appendix contents - sandag.org · mode choice splits total person-trip movements between zones into...

Technical Appendix 15

SANDAG Travel Demand Model Documentation

Appendix Contents

SANDAG Travel Demand Model Documentation ............................ TA 15-2

TA 15-2 Technical Appendix 15: SANDAG Travel Demand Model Documentation

2050 Regional Transportation Plan

SANDAG Travel Demand Model Documentation Technical Appendix 15 includes several

reports that document the SANDAG Travel

Demand Model, as follows:

2050 Regional Travel Demand Model

Documentation: This report presents a

description of the components of the

SANDAG travel demand model used in the

2050 Regional Transportation Plan,

including a general flow of information

and some of the key inputs, assumptions,

and computations for each of the

components.

SANDAG Off-Model Greenhouse Gas

Reduction Methodology for the 2050

Regional Transportation Plan: This paper

documents the SANDAG methodology for

estimating the off-model greenhouse gas

(GHG) reductions for several strategies

included in the 2050 Regional

Transportation Plan (RTP). These strategies

include aspects of the bicycle, pedestrian,

safe routes to schools, vanpool, carpool,

and buspool programs.

SANDAG Travel Demand Model Validation

Report: This report provides a guide to the

accuracy of the SANDAG transportation

model compared to observed data in and

around 2008. It also provides detailed

technical information on how the

transportation model was used to support

the development and decision-making

process for the 2050 RTP and its

Sustainable Communities Strategy, and

the Programmatic Environmental Impact

Report.

SANDAG Transportation Model Sensitivity

Analysis and Report: This report provides

insight into the sensitivity and elasticity of

the SANDAG four-step travel demand

model in terms of policy and operational

adjustment.

401 B Street, Suite 800 • San Diego, CA 92101-4231 • (619) 699-1900

2050 REGIONAL TRAVEL DEMAND

MODEL DOCUMENTATION

SEPTEMBER 2011

i

ABSTRACT

TITLE: 2050 Regional Travel Demand Model Documentation

AUTHOR: San Diego Association of Governments

DATE: September 2011

SOURCE OF COPIES:

San Diego Association of Governments 401 B Street, Suite 800 San Diego, CA 92101 (619) 699-1900

NUMBER OF PAGES:

63

ABSTRACT: The SANDAG regional travel demand model comprises a complex set of assumptions, input data, computations, and model interactions. This report presents a basic description of the components of the SANDAG travel demand model used in the 2050 Regional Transportation Plan, including a general flow of information and some of the key inputs, assumptions, and computations for each of the components.

ii

TABLE OF CONTENTS

EXECUTIVE SUMMARY ................................................................................................................. 1

CHAPTER 1: GROWTH FORECAST BACKGROUND ........................................................................ 3

1.1 OVERVIEW ........................................................................................................................ 3

1.2 DEMOGRAPHIC AND ECONOMIC FORECASTING MODEL (DEFM) .................................. 5

1.3 INTERREGIONAL COMMUTE MODEL (IRCM) ................................................................... 5

1.4 URBAN DEVELOPMENT MODEL (UDM) ............................................................................ 6

1.5 POPULATION BY AGE, SEX, AND ETHNICITY FORECAST (PASEF) ................................... 6

1.6 DATA SOURCES................................................................................................................. 7

CHAPTER 2: TRANSPORTATION MODELS ..................................................................................... 8

2.1 INTRODUCTION ................................................................................................................. 8

2.2 SURVEY INPUTS .............................................................................................................. 11

2.2.1 1995 & 2006 Travel Behavior Surveys ............................................................................................. 12 2.2.2 2001 Caltrans Statewide Survey & 2009 National Household Travel Survey ............................. 12 2.2.3 2001-2003 & 2009 Regional Transit Survey .................................................................................... 12 2.2.4 External Trip Surveys ......................................................................................................................... 13 2.2.5 Traffic Generator Studies ................................................................................................................. 13 2.2.6 1991 Visitor Survey ............................................................................................................................ 13 2.2.7 Census Transportation Planning Package and American Community Survey ........................... 13 2.2.8 2000 Market Research Survey .......................................................................................................... 14 2.2.9 Traffic Counts ..................................................................................................................................... 14 2.2.10 Transit Passenger Counts .................................................................................................................. 14 2.2.11 2006 Vehicle Occupancy and Classification Study ......................................................................... 15

2.3 GROWTH FORECAST INPUTS ........................................................................................... 15

2.4 HIGHWAY NETWORK INPUTS ......................................................................................... 16

2.4.1 Highway Facilities .............................................................................................................................. 17 2.4.2 Highway Attributes ........................................................................................................................... 17 2.4.3 Highway Capacities ........................................................................................................................... 18 2.4.4 Highway Travel Times ....................................................................................................................... 20

2.5 TRANSIT NETWORK INPUTS ............................................................................................ 20

2.5.1 Transit Facilities ................................................................................................................................. 21 2.5.2 Transit Attributes .............................................................................................................................. 21 2.5.3 Travel Times ....................................................................................................................................... 22 2.5.4 Fares .................................................................................................................................................... 23 2.5.5 Transit Walk Access ........................................................................................................................... 23

iii

2.5.6 Transit Auto Access ........................................................................................................................... 24

2.6 TRIP GENERATION .......................................................................................................... 25

2.6.1 Model Structure ................................................................................................................................. 26 2.6.2 Model Calibration ............................................................................................................................. 29

2.7 PATHBUILDING, SKIMMING, AND UTILITY CALCULATIONS .......................................... 29

2.8 TRIP DISTRIBUTION ......................................................................................................... 31


2.9 MODE CHOICE................................................................................................................. 35

2.9.1 Model Structure ................................................................................................................................. 37 2.9.2 Utility Computations ......................................................................................................................... 40 2.9.3 Model Calibration ............................................................................................................................. 50 2.9.4 Mode choice variations..................................................................................................................... 50

2.10 TRUCK MODEL .................................................................................................................... 51

2.10.1 Model Design ..................................................................................................................................... 51 2.10.2 Local Truck Model ............................................................................................................................. 52 2.10.3 Regional Truck Model ....................................................................................................................... 52 2.10.4 Truck Trip Reconciliation .................................................................................................................. 53 2.10.5 Air Quality .......................................................................................................................................... 53

2.11 HIGHWAY ASSIGNMENT ................................................................................................ 53

2.11.1 Model Structure ................................................................................................................................. 54 2.11.2 Post-assignment Processing ............................................................................................................. 55 2.11.3 Model Calibration ............................................................................................................................. 57

2.12 TRANSIT ASSIGNMENT ................................................................................................... 57


2.13 INDUCED DEMAND ......................................................................................................... 58

2.13.1 Land Use ............................................................................................................................................. 58 2.13.2 Trip Generation/Activities ................................................................................................................ 59 2.13.3 Trip Distribution ................................................................................................................................ 59 2.13.4 Trip Mode Choice .............................................................................................................................. 59 2.13.5 Trip Time of Day ................................................................................................................................ 59 2.13.6 Trip Assignment ................................................................................................................................. 60

CHAPTER 3: MODEL CHANGES FOR THE 2050 RTP .................................................................... 61

3.1 SOFTWARE ...................................................................................................................... 61

3.2 ZONE SYSTEM ................................................................................................................. 61

3.3 GROWTH FORECAST ....................................................................................................... 62

3.4 HIGHWAY NETWORK INPUTS ......................................................................................... 62

3.5 TRANSIT NETWORK INPUTS ............................................................................................ 62

iv

3.6 TRIP GENERATION .......................................................................................................... 62

3.7 TRIP DISTRIBUTION ......................................................................................................... 63

3.8 MODE CHOICE MODEL ................................................................................................... 63

3.9 HIGHWAY ASSIGNMENT ................................................................................................ 63

v

LIST OF TABLES

Table 1: Transit Mode Definitions ............................................................................................................ 21 Table 2: Trip Generation Model Inputs .................................................................................................... 25 Table 3: Trip Generation Model Output .................................................................................................. 26 Table 4: Trip Generation Tele-Work Percentages ................................................................................... 28 Table 5: Regional Control Variables ......................................................................................................... 29 Table 6: Trip Distribution Model Inputs ................................................................................................... 31 Table 7: Trip Distribution Model Outputs ................................................................................................ 32 Table 8: Mix Variable Density Factor ........................................................................................................ 33 Table 9: Employment Variable Density Factor ......................................................................................... 33 Table 10: 4D Variable Density Ranges ...................................................................................................... 34 Table 11: Mix Index Density Bins .............................................................................................................. 34 Table 12: Mode Choice Model Inputs ....................................................................................................... 36 Table 13: Mode Choice Model Output ..................................................................................................... 36 Table 14: Mode Choice Time and Cost Coefficients ................................................................................ 41 Table 15: Value of Time ............................................................................................................................. 41 Table 16: Auto Operating Costs ................................................................................................................ 43 Table 17: Parking Costs .............................................................................................................................. 43 Table 18: Mode Choice Constants ............................................................................................................ 47 Table 19: Application of Modal Constants ............................................................................................... 48 Table 20: Highway Assignment Model Inputs ......................................................................................... 54 Table 21: Highway Assignment Model Outputs ...................................................................................... 54 Table 22: Volume Delay Function Parameters ......................................................................................... 55 Table 23: Truck Passenger Car Equivalents .............................................................................................. 55 Table 24: Transit Assignment Model Inputs ............................................................................................. 57 Table 25: Transit Assignment Model Output ........................................................................................... 57

vi

LIST OF FIGURES

Figure 1: 2050 Regional Growth Forecast Models ..................................................................................... 4 Figure 2: Transportation Modeling Process ............................................................................................... 9 Figure 3: Transportation Model Nested Geographies ............................................................................. 16 Figure 4: Nested Mode Choice Model Structure...................................................................................... 38 Figure 5: Local and regional truck trips ................................................................................................... 51

1

EXECUTIVE SUMMARY

SANDAG deals with many complex mobility issues facing the San Diego region, including the development of a long-range Regional Transportation Plan (RTP). Transportation and land use models perform a very basic yet vital set of functions. Models are the principal tools used for alternatives analysis, and they provide planners and decision makers with information to help them equitably allocate scarce resources. The SANDAG transportation model provides a systematic analytical platform so that different alternatives and inputs can be evaluated in an iterative and controlled environment.

For the 2050 RTP, SANDAG uses an enhanced four-step transportation model. Four-step models have been the standard in transportation modeling since the late 1950s, and they are used by nearly every MPO in the United States for the development of transportation plans, corridor studies, Federal Transit Administration New Starts proposals, and air quality analyses. The estimates of regional transportation related emissions analyses meet the requirements established in the Transportation Conformity Rule, 40 CFR Sections 93.122(b) and 93.122(c). These requirements relate to the procedures to determine regional transportation-related emissions, including the use of network-based travel models, methods to estimate traffic speeds and delays, and the estimation of vehicle miles of travel.

The four major steps of the transportation model include:

trip generation;

trip distribution;

mode choice; and

network assignment (highway and transit).

After a first pass through the four steps, a feedback process is used to send congested travel conditions back into trip distribution and through to assignment. After several feedback iterations, a final pass is made through the mode choice and assignment steps to reflect congested travel conditions in mode decision-making. Travel model results are then combined with additional post-process input and output functions to form the complete modeling chain. Additionally, a truck model is run parallel to the four-step model, and truck trip tables are merged with passenger vehicle trip tables for highway assignment and air quality procedures.

A trip generation analysis is the first step in the transportation modeling process. Trip generation estimates the average weekday number of trip productions and attractions, or trip ends, in the region based on land use and demographic information from the series 12 regional growth forecast. Over a 24-hour period, roughly the same number of trips will originate in a zone as are destined there. However, residential zones will generate primarily trip productions while nonresidential zones will generate primarily trip attractions.

After trip generation, trip distribution allocates and balances trip productions and attractions through a gravity approach based on trip end density and location. Trip distribution considers the

2

distance between a trip ends that is based on the assumed highway and public transportation networks that are input for any given future year. The model is designed to modify trip patterns in response to new land use developments and transportation facility changes. For example, the opening of a new shopping center would shift trips from other nearby shopping areas to the new development. Another example would be the introduction of mixed-use development. In this case the model would yield shorter trip lengths by recognizing the increased opportunity for interaction between residential and commercial areas in the development.

Mode choice splits total person-trip movements between zones into different forms of transportation by auto, transit, and non-motorized modes (bicycling and walking). The mode choice step selects the most likely form of transportation for each trip, based on access, traveler’s income, trip purpose, parking costs, fuel price, transit fares, travel time, and other time and pricing parameters.

During network assignment, the model places each trip on the most efficient auto, transit, or non-motorized path based on the mode of transportation that was chosen earlier. Highway assignment produces traffic-volume estimates for all roadway segments in the system. These traffic volumes are an important input to emissions modeling. Similarly, transit trips are assigned to transit routes and segments.

Once these four steps are completed for the millions of trips in the region on an average weekday, the SANDAG model iterates the trip distribution and mode choice step and runs through traffic assignment again based on levels of congestion measured in the previous iteration. The iterations continue until all trips are assigned the most efficient path for their mode. Each step is sensitive to an extensive set of inputs used to prepare a model scenario.

TransCAD 5.0 is the transportation planning computer package used by SANDAG to provide a framework for performing much of the computer processing involved with modeling, and it is used for the trip distribution and assignment steps. ArcInfo, a Geographic Information System (GIS), is used extensively in the modeling process as well to maintain, manipulate, and display transportation, land use, and demographic data. SANDAG has written numerous customized programs that provide a linkage between TransCAD and ArcInfo. Other custom programs perform some modeling functions, such as trip generation and mode choice.

The remainder of the report highlights new features of the series 12 model and comparisons to the previous series 11 model. This model documentation is designed to provide insight into the process and a basic understanding of the relationships among the models and the data requirements

3

CHAPTER 1: GROWTH FORECAST BACKGROUND

1.1 OVERVIEW

SANDAG has produced economic and demographic forecasts for nearly 40 years, and transportation forecasts for nearly three decades. These forecasts are an integral part of SANDAG’s planning process as well as that of other governmental and private organizations.

Unlike the prior two forecasts, the 2050 Regional Growth Forecast includes assumptions about how local plans and policies may evolve over time in response to the region’s continuing growth. This forecast looks out forty years to the year 2050, while the horizon year of current local plans is typically ten or twenty years (i.e. out to 2020 or 2030). To bridge this gap, SANDAG began the forecast with adopted general plans and policies from the 18 incorporated cities. Then local jurisdictions were asked to provide detailed feedback on how land use plans might change in the future. Hence, the 2050 Regional Growth Forecast provides an assessment of where change may occur in the coming decades.

SANDAG uses four integrated models in its demographic, economic, and land use forecasts: (1) the Demographic and Economic Forecasting Model (DEFM), (2) the Interregional Commute Model (IRCM), (3) the Urban Development Model (UDM) and (4) the Population Age, Sex, and Ethnicity Forecast (PASEF), in conjunction with the Transportation Model. The 2050 Regional Growth Forecast is spatially linked to the transportation model via SANDAG’s Master Geographical Reference Areas (MGRAs). Employing the MGRA geography as a touchstone between models ensures data feedback between the Forecast and Transportation Model.

A noteworthy feature of the forecasting process is the feedback of information from one model to another (See Figure 1). For example, regionwide projections of jobs and housing from DEFM are used in the IRCM and then the output from the IRCM is used to adjust the output from DEFM. DEFM then provides the regionwide projections that serve as the basis for UDM and PASEF. Similarly, data from UDM and PASEF are major inputs to the transportation model, and then transportation model data are used in subsequent UDM calculations. A key feature of the modeling system is the central role that land use and transportation policies play in determining future travel patterns and the associated location of people, houses, and jobs.

These interrelated models satisfy the federal requirements specified in the Clean Air Act Amendments of 1990 and the Safe, Accountable, Flexible, Efficient, Transportation Equity Act: A Legacy for Users (SAFETEA-LU). These legislative acts mandate that transportation plans consider the long-range effects of the interaction between land uses and the transportation system.

4

Figure 1: 2050 Regional Growth Forecast Models

5

1.2 DEMOGRAPHIC AND ECONOMIC FORECASTING MODEL (DEFM)

DEFM is designed to forecast population and economic variables for the region. To forecast demographic variables, DEFM considers factors such as birth rates, survival rates, and the age, sex, and ethnic distributions of the resident population. Economic variables including employment, income, and housing supply are forecast based on assumptions about national, state, and local growth patterns and inter-industry relationships.

There are many linkages, both direct and indirect, between the demographic and economic variables that are accounted for and modeled by DEFM. For example, the population determines housing demand, demand for public facilities, and associated public finance projections. Economic activity, as measured by employment and output, depends in part on the size of the local population and income level. Income, in turn, depends in part on employment and labor market conditions. Over time, the population responds to economic conditions as is evident from net migration levels. Thus, the region’s economic activity depends on the local population, but the local population also depends on economic activity. DEFM is designed to capture the main interdependencies and interactions that exist in the region’s economy.

1.3 INTERREGIONAL COMMUTE MODEL (IRCM)

The 2050 Regional Growth Forecast is the third SANDAG forecast to include an Interregional Commute Model (IRCM). The purpose of the model is to account for individuals who work in the region but live outside its boundaries. Historically, the amount of interregional commuting into and out of the San Diego region had been relatively small. However, recent evidence indicates that interregional commuting is increasing rapidly. Between 1990 and 2000, for example, the number of workers commuting from Riverside County to job sites in the San Diego region has increased four-fold.

The IRCM predicts, using a gravity model, the future residential location of the workers holding new jobs created in the San Diego region. The residential location can be either inside the San Diego region, in Orange County, southwest Riverside County, Imperial County, or in Tijuana/Northern Baja California.

The IRCM assigns the residential location of workers based upon the accessibility of potential residential sites to job locations, the availability of residential land for development, and the relative price of homes. There are three basic tenets of the IRCM. First, as commuting time from work to possible residential locations increases, the probability of choosing those locations decreases. Second, more land available for residential development increases the potential for residential growth. Third, lower home prices also are an attraction factor in residential location. These three basic tenets also underlie the gravity model used in the Urban Development Model (UDM).

The results from the IRCM are used to modify the DEFM regional forecast. The initial regional forecast, referred to as the Baseline, is modified to reflect the fact that not all housing units, population, employment, and other elements predicted by the Baseline forecast will occur in the region. Rather, some residential and economic activity will occur in nearby areas outside the region.

6

1.4 URBAN DEVELOPMENT MODEL (UDM)

The Urban Development Model (UDM) allocates employment, population, housing and income from the regional forecast produced by DEFM to neighborhoods and jurisdictions within the region. The model is designed to forecast the location of residential and non-residential activity within the region for 5 year periods. Major model inputs include the current spatial distribution of jobs, housing units, income, and population. Land use data collected from local jurisdictions including general plans, policies, and current and future transportation infrastructure are also critical to the model.

UDM combines the transportation and land use factors mathematically to determine the likelihood that an employee at his or her place of work will reside in alternative residential locations around the region. In general, areas closer to employment opportunities are more attractive to employees as potential residences than areas further away from the place of employment. Therefore, as available residential capacity closer to work places is consumed, new employees are forced to travel longer distances to find suitable residential locations. Residential growth in a jurisdiction is influenced by growth within that jurisdiction as well is in surrounding areas and other parts of the region.

1.5 POPULATION BY AGE, SEX, AND ETHNICITY FORECAST (PASEF)

The program for forecasting detailed demographic characteristics (age, sex, and ethnicity - PASEF) is a demographic model designed to forecast detailed demographic characteristics at a neighborhood level. The detailed demographic forecast comes directly from DEFM, but requires aggregating the single year of age detail into the five-year age groups used in PASEF, and an adjustment for special populations. The model projects population for 18 five-year age groups (0-4, 5-9…,80-84, and 85+) broken down by gender and ethnicity for the region and smaller geographies.

The final stage in PASEF distributes the demographic characteristics estimates from the census tracts to the MGRAs. The model assumes that each MGRA has the same demographic characteristic distribution as the census tract in which it lies.

7

1.6 DATA SOURCES

Data Source(s) Model(s)

Housing U.S. Census Bureau, San Diego County Assessor, local jurisdictions

DEFM, UDM

Jobs (by industry) U.S. Bureau of Labor Statistics, California Employment Development Department, U.S. Department of Defense, local jurisdictions

DEFM, UDM

Labor market (employment, unemployment, labor force participation)

U.S. Bureau of Labor Statistics DEFM

Population and demographic characteristics

U.S. Census Bureau, California Department of Finance

DEFM, UDM, PASEF

Price levels and inflation U.S. Bureau of Labor Statistics, National Association of Realtors, DataQuick Information Systems

DEFM, IRCM

Public finance California Department of Finance DEFM

Travel times SANDAG transportation model IRCM, UDM

United States projections U.S. Census Bureau, and economic projections purchased from private-sector vendor (varies depending on series)

DEFM

Vital records (births, deaths) California Department of Health DEFM

8

CHAPTER 2: TRANSPORTATION MODELS

2.1 INTRODUCTION

Transportation models are designed to compute transportation system impacts such as traffic volumes, traffic speeds, and transit ridership for transportation network and policy alternatives given land use and demographic forecasts from the IRCM, UDM, and DEFM. SANDAG makes use of an advanced four-step transportation modeling process of trip generation, trip distribution, mode choice, and assignment to forecast personal travel activity in the San Diego region. Figure 2illustrates how the four-step process is run in iterations or stages and combined with additional input and output functions to form the complete modeling chain.

TransCAD, created by Caliper Corporation, is a transportation planning computer package used by SANDAG to provide a framework for performing much of the computer processing involved with modeling. Another software package used extensively in the modeling process is ArcInfo, distributed by Environmental Systems Research Institute, Inc. This geographic information system (GIS) maintains, manipulates, and displays transportation, land use, and demographic data. SANDAG has written numerous FORTRAN and Visual Basic programs that provide linkages between TransCAD and ArcInfo. Other programs manipulate data and perform some modeling functions such as trip generation and mode choice.

SANDAG has extensive experience with both transportation modeling software and ArcInfo. SANDAG used TRANPLAN between 1981 and 2004 for a wide range of modeling applications, and then switched to TransCAD in 2004. ArcInfo first was installed at SANDAG in 1985. TRANPLAN and ArcInfo have been used in conjunction for transportation modeling since 1987.

The SANDAG transportation modeling and database maintenance is performed on a mix of Windows servers and individual personal computers. The time necessary to execute the entire transportation modeling process on these machines is about 16 hours, which is felt to be the maximum run time in order to provide reasonable turnaround on modeling projects. Turnaround time is important since SANDAG performs hundreds of model runs each year that range in scope from quantifying traffic impacts of individual development projects to evaluating system level impacts of alternative growth scenarios and transportation facilities for the Regional Transportation Plan (RTP). All of these modeling projects make use of the same basic procedures and data sets. The complexity of modeling procedures and the number of zones, time periods, iterations, modes and other factors determine model execution time. All of these factors have been evaluated so that the model functions within the 16-hour limit.

9

Fig

ure

2:

Tran

spo

rtati

on

Mo

deli

ng

Pro

cess

10

Before running the models in production, a considerable amount of time is spent calibrating model parameters and validating model accuracy. The purpose of calibration is to develop model relationships that can accurately reflect existing travel behavior, so there is confidence the models can be used to forecast future travel behavior. For example, the models correctly estimate current trolley ridership so they should be able to forecast future ridership on proposed trolley extensions and on new bus rapid transit service. Most recently the models were recalibrated to year 2008 conditions before use in the 2050 RTP. This process included a separate validation exercise where modeled results were compared to a variety of observed data sources from the last two decades. A detailed report of the validation exercise was released by SANDAG in June of 2011 (San Diego Association of Governments Travel Demand Model Validation Report). Additionally, the travel model was subjected to a variety of sensitivity tests to assess the response to potential policy questions. The SANDAG Transportation Model Sensitivity Analysis and Report is available on the SANDAG Regional Models website. The next section of this chapter (Section 2.2) describes the survey data used in this calibration and validation process.

As indicated in Figure 2, the modeling process can be broken down into four steps and four phases. In the model input phase, growth forecast data files are assembled, and highway and transit networks are coded. Preparing inputs to the models is often the most time consuming part of a modeling project. Several sections of this chapter document the three major inputs for the rest of the modeling process:

growth forecast inputs used to describe existing and planned land use patterns and demographic characteristics (Section 2.3)

highway networks used to describe existing roadway facilities and planned improvements to the roadway system (Section 2.4)

transit networks used to describe existing and planned public transit service (Section 2.5)

After preparing model inputs, there are four major steps of trip generation, trip distribution, mode choice, and assignment, along with a minor function of path-building and skimming. Additionally, there is a parallel process for the truck model before the results are combined in for highway assignment. There is a section describing each of the modeling steps listed in the order that they are executed as follows:

trip generation (Section 2.6)

path-building and skimming (Section 2.7)

trip distribution (Section 2.8)

mode choice (Section 2.9)

truck model (Section 2.10)

highway assignment (Section 2.11)

transit assignment (Section 2.12)

11

One of the complexities of the modeling process is that transportation measures needed as input to a modeling step may not be produced until later in the modeling process. For this reason there are numerous iterations through the modeling process. As a starting point, the first-stage of the modeling process makes use of simplified trip distribution, mode choice, and highway assignment procedures to produce initial highway travel time forecasts for use in the subsequent feedback loop phase.

Processing may stop after the first stage for small scale modeling projects to reduce costs for outside clients without seriously compromising mode accuracy. However, regional planning studies proceed on to a feedback loop phase and a final stage of mode choice and assignment. These additional stages incorporate the effects of traffic congestion on destination choice and mode choice. For example, people in heavily congested corridors may choose shopping locations closer to home rather than contend with traffic delays. Conversely, widening a congested freeway may make fringe housing more accessible and increase average commute trip lengths. This relationship between trip length and congestion is one aspect of induced travel (Section 2.13).

The details of performing the trip distribution, mode choice, and assignment steps vary depending on which stage of the modeling process is being executed. These variations are described in the sections of the report dealing with each modeling step.

2.2 SURVEY INPUTS

The transportation models make use of survey data to establish relationships between input variables and model-estimated results. For example, trip generation rates are applied to dwelling units from the growth forecasting process to determine the number of trips generated from residential areas. Data collection is costly and time consuming, so surveys are conducted relatively infrequently. This normally does not create a problem since underlying model relationships are relatively stable over time.

The following eight survey groups provide most of the calibration data for the transportation models.

1995 & 2006 Travel Behavior Survey

2001 Caltrans Statewide Survey & 2009 National Household Travel Survey

2001-2003 & 2009 San Diego Regional Transit Survey

External Trip Surveys

Traffic Generation Studies

1991 San Diego Visitor Survey

Census Transportation Planning Package & American Community Survey

2000 Market Research Survey

12

Additional data sources are used to verify model estimates with observed data. Major sources of validation data are traffic counts from Caltrans and local jurisdictions, transit passenger counts from the SANDAG Transit Passenger Counting Program, and SANDAG Vehicle Occupancy and Classification Study.

2.2.1 1995 & 2006 Travel Behavior Surveys

Every ten years SANDAG conducts an extensive travel behavior survey which serves as the primary source for model calibration data. The 2006 and 1995 Travel Behavior Surveys are the basis for the existing models. The 1995 and 2006 surveys had 2,050 and 3,670 San Diego households interviewed, respectively. Survey respondents provided a complete listing of trips made on a survey data with information such as start and end location, start and end time, trip purpose, and trip mode. Information also was collected about household, household member, and household vehicle characteristics. Survey responses were expanded to regional totals and tabulated to develop the following calibration data.

Trip generation rates for the trip generation model

Trip length frequency distributions for the trip distribution model

Non-transit mode use percentages for the mode choice model

2.2.2 2001 Caltrans Statewide Survey & 2009 National Household Travel Survey

In 2001 Caltrans conducted a statewide activity-based household travel survey which had a San Diego sample size of 1,200 households. Of these households, 104 also had vehicles instrumented with GPS recorders. The resulting GPS vehicle trip sample was extensively analyzed to develop underreporting correction factors. These GPS correction factors were applied to both the 1995 and 2001 Travel Survey results prior to model estimation. In 2009 the National Household Travel Survey was conducted and included an oversample of San Diego households funded by Caltrans. Data from the survey will be used for validation purposes.

2.2.3 2001-2003 & 2009 Regional Transit Survey

Every five years SANDAG, in cooperation with transit operators, conducts an on-board transit survey to obtain transit trip and transit user characteristics. The most recent survey, conducted between 2001 and 2003 and also in 2009, provides data used to calibrate the transit portion of the mode choice model.

In the transit survey, surveyors stationed on-board buses, trolleys, and the COASTER distributed questionnaires to passengers over 12 years of age as they boarded the vehicle. Passengers filled out forms while they completed their trip and dropped off forms as they got off vehicles. About 50,000 surveys were returned with useable information, which were tabulated to obtain the following calibration and validation data:

Transit trip shares by income level, trip purpose, and trip length for mode choice calibration

Park-and-ride locations for coding transit network park-and-ride nodes

13

Walk access distance distribution to set maximum walk access distances

External transit trip table for external trip modeling

Relationship of total boardings to linked trips for transit assignment validation

Access mode percentages for transit assignment validation

Zone-to-route trips for transit network validation

Zone-to-zone trip tables for transit network calibration

2.2.4 External Trip Surveys

Roadside interview surveys are conducted periodically to determine the travel characteristics of trips coming into or passing through the San Diego region from outside the region. These surveys are difficult to collect since motorists must be stopped as they are entering or leaving the region and asked a series of questions about trip characteristics. Surveys conducted between 1986 and 1999 and also in 2006 are used to obtain the following parameters:

trip purpose distributions for the trip generation model

external trip lengths for the trip distribution model

through trips which are added to internal and internal-external trips

2.2.5 Traffic Generator Studies

These studies are conducted periodically to collect site level traffic data. The last major study, completed in 1999, placed traffic counters and video cameras at all entrances and exits to 26 survey sites, which included shopping centers, offices, schools, and housing developments. Traffic counts were totaled and averaged over five days to obtain average weekday trip generation totals for the sites. Trips rates then were calculated based on site characteristics such as number of employees, acres, and dwelling units. Travel behavior survey trip rates for nonresidential uses were adjusted to agree with traffic generator trip rates to correct for under-reporting of trips in travel behavior surveys.

2.2.6 1991 Visitor Survey

San Diego is a major convention and vacation destination. A small-scale visitor survey was conducted during the months of July, August and September 1991 to obtain a more complete picture of visitor travel patterns. Surveyors stationed outside selected hotels and tourist attractions questioned passers-by about their trips made on the previous day. Visitor trip generation rates and visitor trip lengths for gravity model calibration were obtained from this survey.

2.2.7 Census Transportation Planning Package and American Community Survey

Since 1960, the decennial census “long form” has included a series of transportation related questions about work trips, including travel time, travel mode, and employment location. The Census 2000 Transportation Planning Package (CTPP) data had limited usefulness for model

14

calibration due to Census Bureau data suppression procedures for protecting confidentiality. The American Community Survey (ACS) is a continuous survey process that is replacing the long form. ACS results will be available more frequently although suppression issues are expected to continue. Early ACS information was used for interregional commute information.

2.2.8 2000 Market Research Survey

In 2000 the Metropolitan Transit Development Board conducted a stated preference survey of 858 San Diego households to identify traveler attitudes towards new forms of public transit by market segments. The resulting datasets were used to estimate mode choice model parameters for comparison with models estimated from traditional revealed preference travel surveys.

2.2.9 Traffic Counts

Traffic counts, used in model validation, are obtained from a variety of different sources. The Caltrans program called PeMS (Performance Monitoring System) which is a system that provides counts at 630 directional freeway locations. Counts at additional locations will be available as the ramp metering system is expanded. PeMS outputs more detailed counts by five minute intervals and also outputs speed estimates which can be compared with model-estimated speeds. A third Caltrans program counts freeway on and off-ramps on a three-year cycle.

The City and County of San Diego and some of the other cities conduct comprehensive traffic count programs and maintain computerized count files. SANDAG converts traffic count stations into an ArcInfo point coverage and subsequently matches counts to network links. SANDAG also collects counts to produce a biennial Traffic Flow Map. Counts from this program are used for cities without computerized count files.

2.2.10 Transit Passenger Counts

SANDAG has operated a Passenger Counting Program since 1979, within which every bus route is counted once a year. Trips are counted by stationing surveyors on-board transit vehicles. Surveyors record the number of passengers boarding and alighting at each transit stop. The number of passengers on board vehicles between stops is computed from the boarding and alighting data. Surveyors also record arrival and departure times at selected time points along a route. An up-to-date transit route and stop inventory is maintained as part of the Passenger Counting Program. Screenlines (imaginary lines that run across multiple transit paths) can be created to look at passenger flows through a corridor. Summing the number of passengers on-board each route that crosses a screenline yields the total screenline count. This helps equalize factors, such as changes in frequency or route path to look at corridor passenger flows over time.

Bus stop inventories from the Passenger Counting Program provide bus stop locations for transit network coding. The Passenger Counting Program also produces the following validation data for checking the accuracy of transit assignment estimates.

Ons and offs at stops

Screenline counts

15

Boardings by route and mode

Transit link passenger volumes

2.2.11 2006 Vehicle Occupancy and Classification Study

SANDAG and Caltrans station surveyors at 22 freeway locations to monitor trends in vehicle occupancy and vehicle classification. These counts are taken every five years. Vehicle occupancies from the most recent 2006 study were used to verify mode choice model estimates and vehicle classifications were input to air pollution emission modeling.

2.3 GROWTH FORECAST INPUTS

The number and location of people living and working in the San Diego region largely determine the amount of travel activity that occurs. The previous chapter described how population, employment, and land use forecasts are produced at the SANDAG smallest unit of geography, the over 800,000 parcel polygons.

Parcel polygons provide flexibility for designing zone boundaries, however, the current generation of transportation models is unable to efficiently use parcel level forecasts. Therefore, the parcel forecasts are aggregated to three different levels of a nested zone system that has been designed to maximize model accuracy while minimizing execution time. There are 21,633 non-motorized zones (MGRA) that are the smallest unit of geography used by the transportation models. This detailed zone system enables the models to accurately compute the amount of activity within walking distance for input to the mode choice model. The highway assignment model makes use of a 4,682 traffic analysis zone system (TAZ) to obtain link level traffic volumes. Finally, a 2,000 trip distribution zone system (TDZ) is used within the feedback loop process to determine trip flows between zones. Figure 3 shows the nested relationships between MGRA (thin black line), TAZ (grey line), and TDZ (red line).

16

Figure 3: Transportation Model Nested Geographies

TAZs range in size from individual blocks in Centre City San Diego up to 150 square miles in sparsely developed rural areas. SANDAG uses a relatively large number of TAZs to reduce the need for developing sub-zones to address local planning needs. TAZ boundaries attempt to group areas with similar land uses and access to the transportation system.

The zone systems include 12 external zones located where major roads cross the county line. These external zones are used to represent travel between the San Diego region and other areas, such as Riverside County, Orange County, and Mexico.

Most studies use regional growth forecasts directly, however the purpose of some studies is to evaluate impacts on the transportation system of proposed land use changes. For example, Environmental Impact Reports may identify traffic volumes on roadways in the vicinity of proposed development projects with and without the proposed project. Cities may use the models to evaluate traffic impacts of proposed General Plan changes. Recently there has been increasing interest in smart growth development and SANDAG has evaluated regional land use alternatives that implement smart growth principles to varying degrees. These land use studies modify the SANDAG standard forecasts within the study area to reflect the proposed changes.

2.4 HIGHWAY NETWORK INPUTS

At many points in the modeling process a computerized representation of the highway system is needed for obtaining inputs to the models. SANDAG uses GIS software to maintain highway information in an ArcInfo master transportation coverage. Coverage is an ArcInfo term used to describe all the individual files which together represent a geographic system in digital form. This

17

network coverage includes existing and planned freeways, toll lanes, HOV lanes, managed lanes, ramps, surface streets classified on general plan circulation elements, and some local roads needed for network connectivity. The network coverage also includes zone connector links, which are used to schematically represent how traffic from zones accesses the street system.

The SANDAG master network files reflect facility improvements proposed in the most recent RTP and General Plan circulation elements from each jurisdiction in the region. The purpose of many planning studies is to look at the impacts of alternative highway improvements. Once these planning studies are completed, recommended facility changes are coded into the master network. At a more local level, the models are often used to evaluate traffic volume impacts of subdivision plans that fine-tune standard highway assumptions.

2.4.1 Highway Facilities

Alignments for existing roads were originally obtained from SanGIS (another agency responsible for maintaining various geographic databases) and have been updated extensively based on high resolution digital aerial photography. Alignments for planned roads are derived from a number of different sources including Caltrans route location studies, local general plan circulation elements, environmental impact reports, and corridor studies.

ArcInfo automatically creates nodes at at-grade intersections. Coders insert additional nodes where traffic signals, stop signs, and ramp meters occur in between street intersections. The ArcInfo dynamic segmentation function also is used to code routes that indicate where turns are prohibited by physical barriers or signs.

2.4.2 Highway Attributes

Once highway alignments have been determined a large number of attributes are coded about each highway segment and node. A number of attributes are informational, such as street name, node numbers, link ID numbers, and functional classification. Other attributes, used to calculate travel time, include segment length (computed by ArcInfo from highway alignments), posted speed, one/two-way operation, and type of intersection control. Another set of attributes used to calculate capacity includes number of lanes, median condition, number of freeway auxiliary lanes, type of operation (mixed flow or high occupancy vehicle only), type of intersection control, and the number of through, left turn, and right turn lanes at intersection approaches. The phasing of new roads, improvements to existing roads, and in some cases the deletion of existing roads is identified using another set of attributes. Finally, additional attributes provide cross-references to traffic count files that are used for model calibration.

Many base year physical attributes can be obtained from high resolution digital photography. These include one/two way operation, location and type of intersection controls, median condition, and the number of main lanes, auxiliary lanes and through, right turn, and left turn intersection approach lanes. Planned highway improvements are obtained from local circulation elements, Regional Transportation Improvement Programs, local Capital Improvement Programs, and long range Regional Transportation Plans.

18

2.4.3 Highway Capacities

Highway network coverages for specific model years and alternatives are selected from the master transportation coverage. Computer programs convert these ArcInfo coverages to TransCAD highway networks by reformatting data items and computing additional attributes needed in the modeling process, such as capacities, travel times, distances, and costs from attributes coded on coverages.

Two capacities are calculated for each direction of a highway link: (1) intersection capacity which is the amount of traffic that can be accommodated by an intersection approach at the end of a link, and (2) mid-link capacity which is the amount of traffic a link could accommodate without intersection controls. Both intersection approach and mid-link capacities are computed on an hourly basis and then factored to A.M. peak period, P.M. peak period, and off-peak period capacities using hourly to time period expansion factors of 2.25, 2.85, and 11.1, respectively. These expansion factors are overridden with location specific expansion factors on freeways where hourly traffic counts are available.

Mid-link Capacity

Mid-link capacity calculations vary for four different types of facilities: freeways, freeway HOV/managed lanes, urban streets, and rural highways. Hourly directional freeway capacities are calculated using the equation below which multiplies the number of main lanes by a per lane carrying capacity supplied by Caltrans that varies between 1,900 and 2,100 vehicles per hour per lane. Auxiliary lane capacity, assumed to be 1,200 vehicles per hour per lane, is added to main lane capacity. A capacity increase of 10 percent is phased in between 2010 and 2030 on freeway segments that are currently not metered, but are slated for ramp metering in the future.

where: fwyc = hourly directional freeway capacity (vehicles per hour) for link ml = number of mixed-flow main lanes on link mlc = capacity per lane for link (vehicles per hour per lane) al = number of auxiliary lanes on link

It is assumed that single occupancy vehicle (SOV) usage of managed lanes will be controlled to keep speeds from falling below level-of-service “D” or 1,600 vehicles per lane. The number of lanes on managed lane facilities also can vary by time period, such as the existing Interstate 15 (I-15) HOV lanes which operate as two lanes southbound in the morning and two lanes northbound in the afternoon.

hovc(tm) = hl(tm) x 1600

where: hovc = hourly directional HOV capacity (vehicles per hour) for link in time period “tm” hl = number of HOV lanes on link in time period “tm”

Mid-link capacities for urban street segments are calculated using the equation below. Two-lane rural highways typically can accommodate much less traffic and a lower capacity of 950 vehicles per hour per direction is assumed for these facilities.

)m(lnurbc 2<2003001800×= --

)1.10.1()1200( almlcmlfwyc

19

where: urbc = urban street mid-link capacity for link ln = number of mid-block lanes on link m = median code (0 or 1 indicates no median)

Intersection Approach Capacity

Because the most significant traffic congestion on urban streets often occurs at traffic signals, procedures have been developed to represent individual signal approach capacity within the model using the following equation.

xc = (tl x 1800 x gc(fc,xfc,napp) + (rl + ll) x tlc(fc)) x 1.0 1.1

where: xc = intersection approach capacity for link tl = number of through lanes at intersection approach gc = green-to-cycle time ratio fc = functional classification of street xfc = functional classification of cross street napp = number of intersection approaches rl = number of right turn lanes at intersection approach ll = number of left turn lanes at intersection approach tlc = per lane turn lane capacity that varies by functional classification

While actual signalized operation is very complex, this equation captures the primary factors that determine capacity. A through lane capacity of 1,800 is multiplied by the number of approach lanes that have been coded. The green-to-cycle time (GC) ratio is a traffic engineering term that quantifies the fraction of total cycle time that is in the green phase for each intersection approach. Within the model, GC ratios vary between 0.09 and 0.84 depending on the functional classification of intersecting streets and number of approaches. For example, a prime arterial that intersects with another prime arterial would have a lower capacity than one with the same approach lane configuration that intersects with a local street. Similarly, two and three legged intersections have higher capacities than four legged intersections because total cycle time is apportioned to fewer phases.

A turn lane capacity that varies between 100 and 250 vehicles per lane per hour depending on the functional classification of the street is multiplied by the number of coded right and left turn lanes and added to through lane capacity. Finally, future capacity increases of up to 10 percent are phased in on regionally significant arterials as a result of improved signal coordination assumed in the 2050 RTP.

A ramp meter is a special type of signal that controls the number of vehicles that can get on a freeway during peak periods. Metering rates are determined by Caltrans and vary from ramp to ramp depending on the location of the ramp and the severity of up-stream freeway congestion. An average capacity of 1,000 vehicles per ramp meter is assumed unless location specific metering rates are available.

20

Stop signs also impose significant reductions in the capacity of surface streets. The model computes capacities of two-way and all-way stop-sign controlled approaches using equations similar to the signalized intersection equation shown above. In a few locations capacities are affected by toll booths and rail crossings where special capacity calculations are made.

Intersection capacity considerations are turned off for freeways and other links that have no intersection controls by setting the capacity to a maximum value.

2.4.4 Highway Travel Times

As with capacities, separate link times and intersection times are computed for each highway segment. Highway link travel times are computed using the following equation:

tt(tm)= lg s

where: tt(tm) = travel time each link and time period (A.M. peak period, P.M. peak period, and off-peak) lg = length on link s = posted or adjusted speed on link

Travel times represent the free-flow link time (ArcInfo computed link length divided by the posted speed). During the calibration process posted speeds may be varied by up to plus or minus 10 miles per hour to better match model estimated traffic volumes with traffic counts. Adjusted speeds replace posted speeds where coded.

Intersection times represent the delay time encountered at traffic signals and other intersection controls under uncongested conditions. An intersection delay time of 10 seconds per signal or stop sign accounts for idling time, acceleration/deceleration time, and the likelihood of being stopped at a signal. Baseline ramp meter times of one minute are assumed for peak period networks. Ramp meters are assumed to be turned off during off-peak hours and so no off-peak ramp meter delays are added.

These input link and intersection travel times reflect free-flow conditions without congestion. Individual link and intersection congestion delays are computed later in the highway assignment step based on forecasted, link-specific traffic volumes.

2.5 TRANSIT NETWORK INPUTS

Transit modeling requires coded transit networks that represent existing and planned conditions. Like highway networks, transit networks are maintained in the master transportation coverage using ArcInfo. However, transit network coding is more complicated than highway coding because of the need to describe how individual transit routes operate over the transit system. Also added is the concept of transit modes which group transit routes with similar operating characteristics.

Table 1 describes the seven transit modes and gives examples of existing routes in each category. Bus rapid transit (BRT) and rapid bus modes represent new types of transit service that will soon be implemented. BRT service would have stations similar to commuter rail and trolleys, and operating characteristics midway between rail and bus service. BRT service would be provided by advanced

21

design buses operating mostly on HOV lanes with Direct Access Ramps. Rapid Bus service would also be provided by advanced design buses operating largely on surface streets with priority transit treatments.

Table 1: Transit Mode Definitions

Mode Number Description Examples

4 Commuter Rail COASTER

5 Light Rail/Street Car Trolley, Sprinter

6 BRT (Regional) None (Proposed)

7 Rapid Bus (Corridor) None (Proposed)

8 Premium Express Bus MTS Routes 810,820,830

9 Express Bus SDTC Routes 20, 50, 150

10 Local Bus SDTC Routes 1-9

2.5.1 Transit Facilities

Most transit routes run over the same streets, freeways, HOV lanes and ramps used in the highway networks. As a result the only additional facilities that are added to the transportation coverage for transit modeling purposes are:

Trolley, streetcar, and commuter rail lines,

Streets used by buses that are not part of local general plan circulation elements,

Transit exclusive right of way (transitways) that have been proposed as part of the future transportation system.

Nodes are located at each transit stop. The ArcInfo dynamic segmentation feature is used to maintain historical, existing, and planned transit routes. Existing routes and stops are modified up to several times a year as new time tables are published. A transit scheduling system (HASTUS) provides accurate existing bus transit stop information. Near term transit route changes are drawn from short-range plans produced by transit agencies. Longer range improvements are proposed as a part of the RTP and other transit corridor studies.

2.5.2 Transit Attributes

Transit node attributes describe stop type and park-and-ride availability at each node. Most transit network attributes are associated with routes. These attributes include transit operator, mode, and most importantly, frequency of service by time period (A.M. peak period, P.M. peak period, mid-day, and night). Initial wait time and transfer time are significant factors that affect transit use and are computed from service frequencies. Existing frequencies are calculated based on published time schedules. Planned service frequencies may be policy based, such as establishing a minimum 15-minute frequency. Alternatively, future frequencies may be computed using vehicle capacity assumptions and forecasted ridership from previous model runs. For example, passenger demand may indicate that a 10-minute frequency is needed on some routes to avoid bus overcrowding.

22

2.5.3 Travel Times

Transit networks for different years and alternatives are selected from the master transportation coverage. Transit travel times on links between rail stations and bus stops are computed at this time. The following equation is used to determine bus speeds.

)()()( mdtbstmhttmbt

where: bt = bus travel time on link during time period “tm” ht = congested highway travel time bs = number of bus stops on link dt = per stop delay time by mode “m”

Bus travel times are assumed to be a function of the number of bus stops on a link and highway travel time. Since highway times include congestion effects from the highway assignment step (Section 2.10), bus travel times are recomputed at different stages of the modeling process. Highway travel times are modified for the following special conditions before computing bus times:

Ramp meter delays at meters with HOV bypass ramps are assumed to be one-third of SOV times.

The maximum legal speed limit is used for the free-flow bus speed on freeways, whereas highway free-flow freeway speeds are set at five MPH above the speed limit to reflect observed speeds from survey data.

Bus speeds of 35 MPH are assumed on selected freeways that allow buses to run on shoulder lanes when speeds on adjacent general lanes fall below 35 MPH.

Travel times on arterial streets used by BRTs are reduced by 10 percent to reflect the effects of bus priority treatments.

Stop delay times of 30 seconds for BRT, Rapid, and Express Bus service and 18 seconds for local bus routes are assumed. Express and local bus stop delays were calculated from observed data and include the effects of acceleration/deceleration, dwell time for boarding passengers, and likelihood of stopping at an individual stop. BRT and Rapid Bus stop delays were assumed to be similar to express bus based on existing systems in other regions.

Travel time procedures for rail service differ from the bus procedures described above. Where COASTER and trolley routes already exist, speeds are obtained from published time schedules. Since rail service is normally not affected by highway congestion, base year station-to station travel times are assumed to remain unchanged over the forecast period. The travel time effects of a proposed COASTER tunnel and several additional COASTER and trolley stops are obtained from outside studies.

Travel times for proposed trolley extensions are computed based on an assumed peak operating speed, a station dwell time of 20 seconds, and acceleration/deceleration times that vary according to station spacing and peak speed.

23

2.5.4 Fares

In addition to transit travel times, transit fares are required as input to the mode choice model. TransCAD procedures have been augmented to replicate the San Diego region’s complicated fare policies which differ among:

buses which collect a flat fare of between $2.00 and $5.00 depending on the type of service,

trolleys which charge a Flat fare of $2.50 ($2.00 for Sprinter), and

commuter rail which has a zone-based fare of between $4.00 and $5.50.

When transfers occur, the overall fare for the trip is set to the highest fare encountered. These fares represent cash fares and are factored later in the mode choice model to account for pass usage based on an analysis of survey data. Fares are converted to 1999 dollars for consistency with income data in the model and are assumed to remain constant over the forecast period.

2.5.5 Transit Walk Access

Accurately specifying transit access opportunities is an important part of the transit forecasting process. A series of programs generates access files based on ArcInfo transit networks, trip generation forecasts by NMZ, elevation grids, and walk barriers. First, transit stops within walking distance of each NMZ are determined using the following formula.

3*)()()()(),( 22 zszzmzabsnmzwfacysynmzxsxnmzsnmzd

where: d = distance between NMZ and stop “s” xnmz = x coordinate of NMZ centroid ynmz = y coordinate of NMZ centroid znmz = elevation of NMZ centroid wfac = walkability factor (1.1 or 1.5) xs = x coordinate of transit stop ys = y coordinate of transit stop zs = elevation of stop

Straight line distances arc computed between each MGRA centroid and transit stop. The region has been broken down into two area types: traditional neighborhoods with grid pattern streets; and suburban neighborhoods with curvilinear street patterns. Walkability factors of 1.1 for traditional neighborhoods and 1.5 for suburban areas are applied to straight line distances so that walk distances better approximate real world conditions. Elevation differences between NMZ centroids and stops are added to horizontal distances after weighting differences by a factor of three to account for the additional difficulty of walking uphill. A maximum walk distance of three-quarter mile is assumed. MGRA-stop connections with a distance greater than the maximum walk distance are discarded as are connections that cross walk barriers. A walk barrier coverage has been developed that contains features such as ridge lines, steep slopes, water body boundaries, freeways, and fenced property lines that could block walk access.

24

All rail stations and BRT stops, and selected local and express bus stops are designated as transit access points (TAP) for use in transit modeling. TAPs are located approximately every one-half mile along a bus route. The mode choice model uses a walk access file of MGRA-TAP connections that are obtained by generalizing NMZ-transit stop connections found in the previous step.

Access opportunities within Centre City are too complicated to represent with these procedures. Instead, the closest TAP to each Centre City zone is identified. A Centre City walk network is coded that allows access to other TAPs in Centre City without explicitly coding each MGRA-to-TAP connection.

2.5.6 Transit Auto Access

Many transit users, who are outside walking distance of transit or have inconvenient feeder bus service, drive to park-and-ride lots or are dropped-off at transit stops. A transit auto access file is generated that represents these drive to transit opportunities.

Transit ridership survey results were analyzed to identify transit stops with significant auto access activity. The vast majority of auto access trips go to formal park-and-ride lots at rail stations and other bus transit centers. Trolley stations without parking lots also have significant auto access usage since passengers can be dropped off or possibly park on-street. The remaining auto access trips occur at shopping center and church lots adjacent to bus routes where informal park-and ride usage can occur.

Formal park-and-ride lots, trolley stations without parking lots, and other informal lot locations are coded in an ArcInfo coverage. The parking space capacity of formal park-and-ride lots is coded, while a maximum of 25 spaces is assumed at other auto access locations. The mode choice model is able to capacity constrain auto access trips to produce more realistic ridership forecasts.

Within the modeling process, a computer program creates connections between TDZs and auto access locations, and calculates the peak period highway travel time and distance from the first stage highway assignment process. Rather than connecting all zones to all auto access locations, the following logic is used to limit the number of auto access connections to only those that are reasonable.

For each TDZ the program first analyzes formal park-and-ride lots along each commuter rail, light rail, BRT, and express bus transit route by direction. Auto access connections are created for the closest two upstream and downstream park-and-ride lots within a maximum eight mile drive distance.

For each TDZ the program next creates connections to the closest informal lot along each directional transit route that is within a maximum drive distance of two miles for local bus routes or four miles for commuter rail, light rail, BRT and express bus routes.

TDZs are connected to the closest formal park-and-ride lot when no other connections have been found.

25

The mode choice model treats trips that drive to transit and trips that are dropped off as two different modes. However, transit survey results did not show significant differences in the locations where these two types of transit access trips occur. Therefore a single auto access file is used to represent connections for both transit auto access modes.

2.6 TRIP GENERATION

The purpose of trip generation is to estimate the number of trips entering and leaving each zone on an average weekday for each forecast year. These trip end forecasts reflect new development, redevelopment, demographic, and economic changes that occur over time, using the inputs shown in Table 2.

The model computes person trips, which account for trips by all forms of transportation including automobiles, light-duty trucks, taxicabs, motorcycles, public transit, bicycling, and walking. Trips are generated for ten trip types: home-work, home-college, home-school, home-shop, home-other, work-other, other-other, serve passenger, visitor, and regional airport. These trip types are designed to group together trips with similar travel patterns. Trips are generated for each parcel polygon from the growth forecast model and then aggregated into the three zone systems used in the transportation models: non-motorized zones (MGRA), traffic assignment zones (TAZ), and trip distribution zones (TDZ).

Table 3 illustrates the output from the trip generation model.

Each trip has two trip ends and the trip generation model calculates trip ends separately. One end is classified as a trip production and the other end as a trip attraction. The home end of home-based trips is defined as the production end and the other end is defined as the attraction end. The work end of work-other trips is defined as the production end and the other end as the attraction end. Other-other trip ends are split evenly into trip productions and trip attractions. Over a 24-hour period, roughly the same number of trips will originate in a zone as are destined there. However, residential zones will generate primarily trip productions while nonresidential zones will generate primarily trip attractions. The production/attraction distinction is important for the trip distribution model discussed in the next chapter.

Table 2: Trip Generation Model Inputs

Input Source

Dwelling units by structure type UDM

Population by age category UDM

Land use acres by land use type UDM

Employment by land use type UDM

Unique generator trips Traffic counts

External trips Traffic counts, IRCM

Trip rates Travel behavior surveys

Traffic generator studies

Regional control variables DEFM, other studies

26

Table 3: Trip Generation Model Output

Output

Person trip productions and attractions by zone and trip purpose

Income distribution of trip productions

Highway terminal times

Trip generation reports

2.6.1 Model Structure

Trips from residential areas are calculated by applying trip rates to the number of dwelling units in each parcel polygon categorized by three structure types: single family, multifamily, and mobile home. Total trips are split into productions and attractions by trip type and then balanced to regional control totals. The residential trip generation equation for non-school purposes has the following form.

a)pbfac(pu,ppustppufstpfrate(st)du(zone,st)tp(zone,pu ),()()

a)abfac(pu,ppustapufstafrate(st)du(zone,st)ta(zone,pu ),()()

where: tp(zone,pu) = number of person trip productions in zone for each purpose du(zone,st) = number of dwelling units in zone by structure type rate(st) = person trip rate for structure type pf(st) = fraction of trip ends that are trip productions by structure type ppuf(st,pu) = fraction of trip productions in each purpose by structure type pbfac(pu) = trip production balancing factor for each purpose ta(zone,pu) = number of person trip attractions af(st) = fraction of trip ends that are trip attractions by structure type apuf(st,pu) = fraction of trip attractions in each purpose by structure type abfac(pu) = trip attraction balancing factor for each purpose

Special procedures have been developed to estimate home-based university and home-based school trip productions that are largely generated by school age population subgroups.

a)pbfac(pu,pagerate(puge)pop(zone,atp(zone,pu ),)

where: tp(zone,pu) = number of person trip productions in zone for each purpose pop(zone,age) = zone population by five year age category rate(age,pu) = person trip rate by age category and purpose pbfac(pu) = trip production balancing factor for each purpose

27

Nonresidential trip rates are usually computed as trips per acre by 80 different land use categories. The form of the nonresidential trip generation equation is similar to the residential equation.

a)pbfac(pu,ppultppufltpfltrateluzone,lt))tp(zone,pu ),()()(

a)abfac(pu,ppultapufltafrate(lt)lu(zone,lt)ta(zone,pu ),()()

where: lu = land use acres lt = land use type

Employee-based nonresidential trip rates have also been calibrated for the 80 land use categories. Employee rates yield the same number of regional trips as acres rates, although they produce a different pattern of trip making across the region. Employee-based trip rates are used for home-work trips to better match CTPP data. In addition, Centre City San Diego uses employee-based rates for all purposes. Centre City densities are much higher than regional averages so that acre-based rates underestimate travel. Centre City also has a greater potential for growth than is indicated by land use redesignation that determines acre-based trips.

Most model applications use the acre measure of nonresidential activity elsewhere in the region since acre-based forecasts are easier to understand, are subject to less uncertainty than employment forecasts, and produce traffic volumes that better match ground counts. Building square footage is expected to be added to future growth forecasts which should overcome problems with both the acre-based and employee-based procedures.

Trips from elementary schools, junior high schools, senior high schools, and golf courses are generated on a per site basis rather than the acre-based methodology described above. Trips from external zones are based on traffic counts on roads crossing the county boundary. These base year trips are factored to future years using output from the Interregional Commuting Model and trends in past traffic counts.

Trip ends calculated by these equations are superseded for a small number of MGRAs that contain unique generators. Unique generators are major traffic generators where traffic counts indicate that applying standard trip rates would misrepresent actual trip making. Unique generators include all military bases, major tourist attractions, major beaches, Indian casinos, the University of California San Diego (UCSD), San Diego State University (SDSU), and Lindbergh Field.

Future year person trips are reduced by a small amount to reflect increased use of tele-working and e-commerce. Reduction factors of between one and eleven percent are applied to selected trip purposes and land use categories. The assumption is that some land use types such as office uses would be more likely to take advantage of tele-working than other uses such as hospitals. Table 4 shows the tele-work and e-commerce eligible land uses and their corresponding reduction percentages.

28

Table 4: Trip Generation Tele-Work Percentages

Land Use Type Series 12 Reduction

Industrial Park 7%

Planned Industrial 7%

Light Industry 7%

Regional Commercial 2%

Community Commercial 1%

Neighborhood Commercial 1%

Specialty Commercial 1%

Streetfront Commercial 1%

Other Commercial 1%

High Rise Office 11%

Low Rise Office 11%

Gov't Office or Center 11%

High Rise Office 11%

Other Public Service 2%

Other Heath Care 3%

SDSU or UCSD 3%

University or College 3%

Junior College 3%

School Dist. Office 11%

Mixed Use 67% 1%

Mixed Use 77% 1%

Mixed Use 67% 1%

Mixed Use 25% 1%

By definition total trip productions must equal trip attractions at the regional level. However, imbalances between the two trip ends can occur since productions are estimated independently from trip attractions. Residential trip generation equations at the zone level also are relatively straightforward and ignore demographic trends such as the aging of the population over time. To overcome these problems, zone level trip productions and attractions are totaled for the region and balancing factors are calculated to match regional control totals for each trip purpose and forecast year.

Regional control totals are obtained by factoring base year trip totals by changes in logically related demographic, economic, and trip variables produced by DEFM and other models. These variables are listed in Table 5.

29

Table 5: Regional Control Variables

Trip Type Variable

Home-work Total employment

Home-college Trip productions

Home-school Trip productions

Home-shop Retail sales

Home-other Service sector output

Work-other Total employment

Other-other Total trips

Serve passenger Auto passenger trips

Visitor Visitor sector output

Regional airport Air passenger trips

A TDZ income level distribution of home-based trip productions is needed as input to the mode choice model. While residential trips are being generated, dwelling unit trip rates by income level are applied and zone level trips by income level are accumulated.

Another mode choice model input is a set of production and attraction end highway terminal times which represent the time spent looking for a parking space and traveling from a parking space to a final destination. The trip generation model processes dwelling unit and land use data which enables weighted terminal times to be calculated for each zone. For example, multifamily dwelling units tend to have less convenient parking than single family houses, so the model assumes a one-minute single family terminal and a two-minute multifamily terminal time. Thus zones with larger concentrations of apartments will have longer production end terminal times.

2.6.2 Model Calibration

Trip generation rates are established from travel behavior survey data and traffic generator studies. Trip rates were initially calculated by expanding trips reported in the 1995 Travel Behavior Survey to the region, tabulating trip ends by purpose and land use, and dividing total trip ends by either the total number of acres, total number of employees, or total number of dwelling units in each land use category. A significant underreporting of trips is found with travel behavior surveys. Trip rates computed from the 1995 Travel Behavior Survey were adjusted to match trip rates found in the 2001 Caltrans GPS survey and traffic generator studies.

2.7 PATHBUILDING, SKIMMING, AND UTILITY CALCULATIONS

The purpose of this part of the modeling process is to find distances, times, costs and other impedances between zones for input to the trip distribution and mode choice models based on coded highway and transit networks described in Sections 2.4 and 2.5. As indicated in Figure 2, the skimming process is repeated numerous times during a model run in order to reflect changes in travel times due to the interaction between travel demand and highway congestion.

30

TransCAD makes use of a “generalized cost” measure, which combines the effects of time valued at $0.35 per minute and distance valued at $0.15 per mile. In the highway skimming process a minimum path algorithm finds the shortest generalized cost path between each zone pair and then calculates network impedances over the minimum cost path. Different network impedances are skimmed at different points in the modeling process. Initially TAZ level free-flow generalized costs are output for the first stage trip distribution model. In both the feedback loop process and in the final modeling stage, TDZ level distances, times, and toll costs are output for up to eight highway modes that are described in the mode choice section. Peak and off-peak period skims are created to reflect the varying level of congestion in the two time periods. After the skimming process, intra-zonal times and costs are added where appropriate using a “nearest neighbor” technique.

Two different techniques are used to determine non-motorized skims. Since most non-motorized trips are short distance trips, SANDAG detailed 21,633 non-motorized zones (MGRA) are used where possible to reflect the increased likelihood of walking and biking offered by mixed use developments. MGRA-to-MGRA distances are calculated using the equation shown in Section 2.5.5 for nearby MGRAs within 1.5 miles. For trips longer than 1.5 miles, non-motorized distances are calculated at the more generalized TDZ level from a subset of coded highway networks without freeways and ramps. This two level approach is used to properly account for most non-motorized trips without seriously increasing model execution time.

The transit skimming process calculates minimum cost paths between transit access points (TAPs) instead of zones. TAPs, described in Section 2.5.5, are selected transit stops that are used to represent access to the transit system. The following four sets of paths are created for modes defined in Table 1

A.M. peak period local bus (mode 10)

A.M. peak period premium service (all modes 4-10)

Mid-day local bus

Mid-day premium service

By creating separate premium service and local bus paths, the model is able to split trips between the two types of service. For example, some transit riders may choose local bus routes over faster, but more expensive rail service serving the same corridor. The A.M. peak period is used to represent peak period transit service and mid-day service is used to represent off-peak conditions.

Next, TransCAD determines the following transit impedances over minimum generalized cost paths between each pair of TAPs.

Number of transfers

Cash fare

First wait time

Transfer wait time

Transfer walk time

Commuter rail In-vehicle time (mode 4 )

Light rail In-vehicle time (mode 5 )

BRT in-vehicle time (modes 6 and 7)

Express bus in-vehicle time (modes 8 and 9)

Local bus In-vehicle time (mode 10)

Main mode indicator

31

First wait and transfer wait times are computed as one-half of the headway on routes serving the TAP pair. For example, a trip that gets on a bus arriving every 30 minutes and transfers to a trolley arriving every 15 minutes, would have a first wait time of 15 minutes and a transfer time of 7.5 minutes. Long first waits are factored down later in the mode choice model to reflect the inconvenience of infrequent service without over penalizing this service. A premium transit path could use a number of different modes. The main mode indicator indicates the mode that is used for the longest distance.

In the feedback loop process, trips are distributed based on a composite utility measure that combines peak and off-peak times and costs for highway, transit, and non-motorized modes. A special version of the mode choice model, described in Section 2.9, computes composite utility values by weighting utilities for individual time periods and modes by their share of total trips.

2.8 TRIP DISTRIBUTION

The trip distribution model links together person trip productions and attractions from the trip generation model to determine trip movements between zones. The model produces trip tables that contain a row for each production zone and a column for each attraction zone. Cells in the table contain the number of trips estimated between zone pairs. Table 6 and Table 7 list inputs to the trip distribution model and some of the major outputs.

The model is designed to modify trip patterns in response to new land use developments and transportation facility changes. For example, the opening of a new shopping center would shift trips from other nearby shopping areas to the new development. Another example would be the introduction of mixed-use development. In this case the model would yield shorter trip lengths by recognizing the increased opportunity for interaction between residential and commercial areas in the development. The model also modifies trip patterns as new roadways are added, accounting for one component of induced traffic.

The trip distribution model is also designed to account for land use characteristics that relate to transportation and behavior known as the “4Ds,” which stand for Diversity, Density, Design, and Destination Accessibility. The 4Ds measures are used to quantify urban characteristics such as compact mixed use areas associated with smart growth planning. More information on 4Ds in the SANDAG model can be found in the 4D Model Development report on the SANDAG Regional Models website.

Table 6: Trip Distribution Model Inputs

Input Source

Trip productions and attractions by zone and trip purpose Trip Generation Model

Impedances between zones Network Skimming

Highway terminal times and parking costs Trip Generation Model

Friction factors gamma function parameters Travel Behavior Surveys

32

Table 7: Trip Distribution Model Outputs

Output

Daily production-attraction person trips between zones by purpose

Trip distribution reports


A gravity model function distributes trips between zones. The model allocates trip productions from each zone to other zones in direct proportion to the number of attractions in other zones and in inverse proportion to the impedance between zones using the following formula.

),()),,((),(

),((

1

zpzabfacpuzxzptiffpuzxta

,pu)zazptiffta(za,pu)tp(zp,pu))t(zp,za,pu

n

where: t = Number of trips between zones “za” and “zp” for each purpose tp(zp,pu) = Trip productions in zone "zp" for purpose “pu” ta(za,pu) = Trip attractions in zone "za" for purpose “pu” ff = Friction factor for travel impedance “ti” and purpose “pu” ti(za,zp,) = Travel impedance between zones "zp" and "za" bfac(za) = Balancing factors for zones "za" and “zp’ n = number of zones

The term “impedance” is a measure of the difficulty of travel between two zones. The trip distribution model is run repeatedly in the overall modeling process shown in Figure 2. Different impedance measures are used at different stages of the modeling process. The previous section describes how impedance measures are calculated. First stage trip distribution uses generalized costs based on free flow times and distances as the impedance measure to distribute trips between 4,682 TAZs. The first stage model is designed to be the starting point for the feedback loop process or can be run as a standalone application for simple model applications.

The feedback loop process makes use of a more complicated measure of impedance which is the composite utility from the mode choice model based on times and costs of travel for peak and off-peak conditions. Using composite utilities instead of free flow generalized costs enables the feedback process to reflect more subtle effects of changes to the transportation system on travel patterns. For example, widening a freeway may reduce congestion and lead to longer distance trip lengths. Major transit investments could also produce transit travel times that are more competitive with the automobile and cause more trips to be made within the transit corridor. The feedback loop trip distribution model distributes trips between the more aggregated 2,000 TDZs so that reasonable model execution times are maintained.

33

Each trip purpose has a different average trip length that ranges from 3 miles for other-other trips to 19 miles for regional airport trips. A gamma function generates friction factors for each trip purpose using 4D factors that account for variations in density. These friction factors determine the likelihood of a trip being made in each impedance increment and are used in the trip distribution model to reflect trip length differences by trip purpose. For example shop trips, which are much shorter than commute trips, have friction factors that diminish more rapidly than work friction factors.

),(),(),( zazpticpumimfb ezazptiaff

where: a, b, c are model parameters mf = 4D mix (or employment) variable density factor by mixed index “mi” for purpose “pu” shown in Table 8 & Table 9

Table 8: Mix Variable Density Factor

Purpose Low Density Medium

Low Density Medium

High Density High Density

Home-Work 1.42 1 1 1

Home-College 1 1 1 1.4

Home-Education 1.2 1.2 1.2 1.3

Home-Shop 1.1 1.1 1.2 1.4

Home-Other 0.9 1 1.05 1.05

Serve Passenger 0.98 1 1 1

Table 9: Employment Variable Density Factor

Purpose Low Density Medium Density High Density

Work-Other 1 1.2 1.2

Other-Other 1.05 1.05 1.05

The 4D variables are defined by two category systems. One system categorizes the individual density variables by range: employment density, dwelling unit density, and total intersections. The employment ranges in Table 10 are used for identifying which factors shown in Table 9 are used for work-other and other-other trip purposes. Additionally, all 3 density variables are used for mode choice non-motorized and transit constants. The second category system uses ranges for a mixed index density variable whose formulation is based on previous work by Portland Metro. The mix index is then categorized using the breakpoints shown in Table 11.

34

Table 10: 4D Variable Density Ranges

Employment (floating number of employees per acre)

Dwelling Unit (floating number of

all dwelling units per acre)

Intersection (floating number of intersections per ½ mile buffer around

MGRA)

Low 0-9.99 0-4.99 0-79.99

Medium 10-29.99 5-9.99 80-129.99

High 30+ 10+ 130+

21sec

21sec

FactorEmploymentFactoritsDwellingUntionsInter

FactorEmploymentFactoritsDwellingUntionsInterMI

where: MI = Mix Index Dwelling Units = dwelling units density within ½ mile of TDZ centroid Employment = employment density within ½ mile of TDZ centroid Intersections = intersections within ½ mile of TDZ centroid Factor 1 = Mean Intersections / Mean Dwelling Unit Density Factor 2 = Mean Intersections / Mean Employment Density

Table 11: Mix Index Density Bins

Density Category Mix Index

Low 0-20

Medium Low 21-1750

Medium High 1751-6500

High 6500+

The model uses a set of friction factors for each trip purpose for the two stages of the modeling process. These friction factors remain unchanged over the forecast period; however, this does not mean that the model produces the same average trip length for all areas of the region and all forecast years. Rather the model reflects trip length differences resulting from different spatial locations of activities. Thus, rural areas that are far from most employment centers would have longer home-work trips than more centrally located areas near Centre City and other major employment sites.

A number of gravity model iterations are performed to bring gravity model attraction estimates in line with trip generation estimates. The first model iteration typically overestimates trips to highly accessible areas and underestimates trips to inaccessible areas. The program computes doubly-constrained balancing factors by dividing estimated productions and attractions into input productions and attractions. The resulting factors are applied to estimated trips in the next iteration. This process continues until the model closes with a criterion of 0.01.

35


Gravity model calibration involves adjusting gamma function parameters until model estimated trip length frequency distributions agree with observed trip length frequencies from the 1995 Travel Behavior Survey and 2001 Caltrans Statewide Survey. A TransCAD procedure estimates initial gamma function parameters that best fit an estimated trip length distribution to observed data. Manual adjustments to these gamma function parameters are necessary before finalizing the calibration process. After this calibration process is complete, K-factors are applied to selected zonal interchanges so that model-estimated trip patterns better match observed data for the following conditions:

Home-college trips to San Diego State University and University of California, San Diego

Coronado Bridge crossings

Inter-zonal other-other and work-other trips

High transit use corridors

2.9 MODE CHOICE

The mode choice model splits total person trips from the trip distribution model into trips by individual forms of transportation called modes. The mode choice model is designed to link mode use to demographic assumptions, highway network conditions, transit system configuration, land use alternatives, parking costs, transit fares, and auto operating costs. Table 12 and Table 13 list major mode choice model inputs and outputs.

36

Table 12: Mode Choice Model Inputs

Input Source

Person trips between zones by purpose Trip Distribution Model

Trip distances between zones Highway Network

Peak/off-peak period highway travel times Highway Network

Peak/off-peak period highway distances Highway Network

Peak/off-peak period highway tolls Highway Network

Non-motorized distances Non-motorized Network

Peak/off-peak period transit times Transit Network

Peak/off-peak period transit fares Transit Network

Transit walk and auto access assumptions Transit Network

Automobile operating cost per mile National Forecasts

Parking costs Parking Surveys

Highway terminal times Trip Generation Model

Trips by income level Trip Generation Model

School bus percentages Travel Behavior Survey

Model parameters Model Calibration

Daily to time period factors Travel Behavior Surveys

Table 13: Mode Choice Model Output

Output

Origin-destination highway vehicle trips between zones by time period

Transit trips between TAPs by time period

Mode choice reports

Many applications make use of the mode choice model. Generating transit patronage forecasts for trolley extensions and other transit improvements is a common use. This model also produces performance measures used to evaluate alternative transportation network and land use scenarios.

Proposed transit improvements reduce the time of making a trip by transit and the model shifts trips to transit from other modes in response to these improvements. Conversely, new or widened highways would increase speeds for motorists and the model would increase automobile travel at the expense of transit usage. This mode shift that occurs with highway improvements is one component of induced demand.

37

The model reduces automobile travel and increases transit and non-motorized travel in response to smart growth land use alternatives that guide development into areas around transit stations and into mixed-use areas. These land use changes reduce trip lengths and make non-motorized travel more attractive. Increasing development around transit stations shortens station access times and increases the number of trips accessible to transit.

The model assumes that travelers make logical and systematic decisions about which form of transportation to take based on knowledge of the time and cost of completing a trip by alternative modes. The model is sensitive to a wide range of facility improvements and policies; however, the model is currently insensitive to programs designed to alter mode use without altering times or costs, such as:

Advertising campaigns to increase the use of transit, bicycling, or ridesharing,

Rideshare matching programs,

Construction of bicycle lanes,

Replacing older buses to increase the attractiveness of transit, and

Providing safer and more comfortable transit stops.

A common misconception about the mode choice model is that it underestimates future transit use for expanded transit alternatives because model calibration is based upon current conditions. This should not be the case. The model estimates transit use for each zone-to-zone movement based upon the quality of transit service relative to other modes. Existing trolley corridors provide a basis for determining potential transit use with high quality transit service. As more light rail, BRT, and bus service is provided, the model recognizes the resulting transit service improvements and shifts travel to transit from other modes. As a result, the model forecasts a 46 percent increase of the work trip transit mode share between 2008 and 2035, when an expanded transit system is expected to be in place. This forecasted increase in transit mode share exceeds historical changes in transit mode share. For example, US Census Bureau statistics show San Diego’s work trip transit mode share increased by only 5 percent between 1980 (Census) and 2005-2009 (ACS) a time in which COASTER and Sprinter service was added and trolley service was expanded. The mode choice model was recalibrated as part of the Mid-Coast FTA New Starts project. This included a thorough review of the mode choice code by an outside consultant and update of the code and parameters to meet FTA guidelines.


SANDAG uses a nested mode choice mode that splits total person trips into 25 different submodes in a hierarchical fashion as illustrated in Figure 4. Initially, school bus trips are factored out of total trips. These trips affect a small percentage of all trips, but are a significant proportion of school trips. Next trips are split into auto, non-motorized, and transit modes. The term "auto" is used generically to include travel by automobiles, pick-ups, vans, light-duty trucks, taxicabs, and motorcycles. Auto trips are further divided into submodes based on vehicle occupancy (drive alone, two person shared-ride, and three or more person shared-ride vehicles) and type of facility.

38

Figure 4: Nested Mode Choice Model Structure

The RTP proposes a number of different types of restricted use freeway facilities. In addition to general purpose freeway lanes the RTP includes:

High Occupant Vehicle (HOV) Lanes. HOV lanes can only be used by buses and shared-ride vehicles with two or more occupants. Under some conditions HOV lanes may be restricted to vehicles with three or more occupants. HOV lanes now exist on I-5 north of I-805 and are planned for many other freeways.

Managed Lanes. These multi-lane HOV facilities are sometimes called high occupancy toll (HOT) lanes. In addition to shared-ride vehicles, managed lanes are open to drive alone vehicles that are willing to pay a toll. There are currently managed lanes on I-15 between SR 163 and the Del Lago Direct Access Ramp (DAR). Managed lanes are being extended on I-15 and are planned for portions of I-5, I-805, SR 52, SR 54, SR 78, SR 94, and SR 125. In certain situations a managed lane will restrict the free HOV usage to shared-ride vehicles with three or occupants. In this case, shared-ride vehicles with two or more occupants pay the same toll as drive alone vehicles.

Toll Roads. The SR 125 toll road is an example of this type of facility. Toll roads require all vehicles, both drive alone and shared-ride, to pay a toll in order to use the facility. Toll Roads are planned for SR 11 and for segments north of SR 78 to the county line on I-5 and I-15.

39

When vehicle occupancy categories are matched with facility types, the following eight auto submodes are possible.

Drive Alone/Non-Toll Trips. These trips, restricted to general purpose lanes, are made up of drive alone trips that choose not to pay to use toll roads or managed lanes.

Drive Alone/Toll Trips. These are toll-paying drive alone trips can use all types of facilities except HOV-only lanes.

Two Person Shared-Ride Non-Toll/Non-HOV Trips. These trips, restricted to general purpose lanes, include shared-ride trips that choose not to pay to use tolled facilities, and trips that are eligible to use HOV lanes but choose not to do so.

Two Person Shared-Ride Non-Toll/HOV Trips. These are shared-ride trips that choose to use an HOV lane, but are unwilling to pay a fee to use tolled facilities.

Two Person Shared-Ride Toll/HOV Trips. These are toll-paying shared-ride trips that can use all types of facilities.

Three or More Person Shared-Ride Non-Toll/Non-HOV Trips.

Three or More Person Shared-Ride Non-Toll/HOV Trips.

Three or More Person Shared-Ride Toll/HOV Trips.

HOV lanes in the RTP are proposed as two person HOV lanes until 2035, so the two person and three or more person shared-ride modes are allowed to use the same facilities. In 2035 and beyond only three or more person shared-ride modes are allowed on HOV lanes or ML/HOT lanes for free.

Non-motorized trips (pedestrian and bicycle trips) reflect the effects of land use, trip purpose, and competing transportation modes. However, estimation procedures do not allow non-motorized facility issues to be addressed. For example, bicycle paths are not explicitly coded and thus do not affect non-motorized trip forecasts. A more rigorous approach is prevented by the small scale of non-motorized facilities and the lack of before and after data showing the effects non-motorized improvements have on non-motorized travel.

Transit trips are subdivided by transit access modes (walk, drive, and drop-off) and transit ride modes (commuter rail, light rail, BRT, express bus, and local bus). The three transit access modes determine how transit riders get to and from transit stops. Riders who walk at both trip ends are classified as transit-walk trips. Drive access trips are those trips where the rider gets to or from a transit stop by either driving or carpooling. Trips where the rider is dropped-off or picked up are classified as drop-off trips. Access mode is important in quantifying transit level-of-service. Drive access usually shortens overall transit trip times, but is limited to locations with park-and-ride access.

Transit ride modes represent travel behavior and non-quantifiable service differences between the five modes. Parameters and coefficients for existing modes are estimated from survey information and asserted for modes that do not exist in the region. During the Urban Area Transit Strategy project it was concluded that the BRT service being proposed in the 2050 RTP would operate midway between bus and rail service.

40

The model computes mode use separately for two time periods, three income levels, and six trip purposes. The two time periods split travel into peak and off-peak hours. The peak period extends from 6:00 a.m. to 9:00a.m. and 3:00 p.m. to 6:00 p.m. The off-peak period covers the remaining 18 hours of the day. It is important to evaluate mode use separately for the two time periods because the quality of service can vary dramatically by mode. For example, transit operators often provide more frequent transit service during peak hours, reducing wait times for transit riders. Conversely, highway congestion is at its worst during peak hours making auto modes less attractive relative to transit.

Mode use also varies by income level. People in low-income households tend to own fewer automobiles and hence make more trips by transit and carpooling. People in upper-income households tend to be mode time sensitive and as a result choose modes based on time and convenience rather than cost. Households are split into three income categories:

low income households with annual incomes less than $30,000 (constant 1999 dollars)

high income households with incomes more than $60,000

middle income households made up of the remaining households

There also is a strong relationship between mode use and trip purpose. For example, most students are below driving age so school trips generate almost no drive alone trips but have a very high rate of transit, school bus, and non-motorized mode use. Home-other trips tend to be made with other household members, so that two and three or more person auto modes are more heavily used than other trip types. In order to reduce computer processing time, the ten trip purposes from trip generation and distribution are collapsed into six purposes for mode choice by combining home-shop and home other trips, and combining work-other, other-other, visitor, and regional airport trips.

2.9.2 Utility Computations

Before determining mode shares the model computes utility measures for each mode based on a combination of a number of time and cost components that are weighted using the coefficients shown in

41

Table 14: Mode Choice Time and Cost Coefficients

Table 15: Value of Time

Income Category Home-Based Work Home-Based Other Non-Home Based

Low $2.02 $1.02 $1.01

Mid $5.42 $2.74 $2.69

High $12.92 $6.40 $6.60

Utilities for auto modes are computed for each TDZ pair, time period, and income group and purpose using the following equation.

nctopncmidnclowipccoefcpmjziztmaudistjziztmtolljzptp

ptcoefjztermaiztermpptcoefjziztmauivtipmauu

/)),(**),,,(),,,(),(cos

),2()()(),1(),,,((),,(

where: auu = auto utility for mode “m” purpose “p” and income group “i” auivt = auto in-vehicle travel time between zones for mode and time period “t” tcoef = time coefficient for purpose from Table 14 termp = terminal time at production end terma = terminal time at attraction end pcost = parking cost at attraction end toll = toll facility cost (if any)

Time/Cost Component

Purpose/Income

Home-Work/College Home-Other Other-Other

Time

1. Auto In-Vehicle -0.0280 -0.0160 -0.0220

2. Auto Terminal -0.0560 -0.0320 -0.0440

3. Transit In-Vehicle -0.0280 -0.0160 -0.0220

4. Transit First Wait -0.0420 -0.0256 -0.0352

5. Transit Transfer Wait

-0.0840 -0.0400 -0.0550

6. Transit Auto Access

-0.0560 -0.0320 -0.0440

7 Bicycle -0.0980 -0.0640 -0.0880

Cost Low Mid High Low Mid High Low Mid High

All Modes -0.0083 -0.0031 -0.0013 -0.0094 -0.0035 -0.0015 -0.0131 -0.0049 -0.0020

42

audist = auto distance between zones cpm = cost per mile to operate an automobile ccoef= cost coefficient for purpose and income group from Table 14 nclow = nesting coefficient at lowest level of nest (0.55) ncmid = nesting coefficient at middle level of nest (0.65) nctop = nesting coefficient at top level of nest (0.85)

Although the form of the equation is the same for all auto modes, in-vehicle times, tolls, and distances may differ for each mode depending on the highway network being evaluated. For example in a congested corridor with a freeway HOV lane, the in-vehicle travel time for drive alone non-toll vehicles would be longer than the time for two person shared-ride HOV/non-toll vehicles which are allowed to use the less congested HOV lane.

In-vehicle highway times are calculated using procedures described in the next highway assignment section and include the effects of congestion. Terminal times represent walking time from home to parking location, walking time from parking location to final destination, and other miscellaneous time. Terminal times at the residential end are assumed to be one-quarter minute for single family units and two minutes for multifamily units since higher density developments tend to have less convenient shared parking. The trip generation step calculates average TDZ residential terminal times based on the number of units by each type. A one minute terminal time is assumed at most nonresidential locations. Higher nonresidential terminal times are assumed in Centre City, at universities, military bases, regional shopping centers, and some older outlying business districts. These terminal times vary between three and ten minutes depending on the location.

The model considers three types of auto costs: auto operating costs, parking costs, and tolls. Auto operating costs are based on the SB 375 Regional Targets Advisory Committee (RTAC) method. RTAC uses the 2009 U.S. Department of Energy annual energy outlook and forecasted fuel efficiency from the California Air Resources Board. All costs are in 1999 dollars and are shown in Table 16. Vehicle depreciation costs are not included since these costs are not usually considered when making a mode choice decision. Transit fares for future-year forecasts are assumed to remain constant. It should be noted that cost and fare assumptions can be varied for scenario testing purposes.

43

Table 16: Auto Operating Costs

Year

Fuel Component of Auto Operating Cost (Cost per Mile, $1999) Fuel Efficiency (MPG) Fuel Price ($1999)

2008 $0.138 20 $2.70

2020 $0.159 23.2 $3.68

2025 $0.152 24.8 $3.77

2030 $0.149 26.4 $3.93

2035 $0.153 26.7 $4.07

2040 $0.155 27.2 $4.21

2050 $0.159 28.3 $4.51

Source:

Annual Energy Outlook 2009, Energy Information Administration, U.S. Department of Energy (2020, 2025, 2030)

Extrapolated 2035, 2040, 2050 Prices

Fuel Efficiency from EMFAC w/ Pavley Post Processor (2020, 2030, 2035)

Fuel Efficiency extrapolated for 2040 and 2050

All fuel prices add a 2007$ $0.25 California surcharge

Parking costs shown in Table 17 are applied in Centre City San Diego, universities, Lindberg Field, and business districts in Oceanside, Escondido, La Jolla, and Hillcrest. Although some individual lots charge more, the highest average parking rates of $12.50 per day are found in Centre City San Diego. These rates are made into a per trip basis by dividing by two and are applied by trip purpose. Home to work trips have a longer duration than other trip purposes leading to higher parking costs on average.

Table 17: Parking Costs

Area Type Home-Work Home-Other Non-Home Based

1 $ - $ - $ -

2 $1.00 $0.20 $0.20

3 $3.50 $0.50 $0.50

4 $7.00 $4.50 $4.50

5 $12.50 $8.00 $8.00

The RTP contains two types of toll facilities: (1) managed lanes that charge a fee for use by drive alone vehicles and (2) toll roads where tolls will be charged for all vehicles using the facility.

The managed lane facilities are coded with a per mile toll cost that can vary by time period and location. Data from the existing managed lanes on I-15 were analyzed to determine default managed lane tolls of $0.26 per mile in peak periods and $0.10 per mile in the off-peak. These default toll rates can be adjusted to match model-estimated managed lane demand to available capacity. The operational goal of the managed lanes is maintain a level of service no worse than ‘D’ for any temporal period.

44

One toll facility in the region already exists. On the SR-125 South Bay Expressway, the toll coding replicates the current tolling schedule that exists for autos. A separate tolling schedule is coded for truck traffic as well. The 2050 RTP has 4 additional toll facilities planned. Since no tolling schedule exists, tolls are coded on a per mile basis. These tolls vary from 10 to 27 cents per mile in during off-peak and 15 to 33 cents per mile in the peak.

Transit utility computations are more complicated than auto computations because there are more travel time components associated with transit. These utilities also need to be computed for each transit access and transit ride mode combination, and for each MGRA-MGRA pair, time period, and income group. Walk access computations use the following equation.

nctopncmid

ipccoefrmipfpctatapptaptrmfare

wkfacptcoefetimewk

ptcoefatapptaptrmtrxferptcoeflfw

ptcoefrmsfwptcoefatapptaptrmtrivt

iprm

/

),(),,(),,,(

),3(

),5(),,,(),3(

),4()(),3(),,,(

),,(twu

where: twu(rm,p,i) = transit-walk utility for ride mode “rm”, purpose “p”, and income group “i” trivt = in-vehicle transit time between TAPs for ride mode and time period “t” tcoef = time coefficient for purpose from Table 14 sfw = short first wait time lfw = long first wait time trxfer = transfer wait timewk = walk time etime = station elevation walk time wkfac = walk factor of 2.5 fare = transit cash fare fpct = cash fare discount factor for pass usage As indicated in Table 14, time coefficients on wait times are higher than in-vehicle time coefficients to reflect the fact that surveys show transit riders perceive this out-of-vehicle time component to be more onerous than time spent riding in a vehicle. Transit skimming procedures calculate first-wait times based on one-half the headway of coded transit routes. In the mode choice model these wait times are split into short and long first-wait times. The first 5 minutes of first-wait times are considered short and the remainder is considered long first-wait time. Long first-wait times use a smaller time coefficient to discount their impact as people may time their station arrival better to coordinate with the transit schedule.

The transit skimming process also calculates transfer wait times, where applicable, based on one-half the headway of coded transit routes. Adjustments to these standard transfer time calculations are made for a small number of stations assumed to have bus and rail service that will be sufficiently coordinated to replace the standard transfer times with lower timed-transfer times.

Walk time to transit is another important factor in mode choice. Section 2.5.5 described how walk distances between MGRAs and TAPs are calculated. The following equation is used in the mode choice model to add walk time at the start of a trip, walk time at the end of a trip, and any time spend walking between transit routes when making a transfer.

45

),,,(/),(),()(w atapptaptrmxfertimewkspdamgraatapwkdistptappmgrawkdistrmktime

where: wktime = overall walk time between production and attraction zones wkdist = walk distances between MGRAs and TAPs wkspd = average walking speed of three MPH xfertime = walk time when transferring between routes.

Additional times are added to account for transit station access. Lot time is added to walk from the center of a park and ride lot to the station platform. Drop-off access at park and ride lots is discounted to 25% of the lot time. For transit stations that are above or below grade, additional time is added to account for elevator or stair access.

The final component of the transit walk utility equation is the transit fare. TransCAD transit skimming procedures find the cash fare between each TAP pair based on existing fare schedules. In the mode choice model these cash fares are discounted to account for pass usage. Transit surveys show pass usage varies by a number of different factors. For example, commuter rail users are more likely to use passes than are local bus riders, incidental non-work trips are less likely to use passes than work trips, and pass usage declines with higher incomes. Thus cash fare discount factors are computed and applied by ride mode, trip purpose, and income level.

Transit drive access (park-and-ride) and drop-off access (kiss-and-ride) utilities are computed in a similar manner as the previous transit walk utilities. Times and fares will differ from those for walk access because drive access connections are restricted only to transit stops with formal or informal parking available, and because drive access connections are usually faster than access by feeder bus or long distance walks. Drive access only is allowed at the production end of home-based trips since users generally do not have access to a car at the nonresidential end of a trip.

The transit auto utility equation shown below replaces walk time at the production end with drive time, and adds the cost of driving and a terminal time the end of the auto access leg of the trip.

)/(

),2(),(

),6(),,,,(

),(),,,,(

),,(),,,,(

),3()(

),5(),,,,(

),3(),4(

),3(),,,,(

),,,(tau nctopncmid

ptcoefptapamterm

ptcoefptappztamrmdt

ipccoefcpmptappztamrmdd

rmipfpctatapptaptamrmfare

wkfacptcoefetimewk

ptcoefatapptaptamrmtrxfer

ptcoeflfwptcoefsfw

ptcoefatapptaptamrmtrivt

ipamrm

where: tau(rm,am,p,i) = transit-walk utility for ride mode “rm”, access mode “am” (drive or drop-off), purpose

“p”, and income group “i” dd = drive access distance cpm = drive cost per mile dt = drive access time term = terminal time at the end of the drive access trip

46

Non-motorized utilities for bike and walk modes are computed by applying a time coefficient to non-motorized travel time. These times are derived from non-motorized distance skims described in Section 2.7 by using an average three MPH speed for walk trips and 12 MPH for bike trips. Walk trip times are categorized into short-walk time (less than 5 minutes), mid-walk time (5 to 20 minutes), and long-walk time (greater than 20 minutes). The model does not consider non-motorized operating costs.

nctopptcoefbksnmdistpbku /),10(/)(

nctopptcoefwkfaclwkwkfacmwkwkfacswkpwku /),3()3()2()1()(

where: bku = bicycle utility dist = distance between zones bks = bicycle speed (12 MPH) wku = walk utility

After computing utilities for each mode, the program then cycles through the six trip purposes and three income levels used in the mode choice model. Exponentiated utilities for each of the 26 lower level modes are calculated using the following equation.

),,int,,(4),,2(),,1(),,(uexp),,( memplerdenpdcnstipccnstipccnstipmpimeu

where: eu(m,i,p) = exponentiated utility for mode “m”, income group “i”, and purpose “p” exp = exponential function u(m,p,,i) = utility for mode, purpose, and income group cnst = modal constant for income group, and purpose 4dcnst = 4D density modal constant for purpose, mode, density, intersection density, and employment den = land use density group inter = intersection density group empl = employment density group

Modal constants (Table 18), calculated during the calibration process, are used to bring model-estimated mode shares into agreement with observed mode shares for each market segment. Market segments for most modes are defined using income group and trip purpose. Federal guidelines prevent income stratification for transit ride modes. Table 19 shows how the 13 modal constants are applied to the 25 submodes. 4D constants were derived for transit and non-motorized modes. The constants were developed using the same 4D variable formulations described in the trip distribution steps. The addition of 4D constants allows the model to better reflect the propensity of using non-motorized and transit modes in mixed use and dense land use.

47

Table 18: Mode Choice Constants

Modal Constant Income

Purpose

Home-Work

Home-College

Home-School

Home-Other

Other-Other

Serve Passenger

1. Non-motorized

Low 1.4 -1.46 1.66 -2.51 -6.68 n/a

Middle -1.01 -1.46 0.53 -4.03 -7.95 n/a

High -2.78 -1.46 0.88 -4.76 -6.72 n/a

2. Transit

Low 2.87 -1.43 -1.55 -5.5 -12.51 n/a

Middle -2.76 -7.51 -7.25 -12.59 -18.37 n/a

High -9.15 -12.81 -11.8 -19.08 -21.58 n/a

3. Shared-Ride

Low -5.96 -4.33 0.4 -0.78 -1.05 1.29

Middle -7.49 -2.6 1.06 -0.68 -1.97 2.39

High -8.51 -4.95 2.92 -1.32 -2.26 1.88

4. Bicycle

Low -4.11 -1.43 -4.5 -4.15 -3.31 n/a

Middle -3.36 -1.43 -3.01 -4.93 -4.45 n/a

High -2.54 -1.43 -4.98 -2.78 -4.45 n/a

5. 3+ Shared-ride

Low -1.5 -4.17 1.5 -1.36 -0.91 -0.04

Middle -3.17 -3.38 1.75 -1.56 -0.85 0.07

High -2.8 -2.43 0.74 -0.98 -0.56 -0.15

6. Toll

Low -1.04 0.05 n/a -6.79 -4.94 n/a

Middle -0.25 0.05 n/a -7.71 -7.58 n/a

High 2.27 1.18 n/a -0.72 -1.82 -6.84

7. Shared-ride HOV

Low 0.58 -1.26 -4.4 0.21 -0.58 -1.05

Middle -0.2 -1.43 -4.28 0.04 -0.85 0.7

High 1.8 -0.67 -3.34 0.32 -0.68 0.66

8. Transit Auto

Low -6.85 -6.81 -9.72 -6.16 n/a n/a

Middle -5.1 -5.12 -11.23 -4.33 n/a n/a

High -2.4 -5.49 -7.4 -3.97 n/a n/a

9. Transit Drop-off

Low -1.89 -2.36 1.4 -1.38 n/a n/a

Middle -2.63 -0.22 3.39 -2.77 n/a n/a

High -3.94 -0.1 0.56 -1.28 n/a n/a

10. Transit BRT All 1.12 0.64 0.64 0.64 0.88 n/a

11. Transit Light Rail All 1.96 1.12 1.12 1.12 1.54 n/a

12. Transit Auto BRT All 0 0 0 0 n/a n/a

13. Transit Auto Light Rail All 0.97 0 0 0.96 n/a n/a

14. Transit Commuter Rail All 2.52 1.44 1.44 1.44 1.98 n/a

15. Transit Express All 0.17 -1.79 -1.79 -1.79 -2.46 n/a

Note: n/a indicates mode not available for purpose.

48

Table 19: Application of Modal Constants

Mode

Number of Modal Constant

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 4D

Drive Alone Non-Toll

Drive Alone Toll +

2 Person Shared-ride Non-Toll/ Non-HOV

+

2 Person Shared-ride Non-Toll/ HOV + +

2 Person Shared-ride Toll/ HOV + + +

3+ Person Shared-ride Non-Toll/ Non-HOV

+ +

3+ Person Shared-ride Non-Toll/ HOV + + +

3+ Person Shared-ride Toll/ HOV + + + +

Walk + +

Bicycle + + +

Commuter Rail/Walk Access + + +

Commuter Rail/Drive Access + + + +

Commuter Rail/Drop-Off Access + + + + +

Light Rail/Walk Access + + +

Light Rail/Drive Access + + + + +

Light Rail/Drop-Off Access + + + + + +

BRT/Walk Access + + +

BRT/Drive Access + + + + +

BRT/Drop-Off Access + + + + + +

Express Bus/Walk Access + + +

Express Bus /Drive Access + + + +

Express Bus /Drop-Off Access + + + + +

Local Bus/Walk Access + +

Local Bus /Drive Access + + +

Local Bus /Drop-Off Access + + + +

Note: + indicates constant is applied to exponentiated utility for mode

.

49

Once exponentiated utilities have been calculated for all submodes upper level modes are obtained by combining utilities from lower level modes as follows.

nclowdateudaneudaeu lnexp

nclowthsrnhsrnnsreusr 222lnexp2

nclowthsrnhsrnnsreusr 333lnexp3

ncmiddaeudreu lnexp

ncmideusreusrsreu 32lnexp

nctopsreudreuaueu lnexp

nctopbkeuwkeunmeu lnexp

ncmidlbwkeuebwkeubrtwkeulrwkeucrwkeutwkeu lnexp

ncmid

lbkreuebkreubrtkreulrkreucrkreu

lbpreuebpreubrtpreulrpreucrpreutaueu lnexp

nctoptaueutwkeutreu lnexp

where: ln = natural log function eu = exponentiated utility da = drive alone dan = non-toll dat = toll sr2 = 2 person shared-ride sr2nn = non-toll/non-hov sr2nh = non-toll/ hov sr2th = toll/hov sr3 = 3+ person shared-ride toll/hov dr = drive au = auto sr = shared-ride nm = non-motorized wk = walk bk = bicycle twk = transit-walk cr = commuter rail lr = light rail brt = bus rapid transit eb = express bus lb = local bus tau = transit auto access pr = park-ride access (drive) kr = kiss-ride access (drop-off) tr = transit nclow = nested logit coefficient for low nest ncmid = nested logit coefficient for middle nest nctop = nested logit coefficient for top nest

50

The ratio of exponentiated utilities, shown in the equation below, computes mode shares. Trips by mode then are obtained by applying mode shares to total person trips in each market segment. The resulting trips are accumulated and output in summary reports and as trip tables for the highway and transit assignment process

)()()()( mbeumaeu

maeumams

where: ms(ma) = mode share for mode “a” eu(ma) = utility for mode “a” eu(mb) = utility for mode “b”


A major project to recalibrate the mode choice model was completed in 2006 and is documented in a report Mode Choice Model Improvements, November 2006. Results from the 1995 Travel Behavior Survey, 2000 Market Research Survey, and 2001 Caltrans Statewide Survey, and 2001-2003 Transit Ridership Survey were analyzed to estimate time and cost coefficients and determine a nesting structure. The mode choice model was updated in 2009 for use in the Mid-Coast FTA New Starts project.

After model estimation, observed base year mode shares were tabulated from the Travel Behavior Survey and Transit Ridership Survey which were used to calibrate modal constants so that model-estimated mode shares agree with observed mode shares by market segment.

Transit ridership forecasts from the transit assignment model were extensively evaluated to determine the accuracy of mode choice estimates and adjust model parameters to correct problem areas. Regional level Census 2000 work-trip mode shares also were used to fine-tune mode share estimates.

2.9.4 Mode choice variations

A number of different applications are derived from the mode choice procedures described above including:

Creating input files to the Federal Transit Administration’s (FTA ) Summit program

Computing composite utilities for the feedback loop trip distribution model

Calculating and applying person trip to vehicle trip conversion factors

Summit Input Files. FTA requires a rigorous cost effectiveness analysis when applying for federal funding for major transit investments under the New Starts program. A special version of the mode choice program creates an output file for each time period and purpose, which contains trips and exponentiated utilities for each TAZ pair and income group. These files then are used by the FTA Summit program to calculate and summarize user benefits that are used in a subsequent measure of cost effectiveness.

51

Composite Utilities. The utility calculations described above can be used to provide a trip distribution impedance measure that factors in times and costs by all modes of transportation. A special version of the mode choice model cycles through each time period, zone pair, purpose, and income group and calculates average utilities for three major modes: automobile, transit, and non-motorized. The lowest utility value then is used to compute a daily composite utility for each TDZ pair by three generalized purposes: home-work/college, home-other, and non-home based.

Person Trip to Vehicle Trip Factoring. First stage and feedback loop model applications use a person trip to vehicle trip conversion program to obtain vehicle trip tables for input to the highway assignment process. This conversion process is intended to reflect the effects of smart growth, transit improvements, and other conditions that are forecasted by the mode choice model, while avoiding the time and complexity of transit network modeling and applying the standard mode choice model.

Vehicle factors are obtained from a previous final stage mode choice model which computes vehicle fractions for four highway modes (non-toll/non-HOV, drive alone toll, person shared-ride non-toll/HOV, and shared-ride toll/HOV) by time period (peak period and off-peak period), mode choice purpose (home-work, home-college, home-school, home-other, serve passenger, and non-home based), TDZ pair, and forecast year (for example 2008, 2020, or 2035).

2.10 TRUCK MODEL

In order to address a growing number of freight related policy questions, the San Diego Association of Governments (SANDAG) in association with Parsons Brinkerhoff developed a truck model for the San Diego region. While the main intention of the model is to focus on how transportation network decisions impact regional truck flows, a secondary but critical outcome is the ability to provide more inclusive impacts to regional air quality. The truck model was designed as a first step to address freight issues in San Diego. A model design was conceived to address both the external and internal influences on truck movements in and thru San Diego. Truck trips in San Diego are greatly influenced by shipping through the US-Mexico border and the Port of San Diego. Separating truck travel out from the passenger travel model allows for increased analysis of benefits for freight related projects. Port accessibility or border related infrastructure projects can be assessed for the relative benefit to goods movement in the region and the air quality from heavy duty trucks. More truck model documentation (Development of a truck model for the San Diego Region) can be found on the SANDAG Regional Models website.

2.10.1 Model Design

The SANDAG Truck Model consists of two truck models. A local truck model simulates truck trips within San Diego County (Internal-Internal or II trips), and a regional truck model simulates truck trips that have an origin and/or destination outside San Diego County (Internal-External or IE/EI/EE trips) (see Figure 5). This two-layer Figure 5: Local and regional truck trips

IE

EI

EE

II

I Internal

E External

San Diego County

52

approach allows differentiating the level of detail for simulating trucks. For a truck trip from, for example, Encinitas to Escondido within San Diego County, the detailed location of both the origin and the destination is of interest to assign truck trips to the right highway link. For a truck trips from San Diego to the Bay Area, however, only the detailed location of the origin is of interest. Whether this long-distance truck trip has San José, Oakland or Berkeley as its final destination is irrelevant to assign the trip to the right network link within San Diego County. Thus, a two-layer approach allows simulating truck trips at the necessary level of detail without carrying a large overhead of unnecessary detail.

2.10.2 Local Truck Model

The model simulating local truck trips within San Diego County is based on the SCAG Truck Model developed by Cambridge Systematics (2008) for the Southern California Association of Governments (SCAG). This trip-based model generates truck trips based on employment and household trip rates. Given the close proximity to the SCAG region and San Diego County, the truck generation factors are expected to be transferable to the SANDAG region. In addition to SCAG factors, trucks generated by special generators are added explicitly. Special generators are facilities that generate a significant number of trucks that cannot be explained by the employment at that facility. A cruise ship terminal, for instance, attracts a large number of trucks even though it has a comparatively small number of employees. In addition, military sites and express mail sent through the airport are treated as special generators. Three truck types are distinguished: Light-Heavy Duty Trucks (8,500 to 14,000 lbs), Medium-Heavy Duty Trucks (14,000 to 33,000 lbs) and Heavy-Heavy Duty Trucks (more than 33,000 lbs). After generating truck trip productions and attractions, a gravity model is used to distribute truck trips. Off-peak travel times are used as impedance in the trip distribution function, since a larger number of truck trips happen in the off-peak hours.

2.10.3 Regional Truck Model

Truck trips having their origin and/or destination outside of San Diego County are simulated by the regional truck model. These truck trips are generated based on goods flows reported by the Freight Analysis Framework 2 (FAF2), which is published by the Federal Highway Administration (2002). Freight flows are reported by STCC commodity between 130 domestic FAF regions as well as international flows from and to the U.S. To increase spatial resolution, flows between 130 FAF zones are disaggregated to flows between 3,241 counties. Employment is used as a weight for this disaggregation: counties with more employment are assumed to produce and attract more trucks than counties with less employment. Using commodity-specific Payload Factors (Battelle 2002: 29) goods flows in tons are converted into number of truck trips. An average vacancy rate of 19.4 percent (U.S. Census Bureau 2008) was added to all flows. FAF2 covers the base year 2002 and the future years 2010 to 2035 in five-year increments. This forecast as published in 2002 is quite optimistic in terms of growth of goods flows. Flows into and out of San Diego County are forecast to grow exponentially from 4.9 Mio. trucks per year in 2002 to 15.7 Mio. trucks in 2035. The current economic downturn as well as the future gas price development may result in a smaller growth. For the San Diego Truck Model, the FAF2 flows were scaled down to grow only linearly from 2020. A further adjustment was made to flows crossing the border with Mexico. Many truck trips across the border serve Maquiladoras in Mexico, which

53

produce goods mostly for the U.S. markets at lower labor rates. Truck trips serving the Maquiladoras tend to be short-distance trips, for which on-time delivery is more important than low vacancy rates. Accordingly, the number of truck trips across the Mexican border was inflated to match truck counts in the base year.

2.10.4 Truck Trip Reconciliation

The regional truck trips are temporarily assigned to a U.S. highway network to generate truck trip production and attraction at the external stations of San Diego County. The local truck model generates the II truck trips, while the regional truck model provides EI, IE and EE trips. A temporal allocation splits trips into three time-of-day periods, namely AM Peak, PM Peak and Off-Peak. Hourly counts at Weigh-In-Motion (WIM) stations provide factors to split truck flows into the three time-periods. Since border crossings with Mexico are closed for trucks from 8 pm to 5 am, trucks trips to and from the Mexican border crossings are reduced in the Off-Peak period. The truck model is validated against truck counts at WIM stations. The challenge of validating the model is that trucks are defined differently at the WIM stations than in the truck model. While the WIM stations count trucks by FHWA truck classification based on axles, the model simulates truck by weight class. The WIM counts include FHWA classes 4 through 14 and exclude the large number of pick-up trucks/vans (FHWA class 3). The simulated volumes, however, include trucks and those pick-up trucks with a weight of more than 8,500 lbs. Consequently, it is expected that simulated volumes are slightly higher than observed truck counts.

2.10.5 Air Quality

SANDAG uses the California Air Resources Board EMFAC2007 for regional air quality analysis. EMFAC2007 (California Air Resources Board 2006) has 4 vehicle classes for heavy duty trucks: Light-Heavy Duty 1 (8501-10000), Light-Heavy Duty 2 (10001-14000), Medium-Heavy Duty (14001-33000), and Heavy-Heavy Duty (33001-60000). These four vehicle classes correspond to the vehicle classes from the heavy duty truck model except for light-heavy duty which is combined into one category. The addition of the truck model enables direct modifications of VMT, vehicle trips, and speed breakdowns to the four EMFAC heavy-duty truck vehicle classes. Proposed infrastructure improvements or policies that improve mobility for heavy-duty trucks can then be assessed for air quality benefits.

2.11 HIGHWAY ASSIGNMENT

Highway assignment is the process of loading vehicle trips between zones onto specific segments of highway. Trips are apportioned to links based on the generalized cost and capacity associated with each link from the highway network coding process described earlier. Major inputs and outputs are listed below in Table 20 and Table 21.

As congestion builds over time, the highway assignment model shifts traffic to adjacent facilities having excess capacity. Similarly, corridors where new roadways or roadway improvements are planned will see traffic diversions to the new facilities from parallel facilities having slower speeds or higher congestion. These shifts in traffic between facilities are a major component of what can be characterized as induced demand.

54

Table 20: Highway Assignment Model Inputs

Input Source

Vehicle trips between zones Mode choice model

Truck trips between zones Truck Model

Highway network Network coding

Volume-delay functions Highway Capacity Manual

Traffic counts Caltrans and local jurisdictions

Table 21: Highway Assignment Model Outputs

Output

Daily and peak hour traffic volumes by highway segment

Congested highway times and speeds by segment and time period

Level-of-service by segment and time period

Air pollution emission model inputs

Summary reports and plots


SANDAG loads traffic using the TransCAD “Multi-Modal Multi-Class Assignment” function with the bi-conjugate load method. This is an iterative technique for balancing estimated link volumes with available capacity by minimizing an overall congestion index. The model first finds minimum paths between zones based on input generalized cost assumptions, including tolls, for each roadway segment and the associated value of time for that mode (Auto and Light Heavy-Duty Trucks 50 cents/min, Medium Heavy-Duty Trucks 51 cents/min, and Heavy Heavy-Duty Trucks 72 cents/min). Trips between zones are accumulated on links making up the minimum time path between each zone pair. Once all trips have been assigned, congestion levels are determined by computing mid-link and intersection volume-to-capacity (V/C) ratios. Input link times are revised using a logit-based volume delay function (VDF) that adjusts speeds based on the mid-link V/C ratio and an intersection delay function that adjusts intersection delays based on the intersection V/C. Additional assignment iterations are performed making use of revised link speeds from the previous iteration until equilibrium is reached. Equilibrium is considered reached when the convergence criterion reaches 0.001.

The logit-based VDF is available in TransCAD and was developed and calibrated by the Israel Institute of Transportation Planning and Research. Since intersection impacts are implicit in the VDF function, the function is used for both highway and arterial facility types.

55

Table 22: Volume Delay Function Parameters

C1 C2 C3 C4 P1 P2 P3 P4

Parameter 0.9526 1 3 3 0.09 350 3.5 2.3

A.M. peak period, P.M. peak period, and off-peak trip tables are loaded separately since motorists may choose different travel paths during different time periods. For example, heavily-congested facilities and metered ramps would tend to be avoided during peak periods, but not during off-peak hours. Daily volumes are obtained by adding together assignments for the three time periods.

TransCAD’s ability to perform simultaneous assignments for multiple assignment modes also is used. As indicated above, the mode choice model outputs up to eight separate vehicle trip tables plus an additional six truck trip tables. Highway networks include existing and proposed HOV lanes as separate facilities. During the assignment process trips by the fourteen modes are matched with facilities that are available for use by each mode. Truck modes are loaded using passenger car equivalents to replicate the increased impact they have due to size, acceleration, and deceleration.

Table 23: Truck Passenger Car Equivalents

Truck Mode PCE

Light Heavy-Duty 1.3

Medium Heavy-Duty 1.5

Heavy Heavy-Duty 2.5

2.11.2 Post-assignment Processing

After completing the highway assignments, additional processing is needed to produce reports, data files, and plots tailored to SANDAG needs. Some of the major functions of this post-assignment processing are described below.

Calibration Error Volume Adjustment. Base year model-estimated link volumes for individual links differ from ground counts, even after model calibration is complete. An automated adjustment procedure has been developed to adjust future year traffic volumes to compensate for calibration errors.

Managed Lane Volume Adjustments. The 2050 RTP proposes managed lanes on many freeways. Managed lanes allow solo drivers who pay a toll to make use of HOV lanes. However, traffic on managed lanes would be controlled so that level-of-service “D” is maintained. Post-assignment procedures simulate optimal managed lane operation by: (1) shifting traffic from over-capacity main freeway lanes to adjacent managed lanes that have excess capacity; and (2) shifting traffic from over-capacity managed lanes to adjacent freeway main lanes.

56

Bus Volume Adjustment. Public transit bus volumes are determined in the transit network coding process. These volumes are added to private vehicle traffic volumes from the highway assignment model.

Hourly Distribution Factors. The TransCAD assignment process outputs A.M. peak period, P.M. peak period, and off-peak period traffic volumes, while many applications need hourly traffic volumes. Caltrans collects hourly traffic counts at its permanent count stations. These counts are used to compute the fraction of traffic in each hour of a time period for each direction and count station. Resulting hourly distribution factors are applied to freeway segments after assigning freeway segments to the closest count station.

Level-of–Service Computations. Highway Capacity Manual procedures are used to compute the level-of-service (LOS) for each highway segment based on a V/C ratio using coded capacities and volumes from the adjustment process described above. LOS computations differ by type of facility. On freeways, an LOS rating “A” through “F” is assigned based on the hour with the highest V/C ratio within each time period. When a V/C ratio exceeds 1.0, the hours at LOS “F” are accumulated. While sophisticated traffic queuing procedures do not now exist as a part of long range transportation models, an attempt is made to account for some queuing impacts in the post-assignment process by assigning freeway links up to one mile in back of chokepoints the chokepoint LOS. This results in somewhat higher levels of freeway congestion and lower freeway speeds which should better reflect reality.

LOS computations on surface streets also are based on volume to capacity ratios. Street segments between major intersections are grouped into sections. The highest V/C on any segment within a section determines LOS for the entire section.

Travel Time Computations. Link travel times are re-computed based on adjusted link volumes using Highway Capacity Manual procedures. The same V/C calculations used to assign LOS also are used to look up speeds and intersection delay times associated with VC ratios.

Freeway speeds vary between three MPH and 75 MPH depending on the V/C ratio. Link travel times are recomputed for each hour and an average travel time for each time period is calculated by weighting the hourly travel time by the hourly volume. Again, freeway HOV lane LOS is assumed to be no worse than LOS “D,” so HOV speeds vary between 62 MPH and 75 MPH.

Travel times on surface streets consider both mid-block congestion and delays at signals and stop signs. Mid-block speeds vary between three MPH and the posted speed, depending on the V/C ratio. Intersection delays vary between ten seconds and two minutes and are added to the link travel time computed from mid-block speed. Delays at ramp meters ramp meters range between one minute and 15 minutes. HOV times on ramp meters with HOV bypass ramps are assumed to be one-third of SOV delays.

Emission Model Inputs. The California Air Resources Board’s EMFAC2007 model is used to estimate on-road motor vehicle emissions and fuel consumption. This model requires as input the vehicle miles of travel in each hour by 13 vehicle types (light-duty automobiles, light-duty trucks, heavy-duty trucks, etc.) by 18 speed ranges. Link level traffic volumes are split into vehicle types by applying vehicle type percentages that vary by facility type and hour of the day. These percentages were obtained from the 2000 Vehicle Classification and Occupancy Study. Resulting vehicle type volumes are multiplied by link distances and accumulated by speed range to produce Burden inputs.

57

Reporting and Plotting. Summary statistics such as vehicle miles of travel (VMT), vehicle hours of travel (VHT), and average speeds are computed which are key performance indicators for transportation and land use alternatives. Model outputs also are linked with GIS coverages to produced computer plots and data sets used to display traffic volumes on the Internet.


Comparisons of base year traffic volumes from the highway assignment model with observed traffic counts provide key measures of the model accuracy. Differences between model-estimated and observed VMT at the regional level indicate whether the overall amount of travel in the region is correct. VMT comparisons by facility type may indicate problems with speed assumptions between facility types. Traffic estimates on individual links also are compared with traffic counts. Large discrepancies may indicate network coding errors, miscoding of access opportunities, land use coding errors, or inappropriate trip generation rates.

2.12 TRANSIT ASSIGNMENT

The transit assignment step determines route, link, and stop level ridership using inputs listed in Table 24. These transit assignment results, as shown in Table 25, are important when evaluating model accuracy and the effectiveness of proposed transit improvements.

Table 24: Transit Assignment Model Inputs

Input Source

Transit trips between TAPs by time period Mode choice model

External transit trips Transit Ridership Surveys

Transit network Network coding

Table 25: Transit Assignment Model Output

Output

Daily and peak period boardings by route, link, and stop

Summary reports and plots


TransCAD software assigns TAP-to-TAP transit trips to the network. Eight separate transit assignments are produced for peak and off-peak periods; walk and auto access; and local bus and premium service. These individual assignments are summed to obtain total transit ridership forecasts.

Before assigning transit trips, external transit trips coming into San Diego from outside the region need to be added to the internal transit trips estimated by the mode choice model. Currently few transit trips enter from the north or east, however, over 20,000 transit trips cross the Mexican border each day. An external transit trip table for the base year is developed from on-board transit ridership surveys and factored to future years based on border crossing trends to account for these trips.

58


Transit ridership forecasts from the transit assignment model are compared with transit counts from the SANDAG transit passenger counting program to determine whether transit modeling parameters need to be adjusted. The transit assignment model itself has few parameters so the calibration process is primarily a check on transit network coding and mode choice procedures.

Some of these comparisons of model-estimated boardings with actual boardings include

system level boardings, which may reveal transfer rate problems and lead to changes to the transfer wait time factor in the mode choice model

boardings by mode, which may reveal modal biases and lead to changes in mode choice modal constants

boardings by frequency of service, which may show biases that lead to changes in the first wait factor in the mode choice model

Centre City screenline crossings, which may lead to changes in parking costs

boardings by stop location, which may indicate problems which specific generators such as a university

2.13 INDUCED DEMAND

The SANDAG regional travel demand model attempts to simplify the chaotic travel decision making process people go through in their daily lives. The regional model employs a number of behavior-based equations to replicate this process. New infrastructure investment can influence travel decision making by a multitude of factors. Induced demand (also known as induced travel) occurs when changes in travel demand are a direct or indirect result of the new infrastructure investment. The term is commonly used to describe a situation where highway expansion results in additional vehicle trips and VMT. The SANDAG regional travel demand model accounts for many of the potential ways travel could be induced. Categories of induced demand are listed below.

2.13.1 Land Use

New infrastructure development can change the near term and future distribution of housing and employment by increasing the accessibility to undeveloped or redevelopable land. This could be through new freeway access to greenfield development or transit oriented redevelopment (TOD) around a new trolley stop. The extent of induced demand depends on the geographic reference area. A change in development from one area of a city to another would cause local travel impacts, but would cause no impact at the regional level since the development does not occur in another area. Transit infrastructure may shift development from greenfields to TODs, or vice-versa for new freeway development. Any infrastructure project that shifts development from outside the county or state to internal would be regional induced demand. SANDAG uses an iterative approach to account for this impact. The transportation model is run with the base year land use, and accessibilities are fed back to the land use model for the first forecast increment. The new travel time accessibilities influence the land use forecast increment, which is

59

then fed back to the transportation model and run against the draft transportation plans. The feedback process is repeated for each increment until the end of the forecast cycle.

2.13.2 Trip Generation/Activities

People engage in activities throughout the day and use the transport system to move from activity to activity. There are several activity pattern impacts that could result from increased accessibility. First, if trip travel times reduce it could create additional time windows for new activities to occur. For example, if the travel time from home to work is reduced by 15 minutes each way, a person may have more time for a new activity such as going to the gym. Second, additional activity time may shift activities made on other days of the week, such as moving a grocery trip from the weekend to a weekday. Third, increased accessibility may shift an in-home activity such as eating, movie watching, exercise, etc. to an out of home activity resulting in a new trip on the transport system. Finally, additional accessibility may reduce the need for trips to be chained together leading to additional trips and person miles traveled. The SANDAG model does not currently reflect these travel inducements however they are being considered during the development of SANDAG’s Activity Based Model (ABM).

2.13.3 Trip Distribution

Changes in accessibility will affect the locations that people choose to travel to. Congested corridors may lead to people choosing locations close to home. If new infrastructure is built people may choose a more desirable but farther destination that is now within the same travel time window. Destination changes could result in changes to person miles traveled. This impact if reflected in the trip distribution model. Modal accessibility that includes roadway congestion and transit availability is used during feedback loops to re-determine choice in trip destinations.

2.13.4 Trip Mode Choice

Infrastructure development will change modal choices available for each trip. A new trolley line will give more travel options to those nearby a station and would influence their choice of what mode to take for a given trip. Likewise, an expanded roadway may give reduced travel time and shift travelers from transit or non-motorized modes to an auto mode. The SANDAG mode choice model accounts for the available travel choices for each trip and resultant shifts in mode.

2.13.5 Trip Time of Day

Increases or decreases in congestion could change the time of departure for a trip. Additional congestion will lead a condition termed as “peak spreading.” People alter their departure time to attempt to travel outside peak congested conditions and thus spread the travel impacts out into the shoulder of the peak period. Conversely, expanding capacity on a roadway could lead to peak sharpening or a remerging of the travel on the shoulders of the peak to the personally desired travel departure time. This second condition is often considered an induced travel impact of new infrastructure.

60

The SANDAG model does not currently account for shifts in trip departure time of day however it is being considered during the development of SANDAG’s Activity Based Model (ABM).

2.13.6 Trip Assignment

Congestion can cause vehicle trips to take less desirable paths to a destination than would occur during uncongested conditions. An expanded facility could cause travelers to change their path to the new facility to save travel time. Diverted trips from one facility to another are often considered induced demand. The changed path, while saving overall travel cost (combined time, distance, and toll costs), does not guarantee the reduction of the length of the trip leading to the potential of additional vehicle miles traveled. The SANDAG highway assignment model accounts for diverted trips due to changes in congestion and infrastructure.

61

CHAPTER 3: MODEL CHANGES FOR THE 2050 RTP

This section contrasts the travel demand model used for the 2050 RTP, using the 2050 Regional Growth Forecast, and the 2030 RTP, using the 2030 Regional Growth Forecast Update. The time period between RTPs is used to update the travel demand model to help answer emerging policy questions in the next RTP. The amount of effort to add new components or recalibrate existing models forces development work into a short time window before the next RTP cycle gets kicked off. Some of the major changes in the travel model used in the 2050 RTP are detailed below. These include:

updating the model base year from 2004 to 2008

updating the underlying zone system for MGRAs, TAZs, and TDZs

thoroughly reviewing the roadway network with member jurisdictions

adding 4D components to the trip distribution and mode choice models

recalibrating the mode choice model to meet FTA guidelines

adding a heavy-duty truck trip model and truck toll diversion model

3.1 SOFTWARE

TransCAD, created by Caliper Corporation, is a transportation planning computer package used by SANDAG to provide a framework for performing much of the computer processing involved with modeling. TransCAD routinely releases software updates. The SANDAG travel demand model software was updated to use and be compatible with TransCAD 5.0 build 1985 for use in the 2050 RTP.

3.2 ZONE SYSTEM

Section 2.3, Growth Forecast Inputs, describes the geographies used throughout the model system. As part of the 2050 growth forecast geographic zone systems were updated. The updated zones followed the same convention of smaller geographies nesting within the larger zone systems. MGRAs nest within TAZs, and likewise TAZs nest within TDZs. MGRAs were realigned to follow parcel boundaries and those with little or no activity were merged with adjacent zones. MGRAs were reduced from 33,353 zones to 21,633. TAZs were realigned to follow the new MGRA boundaries. Additionally, TAZs with further disaggregated from 4605 to 4682 zones. Most new zones were in areas with new land use growth or areas identified for zone splits during community plan updates and other studies. Finally, TDZs were also realigned to follow the new TAZ boundaries. TDZs remained constant in number at 2000 zones.

62

3.3 GROWTH FORECAST

As part of integrating the 2050 Regional Growth Forecast, the regional transportation model was updated to a base year of 2008. New traffic and crossing count information was integrated to update cordon and port of entry forecasts. Airport demand was also revised to match the latest aviation demand from the San Diego International Airport.

3.4 HIGHWAY NETWORK INPUTS

The Series 12 baseline highway network is maintained by SANDAG for use in the transportation model. SANDAG solicited the assistance of local jurisdictions to review the 2008 transportation network data and provide feedback on its ground truth accuracy. In the past, a series of hardcopy maps would be generated and sent out to the local jurisdictions. Typically, only a handful of agencies had the resources to review and markup the hardcopy maps and submit them back to SANDAG within an adequate timeframe. For series 12, staff implemented a Web-based GIS application that allowed local transportation engineers to review and comment on the accuracy of the 2008 baseline network. As a result, the review process promoted greater participation from local jurisdictions and led to improved accuracy of the baseline network.

3.5 TRANSIT NETWORK INPUTS

The 2050 RTP introduces the streetcar as a new transit option for the region. The operating characteristics of streetcar would be similar to local buses in regards to stop spacing but would have slightly better travel times due to priority transit treatments like traffic signal prioritization. Streetcars have been given the same mode as LRT but operate an assumed operating speed of 12 MPH. In addition, emphasis has been placed on precise network coding to capture each aspect of person travel time through the network. Such aspects as the time to walk from one’s car to the trolley platform at a park and ride lot, the added effort of climbing stairs to an elevated station, and the use of aerial photography and GPS data to accurately locate transit stops has been included. Finally, the fares for each transit route were updated, including the recent conversion of the Trolley to a flat fare structure and the reduction of Coaster fare zones from 4 to 3.

3.6 TRIP GENERATION

New trip rates were added for public storage and service stations using a combination of ITE Trip Generation, 8th Edition, trip generation rates and SANDAG trip generation production and attraction trip purpose percentages. Trip generation rates were modified for mixed use land use. Often general plans identify large amounts of acreage for future mixed use development. Without detailed planning of the actual mix and intensity of land uses too many person trips get generated from the sites. Three levels of mixed use trip generation rates were created to accommodate the future plans.

63

3.7 TRIP DISTRIBUTION

In an effort to make the transportation model more sensitive to land use diversity, density, and urban form, 4D segmentation was added to the trip distribution model. For home based trip purposes, the model uses a mixed index to alter the friction factors. For non-home based trips, the model uses an employment density index to alter the friction factors. The resultant model shows average trip length declines as the mixed index or employment density increases (higher values of mixed index are more diverse and dense). While adding the 4D segmentation the trip distribution model was also re-estimated with the 2006 travel behavior survey.

3.8 MODE CHOICE MODEL

The mode choice model was recalibrated as part of the Mid-Coast FTA New Starts project. This included a thorough review of the mode choice code and update of code and parameters to meet FTA guidelines. During the recalibration, 4D related constants were added to transit and non-motorized utility equations to enhance their sensitivity to land use changes. Typically SANDAG takes a conservative approach and assumes no increase in auto operating costs (AOC) throughout the forecast horizon. For the 2050 RTP the RTAC approved process for determining AOC was adopted. The RTAC process results in gas prices that outpace fuel efficiency leading to an increase of $0.02 per mile between 2008 and 2050. This will lead the model to consider non-auto modes as more attractive in the later years of the forecast. Parking costs were also updated by area type based on a parking inventory survey completed in 2010. Parking was not assumed to expand beyond the current pay locations and was conservatively assumed to remain constant over time.

3.9 HIGHWAY ASSIGNMENT

TransCAD highway assignment procedures were updated to use a newer load method that helps the highway assignment converge faster. The user equilibrium Frank-Wolfe algorithm was changed to the bi-conjugate assignment method using the n-conjugate algorithm. This has little impact on the result of the highway assignment other than allowing the model to go through more iterations and reach a tougher convergence criterion in the same amount of time. The convergence criterion was reduced to 0.001. Additionally the script was updated to integrate assignment of 6 additional mode classes for light-heavy duty non-toll/toll, medium-heavy duty non-toll/toll, and heavy-heavy duty non-toll/toll. The assignment value of times were updated to include trucks and to calibrate to toll road count information.

SANDAG Off-Model Greenhouse Gas Reduction Methodology for the 2050 Regional Transportation Plan

Introduction This paper documents the SANDAG methodology for estimating the off-model greenhouse gas (GHG) reductions for several strategies included in the 2050 Regional Transportation Plan (RTP). These strategies include aspects of the bicycle, pedestrian, safe routes to schools (SR2S), vanpool, carpool, and buspool programs.

Bicycle and pedestrian trips, or nonmotorized trips, are accounted for in the traditional travel demand model, but the traditional model is insensitive to investments in additional nonmotorized transportation infrastructure and programs. Through the recent inclusion of the 4-D model components, the traditional model is sensitive to some built environment aspects, such as land use diversity and density.

The traditional transportation model does consider carpooling during mode choice and network assignment. The traditional model, however, is not sensitive to marketing programs related to carpooling and vanpooling considered in the off-model strategies. The traditional model is sensitive to inputs like available carpool facilities, auto operating costs, tolls, and parking pricing.

Background Emissions Factor An off-model emissions factor was developed for 2020, 2035, and 2050 based on a review of the different vehicle types in the EMissions FACtors 2007 model (EMFAC model), which includes the Light-Duty Auto, Light-Duty Truck 1, Light-Duty Truck 2, and Medium-Duty Vehicle classes1. The resulting factors used in the off-model analysis were 1.01 lbs./mile in 2020, 1.00 lbs./mile in 2035, and 1.00 lbs./mile in 2050. Table 1 shows the components used to develop the emissions rates.

Table 1 – Emissions Factor and Supporting Variables

CO2

Emissions Factor (Pounds per Mile)

Supporting Variables

Running Exhaust CO2

(Pounds per Trip)

“Key-On” CO2 (Pounds per Vehicle

Start)

Total CO2 (Pounds per Trip +

Vehicle Start)

Estimated Trip Length

(Miles)

2020 1.01 6.259 0.182 6.441 6.39

2035 1.00 6.505 0.180 6.685 6.71

2050 1.00 6.691 0.180 6.871 6.88

FORMULA: CO2 EMissions FACtors = (Total Regionwide CO2 / Average Estimated Trip Length)

Detailed Strategies The following information is included for each of the strategies to define the off-model methodology: (1) strategy definition and intended effect; (2) assumed level of deployment and GHG reduction calculations; and (3) corresponding GHG emissions reduction estimates as included in the Draft 2050 RTP. Additionally, the documentation includes information from Cambridge Systematics Inc., which provided an independent review and validation of the initial GHG reduction measures2.

1 Pavley/ LCFS factors were developed also by SANDAG modeling staff separate from the off-model analysis. 2 Cambridge Systematics (2010) Senate Bill 375 (Steinberg 2008) Target Setting Methodology, Review and Recommendations: Final Report.

2

Bicycle Network Facilities The bicycle network GHG reduction methodology uses a correlation between regional and local bicycle network investment and anticipated increase in bicycle trips. Using an equation developed by Dill and Carr3, each additional mile of bike facilities (expanded from Class II to include Class I, Class III, Bicycle Boulevards, and Cycle Tracks) per square mile is associated with a roughly one percent increase in the share of all bicycle trips. SANDAG utilized this methodology to quantify the increased bicycle trips from enhanced investments in the regional and local bicycle network based on those improvements included in the approved regional bike plan (“Riding to 2050: The San Diego Regional Bike Plan”). These enhancements also were added to the model data to arrive at an updated bicycle mode share.

To apply the Dill and Carr methodology, four key components were required: (1) Future bike lane miles (regional and local); (2) Study area square miles; (3) Levels of investment for each of the horizon years (2020, 2035, and 2050); and (4) Average bicycle trip length to convert trips into vehicle miles traveled reductions. The study boundaries used were consistent with those developed under the RTP transit strategy (Urban Area Transit Strategy study area). This focused geography was used since the bicycle investments are proposed to be targeted in the urbanized areas of San Diego County. This information is shown in Table 2 below.

Table 2 – Bicycle Methodology Calculation Variables

Variables 2020 2035 2050

Study Area 481 Sq. Miles 481 Sq. Miles 481 Sq. Miles

Level of Investment 22% of 2050 Network 57% of 2050 Network 100% of 2050 Network

Future Bike Lane Miles 311 Miles 807 Miles 1,416 Miles

Average Bicycle Trip Length4 2.13 Miles 2.13 Miles 2.13 Miles

The one percent calculation was performed as follows:

Year 2020: (311 Bike Miles / 481 square miles) * 100 to convert to Dill and Carr % = 0.65% Year 2035: (807 Bike Miles / 481 square miles) * 100 to convert to Dill and Carr % = 1.68% Year 2050: (1,416 Bike Miles / 481 square miles) * 100 to convert to Dill and Carr % = 2.94% Table 3 shows the application of the percentage rates to the overall trips in the urban area for the Draft 2050 RTP.

Table 3 – Draft 2050 RTP – Projected Bicycle Trip Vehicle Miles Traveled Reduction

Scenarios Total Urban Trips*

Dill and Carr Formula (1% increase for every 1 bicycle lane mi. per sq. mi.)

Resulting New Bike Trips

2020 15,318,644 0.65% 99,571

2035 17,276,073 1.68% 290,238

2050 19,220,590 2.94% 565,085

Source: SANDAG Regional Transportation Demand Model, Draft 2050 RTP

3 Bicycle Commuting and Facilities in Major U.S. Cities: If You Build Them, Commuters Will Use Them – Another Look. Dill, J., and T. Carr. 2003.

Transportation Research Board 1828, National Academy of Sciences, Washington, D.C. 4 Based on the NHTS 2009 average bicycle trip length.

3

Effect on GHG emissions: The “New Bicycle Trips” shown in Table 4 were translated to vehicle miles traveled (VMT) reductions through the multiplication of average bicycle trip length (2.13 miles) found in the 2009 National Household Travel Survey (NHTS) report. Table 4 includes the results of this strategy.

Table 4 – Draft 2050 RTP – Bicycle Network VMT and CO2 Results

Scenarios Daily VMT Reduction Lbs. CO2 Reduction Lbs. Per Capita CO2 Reduction

2020 212,087 213,974 0.061

2035 618,207 615,796 0.153

2050 1,203,632 1,203,150 0.274

Pedestrian Network Improvements The pedestrian strategy focuses on shifting some vehicle trips to the pedestrian mode through investments in the pedestrian network beyond what is already included in the regional travel demand model. Enhancements to the pedestrian network were evaluated since SANDAG administers the Smart Growth Incentive Program (SGIP) that awards funding for capital enhancements to the pedestrian realm on a competitive basis. The SGIP funds ($499 million through 2050 in year of expenditure dollars) also are included in the Active Transportation funding (approximately $3.9 billion through 2050 in year of expenditure dollars) proposed for the Draft 2050 RTP.

For the pedestrian network improvement strategy, SANDAG assumed a 10 percent increase in walking trips by 2020, a 20 percent increase by 2035, and a 30 percent increase by 2050. Additionally, it was assumed that every new walk trip would reduce vehicle miles traveled by one mile. These enhancements were added to the model data to arrive at an updated pedestrian trip share. Countywide VMT are projected to decrease by about 160,000 daily miles, as shown in Table 5. This is within the range of VMT reduction that researchers believe can occur in response to pedestrian network investments.5

Table 5 – Draft 2050 RTP – Projected Pedestrian Network VMT Reduction – Draft 2050 RTP

Total Walk Trips 2020 2035 2050

Projected Person Trips (Walk) 485,705 498,283 531,616

10% Increase 48,571 - -

20% Increase - 99,657 -

30% Increase - - 159,485

2020/2035 Walk Trip VMT Reduction 48,571 99,657 159,485

Source: SANDAG Regional Transportation Demand Model, Draft 2050 RTP

This measure is an off-model strategy amenable to the CO2 emissions factors found in Table 1. Table 6 includes the CO2 emissions results of this strategy utilizing the 10/20/30 percent pedestrian increase shown above.

5 Ewing and Cervero report that the elasticity of VMT with respect to changes in a measure of pedestrian network quality “Pedestrian Environment

Factor” (PEF) is -0.03. This indicates, for example, that a 50 percent increase in the PEF (such as systematic investments in sidewalks and pedestrian improvements) would yield a 1.5 percent reduction in VMT. Source: Ewing, R., and R. Cervero (2001) Travel and the Built Environment. Transportation Research Record 1780, 87-114. The Ewing and Cervero report also references the 1000 Friends of Oregon study, “Making the Land Use Transportation Air Quality Connection: The Pedestrian Environment Volume 4A.” Available at:

http://www.friends.org/sites/friends.org/files/reports/LUTRAQ%20Volume%204A%20The%20 Pedestrian%20Enviroment%2C%201993.pdf

4

Table 6 – Draft 2050 RTP – Pedestrian Network VMT and CO2 Results


2020 48,571 49,003 0.014

2035 99,657 99,268 0.025

2050 159,485 159,421 0.036

Safe Routes to Schools Strategies (SR2S) Similar to the pedestrian network, the SR2S strategies are designed to shift some school trips to the pedestrian (and bicycle) modes. SANDAG is proposing to implement SR2S grants programs that encompass at least four of the program’s five Es: Education, Enforcement, Evaluation, Engineering, and Encouragement. This will involve comprehensive planning programs to evaluate need, and determine appropriate improvements and associated programs, pedestrian and bike access capital improvements near schools, and implementation of programs such as walking school buses and educational workshops. This program is intended to reduce the number of vehicle trips associated with parents driving to drop-off and pick-up children at school.

For the SR2S program strategy, SANDAG assumed a 10 percent increase in walk/bike trips by 2020, a 20 percent increase by 2035, and a 30 percent increase by 2050, similar to the pedestrian network improvement strategy. This was based on research conducted by SANDAG and the Metropolitan Transportation Commission (MTC) based on all public school-age trips (K-8). In its most recent Regional Transportation Plan, MTC found6 a 10 percent increase in walk/bike school trips by 2020 and a 20 percent increase by 2035; SANDAG extrapolated these increases to reach 30 percent by 2050. Additionally, it was assumed that every new school walk trip would reduce vehicle miles traveled by one mile and every new school bike trip would reduce vehicle miles traveled by two miles. Regionwide, vehicle miles traveled would decrease by an estimated 29,610 daily miles by 2050, as shown in Table 7 for the Draft 2050 RTP. The following travel demand model output was used to apply the 10 percent, 20 percent, and 30 percent trip adjustments and Census data was used to derive K-8 public school trips from all school trips (public and private K-12) included in the model.

6 Source: Metropolitan Transportation Commission (2008) Transportation 2035 Plan for the San Francisco Bay Area: Performance Assessment Report December 2008. Available at http://www.mtc.ca.gov/planning/2035_plan/Supplementary/T2035Plan-Perf_AssessmentReport.pdf

5

Table 7 – Draft 2050 RTP – Projected School Trips and Estimated SR2S Program Participation

Home-School Walk Trips 2020 2035 2050 Source

Projected K-12 Person Trips (Walk) 196,211 202,301 215,389 1

K-8 Age Public-Only Walk Trips (41% of K-12 Total Walk Trips)

80,447 82,943 88,309 2

10% Increase from SR2S 8,045 MTC research

20% Increase from SR2S 16,589 MTC research

30% Increase from SR2S 26,493

School Walk Trip VMT Reduction

(1 walk trip equals 1 mile)

8,045 16,589 26,493 -

Home-School Bike Trips 2020 2035 2050 Source

Projected K-12 Person Trips (Bike) 11,786 12,114 12,671 1

K-8 Age Public-Only Bike Trips (41% of K-12 Total Walk Trips)

4,832 4,967 5,195 2

10% Increase from SR2S 483 MTC research

20% Increase from SR2S 993 MTC research

30% Increase from SR2S 1,559

School Bike Trip VMT Reduction (1 bike trip equals 2 miles)

966 1,987 3,117 -

Total Walk and Bike Trips

Baseline K-8 Public-Only Walk/Bike Trips (41% of K-12 Total Walk/Bike Trips)

207,997 214,415 228,060 1

K-8 Age Public-Only Walk and Bike Trips (41% of K-12 Total Walk Trips)

85,279

10% SR2S Increase in Walk/Bike Trips 8,528 -

20% SR2S Increase in Walk/Bike Trips 17,582 -

30% SR2S Increase in Walk/Bike Trips 28,051

SR2S Walk/Bike VMT Reduction 9,011 18,575 29,610 - Sources: 1. SANDAG Regional Transportation Demand Model

2. 2008 American Community Survey, U.S. Census Bureau

This measure was an off-model strategy amenable to the CO2 emissions factors found in Table 1. Table 8 includes the results of this strategy.

Table 8 – Draft 2050 RTP - Safe Routes to Schools VMT and CO2 Results


2020 9,011 9,091 0.003

2035 18,575 18,503 0.005

2050 29,610 29,598 0.007

6

Vanpool Strategies The vanpool program strategy is one of the existing programs currently offered by SANDAG that has an impact on GHG reductions. Vanpools have been shown to reduce vehicle miles traveled since only one (albeit larger) vehicle is required to transport the same number of people that single occupant vehicles would normally take 7 to 15 vehicles to transport.

The current vanpool program, which includes 662 vanpools with daily VMT over 74,000 miles7, was assumed to be expanded by about 70 percent by 2020 (to 1,124 vanpools) and by 174 percent by 2035 (to 1,814 vanpools). This future growth is consistent with the vanpool growth rates from 1995 to 2010. Vanpool growth rates fluctuated between 2008 and 2010 primarily due to military deployments and the recession. Military personnel are the largest customer of the vanpool program. The assumption for future growth is based on nominal increase in gas prices, a continuation of a subsidy program (currently $400 per vanpool), van amenities to include Wi-Fi, GPS, travel information and battery charging, and a drop in the vanpool occupancy requirements from 80 percent to 70 percent. Figure 1 shows the expected linear growth trend of the vanpool program.


7 Existing average round trip vanpool mileage of 112 miles route trip per van with an existing occupancy of 8.15 passengers per van was assumed to be

the same in both future horizon years.

Figure 1 – Estimated Vanpool Growth

y = 45.953x - 70.475R² = 0.9505

-500

0

500

1000

1500

2000

FY 95 FY 00 FY 05 FY 10 FY 15 FY 20 FY 25 FY 30 FY 35

Van

pool

s

Fiscal Year

Projected growth trend

Actual growth of vanpools

7

Table 9 – Draft 2050 RTP – Vanpool VMT and CO2 Results


2020 1,143,108 1,153,282 0.326

2035 1,844,838 1,837,643 0.456

2050 2,545,551 2,544,533 0.580

Carpool Strategies SANDAG evaluated the investment in a carpool incentive program to promote the use of fewer vehicles to transport the same number of people to and from work. The program is assumed to begin in 2012 with 5,300 carpools subsidized at $50 per carpooler per month for three months. The assumptions for the number of carpools and the subsidy provided are consistent with the existing regional vanpool program. The approach includes a retention rate assumption after the initial three-month incentive period is discontinued8. The strategy assumes a two percent annual background growth in carpooling based on the assumed expansion of HOV/Express Lanes, the increase in population and travel, the increase in gas prices and increased marketing and outreach efforts for ridesharing9. As a result of these assumptions, there are 30,210 new carpools assumed by 2020 as a result of the program (in addition to the two percent annual background growth [or 24,696 carpoolers] from 2008 to 2020), and an increase of 50,350 new carpools by 2035 (in addition to the two percent annual background growth [or 77,020 carpoolers] from 2008 to 2035). Figure 2 shows the expected growth trend of the carpool program.

Estimated Carpool Growth In terms of mode share, the carpool home-based work mode share reaches approximately 16 percent by 2035 under this strategy (according to trip tables from the Draft 2050 RTP scenario), as opposed to about 11 percent without the program. This deployment level appears reasonable based on the fact that, for example, carpool mode share is already approaching 15 percent in nearby regions.10 The SANDAG iCommute database already has in place a ridematching program and guaranteed ride home program, both of which are essential to operating a successful carpool incentive program. The recommendations of the 2009 Rideshare Incentive Program Study build on these programs and represent a comprehensive approach to implementing a carpool incentive program in San Diego.


8 This retention rate of carpooling was fixed on a sliding scale (assumed to be 90% in the first year and 25% in the tenth year and beyond). The rate was

based on the RideLink Rideshare Incentive Program Study commissioned in 2009. The study looked at other regions in the nation that have implemented a similar incentive program (Riverside County Transportation Commission [RCTC], San Bernardino Associated Governments [SANBAG], Metropolitan Washington Council of Governments [MWCOG], and Atlanta Georgia). These pilot programs typically offered incentives for three months that were valued at approximately $2 per day.

9 The VMT reduction was predicated on the average commuter trip lengths (round trip of 25 miles) from the 2020 and 2030 travel demand models. 10 For example, the 2008 American Community Survey indicates that about 13 percent of commuters traveled in a 2- or 3-person carpool on a typical

work week.

8

Table 10 – Draft 2050 RTP – Carpool VMT and CO2 Results


2020 830,775 838,169 0.237

2035 1,384,625 1,379,225 0.343

2050 1,931,188 1,930,416 0.440

Buspool Programs SANDAG evaluated the deployment of a military buspool program since, like carpools and vanpools, buspools can reduce the number of solo vehicle trips to and from military installations. The benefits of this strategy were calculated by assuming expansion of a new buspool pilot program.

A linear growth trend was assumed for the buspool program, with a goal of transporting 40 percent of military personnel in the region via buspools by 2035. No increase was projected between 2035 and 2050. The program was assumed to be achievable based on the concentrations of military housing, employment locations in the region, and the assumed implementation of mandates requiring military personnel in the region to opt for alternate modes of commute to work. It is expected that the program could result in the participation of 15 percent of regional military personnel (15,766 buspoolers) by 2020 and 40 percent of regional military personnel (41,708 buspoolers) by 2035. The VMT reduction was based on an average round trip length of 32 miles, which is the round trip distance of the Murphy Canyon pilot project. Figure 3 shows the linear regression analysis used to develop the 2020 and 2035 buspool growth rate.

2008 2014 2017 2020 2023 2026 2029 2032 2035 2038 2041 2044 2047 2050

Num

ber

of C

omm

uter

s C

arpo

olin

g to

Wor

k

Year

Figure 2 –Estimated Carpool Growth

9

To reach 15 percent participation of regional military personnel in 2020, about 70 to 100 buspools would be required, or staggered implementation of 8 to 11 buspools per year from 2011 to 2020. It also would require a decrease in the military personnel drive-alone mode share from 94 to 79 percent over a nine-year period. This represents a rapid deployment of the buspools from a pilot into a full-scale program.

This rapid deployment is based on the assumption that the military would enact a trip-reduction mandate in the San Diego region. Military employers have the authority to enact such mandates and, according to SANDAG staff, have shown support for the pilot buspool program. Therefore, aggressive deployment levels appear feasible. This measure was an off-model strategy amenable to the CO2 emissions factors found in Table 1. Table 11 includes the results of this strategy.

Table 11 – Draft 2050 RTP - Buspool VMT and CO2 Results


2020 504,464 508,954 0.144

2035 1,334,624 1,329,419 0.330

2050 1,334,624 1,334,090 0.304

y = 1729.5x - 3E+06

-

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

2005 2010 2015 2020 2025 2030 2035 2040

Figure 3 – Estimated Buspool Growth

10

Summary The six off-model GHG reduction measures described in this report for the Draft 2050 RTP are projected to reduce daily vehicle miles traveled by over seven million miles by 2050, which translates to a daily CO2 emissions reduction of 3,268 metric tons (or 1.642 lbs. CO2 per capita). Table 12 includes the summary of all six off-model strategies for the Draft 2050 RTP. Table 13 summarizes similar data for the Final 2050 RTP for each of the six off-model GHG reduction measures.

Table 12 – Draft 2050 RTP – Summary of Off-Model Strategies VMT and CO2 Results 1


2020 2,748,015 2,772,473 0.784

2035 5,300,526 5,279,854 1.311

2050 7,204,089 7,201,208 1.642 1 Includes Bicycle, Pedestrian, Safe Routes to School, Vanpool, Carpool, and Buspool strategies.

Table 13 – Final 2050 RTP – Summary of Off-Model Strategies VMT and CO2 Results 1

Scenarios Daily VMT Reduction

Lbs. CO2 Reduction

Lbs. Per Capita CO2 Reduction

2020 Bicycle Network 211,672 213,463 0.060

2035 Bicycle Network 617,047 614,572 0.153

2050 Bicycle Network 1,201,406 1,200,146 0.274

2020 Pedestrian Network 47,481 47,883 0.014



2020 Safe Routes to School 9,014 9,090 0.003



2020 Vanpool 1,143,108 1,152,780 0.326

2035 Vanpool 1,844,838 1,837,439 0.456

2050 Vanpool 2,545,551 2,542,883 0.580

2020 Carpool 830,775 837,804 0.237

2035 Carpool 1,384,625 1,379,072 0.343

2050 Carpool 1,931,188 1,929,164 0.440

2020 Buspool 504,464 508,732 0.144

2035 Buspool 1,334,624 1,329,271 0.330

2050 Buspool 1,334,624 1,333,225 0.304

2020 Total 2,746,514 2,769,753 0.784

2035 Total 5,297,354 5,276,107 1.310

2050 Total 7,198,590 7,191,046 1.640 1 Includes Bicycle, Pedestrian, Safe Routes to School, Vanpool, Carpool, and Buspool strategies.

SAN DIEGO ASSOCIATION OF GOVERNMENTS TRAVEL DEMAND

MODEL VALIDATION REPORT

June 2011

SANDAG Travel Demand Model Validation Report ii

TABLE OF CONTENTS

1. EXECUTIVE SUMMARY 3

1.1 REGIONAL MODELING ............................................................................................................................................. 3

2. INTRODUCTION 4

3. TRIP GENERATION 5

3.1 TOTAL TRIPS ........................................................................................................................................................... 5

3.2 TRIPS BY PURPOSE ................................................................................................................................................... 5

3.3 TRIP ENDS BY STRUCTURE TYPE ................................................................................................................................ 6

4. TRIP DISTRIBUTION 7

4.1 COMMUTER TRAVEL FLOWS (3-YEAR ACS: 2006-2008) ............................................................................................ 7

4.2 AVERAGE TRIP LENGTH .......................................................................................................................................... 10

4.3 AVERAGE TRIP DURATION ...................................................................................................................................... 11

5. MODE CHOICE 13

5.1 MODE SHARE ....................................................................................................................................................... 13

5.1.1 MODE SHARE (HOME-WORK) ................................................................................................................................ 14

5.1.2 MODE SHARE (OTHER) .......................................................................................................................................... 15

5.2 MODE SHARE FOR COMMUTING TIME PERIODS ........................................................................................................ 16

5.3 MODE SHARE BY INCOME ...................................................................................................................................... 17

5.4 TRANSIT RIDERSHIP ............................................................................................................................................... 18

6. TRIP ASSIGNMENT 20

6.1 VEHICLE MILES TRAVELED AND HPMS .................................................................................................................... 20

6.2 TRANSIT TRAVEL TIMES .......................................................................................................................................... 20

6.3 AUTO TRAVEL TIMES ............................................................................................................................................. 21

6.4 TRANSIT ROUTING TRAVEL TIME ............................................................................................................................. 22

6.5 TRUCK VOLUMES .................................................................................................................................................. 24

6.6 ROADWAY VOLUMES ............................................................................................................................................ 26

7. WORKS CITED 30

8. APPENDIX A: TRIP LENGTH DISTRIBUTIONS: MODEL (2008) VS. SURVEY (2006) 31

9. APPENDIX B: GLOSSARY 34

SANDAG Travel Demand Model Validation Report 3

1. EXECUTIVE SUMMARY

The current San Diego Association of Governments (SANDAG) transportation and land use models are

consistent with the models used throughout the country, and SANDAG along with the other major California

metropolitan planning organizations are leading the development of the next generation of transportation,

land use, and economic models. SANDAG deals with many complex issues facing the San Diego region,

including the development of the Draft 2050 Regional Transportation Plan (RTP). Transportation and land

use models perform a very basic yet vital set of functions. Models are the principal tools used for alternatives

analysis, and they provide planners and decision-makers with information to help them equitably allocate

scarce resources.

1.1 Regional Modeling

Regional models provide the capability to account for the varied and complex forces that are at work within

the social, economic, and physical aspects of the regional environment. The models provide valuable insights

into many important questions in a short amount of time for more informed decisions. Regional models also

educate planners and policymakers beyond their intuitive judgment. They provide answers to questions like

“If we change this policy, how might that affect the region in the years to come?” These answers can help

policymakers make the important decisions that shape the region’s future.

This report provides a guide to the accuracy of the SANDAG transportation model compared to observed

data in and around 2008. With every statistic, the SANDAG model estimates provide comparable results to

observed data. This report confirms that the SANDAG model is properly calibrated and suitable for use in

evaluating the impacts of the Draft 2050 RTP.

In this report, the SANDAG model is compared to a variety of observed data sources from the last two

decades. The data sources include:

• 1995 SANDAG Household Travel Survey (SDHTS);

• 2006 SANDAG Household Travel Surveys (SDHTS);

• 2009 SANDAG Onboard Transit Passenger Survey (OBS);

• 2001 Caltrans Statewide Household Travel Survey (CHTS);

• 2009 National Household Travel Survey (NHTS);

• 2000 U.S. Census;

• American Community Survey (ACS);

• SANDAG Passenger Counting Program; and

• State and local traffic counts.

This report provides detailed technical information on how the transportation model was used to support

the development and decision-making process for the Draft 2050 RTP and its

Sustainable Communities Strategy, and the Draft Programmatic Environmental Impact Report.


2. INTRODUCTION

For the Draft 2050 Regional Transportation Plan (RTP), the San Diego Association of Governments (SANDAG)

used an enhanced four-step transportation model. Four-step models have been the standard in

transportation modeling since the late 1950s, and they are used by nearly every

metropolitan planning organization in the United States for the development of transportation plans,

corridor studies, Federal Transit Administration New Starts proposals, and air quality analyses. SANDAG is

considered a national leader in transportation modeling best practices. The estimates of regional

transportation-related emissions analyses meet the requirements established in the

Transportation Conformity Rule, 40 CFR Sections 93.122(b) and 93.122(c). These requirements relate to the

procedures used to determine regional transportation-related emissions, including the use of network-based

travel models, methods to estimate traffic speeds and delays, and the estimation of vehicle miles of travel.

This validation report compares a variety of SANDAG model estimates for the 2050 Draft RTP base year 2008

to the 1995 and 2006 SANDAG Household Travel Surveys (SDHTS), 2009 SANDAG Onboard Transit

Passenger Survey (OBS), 2001 Caltrans Statewide Household Travel Survey (CHTS), 2009 National Household

Travel Survey (NHTS), 2000 U.S. Census, 2008 American Community Survey (ACS), passenger counting

program, and traffic counts. This report provides a guide to the accuracy of the SANDAG transportation

model compared to observed data in and around 2008. This report follows the four major steps of the

transportation model:

• Trip Generation (Section 3);

• Trip Distribution (Section 4);

• Mode Choice (Section 5); and

• Trip Assignment (Section 6).

While the SANDAG model estimates are comparable to observations around 2008, a few data themes were

encountered throughout the report. First, none of the observation data used is perfect. For example, the

2006 SDHTS was unable to capture enough college and university students to create a statistically-confident

picture of Home-College trip making behavior. In addition, none of the data provides a comprehensive view

of how Mexican residents impact the San Diego transportation system. Next, in some cases, the validation

data conflict among sources or provide confusing trends. For instance, trends in transit mode share are

different among California Department of Transportation (Caltrans) (declining), Census (flat), and SANDAG

(rising) surveys. Finally, trends in transportation behavior are gradual; however, the latest economic recession

has changed some habits quickly and dramatically. As an example, vehicle miles traveled is down from a high

in 2004 changing a multi-decade trend of year-over-year increases. The new trend is most likely temporary

and due to higher energy prices and unemployment rates. The SANDAG transportation model provides a

systematic analytical platform so that these data conflicts, different alternatives, and inputs can be evaluated

in an iterative and controlled environment.


3. TRIP GENERATION

Trip generation is the first step in the transportation modeling process. Trip generation estimates the

average weekday number of trip productions and attractions, or trip ends, in the region based on land use

and demographic information from the regional growth forecast. Trips ends are calculated for multiple

forms of transportation, including automobiles, public transit, bicycling, walking, and trucks, starting and

ending in each transportation analysis zone, and are categorized by ten trip types (e.g. home to work, home

to shop, other).1

3.1 Total Trips

Table 1 compares total trips generated by the SANDAG model to the 1995 and 2006 San Diego household

travel surveys. The comparison only covers internal (in San Diego County) trips since the standard travel

surveys do not collect information about San Diego travelers living outside of the region. The results from

the travel surveys also do not include travel related to commerce and freight in the region, or about

15 percent to 20 percent of total trips in the region. Underreporting trips also is an issue related to travel

surveys; in traditional surveys, like those conducted by SANDAG, trips are underreported by as much as

30 percent (Pierce, Casas, & Giaimo, 2003), and Caltrans estimated underreporting at 29 percent in California

in its last survey (Caltrans, 2003).

Finally, trips would be expected to grow from 2006 to 2008 in order to keep pace with population growth in

the region. This growth would be slightly tempered due to the travel impacts from the 2007-2009 economic

recession (National Bureau of Economic Research, Inc., 2010). As a guide, the San Diego region population

grew by 17 percent, jobs grew by 30 percent between 1995 and 2006, and reported trips grew by 19 percent.

Between 2006 and 2008, the San Diego region population grew by approximately 2 percent, while

employment growth was flat.

Table 1: Comparison of Total Trips between Travel Surveys and Model Estimates

1995 SDHTS 2006 SDHTS

2006 SDHTS

Low Adjustment*

2006 SDHTS

High Adjustment*

2008 Model

(internal trips)

Total Trips 8,531,158 10,146,846 16,196,800 18,481,800 16,687,379

* Low Adjustment: 2006 SDHTS inflated with 2 percent population change, 29 percent trip underreporting, and 10 percent commercial vehicle trips.

* High Adjustment: 2006 SDHTS inflated with 2 percent population change, 30 percent trip underreporting, and 20 percent commercial vehicle trips.

3.2 Trips by Purpose

A comparison of the distribution of trips by purpose between the travel demand model and various travel

surveys is shown in Table 2. Since modeled commercial trips are largely captured in the non-home-base trips

along with person level non-home based trips, Table 2 excludes non-home-based trips from the comparison

1 Additional information on trip generation is available on the SANDAG Web site. http://www.sandag.org/models.


because commercial vehicle trips are not included in the travel surveys. As noted earlier, commercial trips

account for approximately 15 percent to 20 percent of the travel in the region and would therefore skew the

results of Table 2. The model estimate of each trip purpose is similar to survey proportions.

Table 2: Percentage of Trips by Purpose, excluding Non-Home-Based Trips

1995 SDHTS 2006 SDHTS

2008 Model

(internal trips)

2006 vs.

Model Difference

%Home-Work 20.6% 18.0% 19.5% +1.5%

%Home-College 2.1% 1.7% 2.3% +0.6%

%Home-School 11.5% 11.0% 9.0% -2.0%

%Home-Other 51.8% 51.1% 51.6% +0.5%

%Serve Passenger 14.0% 18.1% 17.5% -0.6%

3.3 Trip Ends by Structure Type

Table 3 compares person trip ends per household between the SANDAG model and various travel surveys.

The travel demand model estimates person trip rates higher than rates from the 2006 survey, because the

model inflates person trips (and therefore household trips) to account for commercial travel in the region.

The trip rates in Table 3 reflect the regional average rate by dwelling unit type. In certain areas, adjustments

are made to reflect actual conditions such as deviations from average household size and vacancy rates. Trip

rates also are affected by household age structure.

Table 3: Person Trip Ends Per Household

1977

SDHTS

1986

SDHTS

1991*

Caltrans

1995

SDHTS

2001*

Caltrans

2006

SDHTS

2008 Model

(Internal Trips)

Single-family DU 14.1 - 11.1 - 9.3 10.7 12.2

Multi-family DU 9.4 - 7.0 - 5.7 7.7 8.7

Mobile home DU 5.4 - - - - 6.5 6.6

Total (weighted average) 11.2 11.6 9.5 9.2 7.9 9.5 10.7

* Caltrans does not include walk, bike, or "other"


4. TRIP DISTRIBUTION

Trip distribution allocates and balances trip productions and attractions through a gravity approach based on

trip end density and location by zone. Trip distribution considers the distance between trip ends based on

the assumed roadway and public transportation networks for a given year. The model is designed to modify

trip patterns in response to new land use developments and transportation facility changes. For example, the

opening of a new shopping center would shift trips from other nearby shopping areas to the new

development. Another example would be the introduction of mixed-use development. In this case the model

would yield shorter trip lengths by recognizing the increased opportunity for interaction between residential

and commercial areas in the development.

The model is calibrated to match observed trip length frequencies from the 1995 and 2006 Travel Behavior

Surveys. Zone-to-zone impedances are a composite measure of peak and off-peak travel times and costs by

auto, transit, and non-motorized modes.

4.1 Commuter Travel Flows (3-Year American Community Survey (ACS):

2006-2008)

Overall, the four-step model replicates observed commuter flows well. Larger cities are better approximated

due to the larger observed sample size and trip flows. Figure 1 illustrates the close correspondence between

scaled 2006-2008 ACS city-to-city flows and the SANDAG modeled flows. The correlation coefficient is .9995

and the R-Squared is .9991. R-Squared is a measure of goodness of fit between 0 and 1 comparing observed

and estimated values with a value of 1 indicating a perfect fit. These high values for correlation coefficient

and R-Squared are partially driven by a large magnitude observation with strong correlation between

observed and modeled (City of San Diego to City of San Diego flows) that skew the results.

Figure 2 shows the same graph after the City of San Diego to City of San Diego observation has been

dropped. The correlation coefficient is .986 and the R-Squared is .972. Census Transportation

Planning Package (CTPP) 2006-2008 flows are highly correlated with the SANDAG model. The R-Squared of

.972 indicates that 97.2 percent of the variation in scaled ACS commuter flows is being accounted for by the

SANDAG model.


Figure 1: Scatter Plot of Worker flows from City to City2 Between CTPP and Model with Linear Trend Line

Figure 2: Scatter Plot of Worker flows from City to City (excl. City of San Diego to City of

San Diego) Between CTPP and Model with Linear Trend Line

Commuter Travel Flows (2000 Census)

2 Excludes Del Mar and Solana Beach due to Census data suppression for small jurisdictions.

Correlation Coefficient: .986

R-Squared: .972

One-to-one correspondence

Trend line


R-Squared: .9991


Trend line


Figure 4 and Figure 5 repeat the analysis of Section 4.1 using scaled 2000 Census flow data at the

Metropolitan Statistical Areas (MSA) to MSA level (See Figure 3). As before, there is a good correspondence

between observed MSA to MSA flows and modeled MSA to MSA flows.

Figure 3: MSA for San Diego County

Figure 4: Scatter Plot of Worker flows from MSA to MSA Between CTPP (2000) and Model (2008) with Linear Trend Line


R-Squared: .9877


Trend line


Figure 5: Scatter Plot of Worker flows from MSA to MSA Between CTPP (2000) and Model (2008) with Linear Trend Line (without outlier)

4.2 Average Trip Length

Table 4 compares average modeled trip length by purpose with the average observed trip length by purpose

in the 2006 household survey. Modeled trip lengths by purpose are close to observed 2006 survey results. The

difference between modeled and observed trip lengths is less than 10 percent for every purpose.

Trip length distribution graphs are included in Appendix A for each trip purpose comparing the model and

the 2006 household survey. For all trips regardless of purpose, the coincidence ratio between modeled and

observed trip length distributions is 0.90. A coincidence ratio of 1 indicates the two distributions are

identical. As the distribution graphs and coincidence ratios indicate, for most purposes, there is a good match

between modeled and observed trip length distributions. The Home-College trip purpose is the most

dissimilar between the modeled and observed trip length distributions having a coincidence ratio of 0.64.

The lower coincidence ratio is a result of the few number of observed Home-College trips in the travel

survey.


R-Squared: .9802


Trend line


Table 4: Average Trip Length by Purpose (miles)

2006 SDHTS 2008 Model

(internal trips)

%Difference Coincidence Ratio

Total 5.9 5.7 -3.4% 0.90

Home-Work 10.9 11.8 +8.3% 0.88

Home-College 9.7 10.1 +4.1% 0.64

Home-School 3.2 2.9 -9.4% 0.74

Home-Other 5.5 5.0 -9.1% 0.91

Non-Home-Based 5.6 5.2 -7.1% 0.89

Serve Passenger 4.1 4.4 +7.3% 0.86

4.3 Average Trip Duration

Table 5 through Table 7 compares average self-reported trip duration in the 2006 household survey and the

2006-2008 ACS with average SANDAG modeled trip duration by mode for internal trips. The self-reported

durations tend to be longer than the modeled durations which reflect traveler perception of longer travel

times plus a reduction in actual roadway congestion between 2006 and 2008 due to the economic recession.

In addition, most people round their self-reported estimates to the nearest next five-minute increment as

opposed to the “exact” real number reported by the model. Finally, as seen in Table 6, the San Diego survey

results differ from the ACS results making model validation difficult. Most likely, the large differences

between the San Diego survey are a result of margin of errors in the ACS data.

Table 5: Average Trip Duration in Minutes for All Trip Types (SDHTS – Self-Reported)

SDHTS

2006

Model

2008 Percent Difference

Drive-Alone 18.78 12.74 -32.2%

Carpool-2 17.60 11.22 -36.3%

Carpool-3+ 15.81 10.68 -32.4%

Walk 15.27 16.96 +11.1%

Bicycle 21.32 15.99 -25.0%

Transit 56.22 51.09 -9.1%

Total 18.99 12.65 -33.3%


Table 6: Average Trip Duration in Minutes for Commute Trip Types (SDHTS / ACS – Self-Reported)

SDHTS

2006

ACS

(2006-2008)

Model

2008

Percent

Difference

(Model / SDHTS)

Percent

Difference

(ACS / SDHTS)

Drive-Alone 25.99 24.0 21.27 -18.2% -11.4%

Carpool-2 28.57 26.3 21.41 -25.1% -18.6%

Carpool-3+ 28.52 29.1 21.31 -25.3% -26.8%

Walk 16.21 9.7 18.2 +12.3% +87.6%

Bicycle 26.06 21.4 18.86 -27.6% -11.9%

Transit 67.53 49.7 57.32 -15.1% +15.3%

Total 27.99 24.9 23.35 -16.6% -6.2%

Table 7: Average Trip Duration in Minutes for Non-Commute Trip Types (SDHTS – Self Reported)

SDHTS

2006

Model

2008

Percent Difference

Drive-Alone 16.50 10.93 -33.8%

Carpool-2 15.12 10.92 -27.8%

Carpool-3+ 14.74 10.92 -25.9%

Walk 11.50 9.59 -16.6%

Bicycle 21.38 10.02 -53.1%

Transit 43.06 34.33 -20.3%

Total 16.18 10.99 -32.1%


5. MODE CHOICE

Mode choice splits total person-trip movements between zones into different forms of transportation by

auto, transit, and non-motorized modes (bicycling and walking). The mode choice step selects the most likely

form of transportation for each trip, based on access, traveler’s income, trip purpose, parking costs, fuel

price, transit fares, travel time, and other time and pricing parameters.

Auto modes include drive-alone non-toll, drive-alone toll, shared-ride non high occupancy vehicle (HOV)/non-

toll, shared-ride HOV/non-toll, and shared-ride HOV/toll. Each HOV mode is further identified as either a two-

person HOV or three-plus person HOV. Transit modes are differentiated by five ride modes (Commuter Rail,

Light Rail, Bus Rapid Transit, Express Bus, and Local Bus) and three access modes (walk, drive, and drop-off).

The mode choice model is designed to link mode use to demographic assumptions, highway network

conditions, transit system configuration, land use alternatives, parking costs, transit fares, and auto operating

costs. Trips between zone pairs are allocated to modes based on the cost and time of traveling by a particular

mode, compared with the cost and time of traveling by other modes.

Income level also is considered, because lower-income households tend to own fewer automobiles and

therefore make more trips by transit and carpooling. People in higher income households tend to choose

modes based on time and convenience rather than cost (Pratt & Park, 2000). The mode choice model is

calibrated using 1995 and 2006 Travel Behavior Surveys trip tables by mode and income, as well as 2001-2003

Regional Transit Survey transit trip characteristics. Regional-level Census 20003 work-trip mode shares also

were used to fine-tune mode-share estimates.

5.1 Mode Share

Table 8 compares modeled mode shares for all trips with mode shares from various surveys. Modeled drive-

alone mode share is higher than in the 1995 and 2006 travel surveys due to the exclusion of commercial

vehicles from the travel surveys. By including commercial vehicles in the model estimates, mode share

estimates skew toward drive-alone and auto in general. As covered in subsequent tables, most of the auto

overages compared to the travel surveys are captured in non-home-based travel where most commercial trips

occur.

3 Necessary ACS data was not readily available at the time of model calibration.


Table 8: Total Trips, Compare Modeled Mode Share (2008) with Travel Surveys4

1991

Caltrans

1995

SDHTS

2001

Caltrans

2006

SDHTS

2008 Model

(internal trips)

%Auto 88.4 95.4 92.7 86.6 94.8

%Drive-Alone - 47.9 - 51.3 52.2

%Carpool - 42.2 - 35.3 42.6

%Transit 1.5 1.4 1.2 2.6 1.4

%SchoolBus 1.9 - 0.7 1.3 0.7

%Walk 7.1 5.5 4.2 8.4 2.8

%Bike 0.8 0.5 0.4 0.7 0.4

5.1.1 Mode Share (Home-Work)

Table 9 compares home-work mode share between the four-step model and various surveys. The model

estimated commute mode shares largely reflect the survey data accurately with the largest absolute

difference of 2.8 percent overestimation overall of carpooling. Modeled drive-alone mode share is slightly

lower than most surveys except the 1991 Caltrans survey. While the surveys show a declining carpool mode

share over time (from 12.4 percent in 1991 to 8.1 percent in 2006), the modeled carpool mode remains

higher at 10.9 percent.

Modeled transit mode share is higher than in the older surveys, but similar to the 2006 survey. Comparing

Table 9 to Table 19, overall transit boardings are over-estimated, so the transit overage is due to transit

ridership favoring commute trips.

Home-work walk mode share in the model is similar to the surveys. The commute walk share is inconsistent

over time comparing Table 9 (Travel Surveys) where walk shares are increasing over time to Table 10 (Census)

where walk mode share is declining over time.

Table 9: Home-Work Trips, Compare Modeled Mode Share (2008) with Travel Surveys

1991

Caltrans

1995

SDHTS

2001

Caltrans

2006

SDHTS

2008 Model

(internal trips)

%Drive-Alone 73.4 83.3 86.2 83.2 80.3

%Carpool 12.4 11.9 9.6 8.1 10.9

%Transit 3.5 2.3 2.2 4.8 6.0

%Walk 1.1 1.7 1.3 2.6 1.8

%Bike 0.8 0.6 0.5 1.1 1.1

Table 10 compares home-work mode shares between the model and Census/ACS sources, which have a larger

sample size than the household behavior surveys. Modeled drive-alone mode share corresponds very closely

4 By including commercial vehicles in the model estimates, mode share estimates skew towards drive-alone and auto in general.


to the 2005-2009 ACS, and underscores a multi-decade trend toward a higher proportion of commuters

driving alone. Modeled carpool mode share is similar to the 2005-2009 ACS, and highlights a multi-decade

trend toward a lower proportion of commuters carpooling. Modeled base-year transit mode share is higher

than all Census and ACS results, dating back to 1970. While walk mode share is lower in the SANDAG model

than in the 2005-2009 ACS, the modeled walk mode share exhibits a multi-decade trend toward a lower

proportion of commuters who walk to work.

Table 10: Home-Work Trips, Compare Modeled Mode Share (2008) with Census/ACS

1980

Census

1990

Census

2000

Census

2005-2009

ACS

2008 Model

(internal trips)

%Drive-Alone 65.1 74.6 77.3 80.2 80.3

%Carpool 17.8 14.5 13.6 11.4 10.9

%Transit 3.4 3.4 3.5 3.5 6.0

%Walk 10.1 4.8 3.6 3.0 1.8

%Bike - 0.9 0.6 0.6 1.1

5.1.2 Mode Share (Other)

Table 11 through Table 13 compares modeled mode shares for other trip purposes with various surveys.

Home-College/School overestimate drive-alone trips largely as a result of underestimating carpool pick-up

and drop-off of non-driving age schoolchildren. Home-based other and non-home-based trips are skewed

toward auto compared to the survey results due to the survey’s exclusion of commercial vehicles and trucks.

Table 11: Home-College/School Trips, Compare Modeled Mode Share (2008) with Travel Surveys

1995

SDHTS

2006

SDHTS

2008 Model

(internal trips)

%Drive-Alone 14.6 12.9 26.7

%Carpool 44.1 50.9 43.1

%Transit 1.7 3.8 2.2

%School Bus - 10.9 9.6

%Walk 22.8 19.9 16.9

%Bike 1.9 1.7 1.3

+


Table 12: Home-Other, Compare Modeled Mode Share (2008) with Travel Surveys

1995

SDHTS

2006

SDHTS

2008 Model

(internal trips)

%Drive-Alone 39.9 48.8 49.0

%Carpool 53.9 37.6 45.5

%Transit 1.1 3.0 1.6

%Walk 4.2 8.9 3.4

%Bike 0.4 0.8 0.4

Table 13: Non-Home-Based, Compare Modeled Mode Share (2008) with Travel Surveys5

1995

SDHTS

2006

SDHTS

2008 Model

(internal trips)

%Drive-Alone 55.4 49.2 55.9

%Carpool 36.9 42.5 42.4

%Transit 1.4 1.6 0.3

%Walk 3.5 5.1 1.2

%Bike 0.1 0.3 0.1

5.2 Mode Share for Commuting Time Periods

Table 14 and Table 15 compare modeled mode shares for commuting in the peak and off-peak periods of the

day with various surveys and ACS results. Note that in Table 14 and Table 15, Caltrans and the ACS report

mode shares by time of day in a way that is not consistent with the four-step model’s definition of peak

period hours (6 to 9 a.m. and 3 to 6 p.m.). The hours for each survey are noted in the column heading.

Table 14: Home-Work Trips in Peak Hours (6 to 9 a.m., 3 to 6 p.m.), Compare Modeled Mode Share (2008) with Travel Surveys and ACS/Census

7 - 9 a.m. 1991

Caltrans

7 - 9 a.m. 2001

Caltrans

6 - 9 a.m.

2006-08 ACS

6 - 9 a.m.

2005-09 ACS

2006

SDHTS

2008 Model

(internal trips)

%Drive-Alone 68.7 87.5 80.7 81.4 83.3 80.2

%Carpool 16.6 7.5 11.7 11.4 7.7 10.8

%Transit 3.0 2.7 3.2 3.1 5.1 6.3

%Walk 1.5 0.9 2.7 2.5 2.4 1.6

%Bike 0.0 1.0 0.5 - 1.2 1.0

5 By including commercial vehicles in the model estimates, mode share estimates skew towards drive-alone and auto in general.


Table 15: Home-Work Trips in Off-Peak Hours, Compare Modeled Mode Share (2008) with Travel Survey

9 - 6 a.m.

2006-2008 ACS

9 – 6 a.m.

2005-2009 ACS

2006

SDHTS

2008 Model

(internal trips)

%Drive-Alone 77.8 78.1 83.1 80.4

%Carpool 11.5 11.4 8.8 10.9

%Transit 4.2 4.2 4.4 5.5

%Walk 3.9 3.8 2.8 2.0

%Bike 0.9 0.8 1.2

5.3 Mode Share by Income

Table 16 through Table 18 compares commute mode shares by income from the 2006 survey and 2006-2008

ACS with modeled commute mode share by income. Note that the income categories vary by source and are

not consistent with the model. A San Diego region-specific CPI was applied to the income categories used in

the 2006 survey and 2006-2008 ACS so that categories would be comparable with those used in the model,

which are expressed in 1999 dollars. For example, in 1999 dollars, low income refers to households making

less than $30,000 a year in the model, less than $34,090 in the 2006 survey, and less than $30,308 in the 2006-

2008 ACS.

Table 16: Home-Work Trips by Low-Income Households, Compare Modeled Mode Share (2008) with CTPP 2006-2008 and 2006 Travel Survey

2006 SDHTS 2006-2008 ACS 2008 Model

%Auto 80.7 85.8 75.1

%Drive-Alone 67.5 73.3 60.7

%Carpool 13.2 12.5 14.3

%Transit 10.1 7.3 17.8

%Walk 6.9 4.4 4.7

%Bike 1.1 0.9 2.5

Table 17: Home-Work Trips by Mid-Income Households, Compare Modeled Mode Share (2008) with CTPP 2006-2008 and 2006 Travel Survey

2006 SDHTS 2006-2008 ACS 2008 Model

%Auto 91.2 91.3 91.1

%Drive-Alone 83.8 78.9 80.0

%Carpool 7.4 12.4 11.1

%Transit 6.1 4.1 6.1

%Walk 2.2 2.5 1.7

%Bike 0.5 0.7 1.0


Table 18: Home-Work Trips by High-Income Households, Compare Modeled Mode Share (2008) with CTPP 2006-2008 and 2006 Travel Survey

2006 SDHTS 2006-2008 ACS 2008 Model

%Auto 95.1 94.7 97.4

%Drive-Alone 88.8 84.0 87.7

%Carpool 6.3 10.7 9.7

%Transit 2.7 2.1 1.4

%Walk 1.0 1.5 0.7

%Bike 1.2 0.5 0.6

5.4 Transit Ridership

Table 19 through Table 21 compares observed transit boarding and passenger miles from the SANDAG

passenger counting program (PCP). Each year, SANDAG personnel ride transit to collect boardings, alightings,

and other information with handheld devices or paper survey forms. The agency also processes a greater

amount of data from automatic passenger counters (APCs) on MTS buses. PCP boardings are single day

counts and do not reflect transit ridership variations throughout the year.

Systemwide the model overestimates transit boardings and passenger miles as seen previously in the mode

choice home-work trips (Table 9). The COASTER, or commuter rail, was the only route largely underestimated

by percent difference. COASTER boardings tend to fluctuate with gas prices and I-5 congestion levels and PCP

counts for the COASTER the following two years were 4,835 in 2009 and 4,709 in 2010. The model is close to

matching the peak to off-peak split in Table 21 with one percent or less variation for both boardings and

passenger miles.

Table 19: Comparison of Modeled to Average Weekday Observed Transit Boardings

Mode Modeled Observed Difference %Diff

Commuter Rail 4,704 6,488 -1,784 -27.5%

Light Rail 124,919 122,773 2,146 +1.7%

SPRINTER 7,719 6,560 1,159 +17.7%

Blue Line 62,976 65,315 -2,339 -3.6%

Orange Line 27,443 27,994 -551 -2.0%

Green Line 26,780 22,904 3,876 +16.9%

Bus (by mode) 267,876 216,015 51,861 +24.0%

Limited Bus 1,445 1,232 213 +17.3%

Express Bus 16,077 10,483 5,594 +53.4%

Local Bus 250,354 204,300 46,054 +22.5%

Bus (by operator)

MTS Bus 230,914 185,784 45,130 +24.3%

NCTD Bus 36,961 30,231 6,730 +22.3%

System Total 397,499 345,276 52,223 +15.1%


Table 20: Comparison of Modeled to Average Weekday Observed Passenger Miles

Mode Modeled Observed Difference %Diff

Commuter Rail 112,333 172,998 -60,665 -35.1%

Light Rail 876,858 744,880 131,978 17.7%

SPRINTER 53,849 53,981 -132 -0.2%

Blue Line 520,458 416,187 104,271 25.1%

Orange Line 154,726 147,473 7,253 4.9%

Green Line 147,825 127,239 20,586 16.2%

Bus (by Mode) 1,077,348 807,009 270,339 33.5%

Limited Bus 37,512 29,761 7,751 26.0%

Express Bus 162,846 80,447 82,399 102.4%

Local Bus 876,990 696,801 180,189 25.9%

Bus (by Operator)

MTS Bus 903,692 666,560 237,132 35.6%

NCTD Bus 173,656 140,448 33,208 23.6%

System Total 2,066,538 1,724,887 341,652 19.8%

Table 21: Comparison of Model vs. Observed Peak/Off-Peak Transit Usage Percentages

Measure Modeled Observed

Boardings, %Peak 44.2 43.8

Boardings, %Off-Peak 55.8 56.2

Passenger Miles Traveled, %Peak 47.1 46.1

Passenger Miles Traveled, %Off-Peak 52.9 53.9


6. TRIP ASSIGNMENT

During network assignment, the model places each trip on the most efficient auto, transit, or non-motorized

path based on the mode of transportation that was chosen earlier. Highway assignment produces traffic-

volume estimates for all roadway segments in the system. These traffic volumes are an important input to

emissions modeling. Similarly, transit trips are assigned to transit routes and segments.

Model accuracy is assessed by comparing model estimated traffic volumes with actual traffic counts obtained

through the SANDAG traffic monitoring program and the Highway Performance Monitoring System (HPMS)

estimates of Vehicle Miles of Travel (VMT) developed by the U.S. Federal Highway Administration (FHWA).

For accuracy, transit ridership forecasts from the transit assignment model are compared with transit counts

from the SANDAG transit passenger counting program to determine whether transit modeling parameters

need to be adjusted.

6.1 Vehicle Miles Traveled and HPMS

Table 22 compares Caltrans HPMS estimate of total daily VMT. HPMS VMT was adjusted up 5 percent for

comparison because HPMS reports daily VMT instead of weekday VMT. The 5 percent adjustment was

derived from traffic count data. SANDAG overestimates VMT by less than 1 percent.

Table 22: Comparison of Weekday VMT between HPMS and Model Estimates

2008 Observed 2008 Modeled Diff %Diff

VMT

VMT/capita

80,317,062

25.6

80,353,686

25.7

36,624

0.1

0.005%

0.4%

6.2 Transit Travel Times

Appendix 3 of the Draft 2050 RTP provides forecasts of peak-period average travel times for 11 corridors

within the region. Table 23 compares modeled transit travel times with transit travel times according to

Google’s online transit planner for each of the 11 corridors. For a representative intersection at each end

point, the online transit planners were asked to suggest transit itineraries with a preferred departure time of

6:30 a.m. The suggested transit itinerary with the shortest duration is selected to compare with modeled

travel time. Any suggested transit itinerary with a departure time past 9 a.m. (which marks the end of the

morning peak, as defined in the model) is excluded. Note that in the case of the last corridor in the table

(Ramona to downtown San Diego), transit travel times suggested by the online transit planners are averaged

over the whole day because Google’s transit planner did not list a transit itinerary for this corridor that

departed during the morning peak. The corridor with the greatest absolute difference and percentage

difference between modeled and Google transit travel time is Western Chula Vista to

Mission Valley (39.1% difference). For this corridor, the majority of the difference comes from differing

transit transfer wait times between the model and Google’s suggested itinerary.


Table 23: Transit Travel Time for Selected Origin-Destination Pairs-Model (2008) vs Google/ SD Commute transit planners (2011)

Origin Destination Modeled Time

(2010)

Google

(2011)

Diff vs

Google

%Diff vs

Google

Oceanside Downtown SD 92.0 92 0.0 0.0%

Escondido Downtown SD 76.5 68 8.5 12.5%

El Cajon Kearny Mesa 70.3 67 3.3 4.9%

Mid-City UTC 47.5 49 -1.5 -3.1%

Western Chula

Vista

Mission Valley 59.8 43 16.8 39.1%

Carlsbad Sorrento Mesa 88.3 99 -10.7 -10.8%

Oceanside Escondido 76.7 70 6.7 9.6%

San Ysidro Downtown SD 37.6 46 -8.4 -18.3%

Otay Ranch UTC 111.6 122 -10.4 -8.5%

Pala/Pauma Oceanside Transit

Center

96.6 100 -3.4 -3.4%

SR 67 (Ramona) Downtown SD 121.8 113 8.8 7.8%

6.3 Auto Travel Times

Table 24 compares modeled highway travel times for the Draft 2050 RTP corridors with observed highway

travel times. Seven of the corridors were identified where there was significant overlap with a pre-defined

Caltrans Performance Measuring System (PeMS) route that has been collecting travel time data since at least

2008. Modeled a.m. peak and p.m. peak travel times are then compared with the average a.m. peak and

p.m. peak travel times according to PeMS for weekdays during the month of April, 2008. For the most part,

the modeled highway travel times are similar to the observed travel times as indicated by the small absolute

difference (in minutes).

The greatest differences between modeled travel times and PeMS data occur for routes involving

I-805 northbound. It is worth noting that the I-805 northbound routes have a.m. peak travel times that vary

more over the course of the peak period than most of the other routes here. For example, for

I-805 northbound between I-8 and I-5, whereas 6 a.m. mean travel time in PeMS is 9.7 minutes, 8 a.m. mean

travel time is 15.3 minutes, which is much closer to the modeled result (shrinking the percent difference to

6.5%). For I-805 northbound between State Route (SR) 54 and I-8, whereas 6 a.m. mean travel time in PeMS is

7.7 minutes, 7:30 a.m. mean travel time is 11 minutes, which also is closer to the modeled result (shrinking

the percent difference to 3.6%). The model is not designed to replicate temporal variability.

PEMs estimates are also affected by the frequency of detector stations. The model's highway links are usually

significantly shorter than the highway segment associated with each detector station, so PeMs may miss

certain local traffic conditions.


Table 24: Highway Travel Time (TT) in Minutes-Model vs. PeMS (pre-defined PeMS routes)

RTP Corridor Model AM

Peak TT

Model PM

Peak TT

PeMS

AM Peak

PeMS PM

Peak

Difference

AM

Difference

PM

Percent Diff

AM

Percent Diff

PM

Oceanside to

Downtown SD 34.0 29.9 35.1 31.9 -1.1 -2.0 -3.1% -6.3

Escondido to

Downtown SD 35.2 32.8 33.7 32.0 1.5 0.8 4.5% 2.5%

Mid-City to UTC 16.3 9.7 12.5 10.0 3.8 -0.3 30.4% -3.0%

Oceanside to

Escondido 15.8 17.5 15.5 21.2 0.3 -3.7 1.9% -17.4%

San Ysidro to

Downtown SD 5.5 5.3 5.5 5.4 0.0 -0.1 0.0% -1.9%

Otay Ranch to UTC 11.4 6.9 9.5 7.8 1.9 -0.9 20.0% -11.5%

SR67 (Ramona) to

Downtown SD 10.6 11.7 11.3 10.5 -0.7 1.2 -6.2% 11.4%

6.4 Transit Routing Travel Time

Table 25 and Figure 6 compare modeled transit travel times for specified transit routes (from the beginning

of the route to the end) with the transit travel times listed in official transit schedules. Figure 6 plots the

modeled versus scheduled transit travel times for the routes. Modeled route durations tend to be lower than

what is listed in the transit schedule.


Table 25: Modeled Transit Travel Times (2008) vs. Transit Schedules (2011) 6

Model Travel Time Schedule Travel Time Difference (min) % Difference

ROUTE DIR PK OP PK OP PK OP PK OP

COASTER IB 55 55 60 61 -5 -6 -8.3% -9.8%

COASTER OB 55 55 61 61 -6 -6 -9.8% -9.8%

SPRINTER IB 50 50 53 53 -3 -3 -5.7% -5.7%

SPRINTER OB 50 50 53 53 -3 -3 -5.7% -5.7%

Blue Line IB 53 53 56 56 -3 -3 -5.4% -5.4%

Blue Line OB 53 53 56 56 -3 -3 -5.4% -5.4%

Orange Line IB 60 60 63 63 -3 -3 -4.8% -4.8%

Orange Line OB 60 60 62 62 -2 -2 -3.2% -3.2%

Green Line IB 41 41 47 47 -6 -6 -12.8% -12.8%

Green Line OB 41 41 48 48 -7 -7 -14.6% -14.6%

810 IB 71 52 67 61 4 -9 6.0% -14.8%

810 OB - 54 - 61 - -12 - -18.2%

20 IB 123 98 100 93 23 5 23.0% 5.4%

20 OB 109 96 103 93 6 3 5.8% 3.2%

2 IB 49 48 40 37 9 11 22.5% 29.7%

2 OB 40 42 43 39 -3 3 -7.0% 7.7%

7 IB 69 68 68 71 1 -3 1.5% -4.2%

7 OB 58 61 76 73 -18 -12 -23.7% -16.4%

11 IB 101 100 93 101 8 -1 8.6% -1.0%

11 OB 100 100 104 103 -4 -3 -3.8% -2.9%

30 IB 83 84 89 89 -6 -5 -6.7% -5.6%

30 OB 99 92 100 96 -1 -4 -1.0% -4.2%

41 IB 37 36 48 45 -11 -9 -22.9% -20.0%

41 OB 36 35 47 42 -11 -7 -23.4% -16.7%

101 IB 88 87 95 98 -7 -11 -7.4% -11.2%

101 OB 90 89 107 100 -17 -11 -15.9% -11.0%

302 IB 58 58 46 49 12 9 26.1% 18.4%

302 OB 59 59 48 47 11 12 22.9% 25.5%

6 IB: In-Bound Route, OB: Out-Bound Route


Figure 6: Modeled vs. Scheduled Transit Route Travel Times

6.5 Truck Volumes

Table 26 and Figure 7 compare observed truck counts on selected roadways to SANDAG model estimates. The

truck model is a new component in the travel demand model and deals with the generation, distribution,

and assignment to roadways for heavy-duty trucks. The observed truck volumes are derived from Automatic

Vehicle Classification (AVC) and Weigh-in-Motion (WIM) detection locations throughout the region. Figure 7

shows a correlation between modeled and observed counts. The model is over estimating traffic along I-5

and underestimating I-8 volumes between San Diego and Imperial County. The largest percent differential

may be a count issue where the observed truck traffic is unbalanced on SR 163 between northbound and

southbound directions by more than 1,000 trucks.


R-Squared: .8638


Trend line


Table 26: Comparing Modeled to Observed Truck Volumes, by Highway

Highway Cross Street Observed

Trucks

Modeled

Trucks Difference %Diff

I-5 NB Leucadia 5,855 6,351 496 8.5%

I-5 SB Leucadia 5,915 7,797 1,883 31.8%

I-8 EB Kitchen Creek

Rd 1,211 854 -357 -29.5%

I-8 WB Kitchen Creek

Rd 1,176 854 -323 -27.4%

I-15 NB Clairemont

Mesa Blvd 5,253 5,128 -125 -2.4%

I-15 SB Clairemont

Mesa Blvd 6,496 5,753 -743 -11.4%

SR 78 EB S Rancho

Santa Fe Rd 3,808 3,082 -726 -19.1%

SR 78 WB S Rancho

Santa Fe Rd 2,844 3,003 159 5.6%

SR 163 NB Kearney Villa

Rd 2,159 3,486 1,327 61.5%

SR 163 SB Kearney Villa

Rd 3,233 3,535 303 9.4%

I-805 NB Naples 4,291 3,527 -764 -17.8%

I-805 SB Naples 5,625 4,058 -1,567 -27.9%

I-805 NB Governor 6,611 6,479 -132 -2.0%

I-805 SB Governor 7,126 7,330 204 2.9%


Figure 7: Comparing Modeled to Observed Truck Volumes, by Highway

6.6 Roadway Volumes

Table 27, Figure 8, and Figure 9 compare observed roadway volumes with the SANDAG model estimates.

Transportation models typically are calibrated from the highest functional classification (Freeway) to the

lowest functional classification. Often on local streets the count location as compared to where driveway

load points can affect the comparison to the modeled roadway. This level of calibration is often worked on

during focused studies for local jurisdictions. Correspondingly, the higher the roadway volume the better the

estimated fit. Ideally in Figure 8 and Figure 9 you would want a tight conglomeration of points around the

45-degree line. The two figures show that the higher volume freeways are each being overestimated and

there is a slight skew overall in the model toward overestimation.


R-Squared: .8373


Trend line


Table 27: Comparing Modeled to Observed Roadway Volumes, by Functional Classification

Functional Classification # Links Abs. Mean % Diff R-Squared

Freeway 252 15.2 0.86

Prime arterial 272 25.5 0.62

Major arterial 1,144 31.6 0.59

Collector 724 49.8 0.34

Local Collector 888 49.9 0.34

Local road (non Circulation Element) 269 64.3 0.14

All 3,551 40.8 0.91

Table 28: Comparing Modeled to Observed Roadway Volumes, by Volume Range

Volume Range # Links Abs. Mean % Diff R-Squared

< 10,000 1,560 58.5 0.22

10,000 to 50,000 1,730 28.7 0.62

50,001 to 100,000 235 15.7 0.72

>100,000 26 16.6 0.75

All 3,551 40.8 0.91


Figure 8: Scatter Plot of Mode led and Observed Roadway Volumes (all roadways)

Absolute Mean Percentage Difference: 40.84224

R-Squared: 0.9077

N = 3551


Trend line


Figure 9: Scatter Plot of Modeled and Observed Freeway Volumes

Absolute Mean Percentage Difference: 15.17589

R-Squared: 0.8587

N = 252


Trend line


7. WORKS CITED

Caltrans. (2003). 2000-2001 California Statewide Travel Survey: Weekday Travel Report. Sacramento: Caltrans.

Caltrans. (2008). 2008 California Public Road Data (HPMS). Sacramento.

Caltrans. (2010). Retrieved from Caltran Performance Measurement System (PeMS): http://pems.dot.ca.gov

National Bureau of Economic Research, Inc. (2010, September 20). US Business Cycle Expansions and

Contractions. Retrieved May 24, 2011, from http://www.nber.org/cycles.html

Pierce, B., Casas, J., & Giaimo. (2003). Estimated Trip Rate Under-Reporting: Preliminary Results from the Ohio

Household Travel Survey. Transportation Research Board, (p. 2). Washington, DC.

Pratt, R. H., & Park, G. (2000). TCRP Project B-12: Traveler Response to Transportation System Changes.

Washington, D.C.: Transportation Research Board, Transit Cooperative Research Program.

San Diego Association of Governements (SANDAG). (2008). 2006 San Diego Household Travel Study. San

Diego.

San Diego Association of Governemetns (SANDAG). (2011). Results of the 2009 Onboard Transit Passenger

Survey for the San Diego Region. San Diego.

U.S. Census Bureau. (n.d.). Retrieved from American Community Survey: http://www.census.gov/acs/www/

U.S. Department of Transportation, Federal Highway Administration. (n.d.). Retrieved from 2009 National

Household Travel Survey: http://nhts.ornl.gov


8. APPENDIX A: TRIP LENGTH DISTRIBUTIONS: MODEL (2008) VS. SURVEY (2006)

Trip length distributions (modeled versus observed) by trip purpose are shown below. These graphs compare

the proportion of trips that fall in each one mile distance increment between the model and the 2006 SDHTS.

Overall, the modeled distribution is quite similar to the distribution observed in the survey. The coincidence

ratio that accompanies each graph is a measure of the proportion of the area under the curves that overlaps.

A coincidence ratio of 1 would indicate that two distributions are identical. Each chart is labeled with the trip

purpose(s) that are being considered.

All Purposes

Coincidence ratio: 0.90

Home-Work



Home-College


Home-School


Home-Other



Non-Home-Based


Serve Passenger



9. APPENDIX B: GLOSSARY

ACS American Community Survey

Caltrans California Department of Transportation

CHTS California Household Travel Survey

CTPP Census Transportation Planning Package

HPMS Highway Performance Monitoring System (Caltrans)

FHWA Federal Highway Administration

FTA Federal Transit Administration

NHTS National Household Travel Survey

OBS Onboard Transit Passenger Survey

PeMS Performance Measuring System

PCP Passenger Counting Program

RTP Regional Transportation Plan

TAZ Transportation Analysis Zone

SANDAG San Diego Association of Governments

SDHTS San Diego Household Travel Survey

SANDAG TRANSPORTATION MODEL SENSITIVITY

ANALYSIS AND REPORT

July 2011

(REVISED – JULY 29, 2011)

SANDAG Transportation Model Sensitivity Analysis and Report 1

TABLE OF CONTENTS

1. EXECUTIVE SUMMARY 3

2. INTRODUCTION 4

3. THE SCENARIOS 5

3.1 BASELINE 5

3.2 AUTO OPERATING COST 5

OVERVIEW ................................................................................................................................................. 5

FINDINGS ................................................................................................................................................... 5

AUTO OPERATING COST SUMMARY STATISTICS ................................................................................................ 6

TRAVEL BY SPEED BIN .................................................................................................................................. 6

AUTO OPERATING COSTS VMT BY FUNCTIONAL CLASS ...................................................................................... 7

AUTO OPERATING COSTS MODE SHARE .......................................................................................................... 8

3.3 PARKING COST 9

OVERVIEW ................................................................................................................................................. 9

FINDINGS ................................................................................................................................................... 9

SUMMARY STATISTICS .................................................................................................................................. 9

MODE SHARE ........................................................................................................................................... 10

SPEEDS BY SELECTED FACILITY TYPES ............................................................................................................ 11

3.4 INCOME DISTRIBUTION 11

OVERVIEW ............................................................................................................................................... 11

FINDINGS ................................................................................................................................................. 12

SUMMARY STATISTICS ................................................................................................................................ 12

3.5 TRANSIT FARES 13

OVERVIEW ............................................................................................................................................... 13

FINDINGS ................................................................................................................................................. 13


MODE SHARE ........................................................................................................................................... 14

3.6 TRANSIT FREQUENCY 15

OVERVIEW ............................................................................................................................................... 15

FINDINGS ................................................................................................................................................. 15


MODE SHARE ........................................................................................................................................... 16

ROUTE RIDERSHIP...................................................................................................................................... 17

3.7 TRANSIT ACCESS – WAIT AND TRANSFER TIME 18

OVERVIEW ............................................................................................................................................... 18

FINDINGS ................................................................................................................................................. 18


MODE SHARE ........................................................................................................................................... 19

3.8 TRANSIT ACCESS – WALK FACTORS 20

OVERVIEW ............................................................................................................................................... 20


FINDINGS ................................................................................................................................................. 20


MODE SHARE ........................................................................................................................................... 21

3.9 NETWORK ASSIGNMENT SENSITIVITY 22

OVERVIEW ............................................................................................................................................... 22

FINDINGS ................................................................................................................................................. 22


VMT BY ROAD CLASSIFICATION ................................................................................................................... 23

HIGHWAY TRAVEL BY SPEED ........................................................................................................................ 24

AVERAGE TRIP LENGTHS ............................................................................................................................. 25

CORRIDOR TRAVEL TIMES (IN MINUTES) BY MODE ......................................................................................... 26

3.10 CAPACITY SCENARIOS 29

OVERVIEW ............................................................................................................................................... 29

FINDINGS ................................................................................................................................................. 29


VMT BY FUNCTIONAL CLASS: ...................................................................................................................... 30

TRAVEL BY SPEED BIN: ............................................................................................................................... 30

TRAVEL IN CONGESTION ............................................................................................................................. 31

MODE SHARE ........................................................................................................................................... 32

3.11 TRIP GENERATION DISCOUNTS 33

OVERVIEW ............................................................................................................................................... 33

FINDINGS ................................................................................................................................................. 33


VMT BY FUNCTIONAL CLASS: ...................................................................................................................... 34

MODE SHARE ........................................................................................................................................... 35

4. GLOSSARY OF ACRONYMS AND TERMS 36


1. EXECUTIVE SUMMARY

Sensitivity tests evaluate the responsiveness of models to systematic changes in input values. The

measure referred to as “elasticity” is used to describe the sensitivity of model results to changes in

model inputs. Elasticity is defined as the ratio of the percentage change in an output to the

percentage change in a model input. Sensitivity testing provides confidence in the use of estimated,

calibrated, and validated travel demand models.

The objective of this report is to provide insight into the sensitivity and elasticity of the

San Diego Association of Governments (SANDAG) four-step travel demand model (or, more simply,

the model) in terms of policy and operational adjustment. SANDAG has one of the most advanced

travel models in the country, and its model is in a state of constant evolution with new data,

procedures, and functions regularly added to the process. The traditional four steps of

transportation modeling include:1

• Trip Generation;

• Trip Distribution;

• Mode Choice; and

• Trip Assignment.

SANDAG is currently developing the 2050 Regional Transportation Plan (RTP), which is expected to

be adopted in October 2011. All of the analysis conducted in this report is done in a model year

2035 projection. This aligns with the horizon year of Senate Bill 375 (Steinberg, 2008) and the

Draft 2050 RTP and its Sustainable Communities Strategy (SCS). More than 40 specific model

scenarios were run and tested for this report. Some of the measures tested in this report are policy

scenarios under consideration in the Draft 2050 RTP while others were not considered but are

potentially important policy variables in the future.

The report indicates the SANDAG model is sensitive to all of the input and parameter changes

analyzed. Transit ridership and vehicle miles traveled are most sensitive to changes in the cost of

travel whether in the form of changing fuel prices, transit fares, or disposable income. The analysis

indicates that while network improvements do affect travel times and mode choice, the impacts are

far less important than economic variables.

1 More information on the SANDAG four-step travel demand model can be found on the SANDAG website (http://www.sandag.org/models).


2. INTRODUCTION

Ten major groups of model parameters are covered in this report. Between one and four actual

model runs were performed for each group to provide the context for the sensitivity analysis.

The purpose of testing several alternatives within each group is to evaluate the range and

sensitivity of each isolated variable. Evaluating only one variable at a time provides sensitivity tests

that are better able to determine the model output changes (elasticity) attributed to that variable.

Each scenario in this sensitivity analysis is compared to the baseline and to the other scenarios

within its group. The baseline scenario is similar to the 2035 Revenue Constrained scenario in the

Draft 2050 RTP. The baseline scenario includes some network changes that will be considered in the

Final 2050 RTP, and it does not include Transportation Demand Management/Transportation System

Management (TDM/TSM) post-processing analysis. A complete list of performance metrics associated

with the baseline scenario is included in Appendix A.

The major sensitivity test groupings include:

� Baseline

� Auto Operating Costs

� Parking Costs

� Income Distribution

� Transit Fares

� Transit Frequency

� Transit Access – Wait Times and Transfers

� Transit Access – Walk Factors

� Network Assignment

� Roadway Capacity

� Trip Generation Discounts

The Draft 2050 RTP Performance Measures are used in most cases to evaluate the model sensitivity

to variables listed above. All of the performance measures for each alternative are presented in

Appendix A. These metrics are intended to provide insight into the model behavior at the

aggregate and disaggregate level. The corridors used for the “Corridor Travel Time” metric were

defined in the Draft 2050 RTP.

Mode choice metrics include analysis for the entire San Diego region as well as the urban areas. The

mode choice analysis focused on 24-hour mode shares for all trip purposes and peak period mode

shares for Home-Work (i.e., commute) trips.


3. THE SCENARIOS

3.1 Baseline

Each scenario in this sensitivity analysis is compared to the baseline and to the other scenarios

within its group. The baseline scenario is similar to the 2035 Revenue Constrained scenario in the

Draft 2050 RTP. The baseline scenario includes some network changes that will be considered in the

Final 2050 RTP, and it does not include TDM/TSM post-processing analysis. A complete list of

performance metrics associated with the baseline scenario is included in

Appendix A. The differences between the baseline scenario used in the Draft 2050 RTP and baseline

used in this report are not significant for purposes of the sensitivity analysis because model

sensitivity tests a variable’s impact in isolation. It is assumed that any shift of an input variable

regardless of the baseline, within reason, would have similar a similar effect on the model outputs.

3.2 Auto Operating Cost

Overview

The scenarios (50, 75, 125, and 150 percent of baseline) focus on testing sensitivity to “auto

operating costs” in the model which is the fuel component of costs associated with operating a

motorized vehicle.

The expectation is that lower auto operating costs will cause an increase in trip lengths, a decrease

in transit share, and a commensurate increase in auto trips. In addition, lower auto operating costs

should result in greater Vehicle Miles Traveled (VMT), Vehicle Hours Traveled (VHT), and more

congested facilities. Note that auto operating cost also affects the drive-to-transit modes, so transit

results may not always be intuitively obvious though the increase in use of drive-transit would be

offset by a shift to auto modes.

Findings

Based on the tables below, the model is sensitive to auto operating cost changes with significant

shifts in output metrics.

The auto operating cost scenarios show changes in VMT and mode shares over all four scenarios.

The VMT differences are directionally and symmetrically intuitive when compared to the baseline as

well as to each other. The 50 percent scenario shows an increase in regional VMT of 12.2 percent,

and the 150 percent scenario shows a decrease in regional VMT of 9.9 percent. Correspondingly, the

75 percent scenario increases regional VMT by 5.8 percent, and the 125 percent scenario shows a

decrease in region-wide VMT of 5.3 percent.

The differences in highway travel speed by speed bin follow the trend with fewer vehicles driving at

55 miles per hour (mph) or greater when the auto operating costs are lower and more vehicles

driving at 55 mph or greater when the auto operating costs are higher as a result of reduced

congestion. The model is not directly sensitive to fuel efficiency, so individual driving habits (e.g.,

slower driving to conserve fuel) in response to changing energy prices are not captured. The


average speed by functional classification follows the same trend with lower average speeds when

the auto operating costs are lower as compared to higher speeds with the higher auto operating

costs. All of the scenarios show significant changes to the travel in congestion metric with the same

trend as noted above.

The changes in mode share are dramatic for all scenarios with single occupant vehicle (SOV) and

high occupancy vehicle (HOV) trips competing with transit and walk/bike trips.

Auto Operating Cost Summary Statistics

50% of Baseline Auto

Operating Costs


Operating Costs

Baseline


Operating Costs


Operating Cost

Total VMT 116,741,761 110,139,169 104,081,786 98,603,783 93,825,178

Total VHT 3,539,185 3,254,242 3,025,199 2,834,694 2,684,386

Commute Transit Share (Peak)

7.300% 7.947% 8.629% 9.307% 9.935%

All Trips Transit Share (Daily)

1.384% 1.517% 1.665% 1.820% 1.970%

Travel by Speed Bin

50% of Baseline Auto Operating

Costs


Costs

Baseline


Costs


Costs

Speeds between 0 and 35 mph

11.5% 8.2% 5.6% 3.7% 2.6%


13.8% 11.2% 9.1% 7.1% 5.6%

Speeds over 55 mph

74.7% 80.6% 85.3% 89.1% 91.7%


Auto Operating Costs VMT by Functional Class

CLASS

50% of Baseline Auto Operating Costs

75% of Baseline Auto Operating Costs

Baseline 125% of Baseline Auto

Operating Costs 150% of Baseline Auto

Operating Costs

VMT VMT DIF DIF PCT VMT VMT DIF DIF PCT VMT VMT VMT DIF DIF PCT VMT VMT DIF DIF PCT

FREEWAY 58,582,889 7,222,349 14.1% 54,871,488 3,510,948 6.8% 51,360,540 48,139,810 -3,220,730 -6.3% 45,261,505 -6,099,035 -11.9%

PRIME 9,601,514 940,162 10.9% 9,100,486 439,134 5.1% 8,661,352 8,288,087 -373,265 -4.3% 7,963,155 -698,197 -8.1%

MAJOR 20,800,412 2,116,827 11.3% 19,674,242 990,657 5.3% 18,683,585 17,803,321 -880,264 -4.7% 17,073,489 -1,610,096 -8.6%

COLLECTOR 7,390,716 780,770 11.8% 6,971,714 361,768 5.5% 6,609,946 6,280,986 -328,960 -5.0% 6,018,239 -591,707 -9.0%

LOCAL COLLECTOR 6,662,884 654,408 10.9% 6,310,510 302,034 5.0% 6,008,476 5,754,071 -254,405 -4.2% 5,552,656 -455,820 -7.6%

RURAL COLLECTOR 543,860 85,792 18.7% 493,548 35,480 7.7% 458,068 429,695 -28,373 -6.2% 412,595 -45,473 -9.9%

LOCAL 1,928,921 166,434 9.4% 1,838,188 75,701 4.3% 1,762,487 1,694,459 -68,028 -3.9% 1,632,855 -129,632 -7.4%

FWY-FWY RAMP 2,390,634 277,822 13.1% 2,250,042 137,230 6.5% 2,112,812 1,985,170 -127,642 -6.0% 1,865,781 -247,031 -11.7%

LOCAL RAMP 3,191,017 310,180 10.8% 3,034,221 153,384 5.3% 2,880,837 2,735,419 -145,418 -5.0% 2,598,956 -281,881 -9.8%

ACCESS 5,648,914 105,231 1.9% 5,594,730 51,047 0.9% 5,543,683 5,492,765 -50,918 -0.9% 5,445,946 -97,737 -1.8%

TOTAL 116,741,761 12,659,975 12.2% 110,139,169 6,057,383 5.8% 104,081,786 98,603,783 -5,478,003 -5.3% 93,825,177 -10,256,609 -9.9%


Auto Operating Costs Mode Share

50% of Baseline Auto Operating Cost

75% of Baseline Auto Operating Cost

Baseline 125% of Baseline Auto

Operating Cost 150% of Baseline Auto

Operating Cost

Trips Percentage Trips Percentage Trips Percentage Trips Percentage Trips Percentage

REGION-WIDE Peak Period COMMUTE Trips

SOV 1,012,534 79.3% 1,005,646 78.8% 997,905 78.3% 989,638 77.8% 980,119 77.2%

HOV 145,741 11.4% 141,883 11.1% 137,697 10.8% 133,479 10.5% 130,770 10.3%

Transit 93,187 7.3% 101,363 7.9% 109,911 8.6% 118,340 9.3% 126,078 9.9%

School Bus 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 25,139 2.0% 26,539 2.1% 28,175 2.2% 30,043 2.4% 32,033 2.5%

TOTAL 1,276,601 100.0% 1,275,431 100.0% 1,273,688 100.0% 1,271,500 100.0% 1,269,000 100.0%

REGION-WIDE

Daily ALL Trips

SOV 11,122,236 52.9% 11,098,261 52.8% 11,069,865 52.7% 11,038,014 52.5% 11,005,451 52.4%

HOV 8,970,488 42.7% 8,936,939 42.5% 8,905,753 42.4% 8,878,203 42.2% 8,853,660 42.1%

Transit 290,963 1.4% 318,828 1.5% 350,025 1.7% 382,531 1.8% 414,027 2.0%

School Bus 140,797 0.7% 134,020 0.6% 128,788 0.6% 124,724 0.6% 121,482 0.6%

Walk & Bike 494,902 2.4% 531,338 2.5% 564,954 2.7% 595,912 2.8% 624,765 3.0%

TOTAL 21,019,386 100.0% 21,019,386 100.0% 21,019,385 100.0% 21,019,384 100.0% 21,019,385 100.0%

URBAN-AREA Peak Period COMMUTE Trips

SOV 834,260 78.2% 827,773 77.7% 822,033 77.2% 813,126 76.5% 804,455 75.9%

HOV 120,212 11.3% 117,037 11.0% 114,066 10.7% 109,751 10.3% 107,430 10.1%

Transit 88,624 8.3% 96,178 9.0% 103,223 9.7% 111,842 10.5% 119,001 11.2%

School Bus 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 23,234 2.2% 24,472 2.3% 24,961 2.3% 27,597 2.6% 29,372 2.8%

TOTAL 1,066,330 100.0% 1,065,460 100.0% 1,064,283 100.0% 1,062,316 100.0% 1,060,258 100.0%

URBAN-AREA

Daily ALL Trips

SOV 9,174,588 53.0% 9,151,331 52.9% 9,134,787 52.8% 9,094,646 52.5% 9,064,405 52.4%

HOV 7,313,432 42.2% 7,283,881 42.1% 7,263,745 42.0% 7,231,788 41.8% 7,210,076 41.7%

Transit 277,858 1.6% 303,837 1.8% 328,328 1.9% 362,561 2.1% 391,414 2.3%

School Bus 102,929 0.6% 97,462 0.6% 93,252 0.5% 89,930 0.5% 87,284 0.5%

Walk & Bike 441,827 2.6% 474,127 2.7% 490,439 2.8% 531,513 3.1% 557,158 3.2%

TOTAL 17,310,634 100.0% 17,310,638 100.0% 17,310,551 100.0% 17,310,438 100.0% 17,310,337 100.0%

SANDAG Transportation Model Sensitivity Analysis and Report (Revised – July 29, 2011) 9

3.3 Parking Cost

Overview

These scenarios (50, 75, 125, and 150 percent of baseline parking costs) focus on the model response to

changes in parking cost. Parking costs are assumed for Centre City San Diego, Lindbergh Field, some

universities, and business districts in Escondido, Oceanside, La Jolla, La Mesa, and Hillcrest. These include

privately owned parking lots as well as on-street parking spaces with meters.

The expectation is that transit mode share will increase to areas with high parking costs although it will

be impacted by trips redistributing to areas with a lower travel cost. VMT and VHT will show modest

changes.

Findings

The model is sensitive to parking costs. The shift in mode share is similar with smaller magnitudes to the

shifts in the previous section related to auto operating costs. The smaller magnitude is attributed to the

limited number of trips that have any parking costs applied due to the limited geographic scope of paid

parking areas. As costs increase, auto mode shares decrease. Speeds are not affected regionally due to the

limited geographic scope of parking zones across the region.

Summary Statistics

50% of Baseline

Parking Costs

75% of Baseline

Parking Costs Baseline

125% of Baseline

Parking Costs

150% of Baseline

Parking Cost

Total VMT 104,086,133 104,081,529 104,081,786 104,071,237 104,057,815

Total VHT 3,027,966 3,027,033 3,025,199 3,023,314 3,020,277


8.393% 8.504% 8.629% 8.718% 8.816%


1.591% 1.631% 1.665% 1.694% 1.721%

SANDAG Transportation Model Sensitivity Analysis and Report (Revised – July 29, 2011) 10

Mode Share

50% of Baseline Parking Cost


Baseline 125% of Baseline Parking Cost




SOV 1,002,852 78.7% 999,882 78.5% 997,905 78.3% 995,674 78.2% 993,774 78.0%

HOV 138,087 10.8% 138,509 10.9% 137,697 10.8% 137,777 10.8% 137,327 10.8%

Transit 106,925 8.4% 108,334 8.5% 109,911 8.6% 111,029 8.7% 112,270 8.8%

School Bus 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 26,176 2.1% 27,173 2.1% 28,175 2.2% 29,128 2.3% 30,078 2.4%

TOTAL 1,274,040 100.0% 1,273,898 100.0% 1,273,688 100.0% 1,273,608 100.0% 1,273,449 100.0%

REGION-WIDE

Daily ALL Trips

SOV 11,097,377 52.8% 11,082,390 52.7% 11,069,865 52.7% 11,058,675 52.6% 11,047,981 52.6%

HOV 8,923,350 42.5% 8,914,419 42.4% 8,905,753 42.4% 8,899,273 42.3% 8,892,921 42.3%

Transit 334,348 1.6% 342,840 1.6% 350,025 1.7% 356,135 1.7% 361,824 1.7%

School Bus 128,821 0.6% 128,854 0.6% 128,788 0.6% 128,904 0.6% 128,915 0.6%

Walk & Bike 535,490 2.5% 550,883 2.6% 564,954 2.7% 576,399 2.7% 587,745 2.8%

TOTAL 21,019,386 100.0% 21,019,386 100.0% 21,019,385 100.0% 21,019,386 100.0% 21,019,386 100.0%


SOV 825,623 77.6% 822,684 77.3% 822,033 77.2% 818,440 76.9% 816,511 76.8%

HOV 113,817 10.7% 114,205 10.7% 114,066 10.7% 113,477 10.7% 113,051 10.6%

Transit 101,050 9.5% 102,468 9.6% 103,223 9.7% 105,187 9.9% 106,439 10.0%

School Bus 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 23,929 2.2% 24,926 2.3% 24,961 2.3% 26,882 2.5% 27,833 2.6%

TOTAL 1,064,419 100.0% 1,064,283 100.0% 1,064,283 100.0% 1,063,986 100.0% 1,063,834 100.0%

URBAN-AREA

Daily ALL Trips

SOV 9,151,795 52.9% 9,136,890 52.8% 9,134,787 52.8% 9,113,164 52.6% 9,102,459 52.6%

HOV 7,273,917 42.0% 7,264,947 42.0% 7,263,745 42.0% 7,249,822 41.9% 7,243,468 41.8%

Transit 317,004 1.8% 325,493 1.9% 328,328 1.9% 338,786 2.0% 344,478 2.0%

School Bus 93,219 0.5% 93,252 0.5% 93,252 0.5% 93,304 0.5% 93,318 0.5%

Walk & Bike 474,581 2.7% 489,967 2.8% 490,439 2.8% 515,482 3.0% 526,822 3.0%

TOTAL 17,310,516 100.0% 17,310,549 100.0% 17,310,551 100.0% 17,310,558 100.0% 17,310,545 100.0%


Speeds by Selected Facility Types

50% of Baseline

Parking Cost

75% of Baseline

Parking Cost Baseline

125% of Baseline

Parking Cost

150% of Baseline

Parking Cost

Highway 56.8 56.8 56.8 56.9 56.9

Prime Arterial 28.8 28.8 28.8 28.8 28.8

Regional Average

34.4 34.4 34.4 34.4 34.5

3.4 Income Distribution

Overview

The income distribution scenarios focus on altering the income range breakpoints. Income is used to

set parameters on how people react to (their sensitivity to) cost factors in the model such as auto

operating, tolls, parking, and transit fares. The regional growth forecast produces data for a total of

ten income groups. The transportation model condenses those ten groups into high, middle, and

low income groups. The baseline scenario uses the regional growth forecast and falls between the

middle income and high income scenario.

Household Income Distribution by Scenario

Scenario Households $0-$29,999 $30,000-$59,999 $60,000 or more

Baseline 1,357,100 233,400 349,200 774,500

Very Low 1,357,100 947,400 143,100 266,600

Low 1,357,100 582,600 507,900 266,600

Middle 1,357,100 87,200 1,003,300 266,600

High 1,357,100 87,200 146,200 1,123,700

Incomes Basis 1999 Dollars

Source: SANDAG, 2050 Regional Growth Forecast (data extracted on: 06/2011)

The model results should be similar to the auto operating costs examined in Section 3.2. In the

scenarios where average income is higher, driving should become more dominant due to larger

disposable incomes. In scenarios where incomes are lower, transit and non-motorized modes should

gain more mode share.2 Increases in people choosing transit and non-motorized versus auto will

cause VMT and VHT to decrease.

2 Pratt, R. H., & Park, G. (2000). TCRP Project B-12: Traveler Response to Transportation System Changes. Washington, D.C.: Transportation Research Board, Transit Cooperative Research Program.


Findings

The model is sensitive to changes in income classification. The income distribution scenarios show

significant changes in VMT and mode shares over all four scenarios. Transit mode shares more than

double in the extreme low-income scale scenario and drop nearly 50 percent in the high-income

scale scenario.

The percent change in travel by speed and the travel speeds by facility type mimic the trend with

the VMT noted above. Travel in congestion is virtually the same between the two low-income

scenarios and the increases in the high-income scenarios fall in line with the overall trend.

The corridor travel times by mode changes are intuitive by mode and by weighted average with the

low-income scale scenarios showing increases in travel time despite the trend. This can be attributed

to the large increases in transit mode share in the low-income scale scenarios.

The mode share changes also make sense when looking at all modes in combination. The transit

and non-motorized shares increase significantly in the low-income scale scenarios with the majority

of the shift coming from the SOV mode. The shift in HOV mode is minimal.

Summary Statistics

Very Low Income Scale

Low Income Scale

Middle Income Scale

Baseline High Income

Scale

Total VMT 77,491,872 87,513,631 99,544,227 104,081,786 112,030,063

Total VHT 2,187,871 2,483,082 2,874,260 3,025,199 3,343,041


19.574% 15.312% 10.970% 8.629% 5.539%


4.137% 3.079% 1.881% 1.665% 1.007%


3.5 Transit Fares

Overview

These scenarios (50, 75, 125, and 150 percent of baseline transit fares) focus on changing the

standard transit fares. These scenarios measure elasticity of transit demand with respect to fare

changes and cross-elasticity of non-transit modes with respect to transit fares. Because the COASTER

uses a different fare structure, the fares for the COASTER were not changed in any of these

scenarios due to the complexities of defining a new zone-fare structure in the model.

The model should increase transit mode shares as a result of lower fares resulting in decreases in

VMT and VHT.

Findings

The model is sensitive to transit fare changes as reflected by changes in VMT and transit mode share

in the summary tables below.

Changes in VMT by facility type change directionally as expected across all scenarios. VMT decreases

for all facility types when the transit fares drop, and the VMT increases for all facility types when

transit fares increase.

Mode share changes between the scenarios are nearly symmetrical. The transit share increases with

lower fares and the transit share decreases with higher fares. These changes are offset by smaller

shifts in all of the other travel modes.

Summary Statistics

50% of Baseline

Transit Fares

75% of Baseline

Transit Fares Baseline

125% of Baseline

Transit Fares

150% of Baseline

Transit Fares

Total VMT 103,524,625 103,816,144 104,081,786 104,294,979 104,493,991

Total VHT 2,999,030 3,011,396 3,025,199 3,034,174 3,043,296


10.130% 9.336% 8.629% 8.059% 7.518%


2.905% 1.864% 1.665% 1.509% 1.398%


Mode Share

50% of Baseline Transit Fares


Baseline 125% of Baseline Transit Fares



REGION-WIDE

Peak COMMUTE Trips

SOV 983,446 77.1% 990,829 77.7% 997,905 78.3% 1,003,708 78.7% 1,008,355 79.1%

HOV 134,763 10.6% 136,665 10.8% 137,697 10.8% 138,448 11.0% 140,311 11.0%

Transit 129,032 10.2% 118,917 9.3% 109,911 8.6% 102,647 8.0% 95,756 7.5%

School Bus 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 26,482 2.1% 27,347 2.1% 28,175 2.2% 28,951 2.3% 29,327 2.3%

TOTAL 1,273,723 100.0% 1,273,758 100.0% 1,273,688 100.0% 1,273,754 100.0% 1,273,749 100.0%

REGION-WIDE

Daily ALL Trips

SOV 11,019,392 52.4% 11,046,403 52.5% 11,069,865 52.7% 11,088,403 52.7% 11,102,251 52.8%

HOV 8,877,437 42.2% 8,893,217 42.3% 8,905,753 42.4% 8,915,251 42.4% 8,922,954 42.5%

Transit 440,545 2.1% 391,713 1.9% 350,025 1.7% 317,257 1.5% 293,832 1.4%

School Bus 128,811 0.6% 128,818 0.6% 128,788 0.6% 128,814 0.6% 128,812 0.6%

Walk & Bike 553,200 2.6% 559,235 2.7% 564,954 2.7% 569,660 2.7% 571,537 2.7%

TOTAL 21,019,385 100.0% 21,019,386 100.0% 21,019,385 100.0% 21,019,385 100.0% 21,019,386 100.0%

URBAN-AREA

Peak COMMUTE Trips

SOV 807,234 75.7% 814,112 76.5% 822,033 77.2% 826,073 77.5% 830,422 78.0%

HOV 110,707 10.5% 112,465 10.6% 114,066 10.7% 114,110 10.8% 115,847 10.9%

Transit 121,893 11.5% 112,450 10.6% 103,223 9.7% 97,275 9.1% 90,824 8.5%

School Bus 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 24,288 2.3% 25,126 2.4% 24,961 2.3% 26,685 2.5% 27,047 2.5%

TOTAL 1,064,122 100.0% 1,064,153 100.0% 1,064,283 100.0% 1,064,143 100.0% 1,064,140 100.0%

URBAN-AREA

Daily ALL Trips

SOV 9,077,033 52.4% 9,102,377 52.6% 9,134,787 52.8% 9,141,733 52.8% 9,154,789 52.9%

HOV 7,229,692 41.8% 7,244,492 41.9% 7,263,745 42.0% 7,265,212 42.0% 7,272,423 42.0%

Transit 418,105 2.4% 372,041 2.1% 328,328 1.9% 301,719 1.7% 279,607 1.6%

School Bus 93,251 0.5% 93,256 0.5% 93,252 0.5% 93,253 0.5% 93,251 0.5%

Walk & Bike 492,461 2.8% 498,380 2.9% 490,439 2.8% 508,629 2.9% 510,467 2.9%

TOTAL 17,310,542 100.0% 17,310,546 100.0% 17,310,551 100.0% 17,310,546 100.0% 17,310,537 100.0%


3.6 Transit Frequency

Overview

These scenarios focus on transit frequency for the COASTER (North County Transit District

Route 398) and Metropolitan Transit System (MTS) Route 7. The COASTER and Route 7 were

chosen due to their regional significance. The COASTER is the only commuter rail service in

the region, and Route 7 is one of the busiest local bus routes in the region. The changes

include increasing and decreasing the frequency of these routes by 50 percent.

These scenarios should result in changes in ridership on the route tested in the direction of

service frequency. An increase in frequency should result in an increase in ridership.

Changes in VMT should be limited to the scale of VMT on competing facilities.

Findings

The model is sensitive to changes in transit route frequency. The overall route ridership

change is consistent with the direction of the frequency change. COASTER frequency

changes result in larger transit mode share changes for peak period commute trips than

Route 7.

Changes in route frequency have a direct impact on route ridership, but isolated frequency

changes do not significantly impact overall transit ridership.

Summary Statistics

Decrease COASTER Frequency

Increase COASTER Frequency

Baseline Decrease Route 7

Frequency

Increase Route 7

Frequency

Total VMT 104,111,608 104,068,051 104,081,786 104,076,808 104,078,393

Total VHT 3,026,059 3,024,544 3,025,199 3,024,858 3,024,527


8.556% 8.712% 8.629% 8.626% 8.632%


1.659% 1.674% 1.665% 1.665% 1.667%


Mode Share




Frequency Increase Route 7

Frequency


REGION-WIDE

Peak COMMUTE Trips

SOV 998,873 78.4% 996,729 78.2% 997,905 78.3% 997,949 78.3% 997,845 78.3%

HOV 137,770 10.8% 137,950 10.8% 137,697 10.8% 137,786 10.8% 137,825 10.8%

Transit 108,983 8.6% 110,970 8.7% 109,911 8.6% 109,869 8.6% 109,956 8.6%

School Bus 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 28,173 2.2% 28,145 2.2% 28,175 2.2% 28,161 2.2% 28,163 2.2%

TOTAL 1,273,799 100.0% 1,273,794 100.0% 1,273,688 100.0% 1,273,765 100.0% 1,273,789 100.0%

REGION-WIDE

Daily ALL Trips

SOV 11,071,210 52.7% 11,068,269 52.7% 11,069,865 52.7% 11,069,963 52.7% 11,069,886 52.7%

HOV 8,905,981 42.4% 8,905,774 42.4% 8,905,753 42.4% 8,905,900 42.4% 8,905,891 42.4%

Transit 348,733 1.7% 351,937 1.7% 350,025 1.7% 350,058 1.7% 350,085 1.7%

School Bus 128,889 0.6% 128,890 0.6% 128,788 0.6% 128,891 0.6% 128,886 0.6%

Walk & Bike 564,573 2.7% 564,516 2.7% 564,954 2.7% 564,572 2.7% 564,639 2.7%

TOTAL 21,019,386 100.0% 21,019,386 100.0% 21,019,385 100.0% 21,019,384 100.0% 21,019,387 100.0%

URBAN-AREA

Peak COMMUTE Trips

SOV 821,603 77.2% 819,566 77.0% 822,033 77.2% 820,716 77.1% 820,635 77.1%

HOV 113,497 10.7% 113,627 10.7% 114,066 10.7% 113,506 10.7% 113,523 10.7%

Transit 103,153 9.7% 105,085 9.9% 103,223 9.7% 104,014 9.8% 104,097 9.8%

School Bus 0 0.0% 0 0.0% 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 25,926 2.4% 25,899 2.4% 24,961 2.3% 25,914 2.4% 25,917 2.4%

TOTAL 1,064,179 100.0% 1,064,177 100.0% 1,064,283 100.0% 1,064,150 100.0% 1,064,172 100.0%

URBAN-AREA

Daily ALL Trips

SOV 9,125,627 52.7% 9,122,832 52.7% 9,134,787 52.8% 9,124,445 52.7% 9,124,372 52.7%

HOV 7,256,554 41.9% 7,256,291 41.9% 7,263,745 42.0% 7,256,448 41.9% 7,256,432 41.9%

Transit 331,420 1.9% 334,531 1.9% 328,328 1.9% 332,706 1.9% 332,725 1.9%

School Bus 93,289 0.5% 93,289 0.5% 93,252 0.5% 93,289 0.5% 93,285 0.5%

Walk & Bike 503,657 2.9% 503,601 2.9% 490,439 2.8% 503,657 2.9% 503,723 2.9%

TOTAL 17,310,547 100.0% 17,310,544 100.0% 17,310,551 100.0% 17,310,545 100.0% 17,310,537 100.0%


Route Ridership

ROUTE MODE




Frequency Increase Route 7

Frequency

Peak Off Peak Daily Peak

Off Peak Daily Peak

Off Peak Daily Peak

Off Peak Daily Peak

Off Peak Daily

398 COASTER 2,380 234 2,615 8,020 2,684 10,704 5,001 1,198 6,199 5,007 1,202 6,209 5,012 1,203 6,215

7 Local Bus 2,809 3,175 5,983 2,796 3,172 5,969 2,807 3,176 5,982 1,227 1,353 2,580 4,431 6,754 11,185


3.7 Transit Access – Wait and Transfer Time

Overview

These scenarios focus on adjusting wait and transfer times up or down by 50 percent.

The model should result in lower VMT and higher transit mode shares when the wait and transfer

time is reduced, and in higher VMT and lower transit mode shares when the time is increased.

Findings

The model is sensitive to changes in transit access wait and transfer time. Regional mode shares and

VMT change as expected. When the wait and transfer times are reduced, the VMT drops and the

peak period transit commute share increases by 2 percent with all other modes reduced

proportionally. Conversely when the wait and transfer time are increased, the VMT increases and

the peak period transit commute share decreases by 1.5 percent.

Summary Statistics

Reduce Wait and Transfer Times

Baseline Increase Wait and Transfer Times

Total VMT 103,397,441 104,081,786 104,506,243

Total VHT 2,993,525 3,025,199 3,043,513


10.649% 8.629% 7.186%


1.984% 1.665% 1.433%


Mode Share

Reduce Wait and Transfer Times

Baseline Increase Wait and Transfer Times

Trips Percentage Trips Percentage Trips Percentage

REGION-WIDE

Peak Period COMMUTE Trips

SOV 975,935 76.6% 997,905 78.3% 1,012,241 79.5%

HOV 134,847 10.6% 137,697 10.8% 141,230 11.1%

Transit 135,646 10.6% 109,911 8.6% 91,536 7.2%

School Bus 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 27,369 2.1% 28,175 2.2% 28,795 2.3%

TOTAL 1,273,797 100.0% 1,273,688 100.0% 1,273,802 100.0%

REGION-WIDE

Daily ALL Trips

SOV 11,021,339 52.4% 11,069,865 52.7% 11,102,985 52.8%

HOV 8,890,344 42.3% 8,905,753 42.4% 8,919,494 42.4%

Transit 417,056 2.0% 350,025 1.7% 301,162 1.4%

School Bus 128,892 0.6% 128,788 0.6% 128,883 0.6%

Walk & Bike 561,755 2.7% 564,954 2.7% 566,862 2.7%

TOTAL 21,019,386 100.0% 21,019,385 100.0% 21,019,386 100.0%

URBAN-AREA

Peak Period COMMUTE Trips

SOV 800,454 75.2% 822,033 77.2% 833,982 78.4%

HOV 110,777 10.4% 114,066 10.7% 116,633 11.0%

Transit 127,789 12.0% 103,223 9.7% 87,045 8.2%

School Bus 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 25,156 2.4% 24,961 2.3% 26,524 2.5%

TOTAL 1,064,176 100.0% 1,064,283 100.0% 1,064,184 100.0%

URBAN-AREA

Daily ALL Trips

SOV 9,079,823 52.5% 9,134,787 52.8% 9,155,012 52.9%

HOV 7,242,211 41.8% 7,263,745 42.0% 7,268,756 42.0%

Transit 394,280 2.3% 328,328 1.9% 287,637 1.7%

School Bus 93,292 0.5% 93,252 0.5% 93,283 0.5%

Walk & Bike 500,945 2.9% 490,439 2.8% 505,864 2.9%

TOTAL 17,310,551 100.0% 17,310,551 100.0% 17,310,552 100.0%


3.8 Transit Access – Walk Factors

Overview

These scenarios focus on adjusting the urban and suburban walk factors in the mode choice model.

Walk factors are applied to the walk-to-transit mode and are different between urban and

suburban to account for urban form. The urban walk factors assume a grid street pattern while the

suburban walk factors do not. Since distance to transit is calculated as a straight line, the model

adds time surcharges to access transit to account for the street network people walk to access

transit. The time surcharge in urban areas is a factor of 1.1 while in suburban and rural areas it is a

factor of 1.3.

Since much of the region is considered suburban, the model should result in lower VMT and a

higher transit mode share with urban walk factors only, because the average time surcharge for

walk to transit will decrease. Higher VMT and a lower transit mode share with suburban walk

factors only.

Findings

The model is sensitive to changes in walk factors. In the urban walk factor scenario, transit mode

shares increase more than in the decrease in the suburban walk factor scenario transit share

decrease. In the model, a significant portion of the region uses the suburban walk factors meaning

the weighted average walk factor in the baseline is much closer to the suburban walk factor

scenario.

Summary Statistics

Urban Walk Factors Only

Baseline Suburban Walk Factors

Only

Total VMT 103,730,052 104,081,786 104,148,533

Total VHT 3,009,130 3,025,199 3,027,421

Commute Transit Share (Peak) 9.675% 8.629% 8.357%

All Trips Transit Share (Daily) 1.861% 1.665% 1.610%


Mode Share

Urban Walk Factors Only

Baseline Suburban Walk Factors

Only



SOV 986,570 77.1% 997,905 78.3% 1,000,664 78.2%

HOV 136,161 10.6% 137,697 10.8% 138,068 10.9%

Transit 123,238 10.1% 109,911 8.6% 106,443 8.7%

School Bus 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 27,773 2.2% 28,175 2.2% 28,570 2.2%

TOTAL 1,273,742 100.0% 1,273,688 100.0% 1,273,745 100.0%

REGION-WIDE

Daily ALL Trips

SOV 11,041,639 52.5% 11,069,865 52.7% 11,077,107 52.6%

HOV 8,894,518 42.3% 8,905,753 42.4% 8,908,508 42.4%

Transit 391,112 2.0% 350,025 1.7% 338,501 1.7%

School Bus 128,760 0.6% 128,788 0.6% 128,760 0.6%

Walk & Bike 563,347 2.7% 564,954 2.7% 566,508 2.7%

TOTAL 21,019,376 100.0% 21,019,385 100.0% 21,019,384 100.0%


SOV 811,040 75.8% 822,033 77.2% 823,330 77.0%

HOV 112,149 10.5% 114,066 10.7% 113,773 10.7%

Transit 115,387 11.3% 103,223 9.7% 100,725 9.9%

School Bus 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 25,566 2.4% 24,961 2.3% 26,317 2.4%

TOTAL 1,064,142 100.0% 1,064,283 100.0% 1,064,145 100.0%

URBAN-AREA

Daily ALL Trips

SOV 9,099,800 52.5% 9,134,787 52.8% 9,131,361 52.7%

HOV 7,246,437 41.8% 7,263,745 42.0% 7,258,940 41.9%

Transit 368,572 2.3% 328,328 1.9% 321,526 2.0%

School Bus 93,203 0.5% 93,252 0.5% 93,203 0.5%

Walk & Bike 502,535 2.9% 490,439 2.8% 505,519 2.9%

TOTAL 17,310,547 100.0% 17,310,551 100.0% 17,310,549 100.0%


3.9 Network Assignment Sensitivity

Overview

These scenarios focus on the model assignment algorithm and redistribution of trips due to network

changes. The networks were modified by deleting an arterial or freeway link, and trip flows were

analyzed before and after its removal.

The first scenario removed a section of El Camino Real from Marron Road to Carlsbad Village Drive

in the North County. The second scenario removed the Interstate (I)-805 overpass (Mission Valley

viaduct) over Mission Valley while leaving the affiliated interchanges to I-8. The bus routes on I-805

were allowed to continue to use Mission Valley viaduct. These two particular routes were selected

because they represent an important path for many trips and help connect the region.

The expectation in both scenarios is a redistribution of trips to closer destinations due to reduced

accessibility and a redistribution of auto trips onto the road network surrounding the deleted road.

Longer travel diversions around the deleted road should not fully offset the reduction in travel from

trip distribution changes, resulting in VMT reductions. In areas along I-805, some trips may switch

modes to competitive transit routes along the I-805 corridor when the Mission Valley viaduct is

removed. Even though I-805 is a major transportation corridor, the impacts of these changes should

be localized due to trip redistribution.

Findings

The model is sensitive to network link deletions.

Removing the I-805 Mission Valley viaduct scenario shows a reduction of more than 600,000 VMT on

the highways alone highlighting the regional significance of this highway project. The total

regional VMT decrease is more than 400,000 daily miles as some of the freeway VMT reduction is

redistributed to other facilities across Mission Valley. The remaining highway VMT reduction is a

result of shorter trips and mode shift to transit. In the I-805 viaduct scenario, only corridor 9 (Otay

Ranch to UTC) and corridor 4 (Mid-City to UTC) shows a difference in travel times when compared

to the baseline since corridor 9 uses I-805 and corridor would be affected by rerouted traffic onto I-

15 and SR 163.

The removal of El Camino Real results in little net change in VMT by facility due to rerouting onto

comparable alternative routes in the region such as Monroe Street, Carlsbad Boulevard, and College

Boulevard to the east with increased traffic on SR 78 to reach the alternative north-south routes.

While El Camino Real is a prime arterial and major thoroughfare in the North County, its removal

results in mostly localized changes. Regional speed distributions, mode shares, and trip lengths are

not affected.


Summary Statistics

Delete El Camino Real

Baseline Delete I-805 Mission Valley Viaduct

Total VMT 104,079,386 104,081,786

103,639,718

Total VHT 3,025,046 3,025,199

3,047,566


8.629% 8.629%

8.699%


1.667% 1.665%

1.674%

VMT by Road Classification

CLASS Delete El Camino Real Baseline Delete I-805 Mission Valley Viaduct

VMT VMT DIF DIF PCT VMT VMT VMT DIF DIF PCT

FREEWAY 51,382,099 21,559 0.0% 51,360,540 50,749,833 -610,707 -1.2%

PRIME 8,610,915 -50,437 -0.6% 8,661,352 8,684,262 22,910 0.3%

MAJOR 18,682,686 -899 0.0% 18,683,585 18,727,622 44,037 0.2%

COLLECTOR 6,632,934 22,988 0.3% 6,609,946 6,614,968 5,022 0.1%

LOCAL COLLECTOR 6,006,207 -2,269 0.0% 6,008,476 6,023,205 14,729 0.2%

RURAL COLLECTOR 455,700 -2,368 -0.5% 458,068 457,584 -484 0.0%

LOCAL 1,766,935 4,448 0.3% 1,762,487 1,762,918 431 0.0%

FWY-FWY RAMP 2,115,305 2,493 0.1% 2,112,812 2,212,741 99,929 4.7%

LOCAL RAMP 2,883,565 2,728 0.1% 2,880,837 2,864,243 -16,594 -0.6%

ACCESS 5,543,041 -642 0.0% 5,543,683 5,542,342 -1,341 0.0%

TOTAL 104,079,387 -2,399 0.0% 104,081,786 103,639,718 -442,068 -0.4%


Highway Travel by Speed


Baseline Delete I-805

Mission Valley Viaduct


5.7% 5.6% 6.9%


9.2% 9.1% 8.9%

Speeds over 55 mph 85.2% 85.3% 84.2%


Average Trip Lengths


Baseline Delete I-805 Mission

Valley Viaduct

Minutes Miles Minutes Miles Minutes Miles


SOV 26.04 13.16 26.04 13.15 26.25 13.13

HOV2 26.85 14.10 26.85 14.08 27.18 14.08

HOV3+ 28.36 15.15 28.39 15.16 28.59 15.13

Transit 53.72 11.18 53.72 11.17 53.73 11.25

Walk 16.51 0.72 16.51 0.72 16.52 0.72

Bike 17.94 3.52 17.94 3.52 17.91 3.52

TOTAL 28.34 12.86 28.34 12.85 28.56 12.83

REGION-WIDE

Daily ALL Trips

SOV 14.90 7.25 14.90 7.24 14.97 7.21

HOV2 13.03 6.13 13.03 6.13 13.09 6.10

HOV3+ 12.62 5.90 12.61 5.90 12.66 5.87

Transit 48.30 8.83 48.30 8.83 48.27 8.87

Walk 17.34 0.78 17.32 0.78 17.33 0.78

Bike 15.37 3.00 15.36 3.00 15.36 3.00

TOTAL 14.66 6.59 14.66 6.58 14.72 6.56


SOV 25.07 12.37 25.03 12.36 25.30 12.33

HOV2 25.80 13.30 25.79 13.27 26.15 13.27

HOV3+ 26.91 14.00 26.95 14.02 27.19 14.02

Transit 53.11 11.09 53.11 11.08 53.13 11.16

Walk 15.91 0.70 15.91 0.70 15.92 0.70

Bike 17.72 3.50 17.72 3.50 17.69 3.49

TOTAL 27.71 12.10 27.71 12.08 27.94 12.07

URBAN-AREA

Daily ALL Trips

SOV 12.96 6.44 13.81 6.44 13.88 6.41

HOV2 12.04 5.37 12.03 5.37 12.10 5.34

HOV3+ 11.66 5.16 11.65 5.15 11.70 5.13

Transit 47.45 8.68 47.45 8.67 47.42 8.71

Walk 16.62 0.76 16.61 0.76 16.62 0.76

Bike 15.14 2.97 15.13 2.97 15.12 2.97

TOTAL 13.73 5.83 13.72 5.83 13.79 5.80


Corridor Travel Times (in minutes) By Mode


Baseline Delete I-805 Mission Valley

Viaduct

4 Mid-City – UTC

Auto 29 28 31

Walk to Transit 42 42 43

Drive to Transit 44 44 45

Carpool 26 26 30

Corridor Weighted Average 31 30 34

9 Otay Ranch - UTC Auto 52 52 56

Walk to Transit 55 55 53

Drive to Transit 53 53 51

Carpool 51 51 55

Corridor Weighted Average 52 52 55


Figure 1: Localized ADT Changes around El Camino Real

99.9 Baseline ADT (Thousands)

99.9 Scenario ADT (Thousands)


Figure 3: Localized ADT changes around the I-805 / I-8 interchange (2035rc6)

99.9 Baseline ADT (Thousands)

99.9 2035rc6 ADT (Thousands)


3.10 Capacity Scenarios

Overview

The scenarios focus on network alternatives with the number of lanes doubled up to a maximum of

eight lanes per direction on highway or arterials compared to the baseline network.

The model should move traffic from one facility type to the larger facility type as capacity is made

available. This would be a regional shift that should result in lower congestion levels and faster

speeds overall. Increasing road capacity should reduce transit mode shares.

Findings

Based on the comparison tables below, the model is sensitive to capacity changes.

Both capacity scenarios show significant changes in the output metrics, especially the increased

freeway capacity scenario. The increased arterial capacity shows an increase of 109,000 VMT with

large decreases on freeways and ramps being offset by increases on local streets and roads. The

increased freeway capacity shows an increase of more than 3 million VMT with the freeways

themselves increasing by nearly 4 million VMT with remainder offset by lower volumes on local

streets and roads.

The differences in highway travel speed by speed bin are modest for the increased arterial capacity

scenario, and dramatic for the increased freeway capacity scenario with almost 98 percent of the

vehicles traveling at 55 mph or greater.

Both scenarios show changes to the travel in congestion metric with the increased freeway capacity

being most dramatic where less than 1 percent of travel occurs in congested conditions.

The travel times by mode figures improve or are the same for all 11 corridors for both scenarios (see

Appendix A). The changes in mode share are minimal for all categories with small increases in SOVs

in both scenarios.

Summary Statistics

Increased Freeway Capacity

Baseline Increased Arterial

Capacity

Total VMT 107,132,110 104,081,786 104,182,715

Total VHT 2,950,096 3,025,199 2,996,798

Commute Transit Share (Peak) 8.453% 8.629% 8.619% All Trips Transit Share (Daily) 1.658% 1.665% 1.665%


VMT by Functional Class:

CLASS Increase Freeway Capacity Baseline Increase Arterial Capacity


FREEWAY 55,294,626 3,978,869 7.7% 51,360,540 50,971,836 -388,704 -0.8%

PRIME 8,488,267 -168,111 -1.9% 8,661,352 8,760,258 98,906 1.4%

MAJOR 18,147,537 -543,604 -2.9% 18,683,585 18,936,071 252,486 1.4%

COLLECTOR 6,426,370 -186,605 -2.8% 6,609,946 6,743,739 133,793 2.0%

LOCAL COLLECTOR 5,814,835 -194,077 -3.2% 6,008,476 6,063,998 55,522 0.9% RURAL COLLECTOR 461,483 680 0.1% 458,068 466,062 7,994 1.7%

LOCAL 1,715,399 -45,215 -2.6% 1,762,487 1,752,172 -10,315 -0.6%

FWY-FWY RAMP 2,222,632 107,898 5.1% 2,112,812 2,090,541 -22,271 -1.1%

LOCAL RAMP 3,013,911 131,336 4.6% 2,880,837 2,857,776 -23,061 -0.8%

ACCESS 5,547,051 1,771 0.0% 5,543,683 5,540,263 -3,420 -0.1%

TOTAL 107,132,111 3,082,942 3.0% 104,081,786 104,182,716 100,930 0.1%

Travel by Speed Bin:

Increase Freeway Capacity Baseline Increase Arterial Capacity

Speeds between 0 and 35 mph 0.4% 5.6% 5.4%

Speeds between 35 and 55 mph 2.0% 9.1% 8.7%

Speeds over 55 mph 97.6% 85.3% 85.9%


Travel in Congestion

Increase Freeway Capacity Baseline Increase Arterial Capacity

Travel in LOS F Conditions

Peak Period ALL Trips 1% 6% 5%

Daily ALL Trips 1% 3% 3%

Peak Period FREEWAY Trips

0% 9% 8%

Daily FREEWAY Trips 0% 5% 4%


Mode Share

Increased Freeway Capacity

Baseline Increased Arterial

Capacity



SOV 1,006,064 78.8% 997,905 78.3% 998,766 78.4%

HOV 135,516 10.6% 137,697 10.8% 137,367 10.8%

Transit 107,912 8.5% 109,911 8.6% 109,804 8.6%

School Bus 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 27,141 2.1% 28,175 2.2% 28,069 2.2%

TOTAL 1,276,633 100.0% 1,273,688 100.0% 1,274,006 100.0%

REGION-WIDE

Daily ALL Trips

SOV 11,080,071 52.7% 11,069,865 52.7% 11,071,169 52.7%

HOV 8,901,728 42.4% 8,905,753 42.4% 8,905,228 42.4%

Transit 348,425 1.7% 350,025 1.7% 349,991 1.7%

School Bus 129,192 0.6% 128,788 0.6% 128,944 0.6%

Walk & Bike 559,970 2.7% 564,954 2.7% 564,053 2.7%

TOTAL 21,019,386 100.0% 21,019,385 100.0% 21,019,385 100.0%


SOV 828,100 77.6% 822,033 77.2% 821,410 77.2%

HOV 111,587 10.5% 114,066 10.7% 113,144 10.6%

Transit 102,078 9.6% 103,223 9.7% 103,949 9.8%

School Bus 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 24,948 2.3% 24,961 2.3% 25,837 2.4%

TOTAL 1,066,713 100.0% 1,064,283 100.0% 1,064,340 100.0%

URBAN-AREA ,

Daily ALL Trips

SOV 9,133,528 52.8% 9,134,787 52.8% 9,125,461 52.7%

HOV 7,253,272 41.9% 7,263,745 42.0% 7,255,874 41.9%

Transit 331,007 1.9% 328,328 1.9% 332,644 1.9%

School Bus 93,592 0.5% 93,252 0.5% 93,338 0.5%

Walk & Bike 499,282 2.9% 490,439 2.8% 503,222 2.9%

TOTAL 17,310,681 100.0% 17,310,551 100.0% 17,310,539 100.0%


3.11 Trip Generation Discounts

Overview

These scenarios focus on removing or doubling the trip generation discounts applied to commute

trips for office workers and shoppers. Trip generation discounts are factors that reduce or increase

trip production and attraction rates in the model. Trip generation rate discounts are used to

evaluate telecommuting and e-commerce.

In both scenarios, the model should produce more VMT when trip discounts are removed and less

VMT when trip discounts are doubled. Transit mode share should rise slightly when the trip

discounts are removed due to marginal increases in network congestion.

Findings

Based on the tables below, the model is sensitive to trip generation discount changes. VMT and

transit mode shares increase when the trip generation discount is removed, and VMT and transit

mode shares decrease when trip generation rate discounts are expanded. The travel in congestion

gets slightly worse when the discount is removed, and gets slightly better when the discount is

doubled.

The corridor travel times increase slightly when the discount is removed, and decreases slightly

when the discount is doubled.

Summary Statistics

Remove Trip Generation Discount

Baseline Double Trip Generation

Discount

Total VMT 104,573,076 104,081,786 103,595,561

Total VHT 3,046,659 3,025,199 3,001,931

Commute Transit Share (Peak) 8.669% 8.629% 8.588%

All Trips Transit Share (Daily) 1.682% 1.665% 1.649%


VMT by Functional Class:

CLASS


Baseline Double Trip Generation Discount


FREEWAY 51,553,670 193,130 0.4% 51,360,540 51,146,913 -213,627 -0.4%

PRIME 8,720,267 58,915 0.7% 8,661,352 8,623,943 -37,409 -0.4%

MAJOR 18,789,010 105,425 0.6% 18,683,585 18,579,438 -104,147 -0.6%

COLLECTOR 6,652,914 42,968 0.7% 6,609,946 6,573,258 -36,688 -0.6%

LOCAL COLLECTOR 6,044,535 36,059 0.6% 6,008,476 5,977,215 -31,261 -0.5%

RURAL COLLECTOR 457,391 -677 -0.1% 458,068 453,365 -4,703 -1.0%

LOCAL 1,774,126 11,639 0.7% 1,762,487 1,750,857 -11,630 -0.7%

FWY-FWY RAMP 2,121,090 8,278 0.3% 2,112,812 2,102,893 -9,919 -0.5%

LOCAL RAMP 2,893,373 12,536 0.4% 2,880,837 2,868,906 -11,931 -0.4%

ACCESS 5,566,700 23,017 0.4% 5,543,683 5,518,773 -24,910 -0.4%

TOTAL 104,573,076 491,290 0.5% 104,081,786 103,595,561 -468,225 -0.5%


Mode Share


Baseline Double Trip Generation

Discount


REGION-WIDE Peak COMMUTE Trips

SOV 1,019,326 78.3% 997,905 78.3% 976,263 78.4%

HOV 141,142 10.8% 137,697 10.8% 134,648 10.8%

Transit 112,896 8.7% 109,911 8.6% 106,945 8.6%

School Bus 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 28,895 2.2% 28,175 2.2% 27,436 2.2%

TOTAL 1,302,259 100.0% 1,273,688 100.0% 1,245,292 100.0%

REGION-WIDE

Daily ALL Trips

SOV 11,129,116 52.7% 11,069,865 52.7% 11,010,406 52.6%

HOV 8,927,054 42.3% 8,905,753 42.4% 8,884,924 42.4%

Transit 355,106 1.7% 350,025 1.7% 345,254 1.6%

School Bus 128,873 0.6% 128,788 0.6% 128,898 0.6%

Walk & Bike 566,916 2.7% 564,954 2.7% 562,228 2.7%

TOTAL 21,107,065 100.0% 21,019,385 100.0% 20,931,710 100.0%

URBAN-AREA Peak COMMUTE Trips

SOV 840,272 77.1% 822,033 77.2% 800,846 77.2%

HOV 116,564 10.7% 114,066 10.7% 110,654 10.7%

Transit 106,959 9.8% 103,223 9.7% 101,166 9.7%

School Bus 0 0.0% 0 0.0% 0 0.0%

Walk & Bike 26,608 2.4% 24,961 2.3% 25,230 2.4%

TOTAL 1,090,403 100.0% 1,064,283 100.0% 1,037,896 100.0%

URBAN-AREA

Daily ALL Trips

SOV 9,178,602 52.8% 9,134,787 52.8% 9,069,963 52.6%

HOV 7,275,742 41.8% 7,263,745 42.0% 7,237,354 42.0%

Transit 337,623 1.9% 328,328 1.9% 328,011 1.9%

School Bus 93,275 0.5% 93,252 0.5% 93,297 0.5%

Walk & Bike 505,887 2.9% 490,439 2.8% 501,418 2.9%

TOTAL 17,391,129 100.0% 17,310,551 100.0% 17,230,043 100.0%


4. GLOSSARY OF ACRONYMS AND TERMS

ADT Average Daily Traffic

Elasticity The responsiveness of a dependent variable to changes in influencing model

input variables

HOV High Occupancy Vehicle

LOS Level of Service

MTS Metropolitan Transit System

NCTD North County Transit District

RTP Regional Transportation Plan

SANDAG San Diego Association of Governments

SOV Single Occupancy Vehicle

VHT Vehicle Hours of Travel

VMT Vehicle Miles of Travel

Appendix A

SANDAG Transportaion Model Sensitivity Report Performance Measures

Section 3.2 - Auto Operating Cost

Goals and Performance Measures

50% of

Baseline Auto

Operating

Costs

75% of

Baseline Auto

Operating

Costs Baseline

125% of

Baseline Auto

Operating

Costs

150% of

Baseline Auto

Operating Cost

SYSTEM PRESERVATION AND SAFETY

1 Annual projected number of vehicle injury/fatal collisions per capita 7.74 7.30 6.90 6.54 6.23

2

Annual projected number of bicycle/pedestrian injury/fatal collisions per 1,000

persons (w/o Post Processor) 0.49 0.52 0.55 0.58 0.61

MOBILITY

5 Average peak work trip travel time (in minutes) 30.6 29.4 28.3 27.5 26.8

Drive alone 28.9 27.5 26.0 24.8 23.8

Carpool 31.3 29.1 27.1 25.6 24.5

Transit 51.2 52.6 53.7 54.4 54.6

6 Average work trip travel speed by mode (in m.p.h.)

Drive alone 28.2 29.3 30.3 31.0 31.4

Carpool 29.4 30.6 31.6 32.3 32.7

Transit 12.2 12.4 12.5 12.5 12.5

7

Percent of work and higher education trips accessible in 30 minutes in peak periods

by mode

Drive alone 64% 67% 70% 74% 76%

Carpool 66% 69% 72% 76% 79%

Transit 12% 13% 13% 14% 15%

8 Percent of non work-related trips accessible in 15 minutes by modeDrive alone 62% 65% 68% 71% 73%Carpool 63% 66% 69% 72% 74%

Transit 7% 8% 8% 9% 9%

9 Out-of-pocket user costs per trip $2.45 $2.31 $2.19 $2.08 $1.99

RELIABILITY

15 Congested vehicle miles of travel (VMT)

Peak Period Percent of ALL auto travel at LOS E or F 22% 18% 14% 11% 9%

Daily Percent of ALL auto travel at LOS E or F 14% 10% 8% 6% 4%

Peak Period Percent of FREEWAY auto travel at LOS E or F 35% 29% 23% 17% 13%

Daily Percent of FREEWAY auto travel at LOS E or F 22% 17% 12% 9% 7%

Peak Period Percent of ALL auto travel at LOS F 10% 8% 6% 4% 3%

Daily Percent of ALL auto travel at LOS F 6% 4% 3% 2% 2%

Peak Period Percent of FREEWAY auto travel at LOS F 17% 14% 9% 6% 4%

Daily Percent of FREEWAY auto travel at LOS F 10% 7% 5% 3% 2%

16 Daily vehicle delay per capita (minutes) 6.5 4.8 3.63 2.8 2.4

17 Daily truck hours of delay 17,273 13,266 10,582 8,687 7,552

18 Percent of VMT by travel speed by mode

Drive alone

Percent of VMT traveling from 0 to 35 mph 12% 9% 6% 4% 3%


Percent of VMT traveling greater than 55 mph 74% 80% 85% 89% 91%

Carpool




Truck




HEALTHY ENVIRONMENT

22 Systemwide VMT (all day) per capita 29.00 27.36 25.85 24.49 23.30

23 Transit passenger miles (all day) per capita 0.55 0.65 0.77 0.88 0.99

24 Percent of peak-period trips within 1/2 mile of a transit stop 78% 78% 78% 78% 78%

25 Percent of daily trips within 1/2 mile of transit stop 80% 80% 80% 80% 80%

26 Work trip mode share (peak periods)* (w/o Post Processing)

Drive alone 79.3% 78.8% 78.3% 77.8% 77.2%

Carpool 11.4% 11.1% 10.8% 10.5% 10.3%

Transit 7.3% 7.9% 8.6% 9.3% 9.9%

Bike/Walk 2.0% 2.1% 2.2% 2.4% 2.5%

3.2 - Auto Operating Cost - 1 7/29/2011

Appendix A




50% of

Baseline Auto

Operating

Costs

75% of

Baseline Auto

Operating

Costs Baseline

125% of

Baseline Auto

Operating

Costs

150% of

Baseline Auto

Operating Cost

HEALTHY ENVIRONMENT

27 Daily Commute mode share (w/o Post Processing)

Drive alone 79.7% 79.3% 78.9% 78.3% 77.7%

Carpool 11.2% 10.8% 10.5% 10.3% 10.1%

Transit 6.9% 7.5% 8.2% 8.8% 9.4%

Bike/Walk 2.2% 2.3% 2.4% 2.6% 2.8%

28 Non work trip mode share (peak periods)* (w/o Post Processing)

Drive alone 46.4% 46.2% 46.1% 45.9% 45.7%

Carpool 50.2% 50.0% 49.9% 49.8% 49.7%

Transit 0.7% 0.8% 0.9% 1.0% 1.1%

Bike/Walk 2.8% 3.0% 3.2% 3.4% 3.5%

29 Non work trip mode share (all day)* (w/o Post Processing)

Drive alone 50.1% 50.0% 49.9% 49.8% 49.7%

Carpool 46.8% 46.6% 46.5% 46.4% 46.2%

Transit 0.7% 0.8% 0.9% 1.0% 1.1%

Bike/Walk 2.4% 2.6% 2.7% 2.9% 3.0%

30 Total bike and walk trips (w/o Post Processing) 494,902 531,338 564,954 595,912 624,765

SOCIAL EQUITY

32 Average travel time per person trip (in minutes)

Low-income population 18.5 17.3 16.4 15.7 15.2

Non low-income population 18.4 17.3 16.3 15.6 15.1

Minority population 18.1 17.0 16.1 15.4 14.9

Non minority population 18.4 17.3 16.3 15.6 15.1

Mobility population 19.0 17.8 16.9 16.1 15.6

Non mobility population 18.3 17.2 16.2 15.5 15.0

Community engagement population 18.4 17.3 16.3 15.6 15.0

Non community engagement population 18.5 17.4 16.4 15.7 15.2

33 Percent of work trips accessible in 30 minutes in peak periods by mode

Low-income population 60% 63% 66% 68% 70%

Drive alone 68% 71% 75% 79% 82%

Carpool 69% 73% 77% 81% 84%

Transit 20% 21% 22% 23% 24%

Non low-income population 60% 62% 65% 67% 70%

SOV/Drive alone 63% 66% 69% 72% 75%

Carpool 65% 68% 71% 74% 77%

Transit 10% 10% 11% 11% 12%

Minority population 60% 62% 65% 68% 70%

SOV/Drive alone 66% 69% 72% 76% 79%

Carpool 67% 70% 74% 78% 81%

Transit 15% 15% 16% 17% 18%

Non minority population 60% 62% 65% 68% 70%

SOV/Drive alone 63% 66% 69% 72% 75%

Carpool 65% 68% 71% 74% 77%

Transit 10% 11% 11% 12% 12%

Mobility population

SOV/Drive alone 68% 72% 75% 78% 81%

Carpool 70% 73% 77% 80% 83%

Transit 17% 18% 19% 20% 20%


Appendix A




50% of

Baseline Auto

Operating

Costs

75% of

Baseline Auto

Operating

Costs Baseline

125% of

Baseline Auto

Operating

Costs

150% of

Baseline Auto

Operating Cost

SOCIAL EQUITY

33 Percent of work trips accessible in 30 minutes in peak periods by mode (cont.)

Non mobility population

SOV/Drive alone 63% 66% 69% 72% 75%

Carpool 64% 68% 71% 74% 77%

Transit 11% 11% 12% 12% 13%

Community engagement population

SOV/Drive alone 66% 69% 73% 77% 80%

Carpool 67% 71% 75% 79% 82%

Transit 18% 19% 19% 20% 21%

Non community engagement population

SOV/Drive alone 63% 66% 69% 72% 75%

Carpool 65% 68% 71% 74% 77%

Transit 10% 11% 11% 11% 12%

34 Percent of homes within 1/2 mile of a transit stop





Mobility population 74% 74% 74% 74% 74%

Non mobility population 67% 67% 67% 67% 67%

Community engagement population 89% 89% 89% 89% 89%

Non community engagement population 61% 61% 61% 61% 61%

PEAK PERIOD AVERAGE TRAVEL TIMES BY CORRIDOR

Oceanside - Downtown San Diego

1a By auto 70 63 58 55 521b By transit (walk access) 96 96 96 96 961c By transit (park and ride access) 88 88 88 88 881d By carpool 69 62 57 53 511e Corridor Weighted Average 75 71 67 66 64

Escondido - Downtown San Diego


El Cajon - Kearny Mesa


Mid-City - UTC


Western Chula Vista - Mission Valley


Carlsbad - Sorrento Mesa


Oceanside - Escondido


San Ysidro - Downtown San Diego



Appendix A




50% of

Baseline Auto

Operating

Costs

75% of

Baseline Auto

Operating

Costs Baseline

125% of

Baseline Auto

Operating

Costs

150% of

Baseline Auto

Operating Cost


Otay Ranch - UTC


Pala/Pauma - Oceanside Transit Center


SR 67 (Ramona) - Downtown San Diego

11a By auto 68 65 63 61 5911b By transit (walk access) 114 113 113 113 11211c By transit (park and ride access) 103 102 102 101 10011d By carpool 65 63 63 61 59

11e Corridor Weighted Average 78 77 77 77 77

Total Population 4,026,131 4,026,131 4,026,131 4,026,131 4,026,131


Appendix A


Section 3.3 - Parking Cost


50% of

Baseline

Parking Costs

75% of

Baseline


125% of

Baseline

Parking Costs

150% of

Baseline

Parking Cost



2



MOBILITY


Drive alone 26.0 26.0 26.0 26.0 26.1

Carpool 27.1 27.1 27.1 27.1 27.2

Transit 54.0 53.9 53.7 53.5 53.4


Drive alone 30.3 30.3 30.3 30.4 30.3

Carpool 31.6 31.6 31.6 31.6 31.6

Transit 12.4 12.5 12.5 12.5 12.5

7


by mode

Drive alone 70% 70% 70% 70% 70%

Carpool 73% 72% 72% 72% 72%

Transit 13% 13% 13% 13% 13%


Transit 8% 8% 8% 8% 8%


RELIABILITY













Drive alone




Carpool




Truck




HEALTHY ENVIRONMENT






Drive alone 78.7% 78.5% 78.3% 78.2% 78.0%

Carpool 10.8% 10.9% 10.8% 10.8% 10.8%

Transit 8.4% 8.5% 8.6% 8.7% 8.8%

Bike/Walk 2.1% 2.1% 2.2% 2.3% 2.4%

3.3 - Parking Cost - 1 7/29/2011

Appendix A




50% of

Baseline

Parking Costs

75% of

Baseline


125% of

Baseline

Parking Costs

150% of

Baseline

Parking Cost

HEALTHY ENVIRONMENT

Drive alone 79.2% 79.0% 78.9% 78.7% 78.5%

Carpool 10.6% 10.6% 10.5% 10.5% 10.5%

Transit 7.9% 8.1% 8.2% 8.3% 8.4%

Bike/Walk 2.3% 2.3% 2.4% 2.5% 2.6%


Drive alone 46.1% 46.1% 46.1% 46.1% 46.0%

Carpool 49.9% 49.9% 49.9% 49.8% 49.8%

Transit 0.8% 0.8% 0.9% 0.9% 0.9%

Bike/Walk 3.1% 3.2% 3.2% 3.2% 3.3%


Drive alone 50% 50% 50% 50% 50%

Carpool 47% 47% 47% 47% 46%

Transit 1% 1% 1% 1% 1%

Bike/Walk 3% 3% 3% 3% 3%


SOCIAL EQUITY












Drive alone 75% 75% 75% 75% 75%

Carpool 77% 77% 77% 77% 77%

Transit 22% 22% 22% 22% 22%


SOV/Drive alone 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71%

Transit 11% 11% 11% 11% 11%


SOV/Drive alone 72% 72% 72% 72% 72%

Carpool 74% 74% 74% 74% 74%

Transit 16% 16% 16% 16% 16%


SOV/Drive alone 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71%

Transit 11% 11% 11% 11% 11%

Mobility population

SOV/Drive alone 75% 75% 75% 75% 75%

Carpool 77% 77% 77% 77% 77%

Transit 18% 18% 19% 19% 19%

3.3 - Parking Cost - 2 7/29/2011

Appendix A




50% of

Baseline

Parking Costs

75% of

Baseline


125% of

Baseline

Parking Costs

150% of

Baseline

Parking Cost

SOCIAL EQUITY



SOV/Drive alone 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71%

Transit 12% 12% 12% 12% 12%


SOV/Drive alone 73% 73% 73% 73% 73%

Carpool 75% 75% 75% 75% 75%

Transit 19% 19% 19% 19% 20%


SOV/Drive alone 69% 69% 69% 69% 69%

Carpool 72% 71% 71% 71% 71%

Transit 11% 11% 11% 11% 11%

















Mid-City - UTC










3.3 - Parking Cost - 3 7/29/2011

Appendix A




50% of

Baseline

Parking Costs

75% of

Baseline


125% of

Baseline

Parking Costs

150% of

Baseline

Parking Cost


Otay Ranch - UTC








3.3 - Parking Cost - 4 7/29/2011

Appendix A


Section 3.4 - Income Distribution

Goals and Performance MeasuresVery Low

Income Scale

Low Income

Scale

Middle Income

Scale Baseline

High Income

Scale



2



MOBILITY


Drive alone 21.7 23.4 25.3 26.0 27.4

Carpool 21.7 23.2 25.2 27.1 29.4

Transit 55.9 55.1 52.2 53.7 50.7


Drive alone 31.2 31.6 31.0 30.3 29.2

Carpool 32.1 32.6 32.2 31.6 30.5

Transit 12.1 12.3 12.4 12.5 12.5

7


by mode

Drive alone 83% 79% 73% 70% 67%

Carpool 85% 81% 75% 72% 69%

Transit 17% 15% 14% 13% 13%


Transit 10% 10% 9% 8% 8%


RELIABILITY













Drive alone




Carpool




Truck




HEALTHY ENVIRONMENT






Drive alone 63.5% 70.0% 76.0% 78.3% 81.8%

Carpool 11.7% 11.0% 10.5% 10.8% 11.0%

Transit 19.6% 15.3% 11.0% 8.6% 5.5%

Bike/Walk 5.3% 3.8% 2.5% 2.2% 1.7%

3.4 - Income Distribution - 1 7/29/2011

Appendix A




Income Scale

Low Income

Scale

Middle Income

Scale Baseline

High Income

Scale

HEALTHY ENVIRONMENT

Drive alone 64.0% 70.4% 76.5% 78.9% 82.3%

Carpool 11.7% 10.9% 10.4% 10.5% 10.6%

Transit 18.6% 14.5% 10.4% 8.2% 5.2%

Bike/Walk 5.7% 4.1% 2.7% 2.4% 1.8%


Drive alone 43.1% 44.2% 45.3% 46.1% 46.9%

Carpool 49.3% 49.9% 50.7% 49.9% 49.8%

Transit 2.4% 1.7% 0.9% 0.9% 0.5%

Bike/Walk 5.3% 4.2% 3.2% 3.2% 2.8%


Drive alone 46% 48% 49% 50% 51%

Carpool 46% 47% 47% 47% 46%

Transit 2% 2% 1% 1% 1%

Bike/Walk 5% 4% 3% 3% 2%

30 Total bike and walk trips (w/o Post Processing) 1,019,638 788,997 555,205 564,954 472,181

SOCIAL EQUITY












Drive alone 88% 84% 77% 75% 70%

Carpool 90% 86% 79% 77% 72%

Transit 28% 25% 22% 22% 20%


SOV/Drive alone 81% 77% 72% 69% 66%

Carpool 83% 79% 74% 71% 68%

Transit 14% 12% 11% 11% 10%


SOV/Drive alone 85% 81% 75% 72% 68%

Carpool 87% 83% 77% 74% 70%

Transit 20% 18% 16% 16% 15%


SOV/Drive alone 81% 76% 71% 69% 66%

Carpool 83% 79% 74% 71% 68%

Transit 14% 13% 11% 11% 11%

Mobility population

SOV/Drive alone 87% 83% 77% 75% 71%

Carpool 89% 85% 79% 77% 72%

Transit 24% 22% 19% 19% 17%


Appendix A




Income Scale

Low Income

Scale

Middle Income

Scale Baseline

High Income

Scale

SOCIAL EQUITY



SOV/Drive alone 82% 77% 71% 69% 66%

Carpool 84% 79% 74% 71% 67%

Transit 15% 13% 12% 12% 11%


SOV/Drive alone 87% 83% 76% 73% 69%

Carpool 89% 85% 78% 75% 70%

Transit 24% 22% 19% 19% 18%


SOV/Drive alone 81% 77% 72% 69% 66%

Carpool 83% 79% 74% 71% 68%

Transit 14% 12% 11% 11% 11%

















Mid-City - UTC











Appendix A




Income Scale

Low Income

Scale

Middle Income

Scale Baseline

High Income

Scale


Otay Ranch - UTC









Appendix A


Section 3.5 - Transit Fares


50% of

Baseline Transit

Fares

75% of

Baseline Transit

Fares Baseline

125% of

Baseline Transit

Fares

150% of

Baseline Transit

Fares



2



MOBILITY


Drive alone 26.1 26.1 26.0 26.0 26.1

Carpool 27.2 27.2 27.1 27.1 27.2

Transit 53.7 53.7 53.7 53.8 53.4


Drive alone 30.3 30.3 30.3 30.3 30.3

Carpool 31.6 31.6 31.6 31.6 31.5

Transit 12.4 12.5 12.5 12.5 12.4

7


by mode

Drive alone 70% 70% 70% 70% 70%

Carpool 72% 72% 72% 72% 72%

Transit 13% 13% 13% 13% 13%


Transit 8% 8% 8% 8% 8%


RELIABILITY













Drive alone




Carpool




Truck




HEALTHY ENVIRONMENT






Drive alone 77.2% 77.8% 78.3% 78.8% 79.2%

Carpool 10.6% 10.7% 10.8% 10.9% 11.0%

Transit 10.1% 9.3% 8.6% 8.1% 7.5%

Bike/Walk 2.1% 2.1% 2.2% 2.3% 2.3%

3.5 - Transit Fares - 1 7/29/2011

Appendix A




50% of

Baseline Transit

Fares

75% of

Baseline Transit

Fares Baseline

125% of

Baseline Transit

Fares

150% of

Baseline Transit

Fares

HEALTHY ENVIRONMENT


Drive alone 77.7% 78.3% 78.9% 79.3% 79.7%

Carpool 10.3% 10.4% 10.5% 10.6% 10.7%

Transit 9.7% 8.9% 8.2% 7.6% 7.1%

Bike/Walk 2.3% 2.4% 2.4% 2.5% 2.5%


Drive alone 46.0% 46.0% 46.1% 46.1% 46.2%

Carpool 49.7% 49.8% 49.9% 49.9% 49.9%

Transit 1.1% 1.0% 0.9% 0.8% 0.7%

Bike/Walk 3.2% 3.2% 3.2% 3.2% 3.2%


Drive alone 50% 50% 50% 50% 50%

Carpool 46% 46% 47% 47% 47%

Transit 1% 1% 1% 1% 1%

Bike/Walk 3% 3% 3% 3% 3%


SOCIAL EQUITY












Drive alone 75% 75% 75% 75% 75%

Carpool 77% 77% 77% 77% 77%

Transit 22% 22% 22% 22% 22%


SOV/Drive alone 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71%

Transit 11% 11% 11% 11% 11%


SOV/Drive alone 72% 72% 72% 72% 72%

Carpool 74% 74% 74% 74% 74%

Transit 16% 16% 16% 16% 16%


SOV/Drive alone 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71%

Transit 11% 11% 11% 11% 11%

Mobility population

SOV/Drive alone 75% 75% 75% 75% 75%

Carpool 77% 77% 77% 77% 77%

Transit 19% 19% 19% 19% 19%

3.5 - Transit Fares - 2 7/29/2011

Appendix A




50% of

Baseline Transit

Fares

75% of

Baseline Transit

Fares Baseline

125% of

Baseline Transit

Fares

150% of

Baseline Transit

Fares

SOCIAL EQUITY



SOV/Drive alone 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71%

Transit 12% 12% 12% 12% 12%


SOV/Drive alone 73% 73% 73% 73% 73%

Carpool 75% 75% 75% 75% 75%

Transit 19% 19% 19% 19% 19%


SOV/Drive alone 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71%

Transit 11% 11% 11% 11% 11%


Low-income population 92.0% 92.0% 92.0% 92.0% 92.0%















Mid-City - UTC










3.5 - Transit Fares - 3 7/29/2011

Appendix A




50% of

Baseline Transit

Fares

75% of

Baseline Transit

Fares Baseline

125% of

Baseline Transit

Fares

150% of

Baseline Transit

Fares


Otay Ranch - UTC








3.5 - Transit Fares - 4 7/29/2011

Appendix A


Section 3.6 - Transit Frequency


Decrease COASTER Frequency Baseline


Decrease Route 7 Frequency Baseline

Increase Route 7 Frequency


1 Annual projected number of vehicle injury/fatal collisions per capita 6.90 6.90 6.90 6.90 6.90 6.90

2

Annual projected number of bicycle/pedestrian injury/fatal collisions per 1,000persons (w/o Post Processor) 0.55 0.55 0.55 0.55 0.55 0.55

MOBILITY

5 Average peak work trip travel time (in minutes) 28.3 28.3 28.4 28.4 28.3 28.4

Drive alone 26.1 26.0 26.1 26.1 26.0 26.1

Carpool 27.1 27.1 27.2 27.2 27.1 27.2

Transit 53.8 53.7 53.7 53.7 53.7 53.7


Drive alone 30.3 30.3 30.3 30.3 30.3 30.3

Carpool 31.6 31.6 31.6 31.5 31.6 31.5

Transit 12.4 12.5 12.6 12.5 12.5 12.5

7

Percent of work and higher education trips accessible in 30 minutes in peak periodsby mode

Drive alone 70% 70% 70% 70% 70% 70%

Carpool 72% 72% 72% 72% 72% 72%

Transit 13% 13% 13% 13% 13% 13%

8 Percent of non work-related trips accessible in 15 minutes by modeDrive alone 68% 68% 68% 68% 68% 68%Carpool 69% 69% 69% 69% 69% 69%

Transit 8% 8% 8% 8% 8% 8%

9 Out-of-pocket user costs per trip $2.19 $2.19 $2.19 $2.19 $2.19 $2.19

RELIABILITY


Peak Period Percent of ALL auto travel at LOS E or F 15% 14% 14% 14% 14% 14%

Daily Percent of ALL auto travel at LOS E or F 8% 8% 8% 8% 8% 8%

Peak Period Percent of FREEWAY auto travel at LOS E or F 23% 23% 23% 23% 23% 23%

Daily Percent of FREEWAY auto travel at LOS E or F 13% 12% 12% 12% 12% 12%

Peak Period Percent of ALL auto travel at LOS F 6% 6% 6% 6% 6% 6%

Daily Percent of ALL auto travel at LOS F 3% 3% 3% 3% 3% 3%

Peak Period Percent of FREEWAY auto travel at LOS F 9% 9% 10% 9% 9% 9%

Daily Percent of FREEWAY auto travel at LOS F 5% 5% 5% 5% 5% 5%

16 Daily vehicle delay per capita (minutes) 3.6 3.63 3.6 3.6 3.63 3.6

17 Daily truck hours of delay 10,642 10,582 10,613 10,597 10,582 10,562


Drive alone

Percent of VMT traveling from 0 to 35 mph 6% 6% 6% 6% 6% 6%


Percent of VMT traveling greater than 55 mph 85% 85% 85% 85% 85% 84%

Carpool




Truck




HEALTHY ENVIRONMENT

22 Systemwide VMT (all day) per capita 25.86 25.85 25.85 25.85 25.85 25.85

23 Transit passenger miles (all day) per capita 0.76 0.77 0.78 0.77 0.77 0.77

24 Percent of peak-period trips within 1/2 mile of a transit stop 78% 78% 78% 78% 78% 78%

25 Percent of daily trips within 1/2 mile of transit stop 80% 80% 80% 80% 80% 80%


Drive alone 78.4% 78.3% 78.2% 78.3% 78.3% 78.3%

Carpool 10.8% 10.8% 10.8% 10.8% 10.8% 10.8%

Transit 8.6% 8.6% 8.7% 8.6% 8.6% 8.6%

Bike/Walk 2.2% 2.2% 2.2% 2.2% 2.2% 2.2%

3.6 - Transit Frequency - 1 7/29/2011

Appendix A








HEALTHY ENVIRONMENT


Drive alone 78.9% 78.9% 78.8% 78.9% 78.9% 78.9%

Carpool 10.5% 10.5% 10.5% 10.5% 10.5% 10.5%

Transit 8.1% 8.2% 8.2% 8.2% 8.2% 8.2%

Bike/Walk 2.4% 2.4% 2.4% 2.4% 2.4% 2.4%


Drive alone 46.1% 46.1% 46.1% 46.1% 46.1% 46.1%

Carpool 49.9% 49.9% 49.9% 49.9% 49.9% 49.9%

Transit 0.9% 0.9% 0.9% 0.9% 0.9% 0.9%

Bike/Walk 3.2% 3.2% 3.2% 3.2% 3.2% 3.2%


Drive alone 49.9% 49.9% 49.9% 49.9% 49.9% 49.9%

Carpool 46.5% 46.5% 46.5% 46.5% 46.5% 46.5%

Transit 0.9% 0.9% 0.9% 0.9% 0.9% 0.9%

Bike/Walk 2.7% 2.7% 2.7% 2.7% 2.7% 2.7%

30 Total bike and walk trips (w/o Post Processing) 564,573 564,954 564,515 564,573 564,954 564,639

SOCIAL EQUITY


Low-income population 16.4 16.4 16.4 16.4 16.4 16.4

Non low-income population 16.3 16.3 16.4 16.4 16.3 16.4

Minority population 16.1 16.1 16.1 16.1 16.1 16.1

Non minority population 16.4 16.3 16.4 16.4 16.3 16.4

Mobility population 16.9 16.9 16.9 16.9 16.9 16.9

Non mobility population 16.2 16.2 16.3 16.3 16.2 16.3

Community engagement population 16.3 16.3 16.3 16.3 16.3 16.3

Non community engagement population 16.4 16.4 16.4 16.4 16.4 16.4


Low-income population 66% 66% 66% 66% 66% 66%

Drive alone 75% 75% 75% 75% 75% 75%

Carpool 77% 77% 77% 77% 77% 77%

Transit 22% 22% 22% 22% 22% 22%

Non low-income population 65% 65% 65% 65% 65% 65%

SOV/Drive alone 69% 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71% 71%

Transit 11% 11% 11% 11% 11% 11%

Minority population 65% 65% 65% 65% 65% 65%

SOV/Drive alone 72% 72% 72% 72% 72% 72%

Carpool 74% 74% 74% 74% 74% 74%

Transit 16% 16% 16% 16% 16% 16%

Non minority population 65% 65% 65% 65% 65% 65%

SOV/Drive alone 69% 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71% 71%

Transit 11% 11% 11% 11% 11% 11%

Mobility population

SOV/Drive alone 75% 75% 75% 75% 75% 75%

Carpool 77% 77% 77% 77% 77% 77%

Transit 19% 19% 19% 19% 19% 19%


Appendix A








SOCIAL EQUITY



SOV/Drive alone 69% 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71% 71%

Transit 12% 12% 12% 12% 12% 12%


SOV/Drive alone 73% 73% 73% 73% 73% 73%

Carpool 75% 75% 75% 75% 75% 75%

Transit 19% 19% 19% 19% 19% 19%


SOV/Drive alone 69% 69% 69% 69% 69% 69%

Carpool 71% 71% 71% 71% 71% 71%

Transit 11% 11% 11% 11% 11% 11%


Low-income population 92% 92% 92% 92% 92% 92%

Non low-income population 62% 62% 62% 62% 62% 62%

Minority population 81% 81% 81% 81% 81% 81%

Non minority population 59% 59% 59% 59% 59% 59%

Mobility population 74% 74% 74% 74% 74% 74%

Non mobility population 67% 67% 67% 67% 67% 67%

Community engagement population 89% 89% 89% 89% 89% 89%

Non community engagement population 61% 61% 61% 61% 61% 61%



1a By auto 58 58 59 59 58 581b By transit (walk access) 114 96 91 96 96 961c By transit (park and ride access) 91 88 83 88 88 881d By carpool 57 57 57 57 57 571e Corridor Weighted Average 68 67 68 68 67 68





Mid-City - UTC











Appendix A









Otay Ranch - UTC






Total Population 4,026,131 4,026,131 4,026,131 4,026,131 4,026,131 4,026,131


Appendix A


Section 3.7 - Tranist Access


Reduce Wait

and Transfer

Times Baseline

Increase Wait

and Transfer

Times


1 Annual projected number of vehicle injury/fatal collisions per capita 6.85 6.90 6.93

2


persons (w/o Post Processor) 0.54 0.55 0.55

MOBILITY

5 Average peak work trip travel time (in minutes) 28.6 28.3 28.1

Drive alone 26.1 26.0 26.0

Carpool 27.2 27.1 27.2

Transit 50.5 53.7 56.3


Drive alone 30.3 30.3 30.3

Carpool 31.6 31.6 31.7

Transit 13.8 12.5 11.5

7


by mode

Drive alone 70% 70% 70%

Carpool 72% 72% 72%

Transit 18% 13% 10%

8 Percent of non work-related trips accessible in 15 minutes by modeDrive alone 68% 68% 68%Carpool 69% 69% 69%

Transit 12% 8% 6%

9 Out-of-pocket user costs per trip $2.18 $2.19 $2.20

RELIABILITY


Peak Period Percent of ALL auto travel at LOS E or F 14% 14% 15%

Daily Percent of ALL auto travel at LOS E or F 7% 8% 8%

Peak Period Percent of FREEWAY auto travel at LOS E or F 22% 23% 24%

Daily Percent of FREEWAY auto travel at LOS E or F 12% 12% 13%

Peak Period Percent of ALL auto travel at LOS F 5% 6% 7%

Daily Percent of ALL auto travel at LOS F 3% 3% 3%

Peak Period Percent of FREEWAY auto travel at LOS F 8% 9% 10%

Daily Percent of FREEWAY auto travel at LOS F 4% 5% 5%

16 Daily vehicle delay per capita (minutes) 3.5 3.63 3.7

17 Daily truck hours of delay 10,149 10,582 10,825


Drive alone

Percent of VMT traveling from 0 to 35 mph 5% 6% 6%


Percent of VMT traveling greater than 55 mph 86% 85% 84%

Carpool




Truck




HEALTHY ENVIRONMENT

22 Systemwide VMT (all day) per capita 25.68 25.85 25.96

23 Transit passenger miles (all day) per capita 0.97 0.77 0.63

24 Percent of peak-period trips within 1/2 mile of a transit stop 78% 78% 78%

25 Percent of daily trips within 1/2 mile of transit stop 80% 80% 80%


Drive alone 76.6% 78.3% 79.5%

Carpool 10.6% 10.8% 11.1%

Transit 10.6% 8.6% 7.2%

Bike/Walk 2.1% 2.2% 2.3%

3.7 - Transit Access - 1 7/29/2011

Appendix A




Reduce Wait

and Transfer

Times Baseline

Increase Wait

and Transfer

Times

HEALTHY ENVIRONMENT


Drive alone 77.2% 78.9% 79.9%

Carpool 10.3% 10.5% 10.7%

Transit 10.1% 8.2% 6.8%

Bike/Walk 2.4% 2.4% 2.5%


Drive alone 46.0% 46.1% 46.1%

Carpool 49.8% 49.9% 49.9%

Transit 1.0% 0.9% 0.8%

Bike/Walk 3.2% 3.2% 3.2%


Drive alone 49.8% 49.9% 49.9%

Carpool 46.4% 46.5% 46.5%

Transit 1.0% 0.9% 0.8%

Bike/Walk 2.7% 2.7% 2.7%

30 Total bike and walk trips (w/o Post Processing) 561,755 564,954 566,862

SOCIAL EQUITY


Low-income population 16.4 16.4 16.4

Non low-income population 16.4 16.3 16.3

Minority population 16.2 16.1 16.1

Non minority population 16.4 16.3 16.3

Mobility population 16.9 16.9 16.9

Non mobility population 16.3 16.2 16.2

Community engagement population 16.3 16.3 16.3

Non community engagement population 16.5 16.4 16.4


Low-income population 65% 66% 66%


Carpool 77% 77% 77%

Transit 28% 22% 17%

Non low-income population 64% 65% 65%

SOV/Drive alone 69% 69% 69%

Carpool 71% 71% 71%

Transit 15% 11% 8%

Minority population 64% 65% 66%


Carpool 74% 74% 74%

Transit 22% 16% 12%

Non minority population 65% 65% 66%


Carpool 71% 71% 71%

Transit 15% 11% 8%

Mobility population


Carpool 77% 77% 77%

Transit 24% 19% 15%

3.7 - Transit Access - 2 7/29/2011

Appendix A




Reduce Wait

and Transfer

Times Baseline

Increase Wait

and Transfer

Times

SOCIAL EQUITY




Carpool 71% 71% 71%

Transit 16% 12% 8%



Carpool 75% 75% 75%

Transit 25% 19% 15%



Carpool 71% 71% 71%

Transit 15% 11% 8%






Mobility population 74% 74% 74%

Non mobility population 67% 67% 67%

Community engagement population 89% 89% 89%

Non community engagement population 61% 61% 61%



1a By auto 59 58 591b By transit (walk access) 89 96 1041c By transit (park and ride access) 83 88 931d By carpool 57 57 581e Corridor Weighted Average 68 67 69





Mid-City - UTC










3.7 - Transit Access - 3 7/29/2011

Appendix A




Reduce Wait

and Transfer

Times Baseline

Increase Wait

and Transfer

Times


Otay Ranch - UTC





11a By auto 63 63 6311b By transit (walk access) 104 113 12311c By transit (park and ride access) 97 102 10711d By carpool 63 63 63

11e Corridor Weighted Average 76 77 79

Total Population 4,026,131 4,026,131 4,026,131

3.7 - Transit Access - 4 7/29/2011

Appendix A


Section 3.8 - Transit Access - Walk Factors


Urban Walk

Factors Only Baseline

Suburban Walk

Factors Only



2



MOBILITY


Drive alone 26.1 26.0 26.1

Carpool 27.2 27.1 27.1

Transit 52.9 53.7 54.6


Drive alone 30.3 30.3 30.3

Carpool 31.5 31.6 31.5

Transit 12.8 12.5 12.4

7


by mode


Carpool 72% 72% 72%

Transit 16% 13% 13%


Transit 11% 8% 8%


RELIABILITY













Drive alone




Carpool




Truck




HEALTHY ENVIRONMENT






Drive alone 77.5% 78.3% 78.6%

Carpool 10.7% 10.8% 10.8%

Transit 9.7% 8.6% 8.4%

Bike/Walk 2.2% 2.2% 2.2%

3.8 - Walk Factors - 1 7/29/2011

Appendix A




Urban Walk


Suburban Walk

Factors Only

HEALTHY ENVIRONMENT


Drive alone 78.0% 78.9% 79.1%

Carpool 10.4% 10.5% 10.6%

Transit 9.2% 8.2% 7.9%

Bike/Walk 2.4% 2.4% 2.5%


Drive alone 46.0% 46.1% 46.1%

Carpool 49.8% 49.9% 49.9%

Transit 1.0% 0.9% 0.8%

Bike/Walk 3.2% 3.2% 3.2%


Drive alone 49.8% 49.9% 49.9%

Carpool 46.4% 46.5% 46.5%

Transit 1.0% 0.9% 0.9%

Bike/Walk 2.7% 2.7% 2.7%


SOCIAL EQUITY













Carpool 77% 77% 77%

Transit 25% 22% 21%



Carpool 71% 71% 71%

Transit 14% 11% 10%



Carpool 74% 74% 74%

Transit 20% 16% 15%



Carpool 71% 71% 71%

Transit 14% 11% 11%

Mobility population


Carpool 77% 77% 77%

Transit 22% 19% 18%

3.8 - Walk Factors - 2 7/29/2011

Appendix A




Urban Walk


Suburban Walk

Factors Only

SOCIAL EQUITY




Carpool 71% 71% 71%

Transit 15% 12% 11%



Carpool 75% 75% 75%

Transit 23% 19% 18%



Carpool 71% 71% 71%

Transit 14% 11% 10%

















Mid-City - UTC










3.8 - Walk Factors - 3 7/29/2011

Appendix A




Urban Walk


Suburban Walk

Factors Only


Otay Ranch - UTC








3.8 - Walk Factors - 4 7/29/2011

Appendix A


Section 3.9 - Network Assignment


Delete

El Camino Real Baseline

Delete

I-805 Mission

Valley Viaduct



2



MOBILITY


Drive alone 26.0 26.0 26.3

Carpool 27.1 27.1 27.4

Transit 53.7 53.7 53.7


Drive alone 30.3 30.3 30.0

Carpool 31.6 31.6 31.2

Transit 12.5 12.5 12.6

7


by mode


Carpool 72% 72% 72%

Transit 13% 13% 14%


Transit 8% 8% 8%


RELIABILITY













Drive alone




Carpool




Truck




HEALTHY ENVIRONMENT






Drive alone 78.3% 78.3% 78.3%

Carpool 10.8% 10.8% 10.8%

Transit 8.6% 8.6% 8.7%

Bike/Walk 2.2% 2.2% 2.2%

3.9 - Network Assignment - 1 7/29/2011

Appendix A




Delete


Delete

I-805 Mission

Valley Viaduct

HEALTHY ENVIRONMENT


Drive alone 78.9% 78.9% 78.8%

Carpool 10.5% 10.5% 10.5%

Transit 8.2% 8.2% 8.2%

Bike/Walk 2.4% 2.4% 2.5%


Drive alone 46.1% 46.1% 46.1%

Carpool 49.9% 49.9% 49.9%

Transit 0.9% 0.9% 0.9%

Bike/Walk 3.2% 3.2% 3.2%


Drive alone 49.9% 49.9% 49.9%

Carpool 46.5% 46.5% 46.5%

Transit 0.9% 0.9% 0.9%

Bike/Walk 2.7% 2.7% 2.7%


SOCIAL EQUITY













Carpool 77% 77% 75%

Transit 22% 22% 22%



Carpool 71% 71% 71%

Transit 11% 11% 11%



Carpool 74% 74% 73%

Transit 16% 16% 16%



Carpool 71% 71% 70%

Transit 11% 11% 11%

Mobility population


Carpool 77% 77% 75%

Transit 19% 19% 19%


Appendix A




Delete


Delete

I-805 Mission

Valley Viaduct

SOCIAL EQUITY




Carpool 71% 71% 70%

Transit 12% 12% 12%



Carpool 75% 75% 74%

Transit 19% 19% 20%



Carpool 71% 71% 71%

Transit 11% 11% 11%

















Mid-City - UTC











Appendix A




Delete


Delete

I-805 Mission

Valley Viaduct


Otay Ranch - UTC









Appendix A


Section 3.10 - Capacity


Increase

Highway

Capacity Baseline

Increase

Arterial

Capacity



2



MOBILITY


Drive alone 24.7 26.0 25.8

Carpool 25.1 27.1 26.8

Transit 53.9 53.7 53.6


Drive alone 33.5 30.3 30.7

Carpool 34.9 31.6 32.0

Transit 12.7 12.5 12.5

7


by mode


Carpool 78% 72% 73%

Transit 13% 13% 13%


Transit 8% 8% 8%


RELIABILITY













Drive alone




Carpool




Truck




HEALTHY ENVIRONMENT






Drive alone 78.8% 78.3% 78.4%

Carpool 10.6% 10.8% 10.8%

Transit 8.5% 8.6% 8.6%

Bike/Walk 2.1% 2.2% 2.2%

3.10 - Capacity - 1 7/29/2011

Appendix A




Increase

Highway

Capacity Baseline

Increase

Arterial

Capacity

HEALTHY ENVIRONMENT


Drive alone 79.2% 78.9% 78.9%

Carpool 10.4% 10.5% 10.5%

Transit 8.1% 8.2% 8.2%

Bike/Walk 2.3% 2.4% 2.4%


Drive alone 46.2% 46.1% 46.1%

Carpool 49.8% 49.9% 49.9%

Transit 0.9% 0.9% 0.9%

Bike/Walk 3.2% 3.2% 3.2%


Drive alone 49.9% 49.9% 49.9%

Carpool 46.5% 46.5% 46.5%

Transit 0.9% 0.9% 0.9%

Bike/Walk 2.7% 2.7% 2.7%


SOCIAL EQUITY













Carpool 83% 77% 78%

Transit 21% 22% 22%



Carpool 76% 71% 72%

Transit 10% 11% 11%



Carpool 80% 74% 75%

Transit 16% 16% 16%



Carpool 76% 71% 72%

Transit 11% 11% 11%

Mobility population


Carpool 82% 77% 78%

Transit 18% 19% 19%

3.10 - Capacity - 2 7/29/2011

Appendix A




Increase

Highway

Capacity Baseline

Increase

Arterial

Capacity

SOCIAL EQUITY




Carpool 76% 71% 72%

Transit 11% 12% 12%



Carpool 82% 75% 76%

Transit 19% 19% 19%



Carpool 76% 71% 72%

Transit 10% 11% 11%

















Mid-City - UTC










3.10 - Capacity - 3 7/29/2011

Appendix A




Increase

Highway

Capacity Baseline

Increase

Arterial

Capacity


Otay Ranch - UTC








3.10 - Capacity - 4 7/29/2011

Appendix A


Section 3.11 - Trip Generation Discounts


Remove Trip

Generation

Discount Baseline

Double Trip

Generation

Discount



2



MOBILITY


Drive alone 26.2 26.0 26.0

Carpool 27.3 27.1 27.0

Transit 53.8 53.7 53.7


Drive alone 30.1 30.3 30.5

Carpool 31.3 31.6 31.8

Transit 12.5 12.5 12.5

7


by mode


Carpool 72% 72% 73%

Transit 13% 13% 13%


Transit 8% 8% 8%


RELIABILITY













Drive alone




Carpool




Truck




HEALTHY ENVIRONMENT






Drive alone 78.3% 78.3% 78.4%

Carpool 10.8% 10.8% 10.8%

Transit 8.7% 8.6% 8.6%

Bike/Walk 2.2% 2.2% 2.2%

3.11 - Trip Generation Discount - 1 7/29/2011

Appendix A




Remove Trip

Generation

Discount Baseline

Double Trip

Generation

Discount

HEALTHY ENVIRONMENT


Drive alone 78.8% 78.9% 78.9%

Carpool 10.5% 10.5% 10.5%

Transit 8.2% 8.2% 8.1%

Bike/Walk 2.4% 2.4% 2.4%


Drive alone 46.1% 46.1% 46.1%

Carpool 49.9% 49.9% 49.9%

Transit 0.9% 0.9% 0.9%

Bike/Walk 3.2% 3.2% 3.2%


Drive alone 49.9% 49.9% 49.9%

Carpool 46.5% 46.5% 46.5%

Transit 0.9% 0.9% 0.9%

Bike/Walk 2.7% 2.7% 2.7%


SOCIAL EQUITY













Carpool 76% 77% 77%

Transit 22% 22% 22%



Carpool 71% 71% 71%

Transit 11% 11% 11%



Carpool 73% 74% 75%

Transit 16% 16% 16%



Carpool 71% 71% 71%

Transit 11% 11% 11%

Mobility population


Carpool 76% 77% 77%

Transit 19% 19% 19%


Appendix A




Remove Trip

Generation

Discount Baseline

Double Trip

Generation

Discount

SOCIAL EQUITY




Carpool 70% 71% 71%

Transit 12% 12% 12%



Carpool 74% 75% 76%

Transit 19% 19% 19%



Carpool 71% 71% 72%

Transit 11% 11% 11%

















Mid-City - UTC











Appendix A




Remove Trip

Generation

Discount Baseline

Double Trip

Generation

Discount


Otay Ranch - UTC









appendix contents - sandag.org · mode choice splits total person-trip movements between zones into...

Documents