Overview
Welcome to the official Users Guide for TDM23, the travel demand model for the Boston Region, developed and maintained by the Central Transportation Planning Staff (CTPS). This guide provides comprehensive documentation to help you understand and use TDM23, from initial setup to advanced features. TDM23 supports Destination 2050 , the Boston Region Metropolitan Planning Organization’s current Long-Range Transportation Plan. Covering all of Massachusetts, Rhode Island, and southeastern New Hampshire, TDM23 is a trip-based aggregate model that simulates surface travel patterns on a typical weekday using a static highway and transit assignment.
This user’s guide supports TDM23 version 1.0 (TDM23.1.0). For a detailed walk-through of TDM23.1.0, refer to the Structures and Performance Report.
To access the model inputs and scripts for TDM23, please email model.support@ctps.org with details about your organization and intended use.
Inputs
TDM23 uses demographic and employment data from the Metropolitan Area Planning Council’s (MAPC) UrbanSim land use model. UrbanSim is a disaggregate model that synthesizes and allocates population and employment data, providing annual estimates for Massachusetts from 2010 to 2050. This data is segmented by zones, enabling TDM23 to accurately simulate travel patterns based on residential and employment locations across the state.
For the network, TDM23 offers a higher resolution of transit representation compared to typical statewide models. However, it does not include a freight model component or a long-distance travel component. The network and zone system are more detailed within the Boston Region MPO than in the rest of the model area.
In addition to zonal and network inputs, TDM23 incorporates various parameters related to model components, such as air quality rates, zonal fares, and terminal times. It also includes configuration parameters like feedback loops, highway iterations, and run options. For more about TDM23 inputs, please refer to the Inputs page.
Components and Market Segments
The figure below provides an overview of the model components by trip category, including household, university, airport, trucks, and external trips. Initially, the steps are processed in a disaggregated format, highlighted by the green tabs. During the distribution step, trips are aggregated, and all subsequent steps are processed in an aggregated format, indicated by the blue tabs.
Travel behavior in TDM23 is estimated at various levels of aggregation based on the trip category and attributes. Regular trips estimated at the household level are segmented by vehicle availability while those estimated at the worker level are segmented by both vehicle availability and worker income. The following tables display the levels of aggregation and the market segments by trip purpose for regular and irregular household trips.
| Estimation Level | Description |
|---|---|
| Worker Level | Work-related trips and work-from-home behavior |
| Household Level | Non-work-related regular household trips and vehicle availability |
| Zonal Level | Irregular household trips, non-resident household trips, and truck trips |
| Trip Purpose | Description | Estimated by | Distribution Segments |
|---|---|---|---|
| HBW | Home-based work | Worker | Vehicle Sufficiency, Worker Income |
| HBPB | Home-based personal business | Household | Vehicle Sufficiency |
| HBSR | Home-based social recreation | Household | Vehicle Sufficiency |
| HBSC | Home-based school | Household | Vehicle Sufficiency |
| HBU | Home-based university | TAZ | NA |
| NHBW | Non-home-based work | Worker | NA |
| NHBNW | Non-home-based non work | Household | NA |
| Trip Category | Segments |
|---|---|
| Special Generators | Personal Business, Social Recreation |
| Airport Ground Access | Resident/Visitor, Business/Leisure |
| Externals | Auto, Medium Truck, Heavy Truck |
| Trucks | Medium Truck, Heavy Truck |
| Mode | Mode Abbreviation | Home-Based Trips (Except School) | Home-Based School | Non-Home-Based | Airport |
|---|---|---|---|---|---|
| Drive alone (SOV) | DA | ✔ | ✔ | ✔ | |
| 2 Person shared ride (HOV2) | S2 | ✔ | ✔ | ✔ | |
| 3+ Person shared ride (HOV3+) | S3 | ✔ | ✔ | ✔ | |
| Walk | WK | ✔ | ✔ | ✔ | |
| Bike | BK | ✔ | ✔ | ✔ | |
| Auto access to transit | TA | ✔ | ✔ | ✔ | |
| Walk access to transit | TW | ✔ | ✔ | ✔ | ✔ |
| Ride-sourcing | RS | ✔ | ✔ | ✔ | ✔ |
| School Bus | SB | ✔ | |||
| Drive and park | DP | ✔ | |||
| Pickup or Drop-off | PU | ✔ | |||
| Logan Express | LX | ✔ | |||
| Rental car | RC | ✔ |
Temporal
TDM23 represents a full “average weekday” of travel, dividing the day into times of day for the supply components and time periods for the demand components.
There are four times of day used to estimate volumes and boardings for the roadway and transit networks:
- AM Peak: 6:30 AM - 9:30 AM
- Midday: 9:30 AM - 3:00 PM
- PM Peak: 3:00 PM - 7:00 PM
- Night: 7:00 PM - 6:30 AM
Additionally, there are two time periods to estimate the demand response to roadway congestion and transit service:
- Peak time period: AM Peak time of day
- Non-Peak time period: Midday time of day
These time definitions were chosen to minimize aggregation errors in the static assignment process, informed by an analysis of household surveys, traffic counts, transit services, and data from platforms like Streetlight and the Regional Integrated Transportation Information System.
Implementation
TDM23 combines Caliper’s TransCAD software platform with Python, utilizing TransCAD’s Flowchart User Interface for its primary operation. Python was chosen for TDM23 development due to its efficient management of inputs and outputs and its streamlined development and debugging process. TDM23 is calibrated to a 2019 base year, assuming the transferability of pre-pandemic travel behavior to future scenarios.
| Component | Estimates | Sensitive To |
|---|---|---|
| Vehicle Availability | Household vehicle availability relative to household drivers (zero, fewer than drivers, greater than or equal to drivers) | Household size, income, workers, children Transit access density |
| Work from Home | Share of commute vs. work at home days | Regionally specific inputs of work-from-home levels |
| Trip Generation | Resident average daily trips within region by purpose produced and attracted by zone | Person type Household size, income, vehicles Household children, seniors, non-workers Employment by category |
| Peak/Off-peak | Segmentation of trips into peak period (AM or PM) and off-peak (MD or NT) | Trips by zone, purpose and market segment |
| Trip Distribution | Flow of trips between zones | Trip productions and attractions by peak/off-peak Path impedances Mode choice utilities |
| Mode Choice | Mode shares and flow of trips by mode | Trip tables by purpose, market segment, and peak/off-peak Path roadway and transit level of service |
| University Travel | Generation and distribution of off-campus university student travel | Commuter enrollment Household population |
| Truck Trips | Generation, distribution, and time of day of medium, and heavy truck trips | Employment Path distances |
| Airport Ground Access | Distribution, time of day, and mode of airport traveler trips | Airport non-transferring enplanements and deplanements |
| Special Generator, Externals | Non-average daily trips (airport) and non-resident/outside of region trips (through trips) | Trips produced/attracted by zone |
| Time of Day | Outbound and inbound trip time of day period | Trip tables by purpose, market segment, peak/off-peak, and mode |
| Component | Estimates | Sensitive To |
|---|---|---|
| Access Density | Access density category of TAZ | Population and employment density Transit location by mode |
| Highway Assignment | Congested speed and volumes by roadway segment | Trip tables by vehicle type and occupancy, market segment, and time of day Roadway network |
| Transit Assignment | Transit activity (Park-and-Ride [PnR], boardings, alightings, transfer) by line segment | Trip tables by transit access mode, market segment, and time of day Transit network |
Outputs
TDM23 produces outputs at every step of the model process. These outputs are initially generated at the most detailed disaggregated segment level and can be aggregated as needed. For more about TDM23 outputs, please refer to the Outputs page.
| Component | Estimates |
|---|---|
| Households | Household trips by purpose, Vehicles |
| Persons | Worker trips |
| TAZ | Person trips by mode, Truck trips |
| Roadway link | Vehicle volume by time of day, Congested speeds |
| Transit route / stop | Boardings / Alightings / Loading by time of day |