HomeMy WebLinkAboutSLO_ModelDocumentation_09_28_2020September 28, 2020 www.camsys.com
San Luis Obispo Citywide Travel Model
Model Documentation
prepared for
City of San Luis Obispo
prepared by
Cambridge Systematics, Inc.
with
Central Coast Transportation Consulting
report
San Luis Obispo Citywide Travel
Model
Model Documentation
prepared for
City of San Luis Obispo
prepared by
Cambridge Systematics, Inc.
1801 Broadway, Suite 1100
Denver, CO 80202
with
Central Coast Transportation Consulting
date
September 28, 2020
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
i
Table of Contents
1.0 Roadway Network .............................................................................................................................. 1-8
1.1 Context and Background ........................................................................................................... 1-8
1.2 Roadway Network Development ............................................................................................... 1-8
1.3 Roadway Network Structure ...................................................................................................... 1-8
1.3.1 Input and Output Networks ........................................................................................... 1-9
1.3.2 Facility Type ................................................................................................................ 1-12
1.4 Area Type ................................................................................................................................ 1-16
1.4.1 Existing Fringe, Urban, Suburban, and Rural Area Type Specification ...................... 1-17
1.5 Link Speeds ............................................................................................................................. 1-18
1.5.1 Estimating Link Speeds .............................................................................................. 1-18
1.5.2 Travel Time ................................................................................................................. 1-19
1.5.3 Link Capacities ............................................................................................................ 1-20
1.5.4 Freeways .................................................................................................................... 1-20
1.5.5 Collectors and Arterials ............................................................................................... 1-22
1.5.6 Resulting Capacity Model ........................................................................................... 1-24
1.5.7 Turn Prohibitions and Penalties .................................................................................. 1-24
1.6 Routable Network .................................................................................................................... 1-25
2.0 Traffic Analysis Zone Structure ...................................................................................................... 2-26
3.0 Land Use Dataset Development ..................................................................................................... 3-30
3.1 Data Sources ........................................................................................................................... 3-30
3.1.1 City of San Luis Obispo Land Use Data ..................................................................... 3-30
3.1.2 ACS 2011-2015 Dataset ............................................................................................. 3-30
3.1.3 2010 Census ............................................................................................................... 3-31
3.2 Model Dataset .......................................................................................................................... 3-31
Forecasting of Household Size and Income ............................................................... 3-32
3.3 Household Disaggregation Models .......................................................................................... 3-36
3.3.1 Household Size Disaggregation Model ...................................................................... 3-36
3.3.2 Household Income Disaggregation Model .................................................................. 3-37
3.3.3 TAZ-Level Bivariate Data ............................................................................................ 3-38
4.0 Trip Generation ................................................................................................................................. 4-39
4.1 California Household Travel Survey (CHTS) ........................................................................... 4-39
4.2 Trip Purposes ........................................................................................................................... 4-40
4.3 Production Rates ..................................................................................................................... 4-41
4.3.1 Income Groups ........................................................................................................... 4-41
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
ii
4.3.2 Cross Classified Production Rates ............................................................................. 4-42
4.4 Attraction Rates ....................................................................................................................... 4-43
4.4.1 Non-Home-Based Production Allocation Models ....................................................... 4-44
4.4.2 Trip Rate Factors and Adjustments ............................................................................ 4-45
4.5 CalPoly Trip Generation and Production Allocation ................................................................ 4-46
4.5.1 University Definition .................................................................................................... 4-46
4.5.2 Trip Types at CalPoly ................................................................................................. 4-48
4.5.3 Special Generator Survey Adaptation ........................................................................ 4-48
4.5.4 Employment and Enrollment Data .............................................................................. 4-48
Special Generator Values ........................................................................................... 4-49
Production Allocation .................................................................................................. 4-50
4.6 External Trips ........................................................................................................................... 4-53
4.6.1 Internal/External Trip Methodology ............................................................................. 4-54
4.6.2 External/External Trip Methodology ............................................................................ 4-56
4.6.3 Forecast External Trips ............................................................................................... 4-57
4.7 Trip Balancing .......................................................................................................................... 4-58
5.0 Trip Distribution ............................................................................................................................... 5-59
5.1 Context and Background ......................................................................................................... 5-59
5.2 Peak and Off-Peak Period Definitions ..................................................................................... 5-60
5.3 Roadway Network Shortest Path ............................................................................................. 5-60
5.3.1 Terminal Penalties ...................................................................................................... 5-60
5.4 Intrazonal Impedance .............................................................................................................. 5-61
5.5 Friction Factor Calibration ........................................................................................................ 5-61
5.5.1 Calibration Targets...................................................................................................... 5-61
5.5.2 Income Processing ..................................................................................................... 5-62
5.5.3 Calibration Process ..................................................................................................... 5-62
5.5.4 Calibration Results ...................................................................................................... 5-63
5.6 K Factors .................................................................................................................................. 5-67
6.0 Mode Choice ..................................................................................................................................... 6-69
6.1 Background .............................................................................................................................. 6-69
6.2 Mode Choice Model Structure ................................................................................................. 6-69
6.3 Mode Choice Model Coefficients ............................................................................................. 6-70
6.4 Mode Choice Model Calibration Targets ................................................................................. 6-72
6.4.1 Target Trips and Shares by Mode and Purpose ......................................................... 6-72
6.4.2 Transit Trips ................................................................................................................ 6-72
7.0 Time of Day ......................................................................................................................................... 7-0
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
iii
8.0 Trip Assignment ................................................................................................................................. 8-2
8.1 Parking Garage Allocation Model .............................................................................................. 8-2
8.2 Traffic Assignment Algorithms ................................................................................................... 8-2
8.3 Closure Criteria .......................................................................................................................... 8-3
8.4 Impedance Calculations............................................................................................................. 8-4
8.5 Volume-Delay Functions ............................................................................................................ 8-4
8.6 Transit Assignment .................................................................................................................... 8-5
8.7 Bicycle Assignment .................................................................................................................... 8-6
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
v
List of Tables
Table 1.1. Input Network Link Fields .............................................................................................................. 1-9
Table 1.2. Input Network Node Fields .......................................................................................................... 1-10
Table 1.3. Output Network Link Fields ......................................................................................................... 1-11
Table 1.4. Facility Type Classifications ........................................................................................................ 1-12
Table 1.5. Area Type Categories .................................................................................................................. 1-16
Table 1.6. Speed Limit to Freeflow Speed Conversion Factors ................................................................... 1-19
Table 1.7. Centroid Connector and Transit Freeflow Speeds ...................................................................... 1-19
Table 1.8. Ideal and Adjusted Capacities for Freeways and Expressway Based on HCM 2000 ................. 1-22
Table 1.9. Link Capacity Adjustment Factors and Resulting Capacity ......................................................... 1-23
Table 1.10. Roadway Capacities (Vehicles per Hour per Lane, LOS E) ...................................................... 1-24
Table 3.1. Source of Input Socioeconomic and Land Use Data ................................................................... 3-31
Table 3.2. Model Land Use and Socioeconomic TAZ Input Data ................................................................ 3-31
Table 3.3. Selected Household Size and Income Summary Data ............................................................... 3-33
Table 3.4. Income Group Definitions ............................................................................................................ 3-37
Table 3.5. Bivariate Household Distribution for San Luis Obispo County .................................................... 3-38
Table 4.1. CHTS Weighted and Expanded Trips by Purpose for San Luis Obispo County ......................... 4-40
Table 4.2. Trip Purpose by Trip Origin and Destination Activities ................................................................ 4-41
Table 4.3. Income Categories....................................................................................................................... 4-41
Table 4.4. Household Trip Production Rates – HBW ................................................................................... 4-42
Table 4.5. Household Trip Production Rates – HBS .................................................................................... 4-42
Table 4.6. Household Trip Production Rates – HBO .................................................................................... 4-42
Table 4.7. Household Trip Production Rates – WBO ................................................................................... 4-43
Table 4.8. Household trip Production Rates – OBO ..................................................................................... 4-43
Table 4.9. Household Trip Production Rates – Total ................................................................................... 4-43
Table 4.10. Refined City of SLO Travel Model Attraction Rates .................................................................. 4-44
Table 4.11. Trip Production Rate Adjustment Factors ................................................................................. 4-45
Table 4.12. Trip Attraction Rate Adjustment Factors ................................................................................... 4-46
Table 4.13. CalPoly Activity Allocation by Zone ........................................................................................... 4-47
Table 4.14. CalPoly Employment and Enrollment ........................................................................................ 4-49
Table 4.15. University Special Generator Values ......................................................................................... 4-49
Table 4.16. Dedicated Off Campus Student Housing .................................................................................. 4-51
Table 4.17. IE Trip Attraction Rates ............................................................................................................. 4-55
Table 4.18. Base Year External Station Data ............................................................................................... 4-56
Table 4.19. Base Year (2016) 24-hour EE Vehicle Trip Table ..................................................................... 4-57
Table 4.20. Forecast Year External Station Data ......................................................................................... 4-57
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
vi
Table 4.21. Future Year (2040) 24-hour EE Vehicle Trip Table ................................................................... 4-58
Table 5.1. Terminal Penalties by Area Type ................................................................................................ 5-61
Table 5.2. Coincidence Ratios for Calibrated Trip Time Distribution Curves ............................................... 5-64
Table 5.3. Friction Factors for All Purposes ................................................................................................. 5-67
Table 5.4. K Factors ..................................................................................................................................... 5-67
Table 6.1. FTA Mode Choice Model Coefficient Guidelines ......................................................................... 6-70
Table 6.2. Mode Choice Model Coefficients ................................................................................................. 6-71
Table 6.3. Initial Calibration Target Trips by Mode and Purpose for City of San Luis Obispo Travel
Model .................................................................................................................................. 6-72
Table 6.4. Revised Calibration Target Trips ................................................................................................. 6-72
Table 6.5. Estimated Shares of Transit Trips by Trip Purpose from 2007 SLO Transit On-board Survey .. 6-73
Table 6.6. SLO Transit Boarding Counts ...................................................................................................... 6-74
Table 6.7. Transit Targets (2016) ................................................................................................................. 6-74
Table 7.1. Peak Period Definitions ................................................................................................................. 7-0
Table 7.2. Time of Day Factors ...................................................................................................................... 7-1
Table 7.3. Pre-distribution Time of Day Factors ............................................................................................. 7-1
Table 7.4. Pre-assignment Time of Day Factors ............................................................................................ 7-1
Table 8.1. Parking Garages ............................................................................................................................ 8-2
Table 8.2. Volume Delay Parameters Alpha and Beta ................................................................................... 8-5
Table 8.3. Modeled and Observed Transit Boardings by Route .................................................................... 8-6
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
vii
List of Figures
Figure 1.1. Facility Type- Entire Modeled Area ............................................................................................ 1-13
Figure 1.2. Highway Network Facility Type- City SOI .................................................................................. 1-14
Figure 1.3. CBD Area Type Definition .......................................................................................................... 1-17
Figure 1.4. Area Types within City SOI ........................................................................................................ 1-18
Figure 2.1. Traffic Analysis Zones (Countywide) ......................................................................................... 2-27
Figure 2.2. Traffic Analysis Zones (Citywide) ............................................................................................... 2-28
Figure 2.3. Traffic Analysis Zones (CBD) ..................................................................................................... 2-29
Figure 3.1. Median Income Forecasting Flowchart ...................................................................................... 3-34
Figure 3.2. Average Household Size Forecasting Flowchart ....................................................................... 3-35
Figure 3.3. Household Size Disaggregation Curves .................................................................................... 3-37
Figure 3.4. Household Income Disaggregation Model ................................................................................. 3-38
Figure 4.1. CalPoly Campus Zones .............................................................................................................. 4-47
Figure 4.2. University Production Allocation ................................................................................................. 4-52
Figure 4.3. External Station Locations .......................................................................................................... 4-54
Figure 5.1. Countywide Trip Time Distribution Curves by Trip Purpose ...................................................... 5-62
Figure 5.2. HBW Trip Time Distribution Curve (Countywide) ....................................................................... 5-64
Figure 5.3. HBS Trip Time Distribution Curve (Countywide) ........................................................................ 5-65
Figure 5.4. HBO Trip Time Distribution Curve (Countywide) ....................................................................... 5-65
Figure 5.5. WBO Trip Time Distribution Curve (Countywide) ....................................................................... 5-66
Figure 5.6. OBO Trip Time Distribution Curve (Countywide) ....................................................................... 5-66
Figure 5.7. K Districts ................................................................................................................................... 5-68
Figure 6.1. San Luis Obispo Mode Choice Model Nesting Structure ........................................................... 6-69
Figure 8.1. Bike Comfort Levels ..................................................................................................................... 8-7
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-8
1.0 Roadway Network
1.1 Context and Background
The roadway network contains basic input information for use in the travel demand model and represents
real-world conditions for the 2016 base year. The roadway networks are used in the model to distribute trips
and route automobile trips. The networks in the GIS environment used by the model are databases in which
all kinds of information can be stored and managed. In addition, the networks provide a foundation for
system performance analysis including vehicle miles of travel, congestion delay, level of service, and other
performance criteria. This chapter provides a description of the network attributes and lookup tables for the
roadway networks. The assumptions and parameters identified herein were identified during the
development of the model’s 2016 base year network, but they generally apply to all model year networks.
The roadway network is a GIS-based representation of the street and highway system in the City of San Luis
Obispo and, at a reduced level of detail, San Luis Obispo County. It operates both as an input database
containing roadway characteristics (such as facility type, number of lanes, area type, etc.) and as a data
repository that can be used to store and view travel model results. The roadway network is one of the
foundational components of the travel model as it serves to represent the supply side of the travel
demand/transportation system relationship. As such, the establishment and review of detailed network
attribute data was very important to the model’s development.
1.2 Roadway Network Development
Two primary sources were available for developing the 2016 base year network: The 2008 roadway network
from the City’s previous travel model and the 2010 network from the most recent model of the San Luis
Obispo Council of Governments (SLOCOG). These networks vary in several ways as shown in Table 1.1.
The right-most column of Table 1.1 contains recommendations for addressing the differences.
SLOCOG’s most recent 2010 model roadway network is used as the basis for the outside of the Sphere of
Influence (SOI). The SLOCOG network outside the SOI was simplified by removing all facilities below minor
arterials. The links in the SLOCOG network that fall within the boundaries of the SOI were then removed and
replaced with the City of San Luis Obispo 2008 network. The two networks were then stitched at the
boundary to ensure network connectivity. Once the two networks were combined, the relevant attributes
necessary for travel demand modeling were added to the network file.
1.3 Roadway Network Structure
The roadway network is structured to contain data for multiple timeframes. The roadway network delivered
with the SLO Citywide Travel Model contains the 2016 base year network, a 2040 forecast year roadway
network, and a buildout network. The model includes the capability to represent the base year, existing plus
committed networks, plan forecast networks, interim horizon year networks, and any other network scenarios
that are desired within a single network database. In addition, the network is structured so that localized
alternatives can be represented within the same file. These alternatives can be activated and deactivated
based on the year of analysis and infrastructure scenario desired using the scenario management system
that forms the basis of the travel model user interface.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-9
1.3.1 Input and Output Networks
The roadway network file contains travel model input data, and it also acts as a repository for both
intermediate (e.g., speed feedback data) and final (e.g., traffic volumes) model data. For this reason, a
separate output model network is created for each model scenario. This output network is created by making
a copy of the input network and then modifying this network to contain data and results specific to each
model run. This copy of the roadway network is created and modified automatically by a network initialization
step when the travel model is run. Required attributes present on the input network link and node layers are
listed in Table 1.1 and Table 1.2. Additional attributes that are created on the output network are listed in
Table 1.3.
Table 1.1. Input Network Link Fields
Field Name Description Comments
ID TransCAD Unique ID Maintained automatically by
TransCAD
Length Link Length in miles Maintained automatically by
TransCAD
Dir Link Direction of Flow Direction of Flow
NAME Street Name
HWY_LABEL Route number for state and US
highways
Dir_yy Scenario-Specific Direction Field yy represents a two-digit year code
e.g., 16, 40, BOUT) or the string “AL”
FT_yy Facility Type for year yy (See Table
1.4 for definition)
AT_yy Area Type for year yy (See Table 1.5
for definition)
AB_LANE_yy Number of Lanes for year yy
AB_LANE_yy
SPLM_yy Speed Limit for year yy.
AB_FBAM_yy Fields used to store speed feedback
results – not typically modified by the
user.
AL” versions of these fields are not
present in the network.
BA_FBAM_yy
AB_FBOP_yy
BA_FBOP_yy
ALT Primary Alternative Number
ALT2 Secondary Alternative Number
ADJ_CNT Adjusted Traffic Count representing
2016 conditions.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-10
Table 1.2. Input Network Node Fields
Field Name Description Comments
ID TransCAD Unique ID Maintained automatically by
TransCAD
ZONE Traffic Analysis Zone Number Populated only for centroid nodes
including external station nodes). Null
for all non-centroid nodes.
PNR_yy Park and Ride Identifier (1 for park
and ride nodes, 0 or null for other
nodes)
yy represents a two-digit year code
e.g., 16, 40) or the string “AL”
ModelNode Variable indicating nodes kept in a link
consolidation exercise performed
during model development.
This value can be set to 1 or null on all
nodes and does not effect model
results.
TimePointNode Route time point These fields contain notes used in
building the route system and are not
used by the model. NewNode Indicates nodes added for transit
SOI Identifies nodes within the model
sphere of influence
INT_ID Intersection ID for turn movement
reporting
Turn movement volumes will be
stored for nodes identified by this field.
RPT_ID Intersection report ID ID number for SR-1 MIS reporting
purposes (not required; not used
directly by the model)
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-11
Table 1.3. Output Network Link Fields
Field Name Description Comments
FT Facility Type for selected year and alternative(s) These fields are populated based
on data in fields identified with a
yy” suffix in the input network. AT Area Type for selected year and alternative(s)
AB_LANE Number of Lanes for selected year and alternative(s)
AB_LANE
SPLM Speed Limit for selected year and alternative(s)
AB_FBAM Speed feedback data for selected year
BA_FBAM
AB_FBOP
BA_FBOP
Mode Mode field used to identify walk and drive access links
FF_SPD Calculated freeflow travel speed Based on speed limit
AB_OPSPD Calculated directional off-peak and peak congested
travel speed
Populated based on speed
feedback results
BA_OPSPD
AB_PKSPD
BA_PKSPD
FF_TIME Calculated freeflow travel time
AB_OPTIME Calculated directional off-peak and peak congested
travel time
BA_OPTIME
AB_PKTIME
BA_PKTIME
AB_OPTRSPD Directional off-peak transit speed and time
BA_OPTRSPD
AB_OPTRTIM
BA_OPTRTIM
AB_PKTRSPD Directional peak transit speed and time
BA_PKTRSPD
AB_PKTRTIM
BA_PKTRTIM
LANE_CAP Calculated per-lane capacity Retrieved from lookup table
AB_CAP Calculated directional capacity
BA_CAP
ALPHA Volume delay parameters Retrieved from lookup table
BETA
The model’s directory structure allows multiple model output directories to exist as subdirectories to a single
input directory. Each time the travel model is run, files located in the input directory are not modified by
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-12
model macros. Instead, if a file is to be modified it will be copied to an output directory; only the copy is
modified.
1.3.2 Facility Type
The facility type of each roadway link reflects its role in the system of streets and highways. Facility type is
used in the travel model as a foundational element for describing the roadway system using mathematical
relationships that correspond to the function, operation, and characteristics of the roadway. Roadway speed,
capacity, and volume-delay characteristics are dependent on facility type.
Most local roads are not included in the travel model facility type scheme, except for the following two cases:
Local streets used to accommodate local bus routes, and
Roadways identified as Urban Local roads.
Traffic is assigned to urban local roadways, but is not assigned to roadways included solely for purposes of
the transit route system. The facility types included in the model are listed in Table 1.4.
Table 1.4. Facility Type Classifications
Code Facility Type
1 Freeway
2 Not Used (reserved for Expressway)
3 Principal Arterial
4 Minor Arterial
5 Collector
6 Ramp
7 Urban Local
9 Walk Access Connector (Present only in the output transit line layer)
11 Highway (Outside SOI)
12 Arterial (Outside SOI)
13 Rural Arterial (Outside SOI)
50 Centroid Connector
51 Transit Local (roadway traffic is not assigned to these links)
60 Bike Only Link
null Inactive Link (does not exist in the specified network)
The numbering system used for facility types generally increases as the primary function of roadways moves
from mobility toward access. While freeways function to provide mobility rather than direct access to homes
and businesses, local streets exist almost exclusively to provide access while providing limited mobility at low
speeds. Maps demonstrating facility type on roadways included in the model are included as Figure 1.1 and
Figure 1.2.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-13
Figure 1.1. Facility Type- Entire Modeled Area
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-14
Figure 1.2. Highway Network Facility Type- City SOI
Descriptions of the facility types are provided below.
Freeway – A divided, restricted access facility with no direct land access and no at-grade crossings or
intersections. Freeways are intended to provide the highest degree of mobility serving higher traffic
volumes and longer-length trips. Freeways in Washtenaw County have 4 or 6 travel lanes (2 or 3 in each
direction). All Interstate facilities are freeways. Freeways in Washtenaw County include I-94, US-23, and
M-14. These are represented by the NFC codes for interstate freeways and other freeways.
Ramp – A link that provides connections between freeways and other non-freeway roadway facilities. On
freeway to non-freeway ramps, traffic usually accelerates or decelerates to or from a stop. Therefore, the
speed on freeway to arterial ramps is often coded slower than the ramp speed limit. Most ramp facilities
were identified using the “CSBRANCH” attribute in the Michigan Geographic Framework, but some have
been identified manually. Facility types shown in Attachment 2.1 represent manual identification of ramp
facilities.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-15
Principal Arterial – These permit traffic flow through and within urban areas and between major
destinations. These are of great importance in the transportation system since they provide local land
access by connecting major traffic generators, such as central business districts and universities, to
other major activity centers. Containing 4 travel lanes, principal arterials carry a high proportion of the
total urban travel on a minimum of roadway mileage. They typically receive priority in traffic signal
systems (i.e., have a high level of coordination and receive longer green times than other facility types),
have turn bays at intersections, include medians or center turn lanes, and sometimes contain grade
separations and other higher-type design features. State and U.S. highways are typically designated as
principal arterials unless they are classified as freeways.
Minor Arterial – Minor arterials collect and distribute traffic from principal arterials, freeways, and
expressways to streets of lower classification and, in some cases, allow traffic to directly access
destinations. They serve secondary traffic generators, such as community business centers,
neighborhood shopping centers, multifamily residential areas, and traffic between neighborhoods.
Access to land use activities is generally permitted, but should be consolidated, shared, or limited to
larger-scale users. Minor arterials generally have slower speed limits than major arterials, may or may
not have medians and center turn lanes, and receive lower signal priority than other facility types (i.e.,
are only coordinated to the extent that major arterials are not disrupted and receive shorter green times
than major arterials).
Collector Street – Collectors provide for land access and traffic circulation within and between
residential neighborhoods and commercial and industrial areas. They distribute traffic movements from
these areas to the arterial streets. Except in rural areas, collectors do not typically accommodate long
through trips and are not continuous for long distances. The cross-section of a collector street may vary
widely depending on the scale and density of adjacent land uses and the character of the local area. Left
turn lanes sometimes occur on collector streets adjacent to nonresidential development. Collector streets
should generally be limited to two lanes, but sometimes have 4-lane sections. In rural areas, major
collectors act similarly to minor arterials, while rural minor collectors fit more closely with the
characterizations described here. Minor collectors do not exist in urban areas.
Centroid Connector – These facilities represent local and/or residential street systems that are too
detailed for modeling purposes. Centroid connectors are usually not coded along actual streets, but
rather they are the means through which the trip and other data at the traffic analysis zone (TAZ) level
are attached to the street system.
Walk Access Connector – This facility type is included to provide pedestrian access directly to a transit
stop or to the collector/arterial roadway network where vehicle access is not possible. The transit
network will utilize the roadway network to provide walk access to transit, but this does not in itself
accurately represent walk access. This facility type ensures that walk access to transit is represented by
the shortest realistic distance between a zone centroid and nearby transit stops. Walk access connectors
are not present in the input network, but are instead created automatically when the model is run.
Bikes Allowed – Local streets are not typically represented in regional travel models except where
required to support the transit route system. Some local streets in the urban core of San Luis Obispo are
included as “Urban Local” in the focused City model to better represent detailed access and traffic flow
conditions. Remaining local streets are retained in the model network as bikes only facilities and have
been assigned a facility type of 60. When the traffic assignment is run, links with a facility of 60 are
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-16
ignored by the pathfinding and traffic assignment algorithms. However, they are considered in the bike
assignment and skims.
Transit Local – This facility type is included to support the transit route system and transit network. The
facility type on local streets that serve transit routes is 51 resulting in the model allowing transit routes to
utilize these links. Transit local links are not used in the trip distribution or traffic assignment procedures.
1.4 Area Type
Area type is an attribute assigned to each traffic analysis zone (TAZ) and roadway and is based on the
activity level and character of each zone. Terminal times, speed-limit to freeflow speed conversion factors,
roadway capacity, and volume-delay characteristics are dependent on area type. Area type is first defined at
the TAZ level based on activity characteristics and then transferred to the roadway network.
Area type is an attribute that can vary with time. Therefore, it was important that area type definitions are
specified in a manner that can be updated for future conditions based on available forecast data. While area
type definitions based on external information such as corridor characteristics (e.g., commercial vs.
residential) or the U.S. Census urbanized area boundary are useful in defining existing area type, this
information is not very useful in defining future year area types. Area type definitions were therefore specified
so that area type forecasts can be developed using forecast socioeconomic data.
The area type guidelines presented herein were developed based on a visual analysis of aerial photography
and street characteristics and input from City staff. The SLOCOG model does not define different area types
within the City of San Luis Obispo SOI. The area type categories defined for the SLO Citywide Travel Model
are listed in Table 1.5.
Table 1.5. Area Type Categories
ID Area Type
1 Central Business District (CBD)
2 CBD Fringe (dense area surrounding CBD)
3 Urban
4 Suburban
5 Rural
6 Outside SOI
Note: For each TAZ, the most dense non-CBD area type is applied for which at least one of the criteria is
met.
Due to the more compact and dense nature of the City of San Luis Obispo Central Business District, it has
been included as a separate area type. The definition of the CBD area type is based on a rough
approximation of the San Luis Obispo Downtown Business Association boundary. An approximation of the
boundary was necessary to accommodate TAZ boundaries. The area assigned the CBD area type is shown
in Figure 1.3. Typically, the definition of the CBD area type will not change in forecast year datasets.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-17
Figure 1.3. CBD Area Type Definition
1.4.1 Existing Fringe, Urban, Suburban, and Rural Area Type Specification
Specification of existing area types was performed by reviewing street centerline data and aerial
photography. The area surrounding the CBD that includes a higher density of buildings, and a denser street
grid has been classified as urban or fringe, with urban including the area with a higher level of density and
greater commercial uses. The suburban area type was assigned to areas with lower building and street
density. Undeveloped areas, or areas with very sparse development, were identified as rural. Resulting area
type definitions are shown in Figure 1.4.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-18
Figure 1.4. Area Types within City SOI
1.5 Link Speeds
Network speeds are used in trip distribution to distribute trips throughout the region, in mode choice to
determine mode-specific travel times, and in trip assignment to route traffic on the roadway network.
Link speeds represent average travel time, including intersection delay, needed to traverse the distance of a
link with little or no traffic (i.e., no congestion effects). These speeds are generally similar to the speed limit
and are calculated as a function of the speed limit, functional class, and area type. Freeflow speeds are
typically lower than the speed limit to account for intersection delay on arterials, collectors, and ramps. On
other facility types, the speed limit and freeflow speed may be the same.
1.5.1 Estimating Link Speeds
Speed limit data is available for the model-level roadway links in the network. This speed limit data can be
used in combination with corridor travel time survey data to approximate a freeflow speed on each network
link. Because the travel model freeflow speed must include intersection delay experienced in uncongested
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-19
conditions, freeflow speed is typically lower than the posted speed limit. The relationship between speed limit
and freeflow speed has been observed to vary by characteristics such as facility type and area type.
No local data is available to facilitate the development of a model relating posted speed limit, facility type,
and area type to freeflow speed. To facilitate estimation of such a model using local data, a comprehensive
and current travel time survey would be necessary. Freeflow speed factors from the previous version of the
model were initially carried forward to the updated model. Adjustments were made during the model
calibration and validation process to better match count volumes on arterials and freeways. The model
freeflow factors are shown in Table 1.6.
Table 1.6. Speed Limit to Freeflow Speed Conversion Factors
ID Class CBD Fringe Urban Suburban Rural Outside
SOI
1 Freeway 1 1 1 0.95 0.95 0.95
3 Principal Arterial 0.98 0.97 0.92 0.9 0.95 0.95
4 Minor Arterial 0.97 0.95 0.9 0.9 0.85 0.85
5 Collector 0.95 0.92 0.88 0.9 0.75 0.75
7 Urban Local 0.75 0.75 0.75 0.75 0.75 0.75
6 Ramp 0.92 0.9 0.85 0.9 0.75 0.75
11 Highway (Outside SOI) 0.85 0.85 0.85 0.9 0.95 0.95
12 Arterial (Outside SOI) 0.85 0.85 0.85 0.9 0.95 0.95
13 Rural Arterial (Outside SOI) 0.85 0.85 0.85 0.9 0.95 0.95
50 Centroid Connectors 1 1 1 1 1 1
51 Transit Links 1 1 1 1 1 1
For centroid connectors and transit only links, values in Table 1.7 are used if speed limit data is not
populated on the network. Speed limits must be provided for all other roadway links to successfully run the
travel model.
Table 1.7. Centroid Connector and Transit Freeflow Speeds
ID Functional Class Area Type
CBD Fringe Urban Suburban Rural Outside
SOI
50 Centroid Connector 15 20 20 20 35 40
51 Transit Local 15 15 15 15 15 15
1.5.2 Travel Time
Freeflow and congested speeds in the roadway network are used to compute travel time for each link. Travel
time is computed in minutes.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-20
1.5.3 Link Capacities
Traffic assignment, especially capacity constrained traffic assignment, requires accurate roadway capacity
values. Capacity is used in the model to measure congestion and to determine route diversion due to
congestion. This is accomplished through the use of volume-delay equations that will be defined and applied
in the traffic assignment model.
In the model, per-lane capacity values are retrieved from a lookup table based on the facility type and area
type of each link in the roadway network. This approach eliminates opportunities for error in defining
capacities at the link level and enforces consistent application of capacity values. Hourly per-lane capacities
are retrieved from a lookup table that is stored in an Access database. These hourly lane capacities are used
in combination with the number of lane information present on the network to define hourly directional
capacity.
The Highway Capacity Manual (HCM or HCM 2000)1 provides guidance on the definition of roadway
capacity. The HCM provides link-level capacity guidelines for freeways and rural highways, but does not
provide detailed link-level capacity guidelines for urban and suburban collector and arterial streets.
Therefore, HCM intersection capacity was used in place of link capacity to develop capacities for these other
facilities.
1.5.4 Freeways
Capacity guidelines for freeways and expressways are provided in Chapters 21 and 23 of HCM 2000.
Unadjusted, or ideal, per-lane capacities based on freeflow speed are provided. These capacities must then
be adjusted for various conditions. The conditions for which adjustments can be applied are described below.
Heavy Vehicle Adjustment Factor – The heavy vehicle adjustment factor accounts for passenger car
equivalents for trucks, buses, and recreational vehicles. HCM 2000 recommends default values of 10%
heavy vehicles in rural areas and 5% heavy vehicles in non-rural areas unless additional data is
available. However, for regional modeling purposes, a heavy vehicle adjustment factor of 1.0 has been
used.
Driver Population Factor – The driver population factor represents the familiarity of drivers with
roadway facilities. Because the model represents traffic on a typical weekday when school is in session,
normal driver familiarity was assumed. Driver population factors are typically used for weekend
conditions or in areas with a high amount of tourist/recreational activity.
Peak Hour Factor – A peak hour factor (PHF) represents the variation of traffic volumes within an hour.
Default values of 0.88 for rural area types and 0.92 for non-rural area types were applied2.
The HCM suggests adjusting flow rate (traffic volume) according to the equation below.
1 Highway Capacity Manual. Transportation Research Board, 2000.
2 HCM 2000, p. 13-11
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-21
Where: 15-min
passenger equivalent flow rate (
pc/hr/ln)
hourly volume (
veh/hr) peak-
hour factor
number of lanes heavy-vehicle adjustment factor driver population factor For travel model application, it is
more practical to adjust capacity than vehicle flow rate. This eliminates the need to adjust vehicle trip
tables prior to and subsequent to traffic assignment. By replacing with ideal capacity (and V with
hourly capacity (C), the equation above can be used to adjust ideal capacity to effective hourly capacity. Furthermore,
it is useful to consider capacity on a per lane (veh/hr/ln) basis, allowing number of lane calculations
to be applied at the link level. The resulting equation can be used to compute per
lane capacity for freeways and expressways.
The
equation below was used to compute
hourly capacities for rural
and freeway facilities.
Where:
Ideal (unadjusted) capacity (pc/hr/ln) link capacity (veh/hr) peak-hour factor
heavy-vehicle adjustment factor driver population factor Ideal capacities are defined in HCM according to
freeflow speed3. Ideal capacities based on typical freeflow speeds are shown in Table 1.8, along with
adjusted capacities computed using the methodology described above. Adjusted capacities have been rounded to 100 vehicles per hour. It
is noted that these calculations result in a lower capacity on rural freeways than on suburban
and urban freeways. This is due to the difference in peaking factors associated with rural facilities.
In practice, it is unlikely that rural freeway facilities will reach capacity. Instead, rural facilities
are likely to become suburban or urban facilities before
nearing capacity. As this occurs peaking
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-22
Table 1.8. Ideal and Adjusted Capacities for Freeways and Expressway Based on
HCM 2000
Facility
Type
Area
Type
Freeflow
Speed
mph)
Ideal
Capacity
Upper Limit
LOS E,
pc/h/ln)
PHF FHV FP Adjusted
Capacity
Upper Limit
LOS E,
pc/h/ln)
Freeway Rural 70 2,400 0.88 1 1 2,100
Freeway Suburban 70 2,400 0.92 1 1 2,200
Freeway Urban 65 2,350 0.92 1 1 2,200
1.5.5 Collectors and Arterials
For non-rural arterial and collector streets, HCM recommends identifying capacity on an intersection basis,
with the intersection with the lowest capacity determining the overall arterial link capacity. The link capacity at
each intersection can be computed using the equation below.
0
Where: Capacity 0 = base saturation flow per lane (pc/h/
ln) – assumed at 1900 number of lanes in
lane group (intersection approach lanes, not bid-block lanes) adjustment factor
for lane width– assumed at 1.0 adjustment
factor for heavy vehicles in traffic stream assumed at 1.0 adjustment factor for
approach grade – assumed at 1.0 adjustment factor for existing of
a parking lane and parking activity –
assumed at 1.0 adjustment factor for blocking
effect of local busses – assumed at 1.0 adjustment factor for
CBD area type adjustment factor for lane utilization – assumed at 0.
95 adjustment factor for left turns in lane group – assumed
at 1.0 adjustment factor for right turns in lane group –
assumed at 1.0 pedestrian adjustment
factor for left-turn movements –
assumed at 1.0 pedestrian-bicycle adjustment factor for right turn movements – assumed at 1.0
peak-hour factor – assumed at 0.92 effective green time per cycle The equation above accounts for
specific details that are not practical to maintain in a regional travel model. Therefore, a number of adjustment factors
can be assumed constant or set to 1.0 for all cases. Some variables that have been
set to 1.0, such as lane width, parking, turns, bus blocking, and pedestrian/bicycle effects, are instead captured
in the area type adjustment. Other variables can be approximated based on the facility type and area type
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-23
addressed by an intersection widening factor that varies by facility type and accounts for the presence of left
and right turn lanes at intersection approaches.
The simplified equation below is useful in a regional travel modeling context. Assumed values for adjustment
factors that vary by facility type and area type are shown in Table 1.9, along with the resulting capacity
values.
0
Where: Capacity 0 = base saturation flow
per lane (pc/h/ln) – assumed at 1900 number
of through (mid-block) lanes,
excluding center turn lanes adjustment factor for area
type adjustment factor for lane utilization –
assumed at 0.95
peak-hour factor – assumed at
0.92 effective green time per cycle
adjustment factor for
intersection widening Table 1.9. Link Capacity Adjustment Factors and
Resulting Capacity FT AT Capacity Expressway
CBD 0.90 0.55 1.30 1,100
Urban / Fringe 0.97 0.55 1.30 1,200
Suburban 0.99 0.55 1.30 1,200
Principal Arterial CBD 0.76 0.45 1.
30 740 Urban / Fringe 0.95 0.45 1.30
920 Suburban 0.99 0.45 1.30
960 Minor Arterial CBD 0.76 0.45
1.15 650 Urban 0.95 0.42 1.
15 760 Suburban 0.99 0.42 1.15
790 Collector CBD 0.75 0.45 1.
05 590 Urban / Fringe 0.95 0.41 1.
05 680 Suburban 0.99 0.41 1.05
710 Local CBD 0.74 0.45 1.
00 550 Urban / Fringe 0.95 0.40 1.00 630 Suburban 0.99 0.40 1.00
660 Presence of a center left turn lane, median, or left turn prohibitions can also impact
link capacity. The intersection widening factors assumed above account for the presence of frequent left turn
lanes or medians on principal arterials, with occasional left turn lanes and medians on minor
arterials. The Corridor MPO roadway network contains a specific variable that identifies roadway corridors where medians or center left turn
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-24
by this variable. To account for center left turn lanes, the number of lanes used to compute total directional
flow is adjusted as follows:
Principal Arterial:
Left turn lane present: Add 0.25 lanes (0.125 lanes in each direction)
No left turn lane present: Subtract 0.5 lanes (0.25 lanes in each direction)
Minor Arterial:
Left turn lane present Add 0.5 lanes (0.25 lanes in each direction)
No left turn lane present: Subtract 0.25 lanes (0.125 lanes in each direction)
No adjustments are made on expressway, local, or collector facilities.
1.5.6 Resulting Capacity Model
The calculations above provide capacity values that can be applied based on the facility type, area type,
number of lanes, and center turn lane presence of each link in the network. The model begins by applying
the hourly lane capacities shown in Table 1.10.
Table 1.10. Roadway Capacities (Vehicles per Hour per Lane, LOS E)
1.5.7 Turn Prohibitions and Penalties
There are two primary types of turn penalties that can be included in the network. Specific (localized) turn
penalties or prohibitions represent known turn penalties or prohibitions at individual locations. Global turn
Functional
Classification
CBD Fringe Urban Suburban Rural Outside SOI
1 Freeway 2,100 2,200 2,200 2,200 2,100 2,100
3 Principal Arterial 740 920 920 960 1,162 1,162
4 Minor Arterial 650 760 760 790 956 956
5 Collector 590 680 680 710 850 850
7 Urban Local 590 680 680 710 850 850
6 Ramp 650 750 750 800 800 800
11 Highway (Outside SOI) 740 920 920 960 1,162 1,162
12 Arterial (Outside SOI) 740 920 920 960 1,162 1,162
13 Rural Arterial (Outside
SOI)
740 920 920 960 1,162 1,162
50 Centroid Connectors 10,000 10,000 10,000 10,000 10,000 10,000
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
1-25
penalties represent the generally increased amount of time required to make a left or right turn rather than
travel straight through an intersection.
Inclusion of specific turn penalties in the roadway network was not recommended for the SLO Citywide
Travel Model. When used, individual turn penalties often represent existing congestion at particular
intersections that may or may not exist in the future, especially if operational improvements are made. While
it is possible to adjust specific turn penalties for future conditions based on planned intersection or signal
timing improvements, this task is beyond the capability or desire of most jurisdictions. Not only can
maintenance of specific turn penalties be a time consuming task, but detailed plans for intersections and
traffic signal timings in a 25-year forecast scenario do not often exist. As a result, turning penalties are not
used.
Turn prohibitions, meanwhile, are a valuable addition to a travel model. Turn prohibitions are used in
locations where turns are prohibited entirely. Specific turn prohibitions often include left turns or prohibited
movements in interchanges. When future changes are made to the roadway network, the practitioner should
be aware of the need to potentially add or remove turning prohibitions.
1.6 Routable Network
Many functions in TransCAD require the creation of a routable network file, identified by a “.net” extension.
Of particular interest for the SLO Citywide Travel Model, pathbuilding/skimming and traffic assignment
procedures require a routable network that contains only model-level links (i.e., no transit-only or local links).
Length and travel time information for each link is stored in the routable network file, as are turn prohibitions.
Specific turn prohibitions are initially stored in a separate file that is referenced when creating the routable
network. An appropriate routable network file is created during the automated network initialization step.
Routable network files are also required when building a transit network and performing interactive
pathbuilding, but these can be created using the TransCAD interface designed for these purposes.
The routable network used for pathbuilding and traffic assignment includes only model-level links, excluding
all bike and transit only links. To accomplish this, a selection set is created containing only links where facility
type is less than or equal to 50 (FT < 50). A routable network file is then created to include only these links.
The routable network file also contains information about centroid connectors. This prevents the pathbuilder
and traffic assignment algorithms from routing trips through centroids. This is accomplished automatically in
the model by creating a selection set where facility type is equal to 50 (FT = 50) and identifying these links as
centroid links in the routable network file.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
2-26
2.0 Traffic Analysis Zone Structure
Traffic analysis zones (TAZ) are small areas containing the land use data that is used as the foundation for
trip-making in the travel model. For the SLO Citywide Travel Model, the TAZ layer is formatted as a polygon
layer in TransCAD’s GIS structure. The TAZs are attached to the network using zone centroids and centroid
connectors that allow travelers access to the transportation system by simulating local and neighborhood
streets. Once the roadway network and TAZ structure were finalized, the zone centroids and centroid
connectors were created.
TAZs are ideally but not always sized and shaped to provide a relatively homogeneous amount and type of
activity within each zone. TAZ delineations traditionally follow the natural and manmade boundaries that tend
to segregate different land uses. These boundaries include water features, roads, railroads, and other lines
that form logical boundaries. Jurisdictional and census boundaries often do not make for good TAZ
definitions because they can be arbitrary in relation to the needs of the model; but they are usually desirable
for data development and reporting functions.
The definition of traffic analysis zones has implications throughout the travel model. For roadway model
components, traffic analysis zone resolution affects the amount of precision that can be achieved when
loading vehicles onto the collector and arterial roadway network. This precision is obtained by increasing
detail in the roadway network, TAZ structure, and socioeconomic data. The desire for increased detail must,
however, be balanced with the ability to develop and maintain the data at the increased level of detail.
For transit model components, the size of traffic analysis zones affects the accuracy of transit pathbuilding,
particularly the walk to transit component. Algorithms used in the transit network processing and mode
choice model have been designed to work properly with the TAZ structure, including the aggregated zones
outside of the SOI.
The TAZ layer from the 2008 model was carried forward for most of the region. Some TAZs that are
anticipated to experience significant population or employment growth were split. Figure 2.1, Figure 2.2, and
Figure 2.3 show the roadway network and TAZs used by the SLO Citywide Travel Model.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
2-27
Figure 2.1. Traffic Analysis Zones (Countywide)
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
2-28
Figure 2.2. Traffic Analysis Zones (Citywide)
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
2-29
Figure 2.3. Traffic Analysis Zones (CBD)
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
3-30
3.0 Land Use Dataset Development
Land use and socioeconomic data provides the fundamental supply side input to the travel model.
Information about households and about non-residential land uses is used determine the total number of
trips made to and from each traffic analysis zone (TAZ).
3.1 Data Sources
This section identifies the development of the land use dataset for the SLO Citywide Travel Model. Several
sources were considered for development of the input data:
City of San Luis Obispo land use and parcel data (available in the SOI/City areas);
ACS 2011-2015 5-year dataset; and,
2010 Census.
Each of these is discussed separately below. A discussion of the approach used to develop the model
dataset follows.
3.1.1 City of San Luis Obispo Land Use Data
The City provided a parcel database that contains a snapshot of the land uses in the City. This database was
fundamental in developing the base year 2016 land use data for the SLO Citywide Travel Model. After initial
development of the 2016 base year land use data, the City performed a comprehensive update to the land
use database. Base year land use data was updated in 2019 based on this updated database. The City’s
parcel dataset was processed to produce land use data for input to the travel model as discussed in section
3.2.
3.1.2 ACS 2011-2015 Dataset
The ACS dataset provides household income and size data that can be applied to household data at the TAZ
level. While the year ACS household totals are inconsistent with the 2016 base year for the model, the
distribution of households by income and size is still useful. Household classification by size and income is
available at the block group level. Income data is provided in the following groups:
1. Less than $10,000;
2. $10,000 to 74,999 (in $5,000 increments);
3. $75,000 - $99,999;
4. $100,000 - $124,999;
5. $125,000 - $149,999;
6. $150,000 - $199,999; and
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
3-31
7. Over $200,000.
Median income for each TAZ and average household size for each TAZ can also be determined using ACS
data.
3.1.3 2010 Census
Cross-classification of households by income and size is only available in the decennial Census and only at
the Public Use Microdata Areas (PUMAs) geography. There are two PUMAs in San Luis Obispo County, one
containing San Luis Obispo, Pismo Beach, Grover Beach, Arroyo Grande, Oceano, and Nipomo. A
countywide cross-classified distribution of households for 2010 has been obtained from the American
FactFinder. It was then used to determine the number of households in each income and household size
group while maintaining the ACS totals for each income level and household size.
3.2 Model Dataset
The land use and socioeconomic data for the SLO Citywide Travel Model has been derived from the various
data sources as listed in Table 3.1.
Table 3.1. Source of Input Socioeconomic and Land Use Data
Data Requirement Data Source (Existing) Data Source (Future)
City / SOI Non-Residential Land Use City Parcel Dataset City Comprehensive Plan
City / SOI Residential Household
Totals
City Parcel Dataset City Comprehensive Plan
City / SOI Median Household Income
and Size by TAZ
ACS 2011-2015 Area type templates, optional
additional detail
County Data SLOCOG Model SLOCOG Model
The City’s parcel dataset has been processed to produce a TAZ-level dataset that contains the square
footage of non-residential land use by land use categories used by the model. It was also used to obtain the
number of households in each TAZ. Land use categories used by the model are shown in Table 3.2 along
with additional household-based socioeconomic variables. Outside of the SOI, trip generation results
imported from the SLOCOG model are used directly. Therefore, socioeconomic data from the SLOCOG data
is not required outside of the SOI.
In addition to total households, TAZ-level information about household size and income is used to improve
the model’s accuracy in the trip generation, trip distribution, and mode choice components. To provide this
additional accuracy, average household size and median household income for each TAZ are required
model inputs. These variables have been approximated at the TAZ level based on 2011-2015 ACS data.
Table 3.2. Model Land Use and Socioeconomic TAZ Input Data
General Category Travel Model Category Units Field Name
Household Household Households TOT_HH
Median Household Income 1999 dollars MED_INC
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
3-32
General Category Travel Model Category Units Field Name
Average Household Size Persons/HH AVG_HHSZ
Office and Service General Office ksf OFFICE
Service Religious Organizations and Meeting Halls ksf RELIG
Hospitals ksf HOSP
Airport (rate per thousand annual enplanements) Enpl AIRPORT
Lodging Motels and Hotels Rooms MOTEL
Beach Resorts Acres BEACH_RST
Retail High Generation Retail ksf HIGH_RET
Medium Generation Retail ksf MED_RET
Low Generation Retail ksf LOW_RET
Schools Elementary Schools Students ELM_SCH
High Schools Students HIGH_SCH
CalPoly Students Students CP_STUD
CalPoly Employees Jobs CP_EMP
Cuesta College Students Students CC_STUD
Cuesta College Employees Jobs CC_EMP
Industry Light Industrial ksf L_IND
Heavy Industrial ksf H_IND
Other Parks & Recreation Acres PARKREC
Agricultural Acres AG
Undeveloped Acres UNDEV
Forecasting of Household Size and Income
Due to the household income and size data for future year conditions not being available at the TAZ level, a
method has been developed to simplify the input of household income and size data in future year datasets.
Different approaches are recommended depending on the existing amount of development in a zones as
well as the amount of growth forecast. Four options are suggested for development of household size and
income data for each TAZ:
1. Use the base year TAZ median income and/or average household size,
2. Use the regional median income and/or average household size,
3. Specify median income and/or average household size based on a template (e.g., by area type), or
4. Specify the average median income and/or average household size directly.
The simplest case, using the base year data by TAZ, is appropriate in many cases. However, there may be
cases in which additional refinement is desired. The flowcharts provided in Figure 3.1 and Figure 3.2
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
3-33
describe a process through which a more refined approach can be implemented. Relevant numbers that are
useful in refinement of the data are included in Table 3.3.
Table 3.3. Selected Household Size and Income Summary Data
Variable Value
Base Year Citywide Median Income $49,669
1 Standard Deviation $77,019
1 Standard Deviation $22,319
Base Year Citywide Average Household Size 2.46
1 Standard Deviation 3.23
1 Standard Deviation 1.69
Base Year CBD Median Income $48,983
Base Year Fringe Median Income $38,036
Base Year Urban Median Income $48,983
Base Year Suburban Median Income $56,657
Base Year Rural Median Income $35,089
Base Year CBD Average HH Size 1.64
Base Year Fringe Average HH Size 1.92
Base Year Urban Average HH Size 2.21
Base Year Suburban Average HH Size 2.57
Base Year Rural Average HH Size 2.35
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
3-34
Figure 3.1. Median Income Forecasting Flowchart
Is specific income
data available for
projected growth?
Use specific
information
TAZ: Median
Income
Is base year Total
HH > 10 or is
growth a
significant portion
of forecast total*?
Is Median Income
within 2 standard
deviations of the
regional median?
Can median income
be forecast at the
template level?
Use the base year
TAZ median income
Use the median
income from the
appropriate template
Use the base
year regional
median income
Can median
income be
forecast at the
template level?
Yes
No
Yes
No
No
Yes
No
Ye
s
No
Yes
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
3-35
Figure 3.2. Average Household Size Forecasting Flowchart
Is specific size
data available for
projected growth?
Use specific
information
TAZ: Average Household
Size
Is base year Total
HH > 10 or is
growth a
significant portion
of forecast total?
Is average size
within 2 standard
deviations of the
regional median?
Can average size
be forecast at the
template level?
Use the base
year TAZ
average size
Use the average size
from the appropriate
template
Use the base
year regional
average size
Can average size
be forecast at the
template level?
Yes
No
Yes
No
No
Yes
No
Yes
No
Ye
s
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
3-36
3.3 Household Disaggregation Models
The socioeconomic input data includes information about median household income and average household
size. Household disaggregation models are then used to estimate the univariate distribution of households
by size and by income group for each TAZ. These models are based on US Census data at the block and
block-group level.
To apply these models, each known variable is used to look up a distribution of households by classification.
For example, a zone with an average household size of 1 person would be comprised entirely of 1-person
households (by definition). Conversely, a zone with an average household size of 4 people would be
modeled as a combination of 1, 2, 3, 4, and 5+ person households. Distributions are represented by hand-
fitted curves based on US Census and ACS data aggregated to TAZs.
It is important that the distribution curves always sum to 100% and that, for the household size model, the
results are consistent with the input value when averaged. Hand-fitted curves have been adjusted to fit the
observed data points, sum to 100%, and produce the appropriate average.
The household income model is expressed as a percentage of regional income rather than an income value
in dollars. This is done to allow for median income data to be input to the model in any chosen units, as long
as the units are consistent for all zones.
3.3.1 Household Size Disaggregation Model
Model trip rates are classified by 5 household size groups. The portion of households in each group can be
approximated for any given TAZ based on the average household size. Disaggregation curves are shown
along with Census data in Figure 3.3. The resulting model is defined as a lookup table provided with the
travel model input dataset.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
3-37
Figure 3.3. Household Size Disaggregation Curves
3.3.2 Household Income Disaggregation Model
The household income group model was developed in a manner similar to the household size disaggregation
model. Low, medium, and high income groups are defined in Table 3.4. Disaggregation curves are shown
along with Census data in Figure 3.4. The resulting model is defined as a lookup table provided with the
travel model input dataset.
Table 3.4. Income Group Definitions
Income Group Income Range
Low $34,999 and lower
Medium $35,000 – $99,999
High $100,000 and higher
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1 1.5 2 2.5 3 3.5 4 4.5 5 5.5
Average Household Size
1-person model 2-person model 3-person model 4-person model 5-person model
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
3-38
Figure 3.4. Household Income Disaggregation Model
3.3.3 TAZ-Level Bivariate Data
The household income and size disaggregation models produce univariate data for each TAZ. To apply trip
production rates that vary by household size and income, bivariate household data is required at the TAZ
level. The TAZ-level data resulting from the household size disaggregation models is used along with the
regional bivariate distribution of households by size and income to estimate the bivariate distribution of
households for each TAZ. The regional bivariate distribution of households by size and income, shown in
Table 3.5, was obtained from the 2010 Public Use Microsample (PUMS) dataset.
Table 3.5. Bivariate Household Distribution for San Luis Obispo County
Income Group 1 Person 2 Person 3 Person 4 Person 5+ Person Total
Low 11,625 6,106 2,204 2,068 1,632 23,635
Medium 7,420 15,327 5,622 4,383 2,843 35,595
High 1,788 10,154 4,110 3,578 2,571 22,201
Total 20,833 31,587 11,936 10,029 7,046 81,431
Source: 2010 PUMS Dataset for San Luis Obispo County
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
70.0%
80.0%
90.0%
100.0%
0 1 2 3 4 5 6
Low Income Model Medium Income Model High Income Model
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-39
4.0 Trip Generation
Trip generation is the first phase of the traditional 4-step travel demand modeling process. It identifies the trip
ends (productions and attractions) that correspond to the places at which activities occur as represented by
land use data (e.g., dwelling units, square feet of commercial development). Productions and attractions are
estimated for each TAZ by trip purpose, and then balanced at the county level. Production and attraction
allocation sub-models are applied in some cases to better represent the geographic locations at which these
trip ends occur. The resulting productions and attractions by trip purpose and TAZ are subsequently used by
the Trip Distribution model to estimate zone-to-zone travel patterns.
Trip productions and attractions are the fundamental variables for defining the trip ends associated with
travel. With the exception of non-home-based trips and trips external to the modeling area, productions occur
at the home end of a trip and attractions occur at the non-home end of the trip.
The primary data source for estimating trip productions and attractions is the California Household Travel
Survey (CHTS) conducted in 2012. Since the survey is household-based, it provides excellent information
with regard to household trip-making. Therefore, the CHTS is especially well suited for estimating trip
production rates. The CHTS also provides some information about trip attractions, but is not particularly well
suited for this purpose due to the absence of data about land use activities at the attraction end of recorded
trips. Attraction rates were carried forward from the 2008 base year model and adjusted during model
calibration and validation.
4.1 California Household Travel Survey (CHTS)
Caltrans conducted the most recent CHTS between October 2012 and December 2012 among households
located in each of the 58 counties throughout the State. Household socioeconomic data gathered in this
survey includes information on household size, income, vehicle ownership, employment status of each
household member, and housing unit type. Travel information was also collected including trip times, mode,
activity at each trip-end, and vehicle occupancy. The survey was conducted among randomly selected
households using telephone recruitment, followed by a diary mail out. Travel diary information was collected
using a telephone interview.
The CHTS data provided by Caltrans had already been through a quality control process. All household and
trip-end locations in the survey database had been geocoded. The survey process and results are
summarized in 2012 California Statewide Travel Survey published by the California Department of
Transportation.
There are 573 records available for households within San Luis Obispo County. When classified by variables
such as household income and size, trip generation characteristics for the County as a whole are assumed
to be representative of trip generation rates within the City of San Luis Obispo. This assumption allows use
of the countywide dataset to develop trip rates for the SLO Citywide Travel Model.
The CHTS included a combined weighting and expansion factor for each household. Weights are sensitive to
vehicle availability, household tenure (rent/own status), county of residence, household size, and household
income. The expanded number of household from the CHTS were similar to the total number of households
in the PUMS dataset and hence weighted trip and household data from the CHTS were used directly to
estimate the trip rates.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-40
4.2 Trip Purposes
Generally, a trip is defined as a distinct travel movement from one clearly identifiable starting place/activity to
another with a distance of at more than one block. The number of trip purposes was retained from the
previous model including:
Home-Based Work (HBW): Commute trips between home and work, including work trips made by
CalPoly employees.
Home-Based University (HBU): Trips between home and the CalPoly campus for school related
purposes by people not employed by the University.
Home-Based Shop (HBS): Trips between home and shopping locations for the purpose of shopping.
Home-Based Other (HBO): All other trips that have one end at home.
Work-Based Other (WBO): Work-related trips without an end at home.
Other-Based Other (OBO): Trips with neither an end at home nor a work-related purpose.
Survey data was processed to identify 612,527 weighted weekday trips in San Luis Obispo County reported
by survey participants. Because weekend survey participants also reported trip activity on a weekday, only
the trips reported on weekends were dropped from the analysis. Households participating in the weekend
portion of the survey were retained in the household total, as were the weekday trips made by these
households.
Survey respondents were asked to report their primary activity at each place visited during the course of a
day. These primary activities were used to categorize each trip into one of the purposes used in the travel
model, resulting in the total number of trips by each purpose shown in Table 4.1. Because only trips that
occur entirely within San Luis Obispo County are included in the model’s trip rates, trips which begin or end
outside the county were dropped from the analysis. Trip purposes were identified based on the origin and
destination activity for each trip using the relationship shown in Table 4.2. Certain origin/destination trip
activity combinations, such as home to home, have been designated as NA and dropped from the trip rate
analysis. Such occurrences were exceedingly rare and do not have a significant impact on overall trip rates.
Table 4.1. CHTS Weighted and Expanded Trips by Purpose for San Luis Obispo
County
Trip Purpose Expanded Weekday Trips Percent of Total
HBW 105,137 17.2%
HBS 89,056 14.5%
HBU 6,583 1.1%
HBO 233,633 38.1%
WBO 61,115 10.0%
OBO 117,002 19.1%
Total 612,527 100%
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-41
Table 4.2. Trip Purpose by Trip Origin and Destination Activities
Home Intermodal Work /
volunteer
School
K12)
College /
University
Shopping Other
Activities
Home NA HBO HBW HBO HBU HBS HBO
Intermodal
activities
HBO OBO WBO OBO OBO OBO OBO
Work/volunte
er
HBW WBO WBO WBO WBO WBO WBO
School (K12) HBO OBO WBO OBO OBO OBO OBO
College/Univ
ersity
HBU OBO WBO OBO OBO OBO OBO
Shopping HBS OBO WBO OBO OBO OBO OBO
Other
activities
HBO OBO WBO OBO OBO OBO OBO
4.3 Production Rates
A detailed analysis of CHTS was conducted in order to develop trip production rates for the SLO Citywide
Travel Model. Past experience has shown that trip production rates are generally sensitive to household size
and to a measure of wealth (such as income or auto ownership). Analysis of CHTS data for the County of
San Luis Obispo has shown sensitivity to these variables. The production model for the updated model is
sensitive to both income and household size.
4.3.1 Income Groups
The CHTS places each household into one of seven income groups. Although useful, there are not sufficient
records in the dataset to retain all seven groupings as income categories. Hence, the income categories
were grouped as shown in Table 4.3.
Table 4.3. Income Categories
Income Group
Model)
Income
Category
Survey)
Low Under 24K
25 - 34K
Medium 35 - 49K
50 - 74K
75 - 99K
High 100 - 149K
Over 150K
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-42
4.3.2 Cross Classified Production Rates
Cross classified (by household size and income) trip rates can be initially computed as the mean number of
trips per household for each combination of household income and size. However, a sufficient number of
samples are not available for each combination. A review of mean trip rates, trip rate standard deviations,
and trip rate confidence intervals was conducted. Initial trip rates were adjusted through grouping of cells
across multiple income groups or household sizes where appropriate. The resulting trip rates by purpose are
shown in Table 4.4 through Table 4.9.
Trip rates for HBU trips are not shown, because sample sizes were not sufficient to develop meaningful trip
rates. HBU trips are instead handled by a special generator and production allocation model.
Table 4.4. Household Trip Production Rates – HBW
Income Group
Model)
Household Size TOTAL
1 2 3 4 5 or more
Low 0.46 0.85 1.86 3.09 3.39 1.08
Medium 0.71 1.31 2.50 3.09 3.39 1.80
High 1.36 1.75 2.50 3.09 4.52 2.42
Total 0.62 1.29 2.37 3.09 3.67 1.69
Table 4.5. Household Trip Production Rates – HBS
Income Group
Model)
Household Size TOTAL
1 2 3 4 5 or more
Low 1.11 1.11 1.57 1.57 1.57 1.20
Medium 1.11 2.05 2.05 2.25 2.25 1.89
High 1.37 2.05 2.72 2.72 2.72 2.28
Total 1.12 1.8 2.04 2.28 2.22 1.72
Table 4.6. Household Trip Production Rates – HBO
Income Group
Model)
Household Size TOTAL
1 2 3 4 5 or more
Low 0.94 1.82 2.46 5.83 5.83 1.99
Medium 0.94 1.82 2.46 5.83 5.83 2.65
High 0.94 1.82 2.54 7.41 7.41 3.49
Total 0.94 1.82 2.49 6.28 6.24 2.60
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-43
Table 4.7. Household Trip Production Rates – WBO
Income Group
Model)
Household Size TOTAL
1 2 3 4 5 or more
Low 0.37 0.37 0.37 0.37 0.53 0.37
Medium 0.44 1.18 1.60 1.60 1.60 1.16
High 0.21 1.11 1.60 1.60 1.60 1.25
Total 0.39 0.95 1.20 1.42 1.46 0.91
Table 4.8. Household trip Production Rates – OBO
Income Group
Model)
Household Size TOTAL
1 2 3 4 5 or more
Low 0.57 1.30 1.30 2.33 3.90 1.07
Medium 0.57 1.54 1.94 2.33 3.90 1.69
High 0.57 1.54 2.63 8.57 8.57 3.27
Total 0.57 1.48 1.88 3.99 4.55 1.74
Table 4.9. Household Trip Production Rates – Total
Income Group
Model)
Household Size TOTAL
1 2 3 4 5 or more
Low 3.45 5.45 7.56 13.19 15.22 5.71
Medium 3.77 7.9 10.55 15.1 16.97 9.19
High 4.45 8.27 11.99 23.39 24.82 12.71
Total 3.64 7.34 9.98 17.06 18.14 8.66
4.4 Attraction Rates
Attraction rates are used to identify the ends of trips that occur at locations other than the trip-maker’s home.
For home-based trip purposes, the attraction end of a trip occurs at a non-residential location, or occasionally
at another person’s home. For WBO trips, the trip production occurs at the trip maker’s workplace and the
trip attraction occurs at the non-work end of the trip. For OBO trips, the trip production and attraction are
synonymous with trip origin and destination. For both non home-based trip purposes, allocation models and
special procedures are used to properly locate the production and attraction end of each trip.
The CHTS did not identify a land use or employment type at all trip attraction locations. This prevents use of
the data to generate trip attraction rates specific to San Luis Obispo. Many recent surveys conducted in the
State of California suffer from a similar exclusion. Trip rates were developed for the 2008 base year model
using data from a household travel survey conducted in San Diego, along with trip rates published in the
Institute of Transportation Engineers (ITE) Trip Generation manual. These trip rates have been carried
forward into the 2016 base year model. Trip attraction rates used in the model are shown in Table 4.10.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-44
Table 4.10. Refined City of SLO Travel Model Attraction Rates
General
Category
ID Detailed Category Units Person Attraction Rates By
Purpose
WBO
Producti
on
Allocatio
n Rate
HBW HBS HBO WBO OBO
Residential 11 Single-Family Residential DU 0.08 0.00 0.97 0.11 0.37 0.05
12 Multi-Family Residential DU 0.12 0.00 0.37 0.11 0.37 0.08
Office 20 General Office KSF 3.65 0.00 2.32 0.15 0.89 0.88
Service 31 Religious Organizations and Meeting
Halls
KSF 0.43 0.00 7.73 0.51 2.98 0.10
32 General Service KSF 3.65 0.00 3.81 0.25 1.47 0.88
33 Hospitals KSF 5.37 0.00 2.78 0.18 1.07 1.29
34 Airport (rate per thousand annual
enplanements)
Enpl 5.79 0.00 11.59 5.79 5.79 1.39
Lodging 41 Motels and Hotels Rooms 0.97 0.00 0.11 0.74 0.95 0.23
42 Beach Resorts Acres 0.00 0.00 3.36 0.12 0.97 0.00
Retail 51 Drive In Retail KSF 13.40 18.82 28.01 8.94 23.23 1.97
52 High Generation Retail KSF 3.75 14.12 20.17 6.71 17.43 0.55
53 Medium Generation Retail KSF 3.75 9.41 12.33 4.47 11.61 0.55
54 Low Generation Retail KSF 1.64 4.71 6.72 2.24 5.81 0.24
Schools 61 Elementary Schools Students 0.13 0.00 1.36 0.19 0.41 0.03
62 High Schools Students 0.27 0.00 1.60 0.37 0.78 0.07
63 CalPoly Students Students See special generator discussion
64 CalPoly Employees Jobs
65 Cuesta College Students Students 0.00 0.00 1.18 0.00 0.41 0.00
66 Cuesta College Employees Jobs 0.80 0.00 0.00 1.00 1.33 0.19
Industry 71 Heavy Industrial KSF 2.84 0.00 0.62 0.83 0.38 0.60
72 Light Industrial KSF 3.17 2.80 0.67 0.45 0.90 0.67
Other 81 Parks & Recreation Acres 0.03 0.00 0.02 0.01 0.01 0.01
82 Agricultural Acres 0.02 0.00 0.13 0.01 0.01 0.00
83 Undeveloped Acres 0.00 0.00 0.00 0.00 0.00 0.00
4.4.1 Non-Home-Based Production Allocation Models
While WBO trips and OBO are initially generated using household based production rates, these trip
productions occur at non-residential locations. The total number of WBO and OBO productions generated at
households is used as a control total for trip balancing, but production allocation rates are used to move non
home-based productions to the appropriate work locations. For WBO trips, trip productions are defined as
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-45
the work trip end and attractions are defined as the non-work trip end. To accommodate this, a set of WBO
production allocation rates have been developed.
4.4.2 Trip Rate Factors and Adjustments
During model validation, it was observed that traffic volumes in and around the SOI were significantly lower
than traffic counts while the trips into and out of the SOI were higher than counts. Trip rates were adjusted to
account for the following issues:
Although localized, the CHTS dataset includes a relatively small sample size within San Luis Obispo
County.
Trip generation outside the SOI is based on the SLOCOG model instead of trip rates.
Aggregation of traffic analysis zones outside of the SOI results in aggregation bias. Trip rate adjustments
were made to counteract this phenomenon.
Trip rate factors are applied by district, as shown in Table 4.11 and Table 4.12. These factors are applied in
combination with K-factors described in Chapter 5.
Table 4.11. Trip Production Rate Adjustment Factors
District HBW HBS HBO HBU WBO OBO IE
Rural / External 1 1 1 1 1 1 1
SOI - Non-CBD 1.8 1.9 1.8 1.0 2.0 1.8 1.5
Morro Bay (North of SLO) 1 1 1 1 1 1 1
Atascadero (North of SLO) 1 1 1 1 1 1 1
Paso Robles (North of SLO) 1 1 1 1 1 1 1
Shandon (North of SLO) 1 1 1 1 1 1 1
California Valley 1 1 1 1 1 1 1
Pismo Beach (South of SLO) 1 1 1 1 1 1 1
Oceano (South of SLO) 1 1 1 1 1 1 1
SOI - CBD 1.8 1.8 1.8 1 1.8 1.8 1.5
Camp SLO / Cuesta 1.5 1.5 1.5 1 1.5 1.5 1.5
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-46
Table 4.12. Trip Attraction Rate Adjustment Factors
District HBW HBS HBO HBU WBO OBO IE
Rural / External 1 1 1 1 1 1 1
SOI - Non-CBD 2.0 1.9 1.8 1.8 1.8 1.8 1.5
Morro Bay (North of SLO) 1.2 1.2 1.2 1.2 1.2 1.2 1
Atascadero (North of SLO) 1.5 1.5 1.5 1.5 1.5 1.5 1
Paso Robles (North of SLO) 1.5 1.5 1.5 1.5 1.5 1.5 1
Shandon (North of SLO) 1 1 1 1 1 1 1
California Valley 1 1 1 1 1 1 1
Pismo Beach (South of SLO) 1.2 1.2 1.2 1.2 1.2 1.2 1
Oceano (South of SLO) 1.2 1.2 1.2 1.2 1.2 1.2 1
SOI - CBD 1.8 1.8 1.8 1.8 1.8 1.8 1.5
Camp SLO / Cuesta 2.5 2.5 2.5 2.5 2.5 2.5 1.5
4.5 CalPoly Trip Generation and Production Allocation
San Luis Obispo County is home to the California Polytechnic State University, San Luis Obispo (CalPoly)
and Cuesta College, a community college. Because CalPoly is a four-year college, students attending this
university tend to be concentrated at households near the universities or live on campus. This suggests that
a special university trip purpose and allocation model can improve representation of CalPoly in the travel
model.
On the other hand, Cuesta College is a community college and is likely to experience conditions that can be
adequately represented by the trip rates and gravity model used for other non-university purposes.
Furthermore, Cuesta College is located outside of the SOI. Therefore, Cuesta College has been represented
using trips directly from the SLOCOG model.
4.5.1 University Definition
The CalPoly campus is separated into seven traffic analysis zones as shown in Figure 4.1. Based on a
review of aerial photography and the CalPoly On-Campus Housing Map, zones were identified as zones that
produce trips (zones that include on-campus housing) and zones that attract trips (zones that contain uses
such as classes and offices). Production and attraction activity was then apportioned to zones based on
zone size and density. A summary of activity allocation by zone is included in Table 4.13.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-47
Figure 4.1. CalPoly Campus Zones
Table 4.13. CalPoly Activity Allocation by Zone
TAZ Description % of
Productions
of
Attractions
378 Central Campus 0% 20%
380 Campus Activity 0% 20%
380001 Campus Activity, plus
Poly Canyon Village On-Campus Housing
50% 20%
380002 Sports Complex 0% 5%
382 Central Campus 0% 30%
384 Primarily On-Campus Housing 50% 0%
386 Sports Complex / Campus Activity 0% 5%
Total 100% 100%
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-48
4.5.2 Trip Types at CalPoly
Because universities do not fall into the normal trip patterns used by the model, some special considerations
are given to trip types at CalPoly. In particular, the Home-Based University (HBU) trip purpose is defined as
a trip by a university student or visitor between home and any location on the university campus. Trip ends at
the University are associated with University faculty and staff, students living on campus, and students and
visitors living off campus and described as follows:
HBW, HBS, and HBO Productions: These production trip ends can occur only for students living on
campus.
HBW Attractions and WBO Productions: These trip ends can occur only for University faculty and
staff.
WBO Attractions and all OBO Trips: These trip ends can only occur for students and visitors living off
campus.
HBS and HBO Attractions: These trip ends cannot occur at the University. All home-based trips to the
University by students and visitors are considered HBU trips and all home-based trips to the University
by faculty and staff is considered HBW trips.
HBU Productions: Trips within the University are not modeled, so HBU productions cannot occur on
campus.
HBU Attractions: HBU attractions can occur only for students and visitors living off campus.
4.5.3 Special Generator Survey Adaptation
Detailed survey data was not available for CalPoly, so university special generator surveys from outside of
the region have been used to specify a special generator models for CalPoly. Special generator studies
conducted for Colorado State University (CSU) and the University of Northern Colorado (UNC) were used to
estimate a special generator model for the North Front Range (Colorado) Regional Travel Model (NFR
RTM). These studies and special generator models were borrowed and adjusted for application at CalPoly.
4.5.4 Employment and Enrollment Data
Total employment and enrollment data for CalPoly was retrieved from the University website4. Because
employment data includes part-time employees, a factor of 0.90 was applied to convert total employment to
full-time equivalent employees (for consistency with the UNC and CSU surveys). Total enrollment was
divided into on-campus and off-campus enrollment based on the number of on-campus housing units.
Employment and enrollment data for CalPoly is summarized in Table 4.14.
4 http://www.calpolynews.calpoly.edu/quickfacts.html
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-49
Table 4.14. CalPoly Employment and Enrollment
Trip Maker Employment or Enrollment
FTE Faculty and Staff 3,015
On-Campus Students 7,200
Off-Campus Students 13,744
Total Enrollment 20,944
Special Generator Values
Special generator values from the NFR RTM were adapted for use in the model by computing a surrogate
trip rate for each trip type based on FTE employment, on-campus students, or off-campus students. Where
data was available, the CSU special generator values were used because CSU is more similar to CalPoly.
The CSU special generator study grouped WBO and OBO trips into non home-based trips, so UNC values
were used to compute WBO and OBO special generator values. During model validation, cordon counts
surrounding the university were reviewed and special generator values were adjusted accordingly. Trip rates,
initial special generator values, and adjusted special generator values are shown in Table 4.15.
Table 4.15. University Special Generator Values
Trip Purpose Trip Rate Unit Generator Value
HBW Productions 0.22 On Campus Students 1,584
HBW Attractions 1.6 FTE Employment 4,824
HBS Productions 0.2 On Campus Students 1,440
HBS Attractions n/a n/a 0
HBU Productions n/a n/a 0
HBU Attractions 3.80 Off Campus Students 52,227
HBO Productions 0.5 On Campus Students 3,600
HBO Attractions n/a n/a 0
WBO Production 0.37 FTE Employment 1,116
WBO Attractions 0.19 Off Campus Student 2,611
OBO Productions 0.25 Off Campus Student 3,436
OBO Attractions 0.25 Off Campus Student 3,436
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-50
Production Allocation
The production end of each HBU trip will occur at a household, most likely near the university. Analysis of the
CHTS survey data was not possible due to the low capture rate of university trips. Instead, student housing
information was used to allocate HBU trip productions to TAZs.
In 2015, 20,944 students were enrolled at Cal Poly with 34% living in dorms on campus. The remaining
13,744 students lived in off-campus housing including dedicated student housing and regular housing. In
order for the model to allocate these students, a three-step process was followed:
1. Identify dedicated off-campus student housing locations.
2. Determine capacity of dedicated off-campus student housing and assign students to available off-
campus student housing.
3. Allocate the remaining student to other housing throughout the region.
The first step is to identify off-campus student housing where only students are likely to live- this includes
official student apartments as well as apartments located in the campus vicinity. Official student off-campus
housing includes the Mustang Village and Stenner Glen Apartments. In addition, apartment complexes within
one mile from campus and those identified by City staff are also considered dedicated student off-campus
housing since these are very likely occupied only by students.
Next, it is necessary to determine the capacity for the dedicated student off-campus housing. The number of
available units for off-campus housing was previously provided by City of SLO and is used for this analysis.
The list of dedicated off-campus student housing along with the unit count is shown in Table 4.16.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-51
Table 4.16. Dedicated Off Campus Student Housing
Apartment Name Address Units Beds / Unit Beds
Mustang Village 1 Mustang Dr. 534 1 to 4 1,335
Stenner Glen 1050 Foothill Blvd. 420 1 420
Valencia Apartments 555 Ramona Dr. 160 3 480
Murray Station 1262 Murray St 82 1 to 4 164
Cedar Creek Association 75 Stenner St. 68 2 136
Stafford Gardens 1415 Stafford St. 59 1 to 2 89
University Vista Apartments 1230 Murray 34 1 to 4 85
Pine Creek 1185 Foothill Blvd. 36 2 72
College Gardens 284 North Chorro St. 44 1 to 2 66
Segrado Corazon Townhomes 60 Casa Street 32 2 64
Foothill Gardens 1311 Foothill Blvd. 53 1 53
Alta Vista Park 265-305 N. Chorro 25 2 50
Cal Park Apartments 250 California Blvd. 24 2 48
Foothill Hacienda 190 California Blvd. 23 2 46
Lee Arms Apartments 258 California 21 2 42
Sierra Vista 510 Foothill Blvd. 26 1 to 2 39
Glen Mar Apartments 1127 Foothill Blvd. 18 2 36
Alta Vista Woods 355-385 N. Chorro 18 2 36
San Luis Village 204 California Blvd. 23 1 to 2 35
Casa Bonita 28-38 Casa Street 12 1 to 2 14
Fairview Apartments 1629 Johnson 22 1 to 2 33
Garfield Arms 738 Grand Ave. 61 1 to 2 91.5
Irish Hills Apartments 11343 Los Osos Valley 146 1 to 3 292
Kris Kai Townhomes 607 Grand Ave. 20 2 40
Sheri Apartments 721 Johnson 122 1 122
Madonna Rd. Apartments 1550 Madonna Rd. 120 1 to 4 300
Total of Apts (included in off campus housing) 2,203 4,489
Once identified, the dedicated student housing units are added to the land use database as a separate
residential “student housing” category. Instead of using the number of housing units, the number of beds is
included in the database to avoid applying a household size parameter to student housing. When the model
is run, the home end of the home-based university trips will be allocated to these zones first before being
allocated to the rest of the region. Note that student housing is assumed to be fully occupied.
Finally, the students who do not reside in dorms or dedicated off-campus housing are allocated to the
general housing supply in the region based on the distance from the university. Since students are more
likely to reside in multi-family units, these are assigned a higher weight than the single family units during the
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-52
allocation process. Prior to the allocation process, housing that cannot be occupied by students including
retirement homes and mobile homes is removed from the available supply. This is accomplished by splitting
the multi-family units in the land use database into multi-family 1, which includes all multi-family households,
and multi-family 2, which cannot be used by students. Multi-family 2 is a subset of the total multi-family
households included in multi-family 1.
In the model, the students not residing in the dedicated on-campus or off-campus student housing are
allocated to zones based on their proximity to campus and availability of single-family and multi-family units.
Once the allocation is complete, the total number of student households as a share of total households has
been compared to the percent of rentals for the area where the household is located. This check ensures
that a reasonable number of students are assigned to zones neighboring the Cal Poly campus. The resulting
university production allocation is shown in Figure 4.2.
Figure 4.2. University Production Allocation
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-53
4.6 External Trips
The SLO Citywide Travel Model covers all of San Luis Obispo County, but does not extend beyond the
County’s boundary. Therefore, the model must include all trips that traverse the County or have at least one
end in the County. Trips with at least one end outside of the County are called external trips. The model
includes external trips as follows:
External-External (EE) Trips – These are through trips that start and end outside of the county modeling
domain. In processing data, EE trips can also include short convenience stops within the county since
these stops are not true destinations and do not reflect the true purpose of the trip. EE trips are applied
in the travel model as a balanced input table of trips between external stations.
Internal-External (IE) and External-Internal (EI) Trips – These trips have one end within the modeling
domain and the other outside of it. For the purposes of the SLO Citywide Travel Model, these trips are
referred to as IE trips.
External travel is applied at the boundary of the model using the external stations shown in Figure 4.3. These
are nodes that receive special processing for external travel. Traffic counts collected at each external station
provide some of the necessary data for estimating external travel. Traffic counts do not indicate origin or
destination, trip purpose, split of through trips, or much extra information other than total vehicles, time of
day, and possibly vehicle classification.
Household travel diary surveys provide detailed information about Internal-Internal (II) trips and some
information about IE/EI trips, but only those that are generated (produced) by households within the
surveyed area. The CHTS provides some limited information about external travel and was used to develop
a methodology for handling of IE and EE trips for the SLOCOG model.
External travel studies and surveys are relatively expensive and not as common as other surveys such as
household travel diary surveys. External travel cordon studies can utilize noninvasive techniques such as
matching license plates at external stations using camera and radar technology combined with special
software that interprets the license plate data. These studies are particularly useful for obtaining EE split and
orientation data. They do not, however, provide any detailed information about the trip purpose.
External travel surveys, on the other hand, can provide a significant amount of trip-specific data by including
some sort of interaction with the driver through roadside pullover interviews, postcard surveys based on
license plate address data, or other means. These tend to be relatively expensive and generally do not attain
high participation rates from the traveling public.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-54
Figure 4.3. External Station Locations
4.6.1 Internal/External Trip Methodology
The SLO Citywide Travel Model includes internal/external and external/internal trips as a single trip purpose,
referred to in the model as IE trips. IE Trip attractions are defined as the internal trip end, with IE productions
defined as the external trip end. This allows IE trips to be balanced to trip productions, which occur at
external stations and are based on traffic count data.
IE Attractions and are generated for each internal TAZ using attraction rates. The IE attraction rates were
adapted from rates used by the SLOCOG model, which are in turn based on data from the CHTS. The
SLOCOG model uses different IE trip rates for different parts of the county. These rates have been simplified
to two sets of rates for use in the SLO Citywide Travel Model: one set for zones within the SOI and another
for zones outside of the SOI. IE attraction rates are shown in Table 4.17.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-55
Table 4.17. IE Trip Attraction Rates
General
Category
ID Detailed Category Units IE Attraction
Rate (SOI)*
IE Attraction
Rate (non-SOI)**
Residential 11 Single Family Household DU 0.075 0.313
12 Multi Family Household DU 0.050 0.286
Office 20 General Office KSF 0 0
Service 31 Religious Organizations and
Meeting Halls
KSF 0 0
32 General Service KSF 0 0
33 Hospitals KSF 0 0.001
34 Airport 0 0
Lodging 41 Motels and Hotels Rooms 0.155 0.334
42 Beach Resorts Acres 4.695 4.320
Retail 51 Drive In Retail KSF 0 0
52 High Generation Retail KSF 0.014 0.041
53 Medium Generation Retail KSF 0 0.003
54 Low Generation Retail KSF 0 0.001
Schools 61 Elementary Schools Students 0.022 0.057
62 High Schools Students 0.029 0.075
63 CalPoly Students Students 0 0
64 CalPoly Employees Jobs 0 0
65 Cuesta College Students Students 0 0
66 Cuesta College Employees Jobs 0 0
Industry 71 Heavy Industrial KSF 0 0
72 Light Industrial KSF 0 0
Other 81 Parks & Recreation Acres 0.001 0.003
82 Agricultural Acres 0 0
83 Undeveloped Acres 0 0
Total IX+XI production and attraction rates from the SLOCOG model for Area Type 1: the City of San Luis Obispo and
nearest suburbs.
Average total IX+XI production and attraction rates for Area Type 2 through 5.
IE productions are defined in the SLO Citywide Travel Model at the external stations. The number of IE
productions occurring at each external station is based on the two data sources described below.
The total vehicle trips at the external station: Total vehicle trips for 2016 at external stations were
borrowed from the SLOCOG Model. Volumes implied by the external station data are available for the
year 2015.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-56
The assumed percentage of EE trips at each external station: The SLOCOG model assumes a
consistent split between IE and EE trips at each external station for 2015. These assumptions were used
directly.
External station data for the SLO Citywide Travel Model is summarized in Table 4.18.
Table 4.18. Base Year External Station Data
External Station Volume % IE % EE IE
Volume
EE
VolumeIDName
1 California 1 North 2,692 73% 27% 1,953 738
2 US 101 North 23,346 86% 14% 20,148 3,198
3 California 41 7,630 95% 5% 7,247 383
4 California 46 7,070 83% 17% 5,889 1,181
5 Carrisa Hwy 134 100% 0% 134 0
6 Maricopa Hwy East 5,205 9% 91% 490 4,715
7 Maricopa Hwy South 1,647 0% 100% 5 1,642
8 US 101 South 66,822 91% 9% 60,745 6,077
9 California 1 South 8,912 99% 1% 8,819 93
4.6.2 External/External Trip Methodology
External-External trips (EE trips) are input to the SLO Citywide Travel Model directly as a vehicle trip table.
The through trip table is based on the through trip table used by the SLOCOG model for 2015. The resulting
2016 EE trip table is presented as Table 4.19.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-57
Table 4.19. Base Year (2016) 24-hour EE Vehicle Trip Table
External
Station
ID
1 2 3 4 5 6 7 8 9 Total
1 0 0 13 337 0 0 0 18 1 370
2 0 0 0 253 0 0 0 1,308 64 1,625
3 13 0 0 0 0 0 0 164 28 206
4 337 253 0 0 0 0 0 0 0 590
5 0 0 0 0 0 0 0 0 0 0
6 0 0 0 0 0 0 817 1,539 0 2,356
7 0 0 0 0 0 820 0 3 0 823
8 18 1,320 164 0 0 1,539 3 0 0 3,044
9 0 0 0 0 0 0 0 0 0 0
Total 369 1,573 178 591 0 2,359 819 3,033 93 9,014
4.6.3 Forecast External Trips
External trip data for the forecast year was available from the SLOCOG model. Resulting IE/ and EE
assumptions are shown in Table 4.20 and Table 4.21.
Table 4.20. Forecast Year External Station Data
External Station Volume % IE % EE IE
Volume
EE
VolumeIDName
1 California 1 North 4,155 82% 18% 3,417 738
2 US 101 North 32,932 90% 10% 29,733 3,198
3 California 41 10,482 96% 4% 10,099 383
4 California 46 10,873 89% 11% 9,692 1,181
5 Carrisa Hwy 148 100% 0% 148 0
6 Maricopa Hwy East 5,244 10% 90% 529 4,715
7 Maricopa Hwy South 1,656 1% 99% 14 1,642
8 US 101 South 88,458 93% 7% 82,380 6,077
9 California 1 South 10,616 99% 1% 10,523 93
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
4-58
Table 4.21. Future Year (2040) 24-hour EE Vehicle Trip Table
External
Station
ID
1 2 3 4 5 6 7 8 9 Total
1 0 0 18 527 0 0 0 21 1 567
2 0 0 0 395 0 0 0 1,557 74 2,027
3 18 0 0 0 0 0 0 217 37 272
4 527 395 0 0 0 0 0 0 0 922
5 0 0 0 0 0 0 0 0 0 0
6 0 0 0 0 0 0 855 1,832 0 2,687
7 0 0 0 0 0 859 0 3 0 862
8 22 1,571 217 0 0 1,832 3 0 0 3,645
9 0 0 0 0 0 0 0 0 0 0
Total 566 1,966 235 922 0 2,690 859 3,631 112 10,981
4.7 Trip Balancing
Trip productions and attractions are estimated separately by purpose using the trip rates and allocation
models previously described. While an attempt is made to make the initial estimate of productions equal to
the initial estimate of attractions, it is not feasible to make them exactly equal in all scenarios, which is
necessary to ensure conservation of trips in the model. The balancing process provides this conservation by
making the productions and attractions equal.
Balancing depends on the level of confidence associated with the initial estimate of productions and
attractions. Since household survey data was available to estimate trip production rates, the home-based trip
purposes are balanced to trip productions. One exception to this is the HBU trip purpose. The special
generator studies and cordon counts upon which the CalPoly estimates are based provided an increase of
reliability for HBU trip attractions to the university campus, so HBU productions are balanced to attractions.
Non-Home-Based trips (WBO and OBO) are also balanced to productions. These trips are balanced to the
initial estimate of productions from the basic trip rates in the cross-classified trip production model. Then, the
productions are re-allocated using the allocation models previously discussed.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-59
5.0 Trip Distribution
This chapter describes the process used to update the Trip Distribution model for the SLO Citywide Travel
Model. Highway skim parameters and gravity model parameters are defined herein.
5.1 Context and Background
Trip distribution is the second phase of the traditional 4-step travel demand model. Trip distribution is the
process through which balanced person trip productions and attractions from the trip generation model are
apportioned among all zone pairs in the modeling domain by trip purpose. The resulting trip table matrix
contains both intrazonal (e.g., trips that don’t leave the zone) on the diagonal and interzonal trips in all other
zone interchange cells for each trip purpose.
The SLO Citywide Travel Model uses a standard gravity model equation and applies friction factors to
represent the effects of impedance between zones. As the impedance (e.g., travel time, spatial separation)
between zones increases, the number of trips between them will decrease as represented by a decreasing
friction factor. This is similar to the standard gravity model which assumes that the gravitational attraction
between two bodies decreases as they become further apart. The gravity model also assumes that the
gravitational attraction between the two bodies is directly proportional to their masses. The trip distribution
model makes a similar assumption in that the number of trips between two zones is directly proportional to
the number of productions and attractions contained in those zones. The gravity model used by trip
distribution to estimate the number of trips between each zone pair is defined in below.
Pi (
1
Where: Tij = trips from zone
i to zone j Pi =
productions in zone i Aj = attractions in zone j
Kij = K-factor
adjustment from i
to zone j i = production
zone j = attraction zone n = total number of zones Fij = friction factor (
a function of impedance between zones i and j) K-factors are often used in travel demand models
to account for nuances in travel behavior and the transportation system that
cannot be accurately modeled with simplified aggregate modeling techniques. They are often applied at the district or
jurisdictional level to adjust regional distribution patterns. They may be applied by trip purpose or for all trips. In the SLO Citywide
Travel Model, K-factors are used to correct for aggregation bias outside of the SOI and to
account for
the attractiveness of the SLO Central Business District (CBD). Friction factors represent the impedance to
travel between each zone pair. Friction factors have been calibrated for each trip purpose based
on observed trip length (time) frequency distributions and average travel times implied therein. Friction factors
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-60
Household Travel Survey (CHTS) for San Luis Obispo County and the City of San Luis Obispo Sphere of
Influence (SOI).
5.2 Peak and Off-Peak Period Definitions
Trips occurring during the AM and PM peak periods are distributed based on peak congested speeds and
trips occurring during off-peak times are distributed based on off-peak congested speeds. Trip distribution is
performed in Production-Attraction (PA) format rather than Origin-Destination (OD) format. This is because
the majority of trips in the AM peak period travel from production to attraction (e.g., to work) and the majority
of trips in the PM peak period travel from attraction to production (e.g., from work). The model uses
directional AM peak period speeds to compute impedance for both AM and PM peak period trips in PA
format.
Due to limited survey data, trips from the survey were not separated into peak and off-peak time periods for
the calibration exercise.
5.3 Roadway Network Shortest Path
The impedance portion of the gravity model equation is based on shortest paths between each zone pair.
Each shortest path is determined through a process called pathbuilding. This process identifies the shortest
route between two network centroids that minimizes an impedance variable. Shortest paths cannot pass
through other centroid connectors. Various data, such as path distance, can be “skimmed” along the shortest
impedance route. The set of all zone to zone shortest paths is called a “shortest path matrix” and is
sometimes referred to as a “skim matrix” with the understanding that the skimmed variable may differ from
the variable(s) used to determine the shortest path. This section describes the process used to generate
shortest path matrices for use in trip distribution.
The shortest path process is performed for the highway and transit networks. This section focuses on the
roadway network shortest path process. Discussion of transit shortest paths is included in the mode choice
chapter.
5.3.1 Terminal Penalties
Terminal penalties are applied in the model to the shortest paths. They simulate several travel-related
variables, such as the time to locate a parking space, walking to a final destination, paying for a parking
space, etc. Terminal penalties, shown in Table 5.1, are added to both the production and attraction end of
each zone pair based on the area type of each zone.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-61
Table 5.1. Terminal Penalties by Area Type
Area Type Terminal Time
1 CBD 1.5
2 Fringe 1
3 Urban 1
4 Suburban 0.75
5 Rural 0.50
6 Non-SOI 0.50
5.4 Intrazonal Impedance
Impedance, or travel time, for trips within a zone (intrazonal impedance) is not generated in the zone to zone
pathbuilding process because the roadway network is not detailed enough for a sub-TAZ level analysis.
Instead, the nearest neighbor rule is used to estimate intrazonal impedance. The nearest neighbor rule is
applied by taking an average of one or more nearby TAZs and multiplying that average by a factor.
Intrazonal travel time has been calculated by multiplying the travel time to the single nearest neighbor by
75% for zones within the SOI and by 70% for zones outside of the SOI.
5.5 Friction Factor Calibration
Friction factors were calibrated based on trip time distributions derived from CHTS data, using speed-based
shortest path matrices. While travel time information was available directly from the survey data, the gravity
model relies on network-based shortest paths. For consistency, calibration targets were created using the
same travel time data as the model.
5.5.1 Calibration Targets
Trip time distribution curves for each trip type have been generated using CHTS data. Shortest path matrices
used to develop these curves are based on freeflow speed and include both intrazonal travel time and
terminal penalties. The resulting countywide trip time distribution curves are shown in Figure 5.1.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-62
Figure 5.1. Countywide Trip Time Distribution Curves by Trip Purpose
5.5.2 Income Processing
In the SLO Citywide Travel Model, HBW trips are segmented into three income groups. Initially, HBW trip
distribution was calibrated for all income groups combined. Then, an attempt was made to develop separate
friction factors for each HBW income group. For the low and medium income groups, slight modifications
were made. However, there was not sufficient survey data to modify the high income HBW friction factors.
5.5.3 Calibration Process
The trip distribution model was calibrated by applying the gravity model using results of the trip generation
model. Friction factors based on the gamma function, defined in the equation below, were calibrated for each
trip type. The gravity model was applied using an initial set of gamma function parameters followed by
iterative adjustment of parameters. Iteration continued until no further improvement in replication of the
calibration target could be achieved through friction factor adjustments. Terminal penalties and intrazonal
travel times were included in the shortest path matrices for both the calibration targets and modeled trip time
distribution.
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
0 5 10 15 20 25 30 35 4045Share
of TripsTrip Time (
Minutes)HBW HBS HBO WBO
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-63
Where: Friction factor for zone pair i,j
Impedance (i.e., travel
time) between zones i and j a, b, c = Calibration parameters In addition to
friction factor adjustments, other model variables and parameters including terminal penalties, intrazonal travel times, volume/delay
equations,
and K-factors can affect calibration of trip length distribution curves. For each iteration of
the calibration process, parameters were adjusted in one of two ways: 1. For initial iterations, the equation
below was used to predict friction factor values at 1 minute increments. Gamma parameters were
then adjusted to fit a curve to the predicted friction factors. Equation (3) compares the trip
length distribution that results from application of a set of friction factors to the calibration target and
predicts new friction
factor
values
that
are more likely to replicate the calibration target.
1 1 Where: Predicted friction factor value for impedance
range r and iteration i 1 Gamma
function based friction factor value for impedance range r and iteration i-1 The
percentage of surveyed trips
in impedance range r 1 The percentage of trips
in impedance range r resulting from application
of iteration i-1 of the gravity model 2. Once application of the calibration equation stopped producing improvements, the gamma parameters were
manually adjusted for each iteration. Trip length and trip distributions for the county as a whole, for trips
within the SOI, and to/from the SOI were monitored during the calibration
process. Because the focus of
the model is the SOI, increased error in the countywide distribution was accepted in order to improve
trip distribution within the SOI. 5.5.4 Calibration Results A measure that can be used to
quantify the relationship between the observed and modeled trip length distribution curves is the coincidence
ratio. The coincidence ratio is a number between zero and one that specifies the area under both the calibration target and model
result trip length distribution curve. Coincidence ratios for each trip type and period are shown in
Table 5.2.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-64
Table 5.2. Coincidence Ratios for Calibrated Trip Time Distribution Curves
Trip Purpose Average Trip Length Coincidence Ratio
Observed Modeled
HBW 12.34 12.4 54%
HBS 9.02 8.61 49%
HBO 8.39 8.9 54%
WBO 8.99 9.27 36%
OBO 7.92 8.45 59%
Resulting trip time distribution curves for each trip type are shown in Figure 5.2 through Figure 5.6. Friction
factors by trip type are shown in Figure 5.7 and documented in Table 5.3.
Figure 5.2. HBW Trip Time Distribution Curve (Countywide)
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
0 20 40 6080Share
of TripsTrip Time (
Minutes)Model Observed Coincidence Ratio:
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-65
Figure 5.3. HBS Trip Time Distribution Curve (Countywide)
Figure 5.4. HBO Trip Time Distribution Curve (Countywide)
0%
5%
10%
15%
20%
25%
0 10 20 30 40 5060Share
of TripsTrip Time (
Minutes)Model Observed Coincidence Ratio:
49%
0%
5%
10%
15%
20%
25%
30%0 10 20 30 4050
60Share of TripsTrip
Time (Minutes)Model Observed Coincidence
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-66
Figure 5.5. WBO Trip Time Distribution Curve (Countywide)
Figure 5.6. OBO Trip Time Distribution Curve (Countywide)
0%
5%
10%
15%
20%
25%
0 10 20 30 40 5060Share
of TripsTrip Time (
Minutes)Model Observed Coincidence Ratio:
36%
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%0 10 20 30 4050
60Share of TripsTrip
Time (Minutes)Model Observed Coincidence
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-67
Table 5.3. Friction Factors for All Purposes
HBW HI HBW MI HBW LI HBS HBU HBO WBO OBO
A 100 100 100 100 100 100 100 100
B 1.1 0.9 0.7 1.2 1 0.5 0.8 2.37
C 0.2 0.15 0.05 0.9 0.1 0.38 0.22 0.03
Note: HBU friction factors are placeholders, as these trips are distributed using a production allocation model.
5.6 K Factors
While it is generally desirable to calibrate and validate travel models without the use of K-factors, it was
determined that K factors were necessary to produce reasonable base year model results. The K-factors in
the SLO Citywide Travel Model are used to account for the following issues:
CBD Activity: The SLO CBD acts as a regional destination, attracting longer trips than the remainder of
the City of SLO.
Aggregation Bias: Aggregation of TAZs outside of the SOI generally reduces the likelihood that trips will
remain within these aggregated communities. K-factors were used to better represent the amount of
travel remaining within these communities.
During the model validation process, the K-factor values shown in Table 5.4 were implemented. K-districts
are defined as shown in Figure 5.7.
Table 5.4. K Factors
District 1 2 3 4 5 6 7 8 9 10 11 97 98 99
1 Rural / External 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 SOI - Non-CBD 1 2 1 1 1 1 1 1 1 2 2 1 1 1
3 Morro Bay 1 1 3 3 3 3 1 0.5 0.5 1 1 1 1 1
4 Atascadero 1 1 3 3 3 3 1 0.5 0.5 1 1 1 1 1
5 Paso Robles 1 1 3 3 3 3 1 0.5 0.5 1 1 1 1 1
6 Shandon 1 1 3 3 3 3 1 0.5 0.5 1 1 1 1 1
7 California Valley 1 1 1 1 1 1 1 1 1 1 1 1 1 1
8 Pismo Beach 1 1 0.5 0.5 0.5 0.5 1 3 3 1 1 1 1 1
9 Oceano 1 1 0.5 0.5 0.5 0.5 1 3 3 1 1 1 1 1
10 SOI - CBD 1 2 1 1 1 1 1 1 1 2 2 1 1 1
11 Camp SLO/ Cuesta 1 2 1 1 1 1 1 1 1 2 2 1 1 1
97 Hwy 101 North External 1 1 1 1 1 1 1 1 1 1 1 1 1 1
98 Hwy 101 South External 1 1 1 1 1 1 1 1 1 1 1 1 1 1
99 Other External 1 1 1 1 1 1 1 1 1 1 1 1 1 1
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
5-68
Figure 5.7. K Districts
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
6-69
6.0 Mode Choice
6.1 Background
The mode choice model is the third step in the 4-step travel demand model. Mode choice separates the
person trip tables resulting from trip distribution into the various non-motorized, transit, and auto modes. The
process results in the personal vehicle trips, transit trips, and bicycle trips that are used in the assignment
step. The SLO Citywide Travel Model includes a number of different travel modes. Trips using personal
vehicles are subdivided by vehicle occupancy. Non-motorized trips are split into walking and biking. Mode
choice is determined using a nested logit model that uses information from roadway, transit, and bicycle
networks.
6.2 Mode Choice Model Structure
The SLO Citywide Travel Model uses a nested-logit model for all trip purposes as shown in Figure 6.1.
Figure 6.1. San Luis Obispo Mode Choice Model Nesting Structure
Nested-logit mode choice models define the state-of-the-practice for regions employing detailed transit
processing and mode choice. The multinomial logit model is defined as:
Where:
probability of selecting mode i a linear function
describing the utility of mode i In a nested-logit model, the utilities of modes in lower level nests are
included in higher level nests through logsum variables,
where
the
logsum
is defined
as:
ln (Choice
Motorized
Drive
Alone
Shared
Ride
2 Shared
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
6-70
Logsums are multiplied by a logsum coefficient in the range 0.0 to 1.0 when they are added into the
appropriate utilities for the next higher level nest.
6.3 Mode Choice Model Coefficients
The SLO Citywide Travel Model uses an asserted model as opposed to a model estimated using maximum
likelihood estimation techniques. The specification of mode choice model coefficients is based, in part, on
guidelines recommended by the Federal Transit Administration (FTA) for Section 5309 New Starts
Applications. Table 6.1 shows the FTA guidelines for in-vehicle travel time coefficients, out-of-vehicle travel
time coefficients, and cost coefficients. Table 6.1 shows the coefficients at the “motorized level” of the
nesting structure. In the mode choice model structure for the City of San Luis Obispo shown in Figure 6.1,
the motorized level is where choices are made among the auto driver-drive alone, auto driver-shared ride 2,
auto driver-shared ride 3+, and transit modes. In addition to the FTA guidelines, coefficients for the mode
choice models are also based on other models employed in California.
Table 6.1. FTA Mode Choice Model Coefficient Guidelines
FTA Guidelines1
Low Value High Value
Coefficient
In-vehicle travel time -0.03 -0.02
Initial wait -0.09 -0.04
Second wait -0.09 -0.04
Walk time -0.09 -0.04
Cost2 – –
Equivalent Minutes of IVTT
Initial wait 2.00 3.00
Second wait 2.00 3.00
Walk time 2.00 3.00
Home-Based Work Value of Time (Estimated Median Household Income)
Low Income ($20,000) $2.30 $3.10
Middle Income ($55,000) $6.60 $8.70
High Income ($140,000) $16.80 $22.40
1 Information from PowerPoint Presentation by FTA at TRB 83rd Annual Meeting, Session 501, January 13, 2004.
2 Value of Time is determined as Coefficient of In-Vehicle Travel Time/Coefficient of Cost. Actual guidelines are:
average income)/4 < Civt/Ccost < (average income)/3.
Table 6.2 shows the model coefficients for the SLO Citywide Travel Model mode choice models.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
6-71
Table 6.2. Mode Choice Model Coefficients
Coefficient Apply to
Modes
Home-
Based
Work
Home-
Based
Shop
Home-
Based
University
Home-
Based
Other
Work-
Based
Other
Other-
Based
Other
In-Vehicle Time (Minutes) Motorized -0.025 -0.025 -0.025 -0.025 -0.025 -0.025
Terminal Time (Minutes) Auto -0.025 -0.025 -0.025 -0.025 -0.025 -0.025
Walk Time (Minutes) Transit -0.050 -0.075 -0.050 -0.075 -0.050 -0.075
First Wait 7.5 Minutes Transit -0.050 -0.075 -0.050 -0.075 -0.050 -0.075
First Wait > 7.5 Minutes Transit -0.025 -0.025 -0.025 -0.025 -0.025 -0.025
Transfer Wait (Minutes) Transit -0.050 -0.075 -0.050 -0.075 -0.050 -0.075
Number of Transfers Transit 0.1875 0.1875 0.1875 0.1875 0.1875 0.1875
Cost (Cents)
All Households Motorized -0.00508 -0.00254 -0.00508 -0.00254 -0.00508
Low Income Household Motorized -0.00556
Middle Income Household Motorized -0.00195
High Income Household Motorized -0.00077
Walk Time Walk -0.05 -0.075 -0.05 -0.075 -0.05 -0.075
Ride Time Bicycle -0.05 -0.075 -0.05 -0.075 -0.05 -0.075
Cal Poly Destination (0/1)
Transit Transit 0.06 0.009 0.060 0.009 0.009 0.009
Walk Walk 0.009 0.009 0.300 0.009 0.009 0.009
Bicycle Bicycle 0.009 0.009 0.500 0.009 0.009 0.009
CBD Destination (0/1)
Transit Transit 0.009 0.009 0.009 0.009 0.009 0.009
Walk Walk 0.009 0.009 0.009 0.009 0.009 0.009
Bicycle Bicycle 0.009 0.009 0.009 0.009 0.009 0.009
Nesting Coefficients1
Top Level 0.9 0.9 0.9 0.9 0.9 0.9
Bottom Level 0.7 0.7 0.7 0.7 0.7 0.7
OVTT/IVTT Ratio 2.0 3.0 2.0 3.0 2.0 3.0
Value of Time ($/Hr)
All Households $2.95 $5.90 $2.95 $5.90 $2.95
Low Income Household $2.70
Middle Income Household $7.70
High Income Household $19.60
1 All model coefficients are specified at the “motorized” choice level. If utilities for walk and bicycle modes are
calculated at the top level, the appropriate coefficients should be multiplied by the top level nesting
coefficient. If walk to transit and drive to transit utilities are calculated at the lowest level of the nesting
structure, the appropriate coefficients should be divided by the bottom level nesting coefficient.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
6-72
6.4 Mode Choice Model Calibration Targets
6.4.1 Target Trips and Shares by Mode and Purpose
Table 6.3 shows the modal shares for San Luis Obispo County based on the 2012 California Statewide
Household Travel Survey. These mode shares were used to create modal targets and perform the initial
calibration of the mode choice model. However, once the bicycle trips based on the mode choice model
output were assigned (as discussed in Chapter 8) and compared to the observed bike counts, a need to
reduce the bicycle targets became apparent. The initial bike assignment overestimated the observed bike
volumes by more than a factor of 2. The bicycle mode share targets were reduced to those shown in Table
6.4.
Table 6.3. Initial Calibration Target Trips by Mode and Purpose for City of San Luis
Obispo Travel Model
Mode Home-Based Work Home-
Based
Shop
Home-
Based
University
Home-
Based
Other
Work-
Based
Other
Other-
Based
Other
Low
Income
Middle
Income
High
Income
Drive Alone 74.1% 85.8% 90.6% 48.7% 64.0% 37.4% 83.2% 35.7%
Shared Ride 2 19.0% 7.6% 2.7% 32.3% 19.6% 33.9% 7.1% 31.6%
Shared Ride 3 3.1% 2.8% 2.9% 14.4% 3.1% 16.2% 4.7% 24.9%
Bus 337 211 127 127 2,573 633 42 506
Walk 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%
Bicycle 3.2% 3.2% 3.2% 4.3% 5.3% 11.7% 4.9% 7.6%
Total 0.6% 0.6% 0.6% 0.3% 8.0% 0.8% 0.2% 0.3%
Table 6.4. Revised Calibration Target Trips
Mode Home-Based Work Home-
Based
Shop
Home-
Based
University
Home-
Based
Other
Work-
Based
Other
Other-
Based
Other
Low
Income
Middle
Income
High
Income
Drive Alone 73.1% 86.2% 91.9% 48.7% 64.0% 37.3% 83.1% 35.7%
Shared Ride 2 19.0% 7.6% 2.7% 32.3% 19.6% 34.0% 7.1% 31.6%
Shared Ride 3 3.1% 2.8% 2.9% 14.4% 3.1% 16.2% 4.7% 24.9%
Bus 337 211 127 127 2,573 635 42 506
Walk 4.0% 2.9% 2.0% 4.3% 5.3% 11.7% 4.9% 7.6%
Bicycle 0.7% 0.5% 0.5% 0.3% 8.0% 0.8% 0.2% 0.3%
Total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0%
6.4.2 Transit Trips
While targets for most trip purposes were developed based on the 2012 CHTS, due to the small sample of
transit trips in the statewide travel survey, transit targets could not be developed using this data source.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
6-73
Instead the 2007 transit survey updated to 2016 boardings was used.5 Responses were received from 776
passengers out of 5,451 surveys distributed with about three-quarters (71%) of the survey responses
received from students at Cal Poly. Trips by purpose were summarized from the on-board survey data. Trip
purposes were based on the reported purpose for the trip and whether the trip originated or ended at “home.”
Since home-based work trips will be stratified by income group in the mode choice model, estimates of
home-based work trips by income are also required. Income information was reported for 47 of the 58 home-
based work trips surveyed. Twenty-eight (60%) of the home-based work travelers reported incomes less
than $35,000 per year and 19 (40%) reported incomes of $35,000 or more. The maximum income range
included in the survey was $75,000 or more while starting income for the high income group used for the
travel model is $100,000. Home-based university trips include only those trips made to or from Cal Poly or
Cuesta College. Home-based school trips to middle and high schools were aggregated into home-based
other trips. Table 6.5 summarizes the shares by trip purpose shown in the on-board survey documentation.
Table 6.5. Estimated Shares of Transit Trips by Trip Purpose from 2007 SLO Transit
On-board Survey
Trip Purpose Share Estimated Linked Trips
Home-Based Work 8% 368
Low Income (less than $35,000) 5% 230
Middle Income ($35,000 - $99,999)
3% 138
High Income ($100,000 or more)
Home-Based Shop 3% 138
Home-Based University 61% 2,806
Home-Based Other 15% 690
Work-Based Other 1% 46
Other-Based Other 12% 552
Total 100% 4,600
In addition to the survey data, the most recent bus boardings were used to update the 2007 survey data to
2016. Table 6.6 summarizes the boardings by route by time of day from the 2016 boarding and alighting
count data.
5 Technical Memorandum #4, On-Board Survey Results, Short Range Transit Plan Update, Prepared for City of San
Luis Obispo, California, prepared by Urbitran Associates, Inc., February 2008.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
6-74
Table 6.6. SLO Transit Boarding Counts
Route Total Daily Boardings
1 189
2 384
3 371
4 1,123
5 1,173
6a 945
6b 663
Total 4,848
The resulting transit targets by trip purpose are shown in Table 6.7. To match modeled transit trips close to
observed data, actual observed transit trips were used instead of transit shares during model calibration.
Table 6.7. Transit Targets (2016)
Trip Purpose Share Estimated Linked Trips
Home-Based Work 8% 337
Low Income (less than $35,000) 5% 211
Middle Income ($35,000 - $99,999)
3% 127
High Income ($100,000 or more)
Home-Based Shop 3% 127
Home-Based University 61% 2,573
Home-Based Other 15% 633
Work-Based Other 1% 42
Other-Based Other 12% 506
Total 100% 4,218
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
7-0
7.0 Time of Day
In the time of day model component, the vehicle trip tables by trip purpose from the mode split process are
converted from production-attraction (PA) format into origin-destination (OD) format and factored into time
periods for assignment on the roadway network. The time of day process is not considered a separate step
in the 4-step transportation modeling process, but is instead grouped with the traffic assignment model.
In the remaining traffic assignment model steps, vehicle trip tables by time of day are assigned to the
network using an equilibrium procedure for the two peak hours (AM and PM) and for the off-peak period.
Transit trips are also assigned to the transit route system using network settings consistent with mode
choice.
Based on discussions with City staff and analysis of traffic count data, the AM and PM peak hours were
defined as shown in Table 7.1.
Table 7.1. Peak Period Definitions
Period Name Period Definition
AM Peak Hour 7:00 AM – 8:00 AM
PM Peak Hour 5:00 PM – 6:00 PM
Off-Peak Period All Remaining Time (22 hours)
Directional time of day factors are used to convert trips from production/attraction (P/A) format to
origin/destination (O/D) format and into peak and off-peak time periods used in the model. This process is
based on extensive data indicating that trips are made directionally by time of day. For example, HBW trips
generally occur from the production to the attraction (i.e., from home to work) in the AM peak and from the
attraction to the production (i.e., from work to home) in the PM peak. It is also recognized that some trips are
made in the reverse of this pattern and many trips are made outside of the peak periods.
Trip time data from the CHTS was used to develop directional time of day parameters. Each recorded trip
was categorized by direction and by time of travel. Since some trips may begin in one period and end in
another, trips were placed into time periods based on the trip mid-point.
In the travel model, the factors are applied directly to the purpose-specific vehicle trip tables created by the
mode split model. As described in Chapter 4: Trip Distribution, daily trip tables are separated into peak
period (combined AM and PM peak periods) and off-peak period trips prior to trip distribution and mode
choice. The traffic assignment time of day module further separates peak period trips into AM and PM peak
hour trips. At the same time, all trip tables are converted from P/A format to O/D format.
Time of day factors shown in Table 7.2 demonstrate the portion of trips by purpose and direction assigned to
each time period. These factors are applied in a two stage process: first in a pre-distribution time of day
module and second in a pre-assignment time of day module. The pre-distribution time of day parameters are
shown in Table 7.3. The pre-assignment time of day parameters are shown in Table 7.4.
Pre-distribution time of day factors are computed based on the 24-hour time of day factors. For the off-peak
period, the distribution time of day factor is simply the sum of the PA and AP factors. For the peak period, the
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
7-1
distribution time of day factor is the sum of PA and AP factors for the AM and PM periods. Distribution time of
day factors are applied by simple multiplication of time of day factors and trip tables.
Pre-assignment time of day factors are also calculated based on 24-hour time periods. Because they are
applied to trip tables that have already been separated into peak and off-peak periods, pre-assignment time
of day factors are computed by diving 24-hour factors by the pre-distribution factors for each period and trip
purpose. They are applied to the peak and off-peak PA tables using the equation below. Because EE trips
are not processed through trip distribution or mode choice, EE time of day is applied prior to trip distribution.
EE time of day is computed by simply multiplying time of day factors by the 24-hour EE trip tables.
1
2
1 2 Where: OD trip-table for the
AM or PM hour (or for the off-peak period)
PA trip-table for the peak or off-peak
period Transposed PA trip-table for the
peak or off-peak period Pre-assignment time of
day factor for the PA direction
Pre-assignment time of day factor for the AP
direction Table 7.2. Time of Day Factors Period HBW
HBS HBU HBO WBO OBO IE EE PA AP PA AP PA AP PA AP PA AP Off-Peak 42.7% 38.7% 35.1% 53.9% 38.
8% 41.0% 38.0% 39.0% 53.7% 29.5% 87.2% 82.0% 82.0% AM Peak 10.0% 0.5% 0.9% 0.3% 15.
0% 0.0% 11.1% 1.2% 1.0% 6.0% 4.3% 6.0% 6.0% PM Peak 0.5% 7.5% 4.6% 5.2% 1.
5% 3.7% 3.0% 2.0% 9.6%
0.3% 8.5% 12.0% 12.
0% Table 7.3. Pre-distribution Time of Day Factors
HBW HBS HBU HBO WBO OBO IE Off-Peak
81% 89% 80% 77% 83% 87% 82.0% Peak
19% 11% 20% 17% 17% 13% 18.0% Table
7.4. Pre-assignment Time of Day Factors Period HBW
HBS HBU HBO WBO OBO IE EE PA AP PA AP PA AP PA AP PA AP Off-Peak 52.5% 47.5% 39.5% 60.5% 48.
6% 51.4% 49.3% 50.7% 64.6% 35.4% 50.0% 50.0% 82.0% AM Peak 53.9% 2.7% 8.1% 3.0% 74.
2% 0.0% 64.1% 7.1% 5.8% 35.5% 16.8% 16.7% 6.0% PM Peak 2.9% 40.5% 41.6% 47.4% 7.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
8-2
8.0 Trip Assignment
8.1 Parking Garage Allocation Model
In the San Luis Obispo CBD, many businesses are served by a combination of on-street parking and central
parking garages rather than direct on-site parking. These parking garages are located in three zones, with
the number of parking spaces available shown in Table 8.2. Prior to traffic assignment, a portion of vehicle
trips destined to zones with in ¼ mile of a parking garage are moved from the destination zone to a parking
garage zone.
Table 8.1. Parking Garages
Zone Garage Name Parking Spaces
108002 Morro/Palm Parking Structure 1,227
108003 City of San Luis Obispo Parking 775
162001 Marsh St. Parking Structure 1,988
Trips are allocated to each parking garage independently, using the formula below to determine the relative
likelihood of trips from a zone to be allocated to a parking garage. The number of trips that are allocated is
limited to the number of spaces available in each garage.
Where:
distance to parking garage
calibration parameter alpha (
1) calibration parameter beta (0.4) calibration parameter gamma (8) Due to the close proximity of
the three parking garages, trips from some zones are allocated to multiple parking garages. The model is implemented to ensure
that no more than 100%
of the trips to any
one zone are moved to parking garages. 8.2 Traffic Assignment Algorithms The traffic assignment module loads the
travel demand as represented by the time of day vehicle trip tables onto the roadway network, which is
the supply side of the model. There are several different algorithms that have been use in
past and present models. The methods that were considered are as follows: Equilibrium: This is
the most common method, which assumes all travelers use the fastest possible route between origin
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
8-3
travel time for all trip makers is minimized. This method tends to work best for short assignment periods
in which an equilibrium condition can be defined.
Stochastic Equilibrium: This method considers congestion and assumes that most, but not all,
travelers use the fastest possible route between origin and destination. The stochastic component of this
method represents imperfect knowledge of the roadway system.
All-or-Nothing: This method does not consider congestion and assigns all trips to the fastest possible
route between origin and destination. It is not appropriate for congested networks because it does not
consider congestion effects and thus tends to overload some facilities and under-load others.
Stochastic: This method does not consider congestion and assigns most, but not all, trips to the fastest
possible route between origin and destination. For similar reasons as the all-or-nothing assignment, the
stochastic assignment process is not appropriate for congested networks.
Incremental Capacity-Restrained Assignment: With this method, the vehicle trip table is assigned
incrementally. Network travel times are updated after each increment is assigned, so congestion effects
are considered. With a very large number of increments, this method can approximate an equilibrium
assignment. This method is very efficient and includes consideration of congestion effects. However, it
has largely fallen out of favor because modern computing power allows for more widespread application
of the equilibrium assignment process, which is less efficient computationally but is theoretically a more
valid algorithm.
Because SLO experiences congestion, only the equilibrium and stochastic equilibrium assignment methods
were considered. Based on previous experience, the equilibrium assignment method is preferred over the
stochastic equilibrium method except in cases where specific problems are encountered. Therefore, the SLO
Citywide Travel Model uses the equilibrium traffic assignment method.
8.3 Closure Criteria
When equilibrium traffic assignment is used, oscillations between equilibrium iterations can sometimes result
in unstable assignment results. If closure criteria are not sufficient, two very similar model runs (e.g., with
only one small adjustment to the roadway network) can produce un-intuitive results. This generally occurs
when the equilibrium traffic assignment algorithm converges at a different number of iterations – sometimes
only one apart – for each run. Even when equilibrium traffic assignment converges after the same number of
iterations, alternating oscillations in traffic volumes can sometimes be observed in traffic assignment results
based on slightly different model networks.
While oscillations introduced by the equilibrium traffic assignment procedure can be of concern, they can be
managed through introduction of a very tight closure criterion. Traffic assignment is performed with a closure
gap of 0.0001 (10-4) and a maximum number of iterations of 500. Convergence is reached prior to the
iteration limit of 500. Test model runs have also shown that a closure gap of 0.001 may be acceptable for
some applications.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
8-4
8.4 Impedance Calculations
The impedance used for determining the shortest path in the Traffic Assignment model can take many forms,
but typically it includes one or more of the following – travel time, travel distance, and tolls. If more than one
impedance variable is used, a generalized cost function is necessary so that the relevant variables can be
added together into a single impedance function expression. Since tolls are not an issue in the SLO area,
they were not seriously considered for the impedance function. Furthermore, experience has shown that
distance is less important than travel time; and including distance is problematic because it essentially
amounts to double-counting the emphasis on this variable since distance is also inherent in the travel time
calculations.
Therefore, congested travel time, rather than a generalized cost function, is used in traffic assignment
calculations as is done in numerous models around the country.
An example of the generalized cost function is shown below. This is provided for reference only since the
SLO Citywide Travel Model uses travel time as the single impedance value. Use of a generalized cost
function requires that assumptions are made regarding auto operating costs and the value of time. These
can be difficult to obtain; and both of these values can vary by region and would be subject to adjustment
during model calibration and validation. With only one variable used in the impedance equation for the SLO
Citywide Travel Model, there is no need to convert them to common cost units.
Cost = (Distance * AutoCost) + (Time * TimeCost)
Where:
Cost = Total link cost, or generalized cost
Distance = Link distance
AutoCost = Auto operating cost (in dollars per unit distance)
Time = Congested travel time for link
TimeCost = Value of time (in dollars per unit of time)
8.5 Volume-Delay Functions
A volume-delay function represents the effect of increasing traffic volume on link travel time. While several
volume delay functions are available for consideration, the most commonly used function is the modified
BPR function. The modified BPR function is based on the original Bureau of Public Roads (BPR) equation
below.
1 +
Where:
Congested travel
time Freeflow travel time
V = Traffic volume C =
Highway design (practical) capacity
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
8-5
The modified BPR equation uses the same form, but replaces design capacity with ultimate roadway
capacity. Ultimate roadway capacities for links in the SLO Citywide Travel Model roadway network are
defined in Chapter 1: Roadway Network. The modified function also replaces the coefficient alpha and the
exponent beta with calibrated values that vary by facility type and area type.
The Highway Capacity Manual (HCM 2000) provides alpha and beta parameters for the modified BPR
equation that were developed to be consistent with HCM based delay calculations6. These parameters vary
by facility type, freeflow speed, and signal spacing (signals/mi). Detailed signal location data is not available
on each link in the SLO roadway network and, more importantly, is unlikely to be available on forecast
networks. Therefore, the model uses a facility type and area type lookup table for determination of the
parameters alpha and beta. The parameters are shown in Table 8.2.
The parameters in Table 8.2 were calibrated to the extent possible, but the little congestion that is present is
not particularly suitable for a rigorous calibration. On the other hand, since the parameters were developed to
be consistent with HCM-based delay calculations, the parameters are appropriate given the speed and
capacity assumptions by area type and functional class.
Table 8.2. Volume Delay Parameters Alpha and Beta
Functional Class CBD Fringe Urban Suburban Rural Outside
SOI
Freeway 1.5 12 1.5 12 1.5 12 1.5 12 1.5 12 1.5 12
Principal Arterial 0.4 5 0.7 5 0.7 5 0.45 5 1.2 5 1.2 5
Minor Arterial 3.5 5 0.6 5 0.6 5 1 5 1.5 5 1.5 5
Major Collector 3 5 0.95 5 0.95 5 1 5 1.5 5 1.5 5
Ramp 3.7 5 3.7 5 3.7 5 3.7 5 3.7 5 3.7 5
Urban Local 3 5 0.95 5 0.95 5 1 5 1.5 5 1.5 5
Highway (Outside SOI) .4 5 0.7 5 0.7 5 0.45 5 1.2 5 1.2 5
Arterial (Outside SOI) .4 5 0.7 5 0.7 5 0.45 5 1.2 5 1.2 5
Rural Arterial (Outside SOI) .4 5 0.7 5 0.7 5 0.45 5 1.2 5 1.2 5
Centroid Connector 0 1.1 0 1.1 0 1.1 0 1.1 0 1.1 0 1.1
8.6 Transit Assignment
Unlike vehicle trips, transit trips are not converted from PA to OD format. Instead, transit trips are assigned in
PA format. In addition, transit trips are assigned for a peak period (AM and PM) and an off-peak period.
Transit trips are not assigned separately for AM and PM peak periods. While this is considered standard
practice, it does require that transit assignment results are evaluated with this in mind.
TransCAD reports transit stop boarding and alighting data at the stop level and transit route volumes at the
segment level. This information can be useful in evaluating the relative importance of different stops along a
6 Highway Capacity Manual, Transportation Research Board, 2000. p. 30-39, Exhibits C30-1 and C30-2.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
8-6
route and of different route segments. However, it is important that transit assignment results at this level of
detail are viewed critically. Where stop-specific data is available, base year model results should be
compared to base year observed data; this information should be taken into account when evaluating model
forecast data.
Because trips are assigned in the production to attraction direction, it is important to use caution when
viewing directional transit assignment results. In particular, one must use care when reviewing transit
assignment results on loop routes. A loop route that connects a residential area to an activity center (e.g.,
university or employment center) may show high ridership on one side of the loop and low ridership on the
other side of the loop.
A comparison of observed boardings and assigned boardings by route is shown in Table 8.3. The overall
modeled transit ridership matches the observed closely. The largest differences are on routes 4, 6a, and 6b,
all of which serve the Cal Poly campus.
Table 8.3. Modeled and Observed Transit Boardings by Route
Model Boardings Survey Boardings Difference
Route 1 26 189 (163)
Route 2 308 384 (76)
Route 3 494 371 123
Route 4 1,387 1,123 264
Route 5 1,358 1,173 185
Route 6a 418 945 (527)
Route 6b 651 663 (12)
Total 4,642 4,848 (206)
8.7 Bicycle Assignment
Bicycle assignment is a new component of the SLO Citywide Travel Model. It is performed similar to the
traffic assignment and is based on the shortest travel time between an origin and a destination zone.
However, the travel time used in the bicycle assignment is calculated not just based on the link distance but
the type of facility. A Bicycle Comfort Index is assigned to each link to capture how a bicyclist perceives the
level of safety and enjoyment on that facility. Dedicated bicycle trails are assigned the highest level of
comfort while large arterials with no bicycle lanes are assigned the lowest level of comfort. Comfort levels by
facility are shown in Figure 8.1.
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
8-7
Figure 8.1. Bike Comfort Levels
Also unlike traffic assignment, bicycle assignment uses an all-or-nothing assignment algorithm because
travel times are unaffected by the number of bicycles on a network link (i.e., there is no bike congestion). The
bicycle assignment uses a more detailed highway network that includes dedicated bike facilities as well as all
local streets, which are omitted from traffic assignment.
The results of the bicycle assignment were validated using bike counts collected by the City of San Luis
Obispo along with Strava data. Strava data includes records of bike trips made by the GPS-based Strava
app users. Strava is used by both bicyclists making both recreational and commute trips, and requests that
users mark each trip as either recreational or commute. The Strava dataset, limited to commute trips only,
was attached to the model network to provide information on bicycle trip patterns.
Strava data contains only a sample of bicycle trips and may have inherent biases (e.g., by income or student
status). To address these concerns, the Strava dataset was mapped alongside bicycle counts. The Strava
San Luis Obispo Citywide Travel Model
Cambridge Systematics, Inc.
8-8
dataset was then scaled to roughly match bicycle count totals. While this dataset is still not sufficient for a
rigorous validation of commute/utilitarian bicycle trips represented by the model, it provided valuable
information on the magnitude and locations of bicycle trips with in the city. A visual comparison of assigned
bicycle trips to Strava data allowed the following adjustments to the bicycle model:
1. The total number of bicycle trips specified in the mode choice targets were adjusted to better match the
overall bicycle activity indicated by the combined dataset.
2. Bicycle comfort index values were adjusted so that the model assigned more bicycle trips to heavily used
facilities and fewer bicycle trips to less frequently used facilities.
The resulting bicycle assignment provides information on bicycle activity in the base year, and may be useful
in better understanding how different development plans or infrastructure investments impact cyclists.