MissoulaTDME Full Documentation 07051

Model User’s Guide
September 2011
Missoula MPO Travel Demand Model Evaluation and Enhancement
Table of Contents
CONTEXT AND BACKGROUND ....................................................................... 1
SYSTEM REQUIREMENTS ................................................................................ 1
INSTALLING THE MODEL ADD-IN ..................................................................... 2

Recommended Method: Running the Installation Program..................................................................... 2
Alternate Method: Manual Installation .................................................................................................... 3
Data Installation ........................................................................................................................................ 4
Directory Structure ................................................................................................................................... 4
RUNNING THE MODEL ................................................................................. 6

Collecting the Required Data .................................................................................................................... 6
Creating and Running a Scenario .............................................................................................................. 7
MODEL UTILITIES – ROADWAY NETWORK ........................................................ 12

Add/Delete Network Year ....................................................................................................................... 12
Refresh Network Defaults ....................................................................................................................... 13
Edit Network ........................................................................................................................................... 14
Copy Feedback Results............................................................................................................................ 14
Create Select Query ................................................................................................................................ 14
MODEL UTILITIES – ROADWAY NETWORK ........................................................ 15

Link/Verify Route System ........................................................................................................................ 15
Edit Routes .............................................................................................................................................. 15
Constant Calibration ............................................................................................................................... 15
MODEL UTILITIES – MAPS AND REPORTS .......................................................... 15

Create Performance Report .................................................................................................................... 15
Create Maps ............................................................................................................................................ 16
MODEL DATABASE .................................................................................... 17

Database Approach ................................................................................................................................. 18
Database Interface .................................................................................................................................. 19
Database Tables ...................................................................................................................................... 21
TRANSPORTATION NETWORKS ...................................................................... 22

Transportation Network Components .................................................................................................... 22
Example Transportation Network Application ....................................................................................... 23
User’s Guide i
WORKING WITH TRANSIT DATA .................................................................... 33

Editing the Route System ........................................................................................................................ 33
Viewing Transit Assignment Results ....................................................................................................... 36
LIST OF TABLES
Table 1: System Requirements ..................................................................................................................... 2
Table 2: Model Input Files............................................................................................................................. 6
Table 3: Map Types ..................................................................................................................................... 16
Table 4: Map Options.................................................................................................................................. 17
Table 5: Access Database Tables................................................................................................................. 21
Table 6: Transit Assignment Results ........................................................................................................... 36
LIST OF FIGURES
Figure 1: Example Model Directory Tree ...................................................................................................... 5
Figure 2: The Scenario Toolbox ..................................................................................................................... 8
Figure 3: The Scenario Editor (Input Tab) ..................................................................................................... 9
Figure 4: The Scenario Editor (General Tab) ................................................................................................. 9
Figure 5: The Scenario Editor (Output Tab) ................................................................................................ 10
Figure 6: The Scenario Editor (Advanced Tab) ............................................................................................ 10
Figure 7: The Model Dialog Box .................................................................................................................. 11
Figure 8: Sub-Steps Dialog Box ................................................................................................................... 11
Figure 9: Add/Delete Network Year Dialog Box .......................................................................................... 13
Figure 10: Access Database Form Interface ................................................................................................ 20
Figure 11: Data Scenario Management ...................................................................................................... 20
User’s Guide ii
Context and Background

This User’s Guide provides instructions on operation of the Missoula MPO Travel Model. Information is
provided regarding installation of the model, management of model scenario data, and running of the
model.
The model is run from the TransCAD software platform through a customized user interface. This
interface provides access to custom calculations developed specifically for the Missoula MPO. Scenario
and file management is achieved through a scenario management system integrated into the custom
user interface. A basic understanding of the TransCAD software program is required to get the most out
of the model. However, users unfamiliar with the software should be able to perform some modeling
tasks with the assistance of this guide.
Some of the travel model data is stored in an Access database. The database contains tabular
information such as trip rates, socioeconomic data, and other TAZ-specific data. The Access database file
feeds this information to the TransCAD based travel model. Guidance on use and maintenance of the
database is also included in this User’s Guide.
System Requirements
The model must be run on a computer running Windows XP or Windows 7 and the TransCAD software
program. Specific system requirements are shown in Table 1.
The listed requirements are suggested minimums; a computer that does not meet these requirements
may still succeed in running the model. Increased processor speeds, multiple processor cores, and
additional memory will reduce the amount of time required to run the model. The required disk space
for installation must be available on the drive where TransCAD has been installed. The required disk
space for additional scenarios can be on a local or network drive and must be available before
attempting to run the model. However, model run times will be considerably longer if the model is run
from a network drive instead of a local drive.
User’s Guide 1
Table 1: System Requirements
Windows XP or Windows 7
Operating System
A 64-bit operating system is recommended.
Intel Core 2 processor or later
Processor
Note: Multiple cores will significantly improve model run times.
4GB – 12 GB
Memory
Note: At least 8 GB of memory is recommended.
TransCAD Software Version 5.0 r4 (Build 2020 or later recommended)
Version 2007 or later
Microsoft Office (including Access)
(Version 2003 will work with reduced functionality)
Disk Space
35 MB
(Installation and input data)
Disk Space 10 GB for each scenario (Without speed feedback)
(Each scenario) 17 GB for each scenario with speed feedback and saved intermediate files
Installing the Model Add-In

Recommended Method: Running the Installation Program
To install the model, run the provided Setup.exe file. If the model has been previously installed, the
installation program will update the model to the most current version. The installation program will not
overwrite custom scenario lists created by the user.
The model setup file contains an option to install model data as well as the model add-in files. If data is
selected for installation, data in the C:\Missoula Model directory will be overwritten.
To access the Add-In, choose Tools  Add-Ins Missoula Model from the TransCAD menu. Once the add-
in has been used once, Missoula Model will be available in the recently used Add-Ins list shown directly
under the Tools menu.
The installation program does not provide an uninstall function. To uninstall the model, use the
following steps:
1. Delete the “Missoula Model” folder from Program Files (Usually C:\Program Files\Missoula
Model, or C:\Program Files (x86)\Missoula Model on a 64 bit system),
2. Choose Tools  Setup Add-Ins… from the TransCAD menu and remove the entry for the
Missoula model,
3. Remove any data (as desired) from local or network drives.
4. Remove LSA\Missoula Model directory from the All Users Application Data folder (Note this step
is optional, as these files use very little disk space)
User’s Guide 2
Removal of the program files and user settings may delete scenario lists created by the user.
Alternate Method: Manual Installation

In most cases, the model should be installed using the installation program described above. However,
there may be some cases where the user prefers to install the model by copying files manually. The
installation program performs these steps as described below. The install program does not modify the
Windows Registry. The manual installation procedure may be necessary when installing the model on
computers with non-standard security or folder configurations.
The travel model add-in can be installed by copying the “Manual Install \Missoula Model” folder from
the installation media to the local Program Files directory. The “Missoula Model” folder can also be
copied from a computer where the model has already been installed.
If the installation media is located at drive D:\, the OCTAM RTM folder can be found at
D:\Manual Install\Missoula Model.
The directory should be placed in C:\Program Files\Missoula Model on a 32-bit system, or
C:\Program Files (x86)\Missoula Model on a 64-bit system.
o If necessary, these files can be installed in an alternate location, as may be necessary on
MDOT computers.
The user interface retains certain settings so that the user does not need to re-specify information each
time the model interface is started. These settings are stored in a user-specified location. The automatic
installation program installs these files in the “LSA\Missoula Model” directory within the hidden all users
Application Data folder. The file named Appdata.txt in the OCTAM program folder must contain a single
line containing the path of this directory. The line must end with a backslash. By default, settings are
stored in the following location:
Windows XP: C:\Documents and Settings\All Users\Application Data\LSA\Missoula Model

Windows 7: C:\ProgramData\LSA\Missoula Model\
It is possible to store the user settings in any location, but it important to store the settings in a user-
writeable location. This is particularly important when installing the model on a Windows 7 machine or a
machine with non-standard security settings. Prior to running the model from the TransCAD interface,
the directory specified in the Appdata.txt file must exist and must contain the DefaultScenario.ini file
provided with the installation media.
Once the add-in files have been copied, the model must be added to the TransCAD Add-Ins list. To do
this, open TransCAD and:
1. Select Tools  Setup Add-Ins... from the menu,

2. Click “Add” to add a new Add-In,
3. Fill the data in as follows:
User’s Guide 3
a. Type: Dialog Box

b. Description: “Missoula Model”
i. Note: This value can be adjusted based on user preference
c. Name: “Missoula Model”
i. Note: this must be exact and is case sensitive
d. UI Database: Click “Browse…” and locate
C:\Program Files\Missoula Model\model_ui.dbd
i. This location will vary if the program was copied to an alternate location
e. In Folder: None (leave blank)
4. Click “OK”
Data Installation
On initial installation, the installation program may provide the option to automatically install the model
data files. However, this option will be removed after initial delivery to prevent accidental replacement
of user-modified data. Instead, data files can be manually copied from the installation media as
described below. This procedure can also be used to copy model data files when installing the model
manually.
To copy model data files from the installation media, copy the “Data Files\Missoula Model” folder from
the installation media to a location on the local computer. To maximize performance, it is not
recommended that the data files are stored on a network location. However, it is important to make
regular back-ups of model data files.
Directory Structure
The example directory tree shown in Figure 1 is structured to provide efficient and straightforward
organization of travel model input and output files. However, TransCAD and the customized user
interface are flexible enough to allow for nearly any directory structure.
User’s Guide 4
Figure 1: Example Model Directory Tree
User’s Guide 5
Running the Model

The model is controlled through a series of dialog boxes. These dialog boxes allow the user to specify
custom model run settings or to copy settings from a previously defined scenario. Users may also run
the travel model, create reports and maps, and specify model run options. Steps required to complete a
successful model run are described below.
Collecting the Required Data

To successfully run the travel model, various data files are required. Some input files are optional and
will provide additional functionality. Each file is identified by a short keyword as identified in Table 2. All
input files should be collected and placed in a model input directory. Input files will not be modified
when the model is run.
Table 2: Model Input Files
Required /
ID Description and Notes
Optional
Network The Roadway Geographic File. Required
A turn penalty file can be identified to enable specific turn penalties. If this
file is not present, no turn prohibitions or penalties will be applied. If used,
TurnPen Optional
this file must be formatted as described in the TransCAD software
documentation.
The Model Database contains various information items and is further
Database Required
described later in this document.
The TAZ geographic file is not used directly by the model, but must be
TAZ Required
included to support automated mapping.
Optional
KFAC K-factor matrix file. (use not
recommended)
Routes A TransCAD Route System based on the roadway network geographic file. Required
Select link/node query file. If this file is present, select link analysis will be
SelQry Optional
performed when traffic assignment is run.
OPpaths
OUE path files used for traffic assignment warm start. (Not currently used –
AMpaths Optional
may be enabled for TransCAD 6)
PMpaths
User’s Guide 6
Creating and Running a Scenario

After the input data has been collected, a scenario must be defined from the model dialog box. Model
scenarios are accessible from the scenario toolbox and contain information about the following for each
scenario:
Input and output directories,

Filenames,
Network year/alternative,
Data year/alternative,
Individual alternatives, and
Advanced settings and parameters.
Scenarios can be copied based on existing scenarios or can be created from scratch using default
settings. Figures 2 through 6 show the scenario toolbox and editor that are used to manage scenarios
along with annotations describing the available functions.
When creating or editing a scenario, use the steps listed below. It is recommended that these steps are
performed in order.
1. Specify a scenario name and identify the scenario input and output directories.
2. As necessary, identify input files by name. Most files will be found automatically, but some files
may need to be located manually.
3. After the status for all required files is shown as “Exists,” edit the scenario settings on the
General tab. Note that network and data year settings do not need to match. It is possible to run
a scenario based on the 2010 roadway network and 2040 socioeconomic data.
4. Optional: Review the output filenames and modify if desired.
5. Optional: Review the advanced settings and modify if desired.
WARNING: The Advanced tab in the Scenario Editor allows the user to edit values that
are not often changed. The advanced interface does not prevent the user from
entering invalid or inconsistent data, which may cause the model to crash or produce
invalid results.
User’s Guide 7
The model dialog box, shown in Figure 7, provides a great deal of flexibility in how the model is run, but
in most cases a very simple approach can be taken.
To run a standard, complete model run simply start the model dialog box, create a scenario, and
click on Step 1 – Prepare Networks. The model will be run with the standard default settings.
To automatically create a performance report when the model run is complete, select the
appropriate checkbox.
If buttons are grayed out and cannot be used, this is usually due to missing input files or invalid
settings.
Figure 2: The Scenario Toolbox
The status will read

All scenarios in the “Missing,” “Ready”
scenario file are listed “Partial,” or “Done.”
here. Double click a The model cannot be
scenario to edit it. run if the selected
Select one or more scenario has a
scenarios before “Missing” status.
running the model.
Change the order in

which scenarios are
Add, copy, and delete displayed.
scenarios using these
buttons.
Load or save a scenario
list.
Create a new blank
scenario list. Show the current
scenario filename.
User’s Guide 8
Figure 3: The Scenario Editor (Input Tab)
Enter a
scenario name.
All input files are
searched for in the
input directory
Identify the scenario
when the input
directories.
directory is changed.
Filenames and file

When a file is
status are displayed
selected, its
here. Double-click
description will
an item to change
be shown here.
the filename or
location.
Figure 4: The Scenario Editor (General Tab)
A description
of the scenario
can be entered
here.
Choose
assignment Set the network
settings year, data year, and
individual
The default alternatives
settings are
appropriate for
most uses
Set speed feedback
options
Select mode
choice or split
User’s Guide 9
Figure 5: The Scenario Editor (Output Tab)
Different
model stages
are listed here.
Files for the
selected stage Filenames and file
are shown. status are displayed
here. Double-click
an item to change
the filename or
When a file is location.
selected, its
description will Note: Files will be
be shown here. missing until the
model has been run.
Figure 6: The Scenario Editor (Advanced Tab)
Available data is
shown here. Some
data can be edited
Different directly in the grid.
model stages Arrays will be edited
are listed here. in a separate dialog.
Subarray data can

be displayed by
Tables, clicking in a cell and
Parameters, or selecting Edit…
Access Data
(i.e., table
names in the
This button will
access
reset all parameters
database) can
currently shown
be selected
(including subarrays)
here.
to default values
Note: Advanced model parameters should not typically be changed.
User’s Guide 10
Figure 7: The Model Dialog Box
If checked, only
the selected
A performance
model step will
report can be
be run. If
created when the
unchecked,
model run is
subsequent
complete.
steps will be
run as well.
Debug mode
disables error
Click on a model handling. This can
step to run that help with
step. troubleshooting, but
prevents TransCAD
Sub-steps can be from “cleaning up”
deactivated. after a crash.
Utilities can be run

from this area.
Running Selected Model Steps
The user interface can be set to run only selected model steps or sub-steps. To run only a single step,
click the “Stop after each step” checkbox in the main model dialog box. When this box is checked, the
selected step will be run, but subsequent steps will not. When this checkbox is cleared, subsequent
steps will be run automatically.
To exclude certain sub-steps or to run only selected sub-steps, the dialog shown in Figure 8 can be used.
By clicking on the button to the left of each model step, the user can enable or disable specific steps.
The behavior of the “Stop after each step” checkbox is not changed when sub-steps are enabled or
disabled.
Figure 8: Sub-Steps Dialog Box
User’s Guide 11
Running Speed Feedback
Speed feedback can be enabled for a scenario from within the scenario editor. When enabled, speed
feedback will only be run if:
1. The model is not set to stop after each step, and

2. The model is started from Step 1, 2, or 3.
Otherwise, the model will be run as if the selected scenario is not set to run speed feedback. When the
model is run with speed feedback enabled, a file named “Feedback.txt” is created in the model output
directory. This file is updated as the model runs and contains a history of speed feedback convergence
process. The file can be used to determine whether speed feedback has converged successfully or
additional iterations are needed. Furthermore, the file can be opened while the model is running to
check speed feedback convergence progress in real time.
When performing alternatives analysis exercises, it is often preferable to run the model without
including the effects of speed feedback. In such cases, it is still important that trip distribution patterns
are consistent with a baseline scenario (e.g., an existing plus committed model run). Running the travel
model with speed feedback enabled also requires considerably longer than running the model with
speed feedback disabled. It is possible to run the model without speed feedback and utilize speed
feedback results from a previous model run to produce consistent trip distribution results. To do this,
follow the steps listed below:
1. Run a complete model run with speed feedback enabled.

2. Use the Copy Feedback Results utility to save resulting speeds to the input network file.
3. Create a new scenario that uses similar roadway and socioeconomic data assumptions.
a. This scenario should reference the same network year as the original run.
b. The scenario may include network alternatives or changes to socioeconomic data.
4. Set the new scenario to run without speed feedback and without initializing speeds.
5. Run the new scenario.
Model Utilities – Roadway Network

The model dialog box includes several utilities that can be used in preparation of model inputs. These
utilities, described below, will only be available if all required input files for a scenario have been
identified and are present.
Add/Delete Network Year

The model roadway network is designed to contain data for various distinct scenarios. This tool will
allow network years to be added or deleted and can be operated as described below.
1. Select a model scenario that references an input network. The referenced input roadway
network will be modified.
User’s Guide 12
2. Click the Add/Delete Network Year button in the main model dialog box (Input tab); the dialog
box shown in Figure 9 will appear.
3. To add a network year:
a. Select a year from the drop-down list.
b. Click the Copy button. The tool will make an exact copy of the selected year. If the
Alternatives option is enabled, you will be prompted to select alternatives to be
included in the new network year.
c. Attributes for the new network year can be modified by opening the network file and
using tools available in the TransCAD software.
4. To delete a network year:
a. Select a year from the drop-down list. Note that the base year network cannot be
deleted.
b. Click the Delete button. The tool will delete all data fields associated with the selected
year.
Figure 9: Add/Delete Network Year Dialog Box
Network years can contain up to four digits. A recommended practice is to use a two to four digit code
formatted as follows:
YYXX
Where YY represents the network year (e.g., 09 for 2009 or 40 for 2040) and XX is an optional descriptor
(e.g., 09A, 09B, 40A, 40B).
Refresh Network Defaults

When opening the roadway network geographic file, TransCAD displays the network using a color theme
based on link facility types. When displaying the node layer, centroid connectors, park and ride nodes,
and other nodes are each displayed differently. If the default settings on a roadway network file stop
working correctly, the Refresh Network Defaults utility will restore the default settings. This utility
operates on the input roadway network for the selected scenario.
User’s Guide 13
Edit Network
This utility creates a map containing the roadway network and route system, and then opens the
network editing toolbox. This map is set up and formatted to facilitate making changes to the roadway
network layer. This utility automates the process of setting up a workspace to edit the roadway
network. The resulting map operates on the input network associated with the selected model scenario.
Copy Feedback Results

This tool will copy speed feedback results from a completed model run to the input roadway network
file. This will allow a subsequent model run to produce trip distribution results that are consistent with
the completed model run as described above in the section titled Running Speed Feedback.
Create Select Query

A select link or node query file (*.qry) can be created for a scenario using the Select Link/Zone Query
Builder provided with the TransCAD software. This toolbox, accessed from Planning  Assignment
Utilities  Select Link/Zone Query Builder, is explained in the TransCAD software documentation. This
tool interactively guides creation of a query file. However it cannot be used to create a select zone query
based on a node selection set. The Create Select Query tool can be used to create a select zone query
based on a node selection set. To use this tool, follow the steps listed below.
1. Add the attributes as needed to the input network node layer (e.g., use a subarea ID).
2. Create a scenario that references the modified input network and select this scenario.
3. Click the Create Select Query button in the main model dialog box (Input tab). The system will
prompt the user if an existing select link/query file is specified for the selected scenario.
4. Enter a name for a new select zone query.
5. Select the query method:
a. To or from: Track trips departing or arriving,
b. From: Only track departing trips, or
c. To: Only track arriving trips.
6. Enter a selection condition when prompted.
7. When prompted, choose whether to add an additional query to the query file.
Once the query file has been created, it can be viewed and edited using TransCAD’s Select Link/Zone
Query Builder or can be used as input to a travel model scenario.
User’s Guide 14
Model Utilities – Roadway Network

Link/Verify Route System
This utility links a route system to a particular roadway geographic file and checks the route system for
errors. More information on this tool can be found in the section titled Editing the Route System.
Edit Routes
The Edit Routes utility creates a map containing the roadway network and route system, activates the
route system layer, and then opens the network editing toolbox. This utility also links and verifies the
route system identified in the selected model scenario. The resulting map is set up and formatted to
facilitate making changes to the route system. This utility automates the process of setting up a
workspace to edit the route system.
Constant Calibration
The Constant Calibration utility runs the mode choice model and compares results to a set of mode
targets stored in the model database. The utility then adjusts mode constants and re-runs mode choice
iteratively until model results are consistent with mode choice targets, or an iteration limit is met. The
utility should only be run for scenarios where the roadway network, route system, and socioeconomic
data reflect conditions consistent with the mode choice target values stored in the database.
This utility makes changes to the input database and should only be used when re-calibrating the base
year model. The model database should be backed up prior to running this utility.
Model Utilities – Maps and Reports

The model contains mapping and reporting utilities that can be used to produce additional model
outputs and summary data. These tools, described below, will only be available if all selected scenarios
have been successfully run and read “done” in the status column. These utilities can only operate on
one scenario at a time and will be disabled when multiple scenarios are selected.
Create Performance Report

This utility creates an HTML performance report that summarizes model results. The user may select to
create some or all of the various reports and can select one or more geographic areas to summarize. The
report contains a linked table of contents that can be used to immediately jump to a specific section of
the report.
User’s Guide 15
Create Maps
The Create Maps utility creates maps that display model results. The utility opens a new toolbox and can
only be opened when a single completed model scenario is selected. The toolbar allows the user to
specify a map type and then choose options described in Tables 3 and 4. Once settings have been
chosen, the Create button sill generated the specified map.
Table 3: Map Types
Map Category Map Type Description

The validation map displays model volumes and traffic counts. Volumes are
shown in thousands, with counts also shown in thousands, but placed in
parentheses. Model results are shown on all links, while count data is only
shown on links where validation traffic count data is available. If the
Validation
highlight option is selected, a color-coding scheme will identify model data
that is similar, higher than, or lower than count data.
This map is only meaningful when created for a base year model run.
The volume map displays model volumes in thousands. The user can opt so
Volume
display daily, peak hour, or off-peak volumes.
Traffic Maps The LOS map displays a planning-level peak-hour level of service estimate
LOS
based on 24-hour volumes.
The select link/node map displays the results of a select link or node
Select Link/Node analysis. This map can only be created successfully if select link or node
analysis is enabled for the selected model run.
The traffic comparison displays the differences between two model runs.
Traffic
When creating this map, the user will be prompted to select a scenario for
Comparison
comparison.
The volume/capacity map displays the volume to capacity ratio using a
Volume/Capacity color theme. While similar to the LOS map, this map uses modeled volume
and capacity data rather than a planning-level analysis.
The transit stops map displays the total activity at each transit stop location.
Transit Stops In places where multiple stops are adjacent to a single network node,
results are combined to show the total node activity.
The route flows map displays transit route flow data. Flow is not combined
Transit Maps Route Flows
in places where more than one route segment share a link.
The aggregated flows map displays aggregated transit flow data. In places
Aggregated Flows where more than one route segment share a link, the map displays the sum
of ridership data for all co-linear routes.
User’s Guide 16
Table 4: Map Options
Map Category Map Option Description

If selected, the map will display NCHRP-255 adjusted volumes instead
NCHRP
of raw model results.
If selected, the map will include traffic volume labels. If left blank, the
Volumes
model will not include labels.
If selected, the map will show centroid connectors. If left blank,
Connectors
centroid connectors will be hidden.
If selected, a validation map will differentiate between links where
Traffic Maps
Highlight model results are similar to, higher than, or lower than traffic count
data. This option can only be selected when creating a validation map.
If selected, volume labels will be plotted on centroid connectors.
Label Connectors However, labels will only be shown in the Volumes option is also
selected.
If selected, volume labels will be plotted on local streets. However,
Label Local
labels will only be shown in the Volumes option is also selected.
Transit Maps Bandwidths If selected, a bandwidth theme will represent aggregated flow totals.
Model Database
The model requires a large and varied set of input data for each mode run. Specific data items are
required inputs for each step of the travel modeling process. The data is contained in three primary
places:
Spatial Data: The roadway line layer and route system contain the supply side information used
by the travel model. The TAZ layer is also input to the travel model, but zone data is not stored
directly in the TAZ layer.
Model Database: The model database contains socioeconomic data and other demand side
information used by the travel model. The database also contains model parameters such as trip
rates and other zonal data such as area type.
Scenario Manager: Some model parameters are stored directly in the scenario manager. Only a
few of these parameters need to be changed in normal use of the model.
This chapter provides a detailed description of the data and parameters contained in the model
database.
User’s Guide 17
Database Approach
The Missoula MPO Travel Model relies on a large amount of data and numerous parameters and lookup
tables. The TransCAD software provides a table format that can be used to store this type of
information. The TransCAD table format is relatively efficient, very stable, and allows for sufficient
precision in storage of decimal numbers. This format, Fixed Format Binary (FFB), has been used to store
all data output from the travel model in table format. However, an Access database has been used to
store the majority of data that is input to the model. The Access format has been used rather than the
FFB format for the following reasons:
The TransCAD table format cannot be read or edited except with the TransCAD software;
The Access database can be used to store nearly all of the input data required for the travel
model. This prevents the need to manage a large number of input files that contain data for
various model steps;
SQL queries within the Access software can be used to transform data from a human-readable
format into a format that is readily used by the travel model; and
The Access database format is designed to allow multiple data scenarios to be managed within a
single consolidated database file.
The model has been designed to support two types of scenarios: network scenarios and data scenarios.
Network scenarios are stored in the TransCAD geographic line layer, while data scenarios are stored
within the model database. A virtually unlimited number of data scenarios can be maintained within a
single database, but in practice it may be useful to maintain different databases for different purposes.
For example, one database may be desired for use in the regional planning process while a different
database could be maintained to facilitate testing of minor socioeconomic data alternatives associated
with proposed development.
The database contains some information that is static (does not change when a different data scenario is
selected) and other data that is dynamic (varies by data scenario). The static and dynamic data items are
listed below. A detailed description of each data item is provided in the sections that follow.
Static Data:
Roadway Parameters (lookup tables by facility type and area type)

Household Size, Income, and Worker Disaggregation Curves
Trip Generation Rates (production and attraction rates)
Friction Factors (gamma parameters)
Terminal Times
Mode Choice Parameters
Time of Day Parameters
User’s Guide 18
Dynamic Data:
Socioeconomic Data
Regional Bivariate Data (household size and income)
Other TAZ Data (e.g., area type, parking cost, districts)
Special Generator Data
External Station Data
Database Interface
When opened, the model database will present the user with a request to enable VBA macros. Once
macros are enabled, the database interface form will appear. This form provides automated
management of data scenarios and guided access to key datasets. The interface is annotated in Figure
10.
To modify dynamic data for a specific data scenario, set the active scenario to the desired year and open
the scenario specific datasets from the main interface dialog box. Data can be edited directly in Access.
Alternately, data can be copied from Access and pasted in Excel. Once data has been modified, it can be
pasted back into the Access database.
Not all datasets can be accessed directly from the database interface form. Some datasets are only
edited during a model update and re-calibration and can be accessed by opening data tables directly.
The user interface provides functionality that will copy dynamic datasets to a new data scenario, create
a new blank data scenario, or delete an existing data scenario. The dialog box that provides this
capability is accessed using the Manage Data Scenarios button and is shown in Figure 11. Once a new
dataset has been created, socioeconomic data, special generator data, and external station data can be
modified for the new scenario. Bivariate data can also be edited, but does not need to be changed in
most cases.
User’s Guide 19
Figure 10: Access Database Form Interface
These buttons
can be used to
import data
from a
properly Use these
formatted buttons to
spreadsheet. interactively
Example edit roadway
spreadsheets lookups and
are provided trip rates/
with the model
These buttons Use this button

will open data to add or
tables delete data
containing scenarios.
dynamic The active
datasets. With scenario and a
the exception description are
of EE data, also shown in
these tables this section.
can be edited.
When loading
data, browsing
will start from
the default
directory.
Figure 11: Data Scenario Management
User’s Guide 20
Database Tables
Information is stored within tables in the Access database file. A list of the tables and a description of
the contents are included in Table 5. For some tables, SQL queries are used to re-format data from
human-readable format to a format compatible with the model. SQL queries are also used to filter
dynamic datasets to show data only for the selected year.
All tables that contain model data are prefixed with the letter a. Queries based on tables use the same
name as the source table, but include a suffix consisting of an underscore and a number (e.g., _1).
Tables, queries, and forms prefixed with an x, y, or z are present only for use with the program interface
and are not listed in Table 5.
Table 5: Access Database Tables
Category Table Name Description

aRoadway Lookup Roadway lookup tables based on link facility type and area type.
Roadway Network
aRoadwayVariables List of variables contained in the lookup table.
aSEData Household and employment data at the TAZ level.
Socioeconomic Data Data at the TAZ level in addition to household and employment
aZoneData
data.
aDisaggIncome
Household size and income disaggregation curves.
aDisaggSize
Regional bivariate distribution of households by income and size.
Household Models
Note: the model adjusts the regional bivariate distribution of
aRegBivarPct
households from this table so that the total number of
households is consistent with the data specified in aSEData.
aAttractionRates
Production and attraction rate tables
aProductionRates
Trip Generation Special generator table. Note: null values indicate results from
aSpecialGen
trip generation will be retained
aTripRateFactors Jurisdictional trip rate factors.
aEETrips External/External trip table.
External Trips Internal/External and External/Internal trip-ends at external
aIETrips
stations.
Trip Distribution aFrictionFactors Friction factor gamma parameters.
aModeCoefficients Mode choice model coefficients.
aModeConstants Alternative specific constants.
Mode Choice aModeNests Mode choice model nesting coefficients.
Mode targets. Note: This table is only used by the mode choice
zModeTargets
calibration utility.
Traffic assignment loading factors. Note: All factors in this table
aLoadingFactors are set to 1.0. The Missoula MPO Model does not use traffic
assignment loading factors.
Assignment
Planning-level LOS capacities for use in post processing and
aLOS_Cap
mapping.
aPeriodFactors Directional time of day factors for use in PA to OD conversion.
User’s Guide 21
Transportation Networks
The Missoula MPO Model requires both roadway and transit networks to operate. These networks are
described in detail in the model documentation. In order to work with the roadway and transit
networks, it is important to understand the connections and relationships between different
components of the transportation networks and the TAZ data.
Transportation Network Components

The transportation network is made up of the components listed below:
Links: Links are stored in the line layer in the roadway geographic file. Different types of links
include:
o Active roadway links: Each roadway segment is represented by a link (or line) in the
roadway network layer.
o Non-Motorized links: Non-motorized links in the roadway geographic file represent
facilities that allow bicycle and pedestrian use, but are not open to motorized vehicles.
Non-motorized travel is also permitted on motorized links unless specifically prohibited.
o Inactive links: The roadway geographic file can also contain roadway and non-motorized
links that are inactive for a given scenario. These links are ignored by the model.
o Centroid Connectors: Centroid connectors attach centroid nodes to roadway links
Nodes: Nodes are attached to links. Types of nodes include:

o Centroids: Centroid nodes connect to the model database and the associated
socioeconomic data. These nodes are identified by a non-zero value in the “ZONE” field
on the node layer. Zone numbers must be unique to each zone, but do not necessarily
need to be sequential.
o External Stations: External stations are similar to centroids, but are located on the
model boundary. Database information for external stations is maintained in a different
format.
o Park-and-Ride Nodes: Park-and-Ride nodes are located at locations where trip-makers
can transition from a personal vehicle to a transit vehicle. Park and ride nodes cannot
be centroid or external station nodes and must be matched with a transit stop.
o Other Nodes: All other nodes are used solely to define link start and end points. These
nodes may be located at intersections, where centroid connectors connect to roadway
links, or mid-block.
Routes: The TransCAD Route System represent transit routes. Routes start and end at a node
and must traverse roadway links in between. Routes utilize information stored on the roadway
links, such as freeflow and congested travel time.
User’s Guide 22
Transit Stops: The transit stop layer is attached to the transit route system. Transit stops provide
a connection between the link and node layers, which is used to represent access to transit (via
the walk or drive modes). All stops must be placed on a transit route and must also be placed at
a network node.
o Note: TransCAD does not automatically enforce consistency between the node and stop
layer. It is important to follow the route system editing guidelines provided later in this
document.
Example Transportation Network Application

The example that follows includes a complete description of the process that would be used to add a
roadway segment to the network, modify a transit route to traverse the new link, and specify non-
motorized facilities on the new facility. The example adds a new connection between Wyoming St. and
Cregg Lane through the Millsite. The connection is added as a two-lane collector street with complete
bicycle and pedestrian facilities (i.e., sidewalks and bike lanes on both sides). In addition, Wyoming St.
and Cregg Lane are converted from local streets to collector streets, resulting in a complete collector
connection between Russell St. and Orange St.
Note: This example is presented for illustrative purposes only and is not necessarily consistent with
Missoula MPO planning documents.
Part 1: Prepare Workspace
Begin this section by starting the TransCAD software program. Once TransCAD is open, perform the
following actions:
1. Start the Missoula MPO Model Dialog Box from Tools  Add-Ins  Missoula Model.
2. Make sure that a valid scenario is selected in the Scenario Toolbar.
a. If the model was installed directly from the installation media, the included base year
scenario should work correctly.
b. If the model data was not installed to the default location, it may be necessary to re-
specify the scenario input and output directory.
3. From the Utilities section of the dialog box (bottom section), select the Roadway Network tab
and click the Edit Network button.
a. This action will create a new map showing the roadway network and route system.
b. The Missoula Model dialog box will be closed.
4. Verify that the intended network file has been loaded into the map.
a. Click the map layers button ( ) in the TransCAD toolbar.
b. In the window that appears, verify that the Missoula Links entry is selected.
c. Verify that the Geographic File named at the bottom of the window is indeed the
geographic file to be edited.
i. Note: If the wrong file is listed, edit the selected scenario to reference the
correct input network file.
d. Click Close.
User’s Guide 23
5. Use the zoom tool ( ) to zoom to the vicinity of the Millsite.

6. Verify that the Map Editing toolbar is present in the
workspace. If it is not, it can be enabled from the
TransCAD toolbar ( ) or menu (Tools  Map Editing
 Toolbox).
Part 2: Add New Roadway Link
Once the workspace has been set up, use the following steps to add the new roadway segment:
1. Review the area where the new roadway will be added. Observe that:
a. The new roadway will start and end at existing nodes, and
b. The new roadway will cross an existing non-motorized facility.
2. Because the new roadway will cross and connect to an existing facility (rather than be
completely grade separated), it will be necessary to:
a. Split the existing facility into separate parts to create a node at the new intersection;
and
b. create the new roadway segment in two steps, creating two links.
Note: The procedure described below is similar to the procedure that would be used to create an
intersection between a new facility and an existing roadway facility.
3. Add the first segment of the new link:
a. Select the Add Line tool ( ) from the Map Editing Toolbox.
b. Click once on the node on the west side of the new segment, then
c. Double-click on the non-motorized link where the new link will intersect.
d. Click the green light ( ) on the Map Editing Toolbox to save the new link to the line
layer file and create an undo point.
1 – Single-click on 2 – Double-click on the non-motorized

the existing node link to split the link and create a new
node.
Note: This action will split the non-motorized link into two separate links and will retain
important attributes on both halves. It is a good idea to verify that all important attributes have
been retained on both halves of the split link until the user has become comfortable with the
network editing process.
User’s Guide 24
4. Add the second segment of the new link:

a. Select the Add Line tool ( ) from the Map Editing Toolbox.
b. Single-click on the newly created node,
c. Single-click as needed to create shape points along the new link, and then
d. Double-click on an existing node to finish the new link.
e. Click the green light ( ) on the Map Editing Toolbox to save the new link to the line
layer file and create an undo point.
2 – Single-click to
create shape nodes
(Optional)
3 – Double-click to
finish the line at an
existing node
1 – Single-click on this
node from step 3
5. Refresh the route system:

a. Set the active map layer to Route System
b. If prompted to update the route system, click Yes.
c. Set the active layer back to the Missoula Network layer
Note: This step corrects any inconsistencies between the roadway line layer and route system
layers that may have been introduced by modifying the roadway line layer.
User’s Guide 25
Part 3: Edit the Roadway Segment
The new roadway links do not initially contain any data. Attributes must be added to the new links using
the steps below.
1. Identify a nearby link (a “source link”) that has the same facility type and area type as the new
link.
2. Click the Edit Line Attributes tool ( ) from the Map Editing Toolbox.
3. Copy attributes from the source link to the new link:
a. Click on the source link,
b. hold shift, then click on each new link,
1 – Click on the
source link
2 – Shift-click on each
target link
c. In the window that appears right-click in the first column and select Copy Values.
Note: You may need to resize the window to see the data for the source link. The column
with source link data should be the only column that contains values.
User’s Guide 26
d. Replace the street name (Name) for the new links by editing the values for Name in the
second two columns. Take care not to modify the name of the source link.
e. Click the green light ( ) on the Map Editing Toolbox to save the updated attributes
and create an undo point.
4. Modify attributes so this link will only be active in specific scenarios:

a. Use the standard information tool ( ) to edit attributes for both links. Click on one
link, then shift-click on the second.
b. Copy 2010 (i.e., *_10) attribute values to the alternative link value attributes (use Ctrl-C
and Ctrl-V, or right click to select copy and Paste). Do not copy the fields including the
letters FB, as they are not used for alternatives.
User’s Guide 27
Copy
c. Select an alternative number (e.g., 7) and enter the value in the ALT field
d. Clear values from the base year attributes and from other network year attributes as
desired. In this example, clear all attributes for all sets of network attributes that do not
end in _AL.
e. Check all remaining attributes, paying particular attention to count data. In this
example, it is recommended that the fields named EST_CNT, BASEVOL, and DO_NCHRP
are all cleared since this link does not exist in the base year.
5. Carefully review the alternative attributes for the new links, making changes as needed:
a. The source link has good pedestrian and only moderate bicycle facilities. Change the
value for BIKE_AL to 1 to indicate good bicycle facilities on the new segment.
b. The source link has a speed limit of 25 mph. Change the value to 35 mph on the new
links.
Par 4: Modify Connecting Links
To build a connected network, an alternative for consideration might upgrade existing segments of
Wyoming St. and Cregg Ln. to collector facilities. The steps below describe the process.
1. Modify the Wyoming St. links:
User’s Guide 28
a. Use the standard information tool ( ) to edit attributes for links on Wyoming St. and
Cregg Ln. Click on one link, then shift-click on remaining links. Review the link display
style to ensure that one or more links was not missed.
b. Select an alternative number (e.g., 8) and set the value for ALT on all selected links to
this value. A good approach is to right-click on the row header for the ALT field and
then select Fill…
c. Fill in _AL attributes using the new links as a guide. Suggested settings are as follows:
i. DIR_AL = 0
ii. FT_AL = 4 (Collector)
iii. AT_AL = 2 (Urban)
iv. AB_LN_AL, BA_LN_AL = 1
v. CTLMED_AL = 0 or blank
vi. SPLM_AL = 35
vii. TIMEPEN_AL = blank (Use of this field is not recommended in most cases)
viii. PED_AL = 1
ix. BIKE_AL = 1
Part 6: Create a New Model Scenario
Once roadway edits are complete, the user can create a new model scenario that includes the new
alternatives. The steps required to do this are included below.
1. Create a new scenario that includes the new roadway alternatives:

a. Close all files and windows in TransCAD (e.g., File  Close All).
b. Start the Missoula Model dialog box from Tools  Missoula Model.
c. In the Scenario Toolbox, make a copy of the base scenario.
2. Edit basic scenario information.
a. Scenario Name: Base 2010 (Alt 7)
b. Input Dir: No change (e.g., C:\Missoula Model\Input\)
c. Output Dir: Reference a new folder (e.g., C:\Missoula Model\Output\Base 2010).
3. Input Tab – No changes at this time.
User’s Guide 29
4. General Tab:
a. Click Alts under Scenario Settings.
b. Move alternatives 7 and 8 to the right side of the window.
5. Output Tab – No changes at this time.
6. Click OK when done to exit the Scenario Editor.
Part 5: Realign a Transit Route
To demonstrate the process of modifying the route system, this example includes realignment of an
existing transit route to use the newly added roadway. Detailed route system editing guidelines are also
provided later in this guide.
Warning: Alternative changes to the transit route system are made in separate copies of the route
system. When working with multiple route systems, the user should take care that all route systems are
updated after making changes to the roadway line layer.
1. Make a copy of the route system:

a. From the model dialog box, select the Transit Utilities tab in the Utilities section.
b. Click the Edit Routes button to close the dialog box and open the route system in a map.
c. Select Tools  Geographic Utilities  Geographic File from the TransCAD menu.
d. Click the Copy button.
e. Save the copy of the route system using a new name (e.g., MissoulaRoutes2009Alt7).
2. Edit a scenario to reference the route system copy:
b. Start the Missoula Model dialog box from Tools  Missoula Model.
c. In the Scenario Toolbox, double-click on the desired scenario (e.g., Base 2010 (Alt 7)
d. Double-click the Routes item in the input file list.
e. Select the new alternative route system.
f. Click OK in the Scenario Editor to return to the model dialog box
3. Open a map to edit the route system:
a. From the model dialog box, select the Transit Utilities tab in the Utilities section.
b. Make sure that the new scenario referencing the route system copy is selected.
c. Click the Edit Routes button to close the dialog box and open the route system in a map.
d. Verify that the Route System Toolbox has appeared on the screen.
e. Zoom the map to the vicinity of the new links
added earlier in this example.
4. Verify that the intended route system file has been
loaded into the map.
a. Click the map layers button ( ) in the
TransCAD toolbar.
b. In the window that appears, verify that the
Route System entry is selected.
c. Verify that the route system file named at the bottom of the window is indeed the route
system to be edited.
User’s Guide 30
i. Note: If the wrong file is listed, check the selected scenario and make sure it
references the correct input route system.
5. Realign Route 9 (Eastbound) to use the new links:
a. Use the Select tool ( ) on the Route System toolbox.
b. Click anywhere along Route 9
c. When prompted, select Route 9.
d. Use the Realign Routes ( ) tool on the Route System
Toolbox.
e. Select a new eastbound route alignment as shown in the
example below.
1 – Click where the route

change will begin
2 – Click on several points

along the new alignment; 1 – Double-click at the
make sure to click on nodes. location where the route will
re-join the original alignment
f. Make sure that the realignment route is expected. The new route will be shown as a
black line, with stops placed along the way. The stops will not be placed correctly yet.
g. Click the green light ( ) on the Route System Toolbox to save the updated alignment
and create an undo point.
6. Update stop placement on the realigned portion of the route

a. Use the Select tool ( ) on the Route System toolbox.
b. Click anywhere along Route 9
c. If prompted, select Route 9.
d. Move route stops ( ) placed along the realigned segment so that they are located
directly on network nodes.
e. Delete excess stops ( ) if too many stops are present, or
f. Add new stops ( ) if not enough stops are present.
User’s Guide 31
g. Click the green light ( ) on the Route System Toolbox to save the updated stop
placement and create an undo point
Reminder: Refer to the section of this document named Editing the Route System when choosing
where to place stops.
7. If desired, realign Route 9 in the westbound direction as well. In this example, the westbound
direction remains unchanged.
8. Verify the Route System

b. Start the Missoula Model Dialog Box
c. Click the Link/Verify Route System button under the Transit Utilities tab.
i. Select the modified route system
ii. Select the roadway geographic file associated with the route system
iii. The utility will ask about checking for disabled links; select No.
Note: The route system uses links that are disabled unless a specific alternative
is selected, so the disabled links will fail. It is OK to allow this check and observe
the results.
d. If the verification step fails or produces a warning message, carefully review the route
system and read all warning messages that the utility produces.
9. Verify the Route System – Stage 2 (Optional)

The route verify tool checks for many possible route system problems, but some errors cannot
be identified until the model is run. After making route system edits, the following additional
steps are recommended:
a. Start the Missoula Model dialog box
b. Select a scenario that references the modified route system
c. Un-check the Stop after Each Step option
d. Run the first model step (1 – Prepare Networks)
e. If the model run fails, there may be an error in the route system.
i. Review the resulting Log.xml file – it may provide useful troubleshooting
information.
ii. If the model run succeeds, the route system is probably valid. However, the user
is advised to review the Log.xml file and Report.xml file to check for meaningful
warnings.
Part 6: Run the Model
Once the above steps are complete, the model should be ready to run. Attempt to run the newly
created model scenario in full. If the model produces an error message, verify that all input data was
created, modified, and referenced properly.
User’s Guide 32
Working with Transit Data

The Missoula MPO Model includes the first mode choice model in the region. This section provides some
guidance on working with the transit Route System as an input file, as well as guidance on and viewing
transit assignment results.
Editing the Route System

The TransCAD route system is based on the roadway network geographic file, or the street line layer.
Because the route system relies on the roadway network, it is important to carefully maintain the
linkage between these two files. This can be accomplished by always ensuring that the route system is
correctly linked to the roadway network and that the route system is present in the map when adding,
deleting, splitting, or joining links in the roadway network.
Route System Link and Verify
Prior to making any changes to the route system or roadway network, it is good practice to link the
route system to the roadway network and verify the route system’s integrity. This can be done by using
the Link/Verify Route System utility available from the main model dialog box. When activated,
TransCAD will ask the user to identify a route system, then the corresponding roadway network file.
TransCAD will link the selected files and then check for errors in the route system.
NOTE: A common mistake is to accidentally select the route system stop layer (a “dbd” file) instead of
the roadway network layer. If this is done, the Link/Verify utility will not make any changes to the
route system, but will show an error message instead.
Maintaining Route System Integrity
Because the route system is directly linked to the roadway network, TransCAD must modify the route
system any time a change is made to the roadway network layer. This is particularly important if a
roadway link that is traversed by a transit route is split, moved, or joined to another link. When this
happens, TransCAD will update the route system data to account for the change. The easiest and least
error-prone method of updating the route system is to add it to the map prior to modifying the roadway
line layer. After modifying a link that is traversed by a route, the user must set the route system as the
active layer to cause TransCAD to update the file.
If a user edits the roadway layer while the route system is not present in the map, TransCAD will update
the route system next time it is opened. This update takes place even if none of the links that have been
modified are traversed by a transit route. To ensure that a route system is up to date, the following
procedure should be used regardless of whether the route system was present in a map during network
edits:
User’s Guide 33
1. Close all files in the TransCAD program.

2. Run the Link/Verify Route System utility.
a. If this utility fails, continue to Step 3 anyway.
3. Open the route system in TransCAD.
a. Verify that the correct line layer was opened automatically to support the route system.
4. Close all files.
5. If the Link/Verify Route System utility failed in Step 2, run the utility again. If the route system
verification still fails, the route system must be repaired manually. (This is uncommon and will
typically only occur if the guidelines in this section are not followed).
It is only necessary to follow the guidelines specified above when splitting, joining, adding, and deleting
links from the roadway network. The steps above are not necessary when modifying attributes of
existing links.
Editing the Route System
The route system can be edited with the Route System Editing Toolbox. However, this toolbox requires a
network (*.net) file as well as a geographic (*.dbd) file. Prior to opening the Route System Editing
Toolbox, create a working network as follows:
1. Open the route system and verify that the correct line layer has been opened with it.
2. Set the line (roadway) layer as the active layer.
3. Create a new network from Networks/Paths  Create.
4. Save the network using a generic name (e.g., net.net).
5. Use the default network settings.
After the network file has been created, the route system can be edited. The Route System Editing
Toolbox is activated by setting the route system as the active layer and then selecting Route Systems 
Editing Toolbox from the TransCAD menu.
The TransCAD documentation provides detailed guidance on route system editing tools. However, some
additional guidelines are provided here to ensure that route systems will work properly.
Route stop problems are usually reported in the Log file that is generated during the first model step. If
the route system has been modified in any way, this file should be inspected for warnings and errors.
However, the log file will not indicate all
WARNING: Route stop problems do not always cause the model to crash. Problems
are often reported in the Log file that is generated during the first model step. If the
route system has been modified in any way, this file should be inspected for
warnings and errors.
User’s Guide 34
ROUTE DIRECTION
All routes are coded directionally from start to finish. If a route traverses a street more than once (e.g.,
out and back along the same corridor), it must be coded to travel out and then back. A simple way to
think of this process is to code the route exactly as it is driven.
Each numbered or named transit route should be coded as a single route, usually starting and ending at
the same place. If a layover time is experienced at a particular stop, the route should start and end at
this point.
STOP PLACEMENT EXAMPLE: Incorrectly placed stop at a two-

way/loop transition. Two stops sharing a node
1. With the exceptions noted below, only one must have different values for Pass_Count.
stop for each route should be placed at each
node. This includes locations where the
route transitions from a loop portion to a
two-way portion. In these cases, a two-pass
stop should be located on the two-way
portion of the route. Failure to observe these
requirements may cause stops to work
incorrectly and will cause “Incorrect Tag
Info” warnings in the log file.
Exceptions
 Beginning/End of Route – see Item 2
below
 Multi-Pass Routes – see item 5 below
2. A route stop should be placed at the beginning and end of each route. The model algorithms will
not verify that this condition is met, but failure to properly place route-end stops will result in
incorrect transit paths, but no errors or warnings will be generated. This is especially important
on circular routes!
3. Route stops should be placed on routes at all locations where walk access to transit might be
available, but all stops must be placed at street layer nodes. Route stops are not needed at
nodes where walk access cannot be achieved, or where other nearby stops on the same route
can be reached more quickly from all nearby zones.
4. TransCAD does not “snap” route stops to line layer nodes, but all route stops must be placed
close to a node on the roadway line. If any stops are placed too far from the nearest node, the
travel model will produce a warning and the stop will be ignored.
5. If a route doubles back (e.g. a two-way route), stops should be placed on the route at both first
and second passes. Two-pass stops on a single node are usually displayed as a stop icon with a
“2” in TransCAD.
User’s Guide 35
6. At a point where a route reverses, but does not start or end, only one stop should be placed. If
two stops are placed at a mid-route reversal, the model will not work correctly and may fail to
run.
Viewing Transit Assignment Results

Transit assignment results are stored in several files with varying levels of detail. Transit assignment
results are itemized and described in Table 6. Transit results can be viewed by joining layers to the
route, stop, and node layer as described below, or by using the mapping utility to create a map that
displays transit results.
Table 6: Transit Assignment Results
Filename Contents Description

This table contains segment-level
route ridership data. It can be
tasn_DAY_tflow.bin Route flow data
viewed by attaching it to the route
system as described below.
This table contains walk access and
tasn_DAY_wflow.bin Walk flow data egress flows. It can be joined to the
line layer by link ID.
This table contains segment-level
aggregated ridership data. If
multiple routes traverse the same
segment, this table contains the
tasn_DAY_aflow.bin Aggregated route flow data
sum of ridership on these routes.
This table can be viewed by
attaching it to the route system as
described below.
This table contains boarding and
alighting data by stop. It can be
tasn_DAY_OnOff.bin Boarding/alighting data
joined to the stop layer based on
Stop IDs.
*In addition to the files listed, separate files are available for peak walk (PKW), peak drive (PKD), off-peak walk
(OPW), and off-peak drive (OPD). For these files, the text “DAY” is replaced by the identifiers listed above.
User’s Guide 36
Attaching Route Data
Route data is stored using a linear referencing system that is not

easily interpreted. However, TransCAD can use this data to create
a layer that can be viewed in a map. To do this, follow the steps
below.
1. Open the output route system in a map.

2. Open the tflow file in TransCAD.
3. Activate the map window and ensure that the route
system is the active layer.
4. Select Route Systems  Linear Referencing  Attach from
the TransCAD menu.
5. Fill in the dialog box as shown below. Note: The start and
end fields are not filled in correctly by default.
6. Click OK.
The result will be an additional layer in the TransCAD map window.

Transit flow on each route segment can be viewed with the Info
tool. If multiple routes share a segment, multiple co-linear
segments can be selected with the Info tool by clicking on the segment and then dragging outward to
create a circle.
Creating the Daily Aggregated Flow Table
Transit assignment results are provided by route, but it is sometimes useful to view total ridership on all
routes by roadway segment. This can be done by using the aggregated flow tables. The transit post-
processing step creates this table, but it can also be created manually. To view daily aggregated transit
flows as a map layer using the default TransCAD interface, follow the steps below.
1. Open the output route system in a map.

2. Open the daily tflow file in TransCAD.
3. Activate the map window and ensure that the route
system is the active layer.
4. Select Transit  Aggregate Transit Assignment from
the TransCAD menu.
5. Select the tflow layer.
6. Save the resulting file.
The result will be an additional layer in the TransCAD map window. The layer will show total transit flow
on each segment when queried with the Info tool. It will not be possible to select multiple co-linear
segments.
To view this data later, or to load data that was automatically generated by the transit post-processing
step, the aflow file that was saved in Step 6 can be attached to the route system layer using Route
Systems  Linear Referencing  Attach.
User’s Guide 37
Summary Report User’s Guide
February 2012
Table of Contents
Context and Background............................................................................................................................... 1
Title Page....................................................................................................................................................... 2
Files and Settings .......................................................................................................................................... 2
Input Network Summary............................................................................................................................... 3
Socioeconomic Data Summary ..................................................................................................................... 4
Trip Generation Summary............................................................................................................................. 5
Trip distribution Summary ............................................................................................................................ 6
Mode Choice Summary ................................................................................................................................. 7
Assigned Vehicle Trip Summary .................................................................................................................... 8
Transit Assignment Summary ....................................................................................................................... 9
Traffic Assignment Summary ...................................................................................................................... 10
Vehicle Validation Summary ....................................................................................................................... 11
Summary Report User’s Guide - DRAFT i


This Summary Report User’s Guide provides a brief overview of the contents of the model summary
report. The screenshot below shows the performance report options dialog with all reports enabled.
Under each summary item, a partial screenshot of the summary report contents is shown for reference.
However, these screenshots may not reflect the exact numbers in the final version of the summary
report.
The “Custom 1” and “Custom 2” settings require the fields “CUSTOM 1” and “CUSTOM 2” to be present
and populated both on the roadway network and in the database. On the roadway network, the fields
must contain a value of “1” on links to be included in the subarea. In the database, the “aZoneData”
table must contain a value of “1” in zones to be included in the summary area. This table can be
accessed from the “Other TAZ Data” button in the interface.
If either of these fields are missing or if they do not include a “1” in at least one zone and TAZ, the
performance report will fail to run with a custom summary area selected.
Summary Report User’s Guide – DRAFT 1

Title Page
The title page contains a simple description of the scenario, including the information shown below.
Summary Report for the Missoula MPO Travel Model

Developed by LSA Associates for The Missoula MPO(Missoula)
Scenario Name: Base2009

Scenario Directory: D:\Models\Missoula Model\Output\Base2009\
Report File: D:\Models\Missoula Model\Output\Base2009\Summary.xlsx
Report Created on: Sunday, February 05, 2012 (2:35 PM)
Files and Settings

The Files and Settings report echoes out all file parameter information for the scenario. Each file
selected in the scenario manager is listed by keyword. Each file is defined by the path and filename as
well as a short description describing the file contents.
Input Files
Input
Value/Description
Network D:\Models\Missoula Model\Input\MissoulaNetwork.dbd
Roadway Geographic File
TurnPen D:\Models\Missoula Model\Input\TPEN.bin
Turn Penalty File
Database …
D:\Models\Missoula Model\Input\MissoulaDatabase.mdb
Output Access database containing model input parameters and data
Prepare Networks:
Value/Description
RdNetwork D:\Models\Missoula Model\Output\Base2009\RoadwayNetwork.dbd
Output Roadway Network
Net D:\Models\Missoula Model\Output\Base2009\ini_Network.net
Routable Roadway Network
Walknet D:\Models\Missoula Model\Output\Base2009\Walk_Network.net
Routable Pedestrian Network

Input Network Summary

The Input Network Summary specifies centerline miles and lane miles from the input network. This
report can be used as a quality control measure to verify that network changes were implemented
correctly. For example, a model run with a 2-mile section of road widened from two to four lanes would
have the same number of centerline miles as the base case. The number of lane miles would be
expected to increase by 4.
Entire Model
Network Centerline Summary
CBD Urban Suburban Rural Total

Freeway 0 1 10 51 63
Principal Arterial 1 17 20 155 192
Minor Arterial 0 16 20 46 82
Collector 2 19 53 220 295
Rural Highway 0 0 13 263 276
Local Street 2 154 326 5,902 6,385
Ramps 0 3 9 9 20
Transit Only 0 11 3 0 14
Centroid Connectors 2 45 91 279 417
Total 8 267 545 6,925 7,745
Entire Model
Network Lane-Mile Summary

Freeway 0 6 41 204 251
Principal Arterial 4 60 57 369 491
Minor Arterial 1 38 44 92 175
Collector 5 38 107 441 592
Local Street 5 309 653 11,803 12,769
Ramps 0 3 10 9 21
Total 18 544 1,119 14,003 15,684

Socioeconomic Data Summary

The first set of tables in this report summarizes bivariate distributions of household by income and size.
These distributions result from the household disaggregation model. Values here are not expected to
change and are most useful in performing mode validation.
The last table in this report summarizes socioeconomic data by subregion and by data type. This table is
most useful in verifying that socioeconomic data totals in the model are as expected. For example, when
running different forecast year land use scenarios, the employment and population totals in this table
should match values from the input dataset.
Entire Model
Household Bivariate Distribution
1 Person 2 Person 3 Person 4 Person 5+ Person Total

Low Income 10,312 4,733 615 311 698 16,670
Medium Income 6,979 14,437 5,282 3,537 933 31,168
High Income 538 5,643 3,608 2,316 2,880 14,984
Total 17,829 24,813 9,506 6,164 4,511 62,822
Entire Model
Socioeconomic Data Summary
RET_EMP SVC_EMP TOT_HH BAS_EMP EDUC_EMP HC_EMP LEIS_EMP Total

Entire Model 12,024 27,830 62,822 21,872 6,903 12,531 10,065 154,047
Missoula MPO 8,839 20,684 39,847 13,536 5,204 10,274 7,670 106,054
Missoula County 9,280 21,784 45,889 16,267 5,765 10,481 8,418 117,884
Ravalli County 2,744 6,046 16,933 5,605 1,138 2,050 1,647 36,163

Trip Generation Summary

The first few tables in the trip generation summary reports trip productions and attractions by purpose
for the peak and off-peak periods as well as for the day as a whole. Both balanced and unbalanced
numbers are included. These tables are useful for monitoring changes in the balance of jobs and
households in forecast year scenarios. If the unbalanced production and attraction totals by trip purpose
differ significantly, the socioeconomic dataset should be reviewed to ensure the forecast balance of
housing and employment is reasonable. Note that the unbalanced HBU trip productions and attractions
are not expected to balance due to implementation of a production allocation model.
The second set of tables provides values that are commonly used as a “reasonableness check,” such as
trip productions and attractions per employee and per household. These tables are most useful in model
validation.
Person Trip Productions (Unbalanced)
LIHBW MIHBW HIHBW HBS HBU HBO WBO OBO Total

Entire Model 9,824 65,244 38,791 68,390 1,278,340 199,503 49,133 140,283 1,849,507
Missoula MPO 6,743 40,288 24,405 45,764 1,213,427 139,928 38,460 116,135 1,625,148
Missoula County 7,082 44,711 29,411 51,543 1,235,860 158,793 41,095 121,652 1,690,148
Ravalli County 2,742 20,533 9,380 16,846 42,480 40,709 8,039 18,631 159,360
Person Trip Attractions (Unbalanced)

Entire Model 8,159 54,519 32,707 152,415 42,526 183,769 55,992 140,283 670,369
Missoula MPO 5,598 37,536 22,634 121,244 42,526 140,052 46,661 116,135 532,386
Missoula County 6,132 41,074 24,732 126,255 42,526 150,351 48,902 121,652 561,623
Ravalli County 2,027 13,445 7,975 26,160 0 33,417 7,090 18,631 108,746
Person Trip Productions (Balanced)

Entire Model 9,824 65,244 38,791 68,390 42,526 199,503 49,000 135,171 608,449
Missoula MPO 6,743 40,288 24,405 45,764 40,367 139,928 38,355 111,880 447,729
Missoula County 7,082 44,711 29,411 51,543 41,113 158,793 40,982 117,191 490,827
Ravalli County 2,742 20,533 9,380 16,846 1,413 40,709 8,018 17,981 117,622
Person Trip Attractions (Balanced)

Entire Model 9,824 65,244 38,791 68,390 42,526 199,503 49,000 135,171 608,449
Missoula MPO 6,742 44,901 26,812 54,197 42,526 152,136 40,820 111,880 480,013
Missoula County 7,388 49,166 29,322 56,437 42,526 163,323 42,766 117,191 508,120
Ravalli County 2,436 16,078 9,468 11,953 0 36,179 6,234 17,981 100,329
Person Trip Productions per Household (Unbalanced)

Entire Model 0.16 1.04 0.62 1.09 20.35 3.18 0.78 2.23 29.44
Missoula MPO 0.17 1.01 0.61 1.15 30.45 3.51 0.97 2.91 40.78
Missoula County 0.15 0.97 0.64 1.12 26.93 3.46 0.90 2.65 36.83
Ravalli County 0.16 1.21 0.55 0.99 2.51 2.40 0.47 1.10 9.41
Person Trip Productions per Household (Balanced)

Entire Model 0.16 1.04 0.62 1.09 0.68 3.18 0.78 2.15 9.69
Missoula MPO 0.17 1.01 0.61 1.15 1.01 3.51 0.96 2.81 11.24
Missoula County 0.15 0.97 0.64 1.12 0.90 3.46 0.89 2.55 10.70
Ravalli County 0.16 1.21 0.55 0.99 0.08 2.40 0.47 1.06 6.95

Trip Distribution Summary

This report summarizes trip distribution characteristics such as average trip length, distance, and speed.
The summary also reports the number of intrazonal trips (i.e., trips that start and end in the same TAZ)
and provides a set of subregion-to-subregion trip tables. While mostly useful for model validation, this
table may be useful in evaluating the impact of different land use strategies on average trip lengths.
Average Trip Lengths for Off-Peak
HBW HBS HBU HBO WBO OBO

Miles 12.43 9.00 4.39 7.68 6.46 4.02
Minutes 18.96 14.57 9.69 13.73 11.43 8.56
Min. (no term) 16.35 11.88 7.70 10.80 8.64 5.70
Speed (w/ term) 39.31 37.06 27.16 33.57 33.88 28.17
Speed (no term) 45.59 45.44 34.19 42.68 44.83 42.36
Modeled Person Trip Totals for Off-Peak

Intrazonal 3,176 3,114 0 10,118 2,184 9,908
Interzonal 70,947 60,010 33,936 150,482 35,595 113,098
Total 74,122 63,124 33,936 160,600 37,779 123,006
% Intrazonal 4.28% 4.93% 0.00% 6.30% 5.78% 8.06%
Off-Peak SubRegion/SubRegion Trips (HBW)
MPO MissoulaCounty
RavalliCounty Total
MPO 42,210 1,748 2,546 46,504
MissoulaCounty 3,622 2,282 456 6,360
RavalliCounty 5,239 801 15,218 21,258
Total 51,070 4,832 18,220 74,122

Mode Choice Summary

The mode choice summary includes a tabulation of person and vehicle trips by mode. This table can be
useful in monitoring changes in mode share over different scenarios. Note that this report only applies
to trips made within TAZs identified as MPO TAZs (i.e., SUBREGION = 1), as Mode Choice is not run for
zones outside of this area.
Mode Choice Summary
HBW HBS HBU HBO WBO OBO Total

Auto DA Peak 17,705 1,640 3,198 8,983 6,622 3,751 41,899
Auto SR Peak 2,650 1,696 1,214 14,206 1,203 5,355 26,324
Auto DA Offpeak 33,093 19,425 12,714 37,110 22,413 37,955 162,711
Auto SR Offpeak 4,953 20,084 4,828 58,682 4,073 54,187 146,807
Total Auto Person 58,401 42,845 21,954 118,981 34,311 101,248 377,741
Transit OPW 346 135 1,404 284 104 207 2,481
Transit OPD 0 0 1,117 0 0 0 1,117
Transit PKW 232 15 407 87 39 27 808
Transit PKD 0 0 275 0 0 0 275
Total Transit 578 150 3,203 372 143 234 4,681
Direct Walk 1,381 1,541 5,974 14,582 2,213 7,384 33,075
Direct Bike 4,479 891 9,235 3,747 1,027 1,520 20,900
Total Person Trips 64,839 45,426 40,367 137,683 37,695 110,387 436,396

Assigned Vehicle Trip Summary

This table summarizes assigned vehicle trips in OD format by time of day and for the day as a whole. It is
primarily used as a validation check.
Entire Model
Assigned Vehicle Trips - AM Peak Period
HBW HBS HBU HBO WBO OBO EE Total

Intrazonal Trips 843 50 0 1,282 264 477 0 2,916
Interzonal Origins 17,373 851 3,043 12,984 4,746 3,893 312 43,202
Total Origins 18,216 901 3,043 14,266 5,010 4,371 312 46,118
Interzonal Destinations 17,373 851 3,043 12,984 4,746 3,893 312 43,202
Total Dest. 18,216 901 3,043 14,266 5,010 4,371 312 46,118
Entire Model
Assigned Vehicle Trips - PM Peak Period
HBW HBS HBU HBO WBO OBO EE Total

Intrazonal Trips 800 154 0 1,099 214 381 0 2,648
Interzonal Origins 16,492 2,641 1,053 11,129 3,843 3,109 312 38,579
Total Origins 17,293 2,794 1,053 12,228 4,057 3,490 312 41,227
Interzonal Destinations 16,492 2,641 1,053 11,129 3,843 3,109 312 38,579
Total Dest. 17,293 2,794 1,053 12,228 4,057 3,490 312 41,227

Transit Assignment Summary

This table lists the number of boardings on each trout in the transit route system. The average number
of boardings per one-way trip is also included as the last item in the table.
This report is useful in comparing transit ridership between different scenarios. It is especially useful
when testing impacts of route system changes with all other model inputs held constant.
Transit Boardings
Boardings
COT PnR 29
East Broadway PnR 544
Route 1 867
Route 10 84
Route 11 215
Route 12 464
Route 2 282
Route 3 177
Route 4 133
Route 5 117
Route 6 288
Route 7 264
Route 8 482
Route 9 221
South Campus PnR 1,535
Total 5,702
Total Transit Trips 4,681
Avg. Boardings/Trip 1.22

Traffic Assignment Summary

The traffic assignment summary reports commonly used statistics resulting from traffic assignment.
These numbers can be used in a performance measure process, either directly or in the development of
measures based on model results. Each statistic is reported by facility type and area type, which is most
important for use in model validation. In review and analysis of model outputs, it is typically sufficient to
consider the total (i.e., network-wide or subarea-wide) value.
Reported statistics include:
Vehicle Miles Traveled (VMT): Total vehicle miles traveled on the roadway network. This number
is often used in comparison of different model scenarios.
Freeflow Vehicle Hours Traveled: Total vehicle hours traveled on the roadway network, based on
the freeflow travel speeds. This number is useful for model validation, but is rarely used for
analysis of model scenario runs.
Congested Vehicle Hours Traveled: Total vehicle hours traveled on the roadway network, based
on the congested travel speeds resulting from traffic assignment. This number is often used in
comparison of different model scenarios.
Vehicle Hours of Congestion Delay: Congestion delay is defined as the amount of time lost due
to network congestion. It is computed as the difference between the congested and freeflow
VHT and is often used in comparison of different model scenarios.
Congested Vehicle Miles Traveled: Congested vehicle miles traveled is the total VMT taking place
on links with a level of service of E or F. While less commonly used, it can be used in comparison
of different model scenarios.
Congested Lane Miles: This reports the total number of lane miles in the network experiencing a
level of service of E or F. While less commonly used, it can be used in comparison of different
model scenarios.
Entire Model
Vehicle Miles Traveled (VMT)

Freeway 0 22,336 216,314 498,801 737,451
Principal Arterial 16,984 409,529 248,739 743,692 1,418,944
Minor Arterial 945 150,682 134,238 89,671 375,536
Collector 8,800 55,064 111,478 266,959 442,300
Rural Highway 0 0 11,656 131,832 143,488
Local Street 1,938 67,590 67,804 385,306 522,638
Ramps 0 11,140 19,755 5,005 35,900
Centroid Connectors 1,725 20,971 13,449 83,671 119,815
Total 30,391 737,313 823,431 2,204,937 3,796,072

Vehicle Validation Summary

This section presents a comparison of traffic count data and travel model volumes. It is useful in model
validation, but is not typically used in scenario comparison. This report is only meaningful for the
existing conditions model run. The model documentation includes a more detailed description of the
various model validation statistics that are included in this report, along with some additional model
validation statistics that were calculated by hand.
Entire Model
Count VMT

Freeway 0 6,999 128,991 264,636 400,626
Principal Arterial 2,681 146,316 34,421 50,039 233,456
Minor Arterial 1,050 47,554 21,387 6,091 76,082
Collector 3,391 20,644 26,318 16,765 67,118
Local Street 125 5,876 2,150 1,104 9,256
Ramps 0 0 0 0 0
Total 7,247 227,389 213,267 339,520 787,423

Model
Documentation
September 2011
Table of Contents
CHAPTER 1: ROADWAY NETWORK ....................................................................... 1
Context and Background ........................................................................................................................... 1
Roadway Network Structure ..................................................................................................................... 1
Input and Output Networks ...................................................................................................................... 2
Multi-Year and Alternative Network Structure ........................................................................................ 3
Representation of Networks by Year .................................................................................................... 4
Representation of New Facilities .......................................................................................................... 4
Representation of Network Alternatives .............................................................................................. 5
Network Attribute Selection ................................................................................................................. 7
Network Attribute List .............................................................................................................................. 9
Functional Classification / Facility Type .................................................................................................. 12
Area Type ............................................................................................................................................ 19
Link Speeds.............................................................................................................................................. 24
Estimating Link Speeds........................................................................................................................ 24
Travel Time.......................................................................................................................................... 26
Link Capacities ......................................................................................................................................... 26
Freeways ............................................................................................................................................. 27
Collectors and Arterials ....................................................................................................................... 28
Resulting Capacity Model ................................................................................................................... 30
Output Network Fields ............................................................................................................................ 31
CHAPTER 2: TRANSIT AND NON-MOTORIZED NETWORKS ................................... 32

Context and Background ......................................................................................................................... 32
Transit Networks ..................................................................................................................................... 32
Transit Route System .......................................................................................................................... 34
Transit Line Layer ................................................................................................................................ 36
Transit Pathbuilding ............................................................................................................................ 40
Non-Motorized Networks ....................................................................................................................... 41
Bicycle Network .................................................................................................................................. 41
Pedestrian Network ............................................................................................................................ 42
CHAPTER 3: TRIP GENERATION.......................................................................... 44

Data Sources ........................................................................................................................................... 44
Traffic Analysis Zones .............................................................................................................................. 45
Household Disaggregation Models ......................................................................................................... 45
Household Size Disaggregation Model ............................................................................................... 46
Household Income Disaggregation Model.......................................................................................... 47
TAZ-Level Bivariate Data ..................................................................................................................... 48
Trip Purposes .......................................................................................................................................... 48
Table of Contents - i
Production Rates ..................................................................................................................................... 51

Income Groups .................................................................................................................................... 51
Attraction Rates ...................................................................................................................................... 56
Non-Home-Based Production Allocation Models ................................................................................... 56
University Trip Generation and Production Allocation ........................................................................... 57
University Definition ........................................................................................................................... 57
University Trip Types........................................................................................................................... 58
Special Generator Survey Adaptation ................................................................................................. 59
External Trips .......................................................................................................................................... 62
External Station Volumes .................................................................................................................... 62
Internal-External and External-Internal Trips ..................................................................................... 64
External-External Trips ........................................................................................................................ 64
Sub-Region Trip Rate Factors .................................................................................................................. 66
Trip Balancing .......................................................................................................................................... 66
CHAPTER 4: TRIP DISTRIBUTION ........................................................................ 67

Peak and Off-Peak Period Definitions ..................................................................................................... 68
Roadway Network Shortest Path ............................................................................................................ 68
Terminal Times .................................................................................................................................... 69
Intrazonal Impedance ......................................................................................................................... 69
Friction Factors ....................................................................................................................................... 70
CHAPTER 5: MODE CHOICE ............................................................................... 73

Observed Mode Shares ........................................................................................................................... 73
Non-Motorized Shares ........................................................................................................................ 73
Vehicle Trip Mode Shares ................................................................................................................... 75
Transit Trip Mode Shares .................................................................................................................... 75
University Mode Shares ...................................................................................................................... 78
Resulting Mode Targets ...................................................................................................................... 79
Auto Occupancy .................................................................................................................................. 79
Model Structure ...................................................................................................................................... 80
Logit Model Background ..................................................................................................................... 80
Missoula Model Structure................................................................................................................... 84
Missoula Model Parameters ............................................................................................................... 85
Cost Variables ..................................................................................................................................... 88
Mode Choice Model Validation Results .................................................................................................. 89
Table of Contents - ii
CHAPTER 6: TRAFFIC ASSIGNMENT .................................................................... 90

Time of Day ............................................................................................................................................. 90
Trip Assignment ...................................................................................................................................... 93
Assignment Algorithms ....................................................................................................................... 93
Closure Criteria ................................................................................................................................... 94
Impedance Calculations ...................................................................................................................... 95
Volume-Delay Functions ..................................................................................................................... 96
Speed Feedback ...................................................................................................................................... 97
Methodology ....................................................................................................................................... 98
Convergence Criteria .......................................................................................................................... 99
Application of Speed Feedback for Alternatives Analysis ................................................................. 101
Traffic Assignment Validation ............................................................................................................... 102
Overall Activity Level ......................................................................................................................... 102
Screenlines ........................................................................................................................................ 103
Measures of Error ............................................................................................................................. 106
LIST OF TABLES
Table 1.1: Input Network Link Fields ...................................................................................................... 9

Table 1.2: Input Network Node Fields .................................................................................................. 11
Table 1.3: Functional Classification / Facility Type Values.................................................................... 13
Table 1.4: Area Type Categories ........................................................................................................... 19
Table 1.5: Area Type Model Criteria ..................................................................................................... 19
Table 1.6: Ratio of Freeflow Speed (Off-Peak) to Speed Limit ............................................................. 24
Table 1.7: Speed Limit to Freeflow Speed Conversion Factors............................................................. 25
Table 1.8: Centroid Connector Freeflow Speeds .................................................................................. 25
Table 1.9: Ideal and Adjusted Capacities for Freeways and Expressways based on HCM 2000 .......... 28
Table 1.10: Link Capacity Adjustment Factors and Resulting Hourly Capacity per Lane ........................ 30
Table 1.11: Roadway Capacities (vehicles per hour per lane, upper-limit LOS E) .................................. 31
Table 2.1: Transit Networks .................................................................................................................. 34
Table 2.2: Route Attributes ................................................................................................................... 34
Table 2.3: Route Headway Assumptions .............................................................................................. 35
Table 2.4: Route Stop Attributes .......................................................................................................... 36
Table 2.5: Key fields in the Transit Line Layer ....................................................................................... 38
Table 2.6: Transit Pathbuilding Weights ............................................................................................... 40
Table 2.7: Bicycle Scoring Values .......................................................................................................... 42
Table 2.8: Output Roadway Network Fields ......................................................................................... 43
Table 3.1: Income Group Definitions .................................................................................................... 47
Table 3.2: Bivariate Household Distribution for Missoula Ravalli, and Mineral Counties .................... 48
Table 3.3: Trip Purposes........................................................................................................................ 48
Table 3.4: Weighted and Expanded Trips by Purpose .......................................................................... 49
Table 3.5: Trip Purpose Definitions based on Reported Activity .......................................................... 50
Table 3.6: Household Trip Production Rates by Income Category ....................................................... 51
Table of Contents - iii

Table 3.7A: Initial HBW Trip Production Rates........................................................................................ 52

Table 3.8A: Initial HBS Trip Production Rates ......................................................................................... 52
Table 3.9A: Initial HBO Trip Production Rates ........................................................................................ 52
Table 3.10A: Initial WBO Trip Production Rates ....................................................................................... 53
Table 3.11A: Initial OBO Trip Production Rates ........................................................................................ 53
Table 3.12A: Initial Trip Production Rates – All Purposes ......................................................................... 53
Table 3.13A: Initial Trip Attraction Rates .................................................................................................. 54
Table 3.7B: Final HBW Trip Production Rates ......................................................................................... 54
Table 3.8B: Final HBS Trip Production Rates ........................................................................................... 54
Table 3.9B: Final HBO Trip Production Rates .......................................................................................... 55
Table 3.10B: Final WBO Trip Production Rates ......................................................................................... 55
Table 3.11B: Final OBO Trip Production Rates .......................................................................................... 55
Table 3.12B: Final Trip Production Rates – All Purposes .......................................................................... 55
Table 3.13B: Final Trip Attraction Rates.................................................................................................... 56
Table 3.14: WBO Production Allocation Rates ....................................................................................... 57
Table 3.15: UM Employment and Enrollment ........................................................................................ 60
Table 3.16: University Special Generator Values .................................................................................... 60
Table 3.17: External Travel Assumptions ................................................................................................ 63
Table 3.18: IE/EI Trips by Trip Purpose and Direction ............................................................................ 64
Table 3.19: EE Trip Table Seed Values .................................................................................................... 65
Table 3.20: 24-Hour EE Trip Table .......................................................................................................... 65
Table 3.21: Jurisdictional Trip Rate Factors ............................................................................................ 66
Table 4.1: Peak and Off-Peak Trip Percentages by Purpose ................................................................. 68
Table 4.2: Terminal Penalties by Area Type .......................................................................................... 69
Table 4.3: Friction Factor Gamma Parameters ..................................................................................... 72
Table 5.1: Non-Motorized Mode Share Targets ................................................................................... 74
Table 5.2: Bicycle and Pedestrian Mode Share Targets ........................................................................ 75
Table 5.3: Drive alone and Shared Ride Mode Share Targets .............................................................. 75
Table 5.4: Fixed Route Boardings ......................................................................................................... 76
Table 5.5: Transit Trip Targets .............................................................................................................. 76
Table 5.6: Drive and Walk Access ......................................................................................................... 77
Table 5.7: Transit Trips by Purpose ....................................................................................................... 78
Table 5.8: HBU Mode Shares (Colorado State University) .................................................................... 78
Table 5.9: Resulting Mode Targets ....................................................................................................... 79
Table 5.10: Average Auto Occupancy ..................................................................................................... 79
Table 5.11: New Starts Coefficient Guidelines ....................................................................................... 85
Table 5.12: Median Incomes and values of time ranges ........................................................................ 86
Table 5.13: Mode Choice Model Coefficients ......................................................................................... 86
Table 5.14: Mode Choice Model Variables ............................................................................................. 87
Table 5.15: Alternative Specific Constants ............................................................................................. 88
Table 5.16: Cost Variables ....................................................................................................................... 88
Table 5.17: Fixed Route Boarding Totals ................................................................................................ 89
Table 6.1: Peak Period Definitions ........................................................................................................ 91
Table 6.2: Time of Day Factors (Based on 24 hours) ............................................................................ 92
Table 6.3: Pre-Distribution Time of Day Factors ................................................................................... 92
Table 6.4: Pre-Assignment PA to OD Time of Day Factors.................................................................... 92
Table of Contents - iv
Table 6.5: Volume Delay Parameters Alpha and Beta .......................................................................... 97

Table 6.6: Regional Activity Validation ............................................................................................... 102
Table 6.7: VMT and VHT Totals ........................................................................................................... 103
Table 6.8: Screenline Data .................................................................................................................. 106
Table 6.9: Model % Root Mean Square Error ..................................................................................... 107
Table 6.10: % Root Mean Square Error by Volume Group ................................................................... 107
LIST OF FIGURES
Figure 1.1: Example Model Run Directory Structure ............................................................................... 3

Figure 1.2a: 2010 Facility Type Designations (Regional) .......................................................................... 14
Figure 1.2b: Facility Type Designations (MPO) ........................................................................................ 15
Figure 1.2c: Facility Type Designations (Urban Detail) ............................................................................ 16
Figure 1.3: Roadway Facility Type Hierarchy ......................................................................................... 17
Figure 1.4a: Area Type Designations (Regional) ...................................................................................... 21
Figure 1.4b: Area Type Designations (MPO) ............................................................................................ 22
Figure 1.4c: Area Type Designations (Urban Detail) ................................................................................ 23
Figure 2.1: Connections between the Route System and Transit Line Layer......................................... 32
Figure 2.2: Roadway and Transit Line Layer Processing ........................................................................ 33
Figure 2.3: Off-Peak Transit Time factors .............................................................................................. 37
Figure 2.4: Peak Transit Time Factors .................................................................................................... 37
Figure 2.5: Example Walk Access Paths ................................................................................................. 39
Figure 3.1: Household Size Disaggregation Curves ................................................................................ 46
Figure 3.2: Household Income Disaggregation Model ........................................................................... 47
Figure 3.3: UM Campus Zones ............................................................................................................... 57
Figure 3.4: University Production Allocation Model Results ................................................................. 61
Figure 3.5: External Station Locations .................................................................................................... 63
Figure 4.1: HBW Trip Time Distribution ................................................................................................. 71
Figure 4.2: Friction Factors ..................................................................................................................... 72
Figure 5.1: Example Multinomial Logit Structure .................................................................................. 81
Figure 5.2: Example Nested Logit model ............................................................................................... 83
Figure 5.3: Mode Choice Model Structure ............................................................................................. 84
Figure 6.1: Screenline Locations .......................................................................................................... 104
Figure 6.2: Screenline Error Values ...................................................................................................... 105
Figure 6.3: Model Count/Volume Comparison .................................................................................... 106
Table of Contents - v
Chapter 1: Roadway Network

The roadway network contains basic input information for use in the travel demand model and
represents real-world conditions for the 2010 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 2010 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
Missoula and, at a reduced level of detail in Missoula and Ravalli counties. 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.
The roadway network is structured to contain data for multiple timeframes. The roadway network
prepared for the Missoula MPO Model contains the 2010 base year network and can also store forecast
year improvements or alternatives. It is designed to accommodate future horizon year networks,
including 2040 and other interim years as desired. The model includes the capability to represent the
2010 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.
Roadway Network Structure

The Missoula MPO roadway network structure was designed to be a flexible data repository and to host
input and output data required by the travel model. This section describes the network file structure
and defines attributes that are populated on the network. Input attributes and some output attributes
are discussed herein. Additional output variables created by subsequent model steps are discussed in
the associated technical memoranda.
Chapter 1: Roadway Network - 1

Input network attributes used by the travel model include facility type, area type, number of lanes
(including presence of a center turn lane), speed limit, and direction of flow. Each of these variables is
addressed in the sections that follow. Values for these attributes have been populated on the roadway
network file for the year 2010.
The roadway network is structured to consolidate data from multiple years and scenarios in a single
TransCAD geographic file. A description of the organizational scheme used to accomplish this
consolidation is provided. In addition, several illustrative examples are provided.
Year-specific input data is used to compute freeflow speed, travel time, and capacity on each link in the
roadway network. Methods used to develop and compute these values are discussed and specific values
are documented herein. This information is placed on a copy of the network rather than the original
input file. Creation of a routable network required by several TransCAD processes is also discussed.
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.
The model’s directory structure allows multiple model output directories to exist alongside a single
input directory. When the travel model is run, files located in the input directory are not modified by
model macros. Instead, if a file is to be modified it will be copied to an output directory and only the
copy is modified.
This approach has several benefits, including the following:
1. All input files are located in one standardized location, making identification of files easy when
edits are required.
2. Because input files are not modified by the travel model macros, it is unlikely that important
data present within input files will be inadvertently overwritten by travel model macros.
3. Since all output files related to a particular model run will be maintained in a single directory,
there will be no confusion about which model scenario is represented by each file.
An example directory structure that would contain travel model input and output files is shown in Figure
1.1.

Figure 1.1: Example Model Run Directory Structure
Multi-Year and Alternative Network

Structure
The Missoula MPO roadway network is designed to store roadway data representing different years in
one consolidated network layer. To accomplish this, selected network attribute names are appended
with a two- through four-digit suffix representing a particular year. By representing multiple networks in
one network file, consistency between baseline and forecast networks is enforced. Furthermore, this
approach eliminates the need to edit multiple network files when making a change in a baseline or
interim year network.

In addition, the network structure allows for the representation of alternative roadway projects such as
roadway widening, realignments, and new facilities that are not tied to a specific network year. These
alternatives can be activated or deactivated individually or in groups, regardless of the network year
that has been selected. While there are some limitations with respect to alternatives sharing the same
link, this capability can be a valuable tool when performing alternatives with the travel model. These
limitations and strategies to overcome them are described below.
Representation of Networks by Year

Each attribute that can vary from year to year (e.g., facility type, area type, number of lanes, direction of
flow, etc.) is represented in the roadway network by an attribute containing a two- through four-digit
numerical suffix. When a particular network is selected for use in the travel model, only attributes with a
suffix matching the selected year are used by the travel model. Of utmost importance is the facility type
attribute. If this attribute is blank on a link for a particular year, that link will be “closed” to traffic (i.e.,
will not exist) in the network when that year is selected. If a valid facility type value is found, then the
remaining attributes specified for that year will be referenced by the travel model.
The roadway network contains data for the calibration year 2010 and can be modified to include
additional data. Missoula MPO staff will be able to and or modify forecast year roadway data in the
network file to represent attributes consistent with the MPO’s Long Range Transportation Plan (LRTP). It
is often necessary to consider multiple interim or buildout year networks (e.g., 2015 or 2050) in addition
to the existing and plan forecast networks. Additional network years can be added at any time through
the following steps:
1. Add new columns to the network link and node tables that will
represent the additional network year (e.g., FT_15, AT_15, etc.);
2. Move these columns so that they are in a convenient location (e.g.,

between the 2010 and 2040 data columns);
3. Fill these columns with data from the corresponding attributes for
either 2010 or 2040; and
4. Adjust the data as necessary.
Because this is a commonly performed task, the model includes a utility that automatically performs
Steps 1 thorough 3 listed above. If alternatives are present in the network file, the utility will also allow
the user to select alternatives to be included in a newly created network year. The utility can also be
used to delete all attributes associated with a particular year. The “Edit Network Year” utility is
accessible from the model dialog box.
Representation of New Facilities

This network structure can represent roadway facilities that do not exist in the current network, but are
planned for future construction. For example, if a new roadway is planned to be built by 2040, it could
be represented in the 2040 roadway network, but not in the base year roadway network. To implement
this, the roadway is added as a new link to the network layer, but is not be assigned a facility type for
the base year. A 2040 facility type would be assigned for this link. When the travel model is run, only
links with a valid facility type are considered by model components that reference the roadway network.
Representation of Network Alternatives

Roadway network alternatives provide a mechanism for testing localized network changes individually
or in combination without the need to create an additional network. Roadway network alternatives are
specified by a set of attributes with the suffix AL (e.g., FT_AL, AT_AL, etc.) and by attributes named ALT
and ALT2, as follows:
 The fields with an AL suffix represent the network attributes used when an alternative is
activated, and
 The “ALT” and “ALT2” fields identify the alternative number associated with each link.
If a particular alternative has been activated prior to a model run, the values in fields containing the AL
suffix will override other network attributes on links where ALT or ALT2 match a selected alternative.
The network structure example sidebar further illustrates application of network alternatives. The
Network Attribute Selection section describes the stepwise procedure used to process network
attributes.

NETWORK STRUCTURE EXAMPLE

To illustrate the concept behind the network structure, a simplified example link data table is shown below.
This table only shows facility type information. Lane, speed override, and area type information follow a similar
theme. In this example network:
 Link 100 exists as a principal arterial (FT = 2) in 2010 and all subsequent years.
 Link 200 is programmed as a principal arterial (exists in 2012 and later).
 Link 300 is planned to be built as a minor arterial (FT = 3) by 2040.
 Link 300 is instead built as a collector (FT = 4) if Alternative 1 is activated.
 Link 400 is a new facility to be built as a minor arterial if Alternative 2 is activated.
 Link 500 exists in 2010 and all future years as a minor arterial, but is closed if Alternative 3 is activated.
EXAMPLE LINK DATASET
ID FT_10 FT_12 FT_40 FT_AL ALT

100 2 2 2 -- --
200 -- 2 2 -- --
300 -- -- 3 4 1
400 -- -- -- 3 2
500 3 3 3 -- 3
Network alternatives can represent scenarios in which roadway attributes differ or scenarios in which
roadways are constructed or removed. For example, an alternative might represent a proposed roadway
widening project that is not included in the 2040 roadway network. This improvement could be included
as an alternative for testing purposes. After adding this one alternative, model scenarios could then be
created that:
1. Represent the base-year network without the roadway widening,

2. Represent the base-year network plus the roadway widening,
3. Represent the 2040 network without the roadway widening, or
4. Represent the 2040 network plus the roadway widening.
As with network attributes that vary by year, absence of facility type data will result in a link being
omitted from consideration in the travel model. It is possible to represent the closure of a roadway by
activating an alternative with a null value for FT_AL on a particular roadway link. This is also useful when
simulating a roadway that is realigned.
This structure does have some limitations. Only two alternatives can occupy the same link, as limited by
the two fields “ALT” and “ALT2.” Also, only one set of alternative attributes can occupy the same link,
limited by the one set of attributes with an “AL” suffix.

These limitations are of particular concern in a scenario where a road exists as a 2-lane facility and
consideration is being given as to whether it should be widened to 4 lanes or 6 lanes. While this scenario
cannot be readily represented in the network alternative structure, this scenario can be represented
through use of either of two suggested options:
1. Create a separate network year (e.g., “09W4” or "40W4”) that represents the road as a 4-lane
facility. Create an alternative that represents the road as a 6-lane facility; or
2. Create an alternative that represents the facility as a 4-lane facility. To run the alternative as a 6-
lane facility, make a copy of the network and change the number of lanes (in the “AL”
attributes) to six before running the model.
Network Attribute Selection

Year and alternative specific network attributes are selected for use in the travel model based on user
selections. The scenario manager that drives the travel model interface maintains user selections
regarding network year and network alternatives. Once these selections have been made, the
automated network initialization step will apply network attributes according to user selections. The
following process is used to assign attribute values to the network for use in the travel model.

When running the travel model, the user must select a network year.
The scenario manager will allow selection of any year where a
complete set of data is present in the roadway network. Specifically,
the user will be able to select any year for which all of the required
year-specific fields are present in the roadway network file. User
selections are saved with a model scenario that is accessible from the
model interface.
1. The user may optionally select to activate specific numbered

alternatives present in the roadway network. A list of available alternatives is generated by
identifying unique values present in the ALT and ALT2 fields. Each unique value is initially
identified as an inactive alternative, but may be set to active by the user. Alternative selections
made by the user are saved with a model scenario that is accessible from the model interface.
2. The network initialization step makes a copy of the input network file and places it in an output
directory specified by the user. One new field is created for each year-specific attribute, but
without the year-specific suffix (e.g., FT, AT, etc.). The field Dir is already present in the network,
so it is not recreated. However, it is modified in the next step.
3. Each new field is populated with data from the corresponding year-specific field matching the
network year selected by the user. For example, if the network year is set to 2012, the field FT
will be filled with data in the field FT_12. Remaining fields will be populated in a similar manner.
4. If any alternatives have been activated, a selection set consisting only of links where either ALT
or ALT2 matches an active alternative is created. Attributes for links in the selection set are filled
with data from the corresponding field ending in _AL. This overwrites any data previously
populated from the year-specific fields. For example, if Alternative 1 is selected, all links where
ALT = 1 or ALT2 = 1 will be selected. For these links only, data in the FT field will be replaced with
data in the FT_AL attribute. This would overwrite data previously read from the FT_12 attribute.
Remaining fields would be populated in a similar manner.
5. Data in the fields that do not include a suffix (e.g., FT, AT, etc.) are referenced for all subsequent
model steps, including the speed, capacity, and volume-delay lookup procedures.

DIRECTION OF FLOW
Direction of flow does not fit within the attribute management scheme, as well as other variables. This is due
to the requirement in the TransCAD software that direction of flow be maintained in the network field “Dir”
at all times. While this fits within the process used to run the model, this requirement can cause difficulties
when editing the network if not addressed. It is important to remember the following points if the direction
of flow varies on a link in different year or alternative networks:
 To display directional arrows for a particular network year, fill the column “Dir” with the value from
the appropriate attribute (e.g., Dir_09).
 The Dir field and year-specific Dir fields should be populated with a 1, -1, or 0 – even for network
years for which links are not active (i.e., year-specific FT is null). The Dir_AL field can be null, but only
if FT_AL is also null.
 When editing route system files, it is vital that the Dir field is set to the appropriate year prior to
opening a route system for editing – especially if any transit routes utilize one-way segments.
Note that these concerns apply only if the Dir attribute varies from year to year.
Network Attribute List

By virtue of the discussions above, the roadway network contains the input attributes listed in Table 1.1.
Additional fields can be added to the network by MPO staff or other users as desired using the standard
tools available in the TransCAD software. Such fields will not be referenced by the travel model, but can
be used to aid in analysis of results.
Table 1.1: Input Network Link Fields
Field Name Description Comments

Maintained automatically by
ID TransCAD Unique ID
TransCAD
Maintained automatically by
Length Link Length in miles
TransCAD
Dir Link Direction of Flow Direction of Flow
Name Street Name
Dir_yyyy Scenario-Specific Direction Field
Scenario-specific facility type (see table 1.3 for
FT_yyyy
definition) yyyy represents a two through
AT_yyyy Scenario-specific area type (see Table 1.4 for definition) four-digit year code (e.g., 09,
Scenario-specific directional number of through lanes 12, 35, 35AA) or the string “AL”
AB_LN_yyyy
(lanes that are used for parking in the off-peak periods
BA_LN_yyyy
are included in this value)


Scenario-specific presence of a center turn lane or
CTLMED_yyyy
median (1 indicates the presence of a center turn lane)
SPLM_yyyy Scenario-specific posted speed limit
This field should be used with
extreme caution. It is intended
TIMEPEN_yyyy Link time penalty (minutes) for use at external stations and
occasionally centroid
connectors.
These fields are described in
BIKE_yyyy
Bicycle and pedestrian facility values the Transit and Non-Motorized
PED_yyyy
Networks section.
AB_FBAM_yyyy
Scenario-specific fields used to hold speed feedback
AB_FBAM_yyyy Fields ending in “AL” are not
results. These fields are optional and usually managed by
BA_FBOP_yyyy present for these fields.
the travel model interface.
BA_FBOP_yyyy
ALT Primary Alternative Number
ALT2 Secondary Alternative Number
Link sub-region:
1 = Within the MPO
SUB_REGION
2 = Missoula County outside the MPO
3 = Ravalli County
CountYY Traffic count volume
Traffic count source: YY represents a two digit year
MDT: Montana Department of Transportation code ranging from 07 through
SourceYY
CITY: Provided by the City containing the link 10.
COUNTY: Provided by the County containing the link
ID field that can be used to match network links to the
CountID
count layer provided by the City of Missoula
FIN_CNT Traffic count selected for model validation
FIN_CNTYR Year that the validation count was collected
Estimated traffic count for use in NCHRP-255
EST_CNT
adjustments
Calibrated base year raw model results for use in NCHRP-
BASEVOL
255 adjustments
NCHRP adjustments will be performed for links with a
DO_NCHRP
“1” in this field
Screenlines Identifies screenline links by screenline number
BIKE_CNT Bicycle count
PED_CNT Pedestrian count
OTHER_CNT Other non-motorized count
CITY City name
Additional Fields These fields are not required by the travel model and have not been fully reviewed.
ZIP Zip code
From regional model dataset –
SPEED_LIMI Speed Limit
may not be up to date.
MDT_RTE MDT Route Number


Retained from the regional
FC Federally designated functional classification
model
URBAN Identifies the Missoula Urbanized Area 13 = Urban Area
CBD Identifies the Missoula CBD
CO County name
GRID Air Quality model grid ID
Identifies links to be included in particulate
PM10
modeling
DEICER Identifies roadways treated with deicer
Added to support air quality
WASH_SAND Identifies roadways treated with sand
model
LOSD_CAP_04 Daily LOS D Capacities from 2004 LRTP
MPO Identifies links within the MPO boundary
In addition to link attributes, several attributes are required on the node layer of the roadway network
file. Centroid nodes are identified by the ZONE attribute on the node layer. Node attributes are listed in
Table 1.2. The PNR and PULSE fields on the node layer are included to support the transit networks.
Table 1.2: Input Network Node Fields
ID TransCAD Unique ID Maintained automatically by TransCAD

Populated only for centroid nodes (including external station
ZONE Traffic Analysis Zone Number
nodes). Null for all non-centroid nodes.
This field must be consistent with the sub-region definition in
SUB_REGION Zone sub-region
the model database. Populated for centroid nodes only.
PNR_yyyy Identifies park and ride nodes Set to 1 for nodes where drive access to transit is permitted
Identifies nodes with pulsed Transfer time in minutes. Overrides the standard transfer time
PULSE_yyyy
transfers at nodes where this field is populated.
Raw modeled turn movements will be saved for nodes on
INT_ID intersection ID (Optional) which a value is present. This ID may be synchronized with a
Synchro network or other traffic database.

WHY SUCH SHORT FIELD NAMES?

Many of the recommended field names (e.g., FT_yy and AT_yy) are very short. This is to facilitate efficient
use of the travel model network and to ensure compatibility with GIS software.
 When exporting TransCAD data for use in ArcMAP and other software packages, an ESRI shapefile is
often used. This file type is limited to 10-digit attribute names. Longer attribute names would are
truncated and can lead to confusion.
 When working with the roadway network, a common task is to select all links with a particular
facility type or area type (e.g., all centroid connectors). It is much more efficient to type “FT=99”
than to type “FAC_TYPE=99, as shown by the keystroke examples below:
o <shift> F T <end shift> = 99  6 keystrokes

o <CAPS> F A C <shift> _ <end shift> T Y P E <CAPS> = 99  15 keystrokes
While this may seem trivial, the increase in efficiency and convenience allowed by short attribute
names is invaluable.
Functional Classification / Facility Type

The functional classification of each roadway link reflects its role in the system of streets and highways.
The term “functional classification” (FC) has specific implications with regards to the administration of
federal-aid highway programs; but travel model networks do not always adhere to these definitions.
Functional classification maintained on the regional model network has been applied to the current
model network and is maintained under the variable FC.
An additional variable named Facility Type (FT) has been added to the network for use in the model to
look up speed, capacity, and volume delay parameters. This will allow facility type to be changed if
necessary during the model calibration and validation process while keeping a record of the functional
class. It will remain possible to summarize model data using either the FT or FC variables. Functional
class / facility type values used in the Missoula MPO Model remain consistent with those used by the
regional model and are listed in Table 1.3. Base year facility type values in the updated model are shown
in Figures 1.2a through 1.2c.

Table 1.3: Functional Classification / Facility Type Values
Functional Classification /
New Code
Facility Type
1 Freeway
2 Principal Arterial
3 Minor Arterial
4 Collector
5 Rural Highway
6 Local Street
7 Ramps
9 Centroid Connector
10 Walk/Bike Centroid Connector

Figure 1.2a: 2010 Facility Type Designations (Regional)



90
$
MISSO ULA
$
$

$
200


12
$
$
$ 


90$


$
38


93
$
RA V A LLI
LEGEND
MPO Boundary County Boundaries
Facility Type
1 - Freeway 2 - Principal Arterial
3 - Minor Arterial 4 - Collector
5 - Rural Highway 6 - Local Street (Not Shown)
7 - Ramp 9 - Centroid Connector (Not Shown)
51 - Non-Motorized Other
0 5 10 15
Miles

Figure 1.2b: Facility Type Designations (MPO)
MISSO ULA



90
$


93
$

$
200


12
$


93
$

12
$


12
$
MPO Boundary
LEGEND
Facility Type


93$
1 - Freeway
3 - Minor Arterial
2 - Principal Arterial
4 - Collector
5 - Rural Highway 6 - Local Street
51 - Non-Motorized
0 1.5 3 4.5
Miles

Figure 1.2c: Facility Type Designations (Urban Detail)
LEGEND
MPO Boundary
Facility Type
1 - Freeway 2 - Principal Arterial
3 - Minor Arterial 4 - Collector
5 - Rural Highway 6 - Local Street
Br 51 - Non-Motorized Other
oa
dw 0 .33 .67 1
ay



St Miles
90 $
Rattlesnake Dr
Mu
llan
Rd


93
$
B ro
adw
a yS St
t n
re
Bu
n
Va
Reserve St
3rd St
5th St
Russell St
e
Av
St
ns
ks
he
oo
ep
Br
St
Higgins Ave
Arthur Ave
South Ave
St


12
$
ks
oo
Br
39th St

Figure 1.3 demonstrates the relationship between the freeway, arterial, collector, and local facility
types. A description of each facility type follows1.
Figure 1.3: Roadway Facility Type Hierarchy
 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. The only freeway included in the Missoula
MPO model is I-90.
 Expressway – The expressway facility type is not present in the Missoula MPO Model.
Expressway facilities are sometimes classified as divided principal arterials, but experience many
features common to freeways. Expressways utilize a higher level of access control than other
arterials and may include some grade-separated intersections. Expressways have higher speed
limits than other principal arterials (e.g., 55 or 65 MPH), provide little or no direct access to local
businesses, may have frontage roads or access roads, and limit signal spacing to at least ½ mile.
The only facility represented in the Missoula MPO Model that might be considered an
expressway is US 93 between Missoula and Stevensville. However, this facility can be
adequately represented as a rural principal arterial.
1
Facility type definitions are adapted from A Policy on Geometric Design of Highways and Streets, 5th
Edition, American Association of State Highway and Transportation Officials (AASHTO), 2004.

 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 freeflow speed on freeway to arterial ramps is often coded as much
slower than the ramp speed limit.
 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. Principal/Major 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). Divided principal arterials 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 and
freeways 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 non-residential 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.
 Centroid Connector – These facilities are the means through which the trip and other data at
the traffic analysis zone (TAZ) level are attached to the street system. Centroid connectors were
generated automatically and are typically connected at mid-block locations.
 Local Streets – Local streets are the lowest roadway classification represented in the travel
model. Local streets provide direct access to residential and some commercial activity. Inclusion
of local streets in a travel model network is exceedingly uncommon. However, the Missoula
MPO Model does include all local streets. This is made possible by the relatively small size of the
Missoula area.

Area Type
Area type is an attribute assigned to each TAZ and roadway and is based on the activity level and
character of the 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 socioeconomic characteristics and then transferred to the roadway network.
Area type is an attribute that can and should vary with time. Therefore, it was important that area type
definitions were 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. Area types used in the Missoula MPO Model include central business district (CBD),
urban, suburban, and rural as shown in Table 1.4.
Table 1.4: Area Type Categories
Area Type Code Area Type

1 CBD
2 Urban
3 Suburban
4 Rural
Zones with the CBD area type were identified using aerial photography and through discussions with the
MPO Staff. Initial identification of non-CBD area types was done at the TAZ level by applying the area
type criteria shown in Table 1.5 to non-CBD zones based on the model socioeconomic dataset.
Table 1.5: Area Type Model Criteria
Area Type Population/ Sq. Mile Employment/ Sq. Mile

1 CBD n/a n/a
2 Urban 4,000 + 4,000 – 19,999
3 Suburban 300 – 3,999 300 – 3,999
4 Rural 0 - 299 0 – 299
Note: For each TAZ, the most dense non-CBD area type is applied for which at least one of the criteria is
met.

After the initial criteria were applied, a manual smoothing process was used to determine the base year
area type designation for each zone. This was supported by overlaying the model TAZ structure on aerial
photography using the Google Earth software program and through discussions with MPO and MDT
Staff. The initially computed area types were then adjusted to:
1. Fill in holes and gaps in contiguous urban and suburban areas,

2. More accurately define existing land uses based on local knowledge, and
3. More accurately define the transition between urban, suburban, and rural area types through a
visual evaluation of the aerial photography and roadway layers.
Once area type values were assigned to each TAZ for the 2010 base year, roadway area type was
assigned to the roadway network. This process started by applying a “Tag” operation to assign area type
to each link based on the closest TAZ. For roadways that are bordered by different area types on either
side, the denser (e.g., urban rather than suburban) area type was assigned. Divided highways bordered
by different area type values were assigned the denser area type for both directions. Area type values
should not differ by direction of travel on such a facility. Furthermore, interchanges occurring on or
directly adjacent to area type borders have been assigned the denser area type. For links crossing an
area type boundary, the most appropriate area type was selected based a visual evaluation.
For roadways that do not lie on an area type boundary, assignment of roadway area type was
straightforward and not further adjusted. Area type designations for TAZs and roadways are shown in
Figures 1.4a through 1.4c.

Figure 1.4a: Area Type Designations (Regional)



90
$
MISSO ULA

$
200


12
$



90$


38$
RA V A LLI


93
$
LEGEND
County Boundaries MPO Boundary
TAZ Area Type
CBD Urban
Suburban Rural
Link Area Type
CBD Urban
Suburban Rural
0 10 20 30
Miles

Figure 1.4b: Area Type Designations (MPO)



90
$


93
$

$
200


12
$


93
$

12
$


12
$
LEGEND
MPO Boundary
TAZ Area Type


93
$ CBD
Suburban
Urban
Rural
Link Area Type
CBD Urban
Suburban Rural
0 1.5 3 4.5
Miles

Figure 1.4c: Area Type Designations (Urban Detail)
Br
oa
dw
ay



St
90 $
Rattlesnake Dr
Mu
llan
Rd


93
$
Bro
adw
ay
S St
t n
re
Bu
n
Va
Reserve St
S 3rd St
S 5th St
Russell St
e
Av
St
s
en
k s
oo
h
ep
Br
St
Higgins Ave
Arthur Ave
South Ave
St


12
$
LEGEND
s MPO Boundary
o ok TAZ Area Type
Br CBD Urban
Suburban Rural
Link Area Type
39th St CBD Urban
Suburban Rural
0 .33 .67 1
Miles

Link Speeds
Network speeds are used in the trip distribution model to distribute trips throughout the region and in
the trip assignment model to route traffic on the roadway network.
Link freeflow 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.
Estimating Link Speeds

Speed limit data is available for 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 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. A comprehensive speed survey
conducted in the Colorado North Front Range1 provides sufficient information to estimate such a model.
The North Front Range survey contains off-peak (approximately freeflow) speed data as well as speed
limit for a large number of corridors around the region. Analysis of this survey’s data results in the
relationships shown in Table 1.6.
Table 1.6: Ratio of Freeflow Speed (Off-Peak) to Speed Limit
Fort Collins CBD Other CBD Urban Suburban Rural

Freeway n/a n/a n/a n/a 1.00
Expressway n/a n/a 0.96* 1.02 1.01
Principal Arterial 0.95* 0.73+ 0.83+ 0.87+ 0.94
Minor Arterial n/a 0.63* 0.94 0.90+ 0.94
Collector n/a n/a 0.82* 0.93* 0.95
* Very small sample size

+ Very large sample size
Not indicated: moderate sample size
Source: 2005 (Colorado) North Front Range Regional Speed Study
1
2005 North Front Range Regional Speed Study (North Front Range Metropolitan Planning Organization, 2005)

A visual review of the data suggests that a set of factors can be applied to speed limits based on facility
type and area type to produce freeflow speeds. However, this data is not specific to the Missoula area.
The Missoula MPO Model uses a set of freeflow to speed limit conversion factors that minimize the
difference between speed limit and freeflow speed. The factors in Table 1.7 are applied in the Missoula
MPO Travel Model.
Table 1.7: Speed Limit to Freeflow Speed Conversion Factors
Area Type
ID Functional Class
CBD Urban Suburban Rural
1 Freeway 1 (no adjustment)
2 Principal Arterial
1
3 Minor Arterial
4 Collectors 1 0.85 0.85 0.85
5 Rural Highway 1
6 Local Street
0.75
7 Ramp
9 Centroid Connector 1 (no adjustment, values may be specified or obtained
10 Walk/Bike Centroid Connector from lookup table)
For centroid connectors, values in Table 1.8 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 1.8: Centroid Connector Freeflow Speeds
Area Type
ID Functional Class
9 Centroid Connector 15 20 20 35
10 Walk/Bike Centroid Connector n/a n/a n/a n/a
Note: Walk and bike speeds are set by the mode choice model.

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 using equation (1). An example application for a one-mile link with a
30 mile per hour freeflow speed is included as equation (ex. 1)
(1)
Where:
= link travel time in minutes
= link velocity (speed) in miles per hour
= link distance in miles
= minutes per hour (unit conversion)
(ex. 1 – Example application)
Freeflow travel time is populated on the roadway network for each link using the freeflow speed.
Congested travel time is discussed in the Traffic Assignment chapter.
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 are 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
Highway Capacity Manual. Transportation Research Board, 2000.

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
applied1.
The HCM suggests adjusting flow rate (traffic volume) according to equation (2).
(2)
( )
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 with hourly capacity ( ), Equation (2) 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 (2) can be
used to compute per lane capacity for freeways and expressways. Equation (3) was used to compute
hourly capacities for rural and freeway facilities.
1
HCM 2000, p. 13-11

(3)
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 speed1. Ideal capacities based on typical
freeflow speeds are shown in Table 1.9, along with adjusted capacities computed using Equation (2).
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 characteristics should be adjusted. This is accomplished
by using updated area type information in forecast-year model runs.
Table 1.9: Ideal and Adjusted Capacities for Freeways and Expressways based on HCM 2000
Ideal
Freeflow Capacity Adjusted Capacity
Facility Area
Speed (Upper Limit PHF FHV FP (Upper Limit LOS
Type Type
(mph) LOS E, E, pc/h/ln)
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
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 Equation (4)2.
(4)
Where:
= capacity
= 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
1
HCM 2000, p. 23-5
2
HCM 2000, p. 30-5

= 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 equations above account for 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 and
bicycle effects are instead captured in the area type adjustment. Other variables can be approximated
based on the facility type and area type of each link. Additionally, a regional travel model must rely on
the number of through lanes on each link, rather than the number of approach lanes at each
intersection. This can be 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.
Equation (4) can be simplified to Equation (5) for use in a regional travel modeling context. Assumed
values for adjustment factors that vary by facility type and area type are shown in Table 1.10, along with
the resulting capacity values. An example application of equation (5) for a suburban principal arterial
with one lane in each direction is shown as equation (ex. 5). Capacities used in the travel model have
been rounded to 10 vehicles per hour per lane.
(5)
Where:
= directional capacity
= 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
961 (ex. 5 – example application)

Table 1.10: Link Capacity Adjustment Factors and Resulting Hourly Capacity per Lane
Capacity
FT AT
per Lane
CBD 0.76 0.45 1.30 740
Urban 0.95 0.45 1.30 920
Principal Arterial
Suburban 0.99 0.45 1.30 960
Rural (Expressway) 0.97 0.55 1.30 1,200
CBD 0.76 0.45 1.15 650
Minor Arterial Urban 0.95 0.42 1.15 760
Suburban / Rural 0.99 0.42 1.15 790
CBD 0.75 0.45 1.05 590
Collector Urban 0.95 0.41 1.05 680
Suburban / Rural 0.99 0.41 1.05 710
CBD 0.74 0.45 1.00 550
Local Street Urban 0.95 0.40 1.00 630
Suburban / Rural 0.99 0.40 1.00 660
Turn Lane Adjustments
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
Missoula MPO roadway network contains a specific variable that identifies roadway corridors where
medians or center left turn lanes are present. Any corridors where all possible left turns are served by a
left turn lane are identified 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/Major Arterial:
o Left turn lane present: Add 0.25 lanes (0.125 lanes in each direction)
o No left turn lane present: Subtract 0.5 lanes (0.25 lanes in each direction)
 Minor Arterial:
o Left turn lane present: Add 0.5 lanes (0.25 lanes in each direction)
o No left turn lane present: Subtract 0.25 lanes (0.125 lanes in each direction)
No center turn lane or median adjustments are made on expressway, collector, or local facilities.
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 1.13. These capacities are then adjusted to account
for the presence of a center turn lane.

Table 1.11: Roadway Capacities (vehicles per hour per lane, upper-limit LOS E)

1 Freeway 2,200 2,200 2,200 2,100
2 Principal Arterial 740 920 960 1,200
3 Minor Arterial 650 760 790 790
4 Collectors 590 680 710 710
5 Rural Highway 590 680 710 710
6 Local Street 550 630 660 660
7 Ramp 700 760 790 790
9 Centroid Connectors 10,000 10,000 10,000 10,000
Example Capacity Calculation
An example capacity calculation for a 2-lane suburban principal arterial link with a median or center turn
lanes is performed as follows:
1. Obtain the basic link capacity from the capacity lookup table: 960 vehicles per hour per lane.
2. Observe that the link has 2 lanes in each direction, plus a median center turn lane. Add 0.125
lanes in each direction to account for the median or center turn lane (as discussed above).
Resulting directional lanes: 2.125
3. Multiply the directional capacity (960) by the directional number of lanes (2.125), resulting in a
directional capacity of 1080 directional vehicles per hour.
Output Network Fields

The travel model network processing macros perform the calculations described above. In addition,
some fields that are required solely for transit and non-motorized processing are added to the output
roadway network – these fields are described further in the transit and non-motorized network section.

Chapter 2: Transit and Non-Motorized

Networks
The travel model uses transit and non-motorized networks to build shortest paths between each zone
pair for pedestrian, bicycle, and transit trips. The resulting shortest paths are used along with paths built
for vehicle trips as inputs to the mode choice model. The Missoula MPO Model uses information stored
on the roadway network layer and a TransCAD route system to represent the transit and non-motorized
networks. For non-motorized pathbuilding, a bicycle and pedestrian scoring system represents the
varying levels of facility quality. For transit pathbuilding, the Missoula MPO model uses the “Pathfinder”
method provided the TransCAD software.
Transit Networks
Transit networks in TransCAD are made up of two separate but connected layers: the transit route
system and the transit line layer. Information from these two layers is combined as shown in Figure 2.1
to allow representation of the walk, drive, and in-vehicle components of a transit trip. Because these
layers are connected, information on the transit line layer, such as link travel times and centroid data, is
available to the route system. However, this also results in a requirement that the roadway and transit
networks are maintained in a manner that prevents them from becoming inconsistent with each other.
Figure 2.1: Connections between the Route System and Transit Line Layer
Chapter 2: Transit and Non-Motorized Networks- 32

To enforce consistency between the roadway and transit line layers, the model input dataset consists of
only one roadway geographic file (roadway/transit line layer). When the travel model is run, separate
copies of this file are made for use in roadway and transit modeling. The roadway line layer includes
information such as link capacity and travel time, as described in the Roadway Network chapter. The
transit line layer includes all of the information present on the roadway line layer, but also includes
information such as walk speed and additional walk access links. The transit line layer and route system
are combined to create a complete transit network. Figure 2.2 demonstrates the process of separating
the input roadway/transit line layer into separate roadway and transit line layers. Since transit routes in
this environment are coded over roadway links with unique IDs, any changes to the input
roadway/transit line layer require that the route system also be modified.
Figure 2.2: Roadway and Transit Line Layer Processing
During transit path-building, the shortest transit path is determined by taking into account the travel
impedance for each part of a traveler’s trip. Different weights are applied to different types of travel
within a trip. For example, out of vehicle travel such as time waiting for a bus is considered to be 2 to 3
times “worse” than in-vehicle travel time. These weights are set to be consistent with mode choice
model parameters. The transit pathbuilder identifies the shortest path between each pair of traffic
analysis zones. The resulting shortest path matrix includes “skimmed” variables such as walk time, wait
time, transit fare, and in-vehicle travel time that form key inputs to the mode choice model.

Once mode choice is complete and the number of transit trips has been computed, the transit network
is used again in transit assignment. In transit assignment, the transit trips computed by the mode choice
model are assigned to the most appropriate route or routes in the transit route system. These
assignment results can be viewed using standard TransCAD tools, which are described further in the
model user’s guide.
Because transit networks include period-specific information and include access and egress links, several
transit network files are created during the modeling process. The Missoula MPO Model creates the four
transit networks shown in Table 2.1.
Table 2.1: Transit Networks
Walk Access Drive Access

Off-Peak Period TransitOPW.tnw TransitOPD.tnw
Peak Period TransitPKW.tnw TransitPKD.tnw
Note: These are default filenames used by the model and can be
changed to user-specified values.
Transit Route System

Transit routes and stops are represented within the TransCAD route system. Contents of the route
system are based on schedule data from the Mountain Line and UM transit systems.
Route System Attributes
Each route is represented as a unique feature in the route system layer. Like the line layer, the route
system layer includes attributes for each feature. These attributes contain route-specific information
such as route name, operator, and headway. Notably absent from the list of route system attributes is
travel time. The TransCAD model computes stop-to-stop travel time using attributes on the underlying
link layer rather than attributes stored directly on the route system. A list of route attributes is included
as Table 2.2.
Table 2.2: Route Attributes

Route_ID TransCAD Unique ID Maintained automatically by TransCAD
Route_Name Descriptive route name These fields are used for informational and summarization
Route_No Route number purposes only
AM_Headway Peak route headway
MD_Headway Off-peak route headway
1 = Mountain Line routes
Mode Transit Mode
2 = University shuttle routes


This value represents the average fare paid by non-
Fare Transit fare in dollars university students. The fare is set to zero for UM shuttle
routes.
FF_Time Route freeflow travel time These fields were used to calibrate the relationship
PKSch_Time between passenger vehicle and bus travel times. They are
Scheduled travel time
OPSchTime not used by the model and are retained for reference only.
ROUTE HEADWAYS
The headway for each transit route is calculated separately for the peak and off-peak time periods. As
discussed in the Trip Assignment chapter, the peak time period includes 7:00 AM through 8:00 AM and
5:00 PM through 6:00 PM. For the transit system, Headway is defined as the average headway for all
busses starting a route within 15 minutes of the peak period. Similarly, off-peak headway is calculated as
the average headway for all remaining busses. Headway assumptions are listed in Table 2.3.
Table 2.3: Route Headway Assumptions
Route Peak Headway (minutes) Off-Peak Headway (minutes)

Mountain Line Route 1 35 35
East Broadway Park and Ride 20 20
South Campus Park and Ride 10 10
COT Park and Ride 15 15
Transit Stops
The transit route system includes transit stop locations coded based on transit schedules and GIS data.
Routes can only be boarded or alighted at these stops. To facilitate a connection to the transit line layer,
all transit stops must be coded to coincide with a distinct node on the input roadway network.
Furthermore, only one stop can be coded per direction, per route, per node. Attributes maintained on
the route stop layer are listed in Table 2.4.

The TransCAD route system structure does not in itself require that transit stops are located at nodes on
the transit line layer. However, when the transit network processing model step is performed, each
transit stop is matched to the closest node on the transit line layer. If the route system contains stops
that cannot be matched to nodes, the model will fail to run.
Note: Detailed transit network coding guidelines are provided in the Missoula MPO Travel Model User’s
Guide.
Table 2.4: Route Stop Attributes

ID TransCAD Unique ID
Route_ID ID of the route associated with the stop
These fields are all maintained
Used to associate a stop with one of multiple times a route
Pass_Count automatically by TransCAD and
passes a particular node.
are read-only.
Milepost Distance from the route starting point
STOP_ID Unique stop ID (identical to ID)
This field is filled in
Identifies the ID of the node on the network layer that
NearNode automatically when the model
matches the route stop
is run.
Transit Line Layer

Some transit variables are maintained on a copy of the roadway network rather than the route system,
allowing for interaction between the roadway and transit networks. In particular, transit travel time is
calculated as a function of vehicle travel time on each link. The transit line layer also provides a
connection between TAZ centroids and route stops. This connection is provided in the form of centroids,
roadway links, non-motorized links, and walk access/egress links and the roadway network.
Transit Travel Time
Transit travel time is computed as a function of congested link time. Congested link times are increased
(and speeds are correspondingly decreased) using a transit time factor. This factor represents the
observed difference between transit route times and congested network times. To compute transit time
factors, published times for each transit route were compared to congested peak and off-peak travel
times present on the roadway network. Regression analysis was then used to develop transit travel time
factors. Network and scheduled transit speeds are plotted along with the resulting best-fit line in Figures
2.3 and 2.4. The model uses a transit time factor of 1.9618 to convert vehicle travel time to bus travel
time.

Figure 2.3: Off-Peak Transit Time factors
Off Peak Schedule

80.00
70.00
60.00
Scheduled Bus Time
50.00
40.00
30.00
20.00
10.00
0.00
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00
Vehicle Travel Time
y = 1.9618x
Offpeak Scheduled Time Linear (Offpeak Scheduled Time)
Figure 2.4: Peak Transit Time Factors
Peak Schedule
80.00
70.00
60.00
Scheduled Bus Time
50.00
40.00
30.00
20.00
10.00
0.00
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00
Vehicle Travel Time
y = 1.9618x
Peak Scheduled Time Linear (Peak Scheduled Time)

During roadway and transit network processing, the fields listed in Table 2.5 are populated with data
required for transit and non-motorized modeling. When speed feedback is run, transit speeds are
calculated based on the congested speeds resulting from speed feedback.
Table 2.5: Key fields in the Transit Line Layer

AB_OPTRTIM Based on the AM congested link time resulting from speed
Off-peak period transit time
BA_OPTRTIM feedback
AB_PKTRTIM Based on the off-peak link time resulting from speed
Peak period transit time
BA_PKTRTIM feedback
AB_OPTRSPD
Off-peak period transit speed
BA_OPTRSPD Calculated based on link time and length (for reference
AB_PKTRSPD only)
Peak period transit speed
BA_PKTRSPD
Used for transit walk access and non-motorized
WALK_TIME Walk travel time
pathbuilding
BIKE_TIME Bicycle travel time Used for non-motorized pathbuilding
Used to identify links that can be used for walk
Mode Non-transit mode field
access/egress
Walk Access and Egress
The transit line layer also represents the connection between TAZ centroids and transit route stops.
With the exception of park and ride trips, all transit trips must start and end on foot1. Several
approaches are available for representing walk access to transit in TransCAD:
 Direct Walk Links: A set of walk access/egress links provides a direct connection between each
TAZ centroid and all transit stops within a specified distance.
 Roadway Network Walk Links: Walk access and egress occurs using the roadway network,
including centroid connectors and most roadways. Walk access cannot occur on links where
walk access has been prohibited, such as on freeway and freeway ramp links.
 Combined Walk Links and Roadway Network: Walk access links are created between transit
stops and immediately adjacent TAZs. Centroid connectors and the local street layer are used to
facilitate walk access and egress for TAZs that are not immediately adjacent to transit stops.
1
Bicycle access and egress to transit is not modeled explicitly, but is instead modeled as walk access
and egress.

The Missoula MPO model connects TAZs to transit stops using the combined walk access link and
roadway network approach. This approach allows representation of direct access to transit stops
adjacent to TAZs while representing the increased walk distance to and from zones that are near, but
not directly adjacent, to transit stops. An example of walk paths from several zones to a specified transit
stop is demonstrated in Figure 2.5.
Figure 2.5: Example Walk Access Paths
Example Paths:
Green Path: 0.06 miles / 1.2 minute walk
Orange Path: 0.37 miles / 7.4 minute walk
The TransCAD model implements this methodology by automatically drawing walk links from each stop
to TAZ centroids within a 1/8 mile radius. Walk access links are created in the transit line layer and are
not present in the roadway line layer. A facility type value of 10 prevents use of walk access links by
vehicles.
A walk speed of 3 mph is assigned to all links on which walk access is permitted. This walk speed is used
to compute a walk time (in minutes). For example, a walk time of 5 minutes would be assigned to a link
one quarter-mile in length.
Timed Transfers
For transfers at most locations, the transfer wait time is computed as one-half the headway of the route
being boarded. However, the Mountain Line transit system is timed to provide quick transfers at the
downtown transfer center. The Missoula Model applies a lower transfer time at this location using a
pulse transfer time value stored on the transit node layer. The node field PULSE will override the default
transfer time for all transfers occurring at a node if it contains a positive value. In the calibrated base
year model, the PULSE field is populated at the downtown transfer center with a value of 5 minutes.
Drive Access
The transit network also connects TAZs to route stops to represent transit trips made using a park and
ride. Drive access connectivity is only provided in the direction from TAZs to route stops. This is

consistent with the mode choice and transit modeling convention that transit pathbuilding and
assignment is performed in Production/Attraction format rather than Origin/Destination format. By
following this convention, it is possible to limit drive access to transit to the production (or home) end of
each trip. Because transit riders do not typically have access to a vehicle at the attraction (or non-home)
end of a trip, transit egress is limited to the walk mode.
Drive access to transit is provided using centroid connectors and roadway links. Zone to park and ride
travel times are computed using peak and off-peak travel times on the roadway network. However,
drive access is only provided to specially designated park and ride nodes. Park and ride nodes are
identified by populating the PNR_yyyy field on the input network node layer with the number 1.
Transit Pathbuilding
The transit networks are built in the TransCAD software for use with the Pathfinder transit shortest path
method. The Pathfinder method is unique to the TransCAD software and builds paths using a weighted
generalized cost approach. Each component of a transit trip is converted into a common unit, allowing
application of different weights to each trip component. Pathfinder weights have been set for
consistency with coefficients in the mode choice model.
The pathfinder evaluates all possible transit paths between each zone pair and identifies the path with
the lowest generalized cost. Path components considered by the Pathbuilder setup in the Missoula MPO
model are listed along with pathbuilding weights Table 2.6.
Table 2.6: Transit Pathbuilding Weights
Variable Description Weight

Time spent walking from the production TAZ centroid to the transit stop (for
Walk Access Time 2.6
walk access trips only)
Time spent driving from the production TAZ centroid to a park and ride (for
Drive Access Time 1
drive access trips only)
Drive Access Cost Auto operating cost associated with drive access (for drive access trips only) 1*
Drive Access Terminal time at the production TAZ consistent with that for an auto trip (for
2.6
Terminal Time drive access trips only)
Time spent waiting for the first bus to arrive, computed as one-half of the
Initial Wait Time
route headway. The short component of the initial wait time includes a wait up 2.6**
(Short)
to 7.5 minutes
Initial Wait Time
Initial wait time exceeding 7.5 minutes 1**
(Long)
In-Vehicle Travel
Time spent riding or waiting in a transit vehicle 1
Time
Transfer Wait
Time spent walking between stops for a transfer (if applicable) 2.6
Time
Transfer Walk
Time spent walking between stops for a transfer (if applicable) 2.6
Time

Variable Description Weight

Transfer Penalty
Additional transfer penalty (calibration parameter) 2.6
Time
Egress Walk Time Time spent walking from the transit stop to the attraction TAZ centroid 2.6
Fare Transit fare paid for the trip 1*
* Travel time variables are converted for consistency with cost variables using the value of time documented in the mode choice
model specification.
** Weighted initial wait time is computed for each route based on the combined short and long wait times. The transit network
weight is set to 1 for initial wait time. Transit shortest path matrices are post-processed to represent un-weighted short and long
initial wait time.
Non-Motorized Networks
The Missoula MPO Model roadway network includes attributes that describe the presence and quality
of non-motorized facilities on roadway links within the MPO. In addition, multi-use paths are included in
the roadway network file to allow inclusion of these facilities in the non-motorized pathbuilding process.
Non-motorized paths are used to build non-motorized shortest paths for use in Mode Choice.
Bicycle Network
The bicycle network has been coded on the roadway layer using a sliding scale system. The sliding scale
will be used to penalize facilities with poor bicycle facilities. Penalties are applied in the form of a travel
time weight, which is consistent with the non-motorized facility score coefficient in the mode choice
model. Bicycle scores have been assigned using a set of rules, followed by review and discussion with
Missoula MPO Staff. The scoring system used is compatible with the Bicycle Compatibility Index (BCI)
and could be updated to include more detailed information in a future model update.
Links in the roadway network are coded with the bicycle scoring values listed in Table 2.7. The scoring
system uses a value from 1 to 5 do indicate the quality of bicycle facilities on each network link. While
decimal values can be used, the model has been implemented using integer values only. Initial bike
scoring was been performed according to the following rules:
 Bicycles are prohibited on freeways and freeway ramps.

 All multi-use paths receive a score of 1.
 All local streets receive a score of 1.
 For Collector and Arterial Streets:
o Adjacent multi-use path: 1
o Collector with Bike Lane: 1
o Arterial with Bike Lane: 3
o Collector with no bike lane: 3

o Arterial with no bike lane: 5
After completion of initial scoring, the bicycle network scoring was reviewed by MPO Staff. Known
problem areas were revised with higher penalties.
Table 2.7: Bicycle Scoring Values
Score Value
1 Good Bicycle Facilities
2 Good to Moderate
3 Moderate Bicycle Facilities
4 Moderate to Inadequate
5 Inadequate Bicycle Facilities, but Bicycle Traffic is Allowed
99 Bicycle Traffic Prohibited
Pedestrian Network
The pedestrian network uses a scoring system identical to the bicycle network scoring system. Initial
coding of pedestrian scores used the following rules:
 Pedestrians are prohibited on freeways and freeway ramps.

 All multi-use paths receive a score of 1.
 Local Streets:
o Streets identified as having deficient sidewalks: 3
o All other streets (includes all local streets outside the City of Missoula): 1
 For Collector and Arterial Streets:
o Sidewalk or adjacent multi-use path: 1
o Collector with a sidewalk on one side: 2
o Collector with no sidewalks: 3
o Arterial with a sidewalk on one side only: 4
o Arterial with no sidewalks: 5
After completion of initial scoring, the pedestrian network scoring was reviewed by MPO Staff. Known
problem areas were revised with higher penalties.

Table 2.8: Output Roadway Network Fields

This field is included to support transit
Mode Transit network mode field
networks.
FF_SPD Freeflow travel speed (mph)
AB_OPSPD
Directional off-peak travel speed (mph) These fields are filled with values
BA_OPSPD
calculated in the speed feedback
AB_PKSPD
Directional peak congested travel speed (mph) process.
BA_PKSPD
FF_TIME Freeflow travel time (minutes)
AB_OPTIME
Directional off-peak travel time (mph) Travel time fields are calculated based
BA_OPTIME
on link speed and length.
AB_PKTIME
Directional peak congested travel time (mph)
BA_PKTIME
These fields are not required for the
OPTRTIM
Off-peak and peak transit time (dummy fields) Missoula MPO model, but are included
PKTRTIM to facilitate future expandability.
AB_OPTRSPD
These times are computed as described
BA_OPTRSPD
Off-peak and peak directional transit time in the Transit and Non-Motorized
AB_PKTRSPD
Networks section.
BA_PKTRSPD
AB_OPTRTIM
BA_OPTRTIM Travel time fields are calculated based
Peak and off-peak transit time
AB_PKTRTIM on link speed and length.
BA_PKTRTIM
LANE_CAP Hourly directional per-lane capacity
AB_PKCAP Total directional capacity for the AM and PM
BA_PKCAP peak hours
BA_OPCAP Total directional capacity for the off-peak Off-peak capacities are described further
BA_OPCAP period in the Traffic Assignment chapter.
Volume delay equations are described
ALPHA
Volume delay equation parameters further in the Traffic Assignment
BETA
chapter.
WALK_TIME
Non-motorized travel time
BIKE_TIME
WALK_IMP Non-motorized impedance reflecting
Non-motorized impedance
BIKE_IMP time adjusted for walk and bike scores
WALK_WTD This field is skimmed and then used to
Non-motorized score multiplied by link length compute a distance-weighted average
BIKE_WTD
transit score.

Chapter 3: Trip Generation

This chapter describes the process used to develop the Trip Generation component of the Missoula
MPO Model. Trip production and attraction rates, special generators, allocation models, balancing
methods, trip purpose, external travel, and other information related to the trip generation model are
addressed herein. Trip generation rates are defined in units of daily person trips for all modes, including
auto/truck, bus/transit, pedestrian, and bicycle.
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 where activities occur as
represented by socioeconomic data (e.g., households, employment). Productions and attractions are
estimated for each Traffic Analysis Zone (TAZ) by trip purpose, and then balanced at the regional level so
that total productions and attractions are equal. In some cases, production and attraction allocation
sub-models are applied to better represent the geographic distribution of trip-ends. 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.
The trip generation model is defined such that trips are produced at home and are generally attracted to
other places of activity (employment). Hence, the terms “productions” and “attractions” are the
fundamental variables for defining the trip ends associated with travel. Productions generally occur at
the home end of a trip; and attractions are typically associated with places of employment. Some
exceptions are described in the following sections, but this method of defining productions and
attractions is generally used for trips internal to the modeling area. External travel is also addressed.
External trips are defined as: 1) external-external (EE) if both trip ends are outside of the modeling area,
and 2) internal-external or external-internal (IE/EI) if one end of the trip is inside and the other end is
outside of the modeling area.
This chapter opens with a discussion of data sources and trip purposes then proceeds to an explanation
of how trip production and attraction rates are established. A section on sub-allocation models, special
generators, and trip balancing follow. Finally, external trips are described, including trips through, to,
and from the region.
Data Sources
The primary data source for estimating trip production and attraction rates is the household-based
travel diary survey, which has three components – household, person, and trip. Since these surveys are
household-based, they provide excellent information with regard to household trip-making and are
especially well suited for estimating trip production rates. Household travel surveys also provide good
information for estimating trip attraction rates since employment data is recorded as part of the person
Chapter 3: Trip Generation - 44

component of the survey. The trip component provides information about the purpose of each trip and
the origin and destination of each trip record.
Several data sources were considered, including the 2009 National Household Travel Survey (NHTS) add-
on dataset for Rapid City, South Dakota, the NHTS add-on dataset for Cedar Rapids, Iowa, a nationwide
application of NHTS data, and the Colorado North Front Range (NFR) Household Travel survey. After
careful review of each potential dataset, the NFR dataset was selected as the primary source for
borrowed data. The NFR is reasonably similar in climate and character to Missoula. Like Missoula, the
NFR includes major university activity and has a higher than average non-motorized (walk and bicycle)
travel share.
Traffic Analysis Zones

Traffic analysis zones (TAZs) are small areas containing the land use data that is used as the foundation
for trip-making in the travel model. For the Missoula MPO Model, the TAZ layer is identical to the 2000
Census block geography. The TAZ structure 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. In most travel models, centroid connectors simulate local and
neighborhood streets. However, the Missoula MPO Model includes local streets in the roadway
network. Because the census block geography is more detailed than typical TAZ geography used in most
models, centroid connectors in the Missoula model connect activity directly to the local street network.
Household Disaggregation Models

The model input data includes information about the number of households in each TAZ. To supplement
this data, the average household size and income for each TAZ has been obtained from US Census data.
The model then uses household disaggregation models to estimate the univariate distribution of
households by size and by income group for each TAZ. Once these distributions have been estimated,
the model uses an iterative proportional factoring process to develop bivariate distributions of
households by income and size for each TAZ.
Household disaggregation models use known variables to establish 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. 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 data.
The distribution curves must always sum to 100% and for the household size model, the results must be
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 to allow median income data to be expressed in any unit, so long as the units are
consistent for all zones. Data may be input in 1999 dollars (consistent with the 2000 Census), or in some
other unit if desired. The Missoula MPO Model has been implemented using 1999 dollars.
Household Size Disaggregation Model

Model trip rates are classified by five 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 3.1. The resulting model is defined as a lookup table provided
with the travel model input dataset.
Figure 3.1: Household Size Disaggregation Curves
100.0%
90.0%
80.0%
70.0%
Percent of Hosueholds
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
1 1.5 2 2.5 3 3.5 4 4.5 5
Average Household Size
1 Person Trendline 2 Person Trendline 3 Person Trendline 4 Person Trendline 5 Person Trendline
1 Person Household 2 Person Household 3 Person Household 4 Person Household 5 Person Household

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 3.1. Disaggregation
curves are shown along with Census data in Figure 3.2. The resulting model is defined as a lookup table
provided with the travel model input dataset.
Table 3.1: Income Group Definitions
Income Group Income Range

Low $19,999 and lower
Medium $20,000 – $74,999
High $75,000 and higher
Figure 3.2: Household Income Disaggregation Model
100.00%
90.00%
80.00%
70.00%
Percent of Households
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4
Ratio of Household Income to Regional Median Income
Low Income HH Medium Income HH High Income HH

Low Income Trendline Medium Income Trendline High Income Trendline

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 3.2, was obtained from the 2008 Public Use Microsample (PUMS) dataset. The
process used to develop TAZ-level household data at the bivariate level is described further in Appendix
2.1.
Table 3.2: Bivariate Household Distribution for Missoula Ravalli, and Mineral Counties
Income Group 1 Person 2 Person 3 Person 4 Person 5+ Person Total

Low 16.4% 7.5% 1.0% 0.5% 1.1% 26.5%
Medium 11.1% 23.0% 8.4% 5.6% 1.5% 49.6%
High 0.9% 9.0% 5.7% 3.7% 4.6% 23.9%
Total 28.4% 39.5% 15.1% 9.8% 7.2% 100.0%
Source: 2008 PUMS Dataset for Missoula, Ravalli, and Mineral Counties (Data is not available for Missoula and
Ravalli counties alone)
Trip Purposes
Trip purpose is used in travel models to categorize various types of household-based trips that have
similar characteristics, such as location of production or attraction end, trip length, auto occupancy, and
others. In this manner, trip rates by trip purpose are sensitive to the specific socioeconomic data
associated with each trip type. In general, it is advisable to disaggregate trips by trip purpose only to the
point that the base and horizon year activity data can support them.
The regional model utilizes three trip purposes to describe personal trip-making. For this model update,
the number of trip purposes was expanded to six by further disaggregating home-based non-work and
non-home-based trips as shown in Table 3.3.
Table 3.3: Trip Purposes
Regional Model Updated Model

Home-Based Work (HBW) Home-Based Work (HBW)
Home-Based Shop (HBS)
Home-Based Non-Work (HBNW) Home-Based University (HBU)
Home-Based Other (HBO)
Work-Based-Other (WBO)
Non-Home-Based (NHB)
Other-Based-Other (OBO)

Generally, a trip is defined as a distinct travel movement from one clearly identifiable starting
place/activity to another with a distance of more than 1 block. In some cases, two or more trips may be
linked to reflect the true trip purpose and to factor out convenience stops, such as stopping for gas on
the way from home to work. In these cases, the travel model represents the linked trip as two separate
trips. The specific trip purpose definitions are as follows:
 Home-Based Work (HBW): Commute trips between home and work and vice versa (e.g.,
includes trips between work and home).
 Home-Based Shop (HBS): Trips between home and shopping locations for the purpose of
shopping.
 Home-Based University (HBU): Trips between home and the university campus for school
related purposes by people not employed by the University (i.e., students and visitors).
 Home-Based Other (HBO): All other trips that have one end at home. These can include trips
between home and appointment, home and recreation, etc.
 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 2,853 unique weekday trips reported by survey participants.
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 described
above, resulting in the total number of trips by purpose shown in Table 3.4. Trip purposes were
identified based on the origin and destination activity for each trip using the relationship shown in Table
3.5. 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.
Analysis of survey data did not include HBU trips, as the number of trips typically captured by household
surveys does not provide sufficient information to estimate HBU production or attraction rates. Instead,
HBU trips are addressed using data borrowed from university special generator studies. Analysis of HBU
trips is included in a separate section of this document.
Table 3.4: Weighted and Expanded Trips by Purpose
Trip Purpose Weighted Trips Percent of Total

HBW 2,724 20%
HBS 1,625 12%
HBO 5,243 38%
WBO 1,097 8%
OBO 3,011 22%
Total 13,700 100%
Source: North Front Range (Colorado) 2009-2010 Household Travel Survey

Table 3.5: Trip Purpose Definitions based on Reported Activity
Buy Services: Video Rentals/Dry Cleaner/Post Office/Car Service/Bank
Go Out/Hang out: Entertainment/Theater/Sports Event/Go to bar
Visit Public Place: Historical Site/Museum/Park/Library

Attend Meeting: PTA/Home Owners Assoc/Other Govt
Use Professional Services: Attorney/Accountant
Buy Goods: Groceries/Clothing/Hardware Store
Use Personal Services: Grooming/Haircut/Nails

Go to Daycare/Before or After School Care
Family Personal Business/Obligations
Trip Activity
School/Day Care/Religious Activity
Pet Care: Walk the Dog/Vet Visits
Go to Gym/Exercise/Play Sports
Attend Business Meeting/Trip
Go to Library: School Related
Rest or Relaxation/Vacation
Go to School As a Student
Coffee/Ice Cream/Snacks
Attend Funeral/Wedding
Medical/Dental Services
Go to Religious Activity
Visit Friends/Relatives
Other Work Related
Transport Someone
Social/Recreational
Drop someone off

Shopping/Errands
Appropriate Skip
Pickup someone
Not Ascertained
Return to Work
Take and wait

Misc Reasons
Get/Eat Meal
Social event
Don't Know
Go to Work
Refused
Buy Gas
Meals
Home
Work
CODE 1 10 20 30 40 50 60 70 80 97 11 12 13 14 21 22 23 24 41 42 43 61 62 63 64 65 51 52 53 54 55 71 72 73 81 82 83 -1 -7 -8 -9
Home 1 HBO HBW HBO HBO HBS HBO HBO HBO HBS HBO HBW HBW HBW HBW HBO HBO HBO HBO HBS HBS HBS HBO HBO HBO HBO HBO HBO HBO HBO HBO HBO HBO HBO HBO HBO HBS HBS HBO HBO HBO n/a
Work 10 HBW WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO n/a n/a n/a
School/Day Care/Religious Activity 20 HBO WBO WBO WBO n/a n/a n/a
Medical/Dental Services 30 HBO WBO WBO WBO n/a n/a n/a
Shopping/Errands 40 HBS WBO WBO WBO n/a n/a n/a
Social/Recreational
Family Personal Business/Obligations
Transport Someone
Meals
50
60
70
80
HBO
HBO
HBO
HBS
WBO
WBO
WBO
WBO
OBO WBO
WBO
WBO
WBO
WBO
WBO
WBO
WBO
OBO n/a
n/a
n/a
n/a
n/a
OBO
n/a
OBO
n/a
n/a
n/a
n/a
Misc Reasons 97 HBO WBO WBO WBO n/a n/a n/a
Go to Work 11 HBW WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO n/a n/a n/a
Return to Work 12 HBW WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO WBO n/a n/a n/a
Attend Business Meeting/Trip 13 HBW WBO WBO WBO WBO WBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO WBO n/a n/a n/a
Other Work Related 14 HBW WBO WBO WBO n/a n/a n/a
Go to School As a Student 21 HBO WBO WBO WBO n/a n/a n/a
Go to Religious Activity 22 HBO WBO WBO WBO n/a n/a n/a
Go to Library: School Related 23 HBO WBO WBO WBO n/a n/a n/a
Go to Daycare/Before or After School Care 24 HBO WBO WBO WBO n/a n/a n/a
Buy Goods: Groceries/Clothing/Hardware Store 41 HBS WBO WBO WBO n/a OBO n/a
Buy Services: Video Rentals/Dry Cleaner/Post Office/Car Service/Bank 42 HBS WBO WBO WBO n/a OBO n/a
Buy Gas 43 HBS WBO WBO WBO n/a n/a n/a
Use Professional Services: Attorney/Accountant 61 HBO WBO WBO WBO n/a n/a n/a
Attend Funeral/Wedding 62 HBO WBO WBO WBO n/a n/a n/a
OBO
Use Personal Services: Grooming/Haircut/Nails 63 HBO WBO WBO WBO n/a n/a n/a
Pet Care: Walk the Dog/Vet Visits
Attend Meeting: PTA/Home Owners Assoc/Other Govt
Go to Gym/Exercise/Play Sports
Rest or Relaxation/Vacation
64
65
51
52
HBO
HBO
HBO
HBO
WBO
WBO
WBO
WBO
WBO
WBO
WBO
WBO
WBO
WBO
WBO
WBO
OBO n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
Visit Friends/Relatives 53 HBO WBO WBO WBO n/a n/a n/a
Go Out/Hang out: Entertainment/Theater/Sports Event/Go to bar 54 HBO WBO WBO WBO n/a n/a n/a
Visit Public Place: Historical Site/Museum/Park/Library 55 HBO WBO WBO WBO n/a n/a n/a
Pickup someone 71 HBO WBO WBO WBO n/a n/a n/a
Take and wait 72 HBO WBO WBO WBO n/a n/a n/a
Drop someone off 73 HBO WBO WBO WBO n/a n/a n/a
Social event 81 HBO WBO WBO WBO n/a n/a n/a
Get/Eat Meal 82 HBS WBO WBO WBO n/a n/a n/a
Coffee/Ice Cream/Snacks 83 HBS WBO WBO WBO OBO n/a n/a
Appropriate Skip -1 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a
Refused -7 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a OBO n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a
Don't Know -8 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a
Not Ascertained -9 HBO WBO OBO OBO OBO OBO OBO OBO OBO OBO WBO WBO WBO WBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO OBO n/a n/a OBO n/a

Production Rates
A detailed analysis of the borrowed data was conducted in order to develop trip production rates for the
Missoula MPO Model. Past experience and analysis of survey data has shown that trip production rates
are generally sensitive to household size and to a measure of wealth such as household income.
Therefore, the production model for the updated model is sensitive to both auto household income and
household size.
Income Groups
The borrowed survey dataset places each household into one of several income groups. Although useful,
there are not sufficient records in the dataset to retain all seven groupings as income categories.
Furthermore, analysis of person trip rates (person trips per household) for each of the categories
suggests aggregation to three income groups: low, medium, and high, as shown in Table 3.6.
Table 3.6: Household Trip Production Rates by Income Category
Income
Income Category All
Group HBW HBS HBO WBO OBO
(Survey) Purposes
(Model)
$0 - $14,999 0.61 0.74 2.00 0.46 1.99 5.80
Low
$15,000 - $19,999 0.59 0.67 2.10 0.30 0.75 4.40
$20,000 - $29,999 1.96 1.53 1.56 0.44 1.23 6.72
$30,000 - $39,999 1.74 0.96 2.94 0.80 1.91 8.35
Medium $40,000 - $49,999 1.82 1.11 4.19 0.67 2.01 9.81
$50,000 - $59,999 1.92 1.23 5.07 1.02 4.30 13.54
$60,000 - $74,999 2.60 0.92 4.90 0.88 2.23 11.54
$75,000 - $99,999 2.44 1.25 5.87 1.06 3.01 13.62
$100,000 - $134,999 2.06 1.25 5.36 1.07 2.65 12.39
High
$135,000 - $149,999 2.52 1.48 3.19 1.84 2.30 11.32
$150,000 - More 2.38 1.27 4.10 1.06 1.56 10.37
Not Included Not Reported 1.88 1.58 4.10 0.96 2.40 10.92
in Analysis
Total 1.91 1.14 3.68 0.77 2.11 9.61
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 size and income. 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. As a result, some income and
household combinations with small sample sizes and similar trip rates were grouped together to
determine a group trip production rate. This trip production rate was then applied to each combination

within the group for use in the model. Grouping was performed separately for each trip purpose. The
resulting initial trip rates are shown in Tables 3.7A through 3.12A. Grouped trip rates are identified in
these tables through use of thicker borders. Production rates were increased by 12% during the model
validation process; Tables 3.7B through 3.12B show the adjusted trip rates.
Table 3.7A: Initial HBW Trip Production Rates
Household Household Size

Income 1 2 3 4 5+ Weighted Average
Low 0.36 0.65 1.31 1.31 1.31 0.60
Medium 0.83 2.30 2.00 2.25 3.03 2.01
High 0.94 2.30 2.66 2.20 2.96 2.39
Weighted
0.66 2.20 2.12 2.22 2.93 1.91
Average
Table 3.8A: Initial HBS Trip Production Rates

Low 0.43 1.12 1.22 1.22 1.22 0.70
Medium 0.54 1.41 0.76 1.50 1.50 1.16
High 0.70 1.41 1.10 1.49 1.49 1.35
Weighted
0.50 1.40 0.92 1.49 1.48 1.14
Average
Table 3.9A: Initial HBO Trip Production Rates

Low 1.22 1.66 3.59 4.85 8.75 1.97
Medium 1.22 1.78 3.85 6.60 8.70 3.49
High 1.41 1.89 4.82 7.17 14.92 5.16
Weighted
1.22 1.80 4.12 6.75 10.43 3.68
Average

Table 3.10A: Initial WBO Trip Production Rates

Low 0.27 0.52 0.60 0.60 0.60 0.38
Medium 0.46 0.57 0.96 0.96 0.96 0.71
High 0.80 1.00 1.05 1.39 1.44 1.14
Weighted
0.40 0.69 0.95 1.09 1.08 0.77
Average
Table 3.11A: Initial OBO Trip Production Rates

Low 1.26 1.26 1.83 1.83 1.83 1.38
Medium 1.13 1.35 2.41 3.64 3.64 2.09
High 1.13 1.34 2.78 4.12 4.12 2.60
Weighted
1.18 1.34 2.45 3.76 3.69 2.11
Average
Table 3.12A: Initial Trip Production Rates – All Purposes

Low 3.53 5.21 8.55 9.82 13.72 5.03
Medium 4.17 7.42 9.98 14.95 17.82 9.46
High 4.98 7.95 12.41 16.36 24.93 12.64
Weighted
3.95 7.43 10.55 15.32 19.62 9.61
Average

Table 3.13A: Initial Trip Attraction Rates
Model Employment Category HBW HBS HBO WBO OBO

Basic 1.49 0.03 0.27 0.05 0.13
Retail 1.22 7.40 2.59 1.74 5.55
Service 1.27 0.15 1.61 0.26 0.68
Education 1.23 1.04 14.78 0.89 2.98
Health 1.12 0.04 2.80 0.39 0.94
Leisure 1.05 5.44 1.28 1.98 2.82
Households 0 0.03 0.61 0.08 0.28
Average (per employee) 1.28 1.48 2.70 0.61 1.53
Source: WATS (Ann Arbor, MI) Travel Demand Model
Table 3.7B: Final HBW Trip Production Rates

Low 0.40 0.73 1.47 1.47 1.47 0.67
Medium 0.93 2.58 2.24 2.52 3.39 2.25
High 1.05 2.58 2.98 2.46 3.32 2.68
Weighted
0.74 2.46 2.37 2.49 3.28 2.14
Average
Table 3.8B: Final HBS Trip Production Rates

Low 0.48 1.25 1.37 1.37 1.37 0.78
Medium 0.60 1.58 0.85 1.68 1.68 1.30
High 0.78 1.58 1.23 1.67 1.67 1.51
Weighted
0.56 1.57 1.03 1.67 1.66 1.28
Average

Table 3.9B: Final HBO Trip Production Rates

Low 1.37 1.86 4.02 5.43 9.80 2.21
Medium 1.37 1.99 4.31 7.39 9.74 3.91
High 1.58 2.12 5.40 8.03 16.71 5.78
Weighted
1.37 2.02 4.61 7.56 11.68 4.12
Average
Table 3.10B: Final WBO Trip Production Rates

Low 0.30 0.58 0.67 0.67 0.67 0.43
Medium 0.52 0.64 1.08 1.08 1.08 0.80
High 0.90 1.12 1.18 1.56 1.61 1.28
Weighted
0.45 0.77 1.06 1.22 1.21 0.86
Average
Table 3.11B: Final OBO Trip Production Rates

Low 1.41 1.41 2.05 2.05 2.05 1.55
Medium 1.27 1.51 2.70 4.08 4.08 2.34
High 1.27 1.50 3.11 4.61 4.61 2.91
Weighted
1.32 1.50 2.74 4.21 4.13 2.36
Average
Table 3.12B: Final Trip Production Rates – All Purposes

Low 3.95 5.84 9.58 11.00 15.37 5.63
Medium 4.67 8.31 11.18 16.74 19.96 10.60
High 5.58 8.90 13.90 18.32 27.92 14.16
Weighted
4.42 8.32 11.82 17.16 21.97 10.76
Average

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 trips, the attraction end of a trip occurs at a non-residential location, or
occasionally at another person’s home. For WBO trips, trip productions occur 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 non-home-based trip
purposes, allocation models and special procedures are used to properly locate the production and
attraction end of each trip.
Attraction data from the NFR survey dataset was not yet available at the time of model development.
Furthermore, NHTS datasets considered as source data did not fully capture trip attraction-end activity.
However, this information was collected by the Michigan Department of Transportation (DOT) and the
Southeast Michigan Council of Governments (SEMCOG) in a survey conducted for the greater Detroit
area. While the data for Detroit itself is not likely to be appropriate for use in Missoula, an analysis
performed for the Washtenaw Area Council of Governments (WATS) focused on Ann Arbor, Michigan is
more appropriate. The trip rates shown in Table 3.13A, borrowed from the WATS Regional Travel Model,
have been borrowed and applied in the Missoula MPO Model. Attraction rates were increased by 12%
during model validation to arrive at the rates shown in Table 3.13B.
Table 3.13B: Final Trip Attraction Rates
Model Employment Category HBW HBS HBO WBO OBO

Basic 1.67 0.03 0.30 0.06 0.15
Retail 1.37 8.29 2.90 1.95 6.22
Service 1.42 0.17 1.80 0.29 0.76
Education 1.38 1.16 16.55 1.00 3.34
Health 1.25 0.04 3.14 0.44 1.05
Leisure 1.18 6.09 1.43 2.22 3.16
Households 0 0.03 0.68 0.09 0.31
Average (per employee) 1.43 1.66 3.02 0.68 1.71
Non-Home-Based Production Allocation

Models
While WBO and OBO trips are initially generated using household-based production rates, these trip
productions occur primarily at places of employment. 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 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 borrowed from the WATS

Regional Travel Model. WBO production allocation rates used in the Missoula MPO Model are shown in
Table 3.14. Final rates were increased by 12% during the model validation process.
Table 3.14: WBO Production Allocation Rates
WBO PA WBO PA
Model Employment Category
(Initial) (Final)
Basic 0.79 0.88
Retail 0.53 0.59
Service 0.99 1.11
Education 0.86 0.96
Health 0.73 0.82
Leisure 0.37 0.41
Households 0 0
Average (per employee) 0.79 0.88
Source: WATS (Ann Arbor, MI) Travel Demand Model
A simpler approach was taken for OBO trips. OBO production allocation rates are identical to OBO
attraction rates. This approach is possible because there is no distinct difference between OBO
productions and attractions. OBO productions and attractions all occur at non-home, non-work
locations.
University Trip Generation and

Production Allocation
Missoula is home to the University of Montana (UM). Because the university is a four-year college,
students tend to live on campus or in households concentrated near the university. This suggests that a
special university trip purpose and allocation model can improve representation of university trips in the
travel model.
University Definition
The UM campus is separated into four traffic analysis zones as shown in Figure 3.3. Based on a review of
aerial photography and the UM Campus Map1, Zones were delineated into two categories: 1) zones that
produce trips (zones that include on-campus housing) and 2) zones that attract trips (zones that contain
uses such as classes and offices). Production and attraction activity was then apportioned evenly among
each zone.
Figure 3.3: UM Campus Zones
1
http://umt.edu/home/map/campusmapCLR.pdf

University Trip Types

Because universities do not follow the normal trip patterns used by the model, some special
considerations were given to trip types at UM. 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.
Special Generator Survey Adaptation

Detailed survey data was not available for UM, so university special generator surveys from outside of
the region have been used to specify a special generator model. 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 UM.
Employment and Enrollment Data
Total employment and enrollment data for UM was retrieved from the University. 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 UM is summarized in Table 3.15.

Table 3.15: UM Employment and Enrollment
Faculty 771
Staff 1,300
Total Faculty and Staff 2,071
On-Campus Students 3,730
Off-Campus Students 11,191
Total Enrollment 14,921
Special Generator Values
Special generator values from the NFR RTM were adapted for use in the Missoula MPO Model by
computing a surrogate trip rate for each trip type based on FTE employment and student housing (e.g.,
on-campus and off-campus). Where available, the CSU special generator values were used because of
the similarities between CSU and UM. Because the CSU special generator study grouped WBO and OBO
trips into non-home-based trips, UNC values were used to compute WBO and OBO special generator
values. Trip rates and special generator values are shown in Table 3.16.
Table 3.16: University Special Generator Values
Initial Special
Trip Purpose Trip Rate Unit
Generator Value
HBW Productions 0.22 On Campus Students 821
HBW Attractions 1.6 FTE Employment 3,314
HBS Productions 0.2 On Campus Students 746
HBS Attractions n/a n/a 0
HBU Productions n/a n/a 0
HBU Attractions 3.80 Off Campus Students 42,526
HBO Productions 0.5 On Campus Students 1,865
HBO Attractions n/a n/a 0
WBO Production 0.37 FTE Employment 766
WBO Attractions 0.19 Off Campus Student 2,126
OBO Productions 0.25 Off Campus Student 2,798
OBO Attractions 0.25 Off Campus Student 2,798
Production Allocation
The production end of each HBU trip will occur at a household, most likely near the university. Student
address data was geocoded and aggregated to traffic analysis zones to assist in development of a HBU
production allocation model. The production allocation model was calibrated to match the allocation of
geocoded student address data by TAZ. The calibration parameters in Equation (1), shown below, were
adjusted iteratively until the resulting production allocation model matched the allocation of geocoded
student address data. Geocoded student address data allocated to TAZs are shown in Figure 3.4 as a dot
density map.

Figure 3.4: University Production Allocation Model Results

[Allocation Factor] = HH * a*(Db )*(edistance *c ) (1)
Where:
HH = Total households in zone
D = Right angle distance to university (mi)
a = Calibration Parameter (70)
b = Calibration Parameter (-0.9)
c = Calibration Parameter (-0.5)
External Trips
In addition to the internal-internal trips that occur entirely within the modeling area, the model must
include external travel from outside of the region. Trips with one end inside the modeling area and the
other outside of the area are called Internal-External (IE) and External-Internal (EI) trips. Through trips,
or External-External (EE) trips, are those that pass through the modeling area without stopping (or with
only short convenience stops).
External travel is modeled explicitly at the external stations where roadways cross the model boundary.
The seven (7) external stations in the MPO model are consistent with the external stations in the
regional model as shown in Figure 3.5.
External Station Volumes

The first step in estimating external travel for the model is to determine the average weekday traffic at
each location in the base year. Next, it is necessary to determine the split between the EE and IE/EI trips
at each external station. This was accomplished using guidance provided in NCHRP Report 3651 along
with a manual review of external station locations, volumes and connections to other regions. The
resulting split between EE and IE/EI trips for each external station is shown in Table 3.17. Only a few
external stations are assumed to carry a significant number of EE trips. In this model adjustment,
external stations have been numbered as zones 6434 through 6440.
1
National Cooperative Highway Research Program (NCHRP) Report 365: Travel Estimation Techniques for
Urban Planning, Transportation Research Board, 1998.

Figure 3.5: External Station Locations
Table 3.17: External Travel Assumptions
External Station Location Total Volume % EE % IE/EI EE Trips IE/EI Trips

6434 Hwy 93 S 670 6% 94% 39 631
6435 I-90 East 7,690 48% 52% 3,714 3,976
6436 I-90 West 7,110 48% 52% 3,429 3,681
6437 Hwy 93 N 7,100 7% 93% 465 6,635
6438 Hwy 200 E 1,910 6% 94% 113 1,797
6439 Hwy 83 N 1,330 0% 100% 0 1,330
6440 Hwy 12 W 750 6% 94% 43 707

Internal-External and External-Internal Trips

IE/EI trips are processed in the travel model using the internal trip purposes described previously. Trips
with a production at the external station are defined as EI trips, while trips with an attraction at the
external station are defined as IE trips. IE/EI trips are allocated among the various trip purposes and by
direction using the distributions shown in Table 3.18.
Table 3.18: IE/EI Trips by Trip Purpose and Direction
Percent By Percent of
Purpose P/A
P/A Purpose Total IE Trips
P 80% 24.0%
HBW A 20% 6.0%
Total 100% 30%
P 90% 22.5%
HBS A 10% 2.5%
Total 100% 25%
P 70% 17.5%
HBO A 30% 7.5%
Total 100% 25%
P 50% 3.0%
WBO A 50% 3.0%
Total 100% 6%
P 50% 7.0%
OBO A 50% 7.0%
Total 100% 14%
External-External Trips
A significant number of EE trips are only assumed to occur at a subset of external stations. EE trips are
further restricted to only occur between a subset of all remaining external station pairs. For example, it
would be exceedingly unlikely for trips to occur between the eastern I-90 external station and the
eastern state highway 200 external station. Therefore, trips between these external station pairs are not
modeled. Conversely, it is expected that a large number of external station trips occur between the east
and west I-90 external stations. Each pair of external stations is assigned one of the following values:
 0 = EE trips are not expected and are therefore not modeled,

 1 = EE trips are expected, or
 2 = EE trips are expected to occur more frequently than for other external station pairs.
These values are entered into an EE trip seed table, shown in Table 3.19. Over the course of a day, the
total number of EE trips at each external station is assumed to be equal for both directions (inbound
trips = outbound trips). This assumption is used to develop total inbound and outbound trips at each
external station. The seed table and total trips are used in an iterative proportional factoring process
(also called a Fratar process) to develop an EE trip table for input to the travel model, shown in Table
3.20.

Table 3.19: EE Trip Table Seed Values
6434 6435 6436 6437 6438 6440
Hwy 200 E
Hwy 12 W
Hwy 93 N
I-90 West
Hwy 93 S
I-90 East
6434 Hwy 93 S 0 0 1 1 1 1
6435 I-90 East 0 0 2 1 0 1
6436 I-90 West 1 2 0 1 1 0
6437 Hwy 93 N 1 1 1 0 0 1
6438 Hwy 200 E 1 0 1 0 0 1
6440 Hwy 12 W 1 1 0 1 1 0
Table 3.20: 24-Hour EE Trip Table
6434 6435 6436 6437 6438 6440
Hwy 200 E
TOTAL
Hwy 12 W
Hwy 93 N
I-90 West
Hwy 93 S
I-90 East
6434 Hwy 93 S 0 0 11 3 6 0 19
6435 I-90 East 0 0 1,631 206 0 20 1,857
6436 I-90 West 11 1,631 0 23 50 0 1,715
6437 Hwy 93 N 3 206 23 0 0 1 233
6438 Hwy 200 E 6 0 50 0 0 1 57
6440 Hwy 12 W 0 20 0 1 1 0 22
TOTAL 19 1,857 1,715 233 57 22 3,902

Sub-Region Trip Rate Factors

Review of traffic count data and trip balancing characteristics suggests that residents of rural Missoula
and Ravalli Counties tend to make fewer home-based trips. Instead, these residents tend to link multiple
trips together, resulting in more non-home based trips. During model validation, several regional
commercial centers were observed to be under-producing trips with. The model applies trip rate factors
to increase trip attractions at these locations. The Missoula MPO Model applies the trip rate factors
shown in Table 2.21, which were developed during the model validation process.
Table 3.21: Jurisdictional Trip Rate Factors
HBW HBS HBO WBO OBO

Sub-Region
P A P A P A P A PA P A
1 CBD 1 1 1 1 1 1 1 1 1 1 1
2 Urban MPO 1 1 1 1 1 1 1 1 1 1 1
3 Suburban MPO 1 1 1 1 1 1 1 1 1 1 1
4 Rural MPO 0.75 0.875 0.75 0.75 0.75 0.75 0.875 0.75 0.875 0.875 0.75
5 Missoula County (Non-MPO) 0.75 1 0.75 0.75 0.75 0.75 1 0.75 0.75 1 0.75
6 Ravalli County 0.875 1.125 0.625 1 0.625 0.75 1 0.625 0.625 1.125 0.625
99 Regional Commercial 1 1 1 3.75 1 3.125 1 3.75 1 3.75 3.75
Trip Balancing
Trip productions and attractions were 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. The balancing process is used to ensure conservation of trips in the model by making the
number of 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 used to estimate trip production rates, home-based trip
purposes are balanced to trip productions. One exception is the HBU trip purpose. The special generator
studies and cordon counts upon which the UM estimates are based provided a more reliable estimate
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.
Chapter 4: Trip Distribution - 66

Chapter 4: Trip Distribution

This chapter describes the process used to develop the Trip Distribution for the 2009 Missoula MPO
Model. The pathbuilding process, trip distribution process, highway skim parameters, and gravity model
parameters, are defined herein.
Trip distribution is the second phase of the traditional 4-step 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 Missoula MPO 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 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 Equation
(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. Use of K-factors was explored during
model validation, but was ultimately found to be unnecessary.
Friction factors represent the impedance to travel between each zone pair. Friction factors have been
calibrated for the HBW trip purpose based on observed trip length (time) from the 2000 CTPP data as no
other local data was available. Friction factors for other trip purposes were developed using a pivot
point analysis which is described later in this chapter.
Peak and Off-Peak Period Definitions

Trips occurring during the AM and PM peak hours 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 hour travel from production to attractions
(e.g., to work) and the majority of trips in the PM peak hour travel from attraction to productions (e.g.,
from work).
To implement trip distribution by time of day, factors representing the portion of trips occurring in the
peak (combined AM and PM peak hours) and off-peak (all other times) are necessary. Peak hour trips
are further separated in the time of day step prior to traffic assignment. The Missoula MPO Model
borrows time of day factors from the Rapid City, SD area1. These factors are based on a National
Household Travel Survey (NHTS) over-sample performed for the area. Trip distribution time of day uses
the factors shown in Table 4.1.
Table 4.1: Peak and Off-Peak Trip Percentages by Purpose

Off-Peak 0.651 0.923 0.851 0.805 0.771 0.910
Peak 0.349 0.078 0.149 0.195 0.228 0.090
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
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
1
The Rapid City area was selected because its size is similar to the Missoula area.

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 Missoula MPO Model finds shortest path between each zone pair based on peak or off-peak
congested travel time. Peak travel time is defined as the AM peak hour directional travel time, while off-
peak travel time is defined as the off-peak period congested travel time. Travel times are calculated
using a speed feedback process, which will be described further in the Traffic Assignment chapter.
Terminal Times
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 4.2, are added to both the production and attraction end
of each zone pair based on the area type of each zone.
Table 4.2: Terminal Penalties by Area Type
Area Type Terminal Time

1 CBD 1.5
2 Urban 1
3 Suburban 1
4 Rural 0.75
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 a number of nearest TAZs and multiplying that average
by a factor. Rather than using the average travel time to multiple nearby zones, intrazonal travel time
for the Missoula MPO Model has been calculated by multiplying the distance to the single nearest
neighbor by 75%.

Friction Factors
Friction factors represent the impedance to travel between each zone pair. The Missoula MPO Model
applies the friction factors in the form of gamma functions for each trip purpose. The gamma function is
defined by Equation (2).
Fij  t  et (2)
Where:
Fij = friction factor between zones i and j
T = travel time
 ,  ,  = calibration parameters
Friction factors for the HBW trip purpose were calibrated to the worker flow and reported work
commute time data obtained from the 2000 Census Transportation Planning Package. For Missoula and
Ravalli counties, worker flow data is available at the block group level for some areas and the tract area
for other areas. This data was disaggregated to the TAZ level and then used in combination with a
roadway shortest path matrix to create a trip length distribution calibration target. Due to the aggregate
nature of the data, the reported commute time and worker flow data do not necessarily match.
Therefore, the Missoula MPO Model was calibrated to fit between the two available datasets. The
calibration targets and model results for HBW trips are shown in Figure 4.1.
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. These variables were monitored during model validation.
Although local data from the CTPP can be used to estimate HBW friction factors, no other local data is
available to calibrate friction factors for the other trip purposes. Trip lengths for non-HBW trip purposes
are based on a pivot-point analysis using data from the Colorado North Front Range (NFR). This was
accomplished by using the calibrated HBW friction factors as a baseline and specifying remaining trip
purposes based on trip-length relationships observed in the NFR data. Figure 4.2 shows the friction
factors for each trip purpose. Table 4.3 contains the calibrated gamma function parameters.
The shape of the calibrated friction factors is different than is often seen in aggregate travel models.
Typically, each TAZ encompasses multiple census blocks and are bordered by collectors, arterials,
freeways, and other natural or man-made barriers. The Missoula model instead uses census blocks
directly as TAZs. To avoid the tendency of over-representing short trips in the trip distribution model,
the friction factors were adjusted to peak at a value greater than 1. This was accomplished by setting the
beta parameter to a negative value.

Figure 4.1: HBW Trip Time Distribution
55
50
45
40
Calibrated Trip Length Distribution
35 CTPP TAZ Worker Flow

CTPP Reported Commute Time
Percent of Trips
30
25
20
15
10
0
0 10 20 30 40 50 60
Travel Time (min)

Figure 4.2: Friction Factors
1500
1300
1100
Friction Factor Values
900 HBW
HBS
700 HBO
WBO
500
OBO
300
100
0 10 20 30 40 50 60
-100
Travel Time (min)
Table 4.3: Friction Factor Gamma Parameters
Trip Purpose Alpha Beta Gamma

HBW 1000 -0.301 0.096
HBS 1000 -0.776 0.205
HBO 1000 -0.776 0.205
WBO 1000 -0.550 0.200
OBO 1000 -0.900 0.300

Chapter 5: Mode Choice

The Missoula model produces and distributes all person trips including non-motorized, carpool, and
transit trips. The mode choice model separates the resulting person trip tables into the drive alone,
shared ride (i.e., carpool), transit (walk access and drive access), and non-motorized (bicycle and walk)
modes. Information about transit routes and the quality of bicycle and pedestrian facilities provides
important input to the mode choice model. The mode choice model also considers trip lengths produced
by the gravity model, resulting in sensitivity to higher density and mixed use areas. Such areas will
produce shorter trips which are more likely to be made using non-motorized modes.
Observed Mode Shares

The Missoula mode choice model has been calibrated to reproduce observed mode shares. Observed
mode share data is largely based on transit route boarding data and CTPP data indicating journey to
work mode of travel. Because the CTPP data only identifies travel modes for work trips, data from the
Colorado North Front Range (NFR) was borrowed to develop mode share targets for non-work trips. A
pivot point analysis was used to adjust borrowed data based on locally specific CTPP data.
In addition to trip data, the Missoula MPO has conducted numerous non-motorized counts at locations
throughout the City. These counts were paired with an intercept survey that identified the origin and
destination of trip segments. This information provides validation data that ensures that non-motorized
trips are modeled with reasonable lengths and in the correct locations.
Non-Motorized Shares
Non-motorized mode shares for commute trips were obtained from the 2000 Census Transportation
Planning Package (CTPP)1. The CTPP dataset indicates that 9.5% of commute trips are regularly made
using non-motorized modes. The CTPP data does not separate pedestrian and bicycle trips, however, so
it was necessary to borrow household survey data to determine an approximate split. In addition, the
CTPP dataset is limited to commute trips, which are approximated by home-based work trips in the
travel model. Therefore, borrowed data was also necessary to determine mode splits for the remaining
trip purposes.
1
Updated CTPP data based on the American Community Survey was not yet available for Missoula at the time of
model development.
Chapter 5: Mode Choice - 73

Trip shares for home-based shop (HBS), home-based other (HBO), work-based other, (WBO) and other-
based other (OBO) trips were approximated based on data borrowed from the 2009 Colorado North
Front Range (NFR) Household Travel Survey. The household travel survey did not collect sufficient
information for home-based university (HBU) trips, so non-motorized mode shares for HBU trips are
addressed separately.
Where the Missoula CTPP data indicates a 9.5% non-motorized mode share, the NFR dataset indicates
an 8.8% non-motorized data share. North Front Range non-motorized mode share data has been
adjusted to account for the slight difference in HBW non-motorized trip shares as shown in Table 5.1.
Table 5.1: Non-Motorized Mode Share Targets
Trip Purpose NFR Household Travel Survey Missoula MPO

HBW 8.8% 9.5% (CTPP)
HBS 5.5% 5.9%
HBO 13.5% 14.6%
WBO 8.7% 9.4%
OBO 8.3% 8.9%
The CTPP dataset does not distinguish between bicycle and walk trips, but it is important to have
independent targets for these two modes. The Missoula Model addresses these two non-motorized
modes independently and accounts for differences in facility quality and trip length. Two sources were
considered in the development of bike/walk mode share targets. After consideration, it was determined
that the NFR data was the most appropriate representation of the non-motorized split between walk
and bicycle trips.
An analysis of the non-motorized count data provided by the Missoula MPO suggests that of non-
motorized trips, bike and walk shares would be 47% and 53% respectively. However, it is much more
likely that a bicycle trip would be counted more than once since bicycle trips tend to cover longer
distances. It is also much more likely that a pedestrian trip would not be counted at all, as many
pedestrian trips are very short and may not cross a count station. While the non-motorized count data is
extremely useful in validation of travel model data, it is not the most suitable dataset for development
of regional mode share targets.
The NFR Regional Household Travel survey does not suffer from some of the sample bias problems
inherent in count data and is therefore more suitable for development of regional mode share targets.
The regional OD survey includes a random sampling of households and is therefore not sensitive to trip
length. Furthermore, the survey has been weighted and expanded to control for sampling bias related to
income or household size. The NFR dataset indicates an overall bicycle share of 31% and an overall walk
share of 69%. This survey also collected detailed information about trip purpose, allowing different
mode share targets to be developed for each trip purpose. Data from the NFR survey was used along
with the overall non-motorized mode share targets listed above to develop the bicycle and pedestrian
mode share targets shown in Table 5.2.

Table 5.2: Bicycle and Pedestrian Mode Share Targets
NFR Shares Missoula Non- Missoula bike and walk shares

Trip Purpose (As a % of non-motorized) Motorized (As a % of all trips)
Bike Walk Share Bike Walk
HBW 76% 24% 9.5% 7.2% 2.3%
HBS 35% 65% 5.9% 2.1% 3.8%
HBO 20% 80% 14.6% 2.9% 11.7%
WBO 31% 69% 9.4% 2.9% 6.5%
OBO 16% 84% 8.9% 1.4% 7.5%
Vehicle Trip Mode Shares

Once non-motorized trips have been identified, the remaining person trips in automobiles are separated
into drive alone, shared ride (i.e., carpool), and transit trips. The drive alone and shared ride trip targets
are also based on a combination of CTPP data and data borrowed from the NFR. As with non-motorized
shares, CTPP data for the HBW trip purpose was expanded to all trip purposes using relationships
present in the NFR household travel survey data, as shown in Table 5.3. As with non-motorized trips, the
NFR survey does not include enough HBU trips to develop HBU mode share targets.
Table 5.3: Drive alone and Shared Ride Mode Share Targets
NFR Household Travel Survey Missoula MPO

Trip Purpose
Drive Alone Shared Ride Drive Alone Shared Ride
HBW 80% 11% 78% (CTPP) 12% (CTPP)
HBS 49% 46% 47% 48%
HBO 34% 49% 33% 51%
WBO 79% 13% 76% 14%
OBO 37% 48% 36% 50%
Transit Trip Mode Shares

The CTPP dataset indicates that approximately 1% of commute trips in Missoula are made using transit.
However, the NFR survey data cannot be used to expand this data point to other trip purposes. The NFR
MPO household travel survey, as with most household travel surveys, did not capture enough transit
observations to produce a reliable estimate. Instead, transit boarding data forms the foundation for the
transit ridership targets. This data is local to Missoula and is very reliable as a source for a mode share
target. However, the transit boarding data does not indicate trip purpose.
To supplement the household travel survey, the North Front Range MPO conducted a transit on-board
survey. The on-board survey identifies the distribution of transit trips among the various trip purposes in
the travel model. This breakdown of transit trips by purpose has been applied to the overall transit
boarding totals provided by transit operators in Missoula.

Two transit operators provide fixed route service within Missoula. Mountain Line provides citywide fixed
route service throughout the city, while the Associated Students of the University of Montana (ASUM)
provides transit service targeted to university students. Each of these route operators collects boarding
data for each route. Total route boarding data, shown in Table 5.4, includes transfers. Surveys
conducted by The Mountain Line indicate that 46% of trips involve a transfer and that no trips were
observed with more than one transfer. This leads to an average of 1.46 transit boardings for each
complete transit trip. On the UM transit system, data indicating transfer frequency was not available.
While it is assumed that some riders transfer between the Mountain Line and UM transit systems, it is
assumed that transfers are minimal and that the vast majority of trips on the UM system only involve a
single boarding. These assumptions lead to the transit trip target shown in Table 5.5.
Table 5.4: Fixed Route Boardings
Route Average Weekday Boardings*

Mountain Line Route 1 547
UM East Broadway Park and Ride 233
UM College of Technology 902
UM South Park and Ride 1,434
* Average weekday boardings for Mountain Line routes include data from January through May of 2010. The data for UM
transit routes includes school days (excluding finals) in August 2010 through March 2011.
Table 5.5: Transit Trip Targets
Boardings
Transit Provider Boardings Total Trips
per Trip
Mountain Line 3,132 1.46 2,145
UM Transit 2,660 1.0 2,660
Total 5,792 1.2 4,805

Transit Access Mode
The Missoula Model represents two transit access modes: walk access and drive access. Walk access
includes all non-motorized transit access, while drive access includes transit trips that make use of a
formal park and ride. Information provided by the transit operators was combined with some simplifying
assumptions to separate transit trips into walk and drive access modes as shown in Table 5.6.
Table 5.6: Drive and Walk Access
Transit Provider Total Trips % Drive Access Walk Access Trips Drive Access Trips
Mountain Line 2,145 0% 2,145 0
UM Transit 2,660 54% 1,224 1,436
Total 4,805 30% 3,369 1,436
Most riders on the Mountain Line system access transit using non-motorized modes. The Mountain Line
system may experience a small number of informal park and ride users and drop-off access, but this type
of access is not frequent enough to warrant explicit inclusion in the Missoula Model. Therefore, the
simplifying assumption is made that all access to Mountain Line routes occurs as walk access. The walk
access mode implicitly includes bicycle access as well.
The UM transit system is largely designed to serve three park and ride facilities. Due to limited on-
campus parking, the University provides three park and rides with fixed route transit service connecting
the park and ride locations to campus. UM transit routes also serve students who live near the routes
and access transit via the walk mode. A survey of UM transit riders performed in 2008 indicates that
46% of riders bike or walk to transit, with the remainder of riders arriving via another mode. With the
assumption that transfers between the Mountain Line and UM system are minimal, this implies that
54% of UM riders arrive via the drive access mode.
Transit Trips by Purpose
Data collected by the transit providers does not include information about the specific purpose of each
transit trip. However, the on-board survey conducted by the NFR does provide a trip breakdown by
purpose. Data from this survey has been applied to locally specific trip totals to arrive at purpose-
specific transit trip targets as shown in Table 5.7. The NFR on-board survey captured a significant
number of university trips, so HBU shares are included in the analysis.
While the NFR transit service is generally similar in nature to the Missoula transit system, there are
some notable differences. The NFR transit service includes some routes specifically targeted towards K-
12 students. Students can ride these routes for free and the routes are designed specifically to serve
students. Because the Mountain Line transit system does not share this characteristic, student trips
were removed from the HBO trip totals from the NFR survey. The NFR transit service includes bus routes
targeted towards university students, but these routes do not serve park and rides. All access to these
routes is by non-motorized modes, so the HBU trip share from the NFR survey was applied to walk
access transit trips only.

Because the UM park and ride locations are intended primarily for student use, the Missoula Model uses
the simplifying assumption that all drive access transit trips are home-to-university trips made by
students and visitors (i.e., HBU trips). While some faculty and staff may also use the shuttles, this
behavior is understood to be minimal.
Table 5.7: Transit Trips by Purpose
NFR Transit Trips by Missoula Drive Access

Trip Purpose Missoula Walk Access
Purpose Trips
HBW 18.0% 606 0
HBS 4.7% 157 0
HBU 54.2% 1,827 1,436
HBO 11.4% 386 0
WBO 4.4% 150 0
OBO 7.2% 243 0
Total 100% 3,369 1,436
The above analysis results in a total of 1,827 walk access HBU trips. This is consistent with the 1,224
walk access trips expected on the UM transit system, as some HBU trips are expected to occur on
Mountain Line routes as well.
University Mode Shares

The household survey and CTPP datasets do not provide sufficient information to develop non-
motorized and vehicle mode share targets for the HBU trip purpose. Therefore, university trip shares are
instead based on data from a special generator survey conducted at Colorado State University (CSU) in
Fort Collins, Colorado. This survey included a complete cordon count paired with student/visitor and
faculty/staff surveys. This dataset, used as a basis for UM trip generation, also contains information
about university trip mode share. University mode shares from the CSU special generator survey are
shown in Table 5.8.
Table 5.8: HBU Mode Shares (Colorado State University)
Mode % Share of HBU Trips

Drive Alone 38%
Shared Ride 14%
Transit 11%
Walk 14%
Bike 22%

Resulting Mode Targets

The analysis above relies on different datasets to produce mode share targets for non-motorized trips,
trips in vehicles, and transit trips. While this information is all valid, the different data sources are not
always compatible. The transit mode shares are expressed in total trips, while the other targets are
expressed as percentages. Furthermore, the combination of CTPP and household survey data can
produce totals that do not always sum to 100%. To account for these conditions, mode share targets
shown in Table4.9 are expressed in different units for transit and non-transit trips. Non-transit mode
shares are expressed as a percentage of non-transit trips, while transit targets are expressed as a
number of total trips.
Table 5.9: Resulting Mode Targets
Walk to Drive to
Purpose Walk Bike Drive Alone Shared Ride
Transit Transit
HBW 2.3% 7.3% 78.6% 11.7% 606 0
HBS 3.8% 2.1% 46.4% 47.7% 157 0
HBU 16.2% 25.3% 42.5% 16.0% 1,827 1,436
HBO 11.8% 2.9% 33.2% 52.1% 386 0
WBO 6.6% 2.9% 76.7% 13.8% 150 0
OBO 7.9% 1.5% 37.5% 53.1% 243 0
Total n/a n/a n/a n/a 3,369 1,436
Note: Transit targets are in expressed as total trips; non-transit trips are expressed as the percent share of non-
transit trips.
Auto Occupancy
Once the nested logit mode choice model determines mode shares for each trip interchange, the result
is a set of person-trip tables for each mode. For non-motorized and transit trips, this is sufficient to
perform trip assignment. However, these trip tables must be consolidated and converted into vehicle
trips for traffic assignment. This is accomplished through use of auto occupancy factors for the shared
ride trip purpose. As with other mode share data, the only locally available data source is the CTPP
dataset. This dataset has again been supplemented with data from the NFR Household Travel Survey.
Because the CSU special generator study did not collect average auto occupancy data for shared ride
trips, auto occupancy for HBW trips has been applied to HBU trips. Average shared ride auto occupancy
for each trip purpose is shown in Table 5.10.
Table 5.10: Average Auto Occupancy
Trip Purpose NFR Household Travel Survey Missoula MPO

HBW 2.2 2.1 (CTPP)
HBS 2.6 2.5
HBU n/a 2.1
HBO 2.6 2.5
WBO 2.4 2.4
OBO 2.9 2.8

Model Structure
The Missoula MPO model has been asserted based on guidance from the Federal Transit Administration
(FTA) regarding appropriate model coefficients. The asserted approach eliminates the need for
extensive data collection and estimation efforts. While model estimation is often useful in large
metropolitan areas with extensive transit service, the asserted approach is considered more appropriate
for an area the size of Missoula.
Logit Model Background

The Missoula Model includes a logit-based mode choice model that is applied for trips within the MPO
boundary1. Mode choice is applied using a nested logit model that addresses both motorized and non-
motorized modes. Nested logit models represent the current “best-practice” for mode choice modeling.
A nested logit model is effectively a series of multinomial logit models applied sequentially from top to
bottom. The generalized formulation to compute probability in a multinomial logit model is shown in
Equation (1a). The probability of selecting mode 1 in the example multinomial logit model in Figure 5.1 is
shown equation (1b).
(1a)
∑
Where:
The probability of using mode
The utility of mode
The utility of mode
The number of available modes
(1b)
Where:
The probability of using mode
The utility of mode
1
All trip-ends with one or both ends outside of the MPO boundary are generated and distributed as vehicle trips.

Figure 5.1: Example Multinomial Logit Structure
The logit model is based on the concept of utilities (or disutilities) that describe the characteristics of
travel by each mode. The utility function can be made up of impedance variables such as travel time,
wait time, and cost as well as locational and socioeconomic variables. Each variable is multiplied by a
coefficient that describes the relative weight (positive or negative) of each variable. A mode constant
that captures mode preferences that are not measured by the other utility variables is also added to the
utility. Due to the relative nature of the mode constants, the mode constant for one mode must be set
to zero. The standard utility equation is shown in Equation (2a). An example equation for the utility of
mode 1 is shown in Equation 2(b). In this simplified example, the utility variables include terminal time,
drive time, and auto operating costs.
(2a)
Where:
The utility for mode i
Mode coefficients for variables
Values for variables for mode
The mode constant for mode
(2b)
Where:
The utility for mode
Mode coefficient for terminal time
Mode coefficient for drive time
Mode coefficient for travel cost
Trip terminal time
Trip drive time

Auto operating cost

The mode constant for mode
The Missoula model uses a mode choice structure that nests multiple multinomial choices. At the
bottom level of the nested logit structure, utility values are computed using the method described for
multinomial application. Utilities for Intermediate modes are computedas a combination of the utilities
for the nested modes (i.e., modes below the intermediate mode). Utilities for intermediate modes are
based on the natural log of the sum of exponentiated sub-mode utilities. This term, referred to as the
“logsum” variable, is computed as shown in Equation (3).
(∑ ) (3)
Where:
The logsum of intermediate mode
Utility terms for nested mode j
The number of sub-modes under mode
Once the logsum variables have been computed for all intermediate modes, mode probabilities are
calculated in a manner similar to that described for multinomial logit models. However, for nested
modes, utilities are replaced by a nesting coefficient and logsum terms as shown in Equation (4a). The
nesting coefficient has a value between 0 and 1. A nesting a value of 0 indicates that sub-modes are
identical and do not need to be included as separate modes. A nesting value of 1 indicates that sub-
modes are distinctly different and could be represented as separate non-nested modes.
As an example, the probability of selecting the motorized mode in the example nested model in Figure
5.2 can be expressed as shown in Equation (4b).
(4a)
∑
Where:
The probability of selecting intermediate mode

The nesting coefficient for intermediate mode
The nesting coefficient for mode m
The number of modes at the same level as mode
(4b)

Where:
The probability of selecting the motorized mode

The top level nesting coefficient
The logsum for the motorized mode
The logsum for the motorized mode
The logsum for the bicycle mode
In this example, the logsum variable for both non-motorized modes is very simple to calculate, as the
non-motorized modes do not include any additional nests. For example, the logsum for the walk mode
can be computed as shown in Equation (5). This computation is further simplified by the fact that
( ) .
( )
(5)
Figure 5.2: Example Nested Logit model

Missoula Model Structure

The Missoula Model uses the nested logit model structure shown in Figure 5.3 below. This structure first
separates motorized trips from walk and bicycle trips, then separates drive alone, shared ride, and
transit trips. Transit trips are further separated into drive and walk access in a third level.
Figure 5.3: Mode Choice Model Structure
Figure 5.3 Notes:

 The bottom level Drive Alone and Shared Ride “dummy” nodes are included so that bottom level
coefficients can be compared to the drive and walk access transit modes.
 The drive access to transit mode is only available for the HBU trip purpose.
This nested logit model is applied using the formulas described in the previous section. The complete
process used to compute the probability of selecting the drive alone model is demonstrated in the
example below.

Example: Computing the probability of selecting the shared ride mode
Part 1: Compute logsum variables starting at the bottom of the nesting structure
Step Formula
Compute the utility for the
shared ride mode
Compute the logsum for the
( )
shared ride mode
Compute the logsum for the
( )
motorized node
Note: Logsum calculations for shared ride and drive alone are not shown, but follow the same general pattern as
the shared ride logsum.
Part 2: Compute probabilities starting at the top of the nesting structure
Step Formula
Compute the probability of

selecting motorized travel
Compute the probability of
selecting the shared ride if
the motorized mode has
already been chosen
Compute the overall
probability of selecting the
shared ride mode
Missoula Model Parameters

Mode choice model coefficients are based on FTA New Starts guidelines that specify preferred ranges
for certain model coefficients. These ranges were developed based on extensive review of mode choice
model performance and accuracy. The general guidelines are summarized in Table 5.11.
Table 5.11: New Starts Coefficient Guidelines
Coefficient Minimum Value Maximum Value

In-Vehicle Travel Time (IVTT) -0.030 -0.020
Out of Vehicle Travel Time (OVTT) -0.090 -0.040
The coefficient for out of vehicle travel time should be between 2 and 3 times the in-vehicle travel time coefficient.

The logit mode choice model specification implies a value of time that is equal to the coefficient for in-
vehicle travel time divided by the coefficient of cost. FTA guidelines state that the value of time should
be between one-quarter and one-third the median or average income. For HBW trips, the median
income is computed separately for each income group as shown in Table 5.12. For other trip purposes,
the value of time is based on the regional median income.
Table 5.12: Median Incomes and values of time ranges
Market Segment Low value of time ($/hour) High value of time ($/hour)
Low Income ($22,689) $2.70 $3.60
Medium Income ($36,958) $4.40 $5.90
High Income ($57,273) $6.90 $9.20
Regional ($34,542) $4.20 $5.50
This Missoula MPO mode choice model uses the average allowable coefficient for in-vehicle travel time.
For work and university trips, the out of vehicle coefficient is specified to be two times the in-vehicle
coefficient. For other trips, the in-vehicle coefficient is specified to be three times the in-vehicle
coefficient. Value of time assumptions use the highest allowable value of time for work and university
trips, with the lower value of time for the remaining trip purposes. The resulting coefficients are listed in
Table 5.13.
In addition to time and cost coefficients, the mode choice model uses geographic “dummy” variables to
represent the increased likelihood of trips to the central business district or university to be transit trips.
The coefficients for these variables vary by mode and are adjusted during model calibration.
The non-motorized modes also include a transit network score. The transit network score is computed
as the distance-weighted average transit score over the course of a trip. The coefficients for transit
network scores are calibrated based on observed walk and bike data.
Table 5.13: Mode Choice Model Coefficients
Coefficient HBW HBS HBU HBO WBO OBO

IVTT -0.025 -0.025 -0.025 -0.025 -0.025 -0.025
OVTT -0.065 -0.065 -0.065 -0.065 -0.065 -0.065
Cost -0.31 -0.31 -0.31 -0.31 -0.31
Cost (Low Income) -0.47
Cost (Med. Income) -0.29
Cost (High Income) -0.19
CBD Dummy 0.009 0.009 0.009 0.009 0.009 0.009
University Dummy 0.009 0.009 0.06 0.009 0.009 0.009
Walk Score -0.025 -0.025 -0.025 -0.025 -0.025 -0.025
Drive Score -0.025 -0.025 -0.025 -0.025 -0.025 -0.025
Note: Travel time is in units of minutes and cost is in units of dollars.

In model application, the above defined coefficients are multiplied by variables obtained in the transit,
non-motorized, and roadway pathbuilding process. The results of the pathbuilding process produce a
number of variables that are not immediately consistent with the model coefficients. Variables from
pathfinding are each matched with one of the coefficients described above using the information shown
in Table 5.14. Alternative specific constants are shown in Table 5.15.
Table 5.14: Mode Choice Model Variables
Variable Name Units Applicable Modes Coefficient

Drive alone,
Origin terminal time Minutes shared ride, OVTT
drive to transit
Drive alone,
Destination terminal time Minutes OVTT
shared ride
Drive alone,
Drive time Minutes shared ride, IVTT
drive to transit
Parking cost Dollars Drive alone, shared ride Cost
Drive alone,
Vehicle operating cost Dollars shared ride, Cost
drive to transit
Access walk time Minutes Walk to transit OVTT
Walk to transit,
Transfer walk time Minutes
drive to transit
Initial wait time Walk to transit,
Minutes OVTT
(First 7.5 minutes) drive to transit
Initial wait time Walk to transit,
Minutes IVTT
(Time over 7.5 minutes) drive to transit
Walk to transit,
Transfer wait time Minutes OVTT
drive to transit
Walk to transit,
Transfer penalty time Minutes OVTT
drive to transit
Walk to transit,
Transit ride time Minutes IVTT
drive to transit
Walk to transit,
Transit fare Dollars Cost
drive to transit
CBD Dummy n/a All except drive alone CBD Dummy (Transit only)
University Dummy
University Dummy n/a All except drive alone
(Transit only)
Walk time Minutes Walk OVTT
Average Walk score n/a Walk Walk Score
Bike time Minutes Walk OVTT
Bike Walk score n/a Walk Walk Score

Table 5.15: Alternative Specific Constants
Walk to Drive to
Trip Purpose Drive Alone Shared Ride Walk Bike
Transit Transit
HBW 0 -1.9365 -2.0524 0 -0.7645 -1.6117
HBS 0 -0.1911 -2.4002 0 0.1404 -2.4558
HBU 0 -1.1329 -0.4646 -1.683 0.7683 -0.3376
HBO 0 0.1717 -2.2186 0 1.7643 -1.7818
WBO 0 -1.7247 -3.0393 0 -0.2888 -2.7265
OBO 0 0.1478 -2.8515 0 0.1863 -2.8599
Cost Variables
The transit fare, vehicle operating cost, and parking cost variables in the utility functions require
specification in the travel model. The values for each cost variable are computed as described in Table
5.16. Because these variables represent a perceived cost instead of an actual cost, some costs in the
model are lower than might be expected.
Table 5.16: Cost Variables
Variable Value Comments

This is computed at roughly one-half of the auto operating cost
Auto Operating Cost $0.25
defined by the IRS.
Monthly parking passes for CBD zones range in price from $35 to
Parking Cost
$1.00 $55, but some businesses own private parking lots. The model
(CBD zones)
assumes an average per-day cost of $1.00.
Parking Cost University parking passes are priced at $185.00. For a school
$1.00
(University Zones) year with 180 days, the approximate daily parking cost is $1.00.
Mountain Line reports average revenue of $0.41 per trip,
Mountain Line Fare excluding university students. This average revenue accounts for
$0.41
(non-HBU trips) discount passes and employer-paid passes that reduce the
perceived cost.
University students do not pay directly to ride Mountain Line
Mountain Line Fare transit. Although students do pay for Mountain Line service as
$0.00
(HBU trips) part of student fees, this cost is not incurred as a result of an
individual student choosing to ride transit.
The cost variable in the mode choice utility functions is dependent on a number of cost assumptions
specified in the model. These cost assumptions include auto operating cost, parking cost,

Mode Choice Model Validation Results

The model calibration process ensures that the mode choice model will produce transit trip and mode
share totals consistent with the targets described above. To further verify mode choice results, assigned
trips on each transit route are compared to observed transit ridership. While route-level statistics are
reported below, the validation is focused on the total boardings for all Mountain Line and UM routes.
Overall transit boarding totals are nearly identical to observed values.
Table 5.17: Fixed Route Boarding Totals
Observed Average Weekday Modeled Average Weekday

Route
Boardings Boardings
UM East Broadway Park and Ride 233 545
UM College of Technology 902 29
UM South Park and Ride 1,434 1,540
Total Mountain Line 3,131 3,607
Total UM Shuttle 2,569 2,114
Combined Total 5,700 5,723
Average Boardings per Trip 1.26 1.22

Chapter 6: Traffic Assignment

This chapter describes the traffic assignment model for the 2009 Missoula MPO Model. The time of day
process used to convert trips from production-attraction (PA) format into origin-destination (OD) format
is presented first, then the assignment algorithms, and finally the iterative speed feedback process. A
discussion of an alternative assignment process, the origin user equilibrium (OUE) method, is presented
for consideration of future implementation in the model.
In the time of day model component, the vehicle trip tables by trip purpose from the mode choice
process are converted from PA format into OD format and factored into time periods for assignment on
the roadway network. The time of day process is not considered to be 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 AM and PM peak hours and for the off-peak period.
After traffic assignment is completed, resulting travel times are fed back to trip distribution and the
model is run iteratively until speeds input to trip distribution are reasonably consistent with speeds
resulting from traffic assignment.
Time of Day
In the early days of travel demand modeling, models were typically either set up directly as peak hour
models or they were 24-hour models for which either the daily trip table or the daily volumes were post-
processed to obtain peak hour directional design year traffic volumes. With processing speeds and
electronic storage capability increasing dramatically over the years, disaggregation of the models
occurred at a fast pace with more traffic analysis zones, larger modeling areas, and more detail in the
model process. Combined with the need for time-specific traffic volumes (e.g., design hour traffic) and
congested speed detail to support the air quality conformity process, these influences made detailed
time of day modeling commonplace.
Based on the analysis of hourly traffic count data, the AM and PM peak hours were defined as shown in
Table 6.1. The peak hour definitions are consistent with the traditional morning and evening peaks
observed in many similarly sized areas. One-hour peaks are often modeled in regions that don’t
experience significant congestion outside of rather short peak periods during typical weekdays. One-
hour peaks also facilitate reporting of the common performance measure of peak hour level of service.
Chapter 6: Traffic Assignment - 90

Table 6.1: 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 in the Missoula MPO Model to convert trips from PA format to
OD 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,
so the factors represent this activity as well as the predominant movements.
Although traffic count data can be used to identify peak hours and to validate the model for peak hours,
it is not particularly useful in defining time of day PA to OD conversion factors. Traffic count data does
not include information about trip purpose or trip direction which is necessary in development of model
parameters. Therefore, time of day data in the Missoula MPO Model is borrowed from the Colorado
North Front Range MPO (NFR MPO) travel model.
In the travel model, time of day factors are applied directly to the purpose-specific vehicle trip tables
created by the mode choice model. As described in the Trip Distribution chapter, daily trip tables are
separated into peak period (combined AM and PM peak periods) and off-peak period trips during trip
distribution. 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 PA format to OD format.
Time of day factors shown in Table 6.2 identify the portion of trips by purpose and direction assigned to
each time period. These detailed factors were borrowed from the NFR MPO model. The factors in Table
6.2 are actually split up and 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
defined in the Trip Distribution chapter and are repeated in Table 6.3 for reference. The pre-assignment
time of day parameters are shown in Table 6.4.
Pre-distribution time-of-day factors (Table 6.3) are computed from the 24-hour time of day factors
shown in Table 6.2. 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 distribution time of day factor is the sum of PA and AP
factors for the AM and PM periods. The pre-distribution time of day factors shown in Table 6.3 are
applied by multiplying the factors by the vehicle trip tables from mode choice for each trip purpose.

Table 6.2: Time of Day Factors (Based on 24 hours)
HBW HBS HBR HBU HBO WBO OBO

Period
PA AP PA AP PA AP PA AP PA AP PA AP PA AP
Off-Peak 0.328 0.323 0.444 0.479 0.353 0.452 0.388 0.410 0.353 0.452 0.446 0.325 0.455 0.455
AM Peak 0.179 0.000 0.013 0.006 0.098 0.007 0.150 0.000 0.098 0.007 0.071 0.055 0.025 0.025
PM Peak 0.016 0.154 0.013 0.046 0.054 0.036 0.015 0.037 0.054 0.036 0.102 0.000 0.020 0.020
Table 6.3: Pre-Distribution Time of Day Factors

Off-Peak 0.651 0.923 0.805 0.798 0.805 0.771 0.910
Peak 0.349 0.078 0.195 0.202 0.195 0.228 0.090
Table 6.4: Pre-Assignment PA to OD Time of Day Factors

Period EE
PA AP PA AP PA AP PA AP PA AP PA AP PA AP
Off-Peak 0.504 0.496 0.481 0.519 0.439 0.561 0.486 0.514 0.439 0.561 0.578 0.422 0.500 0.500 0.840
AM Peak 0.513 0.000 0.167 0.077 0.503 0.036 0.743 0.000 0.503 0.036 0.311 0.241 0.278 0.278 0.080
PM Peak 0.046 0.441 0.167 0.590 0.277 0.185 0.074 0.183 0.277 0.185 0.447 0.000 0.222 0.222 0.080
Because they are applied to trip tables that have already been separated into peak and off-peak periods,
pre-assignment time of day factors (Table 6.4) are computed by dividing 24-hour factors (Table 6.2) by
the pre-distribution factors (Table 6.3) for each period and trip purpose. The factors for each purpose
and time period sum to 100%. They are applied to the peak and off-peak PA tables using Equation (1)
through which the PA factors are multiplied by the PA formatted trip table and the AP factors are
applied to a transposed PA formatted trip table. This converts the PA format to OD format for each time
period for trip assignment. 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)
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

An example based 100 home-based work (HBW) commute trips produced from a neighborhood would
result in the following (figures are rounded):
 35 peak period commute trips (100 trips x 0.651), further defined as:
o 18 trips from home to work in the morning peak hour (100 trips x 0.179)
o zero trips from work to home in the morning peak hour (100 trips x 0.000)
o 15 trips from work to home in the evening peak hour (100 trips x 0.154)
o 2 trips from home to work in the evening peak hour (100 trips x 0.016)
and
 65 off-peak period commute trips (100 trips x 0.349), further defined as:
o 33 trips in the off-peak hours from home to work (100 trips x 0.328)
o 32 trips in the off-peak hours from work to home (100 trips x 0.323)
Trip Assignment
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. No doubt there will be new algorithms
developed in the future, but for the purposes of the Missoula MPO Model, the selection of assignment
algorithms was based on tried and true methods. 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 and destination, considering the effects of congestion. With this
method, the total 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 the Missoula area experiences congestion which will likely worsen in the 20+ year
transportation planning horizon, 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 Missoula MPO Model uses the equilibrium traffic assignment method.
Origin User Equilibrium (OUE) Assignment Method – Promising Greater Efficiency

A new assignment technique will be available in the upcoming TransCAD version 6 called Origin User
Equilibrium (OUE) Assignment. OUE Assignment employs new features that allow it to converge much
quicker while producing results similar to the Frank-Wolfe method that has been used in equilibrium
models for years. This allows for much shorter model run times.
The unique features of OUE include computing an equilibrium solution for each origin zone in addition to
the overall link flow equilibrium for the entire network so that the results can be used to “warm start” the
assignment process. After the model is finalized with the OUE procedure, a small set of (very long and
time-consuming) base model runs will be necessary to establish the initial assignment solution for
subsequent model runs. For example, the base model runs may be 1) No-Build Model with 2040 travel
loaded on an Existing and Committed (E+C) network and 2) Build Model with 2040 travel loaded on the
2040 financially-constrained Regional Transportation Plan network.
As an added benefit, the weighted impedance paths resulting from an OUE assignment can be queried
after the assignment has been run to obtain select link results. In this manner, the links of concern do not
need to be selected prior to assignment.
There are several reference materials available on the Internet related to the OUE Assignment process and
its implementation in TransCAD. Some of these papers shown charts that demonstrate the significant
operational efficiencies of the OUE method.
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 iteration difference – 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.
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 based on a common unit of
cost or other variable.
Since tolls are not an issue in the Missoula 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 in Equation 2. This is provided for reference
only since the Missoula MPO 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 as 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 Missoula MPO Model, there is no need to convert them to common cost units.
Generalized Cost = (Distance * AutoCost) + (Time * TimeCost) (2)
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)
An example of the generalized cost equation is as follows:
Generalized Cost = (1 mile x $0.50/mile) + ((1 mile/30 mph) x $15/hour) = (1 x 0.5) + (0.333 x 15)
= $0.50 + $0.50 = $1.00

In this example, the distance and time variables results in a similar weight, meaning that they both
influence the impedance function the same amount (notwithstanding the inherent double-counting of
distance with the generalized cost function). As congestion increases, the distance component becomes
much more important as an influencing factor.
Volume-Delay Functions
A volume-delay function represents the effect of increasing traffic volume on link travel time in the
assignment process. While several volume delay functions are available for consideration, the most
commonly used function is the modified Bureau of Public Roads (BPR) function. The modified BPR
function is based on the original BPR equation shown in Equation (3).
( ( ) )
(3)
Where:
= Congested travel time
= Freeflow travel time
V = Traffic volume
C = Highway design (practical) capacity
= Coefficient alpha (0.15)
= Exponent beta (4.0)
An example of the BPR equation applied in uncongested and congested conditions is as follows:
Uncongested Condition (V/C ratio equals zero):

Congested Travel Time = (1 minute free-flow travel time) x ((1 + (0.15) x (0.00)4)) = 1 minute
Congested Condition (V/C ratio equals 1.0):

Congested Travel Time = (1 minute free-flow travel time) x ((1 + (0.15) x (1)4)) = 1.15 minutes
As the example indicates, as the road goes from uncongested to congested conditions, the travel time
increases by only 15% when using the traditional BPR coefficients.
The modified BPR equation uses the same form, but replaces design capacity with ultimate roadway
capacity. Ultimate roadway capacities for links in the Missoula MPO Model roadway network are
defined in the Roadway Network chapter. The modified function also replaces the coefficient alpha and
the exponent beta with calibrated values that vary by facility type and area type. Alpha and beta for
centroid connectors were adjusted to ensure that congestion is not represented on centroid connectors.
Resulting alpha and beta values are shown in Table 6.5. Alpha and beta values are specified based on
experience in other areas, but may be adjusted during model validation.

Table 6.5: Volume Delay Parameters Alpha and Beta

Functional Classification
Alpha (α) Beta(β) Alpha (α) Beta(β) Alpha (α) Beta(β) Alpha (α) Beta(β)
1 Freeway 0.8 5.5 0.83 5.5 0.83 5.5 0.83 5.5
2 Principal Arterial 0.4 5 0.4 6 0.4 6 0.4 6
3 Minor Arterial 0.4 5 0.4 6 0.4 6 0.4 6
4 Collector 1 5 1 5 1 5 1 5
5 Rural Highway 1 5 1 5 1 5 1 5
6 Local Street 1 5 1 5 1 5 1 5
7 Ramp 0.4 5 0.4 6 0.4 6 0.4 6
8 Centroid Connector 0 1.1 0 1.1 0 1.1 0 1.1
Note: Parameters are provided for all FT/AT combinations, even though some do not exist (e.g., CBD Freeway).
Speed Feedback
The gravity model used in the trip distribution process relies on congested zone to zone travel time
information to distribute trips. Later in the model process, the traffic assignment procedure produces
estimated congested travel speeds based on traffic flows and application of a volume-delay equation.
The speeds input to trip distribution and the speeds output are generally not consistent after the initial
model run. To rectify this inconsistency, results from traffic assignment are used to re-compute zone to
zone travel times for input to trip distribution. The model is re-run, and a comparison is then made
between the initial and updated zone to zone travel times. If the travel times are not reasonably similar,
the updated travel times are then used to re-run trip distribution and the subsequent model steps. This
process is repeated iteratively until a convergence criterion is met.
Inclusion of a speed feedback process in the travel model process can have interesting and desirable
effects on the way the travel model represents the effects of network improvements in congested
situations. Without speed feedback, overall regional travel demand remains constant regardless of the
roadway network assumptions because trip distribution patterns are not affected by changing
congestion levels. (As a side-note, vehicle travel routes are always affected by congestion in the traffic
assignment model by virtue of the volume-delay functions.)
When speed feedback is added to the process, heavy congestion results in slower speeds, leading to
shorter trip patterns throughout the region. As roadway improvements are planned, addition of capacity
to the network will initially result in faster travel speeds because of less localized congestion. The speed
feedback process recognizes the additional capacity and higher speeds and allows for longer trip lengths
across the region, which has the effect of incrementally increasing overall travel demand due to
roadway network characteristics. This is consistent with the “build it and they will come” philosophy
that suggests that new roadway capacity can induce travel where roadway access did not previously
exist and/or where conditions change from congested to uncongested conditions.

Methodology
There are various approaches to solving the speed feedback problem. Three well-documented methods
are the naïve method, constant-weight method, and method of successive averages (MSA). The naïve
method is not recommended for use as lack of information sharing between subsequent iterations leads
to an inefficient process that will often fail to converge. Furthermore, the naïve method feeds speed
data directly from traffic assignment to trip distribution; while the constant weight and MSA methods
feed volumes to trip distribution which are then used to compute updated speeds (speed feedback is
sometimes referred to as volume balancing). The Missoula Model implements speed feedback using the
MSA method.
The Method of Successive Averages
The Method of Successive Averages uses a simple average of all flows resulting from previous
assignment runs. Flows can be computed as in Equation (1), or simplified to Equation (1a).
( ) (1)
( ) (1a)
Where:
MSAFlow = Flow calculated using the MSA
n = current iteration
Flow = Flow resulting from traffic assignment
The method of successive averages is commonly used in regional travel models and is the approach
recommended by the TransCAD documentation. The method of successive averages also is supported by
built-in functions in the TransCAD software.
The method of successive averages effectively assigns a weight to the traffic volumes from each traffic
assignment iteration that is equal to the reciprocal of the iteration number. In other words, the volume
results from each previous iteration are weighted equally when computing travel times for trip
distribution. After the new MSA-weighted flows are calculated, speeds on each link in the roadway
network are re-estimated, and the remainder of the model is run to complete the iteration.
Initial Speeds and Borrowed Feedback Results
Use of the MSA feedback procedure produces results that are sensitive to the initial speeds/travel times
input to the first iteration of the trip distribution model. For this reason, it is not advisable to use the
results of a previous model run as initial congested speeds in an attempt to reduce the computation
time required to run the model with speed feedback enabled. Instead, free-flow speeds should always
be used as initial speeds when speed feedback is to be run. This is particularly important when model
results and summary statistics from two alternative model runs are to be compared.

In some cases, it is desirable to run the model to test multiple alternatives without running speed
feedback for each scenario. For these cases, it is possible to run the model once with speed feedback
enabled to establish a baseline forecast scenario (e.g., future growth on existing and committed
network) and then save the final model results with speed feedback for use in alternatives texting runs.
When this approach is taken, it is important that feedback is disabled when using the copied feedback
results. In addition, the baseline scenario should be run a second time using copied speeds as input data
and with speed feedback disabled to ensure consistency between all scenarios.
Convergence Criteria
It is important that a meaningful convergence criterion is specified when running a model with speed
feedback. It is not acceptable to simply run speed feedback for a specified number of iterations and
assume convergence. A meaningful speed feedback convergence measure ensures, either directly or
indirectly, that travel time skims input to trip distribution are reasonably similar to travel times skims
created from traffic assignment output. It also provides much better consistency between similar model
runs so that the differences can be attributable to actual performance and not due to computational
issues.
The convergence criterion used must be specified carefully to prevent unnecessary iterations of the
speed feedback process, as the convergence measure will provide diminishing benefits after a certain
point. The point at which at which the best possible convergence has been met will often vary with the
level of congestion in a network. Therefore, it is necessary to monitor speed feedback convergence
when first running a dataset that is significantly different than previously considered scenarios.
Traffic assignment convergence settings also affect speed feedback convergence. If traffic assignment
does not adequately converge, the speed feedback convergence measure may improve slowly or
inconsistently. Alternately, if traffic assignment is set to converge more thoroughly, the speed feedback
convergence measure may improve more consistently and more quickly. However, closure settings that
are too stringent can result in unreasonably long model run times.
Two common speed feedback convergence measures are documented in the following sections.
Shortest path Root Mean Square Error
Shortest Path Root Mean Square Error (% RMSE) is a common technique for speed feedback
convergence. This measure compares zone to zone travel time matrices from subsequent iterations to
the current one using Equation (3), so %RMSE provides an indication of how similar the two travel time
matrices are to one another. This approach directly satisfies the requirement that inputs to trip
distribution and outputs from traffic assignment are reasonably similar. This method also has the
advantage of measuring convergence criteria without the need to run traffic assignment for the final
iteration. This facilitates a simpler structure for the speed feedback model.

(3)
Where:
%RMSE = Percent Root Mean Square Error
() = Travel time between zones j and k for the current iteration i
( ) = Travel time between zones j and k for the previous iteration (i-1)
= Number of zone to zone pairs
Total Misplaced Flow
Another possible convergence measure is referred to as the Total Misplaced Flow (TMF). Represented in
Equation (4), this measure considers the change in traffic volumes between subsequent iterations on a
link-by-link basis. This measure indirectly satisfies the requirement that inputs to trip distribution and
outputs from traffic assignment are reasonably similar by comparing subsequent sets of assignment
results.
 Flow i  Flowi 1
TMF  Links
(4)
 Flow
Links
i
Where:
Flowi = link volume for the current feedback iteration
Flowi-1 = link volume for the previous feedback iteration
Convergence Measure Used in the Missoula Model
The Shortest Path Root Mean Square Error (%RMSE) was implemented as the convergence measure for
use in the Missoula Model due to the more direct measurement of convergence and the ability to
compute convergence prior to computation of traffic assignment.
The speed feedback convergence criterion is set at 0.1% RMSE and the iteration limit is set to 10.

Application of Speed Feedback for Alternatives Analysis

Speed feedback ensures travel time consistency within the entire modeling structure. It was conceived
as a model enhancement in the early 1990’s largely in response to environmental lawsuits, although it is
good practice and now considered a necessity. Generally, speed feedback is most sensitive to network
changes that provide a significant travel time improvement, such as a new freeway into a relatively
undeveloped area. These types of alternatives warrant running the feedback process because they can
affect regional travel patterns. Less significant improvements can also affect travel times and regional
travel patterns to various degrees and should be considered for feedback.
For any and all interim milestone and horizon years, speed feedback should be executed to closure for
the base network in each of these years. This base network could be defined as a no-build, existing plus
committed, or build network for each of these future years. In any given year, speed feedback should be
considered for any of the conditions listed below.
 Anytime a model run includes a significant change to socioeconomic and/or network

assumptions compared to the base.
 A significant new roadway alternative (i.e., new or greatly improved access) over the base case
would likely warrant speed feedback. This would be true for new or significantly better access to
areas that are undeveloped, developing, or already developed. For undeveloped areas, it is likely
that the effect is more significant in later years. Examples include new freeway interchanges,
new freeways and arterials, and in limited cases new collector roads. New tunnels and bridges
could also meet this criteria for re-running speed feedback.
 Less significant roadway improvements might warrant running speed feedback. These might
include roadway widenings or corridor improvements that imply functional class, speed, or
capacity changes. Improvements limited to a short section of roadway or an intersection
generally would not warrant running speed feedback.
 A significant change to socioeconomic assumptions over the base case. This change is more
likely to be necessary over a larger area involving significant demographic shifts, but could
conceivably be limited to one or a small number of zones with very high activity. Socioeconomic
changes should also include an update to area type assumptions.
 Significant changes to external trip assumptions.
 Significant changes to special generator assumptions.
 Any model run in which a significant change in congestion on any corridor is anticipated that
could affect regional travel times and travel patterns. This criterion is largely covered by those
above.
 Changes to model parameters, factors, coefficients, etc. – Note: These changes should only be
made in conjunction with model calibration and validation, but any tests of changes to
parameters should include running the feedback process.

Traffic Assignment Validation

Roadway volumes resulting from traffic assignment were compared against traffic count data. This
process, called traffic assignment validation, ensures that the model is reasonably representing
observed traffic patterns. Traffic count data was obtained from various sources and placed on the
roadway network. Travel model results were then compared to traffic count data using a variety of
techniques, including regional comparisons, screenline comparisons, and visual inspection of individual
link data.
Overall Activity Level

Overall vehicle trip activity was validated by comparing count data to model results on all links where
count data is available using two statistics: the Model Volume as compared to Count Volume and the
Model VMT as compared to Count VMT. These statistics were reviewed at the facility type, area type,
and regional level and are shown in Table 6.6. In addition, regional daily VMT and VHT are shown in
Table 6.7.
Table 6.6: Regional Activity Validation
Number of Model Volume / Model VMT /

Link Type Target1
Counts Count Volume Count VMT
Freeway 12 8.3% 6.3% +/- 10%
Principal Arterial 75 5.3% 1.3% +/- 10%
Minor Arterial 65 -11.9% -13.8% +/- 15%
Collector 128 -18.8% -27.8% +/- 25%
CBD 16 2.0% 2.3% n/a
Urban 191 -5.4% -7.7% n/a
Suburban 104 -5.0% 9.4% n/a
Rural 55 -2.4% -1.9% n/a
Total 355 -2.4% -3.0% +/- 5%
1
The activity level targets are set of general guidelines – not a rule or regulation. The targets apply to the model
volume to count volume ratio.

Table 6.7: VMT and VHT Totals
Link Type VMT VHT

Freeway 737,451 10,164
Principal Arterial 1,418,944 27,418
Minor Arterial 375,536 8,941
Collector 442,300 11,694
Rural Highway 143,488 3,131
Local Street 522,638 21,469
CBD 30,391 1,208
Urban 737,313 23,160
Suburban 823,431 18,745
Rural 2,204,937 45,339
Total 3,796,072 88,451
Total per Household 60 1.40
Total per Person 25 0.58
Screenlines
Another important validation test is the comparison of modeled volumes and observed traffic counts on
screenlines. Screenlines are imaginary lines that extend across a series of roadway links and form a
logical basis for evaluating regional travel movements in the model. Screenlines can also be drawn to
separate major activity areas, along highways, or natural features, or around an activity area. A map of
screenlines used in the Missoula MPO Travel Model is shown in Figure 6.1. Results of the screenline
analysis are shown in Figure 6.2, along with a recommended maximum acceptable error for screenlines.
The maximum acceptable error is based on guidance contained in the National Cooperative Highway
Research Program (NCHRP) report number 255 – Highway Traffic Data for Urbanized Area Project
Planning and Design. Specific screenline data points are included in Table 6.8.

Figure 6.1: Screenline Locations

Figure 6.2: Screenline Error Values
60%
50%
40%
Percent Error
30%
20%
10%
0%
0 50,000 100,000 150,000 200,000 250,000 300,000 350,000 400,000 450,000 500,000 550,000
Sum of Counts
NCHRP 255 Allowable Curve Counts Error

Table 6.8: Screenline Data
Screenline Count Volume Model Volume % Error

1 48,680 53,215 9%
2 123,260 121,922 1%
3 75,330 66,708 13%
4 30,220 30,457 1%
5 125,380 120,730 4%
Measures of Error
While the model should accurately represent the overall level of activity, it is also important to verify
that the model has an acceptably low level of error on individual links. It is expected that the model will
not perfectly reproduce count volumes on every link, but the level of error should be monitored. The
plot shown in Figure 6.3 demonstrates the ability of the Missoula MPO model to match individual traffic
count data points and notes the resulting R2 value. Table 6.9 lists % RMSE values and target values for
each facility type. General guidelines suggest that % RMSE should be near 40% region-wide, with values
below 30% for high volume facility types such as freeways. The % RMSE measure tends to over-
represent errors on low volume facilities, so values on collectors are not particularly meaningful. Table
6.10 shows % RMSE values by volume group.
Figure 6.3: Model Count/Volume Comparison
50,000
45,000
40,000
35,000
Model Volume
30,000
R² = 0.8874
25,000
20,000
15,000
10,000
5,000
0
0 10,000 20,000 30,000 40,000 50,000
Traffic Counts
Model Volume Y=X

Table 6.9: Model % Root Mean Square Error
Link Type Number of Counts % RMSE Validation Target

Freeway 12 20.7% 30%
Principal Arterial 75 27.4% 40%
Minor Arterial 65 35.0% 40%
Collector 128 61.4% n/a
CBD 14 43.0% n/a
Urban 135 32.3% n/a
Suburban 83 43.5% n/a
Rural 50 34.2% n/a
Total 282 36.8% 40%
Note: RMSE Statistics do not include values on local streets.
Table 6.10: % Root Mean Square Error by Volume Group
Low High Mid-Point Number of Counts % RMSE

0 5,000 2,500 207 81%
5,000 10,000 7,500 66 41%
10,000 20,000 15,000 59 33%
20,000 30,000 25,000 22 23%
30,000 40,000 35,000 10 18%
40,000 50,000 45,000 2 12%
Note: RMSE Statistics by volume groups include values on local streets.

2040 Forecast Data
Development
November 2011
Table of Contents
CONTEXT AND BACKGROUND ...................................................................................... 1
CONTROL TOTALS .................................................................................................... 1

Household Population Control Totals ................................................................................................... 1
Household Control Totals ..................................................................................................................... 2
Employment Growth Totals .................................................................................................................. 4
Urban/Rural Allocation ......................................................................................................................... 5
Employment by Type ............................................................................................................................ 5
SUBAREA (UFDA) ALLOCATION .................................................................................. 6

UFDA Dwelling Unit Allocation ............................................................................................................. 6
UFDA Employment Allocation............................................................................................................. 10
TAZ-LEVEL ALLOCATION.......................................................................................... 17
TAZ-Level Dwelling Unit Allocation in the URSA ................................................................................. 17
TAZ-Level Employment Allocation in the URSA .................................................................................. 19
Downtown Area TAZ Allocation .......................................................................................................... 20
Non-URSA Dwelling Unit and Employment Allocation ....................................................................... 22
FINAL EMPLOYMENT TOTALS BY TYPE ......................................................................... 24
FINAL HOUSEHOLD AND POPULATION FORECASTS .......................................................... 24
EXTERNAL TRAVEL FORECASTS................................................................................... 27

Ravalli County Adjustments ................................................................................................................ 27
SPECIAL GENERATOR FORECASTS ............................................................................... 28
ATTACHMENTS ...................................................................................................... 30
LIST OF TABLES
Table 1: Historical Household Population Growth in Missoula County .................................................. 2
Table 2: Household Population Growth Assumptions in Previous Forecast Datasets ............................ 2
Table 3: Historical Household Growth in Missoula County .................................................................... 3
Table 4: Dwelling Unit Growth Assumptions in Previous Forecast Datasets .......................................... 4
Table 5: Employment Summary .............................................................................................................. 4
Table 6: Employment Growth Assumptions in Previous Forecast Datasets ........................................... 4
Table 7: Household Growth – County and Urban Service Area .............................................................. 5
Table 8: Employment Growth Forecasts for Missoula County by URSA / Non-URSA............................. 5
Table 9: Initial Employment Growth forecasts by Category ................................................................... 6
Table 10: Vacancy Rates............................................................................................................................ 7
Table 11A: UFDA Dwelling Unit Allocation Results ..................................................................................... 8
Table 11B: UFDA Dwelling Unit Calculations............................................................................................... 8
Table 12: Land Use to Employment Conversion ..................................................................................... 11
Table 13A: UFDA-Level Employment Capacities and Control Totals......................................................... 13
Table 13B: UFDA-Level Employment Calculations (Residential Employment) ......................................... 13
Table 13C: UFDA-Level Employment Calculations (Non-Residential Iterations 1-2) ................................ 14
Table 13D: UFDA-Level Employment Calculations (Non-Residential Iterations 3-4) ................................ 15
Table 14: Downtown Growth Assumptions ............................................................................................ 21
Table 15: Growth Rate Assumptions for Non-URSA CDPs ...................................................................... 22
Table 16: Unadjusted Employment by Type ........................................................................................... 24
Table 17: 2040 Vacancy Rate Assumptions............................................................................................. 24
Table 18: Household Size Distributions................................................................................................... 25
Table 19: 2040 Household Bivariate Distribution ................................................................................... 26
Table 20: 2040 External Station Growth Assumptions ........................................................................... 27
Table 21: Considered University Enrollment Forecast Approaches........................................................ 28
Table 22: 2040 Special Generator Values ............................................................................................... 29
LIST OF FIGURES
Figure 1: Historical and Forecast Changes in Average Household size .................................................... 3

Figure 2: UFDA Dwelling Unit Allocation Process..................................................................................... 9
Figure 3: Average Employment per Dwelling Unit in Residential Employment Summary Areas .......... 12
Figure 4: UFDA Employment Allocation Process .................................................................................... 16
Figure 5: TAZ-Level Household Allocation Process ................................................................................ 18
Figure 6: TAZ-Level Household Allocation Process ................................................................................ 19
Figure 7: TAZs with Growth Rates Constrained to 0.1% Per Year .......................................................... 23
Figure 8: Household Size Disaggregation Curves ................................................................................... 25
Figure 9: Total Enrollments from 1895 through 2010............................................................................ 28
Attachment 1: Household Growth Map ..................................................................................................... 31

Attachment 2: Employment Growth Map .................................................................................................. 32
Table of Contents - ii

The Missoula MPO Travel Model includes a base-year (2010) socioeconomic dataset based on US Census
data and employment data obtained from a private vendor. To produce long-term traffic forecasts, the
model requires a similar forecast-year dataset that includes household and employment data in a
format consistent with the base year dataset. This document describes the process of developing the
forecast year dataset.
The primary components of the forecast dataset are household and employment totals by TAZ. To arrive
at these components, the process also considers population, dwelling units and vacancy rates, and
employment by type. The forecast process is held to region-wide and sub-regional control totals to
ensure consistency with past trends and to account for previous analysis at the sub-regional level.
Control Totals
The first step in creation of the forecast dataset was development of regional growth control totals. The
control totals are based on historical growth rates, but previous socioeconomic forecast data was also
taken into account. Control totals at the county level include population, households, and employment.
At the sub-regional level, control totals are limited to household and employment. Average household
sizes are applied later in the process to produce sub-regional and TAZ-level population totals.
Household Population Control Totals

Population control totals are based on an assumed compound annual growth rate (CAGR) that is applied
to the base year population total. A growth rate assumption for the next 30 years has been identified
based on a review of historical growth rates and has been compared to growth rates from previous
forecast efforts. Because the model is household based, the population control totals reflect the
population in households rather than the total population. The household population excludes people
living in group quarters such as university dorms. University students living in off-campus apartments
are generally included in the household population total.
From 2000 through 2010, Missoula County household population experienced a compound annual
growth rate of 1.4% per year. However, there was some concern that this 10-year period may not be
indicative of long term expectations. A longer term look at household population growth since 1950
shows a range of 10-year CAGR values from 0.4% to 2.8%. Cumulatively over the 60-year period from
1950 through 2010, Missoula County experienced a CAGR of 1.9%. In the last 30 years (since 1980),
Missoula County experienced a CAGR of 1.2%. Table 1 provides a summary of historical household
population growth trends for Missoula County. To best capture trends indicative of a 30-year forecast
and due to the large variations in household population growth rates for the 20-, 30-, and 40-year
timeframe, an average of the 20-, 30-, and 40-year growth rates has been selected. This results in a
forecast household population growth rate of 1.5%.
2040 Forecast Data Development - 1

Table 2 compares the updated 2040 household population forecasts with previous forecast datasets.
While the growth rate for 2000 through 2025 rate is quite low compared to the current forecast, the
more recent rate (2007 through 2035) is only slightly lower than the 1.5% per year rate used in the
updated 2040 forecast.
Table 1: Historical Household Population Growth in Missoula County
2010 2000 1990 1980 1970 1960 1950

Population 105,665 92,183 75,975 73,282 55,477 43,168 33,998
10-year CAGR 1.4% 2.0% 0.4% 2.8% 2.5% 2.4% n/a
Cumulative CAGR 1.9% 2.0% 2.0% 2.6% 2.5% 2.4% n/a
CAGR to 2010 n/a 1.4% 1.7% 1.2% 1.6% 1.8% 1.9%
Source: US Census Bureau
Table 2: Household Population Growth Assumptions in Previous Forecast Datasets
Total Population
Forecast Timeframe
Base Year Forecast Year CAGR
1
2000 through 2025 88,750 115,200 1.05%
1
2007 through 2035 93,864 137,900 1.38%
2
2010 through 2040 105,665 165,163 1.50%
1: Source: Missoula MPO 2025 and 2035 Transportation Plans; data represents the MPO area only
2: Data represents Missoula County as a whole.
Household Control Totals

Over the last 60 years, the average household size in Missoula County has been consistently decreasing.
This is consistent with national trends of smaller household sizes. In 1950, the US Census indicated an
average household size of 3.41 persons per household, while the 2010 Census indicates an average
household size of 2.38 persons per household. Because the travel model will generate fewer trips for
smaller households, the forecast year model dataset should reflect expected changes in average
household size. However, recent trends indicate that this decline is slowing.
As shown in Table 3, household growth rates have historically been higher than population growth rates.
The difference in growth rates results in the decreasing average household size indicated in Figure 1. As
household sizes continue to shrink, the rate of decrease appears to be slowing. Therefore, long-term
and recent trends may not be appropriate for use in the forecast dataset. Even when using only the last
10 years, the resulting growth rate of 1.8% per year would result in a 2040 average household size of
2.11 persons per household. Due to the slowing of this trend, the average household size for 2040 has
instead been set to 2.25, a modest decrease from the current 2.30 people per household. This results in
a household growth rate of 1.58%.

Table 4 compares the updated 2040 household forecasts with previous forecast datasets. The CAGR
assumption of 1.58% for households is significantly higher than dwelling unit growth assumptions in
previous forecast datasets. However, this updated number is consistent with household size trends in
the US Census data.
Figure 1: Historical and Forecast Changes in Average Household size
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
1940 1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Table 3: Historical Household Growth in Missoula County
2010 2000 1990 1980 1970 1960 1950

Households 45,926 38,439 30,782 28,019 18,012 13,505 10,400
Average HH Size 2.30 2.40 2.47 2.62 3.08 3.20 3.27
10-year CAGR 1.8% 2.2% 0.9% 4.5% 2.9% 2.6% n/a
Cumulative CAGR 2.5% 2.6% 2.7% 3.4% 2.8% 2.6% n/a
CAGR to 2010 n/a 1.8% 2.0% 1.7% 2.4% 2.5% 2.5%
Source: US Census Bureau

Table 4: Dwelling Unit Growth Assumptions in Previous Forecast Datasets
Total HH/DU
Forecast Timeframe CAGR
Base Year Forecast Year
1
2000 through 2025 37,200 48,700 1.08%
1
2005 through 2035 40,812 62,051 1.41%
2
2010 through 2040 45,926 73,406 1.58%
1: Source: Missoula MPO 2025 and 2035 Transportation Plans; data represents dwelling units within the MPO.
2: Data represents occupied households in Missoula County as a whole.
Employment Growth Totals

Development of an employment control total for 2040 is based on the assumption that the jobs to
population ratio will remain constant. While an alternate approach is to assume a constant jobs to
household ratio, the jobs to population ratio was selected for this effort due to the decreasing
household sizes over time. As shown in Table 5, the jobs to population ratio for 2010 is 0.78. Using the
population growth rates discussed above, a resulting employment total for 2040 is 129,517. Although
this analysis assumes that the jobs / population ratio will remain constant over time, it is possible to
create a modified dataset with a different jobs / housing ratio for testing and analysis purposes. If this is
done, it is important to consider effects of the trip generation balancing process as documented in the
Missoula MPO Model Documentation.
Table 5: Employment Summary
2010 2040
Population 105,665 165,163
Employment 82,860 129,517
Jobs / Pop Ratio 0.78 0.78
Table 6 compares the updated 2040 employment forecasts with previous forecast datasets. The CAGR
assumption of 1.5% for employment is higher than employment growth assumptions in previous
forecast datasets. However, this updated number maintains a consistent employment to population
ratio.
Table 6: Employment Growth Assumptions in Previous Forecast Datasets
Total Employment
Forecast Timeframe CAGR
Base Year Forecast Year
1
2000 through 2025 56,000 76,300 1.24%
1
2005 through 2035 57,078 80,360 1.23%
2
2010 through 2040 82,860 129,517 1.50%
1: Source: Missoula MPO 2025 and 2035 Transportation Plans; data represents employment within the MPO.
2: Data represents employment in Missoula County as a whole.

Urban/Rural Allocation
The urban service area (URSA) in Missoula County has historically experienced a higher rate of growth
than the remainder of the County. The CAGR for households over the past 10 years has been slightly
higher in the URSA than in the County as a whole. The countywide forecast CAGR for households of
1.58% has been prorated to produce a compound annual growth rate of 1.65% for the URSA. The
remaining household growth has been allocated to the portion of Missoula County outside of the URSA,
resulting in a CAGR for the non-URSA of 1.28%.
Table 7: Household Growth – County and Urban Service Area
2000 – 2010 Prorated 2040

2000 2010
CAGR CAGR Forecast
Missoula County 38,439 45,926 1.80% 1.58% 73,405
Urban Services Area 29,639 35,757 1.89% 1.65% 58,497
Non-URSA County 8,800 10,169 1.46% 1.28% 14,909
Source: US Census Bureau (2000 and 2010 data)
As with households, employment control totals have been developed for the URSA and for the non-
URSA county. The countywide and URSA jobs to population ratios have been held constant, while
allowing a slight decrease in the jobs to population ratio in the non-URSA county. Resulting employment
control totals are shown in Table 8.
Table 8: Employment Growth Forecasts for Missoula County by URSA / Non-URSA
2010 2010 2010 Jobs / 2040 2040 2040 Jobs

Dwelling Units
Population Employment Pop Population Employment / Pop
Missoula County 105,665 82,860 0.78 165,163 129,517 0.78
Urban Services Area 82,269 73,977 0.90 131,618 118,353 0.90
Non-URSA County 23,396 8,883 0.38 33,545 11,164 0.33
Employment by Type
The Missoula MPO Model breaks down employment into six categories. Initial forecast year control
totals assume that the distribution of employment by category will remain constant for the county as a
whole and within the URSA (allowing for a slight change in the distribution in the non-URSA County). The
resulting control totals by employment category are shown in Table 9. However, these initial control
totals are replaced later in the process with employment type distributions based on land use to
employment conversion.

Table 9: Initial Employment Growth forecasts by Category
County URSA Non-URSA County

% %
2010 % 2040 2010 % 2040 2010 2040
(2010) (2040)
Retail 12,367 15% 19,331 11,648 16% 18,635 719 8% 676 6%
Service 24,381 29% 38,109 22,592 31% 36,144 1,789 20% 1,911 18%
Basic 17,572 21% 27,466 13,454 18% 21,524 4,118 46% 5,777 53%
Educational 5,891 7% 9,208 5,216 7% 8,345 675 8% 839 8%
Healthcare 12,964 16% 20,264 12,688 17% 20,299 276 3% 276 2%
Leisure/Hospitalit
9,685 12% 15,138 8,379 11% 13,405 1,306 15% 1,685 16%
y
Total 82,860 100% 129,517 73,977 100% 118,353 8,883 100% 11,164 100%
Subarea (UFDA) Allocation

Households within the URSA are allocated to urban fringe development areas (UFDAs). Previous work
has identified dwelling unit capacities for each UFDA based on a thorough review of existing conditions
within each UFDA. The 2035 Long Range Transportation Plan (LRTP) identified dwelling unit growth
forecasts at the TAZ level, which can be aggregated to the UFDA level for analysis. Although
employment data is not specifically available at the UFDA level, detailed land use data has been
aggregated to the UFDA level to aid in the identification of UFDA-level employment control totals.
The subarea allocation process is designed to take advantage of previous work efforts that define the
expected growth characteristics in the URSA. The process assumes that areas forecast to grow fastest by
2035 will continue to grow fastest through 2040. In addition, the process ensures that UFDA-level
dwelling unit capacities are not exceeded.
UFDA Dwelling Unit Allocation

Dwelling unit control totals were developed for each of the 14 UFDAs within the urban service area. For
each UFDA, the following information was used to identify a 2040 household total:
Existing Dwelling Units: Total dwelling units based on 2010 Census data,
Entitled Lots: Households with entitlement, but that have not yet been constructed,
2035 Household Forecast: Total household growth (as compared to 2010 conditions) forecast
for the 2035 long range transportation plan, and
Available Capacity: The capacity for new dwelling units as determined by the City of Missoula
GIS department.

Because the UFDA capacities reflect dwelling units rather than occupied households, UFDA control totals
reflect dwelling units as well. A countywide vacancy rate of 8.34% was used to convert control totals
from households to dwelling units. As shown in Table 10, the vacancy rate in the non-URSA county is
considerably higher than the vacancy rate in the URSA. The significantly higher vacancy rates seen in the
non-URSA County reflect the large number of seasonal and part-time dwelling units present in rural
Missoula County. Vacancy rates for the URSA and non-URSA County were held constant, leading to a
reduction in the countywide average vacancy rate for 2040.
Table 10: Vacancy Rates
2010 Vacancy Rate 2040 Vacancy Rate

Missoula County 8.34% 8.07%
Urban Services Area (URSA) 4.98% 4.98%
Non-URSA County 18.48% 18.48%
Source: Census 2010
Forecast dwelling units were allocated to the 14 UFDAs based on the 2035 forecast dataset growth
assumptions. The process began with the 2035 forecast totals. Remaining dwelling units growth was
proportionally allocated to the UFDAs based on the previous forecast growth (i.e., new households
assumed from 2010 to the 2035 forecast dataset). In cases where the resulting dwelling unit total
exceeded an UFDA’s capacity, the excess units were allocated to UFDAS that still had remaining capacity.
This process was performed iteratively until all growth had been accounted for without any over-
capacity UFDA, producing the results shown in Table 11A. Results were verified to ensure that all
entitled lots are accounted for in the UFDA-level control totals. Detailed calculations are shown in Table
11B. The process is represented graphically in Figure 2.

Table 11A: UFDA Dwelling Unit Allocation Results
Available 2010 Dwelling Entitled Growth Growth

UFDA Capacity Units Lots through 2035 through 2040
CAP A B C G = C + G1 + G2
Wye 3,182 231 720 2,518 3,150
Grant Creek 1,276 562 0 0 0
Rattle Snake 795 2,376 167 350 438
East Missoula 1,218 1,416 332 708 886
Bonner W. Riverside 482 807 2 331 414
Central 4,031 5,773 564 2,864 3,584
N1
University 442 2,766 0 442 442
South Hills 2,885 3,962 101 884 1,106
Reserve to Russell Corridor 5,076 5,326 144 2,634 3,296
Brooks Corridor 3,734 5,778 0 2,387 2,986
East Mullan 6,521 2,345 951 2,757 3,449
Target Range/Orchard Homes 1,655 2,463 381 1,101 1,378
Miller Creek 1,737 1,927 1,550 1,503 1,737
West Mullan 2,234 1,900 829 852 1,066
Total 35,268 37,632 5,741 19,331 TOT = 23,932
Table 11B: UFDA Dwelling Unit Calculations
Iteration 1 Iteration 2
Allocation Allocated Growth Allocation Allocated Growth
UFDA N2
Weight G1 = Min(TOT * W1, Weight G2 = Min(TOT * W1,
W1 = C / ∑C CAP) W2 = G1 / ∑G1 CAP)
Wye 0.13 2,518 0.14 33
Grant Creek 0.00 0 0 0
Rattle Snake 0.02 350 0.02 5
East Missoula 0.04 708 0.04 9
Bonner W. Riverside 0.02 331 0.02 4
Central 0.15 2,864 0.16 38
University 0.02 442 - -
South Hills 0.05 884 0.05 12
Reserve to Russell Corridor 0.14 2,634 0.15 35
Brooks Corridor 0.12 2,387 0.14 31
East Mullan 0.14 2,757 0.16 36
Target Range/Orchard Homes 0.06 1,101 0.06 15
Miller Creek 0.08 1,503 - -
West Mullan 0.04 852 0.05 11
Total 1.00 23,703 1.00 1.00
Table 11 Notes
N1: The UFDA Capacity for the University UFDA was increased from 56 (UFDA Report) to 442 (2035 growth assumptions).
N2: Iteration 2 allocation weights are only computed for zones with remaining capacity.

Figure 2: UFDA Dwelling Unit Allocation Process
2035 DU
Forecasts
Proportionally allocate
Regional
additional growth (beyond
2040 DU
2035) based UFDA-level 2035
Totals
growth
Initial UFDA Totals
Limit UFDA DU
UFDA-Level DU
growth based on
Capacities
capacities
Proportionally allocate excess
growth to UFDAs with
additional capacity based on
interim UFDA totals
Capacity-Limited
UFDA Totals
UFDA sum =
No
regional
control?
Yes
2040 UFDA
Control Totals

UFDA Employment Allocation

Unlike the household dataset, neither 2035 employment forecast data or ultimate capacities were
available. Therefore, it was necessary to develop a method of allocating employment data and
identifying capacities at the TAZ and UFDA level. A GIS layer containing existing and planned land use
designations provided the data to support this process.
2040 employment capacities were developed by converting the land use layer into a set of total
employment capacities. Existing employment was then subtracted, resulting in a remaining employment
capacity for each TAZ. The TAZ-level capacities were then aggregated to the UFDA level. During this
process, employment capacities were also separated into the 6 different employment types used by the
travel model. The land use to employment conversion factors, along with employment type distributions
are shown in Table 12. The resulting UFDA-level employment capacities are shown in Table 14.
The base year travel model dataset includes employment at home-based businesses (residential
employment) as well as employment at traditional commercial uses. To provide consistency with the
base year employment dataset, additional capacity was added to each TAZ and UFDA based on dwelling
unit growth assumptions. A review of the three built-out residential areas shown in Figure 3 suggests an
average of 0.30 employees per dwelling unit in such areas. This employment is a combination of
residential employment and intermixed non-residential uses such as community commercial centers.
Additional capacity for employment was added to each TAZ based on dwelling unit capacity at the TAZ
level (TAZ-level dwelling unit capacity is discussed further in the TAZ-Level Allocation section).
Two approaches for allocation of employment growth by UFDA were explored: proportional allocation
to UFDAs based on available capacity or forecast household growth. Because non-residential growth
tends occur along with household growth, the latter method was selected. Residential employment
growth (i.e., employment at home businesses and small commercial uses intermixed with residential)
was identified first based on the UFDA-level household growth assumptions at a rate of 0.30 employees
per new dwelling unit. Remaining employment was then allocated to each UFDA. The resulting UFDA
employment growth forecasts are shown in Table 13A, with calculations shown in Tables 13B through
13D. The employment allocation process is represented graphically in Figure 4. The following manual
adjustments were made during the process:
1. Grant Creek: The household allocation process used half of the employment capacity (625) for
this UFDA instead of household growth.
2. Target Range/Orchard Homes: This UFDA already contains more employment that the land-use
based capacity, largely due to the presence of Community Medical Center. Future employment
growth for this UFDA consists of residential employment growth and medical center
employment growth. Medical center employment growth is assumed at 75 employees per
decade (225 employees over 30 years).
3. Central: The employment capacity and control total for this UFDA exclude growth in the
downtown area covered by the Missoula Greater Downtown Master Plan. Employment growth
within this area is discussed in further detail in a later section.

Table 12: Land Use to Employment Conversion
Amusement
Hospitality/
Educational
Healthcare
Employees
(KSF/Acre)
Land Use Designation
Service
Retail
/ KSF
Basic
FAR
ADP-Light Industrial and
6.85 0.05 20% 35% 35% 0% 0% 10%
Commercial
Community and Highway/Heavy
9.8 3.50 0% 40% 40% 0% 0% 20%
Commercial
Community Commercial 9.8 3.50 0% 30% 30% 0% 20% 20%
General Commercial 9.8 3.50 0% 30% 30% 0% 20% 20%
Highway/Heavy Commercial 9.8 3.50 0% 40% 40% 0% 0% 20%
Community Crossroads 12.2 3.20 5% 15% 30% 15% 30% 5%
Neighborhood Commercial 12.9 3.50 5% 50% 15% 5% 25% 0%
Special District Commercial 12.2 3.20 10% 15% 30% 10% 30% 5%
Cottage Industrial 8.8 2.00 90% 0% 10% 0% 0% 0%
Heavy Industrial 8.8 1.00 90% 0% 10% 0% 0% 0%
Light Industrial 8.8 2.00 90% 0% 10% 0% 0% 0%
Light Industrial and Commercial 6.85 0.05 20% 35% 35% 0% 0% 10%
Mixed-Use 4 dwelling units per
11.8 3.50 0% 25% 30% 15% 20% 10%
acre
Mixed-Use 16 dwelling units per
11.8 3.50 0% 25% 30% 15% 20% 10%
acre
Mixed-Use 11.8 3.50 0% 25% 30% 15% 20% 10%
Historical Mixed-Use 11.8 3.50 0% 25% 30% 15% 20% 10%
Neighborhood Center 12.9 3.50 5% 50% 15% 5% 25% 0%
Land Use Designation DU / Emp /
(Residential) Acre DU
4 dwelling units per acre 4 0.30 20% 5% 25% 20% 25% 5%
6-8 dwelling units per acre 7 0.30 20% 5% 25% 20% 25% 5%
12-16 dwelling units per acre 14 0.30 20% 5% 25% 20% 25% 5%
Note: The data above are generalized conversion factors based on analysis of ITE trip rates, large-scale development
proposals, and fee program studies. They are used to allocate regional totals to the TAZ level and may not be consistent
with specific or more detailed studies.

Figure 3: Average Employment per Dwelling Unit in Residential Employment Summary Areas
Total Employees
DU Employment /DU Retail Service Basic Educational Health Care Leisure
Area1 1,842 568 0.308 27 148 75 152 154 12
Area2 379 88 0.232 6 17 15 0 49 1
Area3 255 78 0.306 0 16 54 0 0 8
Total 2,476 734 0.296 33 181 144 152 203 21
Percent Distribution 4% 25% 20% 21% 28% 3%
Assumed Distribution
for Home Employment 5% 25% 20% 20% 25% 5%

Table 13A: UFDA-Level Employment Capacities and Control Totals
Remaining 2040 Employment Growth

2010
UFDA Total Capacity Capacity EG = EGR + EGNR1 + EGNR2 +
Employment
ECAP EGNR3 + EGNR4
Wye 5,111 540 4,571 4,571
Grant Creek 1,718 521 1,197 625
Rattle Snake 1,183 1,051 132 132
East Missoula 3,119 1,166 1,953 1,953
Bonner W. Riverside 5,565 445 5,120 925
Central 20,413 15,542 4,871 4,871
University 1,083 4,715 0 133
South Hills 3,353 1,500 1,853 1,853
Reserve to Russell Corridor 44,093 13,048 31,045 7,360
Brooks Corridor 24,064 19,226 4,838 4,838
East Mullan 36,455 4,317 32,138 7,703
Target Range/Orchard Homes 2,349 6,182 0 688
Miller Creek 2,672 1,072 1,600 1,600
West Mullan 40,876 4,652 36,224 2,381
Total 192,057 73,977 125,544 39,633
Table 13B: UFDA-Level Employment Calculations (Residential Employment)
Non-Residential
Dwelling Unit Growth Residential Employment
Employment
UFDA (Adjusted for Allocation) Growth
Capacity
GADJ EGR = G * 0.30
ECAPNR = ECAP - EGR
Wye 3,150 945 3,626
N1 N1
Grant Creek 625 0 625
Rattle Snake 438 131 1
East Missoula 886 266 1,687
Bonner W. Riverside 414 124 4,996
Central 3,584 1,075 3,796
University 442 133 0
South Hills 1,106 332 1,521
Reserve to Russell Corridor 3,296 989 30,056
Brooks Corridor 2,986 896 3,942
East Mullan 3,449 1,035 31,104
N2
Target Range/Orchard Homes 1,378 413 275
Miller Creek 1,737 521 1,079
West Mullan 1,066 320 35,905
Total 24,557 7,180 119,186

Table 13C: UFDA-Level Employment Calculations (Non-Residential Iterations 1-2)
N4
Allocation N4 Allocation Allocated Growth
Allocated Growth N3
UFDA Weight Weight EGNR2 = Min[(ETOTNR
EGNR1 = Min(ETOTNR
EW1 = (GADJ / EW2 = (GADJ / – ∑EGNR1)* EW2,
* EW1, CAP)
∑GADJ) ∑GADJ) CAP]
Wye 0.13 3,626 - 0
Grant Creek 0.03 625 - 0
Rattle Snake 0.02 1 - 0
East Missoula 0.04 1,202 0.09 456
Central 0.15 3,796 - 0
University 0.02 0 - 0
South Hills 0.05 1,499 0.11 22
Reserve to Russell Corridor 0.13 4,469 0.32 1,695
Brooks Corridor 0.12 3,942 - 0
East Mullan 0.14 4,678 0.34 1,774
Target Range/Orchard Homes 0.06 275 - 0
Miller Creek 0.07 1,079 - 0
West Mullan 0.04 1,446 0.10 548
Total 1.00 5,254 1.00 4,707

Table 13D: UFDA-Level Employment Calculations (Non-Residential Iterations 3-4)
N4
Allocated Growth N4
Allocation Allocation Allocated Growth
Weight
N3 EGNR2 = Weight
N3
UFDA EGNR2 = Min[(ETOTNR
Min[(ETOTNR –
EW3 = (GADJ / EW4 = (GADJ / – ∑EGNR1-3)* EW3,
∑EGNR1-2)* EW3,
∑GADJ) ∑GADJ) CAP]
CAP]
Wye - 0 - 0
Grant Creek - 0 - 0
Rattle Snake - 0 - 0
East Missoula 0.10 30 - 0
Central - 0 - 0
University - 0 - 0
South Hills - 0 - 0
Reserve to Russell Corridor 0.36 198 0.40 9
Brooks Corridor - 0 - 0
East Mullan 0.38 207 0.42 10
Target Range/Orchard Homes - 0 - 0
Miller Creek - 0 - 0
West Mullan 0.12 64 0.13 3
Total 1.00 523 1.00 23
Table 13 Notes
N1: Adjusted growth and non-residential capacity for use in employment allocation represents ½ of employment capacity
in Grant Creek.
N2: Capacity was increased for Target Range/Orchard Homes to account for Community Medical Center growth.
N3: Iteration 2-4 allocation weights are only computed for zones with remaining capacity.
N4: ETOTNR represents total non-residential employment growth is calculated by subtracting residential employment
growth (7,180) from total employment growth (39,633), resulting in 32,453 new non-residential employees.

Figure 4: UFDA Employment Allocation Process
Allocate Residential
Employment Growth
UFDA-Level
DU
Capacities 2040 DU
Compute Non-Residential Control
Employment Capacities and Totals
Regional Control Total
Regional
2040 DU
Totals
Proportionally allocate non-
residential employment based
on UFDA-level household
growth
Initial UFDA Totals
Limit UFDA growth

based on capacities

growth to UFDAs with
additional capacity based on
Capacity-Limited
2040 household growth totals
UFDA Totals
No UFDA sum = Yes 2040 UFDA

regional
Control Totals
control?

TAZ-Level Allocation
Total households and employment were allocated to TAZs using a method similar to the UFDA-level
allocation. Outside of the URSA, employment and households were first allocated to census designated
places (CDPs), and then allocated to TAZs. In addition, forecasts for the area covered by the Missoula
Greater Downtown Master Plan were manually adjusted for consistency with this planning document.
TAZ-Level Dwelling Unit Allocation in the URSA

As with UFDA-level allocation, dwelling unit allocation to TAZs began with 2035 forecasts. For each TAZ,
entitled lots were compared to 2035 forecasts to ensure that all entitled households were accounted
for. In cases where forecasted 2035 growth did not accommodate all currently entitled lots, the baseline
TAZ-level 2035 forecasts were increased to include entitled lots.
Next, household capacities were developed at the TAZ level based on the land use GIS layer. Allowable
household densities were converted to capacities, and then existing households were subtracted. When
summed up to the UFDA level, the resulting capacities were not fully consistent with the more detailed
UFDA-level capacity analysis. However, the TAZ-level capacities serve a purpose in guiding allocation of
household growth at the TAZ level. To enforce consistency, TAZ-level capacities were factored so that
the sum of adjusted TAZ capacities within each UFDA is equal to the UFDA capacity.
After identifying initial household numbers (2035 forecast plus entitled as needed), new households
were proportionally allocated to TAZs within each UFDA. The proportional allocation was based on the
forecast growth between 2010 and 2035, inclusive of entitled lots. Household growth was capped at
each TAZ’s capacity, with remaining households iteratively allocated to TAZs with remaining capacity.
The resulting TAZ dataset expands on growth assumptions made in development of the 2035 plan, while
ensuring that household density does not exceed allowable densities identified in the Missoula Land Use
Plan. Resulting TAZ-level growth has been provided as a model dataset and is also presented in maps
accompanying this document. A graphical representation of the process is shown in Figure 5 and
example calculations are shown in Appendix A.

Figure 5: TAZ-Level Household Allocation Process
2035 DU Add entitled lots beyond 2035 Entitled

Forecasts forecasts Lots
Baseline TAZ DU
Totals
Proportionally allocate
2040 UFDA remaining DUs to TAZs within
Control Totals each UFDA based on baseline
TAZ totals
Land Use-
Based TAZ
Capacities
Initial TAZ Totals
Adjust TAZ
Limit TAZ growth
Capacities to match
based on capacities
UFDA Total

growth to TAZs with
additional capacity based on Capacity-Limited
baseline household growth TAZ Totals
totals
No TAZ total = Yes 2040 TAZ DU

UFDA control? Data

TAZ-Level Employment Allocation in the URSA

Employment within each UFDA was allocated proportionally to TAZs in three steps. First, residential
employment was allocated to TAZs at a rate of 0.3 employees per new dwelling unit. Next, specific uses
such as downtown growth and Community Medical Center were assigned to individual TAZs. Finally,
remaining employment in each UFDA was proportionally allocated to TAZs with the most available
employment capacity. With this approach, the process did not result in over-capacity TAZs so there was
no need to iteratively allocate employment from over-capacity TAZs. Employment allocation was
performed by employment type using UFDA-level control totals distributed into employment types
based on available capacity.
Resulting TAZ-level growth has been provided as a model dataset and is also presented in maps
accompanying this document. A graphical representation of the process is shown in Figure 6 and
example calculations are shown in Appendix B.
Figure 6: TAZ-Level Household Allocation Process
2040 DU Add residential employment

Forecasts (by employment type)
Specific Uses
(Community Add specific employment
Medical, growth
Downtown)
Proportionally Allocate
Land Use-
remaining employment
Based TAZ
growth based on remaining
Capacities
TAZ Capacity
Set employment type for new

employment based on TAZ-
level capacity distribution
2040 TAZ
Employment
Data

Downtown Area TAZ Allocation

In the Missoula Downtown Area, both employment and households were allocated based on the
Missoula Greater Downtown Master Plan. The plan identifies planned and proposed residential and
commercial uses in the downtown area, most of which are expected to be completed within the 2040
timeframe. Dwelling unit assumptions are defined explicitly in the plan, while square footage was
converted to employment using the conversion rates discussed previously. Resulting employment and
household totals shown in Table 14 reflect a reduction to the number of dwelling units in the Railyard
development, which has been reduced to 30% of the 1,300 dwelling units reflected in the Missoula
Greater Downtown Master Plan.
UFDA household capacities and dwelling unit control totals are inclusive of the downtown area, as
previous analysis accounted for growth in this area. Conversely, Employment capacities and control
totals explicitly exclude growth in the downtown area. Employment growth assumptions for the
downtown area are in addition to the control total for the Central UFDA, while household growth
assumptions for the downtown area are included in the control total for the Central UFDA.

Table 14: Downtown Growth Assumptions
Development Health- Leisure Total DU

TAZ Retail Service Basic DU
Name care / Hosp. Emp. (Initial)
392 53 40 4 4 11 112 0
393 53 40 4 4 11 112 7
394 71 53 5 5 15 149 0
395 71 53 5 5 15 149 0
West Broadway 402 0 0 0 0 0 0 103 380
403 53 40 4 4 11 112 0
407 0 0 0 0 0 0 210
408 53 40 4 4 11 112 7
418 0 0 0 0 0 0 7
Government 822 0 0 0 0 0 0 5
0
Center 828 0 205 0 0 0 205 0
1st Ave
317 25 4 2 0 5 35 0 0
Commercial
Triangle 836 34 307 2 34 7 385 0
Employment 0
857 51 461 4 50 11 578 11
Anchor
Depot Square 825 153 256 25 34 42 511 62 75
Roundhouse Park 317 0 0 0 0 0 0 0 0
Railyard 317 0 0 0 0 0 0 357 1300
North Higgins 826 29 4 2 0 6 42 5 0
139 E Main St 851 123 18 15 0 79 234 0 0
Madison 788 10 250 1 28 2 290 43
Employment 0
790 15 375 1 41 3 435 25
Anchor
Triangle Housing 835 190 27 23 0 129 369 288 350
Mill Site 1128 55 40 4 4 12 115 632 540
Orange & Main
841 123 18 9 0 26 175 58 70
Retail Anchor
Orange & Main
843 118 17 8 0 25 168 24 0
Pkg. Structure
New Infill Retail
842 81 12 6 0 17 116 2 0
Sites
Caras Park
839 0 0 0 0 0 0 0 0
Expansion
Macy's Anchor & 797 0 0 0 0 0 0 15
Front Street 798 0 0 0 0 0 0 0 0
Parking Structure 851 61 9 4 0 13 88 0
1170 77 11 6 0 17 110 103
Hip Strip 1171 33 5 2 0 7 47 0 125
1154 0 0 0 0 0 0 5
796 2 22 2 0 22 49 3
Cultural District 0
794 2 22 2 0 22 49 16
Total 1,537 2,324 144 214 523 1,988 2,840

Non-URSA Dwelling Unit and Employment Allocation

Outside of the URSA, forecast growth was primarily focused on Census Designated Places (CDPs). Each
CDP in Missoula County was identified and 10-year growth rates were identified. The 10-year compound
annual growth rates were pro-rated to reflect the forecast CAGR for dwelling units and employees for
the non-URSA county as whole. The prorated rates were then applied directly to the base year dwelling
unit and employment data. After CDP growth had been allocated, the remaining non-URSA County
growth was allocated to the remaining TAZs outside of CDPs. Growth in large TAZs consisting primarily
of public lands was constrained to 0.1% per year. Table 16 lists non-URSA County and CDP growth rates,
with rural constrained zones identified in Figure 7.
Table 15: Growth Rate Assumptions for Non-URSA CDPs
Prorated Employment
CDP/Area 2000-2010 DU CAGR Prorated DU CAGR
CAGR
Non-URSA County 1.83% 1.28% 0.76%
Huson CDP 1.50% 1.05% 0.63%
Frenchtown CDP 2.11% 1.48% 0.88%
Evaro CDP 1.21% 0.85% 0.51%
Piltzville CDP 0.64% 0.45% 0.27%
Turah CDP 2.13% 1.49% 0.89%
Clinton CDP 3.58% 2.51% 1.49%
Lolo CDP 1.85% 1.30% 0.77%
Carlton CDP 1.22% 0.85% 0.51%
Condon CDP 1.62% 1.14% 0.68%
Seeley Lake CDP 2.89% 2.03% 1.21%
Non-CDP County 2.00% 1.40% 0.84%

Figure 7: TAZs with Growth Rates Constrained to 0.1% Per Year

Final Employment Totals by Type

The process of allocating employment to UFDAs and TAZs based on land use data results in a regional
distribution of employment by type that is somewhat different than the distribution of employment
computed based on current conditions. Because some variation in regional employment type
distribution is expected over time, the distribution resulting from the allocation process was not
adjusted to fit the base year distribution. Initial and final employment type distributions are shown in
Table 16.
Table 16: Unadjusted Employment by Type
Initial Values based on Final Values based on

Current Conditions Employment Allocation
Employees % Employees %
Retail 19,331 15% 21,590 17%
Service 38,109 29% 37,987 29%
Basic 27,466 21% 26,380 20%
Educational 9,208 7% 9,865 8%
Healthcare 20,264 16% 19,402 15%
Leisure/Hospitality 15,138 12% 14,293 11%
Total 129,517 100% 129,516 100%
Final Household and Population Forecasts

The above discussion focuses employment, which is used directly in the travel model, as well as dwelling
units, which must be converted to household data prior to use in the travel model. In addition, the
forecast dataset assumes that household size decreases over time. This decrease must be reflected in
the regional household size distribution as well as in TAZ-level average household size values.
Household totals by TAZ were developed by applying the 2040 vacancy rate assumptions discussed
previously. These assumptions are repeated in Table 17 for reference.
Table 17: 2040 Vacancy Rate Assumptions
2010 Vacancy Rate 2040 Vacancy Rate

Missoula County 8.34% 8.07%
Urban Services Area (URSA) 4.98% 4.98%
Non-URSA County 18.48% 18.48%
Source: Census 2010

Regional household size is input to the model in the form of a regional distribution of households by
both size and income. The model forecast assumes that real income (i.e., income adjusted for inflation)
will stay constant over time. However, the trend of decreasing household sizes is forecast to continue,
although at a slower rate. The 2040 average household size assumption of 2.25 has been used to
develop an updated distribution of households by size using the travel model’s household size
disaggregation curves, shown in Figure 8. The 2010 and 2040 household size distributions are shown in
Table 18. The resulting 2040 bivariate (size and income) distribution shown in Table 19 was computed by
proportionally factoring the 2010 distribution.
Figure 8: Household Size Disaggregation Curves
100.0%
90.0%
80.0%
70.0%
Percent of Hosueholds
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
1 1.5 2 2.5 3 3.5 4 4.5 5
Average Household Size
1 Person Trendline 2 Person Trendline 3 Person Trendline 4 Person Trendline 5 Person Trendline
1 Person Household 2 Person Household 3 Person Household 4 Person Household 5 Person Household
Table 18: Household Size Distributions
Household Size 2010 2010 (%) 2040 2040 (%)

1 Person Households 13,330 29% 22,381 30%
5+ Person Households 3,405 7% 4,993 7%
Total 45,889 100% 73,405 100%
Average Household Size 2.30 2.25

Table 19: 2040 Household Bivariate Distribution
HH SIZE 1 HH SIZE 2 HH SIZE 3 HH SIZE 4 HH SIZE 5 Total

Low Income 12,820 5,407 716 363 733 20,038
Medium Income 8,804 15,982 5,597 3,741 963 35,087
High Income 776 7,133 4,314 2,755 3,301 18,280
Total 22,400 28,522 10,627 6,859 4,997 73,405
At the TAZ level, forecast year household sizes were adjusted to reflect the lower regional average
household size by factoring base year average household sizes. In TAZs without existing households, the
average household size was set to the regional average. Average households sizes were also set to the
regional average in TAZs with fewer than 5 existing households and 10 or more forecast year
households.

External Travel Forecasts

The 2040 forecast year model run also requires growth assumptions for travel at external stations. 2040
external station forecasts are based on growth rates provided by MDT as shown in Table 20. After
application of growth rates, the split between EE and IE/EI trips was computed using the methodology
documented for the base year model. all remaining assumptions, such as the distribution of IE/EI trips
by direction and purpose, identified for the base year remain.
Table 20: 2040 External Station Growth Assumptions
Growth Adjusted
External Base Year Growth 2040 EE
Location Rate Growth % EE % IE/EI IE Trips
Station Volume* Rate Volume Trips
Timeframe Rate**
1985 to
6434 Hwy 93 S 800 -1.1% 1.0% 809 5.8% 94.2% 47 762
2010
1983 to
6435 I-90 East 5,260 1.6% 1.6% 12,815 48.9% 51.1% 6,269 6,546
2010
1980 to
6436 I-90 West 3,770 1.9% 1.9% 11,907 48.8% 51.2% 5,812 6,095
2010
1980 to
6437 Hwy 93 N 3,720 2.1% 2.1% 12,873 7.2% 92.8% 933 11,940
2010
1980 to
6438 Hwy 200 E 1,500 1.5% 1.5% 3,650 6.1% 93.9% 224 3,426
2010
1980 to 100.0
6439 Hwy 83 N 500 3.6% 2.1% 2,579 0.0% 0 2,579
2008 %
1985 to
6440 Hwy 12 W 850 -0.6% 1.0% 997 5.8% 94.2% 58 939
2010
* Base year volume is for the last year in the growth rate timeframe and may not match the adjusted base year model volume.
** Negative growth rates were adjusted to 1.0 % per year; The rate of 3.6% per year for the Hwy 83 N station was adjusted to
the next-highest rate (2.1% per year).
Ravalli County Adjustments

While the link that crosses the Missoula/Ravalli county line is not a special generator, Ravalli County is
modeled with considerably less detail than Missoula County. Growth in Ravalli County was adjusted so
that the model volume at the county line represents reasonable growth based on historical growth
rates. A timeframe from 2006 through 2009 was selected to be representative of the expected future
growth at this location, resulting in a compound annual growth rate of 0.99 % per year. To achieve this
growth rate at the county line, the overall Ravalli County growth rate was set to 0.5% per year. Data in
Ravalli County was simply increased uniformly across all zones using this compound annual growth rate.

Special Generator Forecasts

Forecast data for the University of Montana main campus was developed based on historical growth
rates at the university. As shown in Table 21, and indicated by lines in Figure 9, the university growth
rate has been increasing over time when considered as a linear function, but decreasing over time when
considered as a compound annual growth rate. Comments from university representatives indicate that
the university expects a slowdown in growth due to a number of factors, including limited space for
expansion. Based on discussions with MPO and MDT Staff, the 2040 forecast enrollment is based on the
average of the compound and linear growth extrapolations using 1969 as a base year. Rounded to the
nearest thousand students, the 2040 enrollment forecast assumes 21,000 students. All special generator
values were factored to represent the forecast enrollment growth as shown in Table 22.
Figure 9: Total Enrollments from 1895 through 2010
Source: University of Montana Data Digest, 2010-2011
Table 21: Considered University Enrollment Forecast Approaches
Linear to 2011
Enrolled CAGR 2040 Linear 2040 Average 2040
Year CAGR to 2011 (Students per
Students Forecast Forecast Forecast
year)
1895 100 4.4% 52,149 128 18,626 35,388
1935 2,000 2.7% 32,124 170 19,851 25,988
1969 8,000 1.5% 22,946 165 19,700 21,323
2011 14,921 n/a n/a n/a n/a n/a

Table 22: 2040 Special Generator Values
Trip Purpose Trip Rate Unit Base Year Trips 2040 Trips
HBW Productions 0.22 On Campus Students 821 1,155
HBW Attractions 1.6 FTE Employment 3,314 4,664
HBS Productions 0.2 On Campus Students 746 1,050
HBS Attractions n/a n/a 0 0
HBU Productions n/a n/a 0 0
HBU Attractions 3.8 Off Campus Students 42,526 59,852
HBO Productions 0.5 On Campus Students 1,865 2,625
HBO Attractions n/a n/a 0 0
WBO Production 0.37 FTE Employment 766 1,078
WBO Attractions 0.19 Off Campus Student 2,126 2,992
OBO Productions 0.25 Off Campus Student 2,798 3,938
OBO Attractions 0.25 Off Campus Student 2,798 3,938

Attachments
The pages that follow include dot-density maps representing existing and forecast household and
employment data.

Attachment 1: Household Growth Map

Attachment 2: Employment Growth Map

Appendix A: Sample TAZ Dwelling Unit Allocation
UFDA Rattlesnake Rattle Snake
Total Capacity 3,171
Available Capacity 795
Entitled Lots 167
New Units 2040 357
Column B C D E F G H I J K L M N O P
Total 2,376 167 299 331 4,247 1,990 790 356 356 586 1 1 585 357
Adjusted
Capacity 2040 New Remaining 2040 New Remaining
Entitled 2030 DU Total Capacity Available (UFDA_CAP / New 2035 Dus Units Capacity Iteration 2 Units Capacity Total 2040
UFDA TAZ 2010 DU Lots (Not Used) 2035 DU (LU Analysis) Capacity (G - C) ΣH) * H [Max(E, F)] (Iteration 1) (Iteration 1) Weight (Iteration 2) (Iteration 2) New Units
Rattle Snake 5 16 0 3 4 9 0 0 4 4 0 0.000 0 0 4
Rattle Snake 6 30 0 0 0 46 16 6 0 0 6 0.000 0 6 0
Rattle Snake 7 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 8 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 11 270 6 2 2 566 296 118 6 6 112 0.107 0 112 6
Rattle Snake 12 2 0 0 0 1 0 0 0 0 0 0.000 0 0 0
Rattle Snake 13 8 0 0 0 20 12 5 0 0 5 0.000 0 5 0
Rattle Snake 14 23 0 4 4 37 14 6 4 4 2 0.071 0 2 4
Rattle Snake 15 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 16 1 0 0 0 1 0 0 0 0 0 0.000 0 0 0
Rattle Snake 17 4 0 0 0 41 37 15 0 0 15 0.000 0 15 0
Rattle Snake 18 13 0 0 0 19 6 2 0 0 2 0.000 0 2 0
Rattle Snake 19 1 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 20 12 0 0 0 20 8 3 0 0 3 0.000 0 3 0
Rattle Snake 21 7 0 0 0 9 2 1 0 0 1 0.000 0 1 0
Rattle Snake 22 42 0 0 0 122 80 32 0 0 32 0.000 0 32 0
Rattle Snake 23 2 0 0 0 5 3 1 0 0 1 0.000 0 1 0
Rattle Snake 24 1 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 25 72 0 10 11 135 63 25 11 11 14 0.196 0 14 11
Rattle Snake 26 69 0 4 5 54 0 0 5 5 0 0.000 0 0 5
Rattle Snake 27 3 0 0 0 3 0 0 0 0 0 0.000 0 0 0
Rattle Snake 28 0 0 0 0 1 1 0 0 0 0 0.000 0 0 0
Rattle Snake 29 16 0 0 0 20 4 2 0 0 2 0.000 0 2 0
Rattle Snake 30 1 0 0 0 2 1 0 0 0 0 0.000 0 0 0
Rattle Snake 32 10 0 0 0 14 4 2 0 0 2 0.000 0 2 0
Rattle Snake 33 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 34 5 0 0 0 9 4 2 0 0 2 0.000 0 2 0
Rattle Snake 35 10 0 0 0 14 4 2 0 0 2 0.000 0 2 0
Rattle Snake 36 8 0 0 0 8 0 0 0 0 0 0.000 0 0 0
Rattle Snake 37 16 0 0 0 18 2 1 0 0 1 0.000 0 1 0
Rattle Snake 38 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 39 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 40 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 41 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 42 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 43 0 0 25 28 0 0 0 28 28 0 0.000 0 0 28
Rattle Snake 45 0 0 18 20 0 0 0 20 20 0 0.000 0 0 20
Missoula Travel Demand Model Evaluation and Enhancement

2040 Forecast Data Development - Appendix A - 1
Total 2,376 167 299 331 4,247 1,990 790 356 356 586 1 1 585 357
Adjusted
Rattle Snake 49 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 56 54 1 22 24 114 60 24 24 24 0 0.000 0 0 24
Rattle Snake 57 0 0 0 0 1 1 0 0 0 0 0.000 0 0 0
Rattle Snake 58 20 0 0 0 13 0 0 0 0 0 0.000 0 0 0
Rattle Snake 59 44 0 0 0 31 0 0 0 0 0 0.000 0 0 0
Rattle Snake 60 40 0 0 0 36 0 0 0 0 0 0.000 0 0 0
Rattle Snake 61 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 62 15 0 0 0 10 0 0 0 0 0 0.000 0 0 0
Rattle Snake 63 161 0 31 34 253 92 37 34 34 3 0.607 1 2 35
Rattle Snake 64 2 0 5 6 10 8 3 6 6 0 0.000 0 0 6
Rattle Snake 65 7 0 4 4 6 0 0 4 4 0 0.000 0 0 4
Rattle Snake 66 2 0 0 0 1 0 0 0 0 0 0.000 0 0 0
Rattle Snake 67 9 0 1 1 10 1 0 1 1 0 0.000 0 0 1
Rattle Snake 68 8 0 3 4 7 0 0 4 4 0 0.000 0 0 4
Rattle Snake 69 2 9 0 0 3 1 0 9 9 0 0.000 0 0 9
Rattle Snake 70 18 0 0 0 24 6 2 0 0 2 0.000 0 2 0
Rattle Snake 71 8 0 0 0 11 3 1 0 0 1 0.000 0 1 0
Rattle Snake 72 8 0 0 0 12 4 2 0 0 2 0.000 0 2 0
Rattle Snake 73 7 0 0 0 12 5 2 0 0 2 0.000 0 2 0
Rattle Snake 74 13 0 0 0 26 13 5 0 0 5 0.000 0 5 0
Rattle Snake 75 78 0 0 0 133 55 22 0 0 22 0.000 0 22 0
Rattle Snake 76 5 0 0 0 18 13 5 0 0 5 0.000 0 5 0
Rattle Snake 77 17 0 0 0 23 6 2 0 0 2 0.000 0 2 0
Rattle Snake 78 2 12 0 0 18 16 6 12 12 0 0.000 0 0 12
Rattle Snake 79 35 0 0 0 33 0 0 0 0 0 0.000 0 0 0
Rattle Snake 80 5 0 0 0 14 9 4 0 0 4 0.000 0 4 0
Rattle Snake 81 26 0 0 0 53 27 11 0 0 11 0.000 0 11 0
Rattle Snake 82 67 0 0 0 143 76 30 0 0 30 0.000 0 30 0
Rattle Snake 83 0 0 0 0 1 1 0 0 0 0 0.000 0 0 0
Rattle Snake 84 13 0 0 0 18 5 2 0 0 2 0.000 0 2 0
Rattle Snake 85 22 0 0 0 32 10 4 0 0 4 0.000 0 4 0
Rattle Snake 101 191 139 166 183 479 288 115 183 183 0 0.000 0 0 183
Rattle Snake 102 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 104 88 0 1 1 301 213 85 1 1 84 0.018 0 84 1
Rattle Snake 105 18 0 0 0 17 0 0 0 0 0 0.000 0 0 0
Rattle Snake 106 19 0 0 0 22 3 1 0 0 1 0.000 0 1 0
Rattle Snake 107 19 0 0 0 18 0 0 0 0 0 0.000 0 0 0
Rattle Snake 109 10 0 0 0 26 16 6 0 0 6 0.000 0 6 0
Rattle Snake 111 1 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 115 0 0 0 0 4 4 2 0 0 2 0.000 0 2 0
Rattle Snake 129 0 0 0 0 3 3 1 0 0 1 0.000 0 1 0
Rattle Snake 130 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 132 0 0 0 0 1 1 0 0 0 0 0.000 0 0 0
Rattle Snake 133 0 0 0 0 3 3 1 0 0 1 0.000 0 1 0
Rattle Snake 134 64 0 0 0 201 137 55 0 0 55 0.000 0 55 0
Rattle Snake 135 20 0 0 0 26 6 2 0 0 2 0.000 0 2 0

Total 2,376 167 299 331 4,247 1,990 790 356 356 586 1 1 585 357
Adjusted
Rattle Snake 136 22 0 0 0 28 6 2 0 0 2 0.000 0 2 0
Rattle Snake 137 5 0 0 0 14 9 4 0 0 4 0.000 0 4 0
Rattle Snake 138 35 0 0 0 77 42 17 0 0 17 0.000 0 17 0
Rattle Snake 142 6 0 0 0 10 4 2 0 0 2 0.000 0 2 0
Rattle Snake 143 9 0 0 0 17 8 3 0 0 3 0.000 0 3 0
Rattle Snake 144 46 0 0 0 77 31 12 0 0 12 0.000 0 12 0
Rattle Snake 145 13 0 0 0 20 7 3 0 0 3 0.000 0 3 0
Rattle Snake 146 24 0 0 0 21 0 0 0 0 0 0.000 0 0 0
Rattle Snake 147 6 0 0 0 13 7 3 0 0 3 0.000 0 3 0
Rattle Snake 148 15 0 0 0 21 6 2 0 0 2 0.000 0 2 0
Rattle Snake 149 11 0 0 0 21 10 4 0 0 4 0.000 0 4 0
Rattle Snake 150 33 0 0 0 46 13 5 0 0 5 0.000 0 5 0
Rattle Snake 151 16 0 0 0 20 4 2 0 0 2 0.000 0 2 0
Rattle Snake 152 72 0 0 0 22 0 0 0 0 0 0.000 0 0 0
Rattle Snake 153 7 0 0 0 10 3 1 0 0 1 0.000 0 1 0
Rattle Snake 154 18 0 0 0 17 0 0 0 0 0 0.000 0 0 0
Rattle Snake 155 7 0 0 0 10 3 1 0 0 1 0.000 0 1 0
Rattle Snake 156 10 0 0 0 10 0 0 0 0 0 0.000 0 0 0
Rattle Snake 157 6 0 0 0 9 3 1 0 0 1 0.000 0 1 0
Rattle Snake 158 15 0 0 0 20 5 2 0 0 2 0.000 0 2 0
Rattle Snake 159 15 0 0 0 20 5 2 0 0 2 0.000 0 2 0
Rattle Snake 160 6 0 0 0 14 8 3 0 0 3 0.000 0 3 0
Rattle Snake 161 17 0 0 0 25 8 3 0 0 3 0.000 0 3 0
Rattle Snake 162 3 0 0 0 11 8 3 0 0 3 0.000 0 3 0
Rattle Snake 163 6 0 0 0 13 7 3 0 0 3 0.000 0 3 0
Rattle Snake 164 4 0 0 0 6 2 1 0 0 1 0.000 0 1 0
Rattle Snake 165 5 0 0 0 6 1 0 0 0 0 0.000 0 0 0
Rattle Snake 166 6 0 0 0 13 7 3 0 0 3 0.000 0 3 0
Rattle Snake 167 6 0 0 0 7 1 0 0 0 0 0.000 0 0 0
Rattle Snake 168 5 0 0 0 10 5 2 0 0 2 0.000 0 2 0
Rattle Snake 169 10 0 0 0 10 0 0 0 0 0 0.000 0 0 0
Rattle Snake 170 4 0 0 0 10 6 2 0 0 2 0.000 0 2 0
Rattle Snake 171 7 0 0 0 10 3 1 0 0 1 0.000 0 1 0
Rattle Snake 172 13 0 0 0 20 7 3 0 0 3 0.000 0 3 0
Rattle Snake 173 7 0 0 0 9 2 1 0 0 1 0.000 0 1 0
Rattle Snake 174 9 0 0 0 8 0 0 0 0 0 0.000 0 0 0
Rattle Snake 175 7 0 0 0 10 3 1 0 0 1 0.000 0 1 0
Rattle Snake 177 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 2578 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 2645 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 2,648 15 0 0 0 17 2 1 0 0 1 0 0 1 0
Rattle Snake 2,649 3 0 0 0 5 2 1 0 0 1 0 0 1 0
Rattle Snake 2650 0 0 0 0 8 8 3 0 0 3 0 0 3 0
Rattle Snake 2651 21 0 0 0 30 9 4 0 0 4 0 0 4 0
Rattle Snake 2652 3 0 0 0 12 9 4 0 0 4 0 0 4 0
Rattle Snake 2653 2 0 0 0 4 2 1 0 0 1 0 0 1 0

Total 2,376 167 299 331 4,247 1,990 790 356 356 586 1 1 585 357
Adjusted
Rattle Snake 2654 0 0 0 0 1 1 0 0 0 0 0 0 0 0
Rattle Snake 2656 0 0 0 0 2 2 1 0 0 1 0 0 1 0
Rattle Snake 2683 10 0 0 0 20 10 4 0 0 4 0 0 4 0
Rattle Snake 2684 8 0 0 0 9 1 0 0 0 0 0 0 0 0
Rattle Snake 2685 15 0 0 0 15 0 0 0 0 0 0 0 0 0
Rattle Snake 2686 14 0 0 0 12 0 0 0 0 0 0.000 0 0 0
Rattle Snake 2687 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 2688 7 0 0 0 48 41 16 0 0 16 0.000 0 16 0
Rattle Snake 2690 0 0 0 0 3 3 1 0 0 1 0.000 0 1 0
Rattle Snake 2691 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 2692 2 0 0 0 4 2 1 0 0 1 0.000 0 1 0
Rattle Snake 2693 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 2694 0 0 0 0 0 0 0 0 0 0 0.000 0 0 0
Rattle Snake 2695 0 0 0 0 1 1 0 0 0 0 0.000 0 0 0
Rattle Snake 2697 2 0 0 0 2 0 0 0 0 0 0.000 0 0 0
Rattle Snake 2698 0 0 0 0 4 4 2 0 0 2 0.000 0 2 0
Rattle Snake 2700 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Rattle Snake 2713 8 0 0 0 24 16 6 0 0 6 0 0 6 0

Appendx B: Sample TAZ Employment Allocation
UFDA East Missoula
TotalCapacity+HomeEmp 3,119
New HomeEmp 266
Total 2010 Employment 1,166 UFDA Totals by Type Retail Service Basic Educational Healthcare Leisure/Hospitality
Total Remaining Employment Cap. 1,687 465 525 58 58 291 292 Computed by applying base year employment distribution to residential
New Remainder 2040 Emp 1,687 employment and allocating UFDA Totals by Type based on capacity by type
Column B C D E F G H I J K L M N O P Q R S T U
Total 1,416 1,166 227 280 415 14 159 71 3,119 855 966 106 106 535 538 332 100 2,072 1,687 469 550 77 77 316 297
2010 2010 2010 2010 Total Emp Capacity Retail Service Basic 2040 2040 New Res. 2040 New
2010 Tot Retail Service Basic Educ. 2010 Health 2010 Leis. (LU Analysis and Emp. Emp. Emp. Educ. Emp. Health Emp. Leis. Emp. New DU Emp Available Non-Res. New 2040 New 2040 New 2040 New 2040 New 2040 New 2040
UFDA TAZ DU Emp. Emp. Emp. Emp. Emp. Emp. Emp. HH growth) Capacity Capacity Capacity Capacity Capacity Capacity Units (R * 0.30) Capacity (K - D) Employment Ret Emp Service Emp Basic Emp Educ. Emp Health Emp Leis. Emp
East Missoula 44 254 761 185 167 262 0 99 48 1,195 358 358 0 0 239 239 (0) (0) 434 353 194 194 0 0 130 130
East Missoula 46 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 50 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 51 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 53 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 54 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 55 1 0 0 0 0 0 0 0 31 9 9 0 0 6 6 - - 31 25 5 5 0 0 3 3
East Missoula 858 144 33 5 17 11 0 0 0 169 19 45 25 25 40 14 16 5 131 107 11 26 15 15 23 8
East Missoula 859 16 12 0 8 4 0 0 0 8 0 2 2 2 2 0 (0) (0) 0 0 0 1 1 1 1 0
East Missoula 860 20 8 0 1 1 0 6 0 135 41 41 0 0 27 27 - - 127 103 22 22 0 0 15 15
East Missoula 861 8 4 0 2 0 0 2 0 5 0 1 1 1 1 0 - - 1 1 0 1 1 1 1 0
East Missoula 862 2 0 0 0 0 0 0 0 3 0 1 1 1 1 0 9 3 0 0 0 1 1 1 1 0
East Missoula 863 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 864 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 865 18 11 0 2 9 0 0 0 8 0 2 2 2 2 0 (0) (0) 0 0 0 1 1 1 1 0
East Missoula 866 18 4 1 1 2 0 0 0 10 1 3 2 2 3 1 - - 6 5 1 2 1 1 2 1
East Missoula 867 33 3 0 0 3 0 0 0 16 1 4 3 3 4 1 (0) (0) 13 11 1 2 2 2 2 1
East Missoula 868 22 2 0 0 0 2 0 0 8 0 2 2 2 2 0 (0) (0) 6 5 0 1 1 1 1 0
East Missoula 869 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 870 25 5 0 4 1 0 0 0 9 0 2 2 2 2 0 (0) (0) 4 3 0 1 1 1 1 0
East Missoula 871 24 3 0 2 1 0 0 0 8 0 2 2 2 2 0 (0) (0) 5 4 0 1 1 1 1 0
East Missoula 872 23 27 0 0 24 0 3 0 8 0 2 2 2 2 0 (0) (0) 0 0 0 1 1 1 1 0
East Missoula 873 21 1 0 0 1 0 0 0 8 0 2 2 2 2 0 (0) (0) 7 6 0 1 1 1 1 0
East Missoula 874 17 1 0 1 0 0 0 0 8 0 2 2 2 2 0 (0) (0) 7 6 0 1 1 1 1 0
East Missoula 875 0 14 0 0 0 12 2 0 8 0 2 2 2 2 0 - - 0 0 0 1 1 1 1 0
East Missoula 876 18 0 0 0 0 0 0 0 8 0 2 2 2 2 0 (0) (0) 8 6 0 1 1 1 1 0
East Missoula 877 7 2 0 0 2 0 0 0 4 0 1 1 1 1 0 - - 2 1 0 1 1 1 1 0
East Missoula 878 24 0 0 0 0 0 0 0 4 0 1 1 1 1 0 (0) (0) 4 3 0 1 1 1 1 0
East Missoula 879 19 2 0 1 1 0 0 0 6 0 2 1 1 2 0 (0) (0) 4 4 0 1 1 1 1 0
East Missoula 880 0 0 0 0 0 0 0 0 44 12 13 1 1 9 8 (0) (0) 44 36 7 7 1 1 5 4
East Missoula 881 22 23 2 7 0 0 14 0 7 0 2 1 1 2 0 (0) (0) 0 0 0 1 1 1 1 0
East Missoula 882 5 25 25 0 0 0 0 0 141 42 42 0 0 28 28 - - 116 94 23 23 0 0 15 15
East Missoula 883 31 2 0 0 2 0 0 0 8 0 2 2 2 2 0 (0) (0) 6 5 0 1 1 1 1 0
East Missoula 884 5 21 2 4 15 0 0 0 94 28 28 0 0 19 19 - - 73 60 15 15 0 0 10 10
East Missoula 885 0 0 0 0 0 0 0 0 26 8 8 0 0 5 5 - - 26 21 4 4 0 0 3 3
East Missoula 886 0 0 0 0 0 0 0 0 12 4 4 0 0 2 2 - - 12 10 2 2 0 0 1 1
East Missoula 887 16 1 0 1 0 0 0 0 78 22 23 1 1 16 15 (0) (0) 77 63 12 12 1 1 9 8
East Missoula 891 8 3 2 1 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 892 20 14 1 3 10 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 893 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 895 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 896 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 897 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 898 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 899 3 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 900 30 5 0 0 4 0 0 1 48 2 12 10 10 12 2 (0) (0) 43 35 1 7 5 5 7 1
East Missoula 901 26 0 0 0 0 0 0 0 20 1 5 4 4 5 1 (0) (0) 20 17 1 3 2 2 3 1
East Missoula 902 5 0 0 0 0 0 0 0 4 0 1 1 1 1 0 - - 4 3 0 1 1 1 1 0
East Missoula 909 99 53 3 20 22 0 0 8 759 267 275 8 8 55 146 (0) (0) 706 575 145 149 4 4 30 79
East Missoula 910 26 1 0 1 0 0 0 0 7 0 2 1 1 2 0 2 1 5 4 0 1 1 1 1 0
East Missoula 911 19 0 0 0 0 0 0 0 4 0 1 1 1 1 0 - - 4 3 0 1 1 1 1 0
East Missoula 912 18 0 0 0 0 0 0 0 4 0 1 1 1 1 0 - - 4 4 0 1 1 1 1 0
East Missoula 913 3 7 0 3 4 0 0 0 1 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 914 3 1 0 1 0 0 0 0 5 0 1 1 1 1 0 (0) (0) 4 4 0 1 1 1 1 0
East Missoula 915 0 0 0 0 0 0 0 0 4 1 1 0 0 0 1 - - 4 3 1 1 0 0 0 1

2040 Forecast Data Development - Appendix B - 1
Total 1,416 1,166 227 280 415 14 159 71 3,119 855 966 106 106 535 538 332 100 2,072 1,687 469 550 77 77 316 297
East Missoula 916 10 12 0 0 1 0 0 11 72 29 29 0 0 0 14 - - 60 49 16 16 0 0 0 8
East Missoula 917 64 42 0 10 30 0 1 1 72 4 18 14 14 18 4 13 4 26 22 2 11 8 8 11 2
East Missoula 918 18 0 0 0 0 0 0 0 4 0 1 1 1 1 0 (0) (0) 4 3 0 1 1 1 1 0
East Missoula 919 6 0 0 0 0 0 0 0 7 0 2 1 1 2 0 (0) (0) 7 6 0 1 1 1 1 0
East Missoula 920 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 921 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 922 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 4 1 0 0 0 0 0 0 0 0
East Missoula 923 4 1 1 0 0 0 0 0 5 0 1 1 1 1 0 - - 4 3 0 1 1 1 1 0
East Missoula 924 17 0 0 0 0 0 0 0 4 0 1 1 1 1 0 - - 4 3 0 1 1 1 1 0
East Missoula 926 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 927 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 928 1 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 929 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 930 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 931 21 11 0 11 0 0 0 0 - 0 0 0 0 0 0 217 65 0 0 3 16 13 13 16 3
East Missoula 932 24 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 933 8 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 934 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 935 1 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 936 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 937 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 938 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 939 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 940 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 941 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 943 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 944 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 945 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 951 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 954 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 955 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0
East Missoula 956 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 957 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 958 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 959 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 960 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 961 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 962 0 0 0 0 0 0 0 0 19 4 5 1 1 4 3 (0) (0) 19 15 2 3 1 1 2 2
East Missoula 963 0 0 0 0 0 0 0 0 8 2 2 0 0 2 2 - - 8 6 1 1 0 0 1 1
East Missoula 964 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 965 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 967 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 968 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 969 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 970 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 971 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 972 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 981 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 987 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 988 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 989 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 991 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 994 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 995 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 996 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 997 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 999 75 39 0 6 2 0 31 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 1000 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 1007 24 9 0 4 3 0 0 2 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 1008 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 1009 43 2 0 2 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 1010 19 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 1011 1 1 0 0 0 0 1 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 2473 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 2475 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 (0) (0) 0 0 0 0 0 0 0 0

Total 1,416 1,166 227 280 415 14 159 71 3,119 855 966 106 106 535 538 332 100 2,072 1,687 469 550 77 77 316 297
East Missoula 2476 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 71 21 0 0 1 5 4 4 5 1
East Missoula 2477 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 2705 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0
East Missoula 2734 26 0 0 0 0 0 0 0 - 0 0 0 0 0 0 - - 0 0 0 0 0 0 0 0

Data Requirements for
Emissions Modeling in
MOVES
November 2011
Table of Contents
MODEL OUTPUT ...................................................................................................... 1
HPMS INPUT ......................................................................................................... 2
OTHER LOCALLY SPECIFIC DATA ................................................................................... 2
Model Output
The Missoula MPO Travel Model processes model results and produces data files for use with EPA’s
Motor Vehicle Emission Simulator (MOVES). These files are produced for two subsets of links within the
modeling area, defined as links inside of the CO and PM air quality maintenance and non-attainment
areas. Links that originally crossed a boundary have been split to allow an accurate accounting of data
within each maintenance area. If necessary, the definition of the boundaries can be changed by
modifying roadway network link attributes. Specifically, the following link attributes identify links
included in the MOVES output files:
CONF_PM10: Identifies links in the PM10 air quality boundary, and

CONF_CO: Identifies links in the CO air quality boundary.
Data is output for links with a value of “1” in the corresponding network attribute.
Each time the model is run, the MOVES utility produces the data files listed below. Each file can be
directly imported into a MOVES run specification using tools available in the user interface.
Average Speed Distribution: This table represents the fraction of total VHT that occurs in each
of 16 speed categories (i.e., speed bins). VHT data is also broken down by time of day, roadway
type, and vehicle type.
o MOVES Table Name: AverageSpeedDistribution
o Utility Output Filename: VHT_CO.csv, VHT_PM.csv
o Dependent Variables and Source Data:
 Speed Bin: Travel model results
 Hour of Day: Travel model results processed using traffic count summary data
 Roadway Type: Travel model facility types, grouped into MOVES roadway types
 Vehicle Classification: Assumed uniform for all vehicle types
Hourly VMT Fraction: This table represents the fraction of total VMT by hour on a typical
weekday as represented by the travel model. VMT data is also broken down by roadway type
and vehicle type.
o MOVES Table Name: HourVMTFraction
o Utility Output Filename: VMT_CO.csv, VMT_PM.csv
o Dependent Variables and Source Data:
 Hour of Day: Travel model results processed using traffic count summary data
 Roadway Type: Travel model facility types, grouped into MOVES roadway types
 Vehicle Classification: Assumed uniform for all vehicle types
Data Requirements for Emissions Modeling in MOVES - 1

HPMS Input
MOVES requires input of an annual on-road VMT estimate, which can be obtained from HPMS data.
Because HPMS data is only available for Missoula County as a whole, annual countywide VMT will be
reduced to reflect the maintenance and non-attainment areas using the proportion of countywide daily
VMT data from the travel model that occurs within the air quality boundaries.
For forecast years, base year VMT data will be factored upwards based on the increase in modeled VMT
for the base and future year scenarios.
The MOVES model can also accept detailed VMT fraction data by month (as a percentage of annual
VMT) and by day (as distributed between weekdays and weekends). The Missoula analysis will utilize
automated traffic recorder (ATR) data only if the available data is shown to be reasonable. If the
available ATR data is not reasonable or sufficient for this purpose, the default MOVES values will be used
instead.
Other Locally Specific Data

In addition to the HPMS and model data, MOVES requires additional information about local conditions.
Key requirements are listed below by MOVES table name. Forecast year model runs assume that base
year distributions remain constant over time, unless data suggests otherwise.
SourceTypeAgeDistribution: Distribution of Vehicles by Age for each vehicle type (provided by MDT).
SourceTypePopulation: Number of vehicles within each of 13 vehicle types for the base year (vehicle
registration data, similar to REG DIST in Mobile 6.2; provided by MDT).
FuelSupply: Market Shares for the different fuel types used in the region (provided by the Montana
Department of Environmental Quality).
FuelFormulation: The composition of different fuels used in the region (RVP, sulfur, oxygenates etc.;
provided by the Montana Department of Environmental Quality).
zoneMonthHour: Meteorological data representing Temperature and humidity. Information will be

specified for consistency with past Mobile model runs.
IMCoverage: Inspection and maintenance programs are not planed or currently in place in Missoula
County.
HPMSVTypeYearData: Total VMT by the 6 HPMS vehicle types for the maintenance and non-attainment
areas in the base year. (Countywide data will be factored to reflect the air quality boundaries using
travel model data).
Data Requirements for Emissions Modeling in MOVES - 2

Turn Movement Module
User’s Guide
May 2014


Table of Contents

PROCESS TURNS ...................................................................................................... 1
The Built‐In Turn Movement Utilities ................................................................................................... 1
Turn Movement Add‐In ........................................................................................................................ 1
Intersection Definition File.................................................................................................................... 2
Unadjusted Turns .................................................................................................................................. 3
Adjusted Turns ...................................................................................................................................... 3

Table of Contents ‐ i


Process Turns

The travel model saves turn movement information for selected intersections during the traffic
assignment routines. No calibration or validation process has been performed to ensure that turn
movements produced by the travel model are reasonable or realistic. Raw movement data should not
be used directly for analysis, so a utility is provided to adjust data using processes defined in NCHRP‐
255. TransCAD also contains built‐in utilities for estimating turn movements based on link flows. These
estimates can be used to get a general sense of activity at an intersection. When using the built‐in utility,
turning movement counts should be compared to base year model results, as well as turning movement
forecasts.

The Built-In Turn Movement Utilities

The built‐in turn movement utility can be accessed after opening or creating a map that includes traffic
assignment results (e.g., in a joined view). To view estimated turn movement results, click the
intersection diagram tool in the main mapping toolbox: and then select a node. After settings are
entered in the dialog box, an intersection diagram will be created. For additional details on operation of
this function, see the TransCAD program documentation.

Alternately, raw modeled turn movements can be viewed using the tool accessible from Planning 
Planning Utilities  Display Intersection Flows. Modeled turn movement volumes for each assignment
are saved in “AMTurns.bin”, “PMTurns.bin” and “OPTurns.bin.” Turn movements are only saved for
intersections with a value in the INT_ID node field of the input roadway network. When saving turn
movements, each node with a value in the INT_ID field should contain a unique positive number.

Turn Movement Add-In

Operation of the intersection processing utility requires additional data and is only run for those
intersections identified by the INT_ID field on the node layer. All intersections with a value in this field
can be included in the analysis. This ID also serves as a link between the TransCAD network and
information contained in other databases, such as a Synchro network. For functions requiring count data
and to export data to in an external format, the INT_ID field on the node layer must match the node ID
of intersections in a turn movement count file and a Synchro network. The turn movement processor is
currently limited to intersections with three or four legs. To access turn movement add‐ins, click the
Process Turns button on the main model dialog box (Maps and Reports tab).

1


Figure 1: The Process Turns Dialog Box

Load, create,
or edit an
Intersection
Definition file.
Save raw
(unadjusted)
turn
movements to
a CSV or HTML
file.
Save adjusted
turn
movement
estimates to a
CSV file.

Intersection Definition File

All turn movement functions require an intersection definition file that identifies the configuration of
each intersection selected for analysis. A new intersection definition file can be created from the turn
movement dialog box, or a previously created definition file can be loaded. Turn movement definition
files reference INT_ID values and link ID values, so a new definition file must be created after certain
input file modifications.

Once an intersection definition file is created, it should be verified for accuracy. For functions requiring
existing traffic count data, the intersection leg definitions must match those in an input data file.
Definition files should be checked manually to ensure that the correct legs are identified at each
intersection. To do this, click the “Edit” button after creating an intersection definition file. If necessary,
correct the intersection definition file by adjusting the definitions of each intersection as shown in
Figure 2.

2


The turn movements utility cannot process intersections with more than 4 legs or with diagonal legs.
To overcome this limitation, all intersection approaches must be renamed as N, S, E or W in a
Synchro network and 5‐legged intersections must be evaluated manually.

Figure 2: Editing the Intersection Definition File

Intersection
approaches are
color‐coded. The Use the arrows to
red leg is the browse through
northern intersections, or the
approach. Jump button to
select an
intersection by
number.
Click on a button
and then click on an
intersection leg to
define its direction.
The initial guess may
not be correct. Here,
the north and east
legs are defined
incorrectly.
Unadjusted Turns

Unadjusted turn movements can be saved to a comma separated variable (.csv) file. This file can then be
read using a program such as Excel or can be imported to Synchro. For import to Synchro, intersection ID
numbers and approach directions must be consistent. When outputting adjusted turns for forecast
scenarios, unadjusted turns from a base year (i.e., 2010) scenario are also required as an input file.

Adjusted Turns

The model can adjust turn movements using NCHRP‐255 intersection procedures. Use the steps
described below to export adjusted turn movement data to a CSV file that can be imported to Synchro.

1. Identify intersections in the TransCAD Network:
Enter the intersection ID into the INT_ID field in the node layer of a TransCAD network. If a separate
network is to be used for the calibrated base year model run, INT_ID values must be entered into
3


this network as well. The INT_ID field should be cleared for nodes or intersections that are not
present in the Synchro network.

2. Run the base and forecast year travel model:
The base and forecast year model scenarios should be run in full with INT_ID information present in
both networks.

3. Create a base year intersection definition file:
Use the Process Turns utility to create an intersection definition file. Ensure that all intersection
approaches are defined in a manner consistent with the Synchro file.

4. Export the unadjusted base year model turn movements:
Save the turn movement data with a name that is easy to remember.

5. Create a CSV file containing observed turn movement data:
This file must contain complete turn movement data for each intersection to be analyzed. The file
can be created by exporting turn movement data from Synchro, or by modifying the unadjusted
base year model turn movements file to contain turn movement count data.

6. Load or create a forecast year intersection definition file:
In most cases, the file created for the base year can be re‐used. If not, a new file must be created. If
in doubt, use the Edit function to verify that a loaded intersection definition file is correct.

7. Export the adjusted forecast turn movements:
CSV files containing base year count data and base year unadjusted modeled turn movements must
be referenced. Exported turn movements can be loaded into Synchro or a spreadsheet program for
additional analysis.

Once the above steps have been followed, an adjusted turns CSV file will be created, which contains
turn movement forecasts based on observed turn movement counts and travel model forecasts.
However, these forecasts are estimates and professional judgment should be used to interpret the
results. Where intersection configurations change or where turn movement count data is suspect,
manual intervention will be required. If more detailed information, such as a traffic study, is available,
this information should be used in addition to or instead of these planning level forecasts.

4

MissoulaTDME Full Documentation 07051

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

MissoulaTDME Full Documentation 07051

Hochgeladen von

Copyright:

Verfügbare Formate

Model User’s Guide

SYSTEM REQUIREMENTS ................................................................................ 1

INSTALLING THE MODEL ADD-IN ..................................................................... 2

RUNNING THE MODEL ................................................................................. 6

MODEL UTILITIES – ROADWAY NETWORK ........................................................ 12

MODEL UTILITIES – ROADWAY NETWORK ........................................................ 15

MODEL UTILITIES – MAPS AND REPORTS .......................................................... 15

MODEL DATABASE .................................................................................... 17

TRANSPORTATION NETWORKS ...................................................................... 22

WORKING WITH TRANSIT DATA .................................................................... 33

Context and Background

Table 1: System Requirements

Installing the Model Add-In

Alternate Method: Manual Installation

Windows XP: C:\Documents and Settings\All Users\Application Data\LSA\Missoula Model

1. Select Tools  Setup Add-Ins... from the menu,

a. Type: Dialog Box

Figure 1: Example Model Directory Tree

Running the Model

Collecting the Required Data

Table 2: Model Input Files

Creating and Running a Scenario

Input and output directories,

Figure 2: The Scenario Toolbox

The status will read

Change the order in

Figure 3: The Scenario Editor (Input Tab)

Filenames and file

Figure 4: The Scenario Editor (General Tab)

Figure 5: The Scenario Editor (Output Tab)

Figure 6: The Scenario Editor (Advanced Tab)

Subarray data can

Note: Advanced model parameters should not typically be changed.

Figure 7: The Model Dialog Box

Utilities can be run

Running Selected Model Steps

Figure 8: Sub-Steps Dialog Box

Running Speed Feedback

1. The model is not set to stop after each step, and

1. Run a complete model run with speed feedback enabled.

Model Utilities – Roadway Network

Add/Delete Network Year

Figure 9: Add/Delete Network Year Dialog Box

Refresh Network Defaults

Copy Feedback Results

Create Select Query

Model Utilities – Roadway Network

Model Utilities – Maps and Reports

Create Performance Report

Table 3: Map Types

Map Category Map Type Description

Table 4: Map Options

Map Category Map Option Description

Roadway Parameters (lookup tables by facility type and area type)

Figure 10: Access Database Form Interface

These buttons Use this button

Figure 11: Data Scenario Management

Table 5: Access Database Tables

Category Table Name Description

Transportation Network Components

Nodes: Nodes are attached to links. Types of nodes include:

Example Transportation Network Application

Part 1: Prepare Workspace

5. Use the zoom tool ( ) to zoom to the vicinity of the Millsite.