travel time is estimated for the same peak and off-peak conditions as the highway travel times using a.m. peak period and off-peak period assignments. Transit travel impedance is comprised of five components:

1. IVTT,

2. Auxiliary travel time for access (walk or drive),

3. Auxiliary travel time for egress (walk or drive),



4. Total wait time, and

5. Boarding time for transfers.

These measures are calculated separately for the two primary modes of transit: 1) walk access and 2) auto access. There are a series of parameters that will affect the development of transit travel times presented in Table 6.14. These reflect constraints on travel time (such as the maximum time to wait), as well as factors that account for different percep-tions of time (such as the difference in perception between time spent waiting for a bus compared to time spent riding a bus). Travel surveys have shown that time spent waiting for a transit vehicle is more onerous that time spent riding on a transit vehicle.

Table 6.14 Transit Travel Time Parameters

Parameter Value

Maximum Wait Time 15 minutes Boarding Penalty 5 minutes Wait Time Weight 2.8 Transfer Time Weight 2.8 Walk Time Weight 2.6

The waiting time at a given transit stop is based on the combined frequency of the routes that serve a particular origin and destination. This provides lower waiting times for areas with frequent service to account for the fact that travelers will choose to board the first bus that arrives (assuming it serves the travelers destination).

Time spent walking to access, transfer, or egress from the transit system is determined using a walking speed of three miles per hour and the distance along the links used. This includes distances on centroid connectors to begin or end the trip.

Distance

Distance is calculated as the weighted average of the link lengths for all paths used in the highway network. This can vary by time period since the path from an origin to destina-tion can be affected by congestion in the system. Distance is estimated in miles. Intra-zonal distance is also created as 50 percent of the distance to travel to the four nearest neighboring zones.



Highway Assignment

The highway assignment uses a stochastic user equilibrium (SUE) procedure to assign carpool and non-carpool trips to the roadway network for different time periods. This is a user optimal procedure that is based on the assumption that each traveler chooses a route that is the shortest time path. The assignment is run for 10 iterations.

Modes

There are six modes assigned in the multi-class assignment:

1. Drive alone,

2. Shared ride 2 persons,

3. Shared ride 3 or more persons,

4. Light trucks,

5. Medium trucks, and

6. Heavy trucks.

Shared ride 3+ trips are allowed to use the full network, including HOV 3+ lanes; shared ride 2 trips are allowed to use the full network, except for HOV 3+ lanes; and drive alone trips are restricted to the use of general purpose (non-HOV) lanes. Trucks are restricted to non-HOV lanes and are further prohibited from selected facilities.

Volume-Delay Functions

The highway assignment procedure is applied in an iterative fashion, where travel times are updated after each iteration to reflect congestion occurring on the network. These updates to travel time are based on a volume-delay function for each link, which are pre-sented in Figure 6.11. Majority of roads use a volume-delay function that was developed by Akcelik and are consistent with those established by MTC. Additional volume-delay functions by facility type were tested during model calibration, but the original MTC functions were retained in the final model. The free-flow time is based initially on the network data provided for each link,2 and then updated in each iteration to represent the travel time from the last iteration.

2 The free-flow travel time is required as input to each of the time period assignments and, in the

case of the daily assignment, is actual free-flow travel time. In the case of the a.m. peak, off-peak, and p.m. peak assignments, the initial free-flow travel time is calculated as the travel time resulting from the daily assignment and output as a function of the volume-delay on the daily network.



Figure 6.11 Volume-Delay Functions by Facility Type

0

10

20

30

40

50

60

70

0 0.2 0.4 0.6 0.8 1 1.2 1.4

Volume/Capacity Ratio

Speed (mph)

FreewayExpresswayCollectorFreeway RampMajor ArterialMetered Ramp

Turn Penalties

Turn penalties are included in the trip assignment model to prohibit certain turn move-ments. These are included in the model by identifying specific turn movements by their node numbers. The current model contains 122 turn prohibitors.

Transit Assignment

Transit Assignment was implemented using an exact replica of MTC transit path building process, which required a unique process in TransCAD that can mimic the TP+ transit path builder routine. Transit assignment uses a multi-path algorithm based on the assess-ment of optimal strategies. The optimal strategy is the path for each traveler that mini-mizes the expected travel time, including time spent walking, waiting, and riding. Time spent waiting for a transit vehicle is calculated based on the fact that there may be many transit vehicles traveling from a specific origin to a specific destination, and the traveler will choose to take the first vehicle that arrives.



Modes

There are three modes used in transit assignment: best path, walk path, and midday. The best path is used to determine the drive access to transit trips. The walk mode provides access to the transit modes (bus, rail, and ferry). Transit assignments are only carried out for walk access to transit, because drive access to transit trips are split into two parts: 1) the auto portion of the trip, which is assigned to the highway network; and 2) the tran-sit portion of the trip, which is assigned from the park-and-ride lot to the final destination.

Transit Time Functions

Transit time is calculated separate from highway time to account for the fact that transit vehicles have to stop and pick up passengers along the route, and typically travel at slightly slower speeds than passenger cars due to their size and weight. MTC uses transit runtime from real transit schedules and highway travel time to calculate a time factor. This time factor is then applied to highway time to calculate transit travel times that match the input run times.



7.0 Model Refinements

7.1 Introduction

The CCTA countywide model has been converted from MTC and implemented in TransCAD software. There are several model refinements that were considered and adopted for use in the CCTA countywide model to enhance the capabilities of the model for local area planning activities. The remainder of this section presents the approach to implementing each of the proposed model refinements and provides information on why the refinement is beneficial to the CCTA countywide model.

There are five model refinements that have been adopted by the Technical Model Working Group (TMWG) for the CCTA countywide model:

1. Modifying the estimation of school trip attractions in the trip generation model based on the enrollment of schools within the study area.

2. Adding special generator trips within the study area to the trip generation model.

3. Identifying transit trips that drive to park-and-ride lots, and assigning these trips to the highway network.

4. Developing time-of-day models for a.m. peak, off-peak and p.m. peak periods, assigning these time-period specific trip tables separately; then summing the volumes to produce average daily volumes.

5. Developing peak-hour assignment processes for a.m. and p.m. peak hours.

The approaches and rationale for these model refinements are discussed in the remaining sections.



7.2 School Trip Attractions Model

Purpose

The home-based school trip attractions model, converted from the MTC model, estimates school trip attractions using the following assumptions:

• Home-based grade school attractions are based on population ages five to 13;

• Home-based high school attractions are based on high school enrollment, and

• Home-based college attractions are based on college full-time employment equivalents (FTE).

Since the home-based grade school (defined to include grade school and middle school trips) trip attractions are based on residential data rather than enrollment data, we pro-posed revising the home-based grade school trip attraction equations to be based on grade/middle school enrollment.

Approach

The MTC school trip attraction equations are provided in Table 7.1. The home-based grade school trip attraction equation was modified to reflect grade and middle school enrollment using the following equation:

HBGSA = GSENROLL * 1.314

Where:

GSENROLL = The grade and middle school enrollment; and

1.314 = The number of school trips per student per day (estimated from the 1990 MTC survey).

Data

Table 7.2 presents grade/middle and high school enrollment for the six super-districts in the study area. Figure 7.1 shows the location of grade, middle, and high schools. A com-prehensive list of all public and private schools is provided in Appendix G.



Table 7.1 MTC School Trip Production and Attraction Equations

Trip Purpose Trip Production Equation Trip Attraction Equation Home-based grade school

HBGSP = POP0513 * 0.923 * 1.314 HBGSA = HBGSP

Home-based high school

HBHSP = POP1417 * 0.943 * 1.314 HBHSA = HSENROLL * 1.314

Home-based college

HBColP = POP1824 * <PCTENR_C> * 1.157 HBColA = COLL_FTE * 1.157

Assumptions Where: POP0513 = Number of persons age 5-13 POP1417 = Number of persons age 14-17 POP1824 = Number of persons age 18-24 0.923, 0.943 = Percent of persons enrolled by age (1990 Census PUMS) 1.314, 1.157 = Trips per student (estimated from 1990 Survey) PCTENR_C = Percent of 18-24 year olds, enrolled in college, by county (PUMS)

Where: HSENROLL = High school enrollment COLL_FTE = College full time equivalent enrollment 1.314, 1.157 = Trips per student (estimated from 1990 Survey)

Table 7.2 School Enrollment by Super-District

Super-District Grade/Middle

School Enrollment High School Enrollment

Richmond/El Cerrito 26,071 9,417 Concord/Martinez 20,028 9,116 Walnut Creek/Lamorinda 16,224 8,506 Danville/San Ramon 13,778 6,321 Antioch/Pittsburg 32,095 12,126 Livermore/Pleasanton 19,925 9,119 Total 128,121 54,605



Figure 7.1 School Locations in Contra Costa



7.3 Special Generators

Purpose

The CCTA countywide model, converted from the MTC model, does not include any spe-cial generators in the Contra Costa model study area. The primary reason for including special generators is to more accurately predict traffic volumes and transit ridership in local areas where special generators exist. At a regional level, special generators are less important to include in the travel demand forecasting model, because the evaluation of traffic volumes and transit ridership is not detailed at the local level to warrant the level of effort necessary to include and maintain special generators. For a countywide model, these become more important and, pending a review of the significance of any particular generator, are recommended for use in predicting local area volumes.

Approach

As part of the countywide model development effort, Cambridge Systematics, Inc. com-piled and reviewed an initial list of special generators. This review focused on the ability of current and future models to capture the travel characteristics for these generators within the available trip purposes defined for the model. The trips associated with these generators were established outside the four-step modeling process, and evaluated using the ITE Trip Generation Manual.3 We evaluated trips estimated by the Contra Costa coun-tywide trip generation model for these generators (using existing trip attraction rates), and compared these to trips calculated from ITE Trip Generation Manual rates to determine which special generators warranted special attention in the countywide travel model. Special generators that were evaluated and included are hospitals, military bases, shop-ping malls, and retirement homes. We also evaluated regional parks and unique indus-trial areas, but these did not pass the threshold established to qualify as a special generator.

Special generators typically apply only to the trip generation portion of the four-step process as the trip generation rates are more distinct than any other portion, so our approach is to evaluate trip generation impacts for all special generators. The following steps were followed to evaluate the trip generation impacts of special generators:

1. Identify an initial list of potential special generators to include hospitals, military bases, shopping malls, retirement homes, regional parks, garbage dumps, and unique industrial areas.

3 Institute of Transportation Engineers, Trip Generation User’s Guide, Sixth Edition, Publication

Number IR-016D, 1997.



2. Obtain input from the TMWG on specific generators that should be included within each category for a total of 31 generators. Following the initial consideration of the full list of potential generators, this resulted in a total of 24 special generators for evalua-tion, seven were dropped due to lack of data or limited overall impacts on traffic.

3. Obtain data on size for each of 24 special generators and identify the number of employees that would be associated with each of these special generators. Use these data, along with ITE trip generation manual trip rates, to estimate daily person trips associated with each special generator.

4. Estimate daily person trips that the countywide model would generate for these facili-ties based on the number of employees in each.

5. Compare the difference between the estimated special generator trips and the modeled trips and recommend special generators based on a difference of more than 2,500 daily person trips. This resulted in a recommendation for 15 special generators to be included in the countywide model.

After the total daily person trips are estimated for each special generator, these were allo-cated for each trip purpose and added to the remaining trips in the trip generation model. At the same time, the trips associated with employees at each generator were subtracted from the trip generation model to avoid double-counting trips at these generators.

Data

Special generators that were identified for addition to the CCTA countywide travel model included hospitals, military bases, shopping malls, retirement homes, regional parks, gar-bage dumps, and unique industrial areas. Data on size of these generators were devel-oped from GIS developed by Environmental Systems Research Institute, Inc. (ESRI) and Geographic Data Technology, Inc. (GDT). The following special generators were consid-ered, but later dropped due to lack of data and limited traffic impacts:

• Garbage dumps at Keller Canyon and Vasco Road were initially considered, but later dropped due to lack of data and expectations of low traffic impacts.

• Livermore Airport with six permanent and nine temporary staff was considered to be too small to be included as a special generator.

• Naval Weapons Station in Concord was decommissioned in the year 2000 and retains only 30 employees. This was considered to be too small to be included as a special generator.

• Richmond Clinics is a medical office facility and not a hospital, so it was dropped from the list of hospitals.

• Springtown and Sunny Glen Retirement Homes were dropped due to lack of data.



Table 7.3 presents the list of special generators that were evaluated. Figure 7.2 presents the distribution of these generators for the Contra Costa model study area by type of generator.

Table 7.3 Special Generators by Type and Size

ID Name Category TAZ Size Units

1 Contra Costa Regional Medical Center

Hospital 20010 164 Beds

2 John Muir Medical Center Hospital 20235 321 Beds

3 Kaiser Permanente Walnut Creek Medical Center


4 San Ramon Regional Medical Center


5 Veterans Affairs Northern California Health Care


6 Doctors Medical Center Hospital 10118 233 Beds

7 Mt. Diablo Medical Center Hospital 20123 259 Beds

8 Sutter Delta Medical Center Hospital 30129 110 Beds

9 Valley Care Health System Hospital 50377 142 Beds

10 Broadway Plaza/Downtown Walnut Creek

Shopping mall 20221 685 Square feet

11 Sun Valley Mall Shopping mall 20109 1,440 Square feet

12 County East Mall Shopping mall 30076 540 Square feet

13 Hilltop Mall Shopping mall 10173 1,700 Square feet

14 Stoneridge Mall Shopping mall 50365 1,285 Square feet

15 Parks Reserve Forces Training Area

Military base 50322 741 Employees

16 Mt. Diablo Regional park 20260 20,000 Acres

17 Black Diamond Mines Regional park 30088-30089 5,717 Acres

18 Briones Regional park 20001-20003 5,756 Acres

19 Wildcat Canyon/Tilden Regional park 10182-10185 4,505 Acres

20 Las Trampas Regional park 40221-40224 3,798 Acres

21 Morgan Territory/Round Valley

Regional park 40251, 40258 6,171 Acres

22 Pt. Pinole Regional park 101061 2,315 Acres

23 Rossmoor Retirement Home Retirement home 20270 6,400 Dwelling units

24 Lawrence Livermore Lab Unique industry 50512-50515 8,000 Employees



Figure 7.2 Distribution of Special Generators

Source: ESRI and GDT, U.S. GDT Institutions and U.S. GDT Retail Centers, April 2002.

Special Generator Development

Special generators were evaluated by comparing trips that were currently generated from facilities in the Contra Costa countywide travel model to trips that were generated from these facilities as special generators. In order to make this comparison, it was necessary to calculate employment that is related to each generator. This employment estimate, pre-sented in Table 7.4, also provides a check on the inclusion of the generator employment in the land use data file. Table 7.4 also provides an identification of the type of employment that is used for comparison, as follows:



Table 7.4 Special Generator Employment Estimates

ID Name

Special Generator Employees

TAZ Employees

Percent of Employees

Type of TAZ Employees


373 1,111 34% Service

2 John Muir Medical Center 731 1,433 51% Service


521 966 54% Service

4 San Ramon Regional Medical Center

280 280 100% Service


326 511 64% Service

6 Doctors Medical Center 530 1,080 49% Service

7 Mt. Diablo Medical Center 590 1,300 45% Service

8 Sutter Delta Medical Center 250 289 87% Service

9 Valley Care Health System 323 2,015 16% Service


1,076 1,622 66% Retail

11 Sun Valley Mall 2,261 4,473 51% Retail

12 County East Mall 848 1410 60% Retail

13 Hilltop Mall 2,669 2669 100% Retail

14 Stoneridge Mall 2,017 2017 100% Retail

15 Parks Reserve Forces Training Area

741 1581 47% Total

29 Mt. Diablo 15 15 100% Other

30 Black Diamond Mines 6 47 13% Other

31 Briones 6 53 11% Other

32 Wildcat Canyon/Tilden 5 362 1% Other

33 Las Trampas 4 760 0% Other

34 Morgan Territory/Round Valley

6 99 6% Other

35 Pt. Pinole 2 535 0% Other

36 Rossmoor Retirement Home 8,000 8261 97% Population

37 Lawrence Livermore Lab 8,000 9,523 85% Service



• Service employment includes hospital employees,

• Retail employment includes shopping center employees,

• Employees in regional parks are compared to other/total employees in the LUIS, and

• Total residents in retirement homes are compared to total population.

Special generator employment is used to calculate trips that are directly related to the spe-cial generator using two methods:

1. ITE Trip Generation Manual provides trip rates for individual generators based on size variables. The size variable is the number of beds for hospitals, thousand square feet (TSF) of gross leasable area for shopping centers, and acres of land for regional parks.

2. The Contra Costa countywide travel model provides trip rates for specific categories of employment. The service employment category contains jobs in hospitals, the retails employment category contains jobs in shopping centers, and the other employ-ment category contains jobs in regional parks.

Table 7.5 provides a comparison of the results of the application of these two methods. The trips generated from the ITE Trip Generation Manual involve a series of assumptions for each type of generator; these are provided for reference in Table 7.6.

We established a threshold of 2,500 trip differences per generator to consider inclusion in the model as a measure of significant impact to the traffic volumes. Under this threshold, all the shopping centers, all but one hospital, and none of the regional parks will be included. In addition, the Parks Reserve Forces Training area (a military base) and the Rossmoor Retirement Home will be included in the model.

Implementation

There were 15 proposed special generators to add to the existing trip generation model. These are presented in Table 7.7. The five shopping center generators were distributed using the home-based shopping gravity model. The hospitals and military base were dis-tributed using the home-based other trip distribution model.



Table 7.5 Special Generator Trip Estimates

ID Name

Special Generator

Trips Model Trips Difference

Special Generator Trips Per Employee

1 Contra Costa Regional Medical Center 3,954 656 3,298 10.59

2 John Muir Medical Center 5,518 1,096 4,422 7.55

3 Kaiser Permanente Walnut Creek Medical Center 4,602 783 3,818 8.83

4 San Ramon Regional Medical Center 3,545 2,353 1,193 12.66

5 Veterans Affairs Northern California Health Care 3,745 982 2,763 11.50

6 Doctors Medical Center 4,641 846 3,795 8.75

7 Mt. Diablo Medical Center 4,901 999 3,902 8.31

8 Sutter Delta Medical Center 3,416 964 2,452 13.64

9 Valley Care Health System 3,735 539 3,196 11.55

10 Broadway Plaza/Downtown Walnut Creek 31,725 8,220 23,505 29.49

11 Sun Valley Mall 51,137 17,313 33,824 22.62

12 County East Mall 27,217 3,443 23,774 32.11

13 Hilltop Mall 56,896 20,972 35,924 21.32

14 Stoneridge Mall 47,526 21,809 25,716 23.56

15 Parks Reserve Forces Training Area 5,395 408 4,987 7.28

29 Mt. Diablo 1,615 61 1,554 107.69

30 Black Diamond Mines 623 24 600 107.69

31 Briones 628 24 604 107.69

32 Wildcat Canyon/Tilden 491 19 473 107.69

33 Las Trampas 414 16 398 107.69

34 Morgan Territory/Round Valley 673 26 647 107.69

35 Pt. Pinole 252 10 243 107.69

36 Rossmoor Retirement Home 18,576 20479 1903 2.14

37 Lawrence Livermore Lab 20,088 20,306 -218 2.5



Table 7.6 Assumptions for Special Generator Categories*

Trips Per Bed Employees

Per Bed Trips Per Employee

ITE Code 610 for Hospitals

Average Size

Hospitals 11.77 2.28 5.17 Trips = 7.381*Beds+1718.324 392 beds

Trips Per TSF Employees

Per TSF1 Trips Per Employee

ITE Code 820 for Shopping Centers

Average Size

Shopping Centers

42.92 1.57 27.34 Ln(Trips) = 0.643 Ln(x) +5.866 331 TSF

Trips Per

Acre2 Employees Per Acre3

Trips Per Employee

ITE Code 417 for Regional Parks

Average Size

Regional Parks

0.0808 0.001013 79.77 79.77 person trips per employee 310 acres

Trips per employee

ITE Code 510 for Military Bases

Average Size

Military Bases

7.28 Ln(Trips) = 0.571* Ln(x) +4.520 7,747

Trips Per Unit ITE Code 252 for

Congregate Care Facility Average

Size Retirement Homes

2.15 2.15 Trips per Dwelling Unit 183

Trips Per Employee

ITE Code 760 for Research and Development Center

Average Size

Unique Industry

2.77 Ln(Trips) = 0.800* Ln(x) +2.418 1,022

* ITE, Trip Generation User’s Guide, Sixth Edition, Publication Number IR-016D, 1997. 1. ITE did not provide an estimate of employees per TSF, so this was derived from a study in California,

which can be found at http://www.ceres.ca.gov/ceqa/cases/1991/shapell_112291.html. 2. ITE trips per acre (4.57) results in too many trips. We calculated trips to Mt. Diablo based on total employ-

ees in the TAZ and divided it by the acres to get a more realistic ‘trips per acre’ value. 3. Employees per acre was derived by using the calculated ‘trips per acre’ value and ITE Code 417 for regional

park.



Table 7.7 Recommended Special Generators

Name

Special Generator

Trips


Remaining TAZ

Employees Type of TAZ Employees


3,954 373 738 Service

2 John Muir Medical Center 5,518 731 702 Service


4,602 521 445 Service


3,745 326 185 Service

6 Doctors Medical Center 4,641 530 550 Service

7 Mt. Diablo Medical Center 4,901 590 710 Service

8 Sutter Delta Medical Center 3,416 250 39 Service

9 Valley Care Health System 3,735 323 1,692 Service


31,725 1,076 546 Retail

11 Sun Valley Mall 51,137 2,261 2,212 Retail

12 County East Mall 27,217 848 -848 Retail

13 Hilltop Mall 56,896 2,669 -1,709 Retail

14 Stoneridge Mall 47,526 2,017 -1,837 Retail

15 Parks Reserve Forces Training Area 5,395 741 -589 Service

7.4 Park-and-Ride Trips

Purpose

In most regional travel demand forecasting models, park-and-ride trips are identified during the mode choice modeling process as “drive access to transit” trips. This is true of the MTC regional travel modeling system. Since these park-and-ride trips are identified as transit trips, they are modeled from origin to destination with the park-and-ride lot as an intermediate stop on the trip. During assignment, these trips are assigned to the transit network, traveling from the park-and-ride lot station (or bus stop) to the destination of the trip. The portion of the trip which involves driving from the origin of the trip to the park-and-ride lot is retained in the trip table, but never assigned to the network. When vehicle miles traveled are determined, the vehicle miles traveled associated with these park-and-



ride lots trips are estimated as a post-process to the model and included in overall vehicle miles traveled estimates.

In the CCTA model study area, there is quite a bit of congestion during peak hours in areas near the park-and-ride lots. This congestion cannot be currently quantified in the CCTA countywide model unless the park-and-ride lot trips are added to the highway assignment process. This process is desired to produce accurate estimates of congestion in and around park-and-ride lots in the study area, especially during peak hours.

Approach

The approach was to retain the current process of assigning drive access to transit trips to the transit network from the park-and-ride lot to the destination of the trip, and then to add an additional process that assigns the remaining portion of the trip from the origin to the park-and-ride lot to the highway network. This additional process included devel-oping trip tables that represent this portion of the overall drive access to transit trip. The process was implemented using an exact replica of MTC transit path building process, which required a unique process in TransCAD that can mimic the TP+ transit path builder routine. This transit path builder was written specifically to mimic the TP+ process and, because it is not a “true” path builder, the drive access trips cannot be easily assigned directly to a park-and-ride lot. Instead, these trips were assigned to the closest highway node or intersection as a simplifying routine.

Once the trip tables to represent the drive access portion of these transit trips were cre-ated, it was incorporated into the highway assignment process as “SOVs.” This is a sim-plifying assumption, because, while we know that not all drive access to transit trips are SOVs, we do not have a submodel that can differentiate these trips into SOVs and HOVs. Since these drive access trips were primarily using local roads or arterials to access the park-and-ride lots, rather than freeways with HOV lanes, this assumption seemed reason-able for our purposes.

Results

Drive access assignment procedure is implemented for the entire model area. Table 7.8 shows the number of people driving to the BART stations inside Contra Costa County.



Table 7.8 Drive Access Trips to Bart Stations inside the CCTA Study Area

Name Drive Access

Trips

1 Orinda 3,155

2 Lafayette 1,387

3 Walnut Creek 8,253

4 Pleasant Hill 6,162

5 Concord 3,158

6 North Concord 2,771

7 West Pittsburg/Baypoint 7,086

7.5 Peak Period and Daily Models

Purpose

The current MTC regional travel forecasting model includes two a.m. peak-period assignments (one for a two-hour period and another one for a four-hour period) and a daily assignment. The daily assignment is primarily used for summarizing regional travel statistics and is not considered reliable to evaluate congestion for individual roadways, since congestion varies widely across the 24 hours in a day. As a result, we recommended that the CCTA countywide model incorporate a time-of-day model that separates the daily highway trip tables output from the mode choice model into two peak periods and one off-peak period. These time period trip tables were then assigned separately to the highway network to produce volumes and speeds for individual time periods. The sum of the three time period volumes were used to estimate average daily volumes for any roadway in the system, rather than assigning a daily trip table to estimate daily volumes. This is a preferred method to estimate daily volumes, because the peak and off-peak travel conditions are more accurately represented.

Approach

MTC has provided estimates of trip by trip purpose and mode for each hour of the day, developed from the 1990 MTC household travel survey. These data can be updated when the 2000 MTC household travel survey data becomes available. The “vehicle driver” mode



from this analysis was used to disaggregate the drive alone, shared ride 2, and shared ride 3+ trip tables from the mode choice model into three time periods:

1. a.m. peak period from 6:00 a.m. to 10:00 a.m.;

2. p.m. peak period from 3:00 p.m. to 7:00 p.m.; and

3. Off-peak period covering all remaining hours.

Each trip purpose time period factor is estimated by direction; that is, whether the trip is from production to attraction (P-A direction) or whether the trip is from the attraction to the production (A-P direction). This directionality is necessary to convert the P-A trip tables into origin-destination (O-D) trip tables. The difference between P-A trip tables and O-D trip tables is that P-A trip tables have all trips beginning at the home end for home-based trips. Thus, a home-based work trip will have two trips beginning at home and ending at work in the P-A tables; and will have one trip beginning at home and ending at work, plus one trip beginning at work and ending at home in the O-D table. This is dem-onstrated in Figure 7.3.

Figure 7.3 Production-Attraction and Origin-Destination Trips

Home

Work2 P-A Trips from Home to Work

1 O-D Home to Work Trip and1 O-D Work to Home Trip

Once the time-period-specific trip tables are developed, these are assigned to the highway network to produce volumes and speeds for each time period and each link in the net-work. Daily volumes are then constructed by summing the a.m. peak, the p.m. peak and the off-peak-period volumes. This results in different traffic volumes than assigning the daily trip table to the network, because traffic diversion due to congestion is represented differently. It is considered advantageous to use the summation process rather than the



assignment process to produce daily volumes because the congestion is represented in a more disaggregate and accurate form.

Data

Peaking factors used in the CCTA countywide model are provided in Table 7.9. Factors are by purpose and direction of trips.

Table 7.9 Peaking Factors for AM and PM Peak Period

Purpose Direction

AM Peak Period

(6:00 a.m. to 10:00 a.m.)

PM Peak Period

(3:00 p.m. to 7:00 p.m.)

Off-Peak Period

Home-based work From home 39.69% 3.07% 11.10% Home-based work To home 1.13% 31.65% 13.37% Home-based school From home 48.25% 2.84% 5.27% Home-based school To home 0.86% 18.34% 24.43% Home-based other From home 11.29% 13.49% 21.85% Home-based other To home 3.75% 19.98% 29.64% No-home-based From/to home 11.34% 28.64% 60.02% Small trucks 24.80% 24.80% 50.40% Medium trucks 28.85% 28.85% 42.30% Combo trucks 23.40% 23.40% 53.20% Total 23.66% 31.55% 44.79%

7.6 Peak Hour Model

Purpose

The current MTC regional travel demand forecasting model does not contain any proce-dures to produce peak-hour traffic volumes. Since the CCTA and various local agencies rely on peak-hour traffic volumes for a variety of planning applications, it was determined that a peak hour model was a necessary model refinement.

The proposed peak hour model generates a.m. and p.m. peak-hour traffic volumes for all roadways in the CCTA model study area. The purpose of this is to extract the peak-hour



volumes from the peak-period traffic in each peak period. Since the peak hour (i.e., the highest hour of volume within the four-hour peak period) is different for each link and subarea in the model study area, it was determined by the TMWG that a “floating” peak hour was preferred over a “fixed” peak hour. This means that the peak hour is the highest peak hour for each link in the study area, rather than specifying a particular hour of the day that represents the peak hour for all links in the study area.

Approach

The recommendation to use a “floating” peak hour requires a review of the process to develop peak-hour trips rather than the “fixed” peak-hour definition. A “fixed” peak-hour approach involves estimating travel demand during a fixed peak hour by applying time-of-day peaking factors to a peak-period trip table, and possibly adjusting for peak-spreading phenomena. A “floating” peak-hour approach cannot estimate travel behavior in a fixed time period, because some parts of the study area will have an earlier peak hour than other parts of the study area. Our recommended approach was to apply a fixed peak-hour factor to the peak-period trip table, then use matrix adjustment procedures to adapt this trip table to the highest peak-hour traffic counts throughout the study area. This process is called origin-destination matrix estimation (ODME) in the TransCAD software.

The ODME process creates trip table adjustment factors from the existing traffic counts that are applied to the trip tables during the peak-hour assignment process. In the fore-cast model runs, these trip table adjustment factors were used with future peak-hour trip tables to simulate the same process. The trip table adjustment factors indirectly address the peak-spreading phenomena in congested areas of the system. It does not address future peak spreading. The Authority’s Technical Procedures include a methodology to address future peak spreading by applying a post-processing technique to the peak-hour volumes generated by the model.

During model calibration, we reviewed the ODME process and reported some very high adjustment factors, especially in areas where the numbers of trips were small. As a result, we also tested applying ODME with constraints on the size of the adjustment factors. These model runs did not produce results that were reasonable compared to the calibra-tion results without running ODME. In addition, the peak-hour assignment models with-out ODME were reasonable for aggregate measures of validation. As a result, the ODME process to adjust for the floating peak-hour volumes was dropped from the approach.

Data

The peak-hour percentages are calculated from the screenline counts collected for use in model validation by PHA Transportation Consultants. There are 414 count locations in this Validation Database. The following peak-hour percentages were calculated for each peak hour:



• 31.1 percent of the traffic in the a.m. peak period occurs in the peak hour; and

• 29.8 percent of the traffic in the p.m. peak period occurs in the peak hour.

The peak hour for any particular link range all the way from 6:00 a.m. to 10:00 a.m. in the a.m. peak period, and from 3:00 p.m. to 7:00 p.m. in the p.m. peak period, indicating a wide range of peak hours throughout the study area.

Results

The validation results of the peak-hour assignment are not as strong in local areas and lower facility type roads, because of the decision to assign trips without any adjustment factors (using the ODME process) to account for the floating peak-hour counts. Since the model is necessarily a fixed demand for a certain time period (such as the peak hour), the model cannot accurately replicate peak-hour counts that are for a floating peak hour. Table 7.10 presents a comparison of select validation tests to demonstrate the difference in validation results that arise from this decision.

Table 7.10 Comparison of Peak-Hour Assignments With and Without ODME

PM Peak-Hour Assignment

Without ODME

PM Peak-Hour Assignment With ODME Target

Link-Based Validation Tests Freeway Links within 20% 81% 59% 75% Freeway Links within 10% 66% 32% 50% Arterials with 10,000+ vehicles within 30% 68% 88% 75% Arterials with 10,000+ vehicles within 15% 53% 59% 50% Intersections with 1000+ vehicles within 20% 46% 73% 50% Intersections with 500-1000 vehicles within 20% 28% 64% 30%

Facility Type Validation Tests Freeway 2% -2% +/-7% Major Arterial -7% -6% +/-10% Minor Arterial -30% -13% +/-15% Collector -25% -10% +/-25% Total -7% -5% +/-5%

Area Type Validation Tests CBD/Urban Bus -11% -5% +/-5% Urban -5% -6% +/-5% Suburban -7% -5% +/-5% Rural -1% -5% +/-5% Total -7% -5% +/-5%



8.0 Model Validation

The primary purpose of model validation is to test the reasonableness of the travel demand forecasting model for the planning applications it will be used to evaluate. The goal is to validate different components of the modeling system separately and to validate different model outputs that will be used in planning evaluations. To that end, the fol-lowing categories of model validation targets were considered:

• Geographic comparisons of daily model volumes and traffic counts to review the rea-sonableness of the trip generation model component.

• Screenline and cordon line comparisons of daily, a.m. and p.m. peak hour model vol-umes and traffic counts to review the reasonableness of the trip distribution, mode choice and trip assignment component. This would also include comparisons of model boardings and observed counts for groups of transit lines to represent corridors.

• Total boardings, transfer rates, and total daily model volumes compared to observed counts by mode to review the reasonableness of the mode choice component.

• Total daily and time period comparisons of model volumes and traffic counts by facil-ity type, area type, and volume group to assess the reasonableness of the trip assign-ment and peak spreading model component. Time periods would include a.m. and p.m. peak periods and a.m. and p.m. peak hours.

• Systemwide comparisons of model and observed measures of performance. For the highway system, this would include vehicle miles traveled, vehicle hours traveled and vehicle hours of delay. For the transit system, this would include total boardings by operator, total boardings by type of service (local or express), number of transfers by operator and total boardings by time period.

• Link-based comparisons of a.m. and p.m. peak hour volumes and traffic counts by facility type and time period to assess the reliability of the volumes.

Validation targets for each of these categories were developed to provide a reasonable level of accuracy at the countywide and subarea planning level expectations. It is not expected that the validated county model would meet the expectations of project-level or local level planning applications without localized refinements to the input data.



8.1 Systemwide Validation

The systemwide validation is completed for the CCTA study area and considers miles traveled, as well as hours traveled to consider the reasonable of speed on the overall sys-tem. Regionwide validation is completed for the nine-county Bay Area in the context of the MTC consistency requirements for trip generation, trip distribution, and mode choice and is presented in the MTC Model Consistency report.

The CCTA travel model is within +/-4 percent of the systemwide vehicle miles traveled, within +/-6 percent of the systemwide vehicle hours traveled and within +/-4 percent of the systemwide vehicle hours of delay for all time periods, except for a.m. peak hour, which is 11 percent over-estimated. In addition, the CCTA travel model predicts an aver-age speed within +/-3 percent of observed for all time periods. Table 8.1 presents a com-parison of the observed and estimated systemwide variables.

Table 8.1 Systemwide Validation Test

Model Results

Variables AM Peak

Period AM Peak

Hour PM Peak

Period PM Peak

Hour Targets Highway Assignments VMT Counts 1,419,679 441,384 1,586,832 449,237 VMT Model 1,381,279 459,201 1,616,217 450,201 Percent Difference -3% 4% 2% 0% +/-5% VHT Counts 33,606 14,095 36,454 13,337 VHT Model 32,346 15,006 36,067 13,365 Percent Difference -4% 6% -1% 0% +/-5% VHD Counts 526,001 375,683 490,047 316,678 VHD Model 521,518 418,535 472,537 322,211 Percent Difference -1% 11% -4% 2% +/-5% Average Observed Speed 42.25 31.32 43.53 33.68 Average Model Speed 42.70 30.60 44.81 33.69 Percent Difference 1% -2% 3% 0% +/-5% Transit Assignments Daily Boardings 10% +/-10%

8.2 Highway Assignment Validation

The highway assignment is assessed in three individual validation tests, comparing traffic counts to model estimated volumes by facility type, area type, and screenline for each time period. There are five time periods in this analysis: a.m. peak period, p.m. peak period, a.m. peak hour, p.m. peak hour, and daily. The daily assignment is a summation of the a.m. peak period, the p.m. peak period and the off-peak period. The off-peak period is not



reported separately for validation purposes, because the average daily traffic is a more critical value for interpretation of results and is therefore reported instead of the off-peak period.

The validation test of highway assignment by facility type is presented in Table 8.2 by time period. This table shows a trend that freeways are slightly over-estimated (but within the validation target by time period) and arterials are slightly under-estimated. Lower level facilities are consistently under-estimated, but this can be improved through local review and judgment of network attributes at a local level. This validation test shows that the model is producing reasonably results by facility type.

Table 8.2 Validation of Highway Assignment by Facility Type and Time Period

Facility Type

Number Of

Counts 2000

Counts 2000

Model Difference Percent

Difference Target A.M. Peak Period (6:00-10:00 a.m.) Freeway 70 1,255,612 1,320,791 65,179 5% +/-7% Major Arterial 220 468,587 427,932 -40,655 -9% +/-10% Minor Arterial 24 16,063 10,698 -5,365 -33% +/-15% Collector 99 94,817 60,706 -34,111 -36% +/-25% Total 413 1,835,079 1,820,128 -14,951 -1% +/-5% P.M. Peak Period (3:00-7:00 p.m.) Freeway 70 1,419,427 1,482,604 63,177 4% +/-7% Major Arterial 220 625,265 571,109 -54,156 -9% +/-10% Minor Arterial 24 18,474 16,781 -1,693 -9% +/-15% Collector 99 130,253 89,697 -40,556 -31% +/-25% Total 413 2,193,419 2,160,190 -33,229 -2% +/-5% Daily Freeway 70 5,073,202 5,165,641 92,439 2% +/-7% Major Arterial 220 2,074,574 1,552,334 -522,240 -25% +/-10% Minor Arterial 24 65,276 47,069 -18,207 -28% +/-15% Collector 99 444,775 251,565 -193,210 -43% +/-25% Total 413 7,657,827 7,016,610 -641,217 -8% +/-5% A.M. Peak Hour Freeway 70 748,073 817,582 69,509 9% +/-7% Major Arterial 220 1,371,778 1,282,162 -89,616 -7% +/-10% Minor Arterial 24 25,918 16,248 -9,670 -37% +/-15% Collector 99 214,635 147,637 -66,998 -31% +/-25% Total 413 2,360,404 2,263,629 -96,775 -4% +/-5% P.M. Peak Hour Freeway 70 696,995 712,376 15,381 2% +/-7% Major Arterial 220 1,538,022 1,426,325 -111,697 -7% +/-10% Minor Arterial 24 27,863 19,603 -8,260 -30% +/-15% Collector 99 233,015 175,026 -57,989 -25% +/-25% Total 413 2,495,895 2,333,331 -162,564 -7% +/-5%



The validation test of highway assignment by area type is presented in Table 8.3 by time period. This table shows that all area types are within the validation target except for rural daily trips and that there are no apparent biases in the model by area type. Rural facilities can benefit from additional local review of roadway attributes.

Table 8.3 Validation of Highway Assignment by Area Type and Time Period

Area Type

Number of

Counts 2000

Counts 2000


Difference Target A.M. Peak Period (6:00-10:00 a.m.) CBD/Urban Bus 31 135,015 129,270 -5,745 -4% +/-10% Urban 128 547,196 551,544 4,348 1% +/-10% Suburban 176 874,322 880,408 6,086 1% +/-10% Rural 78 278,546 258,905 -19,641 -7% +/-10% Total 413 1,835,079 1,820,128 -14,951 -1% +/-5% P.M. Peak Period (3:00-7:00 p.m.) CBD/Urban Bus 31 163,761 181,549 17,788 11% +/-10% Urban 128 656,786 621,900 -34,886 -5% +/-10% Suburban 176 1,056,104 1,060,540 4,436 0% +/-10% Rural 78 316,768 296,202 -20,566 -6% +/-10% Total 413 2,193,419 2,160,190 -33,229 -2% +/-5% Daily CBD/Urban Bus 31 589,692 558,683 -31,009 -5% +/-10% Urban 128 2,305,501 2,121,327 -184,174 -8% +/-10% Suburban 176 3,634,007 3,368,267 -265,740 -7% +/-10% Rural 78 1,128,627 968,333 -160,294 -14% +/-10% Total 413 7,657,827 7,016,610 -641,217 -8% +/-5% A.M. Peak Hour CBD/Urban Bus 31 573,964 537,562 -36,402 -6% +/-10% Urban 128 420,482 408,199 -12,283 -3% +/-10% Suburban 176 930,317 878,701 -51,616 -6% +/-10% Rural 78 435,641 439,167 3,526 1% +/-10% Total 413 2,360,404 2,263,629 -96,775 -4% +/-5% P.M. Peak Hour CBD/Urban Bus 31 663,900 591,828 -72,072 -11% +/-10% Urban 128 431,861 409,563 -22,298 -5% +/-10% Suburban 176 944,984 883,364 -61,620 -7% +/-10% Rural 78 455,150 448,576 -6,574 -1% +/-10% Total 413 2,495,895 2,333,331 -162,564 -7% +/-5%

The third validation test for highway assignment is to compare screenline traffic counts against model volumes. This validation test has been in use in Contra Costa for many years and as such, the screenlines are actively used to assess the reliability of the model in any particular local area. Figures 8.1 through 8.4 present the locations of screenlines in the East County, Central County, West County, and Tri-Valley subareas, respectively.



Figure 8.1 East County Screenlines



Figure 8.2 Central County Screenlines



Figure 8.3 West County Screenlines



Figure 8.4 Tri-Valley Screenlines

The validation test is on the summation of all links in a screenline for the a.m. and p.m. peak hours, presented in Figures 8.5 and 8.6, respectively. In the a.m. peak hour, there is only one screenline that is well outside the maximum desirable deviation and that is the Richmond/San Pablo screenline (Number I-15). I-580 is not included in this screenline and the east-west movements are dominated by the parallel traffic on I-580, which is slightly over-estimated. If I-580 were included in the screenline, the peak direction of travel for this screenline is under-estimated by 12 percent. In the p.m. peak hour, there is only one screenline that is outside the maximum desirable deviation and that is the Richmond/San Pablo screenline (same as in the a.m. peak hour). The screenline summa-ries for the a.m. and p.m. peak hours are provided in Table 8.4.



Figure 8.5 AM Peak Hour Screenline Validation

0

5

10

15

20

25

30

35

40

– 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000

AM Peak Hour Counts

Percent Deviation from Counts

Internal ScreenlinesRegional ScreenlinesMaximum Desirable Deviation

Figure 8.6 PM Peak Hour Screenline Validation

0

5

10

15

20

25

30

35

40

45

– 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 100,000

AM Peak Hour Counts

Percent Deviation from Counts

Internal ScreenlinesRegional ScreenlinesMaximum Desirable Deviation



Table 8.4 Validation of Highway Assignment by Screenline for A.M. Peak and P.M. Peak Hour

Screenline AM Peak Hour PM Peak Hour

No. Name 2000

Count 2000

Model Percent

Difference 2000

Count 2000

Model Percent

Difference Internal Screenlines Central I1 SR 4 23,130 24,715 6.9% 24,701 28,693 16.2% Central I2 Concord 29,027 28,367 -2.3% 31,414 32,477 3.4% Central I3 Orinda 15,167 19,995 31.8% 15,887 19,239 21.1% Central I4 I-680 39,372 35,780 -9.1% 43,621 40,581 -7.0% Central I5 Treat 34,701 30,801 -11.2% 36,840 35,688 -3.1% Central I6 Ygnacio 29,072 25,831 -11.1% 30,089 29,692 -1.3% Central I7 SR24 4,959 4,930 -0.6% 5,716 5,625 -1.6% Central I8 Walnut Creek 26,865 28,698 6.8% 29,267 33,929 15.9% Tri-Valley I9 San Ramon 14,778 15,506 4.9% 15,941 18,223 14.3% Tri-Valley I10 Danville 7,002 4,988 -28.8% 7,241 6,166 -14.9% Tri-Valley I11 Danville 3,564 2,819 -20.9% 3,205 3,956 23.4% East I12 Antioch/Brentwd 6,727 5,748 -14.6% 7,918 6,821 -13.9% East I13 Oakley/Brentwd 5,582 6,358 13.9% 7,004 7,152 2.1% West I14 Richmond 21,174 24,121 13.9% 21,518 24,443 13.6% West I15 Rich/Sanpb 10,869 6,970 -35.9% 14,029 8,361 -40.4% Tri-Valley I16 I-580 23,939 21,496 -10.2% 26,507 25,237 -4.8% Tri-Valley I17 West Livermore 18,533 21,038 13.5% 21,015 22,426 6.7% West I18 Pinole/County 20,699 22,847 10.4% 21,191 25,533 20.5% Total 335,160 331,007 -1.2% 363,104 374,239 3.1% Regional Screenlines Cordon Line 84,753 88,104 4.0% 91,783 96,790 5.5% West R1 West/Central 5,743 5,886 2.5% 6,090 6,089 0.0% Central R2 Lamorinda 22,237 22,150 -0.4% 23,821 22,897 -3.9% Central R3 TriValley 16,820 19,147 13.8% 17,987 21,811 21.3% East R4 Central/East 16,870 16,969 0.6% 17,266 19,011 10.1% Central R5 S.C Central 6,627 6,951 4.9% 7,408 7,830 5.7% East R6 S.C East 12,553 13,863 10.4% 14,355 15,476 7.8% Tri-Valley R7 S.C Tri Valley 13,348 15,069 12.9% 15,780 16,771 6.3% West R8 S.C West 16,525 17,333 4.9% 19,700 19,273 -2.2% Tri-Valley R9 Alameda County 22,611 19,722 -12.8% 18,731 17,946 -4.2% Tri-Valley R10 Sunol 10,489 11,635 10.9% 12,429 12,983 4.5% Tri-Valley R11 Greenville 12,029 11,129 -7.5% 12,300 12,262 -0.3% Total 240,605 247,959 3.1% 257,649 269,137 4.5% Total by Subarea East 41,732 42,938 3% 46,543 48,460 4% Central 247,977 247,366 0% 266,751 278,460 4% West 75,010 77,156 3% 82,528 83,698 1% Tri-Valley 126,293 123,403 -2% 133,148 135,969 2% Cordon Line 84,753 88,104 4% 91,783 96,790 5% Grand Total (Regional + Internal) 575,765 578,966 0.6% 632,241 643,376 1.8%



Screenline summaries for a.m. peak period, p.m. peak period, and daily assignments are presented in Table 8.5. These are reported for information only, since the validation tar-gets were set for peak hour assignments only. Detailed screenline summaries by direction and link are available upon request from the CCTA.

Screenlines are also summarized by subarea in Tables 8.4 and 8.5 to indicate whether there are any biases in the model in these areas. All time periods, except average daily traffic, are within +/- six percent by subarea. The difference in average daily traffic is primarily the result of the off-peak traffic assignment, which under-estimated counts overall by eight percent.



Table 8.5 Validation of Highway Assignment by Screenline for A.M. Peak Period, P.M. Peak Period, and Average Daily Traffic

Screenline AM Peak Period PM Peak Period Daily ADT

No. Name 2000

Count 2000

Model Percent

Difference 2000

Count 2000

Model Percent

Difference 2000

Count 2000

Model Percent

Difference Internal Screenlines Central I1 SR 4 75,278 80,971 8% 91,949 101,338 10% 334,193 317,617 -5% Central I2 Concord 93,450 86,692 -7% 110,895 107,605 -3% 357,052 330,611 -7% Central I3 Orinda 54,210 62,898 16% 58,449 62,746 7% 212,515 218,530 3% Central I4 I-680 127,609 108,249 -15% 160,541 131,511 -18% 513,032 424,711 -17% Central I5 Treat 109,006 97,043 -11% 132,325 120,825 -9% 462,630 382,087 -17% Central I6 Ygnacio 90,074 81,578 -9% 110,765 98,950 -11% 390,740 326,424 -16% Central I7 SR24 14,939 15,881 6% 19,996 19,877 -1% 69,723 63,954 -8% Central I8 Walnut Creek 89,067 88,824 0% 107,451 113,800 6% 341,254 351,994 3% Tri-Valley I9 San Ramon 46,979 44,654 -5% 55,624 58,212 5% 190,782 190,113 0% Tri-Valley I10 Danville(NB/SB) 19,243 15,807 -18% 25,205 21,465 -15% 83,278 64,990 -22% Tri-Valley I11 Danville (EB/WB) 8,995 8,467 -6% 11,151 12,843 15% 35,532 36,984 4% East I12 Antioch/Brentwoo

d 21,955 18,157 -17% 27,383 23,265 -15% 99,570 69,866 -30%

East I13 Oakley/Brentwood 18,594 18,972 2% 25,343 23,728 -6% 89,795 71,371 -21% West I14 Richmond 69,373 79,820 15% 79,695 87,185 9% 303,129 304,555 0% West I15 Rich/Sanpb 33,192 25,298 -24% 50,844 33,794 -34% 180,287 98,068 -46% Tri-Valley I16 I-580 76,410 61,697 -19% 91,773 78,938 -14% 317,681 255,416 -20% Tri-Valley I17 West Livermore 64,555 66,400 3% 76,064 73,782 -3% 269,526 244,464 -9% West I18 Pinole/County 71,449 71,789 0% 87,344 85,703 -2% 295,303 275,025 -7% Total 1,084,378 1,033,197 -5% 1,322,797 1,255,566 -5% 4,546,022 4,026,780 -11%



Table 8.6 Validation of Highway Assignment by Screenline for A.M. Peak Period, P.M. Peak Period, and Average Daily Traffic

Screenline AM Peak Period PM Peak Period Daily ADT

No. Name 2000

Count 2000

Model Percent

Difference 2000

Count 2000

Model Percent

Difference 2000

Count 2000

Model Percent

Difference Regional Screenlines Cordon Line 281,316 286,973 2% 322,052 332,739 3% 1,158,670 1,119,805 -3% West R1 West/Central 17,124 15,718 -8% 19,639 17,657 -10% 62,278 51,959 -17% Central R2 Lamorinda 68,161 68,633 1% 82,710 77,932 -6% 248,051 256,992 4% Central R3 TriValley 53,274 56,982 7% 64,052 71,762 12% 222,621 224,292 1% East R4 Central/East 53,004 50,653 -4% 60,598 61,559 2% 214,751 189,573 -12% Central R5 S.C Central 20,048 23,127 15% 25,180 21,474 -15% 85,374 70,912 -17% East R6 S.C East 39,746 41,776 5% 51,652 50,951 -1% 179,129 153,146 -15% Tri-Valley R7 S.C Tri Valley 44,875 46,488 4% 54,722 55,348 1% 191,527 185,038 -3% West R8 S.C West 51,145 58,946 15% 61,764 71,443 16% 192,191 227,423 18% Tri-Valley R9 Alameda County 73,343 66,748 -9% 67,849 60,581 -11% 264,964 227,185 -14% Tri-Valley R10 Sunol 32,584 34,420 6% 42,526 44,088 4% 147,511 150,328 2% Tri-Valley R11 Greenville 34,976 34,284 -2% 41,957 39,090 -7% 144,736 133,176 -8% Total 769,596 784,748 2% 894,701 904,625 1% 3,111,803 2,989,829 -4% Total by Subarea East 133,299 129,557 -3% 164,976 159,503 -3% 583,245 483,956 -17% Central 795,116 770,878 -3% 964,313 927,819 -4% 3,237,185 2,968,125 -8% West 242,283 251,571 4% 299,286 295,782 -1% 1,033,188 957,029 -7% Tri-Valley 401,960 378,966 -6% 466,871 444,348 -5% 1,645,537 1,487,694 -10% Cordon Line 281,316 286,973 2% 322,052 332,739 3% 1,158,670 1,119,805 -3% Grand Total (Regional + Internal) 1,853,974 1,817,945 -2% 2,217,498 2,160,190 -3% 7,657,825 7,016,610 -8%



8.3 Transit Assignments

Transit assignments were validated by route group and mode for all transit routes within the CCTA study area and for the BART system as a whole. These results are presented in Table 8.7. There are four route groups in the study area: regional BART (all BART lines), local BART (all BART stations within the study area), express bus, and local bus. Table 8.7 presents the individual operators in each of the route groups and presents the subtotals by route group, with a validation target of +/-20 percent. Express bus is the only route group outside the target (at 21 percent over-estimated), but this group only contains four bus routes and this over-estimation only represents 1,252 riders. Validation targets by mode are +/-10 percent and local BART stations meet this target, but all bus routes are 12 percent over-estimated. Most of this over-estimation is caused by the midday service of a single CCCTA local bus route, which can be corrected during local area studies.

8.4 Link-Based Highway Validation Tests

A series of link-based highway validation tests were set up to ensure accurate link-based traffic volumes by time period for higher level facilities (freeways and arterials). The link-based validation tests by time period are presented in Table 8.8. Of these 12 validation tests in the peak period and daily assignments, all but one (arterials with 10,000 or more vehicles within +/-30 percent) meets the established targets.

For the a.m. and p.m. peak hours, the model meets most of the freeway and arterial vali-dation tests but not the intersection level tests. This is a primarily a result of the fact that the models were validated to match higher level facilities (such as freeways and arterials) that were based on screenline counts, all collected at approximately the same time using the same methods, and the intersection validation tests are based on intersection counts, collected at different times by different agencies. Comparisons of these two datasets (in Section 2.0) reveal significant differences in the counts for the same location. As a result, the intersection validation tests are very difficult to match, since they are based on a set of counts that is inconsistent with the screenline counts.



Table 8.7 Transit Assignment Validation Summary

Route Group Operator 1998

Count 2000


Difference Target Regional BART Pittsburg/Bay

Point to Colma 97,728 77,145 -20,583 -21%

Regional BART Richmond to Colma

70,000 58,801 -11,199 -16%

Regional BART Fremont to Colma 47,914 41,637 -6,277 -13% Regional BART Dublin/Pleasanton

to Colma 40,611 40,103 -508 -1%

Regional BART Richmond to Fremont

29,000 25,225 -3,775 -13%

Local BART BART in CCTA Study Area

30,795 33,446 2,651 9%

Express Bus AC EXP 3,847 3,785 -62 -2% Express Bus CCCTA Exp 2,018 3,332 1,314 65% Local Bus LAVTA 1,595 1,972 377 24% Local Bus AC LOCAL 15,408 13,705 -1,703 -11% Local Bus CCCTA Local 13,398 17,329 3,931 29% Local Bus Tri - Delta 3,941 5,396 1,455 37% Local Bus WestCat 3,068 2,875 -193 -6% Subtotal by Route Group Regional BART 285,253 242,911 -42,342 -15% +/-20% Local BART 30,795 33,446 2,651 9% +/-20% Local Bus 37,410 41,277 3,867 10% +/-20% Express Bus 5,865 7,117 1,252 21% +/-20% Subtotal by Mode BART 30,795 33,446 2,651 9% +/-10% All Bus 43,275 48,394 5,119 12% +/-10% Total 74,070 81,840 7,770 10% +/-10%



Table 8.8 Link-Based Validation Tests by Time Period

Facility Type and Criteria Number of

Counts

Links Meeting Target

Percent of Links Within

Target Validation

Target A.M. Peak Period (6:00-10:00am) Freeway Links within 20% 70 61 87% 75% Freeway Links within 10% 70 54 77% 50% Arterials with 10,000+ vehicles within 30% 88 45 51% 75% Arterials with 10,000+ vehicles within 15% 88 45 51% 50% P.M. Peak Period (3:00-7:00pm) Freeway Links within 20% 70 56 80% 75% Freeway Links within 10% 70 44 63% 50% Arterials with 10,000+ vehicles within 30% 88 72 82% 75% Arterials with 10,000+ vehicles within 15% 88 50 57% 50% Daily Period Freeway Links within 20% 70 66 94% 75% Freeway Links within 10% 70 55 79% 50% Arterials with 10,000+ vehicles within 30% 88 77 88% 75% Arterials with 10,000+ vehicles within 15% 88 47 53% 50% A.M. Peak Hour Freeway Links within 20% 70 58 83% 50% Freeway Links within 10% 70 45 64% 75% Arterials with 10,000+ ADT within 30% 88 52 59% 50% Arterials with 10,000+ ADT within 15% 88 24 27% 50% Intersections with 1000+ Vehicles per hr within 20% 366 139 38% 50% Intersections with 500-1000 Vehicles per hr within 20% 358 108 30% 30% All Intersections within 30% of Counts 922 415 45% 75% All Intersections within 15% of Counts 922 187 20% 50% 80% of Freeway Counts below the Curve 114 75 66% 80% 80% of Ramp Counts below the Curve 451 298 66% 80% P.M. Peak Hour Freeway Links within 20% 70 55 79% 50% Freeway Links within 10% 70 42 60% 75% Arterials with 10,000+ vehicles within 30% 88 56 64% 50% Arterials with 10,000+ vehicles within 15% 88 42 48% 50% Intersections with 1000+ Vehicles per hr within 20% 427 199 47% 50% Intersections with 500-1000 Vehicles per hr within 20% 310 113 36% 30% All Intersections within 30% of Counts 922 475 52% 75% All Intersections within 15% of Counts 922 278 30% 50% 80% of Freeway Counts below the Curve 114 85 75% 80% 80% of Ramp Counts below the Curve 451 275 61% 80%



9.0 Forecasts

9.1 Introduction

A key function of the travel demand forecasting model is to generate long-range traffic forecasts that can be used in a variety of applications. The decennial model update uses ABAG Projections 2000 and 2002 land use information to generate new traffic forecasts for near-term (10-year) and long-range (25-year) planning horizons. The model is capable of forecasting still further into the future, provided that an appropriate land use data set is available.4 This section documents the approach, methodology, and results of the future year forecasting effort.

9.2 Future Scenarios

The CCTA model was run for the year 2000 calibration scenario, and five future year sce-narios (see Table 9.1).

Table 9.1 Future Scenarios

Scenario Socioeconomic

Data (Year) Network Improvements

1. Existing 2000 Existing conditions as of 2000/2001, including recently completed major highway projects

2. TIP 2010 The 2002 RTIP/STIP projects

3. CMP* 2010 Authority’s 2003 CMP/CIP

4. RTP #1 2020 2020 MTC 2001 RTP Update Track 1 Improvements.

5. RTP #1 2025 2025 Same as Scenario 4

6. RTP #2 2025 MTC’s Blueprint; the RTP Track 2, plus selected projects from the Authority’s CTPL

* Note that the 2003 CMP scenario has been postponed pending completion of an updated 2003 CMP/CIP pro-ject list.

4 ABAG Projections 2003 is expected to provide a 2030 horizon year.



9.3 Land Use Data

The future year socioeconomic data was generated based upon forecasts developed by ABAG.

Almost all of the forecast socioeconomic input values for the model forecasts were con-tained in the LUIS for the CCTA model area. For the MTC region, the forecasted inputs were obtained from the MTC web site.

Forecasts of population, household, income, and employment data by zone were devel-oped by EPS for inside Contra Costa and Tri-Valley. Forecasts of this data for the rest of the MTC region were taken directly from MTC forecast files for the region.

Residential, commercial, industrial and agricultural acres in a zone were not present in the LUIS database. They were calculated based on their percentage of total acres in the origi-nal MTC zones.

Future-year school enrollment for existing school facilities is not forecasted by MTC or ABAG, and is not readily available from the school districts. In general, enrollment is pre-dicted to remain flat for existing facilities. Increased enrollment is reflected primarily in the construction of new facilities, some of which have been identified for this forecasting effort. Therefore, with the exception of new facilities that were identified, future year enrollment remained at the year 2000 level for existing schools. The model can be adjusted in the future to evaluate the impacts of higher school enrollment should that data become available.

One area where future school locations were identified was in the Dougherty Valley. Enrollment for each new school was estimated using the table below (Table 9.2).5

Table 9.2 Future Schools in Dougherty Valley

New School TAZ Enrollment Year

High School 40196 2,000 2007

Elementary School 40193 700 2010

Middle School 40191 1,150 2005

5 Source: John Cunningham, Contra Costa County Community Development Department,

Martinez, California.



Total acreage for each zone was assumed to be unchanged in future years. It would only change if the zone boundaries were changed.

Auto zonal level of service variables contain information on the parking costs and terminal times for each zone in the CCTA model. This information was obtained from the MTC forecast files for both the base and the forecast years.

9.4 Special Generator Data

The CCTA model special generators include hospitals, military bases, shopping malls, retirement homes and regional parks. They are listed in Table 9.3. The special generator trips and employees were calculated based on the information given in Table 9.4. Note that all the special generators in the base year remained at the same levels for all of the forecast years (no increase in employment or intensity). The Deer Valley Medical Center is the only new special generator included in the forecast years.

Table 9.3 Special Generators

Name

Special Generator

Trips


Remaining TAZ

Employees

Type of TAZ

Employees Year

Added 1 Contra Costa Regional

Medical Center 3,954 373 738 Service 2000

2 John Muir Medical Center 5,518 731 702 Service 2000 3 Kaiser Permanente Walnut

Creek Medical Center 4,602 521 445 Service 2000


3,745 326 185 Service 2000

6 Doctors Medical Center 4,641 530 550 Service 2000 7 Mt. Diablo Medical Center 4,901 590 710 Service 2000 8 Sutter Delta Medical Center 3,416 250 39 Service 2000 9 Valley Care Health System 3,735 323 1,692 Service 2000 10 Broadway Plaza/Downtown

Walnut Creek 31,725 1,076 546 Retail 2000

11 Sun Valley Mall 51,137 2,261 2,212 Retail 2000 12 County East Mall 27,217 848 -848 Retail 2000 13 Hilltop Mall 56,896 2,669 -1,709 Retail 2000 14 Stoneridge Mall 47,526 2,017 -1,837 Retail 2000 15 Parks Reserve Forces Training

Area 5,395 741 -589 Service 2000

16 Rossmoor Retirement Home 18,576 8,000 0 Residents 2000 17 Deer Valley Medical Center 4,195 811 0 Service 2010

Note: All the special generators in the base year remain the same in all the forecast years. The only new spe-cial generator in the forecast years is the Deer Valley Medical Center.



Table 9.4 Assumptions for Special Generator Categories

Trips Per

Bed Employees Per 1,000 SF

Trips Per Employee

ITE Code 610 for Hospitals

Average Size

Hospitals 11.77 16.78 5.17 Trips = 7.381*Beds + 1,718.324 392 beds

Trips Per

TSF Employees

Per TSF1 Trips Per Employee

ITE Code 820 for Shopping Centers

Average Size

Shopping Centers

42.92 1.57 27.34 Ln(Trips) = 0.643 Ln(x) +5.866 331 TSF

Trips Per Acre2

Employees Per Acre3

Trips Per Employee

ITE Code 417 for Regional Parks

Average Size

Regional Parks

0.0808 0.001013 79.77 79.77 person trips per employee 310 acres

Trips Per Employee

ITE Code 510 for Military Bases

Average Size

Military Bases

7.28 Ln(Trips) = 0.571* Ln(x) +4.520 7,747

Trips Per Unit

ITE Code 252 for Congregate Care Facility

Average Size

Retirement Homes

2.15 2.15 trips per dwelling unit 183

Trips Per Employee

ITE Code 760 for Research and Development Center

Average Size

Unique Industry

2.77 Ln(Trips) = 0.800* Ln(x) +2.418 1,022

Source: Institute of Transportation Engineers, Trip Generation User’s Guide, 6th Edition, Publication Number IR-016D, 1997.

1 ITE did not provide an estimate of employees per TSF, so this was derived from a study in California, which can be found at http://www.ceres.ca.gov/ceqa/cases/1991/shapell_112291.html.

2 ITE trips per acre (4.57) results in too many trips. We calculated trips to Mt. Diablo based on total employees in the TAZ and divided it by the acres to get a more realistic ‘trips per acre’ value.

3 Employees per acre was derived by using the calculated ‘trips per acre’ value and ITE code 417 for regional park.



9.5 Supplementary Data Files

The CCTA requires additional information beyond that contained in the Land Use Information System (LUIS) in order to perform forecasts. Most of this supplementary data was obtained from MTC files.

External trip forecasts (trips coming from or going to points outside of the MTC nine-County region) were taken directly from the Metropolitan Transportation Commission for 2000 and 2025. The 2025 external trips were used for all of the forecast years (2010 and on).

Full documentation of the external trip assumptions for each forecast year is provided in Appendix K. The table summarizes the assumed vehicle trip levels for each forecast year, by county of origin and destination.

The model requires calibration files (dhbwg251.dat, dbhwa251.dat ) used in trip generation and ‘delta’ files (HBSMC.DLT, HBSRMC.DLT and NHBMC.DLT) used in Model Split. These files were obtained from MTC for 2025 and used for all forecast years (2010 to 2025).

The intrazonal daily peak and off-peak trip input matrices were obtained from MTC for the year 2025. These files were used for all forecast years (2010 to 2025).

All of the MTC files were converted from the MTC 1099 zone system to the Countywide 2,600-zone system for the year 2025 using files from the MTC 2025 model.

The transit stop fare and link fare files for 2000 and 2025 were obtained from MTC. The MTC 2000 transit network is used for both the CCTA 2000 and 2010 future scenarios; con-sequently the year 2000 MTC fares are used for both those years in the CCTA model. Similarly, since the MTC 2025 transit route system is used for the CCTA forecast years 2020 and 2025, the corresponding transit files (e.g., stopfare, linkfare, modes, modefactors and farematrix) are used for the 2020 and 2025 CCTA model forecasts to avoid errors due to differences in route systems between years.

9.6 Highway Network Data

The year 2000 highway network was created as part of the model calibration and valida-tion process. The future year network builds upon the 2000 network using a “master net-work” approach. New links, link deletions, and link edits are specified for each future scenario using one large data base that includes all link data for all scenarios.

To generate the future network scenarios, two sources were consulted (see Table 9.5). Inside Contra Costa and the Tri-Valley, the future network improvements for each sce-nario were created from the project list provided by CCTA staff (see Table 9.5 at end of document). Outside of Contra Costa and the Tri-Valley, the appropriate future year high-way network from MTC’ 2001 RTP Update was used. Appendix I shows how this worked



for each scenario. Appendix I lists the highway and transit projects included in each future scenario.

Table 9.5 Sources of Each Network Scenario

Scenario Year Description Inside

Study Area Outside

Study Area

1. Existing 2000 Existing conditions as of 2000/2001, including recently completed major highway projects (MTC 2000 outside study area)

CCTA 2000 MTC 2000

2. TIP 2010 Network Scenario 1 + Year 2000 RTIP/STIP projects (MTC 2010 outside study area)

CCTA 2000 + Add from Table

MTC 2010 TIP network for highway, 2000 network for transit

3. CMP 2010 Network Scenario 2 + Authority’s 2001 CMP CIP (MTC 2010 outside study area)

CCTA TIP + 2003 CMP CIP Update*)

MTC 2010 TIP network for highway, 2000 for transit

4. RTP #1 2020

2020 Network Scenario 2 + Track 1 RTP consistent with the 2001 RTP Update (MTC 2025 RTP Track 1 outside study area).

CCTA Track 1 as specified in CCTA database

MTC RTP Track 1 network

5. RTP #1 2025

2025 Same as Network Scenario 4 Scenario#4 Scenario #4

6. RTP #2 2025 Network Scenario 4 + selected projects from the CTPL based upon MTC’s Blueprint (MTC 2025 Blueprint outside study area)

CCTA CTPL MTC RTP Blueprint Network for highway, Track 1 for transit.

*To be completed. Use 2025 CTPL in the interim.

9.7 Transit Network Data

The future year transit network modifications to the master transit network were created in much the same manner as for the future highway networks. Two sources were used as detailed in Table 9.5

Future routes information was provided by transit agencies as part of the review. The future year transit network is financially constrained, and therefore only contains improvements as specified by CCTA staff (See Appendix I) in addition to the base year transit network. Major rail improvements, such as e-BART to Antioch, and BART to San



Jose, were included in the future year network, while very few changes were made to future bus transit routes. The model can be modified to reflect increased bus service, including express bus, should that option become viable.

9.8 Future Intersection Geometry

The model is capable of generating future year intersection level of service based upon a data set of future-year intersection geometry. This information is subject to further analy-sis and review at the subarea and local level. Consequently, intersection LOS reports were not generated as part of this initial effort.

9.9 Forecast Results

The forecasted county-to-county vehicle trip tables for each scenario are shown in Table 9.6 through 9.10. The forecasted a.m. peak hour, p.m. peak hour, a.m. peak-period, p.m. peak-period, and daily traffic volumes (drive alone plus shared ride) are shown in Tables 9.11 through 9.15.

Analysis of Trip Tables

• Total a.m. peak-hour vehicle trips in the Bay Area are projected to increase by about 13 percent by the year 2010, 28 percent by the year 2020, and 32 percent (over 2000) by 2025.

• The MTC Track 2 investment program would trim the growth in peak-hour vehicle traffic by about five percent (the growth in 2025 would be 27 percent, instead of 32 percent over the year 2000).

• The a.m. peak-hour traffic to/from and within Contra Costa County is projected to increase 17 percent by 2010, 34 percent by 2020, and 41 percent by 2025. The MTC Track 2 program would trim two percent off of the projected growth for 2025.



Table 9.6 AM Peak-Hour County-to-County Trip Table Forecasts

San

Francisco San

Mateo Santa Clara Alameda

Contra Costa Solano Napa Sonoma Marin Total

2000 San Francisco 46,910 12,802 2,900 4,860 2,240 398 107 299 1,575 72,091 San Mateo 19,201 75,134 16,467 4,736 1,366 116 26 67 402 117,516 Santa Clara 1,935 10,902 256,856 8,292 1,325 106 24 50 115 279,604 Alameda 6,305 6,931 14,962 135,158 14,897 919 209 298 908 180,588 Contra Costa 4,666 1,684 1,382 19,066 93,929 3,388 462 414 1,338 126,327 Solano 2,345 552 389 2,212 4,517 31,521 1,446 335 416 43,733 Napa 273 87 91 494 683 1,149 12,983 1,152 148 17,059 Sonoma 1,973 526 193 472 433 416 1,442 52,063 4,365 61,882 Marin 6,863 924 254 958 740 274 109 1,423 26,801 38,345 Total 90,469 109,542 293,494 176,249 120,129 38,287 16,807 56,101 36,068 937,146 2010 STIP San Francisco 44,365 13,607 3,063 5,292 2,331 478 153 466 1,832 71,588 San Mateo 20,665 81,104 18,008 5,302 1,348 144 33 92 490 127,185 Santa Clara 2,147 12,235 288,927 9,515 1,086 133 34 70 144 314,291 Alameda 6,557 7,460 16,379 152,649 15,229 1,194 330 565 1,179 201,544 Contra Costa 5,008 1,883 1,679 23,868 110,657 4,421 742 743 1,666 150,668 Solano 2,459 591 430 2,625 5,195 40,247 2,479 490 495 55,012 Napa 307 85 131 572 747 1,399 15,575 1,500 163 20,477 Sonoma 1,959 458 230 547 425 475 1,781 65,648 4,315 75,839 Marin 7,093 972 268 1,079 804 347 159 1,971 29,760 42,453 Total 90,560 118,394 329,115 201,449 137,822 48,840 21,286 71,546 40,045 1,059,058 Growth 13% 2020 Track 1 San Francisco 46,435 15,112 3,402 5,852 2,680 549 173 513 2,028 76,745 San Mateo 24,866 88,960 20,093 5,969 1,473 166 39 102 556 142,224 Santa Clara 2,779 13,847 325,898 10,753 1,146 152 39 80 167 354,860 Alameda 9,399 8,649 18,402 172,191 17,692 1,414 397 651 1,400 230,194 Contra Costa 6,084 2,270 1,988 27,354 125,974 5,183 888 858 1,936 172,535 Solano 2,852 753 517 2,868 6,022 48,844 3,246 621 620 66,345 Napa 310 98 151 589 825 1,778 18,065 1,917 202 23,935 Sonoma 2,151 475 252 510 409 524 1,885 75,647 4,825 86,677 Marin 8,048 1,074 301 1,154 869 397 189 2,357 32,242 46,631 Total 102,924 131,238 371,003 227,242 157,091 59,007 24,920 82,746 43,975 1,200,146 Growth 28% 2025 Track 1 San Francisco 45,393 15,399 3,650 5,743 2,595 835 203 851 2,011 76,679 San Mateo 24,592 89,413 20,006 5,921 1,353 200 44 171 603 142,303 Santa Clara 2,623 13,806 329,789 11,384 1,033 156 49 136 188 359,164 Alameda 9,966 9,922 20,242 176,840 17,949 1,621 452 930 1,372 239,294 Contra Costa 6,459 2,598 2,507 28,247 131,948 6,131 1,054 1,335 1,956 182,235 Solano 3,713 1,014 577 3,351 7,635 52,049 3,635 756 868 73,599 Napa 400 136 164 669 973 1,770 18,877 2,164 263 25,417 Sonoma 2,832 732 333 616 507 490 1,802 78,680 6,101 92,093 Marin 6,751 997 311 976 714 395 179 3,124 31,361 44,809 Total 102,729 134,018 377,577 233,747 164,708 63,648 26,295 88,147 44,724 1,235,593 Growth 32% 2025 Track 2 San Francisco 43,818 14,547 3,380 5,280 2,330 821 202 803 1,940 73,121 San Mateo 21,495 87,710 19,516 5,664 1,297 198 44 170 597 136,691 Santa Clara 1,984 13,167 315,215 10,886 1,026 156 49 136 186 342,804 Alameda 7,416 9,607 19,445 170,998 16,796 1,604 451 920 1,358 228,595 Contra Costa 5,668 2,475 2,479 27,532 131,548 6,114 1,053 1,330 1,947 180,145 Solano 3,535 947 575 3,247 7,610 51,879 3,621 755 867 73,036 Napa 396 136 164 664 970 1,752 18,739 2,164 263 25,249 Sonoma 2,331 722 331 586 506 490 1,802 78,461 6,048 91,276 Marin 6,188 984 308 946 706 395 179 3,066 31,017 43,788 Total 92,830 130,294 361,413 225,802 162,790 63,408 26,142 87,805 44,223 1,194,706 Growth 27%



Table 9.7 PM Peak-Hour County-to-County Trip Table Forecasts

San

Francisco San



2000 San Francisco 67,326 22,952 2,430 8,406 4,386 2,915 355 1,735 6,456 116,962 San Mateo 18,926 108,077 15,162 7,816 1,772 527 91 455 916 153,741 Santa Clara 3,352 19,248 345,194 16,042 2,848 372 69 175 253 387,553 Alameda 5,086 5,028 9,333 180,573 19,371 3,016 654 335 559 223,954 Contra Costa 1,755 1,036 1,828 13,862 138,116 6,328 928 350 475 164,679 Solano 731 143 116 1,332 4,854 45,773 1,080 428 290 54,745 Napa 146 30 25 313 777 1,314 18,170 1,775 114 22,663 Sonoma 520 104 57 308 520 389 1,592 69,423 1,625 74,538 Marin 2,665 519 146 612 892 483 168 3,844 39,225 48,555 Total 100,508 157,136 374,292 229,264 173,535 61,115 23,107 78,520 49,913 1,247,390 2010 STIP San Francisco 59,914 24,665 2,695 8,677 4,592 3,073 388 1,757 6,692 112,452 San Mateo 20,326 116,154 16,704 8,315 1,790 575 94 417 982 165,357 Santa Clara 3,573 21,027 391,546 17,381 2,709 410 88 191 272 437,197 Alameda 5,707 5,796 10,892 203,154 22,038 3,584 758 383 631 252,943 Contra Costa 1,948 1,062 1,555 14,675 162,299 7,426 1,050 376 517 190,909 Solano 841 172 144 1,586 5,945 57,785 1,362 521 368 68,722 Napa 196 38 35 430 1,050 2,188 21,895 2,253 162 28,245 Sonoma 710 121 71 450 662 544 1,997 89,537 2,118 96,211 Marin 2,830 606 175 760 1,046 587 195 3,958 43,393 53,550 Total 96,045 169,641 423,815 255,427 202,130 76,171 27,828 99,392 55,137 1,405,586 Growth 13% 2020 Track 1 San Francisco 62,995 28,993 3,338 11,281 5,370 3,597 419 1,896 7,502 125,392 San Mateo 22,490 126,215 18,529 9,402 2,060 712 109 426 1,072 181,014 Santa Clara 3,977 23,019 436,039 19,227 3,043 471 100 200 293 486,367 Alameda 6,554 6,511 12,260 225,663 24,868 3,972 816 362 674 281,680 Contra Costa 2,287 1,177 1,663 16,848 183,151 8,572 1,165 353 561 215,778 Solano 999 204 167 1,840 6,771 69,249 1,731 569 424 81,954 Napa 221 44 41 499 1,207 2,844 25,278 2,386 191 32,712 Sonoma 726 131 85 482 708 656 2,397 102,348 2,453 109,986 Marin 3,119 675 198 878 1,189 707 235 4,376 46,789 58,167 Total 103,368 186,968 472,319 286,121 228,368 90,781 32,250 112,916 59,959 1,573,050 Growth 26% 2025 Track 1 San Francisco 61,461 28,817 3,122 11,936 5,652 5,286 598 2,787 6,323 125,983 San Mateo 22,776 125,842 18,336 11,056 2,281 975 147 660 1,017 183,090 Santa Clara 4,081 22,800 438,844 21,705 3,550 523 108 261 300 492,172 Alameda 6,669 7,123 13,644 231,414 25,513 4,894 964 465 567 291,252 Contra Costa 2,330 1,138 1,521 17,014 190,676 11,447 1,462 473 461 226,521 Solano 1,954 297 178 2,382 8,507 74,149 1,769 548 563 90,347 Napa 324 58 50 603 1,457 3,173 26,640 2,310 218 34,832 Sonoma 1,357 232 137 697 1,051 763 2,586 106,955 3,803 117,580 Marin 2,894 713 215 848 1,174 1,048 319 6,149 44,393 57,751 Total 103,846 187,019 476,048 297,654 239,860 102,256 34,593 120,607 57,646 1,619,529 Growth 30% 2025 Track 2 San Francisco 59,078 25,883 2,571 9,709 5,092 5,073 593 2,358 5,844 116,199 San Mateo 21,526 123,887 17,787 10,788 2,200 923 147 651 1,007 178,918 Santa Clara 3,761 22,329 422,756 21,034 3,518 522 108 259 299 474,586 Alameda 5,987 6,896 13,151 225,422 24,830 4,770 959 446 551 283,012 Contra Costa 2,118 1,099 1,512 16,222 190,173 11,405 1,458 471 456 224,913 Solano 1,906 291 178 2,356 8,481 73,921 1,752 548 563 89,997 Napa 323 58 50 601 1,454 3,158 26,440 2,310 218 34,613 Sonoma 1,267 231 136 689 1,048 762 2,586 106,700 3,755 117,174 Marin 2,715 703 212 837 1,167 1,046 319 6,083 43,880 56,962 Total 98,680 181,376 458,354 287,660 237,964 101,580 34,362 119,827 56,572 1,576,374 Growth 26%



Table 9.8 AM Peak-Period County-to-County Trip Table Forecasts

San

Francisco San



2000 San Francisco 146,593 40,007 9,062 15,188 7,000 1,243 334 936 4,922 225,284 San Mateo 60,002 234,795 51,460 14,801 4,269 364 80 209 1,258 367,239 Santa Clara 6,046 34,068 802,676 25,912 4,142 330 75 156 360 873,764 Alameda 19,703 21,660 46,758 422,370 46,552 2,872 652 932 2,837 564,337 Contra Costa 14,580 5,261 4,318 59,582 293,528 10,588 1,443 1,293 4,180 394,773 Solano 7,327 1,726 1,217 6,913 14,115 98,503 4,519 1,046 1,300 136,666 Napa 853 271 284 1,543 2,134 3,591 40,572 3,599 463 53,309 Sonoma 6,164 1,645 602 1,474 1,352 1,299 4,507 162,697 13,640 193,381 Marin 21,448 2,886 793 2,994 2,313 856 340 4,446 83,754 119,829 Total 282,716 342,319 917,170 550,777 375,404 119,646 52,521 175,315 112,714 2,928,582 2010 STIP San Francisco 138,642 42,522 9,572 16,537 7,284 1,494 479 1,457 5,727 223,714 San Mateo 64,578 253,449 56,274 16,568 4,212 450 105 287 1,530 397,453 Santa Clara 6,709 38,235 902,896 29,733 3,393 417 105 220 451 982,158 Alameda 20,491 23,311 51,186 477,027 47,592 3,733 1,032 1,766 3,685 629,824 Contra Costa 15,650 5,886 5,248 74,587 345,804 13,817 2,317 2,323 5,207 470,839 Solano 7,684 1,846 1,344 8,203 16,235 125,773 7,747 1,533 1,548 171,913 Napa 960 264 408 1,788 2,333 4,373 48,671 4,687 508 63,992 Sonoma 6,123 1,430 719 1,710 1,328 1,485 5,567 205,150 13,485 236,996 Marin 22,164 3,039 837 3,373 2,512 1,085 498 6,159 93,000 132,667 Total 283,001 369,982 1,028,484 629,527 430,694 152,626 66,520 223,582 125,140 3,309,555 Growth 13% 2020 Track 1 San Francisco 145,110 47,225 10,632 18,288 8,376 1,715 542 1,603 6,337 239,828 San Mateo 77,705 278,001 62,792 18,653 4,604 519 121 319 1,738 444,451 Santa Clara 8,684 43,273 1,018,430 33,604 3,582 474 121 249 521 1,108,938 Alameda 29,371 27,027 57,505 538,096 55,288 4,420 1,240 2,034 4,375 719,355 Contra Costa 19,012 7,095 6,213 85,482 393,667 16,198 2,774 2,681 6,049 539,171 Solano 8,913 2,353 1,615 8,964 18,820 152,638 10,144 1,942 1,939 207,328 Napa 970 306 472 1,842 2,577 5,555 56,453 5,992 631 74,798 Sonoma 6,721 1,484 786 1,594 1,278 1,637 5,891 236,397 15,077 270,865 Marin 25,150 3,356 940 3,607 2,716 1,241 590 7,366 100,755 145,722 Total 321,636 410,120 1,159,385 710,130 490,909 184,397 77,876 258,582 137,421 3,750,456 Growth 28% 2025 Track 1 San Francisco 141,852 48,123 11,405 17,947 8,109 2,611 633 2,659 6,285 239,623 San Mateo 76,851 279,415 62,517 18,502 4,227 626 139 535 1,886 444,698 Santa Clara 8,196 43,144 1,030,591 35,574 3,228 488 154 426 586 1,122,388 Alameda 31,144 31,006 63,256 552,625 56,091 5,065 1,411 2,908 4,288 747,794 Contra Costa 20,186 8,119 7,834 88,272 412,339 19,158 3,293 4,172 6,111 569,483 Solano 11,604 3,169 1,803 10,473 23,860 162,654 11,360 2,362 2,713 229,998 Napa 1,250 425 513 2,092 3,039 5,532 58,990 6,762 823 79,427 Sonoma 8,849 2,287 1,040 1,926 1,586 1,530 5,631 245,874 19,065 287,789 Marin 21,098 3,117 970 3,048 2,232 1,236 561 9,762 98,004 140,028 Total 321,029 418,805 1,179,929 730,460 514,712 198,901 82,172 275,459 139,762 3,861,228 Growth 32% 2025 Track 2 San Francisco 136,932 45,458 10,563 16,499 7,282 2,565 632 2,509 6,062 228,503 San Mateo 67,170 274,093 60,986 17,700 4,054 620 139 532 1,866 427,161 Santa Clara 6,199 41,145 985,046 34,018 3,207 487 154 424 581 1,071,262 Alameda 23,176 30,022 60,766 534,368 52,486 5,012 1,410 2,875 4,245 714,358 Contra Costa 17,711 7,733 7,746 86,038 411,089 19,107 3,290 4,156 6,084 562,955 Solano 11,046 2,958 1,798 10,148 23,781 162,122 11,316 2,359 2,709 228,237 Napa 1,236 425 513 2,075 3,033 5,475 58,560 6,762 823 78,903 Sonoma 7,285 2,256 1,033 1,831 1,580 1,530 5,631 245,191 18,900 285,238 Marin 19,338 3,076 963 2,955 2,205 1,234 561 9,580 96,927 136,839 Total 290,094 407,168 1,129,415 705,631 508,717 198,151 81,693 274,389 138,197 3,733,456 Growth 27%



Table 9.9 PM Peak-Period County-to-County Trip Table Forecasts

San

Francisco San



2000 San Francisco 217,181 74,040 7,838 27,117 14,149 9,403 1,145 5,597 20,827 377,295 San Mateo 61,052 348,636 48,909 25,213 5,715 1,699 293 1,469 2,954 495,940 Santa Clara 10,812 62,091 1,113,530 51,748 9,188 1,199 223 564 818 1,250,171 Alameda 16,408 16,218 30,107 582,492 62,486 9,728 2,109 1,080 1,804 722,433 Contra Costa 5,663 3,341 5,898 44,717 445,537 20,414 2,993 1,129 1,533 531,224 Solano 2,358 460 374 4,296 15,658 147,654 3,484 1,380 934 176,597 Napa 471 95 82 1,009 2,505 4,238 58,612 5,726 367 73,106 Sonoma 1,679 336 184 993 1,677 1,253 5,137 223,945 5,240 240,444 Marin 8,596 1,674 471 1,975 2,877 1,557 542 12,401 126,533 156,628 Total 324,219 506,891 1,207,393 739,561 559,791 197,145 74,538 253,291 161,011 4,023,839 2010 STIP San Francisco 193,272 79,566 8,693 27,989 14,813 9,911 1,252 5,666 21,588 362,750 San Mateo 65,569 374,689 53,884 26,821 5,775 1,854 304 1,344 3,169 533,409 Santa Clara 11,526 67,829 1,263,051 56,069 8,737 1,322 285 616 877 1,410,313 Alameda 18,409 18,697 35,135 655,335 71,089 11,561 2,445 1,237 2,036 815,943 Contra Costa 6,284 3,425 5,016 47,340 523,546 23,956 3,387 1,214 1,669 615,836 Solano 2,711 554 465 5,115 19,176 186,403 4,393 1,680 1,187 221,685 Napa 632 121 112 1,386 3,387 7,058 70,629 7,267 522 91,114 Sonoma 2,290 392 228 1,452 2,136 1,756 6,442 288,828 6,833 310,356 Marin 9,129 1,955 563 2,451 3,373 1,893 630 12,768 139,979 172,741 Total 309,822 547,228 1,367,146 823,958 652,034 245,714 89,767 320,619 177,860 4,534,148 Growth 13% 2020 Track 1 San Francisco 203,211 93,527 10,769 36,389 17,322 11,604 1,352 6,117 24,200 404,491 San Mateo 72,547 407,144 59,770 30,330 6,644 2,297 351 1,373 3,459 583,915 Santa Clara 12,829 74,254 1,406,578 62,021 9,815 1,518 323 645 945 1,568,927 Alameda 21,141 21,002 39,548 727,946 80,219 12,814 2,631 1,169 2,175 908,645 Contra Costa 7,378 3,797 5,365 54,349 590,811 27,653 3,757 1,140 1,810 696,060 Solano 3,222 658 538 5,936 21,842 223,385 5,584 1,835 1,368 264,368 Napa 714 142 132 1,610 3,895 9,175 81,543 7,696 617 105,524 Sonoma 2,342 423 273 1,556 2,285 2,115 7,733 330,153 7,911 354,793 Marin 10,060 2,177 640 2,832 3,837 2,281 759 14,116 150,932 187,635 Total 333,445 603,124 1,523,612 922,970 736,670 292,841 104,033 364,245 193,418 5,074,357 Growth 26% 2025 Track 1 San Francisco 198,263 92,958 10,072 38,502 18,233 17,050 1,930 8,989 20,398 406,395 San Mateo 73,470 405,943 59,149 35,665 7,358 3,145 474 2,128 3,282 590,614 Santa Clara 13,164 73,549 1,415,627 70,016 11,453 1,686 350 841 969 1,587,653 Alameda 21,513 22,976 44,013 746,497 82,299 15,787 3,111 1,499 1,828 939,524 Contra Costa 7,517 3,671 4,907 54,882 615,084 36,926 4,715 1,525 1,486 730,713 Solano 6,303 957 575 7,683 27,443 239,189 5,706 1,769 1,817 291,442 Napa 1,047 186 161 1,944 4,699 10,234 85,934 7,453 705 112,363 Sonoma 4,376 749 442 2,248 3,389 2,460 8,341 345,017 12,269 379,292 Marin 9,335 2,299 693 2,735 3,787 3,379 1,029 19,834 143,202 186,293 Total 334,987 603,288 1,535,639 960,173 773,743 329,858 111,590 389,055 185,955 5,224,288 Growth 30% 2025 Track 2 San Francisco 190,573 83,492 8,294 31,318 16,425 16,364 1,913 7,607 18,850 374,835 San Mateo 69,440 399,635 57,379 34,801 7,098 2,978 474 2,101 3,248 577,153 Santa Clara 12,131 72,029 1,363,728 67,852 11,350 1,683 349 835 964 1,530,921 Alameda 19,314 22,244 42,423 727,168 80,098 15,388 3,092 1,440 1,777 912,943 Contra Costa 6,831 3,544 4,878 52,330 613,461 36,789 4,705 1,520 1,469 725,527 Solano 6,148 940 573 7,601 27,358 238,456 5,652 1,768 1,815 290,312 Napa 1,041 186 161 1,940 4,691 10,186 85,292 7,452 705 111,655 Sonoma 4,086 744 440 2,223 3,381 2,459 8,341 344,193 12,114 377,980 Marin 8,758 2,268 684 2,701 3,766 3,374 1,028 19,623 141,548 183,750 Total 318,322 585,082 1,478,559 927,935 767,627 327,676 110,846 386,539 182,490 5,085,076 Growth 26%



Table 9.10 Off-Peak County-to-County Trip Table Forecasts

San

Francisco San



2000 San Francisco 327,589 95,079 12,073 35,298 25,290 6,917 1,168 4,230 18,051 525,694 San Mateo 91,755 507,317 72,781 28,328 10,506 1,430 304 977 2,777 716,174 Santa Clara 13,096 77,397 1,553,568 56,642 17,130 1,198 305 607 876 1,720,821 Alameda 22,803 22,183 45,999 844,343 109,626 9,574 2,093 1,834 2,335 1,060,789 Contra Costa 13,399 6,219 10,436 73,440 656,424 22,445 3,300 2,227 2,360 790,250 Solano 5,341 963 728 7,443 24,198 211,731 3,437 1,616 1,275 256,733 Napa 857 198 151 1,758 4,084 3,609 82,427 7,521 474 101,079 Sonoma 3,300 759 367 1,572 2,559 1,701 7,392 307,124 8,271 333,045 Marin 15,557 2,518 735 2,227 2,730 1,757 609 9,433 186,413 221,978 Total 493,697 712,632 1,696,837 1,051,051 852,546 260,363 101,035 335,571 222,832 5,726,562 2010 STIP San Francisco 278,371 101,536 13,225 36,864 25,811 7,682 1,429 5,083 18,964 488,966 San Mateo 99,066 544,494 79,209 30,325 9,738 1,633 354 1,025 3,088 768,931 Santa Clara 14,090 84,202 1,767,903 61,914 13,817 1,429 416 779 980 1,945,530 Alameda 25,119 25,068 52,773 949,026 111,818 11,558 2,653 2,388 2,712 1,183,114 Contra Costa 14,710 6,418 9,368 79,855 766,563 26,535 4,023 2,638 2,598 912,708 Solano 5,918 1,109 863 8,610 28,508 265,692 4,936 2,164 1,612 319,412 Napa 1,042 232 193 2,168 5,028 5,385 100,649 9,688 618 125,003 Sonoma 3,773 747 376 1,808 2,599 2,132 8,832 400,656 9,326 430,249 Marin 15,848 2,796 829 2,548 2,918 2,179 770 10,878 205,080 243,845 Total 457,937 766,603 1,924,738 1,173,118 966,802 324,225 124,061 435,298 244,976 6,417,757 Growth 12%2020 Track 1 San Francisco 292,092 114,828 15,180 43,144 28,741 8,947 1,533 4,978 20,695 530,139 San Mateo 110,654 587,894 86,285 33,343 10,610 1,924 405 1,026 3,327 835,467 Santa Clara 15,988 90,904 1,948,066 67,659 14,801 1,627 469 827 1,060 2,141,402 Alameda 30,237 27,964 58,632 1,044,139 122,942 12,850 2,888 2,222 2,951 1,304,825 Contra Costa 17,140 7,178 10,204 88,940 860,393 30,365 4,477 2,498 2,841 1,024,038 Solano 7,023 1,345 997 9,753 32,096 316,389 6,313 2,420 1,880 378,217 Napa 1,176 272 223 2,446 5,688 6,900 116,341 10,430 728 144,204 Sonoma 3,901 790 439 1,825 2,636 2,477 10,121 455,796 10,516 488,500 Marin 17,560 3,062 911 2,787 3,197 2,527 902 11,880 220,240 263,066 Total 495,770 834,237 2,120,937 1,294,036 1,081,104 384,006 143,450 492,078 264,239 7,109,858 Growth 24%2025 Track 1 San Francisco 286,139 114,522 14,148 45,605 30,177 15,939 2,299 8,598 17,846 535,272 San Mateo 112,339 586,163 84,894 40,126 11,232 2,702 518 1,601 3,280 842,855 Santa Clara 16,047 90,685 1,959,877 78,771 15,827 1,793 497 957 1,094 2,165,548 Alameda 31,422 32,083 65,235 1,070,954 124,570 16,749 3,528 2,983 2,647 1,350,171 Contra Costa 17,709 7,125 9,934 89,285 894,401 41,915 5,842 3,571 2,488 1,072,269 Solano 12,195 1,893 1,064 12,508 40,432 339,829 6,813 2,517 2,765 420,016 Napa 1,742 352 246 2,919 6,749 7,533 123,394 10,461 934 154,328 Sonoma 6,735 1,274 591 2,495 3,594 2,671 10,412 478,458 16,347 522,578 Marin 15,869 3,120 939 2,582 2,993 3,968 1,256 19,591 208,365 258,683 Total 500,197 837,216 2,136,927 1,345,244 1,129,975 433,100 154,559 528,738 255,764 7,321,720 Growth 28%2025 Track 2 San Francisco 274,870 106,472 12,856 40,702 28,285 15,513 2,286 7,720 16,760 505,464 San Mateo 105,429 578,458 83,598 39,486 11,004 2,630 519 1,588 3,253 825,965 Santa Clara 14,554 89,138 1,900,051 77,038 15,748 1,791 497 953 1,089 2,100,858 Alameda 27,305 31,433 63,506 1,047,025 119,386 16,479 3,518 2,925 2,595 1,314,172 Contra Costa 16,261 6,918 9,865 86,009 892,300 41,743 5,834 3,558 2,466 1,064,954 Solano 11,827 1,833 1,061 12,329 40,289 338,827 6,764 2,516 2,762 418,209 Napa 1,730 352 246 2,908 6,736 7,477 122,325 10,461 934 153,169 Sonoma 6,092 1,263 589 2,451 3,587 2,670 10,412 477,461 16,206 520,732 Marin 14,659 3,075 924 2,535 2,968 3,962 1,256 19,338 205,683 254,399 Total 472,728 818,941 2,072,696 1,310,483 1,120,302 431,093 153,411 526,519 251,748 7,157,921 Growth 25%



Table 9.11 AM Peak Hour Screenline Results

Vehicle Trips Percent Growth Over 2000 Screenline No. Name 2000

2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2 2000

2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2

I1 SR 4 24,858 32,514 37,686 40,943 40,433 0% 31% 52% 65% 63% I2 Concord 27,756 34,386 38,269 39,877 39,829 0% 24% 38% 44% 43% I3 Orinda 18,642 22,373 26,079 28,020 29,210 0% 20% 40% 50% 57% I4 I-680 34,486 43,134 48,252 52,666 53,606 0% 25% 40% 53% 55% I5 Treat 28,999 37,502 41,379 43,137 43,458 0% 29% 43% 49% 50% I6 Ygnacio 24,009 30,095 32,489 34,410 32,762 0% 25% 35% 43% 36% I7 SR24 4,876 6,336 8,016 8,576 7,054 0% 30% 64% 76% 45% I8 Walnut Creek 26,297 32,793 35,321 37,051 36,857 0% 25% 34% 41% 40% I9 San Ramon 16,311 22,465 24,906 25,525 23,182 0% 38% 53% 56% 42% I10 Danville(NB/SB) 7,415 8,261 9,958 11,547 9,788 0% 11% 34% 56% 32% I11 Danville (EB/WB) 4,094 5,237 6,414 6,756 6,017 0% 28% 57% 65% 47% I12 Antioch/Brentwood 7,169 11,935 12,310 13,558 15,311 0% 66% 72% 89% 114% I13 Oakley/Brentwood 7,602 14,219 12,942 13,259 14,374 0% 87% 70% 74% 89% I14 Richmond 25,505 31,049 31,920 34,698 32,808 0% 22% 25% 36% 29% I15 Rich/Sanpb 8,836 12,972 11,894 14,549 12,118 0% 47% 35% 65% 37% I16 I-580 23,648 38,013 43,789 45,759 42,329 0% 61% 85% 93% 79% I17 West Livermore 23,246 33,227 35,459 33,129 30,647 0% 43% 53% 43% 32% I18 Pinole/County 25,477 31,970 33,849 37,067 34,406 0% 25% 33% 45% 35% Total 339,224 448,482 490,934 520,526 504,190 0% 32% 45% 53% 49% Cordon Line

Cordon Line 92,496 113,543 128,157 138,848 141,380 0% 23% 39% 50% 53%

R1 West/Central 6,376 9,089 10,977 11,582 10,510 0% 43% 72% 82% 65% R2 Lamorinda 20,884 24,911 27,380 28,936 30,444 0% 19% 31% 39% 46% R3 TriValley 19,408 25,824 27,003 27,398 24,677 0% 33% 39% 41% 27% R4 Central/East 17,066 23,923 27,373 28,701 28,750 0% 40% 60% 68% 68% R5 S.C Central 6,675 7,441 7,951 8,665 8,067 0% 11% 19% 30% 21% R6 S.C East 14,401 20,660 20,835 22,247 23,211 0% 43% 45% 54% 61% R7 S.C Tri Valley 13,941 19,768 23,010 23,098 20,787 0% 42% 65% 66% 49% R8 S.C West 19,714 25,099 27,587 30,755 29,379 0% 27% 40% 56% 49% R9 Alameda County 21,695 25,283 27,019 27,617 25,545 0% 17% 25% 27% 18% R10 Sunol 12,459 17,554 20,636 21,836 21,034 0% 41% 66% 75% 69% R11 Greenville 14,384 22,353 22,588 22,441 20,050 0% 55% 57% 56% 39% Total 259,497 335,448 370,516 392,124 383,835 0% 29% 43% 51% 48% Grand Total 598,722 783,930 861,450 912,650 888,025 0% 24% 30% 34% 33%



Table 9.12 PM Peak Hour Screenline Results

Vehicle Trips Percent Growth Over 2000 Screenline No. Name 2000 2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2 2000

2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2

I1 SR 4 28,004 36,027 38,399 43,101 46,648 0% 29% 37% 54% 67% I2 Concord 30,390 36,974 40,556 42,774 44,116 0% 22% 33% 41% 45% I3 Orinda 17,846 20,355 22,198 23,352 24,322 0% 14% 24% 31% 36% I4 I-680 36,852 45,351 50,968 58,349 60,240 0% 23% 38% 58% 63% I5 Treat 33,553 41,780 45,188 47,973 48,923 0% 25% 35% 43% 46% I6 Ygnacio 27,578 34,196 37,544 39,228 38,438 0% 24% 36% 42% 39% I7 SR24 5,741 6,632 7,636 8,065 7,619 0% 16% 33% 40% 33% I8 Walnut Creek 31,048 38,538 42,117 44,409 44,975 0% 24% 36% 43% 45% I9 San Ramon 19,194 22,630 24,576 25,719 23,455 0% 18% 28% 34% 22% I10 Danville(NB/SB) 7,924 9,231 10,511 10,986 10,588 0% 16% 33% 39% 34% I11 Danville (EB/WB) 5,204 6,123 6,355 6,824 6,544 0% 18% 22% 31% 26% I12 Antioch/Brentwood 8,716 13,025 13,520 13,862 16,202 0% 49% 55% 59% 86% I13 Oakley/Brentwood 8,502 14,014 13,023 13,577 15,142 0% 65% 53% 60% 78% I14 Richmond 25,073 29,302 30,867 34,281 34,954 0% 17% 23% 37% 39% I15 Rich/Sanpb 9,458 12,917 15,447 15,934 15,644 0% 37% 63% 68% 65% I16 I-580 25,585 36,863 43,739 46,090 43,681 0% 44% 71% 80% 71% I17 West Livermore 22,697 32,203 34,197 34,260 34,856 0% 42% 51% 51% 54% I18 Pinole/County 25,508 31,033 33,723 37,878 39,973 0% 22% 32% 48% 57% Total 368,874 467,195 510,564 546,662 556,321 0% 27% 38% 48% 51% Cordon Line

Cordon Line 99,187 115,377 129,667 145,756 155,717 0% 16% 31% 47% 57%

R1 West/Central 5,647 7,572 9,639 12,149 12,993 0% 34% 71% 115% 130% R2 Lamorinda 21,712 24,920 27,073 28,722 29,767 0% 15% 25% 32% 37% R3 TriValley 21,078 26,278 29,275 30,066 27,149 0% 25% 39% 43% 29% R4 Central/East 18,366 24,727 27,594 28,749 30,316 0% 35% 50% 57% 65% R5 S.C Central 5,961 6,879 8,790 8,109 9,459 0% 15% 47% 36% 59% R6 S.C East 15,696 21,490 21,431 22,354 24,618 0% 37% 37% 42% 57% R7 S.C Tri Valley 16,276 20,774 23,101 24,425 21,123 0% 28% 42% 50% 30% R8 S.C West 20,618 25,377 27,361 31,836 33,334 0% 23% 33% 54% 62% R9 Alameda County 18,403 21,804 23,143 24,089 23,194 0% 18% 26% 31% 26% R10 Sunol 12,611 15,227 19,839 20,431 19,333 0% 21% 57% 62% 53% R11 Greenville 13,245 20,468 20,772 20,898 21,503 0% 55% 57% 58% 62% Total 268,800 330,894 367,685 397,584 408,505 0% 23% 37% 48% 52% Grand Total 637,673 798,089 878,249 944,246 964,825 0% 20% 27% 32% 34%



Table 9.13 AM Peak Period Screenline Results


2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2 2000

2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2

I1 SR 4 81,413 100,449 113,315 125,748 124,975 0% 23% 39% 54% 54% I2 Concord 86,456 104,539 117,015 123,635 120,799 0% 21% 35% 43% 40% I3 Orinda 58,294 68,855 76,065 79,129 79,284 0% 18% 30% 36% 36% I4 I-680 106,876 123,746 136,453 145,214 140,277 0% 16% 28% 36% 31% I5 Treat 96,134 117,117 131,766 138,807 137,447 0% 22% 37% 44% 43% I6 Ygnacio 79,469 95,434 105,243 111,682 107,552 0% 20% 32% 41% 35% I7 SR24 15,833 20,139 23,538 23,647 24,358 0% 27% 49% 49% 54% I8 Walnut Creek 91,565 114,491 126,672 132,877 129,910 0% 25% 38% 45% 42% I9 San Ramon 48,033 62,996 72,002 74,344 68,522 0% 31% 50% 55% 43% I10 Danville(NB/SB) 16,561 22,200 27,987 31,875 29,172 0% 34% 69% 92% 76% I11 Danville (EB/WB) 10,013 14,802 18,054 19,438 18,043 0% 48% 80% 94% 80% I12 Antioch/Brentwood 22,008 36,792 37,303 39,287 42,895 0% 67% 70% 79% 95% I13 Oakley/Brentwood 21,725 39,218 38,378 40,870 42,780 0% 81% 77% 88% 97% I14 Richmond 83,820 97,930 104,761 113,673 107,869 0% 17% 25% 36% 29% I15 Rich/Sanpb 27,460 37,056 40,554 41,422 38,695 0% 35% 48% 51% 41% I16 I-580 68,520 100,617 123,289 133,383 127,646 0% 47% 80% 95% 86% I17 West Livermore 67,648 92,742 103,840 100,099 104,781 0% 37% 54% 48% 55% I18 Pinole/County 77,231 93,658 101,504 112,381 108,347 0% 21% 31% 46% 40% Total 1,059,058 1,342,782 1,497,739 1,587,513 1,553,354 0% 27% 41% 50% 47% Cordon Line

Cordon Line 301,518 367,420 415,202 450,279 442,629 0% 22% 38% 49% 47%

R1 West/Central 16,468 21,612 27,658 31,699 28,923 0% 31% 68% 92% 76% R2 Lamorinda 68,269 81,595 90,880 93,957 93,430 0% 20% 33% 38% 37% R3 TriValley 57,406 75,872 86,743 88,869 83,813 0% 32% 51% 55% 46% R4 Central/East 51,956 70,285 82,203 86,006 85,601 0% 35% 58% 66% 65% R5 S.C Central 21,354 25,593 26,895 28,050 26,195 0% 20% 26% 31% 23% R6 S.C East 43,895 60,634 61,624 64,653 67,025 0% 38% 40% 47% 53% R7 S.C Tri Valley 45,969 59,220 67,796 68,268 65,218 0% 29% 47% 49% 42% R8 S.C West 62,951 77,149 85,087 94,312 92,891 0% 23% 35% 50% 48% R9 Alameda County 69,903 80,686 84,605 87,715 84,487 0% 15% 21% 25% 21% R10 Sunol 37,780 49,002 61,787 63,384 60,704 0% 30% 64% 68% 61% R11 Greenville 34,880 52,555 53,230 53,672 51,406 0% 51% 53% 54% 47% Total 812,346 1,021,622 1,143,709 1,210,863 1,182,321 0% 26% 41% 49% 46% Grand Total 1,871,403 2,364,403 2,641,448 2,798,376 2,735,675 0% 21% 29% 33% 32%



Table 9.14 PM Peak Period Screenline Results


2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2 2000

2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2

I1 SR 4 96,438 123,126 139,672 149,603 145,895 0% 28% 45% 55% 51% I2 Concord 103,597 120,894 134,699 141,262 139,190 0% 17% 30% 36% 34% I3 Orinda 60,308 72,469 80,525 83,413 81,820 0% 20% 34% 38% 36% I4 I-680 127,900 147,789 165,660 182,583 177,900 0% 16% 30% 43% 39% I5 Treat 115,453 143,279 162,848 166,629 159,670 0% 24% 41% 44% 38% I6 Ygnacio 93,971 116,568 130,668 133,741 125,646 0% 24% 39% 42% 34% I7 SR24 19,787 22,637 27,628 27,262 28,109 0% 14% 40% 38% 42% I8 Walnut Creek 107,980 133,962 150,201 157,136 147,983 0% 24% 39% 46% 37% I9 San Ramon 63,776 75,471 84,195 86,470 80,609 0% 18% 32% 36% 26% I10 Danville(NB/SB) 21,020 30,925 33,785 39,096 34,496 0% 47% 61% 86% 64% I11 Danville (EB/WB) 15,304 20,917 23,125 24,149 21,756 0% 37% 51% 58% 42% I12 Antioch/Brentwood 30,646 43,139 42,856 45,127 51,265 0% 41% 40% 47% 67% I13 Oakley/Brentwood 28,807 44,805 42,362 44,619 47,942 0% 56% 47% 55% 66% I14 Richmond 89,776 102,818 105,306 119,523 115,535 0% 15% 17% 33% 29% I15 Rich/Sanpb 33,385 46,628 50,068 54,577 54,548 0% 40% 50% 63% 63% I16 I-580 85,610 121,192 145,767 154,981 146,774 0% 42% 70% 81% 71% I17 West Livermore 74,394 104,803 116,236 119,654 120,278 0% 41% 56% 61% 62% I18 Pinole/County 87,984 105,509 111,082 131,903 127,158 0% 20% 26% 50% 45% Total 1,256,136 1,576,930 1,746,685 1,861,729 1,806,574 0% 26% 39% 48% 44% Cordon Line

Cordon Line 349,508 414,770 457,055 522,372 516,246 0% 19% 31% 49% 48%

R1 West/Central 18,206 26,534 31,609 39,401 37,852 0% 46% 74% 116% 108% R2 Lamorinda 75,049 87,991 97,769 102,532 99,800 0% 17% 30% 37% 33% R3 TriValley 68,019 89,063 97,958 98,080 92,146 0% 31% 44% 44% 35% R4 Central/East 62,308 79,931 92,175 96,041 94,750 0% 28% 48% 54% 52% R5 S.C Central 21,361 23,651 26,951 28,492 31,866 0% 11% 26% 33% 49% R6 S.C East 52,870 68,816 70,653 73,471 75,577 0% 30% 34% 39% 43% R7 S.C Tri Valley 54,703 68,347 78,215 78,363 73,394 0% 25% 43% 43% 34% R8 S.C West 74,422 89,128 93,080 112,220 110,896 0% 20% 25% 51% 49% R9 Alameda County 63,467 76,121 79,806 83,204 80,832 0% 20% 26% 31% 27% R10 Sunol 46,312 55,376 66,015 68,784 66,812 0% 20% 43% 49% 44% R11 Greenville 39,355 60,661 62,418 62,247 61,761 0% 54% 59% 58% 57% Total 925,580 1,140,389 1,253,702 1,365,206 1,341,930 0% 23% 35% 47% 45% Grand Total 2,181,716 2,717,319 3,000,388 3,226,935 3,148,504 0% 20% 27% 32% 31%



Table 9.15 Daily Screenline Results

Vehicle Trips Percent Growth Over 2000

Screenline No. Name 2000

2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2 2000

2010 STIP

2020 Track 1

2025 Track 1

2025 Track 2

I1 SR 4 304,821 394,763 438,752 479,978 469,613 0% 30% 44% 57% 54% I2 Concord 321,118 383,294 420,309 440,443 436,354 0% 19% 31% 37% 36% I3 Orinda 207,462 248,318 271,117 282,318 276,884 0% 20% 31% 36% 33% I4 I-680 405,236 473,573 519,111 554,150 541,025 0% 17% 28% 37% 34% I5 Treat 363,644 459,048 506,709 527,970 511,328 0% 26% 39% 45% 41% I6 Ygnacio 307,867 387,766 422,991 440,197 419,354 0% 26% 37% 43% 36% I7 SR24 65,525 74,660 84,196 84,453 85,611 0% 14% 28% 29% 31% I8 Walnut Creek 336,348 428,324 468,162 488,417 467,474 0% 27% 39% 45% 39% I9 San Ramon 207,579 249,964 276,609 283,602 263,562 0% 20% 33% 37% 27% I10 Danville(NB/SB) 64,571 86,720 99,716 109,980 102,648 0% 34% 54% 70% 59% I11 Danville (EB / WB) 42,870 57,387 67,354 70,555 66,755 0% 34% 57% 65% 56% I12 Antioch/Brentwood 96,813 143,310 141,748 148,226 166,343 0% 48% 46% 53% 72% I13 Oakley/Brentwood 88,884 145,470 143,715 150,887 162,187 0% 64% 62% 70% 82% I14 Richmond 312,773 360,297 383,574 420,681 405,179 0% 15% 23% 35% 30% I15 Rich/Sanpb 106,017 130,816 143,679 154,755 152,271 0% 23% 36% 46% 44% I16 I-580 281,543 398,866 478,112 508,871 483,651 0% 42% 70% 81% 72% I17 West Livermore 250,561 382,769 419,793 424,174 422,839 0% 53% 68% 69% 69% I18 Pinole/County 282,623 340,873 369,275 421,155 409,025 0% 21% 31% 49% 45% Total 4,046,255 5,146,218 5,654,925 5,990,814 5,842,103 0% 27% 40% 48% 44% Cordon Line

Cordon Line 1,177,103 1,431,498 1,575,886 1,759,788 1,721,624 0% 22% 34% 50% 46%

R1 West/Central 52,923 72,479 89,509 104,167 97,697 0% 37% 69% 97% 85% R2 Lamorinda 247,883 296,273 324,373 337,971 330,649 0% 20% 31% 36% 33% R3 TriValley 215,578 287,419 310,935 318,999 299,588 0% 33% 44% 48% 39% R4 Central/East 193,849 253,960 287,616 299,707 301,892 0% 31% 48% 55% 56% R5 S.C Central 60,945 69,612 75,374 79,076 87,486 0% 14% 24% 30% 44% R6 S.C East 162,015 216,626 228,874 238,584 247,953 0% 34% 41% 47% 53% R7 S.C Tri Valley 183,003 235,866 263,327 265,279 249,201 0% 29% 44% 45% 36% R8 S.C West 234,827 290,139 317,103 367,106 364,451 0% 24% 35% 56% 55% R9 Alameda County 242,186 302,639 320,735 331,003 321,714 0% 25% 32% 37% 33% R10 Sunol 164,274 207,189 240,382 249,706 244,217 0% 26% 46% 52% 49% R11 Greenville 134,724 239,015 243,235 244,032 230,963 0% 77% 81% 81% 71% Total 3,069,311 3,902,713 4,277,348 4,595,418 4,497,433 0% 27% 39% 50% 47% Grand Total 7,115,567 9,048,931 9,932,273 10,586,232 10,339,536 0% 21% 28% 33% 31%



Analysis of Screenlines

• a.m. peak hour traffic on Contra Costa roads is projected to increase 24 percent by 2010, 30 percent by 2020, and 34 percent by 2025 at regional and internal screenlines. The projected growth in a.m. peak period traffic at the screenlines is one or two per-centage points lower for each of the forecast years.

• The MTC Track 2 program trims about one to two percentage points off of the pro-jected growth in the year 2025.

• Peak hour screenline traffic forecasts tend to be more volatile (showing larger and smaller increases for individual screenlines) than the peak period forecasts.

• West Central (R1), traffic is projected to increase between 2000 and 2025 by 65 to 82 percent during the a.m. peak hour, 76 to 92 percent during the a.m. peak period, 85 to 97 percent on a daily basis.

• East-Central (R4) traffic is projected to increase between 2000 and 2025 by 68 percent during the a.m. peak hour, 65 to 66 percent during the a.m. peak period, 55 to 56 percent on a daily basis.

• Ygnacio Treat (R5), traffic is projected to increase between 2000 and 2025 by 21 to 30 percent during the a.m. peak hour, 23 to 31 percent during the a.m. peak period, 30 to 44 percent on a daily basis.

• Route 4 East (R6) traffic is projected to increase between 2000 and 2025 by 54 to 61 percent during the a.m. peak hour, 47 to 53 percent during the a.m. peak period, 47 to 53 percent on a daily basis.

• San Ramon Valley – I 680 (R7) traffic is projected to increase between 2000 and 2025 by 49 to 66 percent during the a.m. peak hour, 42 to 49 percent during the a.m. peak period, 36 to 45 percent on a daily basis.

• Altamont Pass (R11), traffic is projected to increase between 2000 and 2025 by 39 to 56 percent during the a.m. peak hour, 47 to 54 percent during the a.m. peak period, 71 to 81 percent on a daily basis.

• The Richmond-San Pablo internal screenline (I-15), which runs parallel to I-80, just east of San Pablo Avenue and 23rd Street, generally shows higher growth in 2025 Track 1 than in 2025 Track 2. This screenline also exhibits fluctuations in a.m. peak hour traffic between 2010 and 2020 Track 1, and 2025 Tracks 1 and 2. These fluctuations occur only for the a.m. peak hour. All other periods (p.m. peak hour, a.m. peak period, and p.m. peak period show increases in traffic from 2010 to 2020. The fluctuations may be due to network changes, such as the I-580 HOV lane in Track 1. The I-580 freeway and HOV lanes are not included in the original definition of this screenline, which may



account for some of the fluctuations. (This will be investigated and the final conclu-sions presented at the TMWG meeting.)

• The Oakley-Brentwood screenline (I-13) is a north-south screenline running east of SR 160, crossing the Delta Expressway, and then running west of the Expressway. The a.m. peak hour traffic on this screenline increases dramatically – by 87 percent (eight percent per year) – between 2000 and 2010. The growth curve then flattens, with a total growth of 89 percent by 2020. The forecasted traffic crossing this screenline drops slightly between 2010 and 2020 for all periods. This may be the result of the construc-tion of several new east-west streets between 2010 and 2020 that were not included in the screenline totals. (This will be investigated and the final conclusions presented at the TMWG meeting.)

• The Antioch-Brentwood screenline (I-12) also shows dramatic growth – more than 100 percent – between 2000 and 2025. This screenline also shows flat growth between 2010 and 2020 for the p.m. peak period. This flattening phenomenon may be due to the omission of several new north-south streets from the screenline. (This will be investigated and the final conclusions presented at the TMWG meeting.)



10.0 Level of Service (LOS)

10.1 Introduction

Caliper developed an interface that will replace the existing CCTA VCCC LOS program with a new program that can work independently on a Windows OS, or can be invoked within the CCTA TransCAD countywide model. The program is capable of the following:

• Manage base and future year turning movement and approach volume datasets;

• Calculate the adjusted model output for the approach and departure volumes for each link at an intersection;

• Apply Furness adjustment to estimate future turning movements; and

• Compute the LOS for signalized intersection using the CCTALOS method.

From a single toolbox, similar to the one shown in Figure 10.1, the user has access to all of the operations listed above and the freedom to select one or more files to be involved in the analysis.

Figure 10.1 CCTA Interface

CCTA Technical Procedures

Adjust Model Output

Furness Adjust

Compute CCTALOS

Compute ICU

Choose Input Files



The Adjust Model Output, Furness Adjust, Compute CCTALOS, and Compute ICU procedures each will allow the user to perform the respective operation on all intersections in the network or on a selection set of intersections in the network. Therefore, all of the tools, with the exception of the Choose Input Files button, will require that a TransCAD map be open and that a node layer be displayed in the map. This node layer should contain the intersections upon which the analysis will be performed.

In interactive mode, the user will be able to edit the input parameters to the CCTALOS and intersection capacity utilization (ICU) methods through an interactive toolbox. The user will also have the option to create and open an intersection diagram depicting the turning movement flows at the intersection and the LOS results (see Figure 10.2).

Figure 10.2 Turning Movement Diagram

A data flow chart illustrating the interface, inputs, and outputs is provided in Figure 10.3.



Figure 10.3 Data Flow Diagram for CCTA Technical Procedures Implementation

Set of Selected Intersections

CCTA Technical Procedures Dialog Box

Choose Input Files

Adjust Model Output

Compute ICU

Furness Adjust

Compute CCTALOS

TransCAD Map

Turning Movement Tables

• Approach/departure volumes

• Intersection level of service

• Approach/departure volumes

• Intersection level of service



10.2 CCTALOS Validation

The CCTALOS TransCAD Version was tested on approximately 360 intersections (see Appendix J Table 1 for list sorted by jurisdiction with the traffic count date). It was found that some intersections were duplicate locations. These intersections were removed from the analysis. Some intersections did not have driveways or minor streets coded that cre-ated a problem in computing the LOS. In all such cases, short stub dummy links were added in. Further, it was found that, for some intersections, a count was provided in a movement that did not have a geometry associated with it. In such situations, the counts were considered to be null.

The CCTALOS TransCAD Version results for the +/-360 intersections for existing condi-tions were compared to the results obtained from the CCTALOS DOS program. The results of the comparison for a.m. peak hour are shown in Appendix J Table 2. Since it is sorted by jurisdiction, Appendix J Table 1 can be used as a key to look up the intersection number for Appendix J Table 2.

10.3 Choose Input Files

The first button in the toolbox, Choose Input Files, will open a dialog box that will allow the user to browse and select one or more files that belong to each class of input, such as turning movement count tables, turning movement model estimates, and turning move-ment model forecasts. These files will then be available as input tables to the adjustment and LOS operations. Input files can be added and removed later through the Choose Input Files dialog box. The Choose Input Files button circumvents the need to choose input files when each of the technical operations is invoked.

The different types of files that the user will be asked to choose include:

1. Base year turning movement counts,

2. Base year turning movement model estimates,

3. Base year turning movement forecasts, and

4. Desired future year approach and departure volumes.

Turning movement tables will have a record for each turning movement at a single inter-section, at multiple intersections, or at every intersection in a network. Turning move-ment tables will have at least the following fields:

1. Node ID – The ID of the node at which the turning movement occurs,

2. From Link ID – The ID of the approach link for the turning movement,



3. To Link ID – The ID of the departure link for the turning movement, and

4. Volume – The flow in vehicles per hour on the turning movement.

The four fields above are sufficient for the model output and Furness adjustment proce-dures. Additional fields that pertain uniquely to individual turning movements (e.g., number of dedicated lanes, saturation flow, etc.) will be required for the LOS operations and will be discussed later.

10.4 Adjust Model Output

The Adjust Model Output button opens a dialog box prompting the user to select three input files from among those chosen in the Choose Input Files dialog box. These input files include:

1. Base year turning movement counts,

2. Base year turning movement model estimates, and

3. Future year turning movement model forecasts.

Once the input files are chosen and the user clicks “OK” to proceed, the user will choose a name for the output file. The output file is a table containing the total desired approach and departure volumes for each leg of each of the intersections included in the analysis. The output file is automatically added to the list of input approach and departure volume tables.

Furness Adjust

The Furness Adjust button opens a dialog box requiring two inputs: 1) a base year turning movement count table and 2) a desired future year approach and departure volume table. When the user has selected the input files and clicked “OK” to proceed, the user is prompted to choose an output file name to store the future year turning movement counts for each intersection in the analysis. The output file is automatically added to the list of input files available to other operations.

10.5 Compute CCTALOS

The Compute CCTALOS button launches the CCTALOS methodology, which allows the user either to interactively click on intersections in a map to compute signalized intersec-tion LOS or to run a batch LOS computation on a selection set of intersections in a map.



In interactive mode, a toolbox is opened that allows the user to click on individual inter-sections in a map and to modify input values, such as turning movement volume, number of phases, and other parameters and variables that are also editable in the DOS CCTALOS program. The resulting LOS is displayed in the toolbox.

In batch mode, the user is able to execute the CCTALOS computation on a set of intersec-tions in a map. The output results may either be stored in a new table having a field for node ID and a field for LOS, or in a user specified LOS field in the node layer database.

The Compute CCTALOS tool first opens a setup dialog box that allows the user to choose the input files and fields that store certain input variables required by the CCTALOS methodology. The setup dialog box also allows the user to edit global parameters, such as adjustment factors and lane capacities. From the setup dialog box, the user may click “OK” to execute the batch calculation or to open the interactive toolbox. The setup dialog box can be reopened from the interactive toolbox.

Below is a summary of required and optional inputs to the CCTALOS computation and a proposed approach for storing those variables in standard TransCAD data files.

Turning Movement Variables

Turning movement variables are stored in turning movement databases. These databases have a record for each turning movement at each intersection in the analysis. The required fields that relate the turning movement data to line and node features in the TransCAD geographic database are the following:

1. The node ID at which the turning movement takes place,

2. The ID of the link approaching the intersection, and

3. The ID of the link departing the intersection.

Additional fields are required to store CCTALOS input variables. These include inputs that vary at the turning movement level and so require a field in the turning movement table. These include the following:

1. Turning movement flow in vehicles per hour,

2. Number of lanes dedicated to the turning movement, and

3. Lane usage code.

The lane use code is identical to the lane nomenclature used in the CCTALOS software (see Table VI-III, CCTALOS User’s Manual). An example of an input turning movement table is shown in Figure 10.4.



Figure 10.4 Input Turning Movement Table

Approach and Departure Variables

The CCTALOS methodology also requires approach and departure variables that vary at the link level. Directional (e.g., AB and BA) fields in the link layer store variables specific to approaches to and departures from intersections in the network. These include:

1. Split phasing – A character string “Yes” or “No,” indicating whether split phasing is used on the link’s approach. The split phasing option should be identical for the links that represent opposing approaches.

2. Pedestrians – The number of pedestrians crossing the downstream end of the link in pedestrians per hour. This variable is used to convert right turning volumes to passenger car equivalents (PCE). Pedestrian flow will be an optional input to support cases where pedestrian flow is negligible or right-turn volumes are already converted to PCEs.

Variables that are used to adjust turning movement volumes, such as pedestrian volumes, will be taken into account in the calculations. Thus, the user will not have to make those adjustments externally, as is done with the CCTALOS software.



Intersection Variables

Intersection variables are those that vary at the intersection level, particularly variables pertaining to signal operations. For the CCTALOS procedure, the only intersection vari-able is the number of phases.

Input Parameters

There are a variety of parameters used throughout the CCTALOS methodology, including lane capacities and PCE adjustment factors. These parameters are accessible through the CCTALOS setup dialog box, where the user may edit the values.

In summary, the following are inputs needed to compute CCTALOS in TransCAD:

• Turning Movement Inputs:

1. Node ID,

2. Approach Link ID,

3. Departure Link ID,

4. Flow,

5. Number of Lanes, and

6. Lane Usage Codes.

• Link Database Inputs:

7. Split phasing, and

8. Pedestrian flow (optional).

• Node Database Inputs:

9. Number of Phases.

10.6 Compute ICU

Another tool included in the CCTA dialog box is for calculating the ICU. Like the CCTALOS, the ICU is a measure of the LOS at signalized intersections.

The implementation of the ICU methodology in TransCAD is very similar to the proposed approach for the CCTALOS. The ICU can be computed either in interactive mode by clicking on individual intersections, or in batch mode on a group of intersections. Like the proposed CCTALOS approach, input parameters for the ICU (e.g., cycle length, minimum green, lost time, peak-hour factor, etc.) are available for editing in the ICU setup dialog



box. Also, input variables (e.g., turning movement flows, saturation flows, etc.) are drawn from turning movement tables and geographic databases.

FR1 CCTA Travel Model Documentation COMPLETEtechnical details required in the model documentation...

Documents

Transcript of FR1 CCTA Travel Model Documentation COMPLETEtechnical details required in the model documentation...