Travel Forecast

About the Travel Demand Forecast Model

SEMCOG’s Travel Demand Forecast Model estimates current and forecast motorized vehicle travel volumes, speed, and patterns in Southeast Michigan. The data is used to inform policies, develop projects, and identify opportunities for collaborative initiatives contained in SEMCOG’s long-range transportation plan.

For day-to-day operations, the travel model adds value to SEMCOG members by:

  • supporting master plan updates,
  • analyzing traffic safety improvements,
  • considering economic development impacts, and
  • determining air quality conformity.

Travel modeling is both required and essential to planning efforts at SEMCOG. This model is constantly refined. Data inputs such as population, employment, and households are updated to reflect current demographics. Data outputs compare favorably with those of other regional area models of the same size.

To assist decision makers in making informed transportation planning decisions and transportation project investments, SEMCOG maintains three regionwide traffic models for forecasting and simulating transportation activities, and uses the EPA’s latest mobile vehicle emission model (MOVES) for transportation conformity analysis. These three traffic models include: two types (trip-based and activity-based) of travel demand forecast model (TDFM), which uses static traffic assignment, and the third one is a dynamic traffic assignment (DTA) model. Currently, SEMCOG’s TDFM for official use is trip-based (the version of E8Plus). To improve the modeling capabilities, SEMCOG has recently begun transitioning to a newly developed activity-based model (ActSim 1.0). These models differ in their level of detail, the methodologies employed, and the type of output they generate. The table below highlights the features in comparison among these three traffic models.    

Model Feature TDFM - E8Plus TDFM: ActSim 1.0 Dynamic Assignment
Resident Travel Trip-based Activity-based Not Applicable (N.A.)
-Spatial Scale TAZ as Origin and destination (O/D) MAZ as Origin and destination (O/D)
-Temporal Scale 5 time periods 48 half-hour periods
-Simulation level Aggregated Household groups Both person level and household level
-Produced Travel Choices Group travel choices (trip ODs, modes, time) segmented by HH size, workers, children, auto & income Individual travel choices (work/school locations, transit pass, daily activities, tour/trip destinations, scheduling & modes)
Truck Travel Tour-based 
-Spatial Scale TAZ as O/Ds
-Temporal Scale 48 half-hour periods
-Simulation level Disaggregated Firms and Trucks
-Travel choices Firm activities, stops, vehicle type, stop duration, tours, scheduling 
Traffic Assignment Static Dynamic
-Roadway Highway & Transit Highway only
-Network Details Planning network Detailed network, Intersection & signals
-Spatial Scale TAZ as O/Ds
-Temporal Scale 5 time periods 48 half-hour periods
-Vehicle Classes 3 passenger vehicle classes (SOV, HOV, HOV2) and 3 truck classes (Light, Medium, Heavy) 3 classes (SOV & Light Trucks, HOV & HOV2, Medium & Heavy Truck)
-Traffic Output Link level volume, speed, volume/capacity ratio Link level volume, speed, delay, queue
-Transit Modes Walk/PnR/KnR to local/premium buses N.A.
-Transit Output Route level ridership, stop on/off flow N.A.

Each of these four models is described in the following sections, with links provided to detail technical documentation and model reports online.

To request travel forecasts and model data, please fill out the Data Request Form (pdf, 61KB)

1. SEMCOG Official TDFM – E8Plus

As the Metropolitan Planning Organization (MPO) for Southeast Michigan, SEMCOG maintains a fully calibrated and validated regional travel model in order to meet federal requirements of developing Regional Transportation Plan (RTP) and Transportation Improvement Program (TIP). SEMCOG’s regional TDFM, referred to as the version E8Plus, was last updated in 2023 for SEMCOG’s 2050RTP development. The E8Plus model covers both personal trips and commercial vehicle movements within and through the region. 
The person travel demand estimation is trip-based, implemented in TransCAD, and being calibrated with the 2015 household travel survey and the 2019 onboard transit survey. Compared to conventional trip-based models, some advanced model features included in SEMCOG’s TDFM are: a destination choice model, which is used in trip distribution for internal person trips, and a nested logit model structure used in the mode choice component. 
The commercial vehicle (CV) travel demand estimation is tour-based, programmed in R, and simulates the formation of commercial vehicle tours within the SEMCOG region, as well as commercial vehicle trips to, from, and through the region. These CV travel model components were implemented using SEMCOG’s 2017 commercial vehicle survey and other observed truck data. 
The last step of SEMCOG TDFM includes a model of both highway and transit assignments in TransCAD. The vehicle trips in the origin and destination pairs, estimated from person travel and commercial vehicle travel are assigned to the regional highway networks using a static highway assignment procedure. Traffic assignment is performed for six separate vehicle classes: single occupant vehicle, shared ride 2, shared ride 3+, light commercial vehicle, medium commercial vehicle, and heavy commercial vehicle. The transit person trips resulting from the person travel estimation are assigned to the fixed transit route system. The transit assignment does not include capacity constraint, so increasing transit volumes does not result in diversion of transit trips to other transit services. 
The base year 2020 of E8Plus model used the 2020 household/population and the 2019 employment data as model inputs to validate the model output against the travel observed before COVID in the region. The travel behavior changes due to COVID were not reflected in this E8Plus model version. Regional travel was forecasted in five-year intervals from the base year 2020 to the future year 2050, the last year of SEMCOG’s 2050 RTP.

Resource Available Online:

2. SEMCOG's Next Generation TDFM

In the efforts of improving the region’s travel behavior analysis tools, as noted above, SEMCOG has developed a new ABM, which will replace the resident travel demand model in SEMCOG’s official TDFM - E8Plus from trip-based to activity-based. This ABM, which includes an implementation of ActivitySim 2-zone prototype (programmed in Python) for SEMCOG region, addresses resident travel choices at both household-level and person-level, as well as intra-household interactions between household members. It will better support the region’s current and future planning needs by providing rich and detailed information regarding travel patterns under a wide variety of scenarios. 
The current version of SEMCOG’s ABM, referred to as ActSim Version 1.0, was calibrated and validated using SEMCOG’s 2015 household travel survey and 2019 on-board transit survey.
While SEMCOG’s trip-based model is still the production model for supporting SEMCOG’s RTP/TIP process and air quality conformity analysis, SEMCOG staff has been evaluating the newly developed ABM in order to transition from the trip-based model for SEMCOG’s next long-range plan.  

Resource Available Online:

Model improvement plan: 

3. SEMCOG DTA

SEMCOG’s regionally calibrated dynamic traffic assignment (DTA) is a mesoscopic vehicular traffic simulation model using DynusT\ DynuStudio software, which is capable of simulating a wide range of operational planning scenarios. SEMCOG’s regional DTA has been developed for the model years of 2015 and 2020 using the trip matrices produced from SEMCOG’s regional TDFM. 

The big difference comparing to the static assignment in SEMCOG’s TDFM comes from the three parts: 1) traffic control signal timing data is taken into consideration as part of regional network, 2) trip matrices specified in short 30-minutes intervals, 3) the link volumes, travel times and speed are time-variant, which is inherently consistent with actual congestion dynamics. As a result of these improved features, the DTA model can estimate more realistic and precise traffic flow patterns, travel time\speed, delay and queuing information. 

SEMCOG has used the DTA dozens of times to assist MDOT, counties and communities make better informed decisions on projects such as I-375, I-96, Gratiot Avenue, and I-94

Travel Model Improvement Program (TMIP): Practical Guide on DTA applications-SEMCOG case study (pdf, 2.2MB)

4. EPA's MOVES

SEMCOG’s regional transportation conformity is using EPA’s latest Mobile Vehicle Emission Simulator – MOVES4 to analyze the emissions produced from the activities of mobile vehicles. Currently, the seven counties of SEMCOG region are maintenance areas for both the 2006 24-hour PM2.5 and the 2015 8-hour ozone National Ambient Air Quality Standards, so transportation conformity analysis is required for the region. MOVES inputs are developed using local data when available and proper to be used. The MOVES run specification and the methodology used to develop local inputs for SEMCOG’s 2050RTP is included in this report. 

Travel Surveys

Other Reports and Presentations

Other Tools

LOCUS Data

SEMCOG has purchased LOCUS origin-destination flow data, which are derived from location-based services data collected passively from mobile devices. These data provide robust information about how people are moving, which locations they are visiting, and when their travel is occurring. This data can provide a deeper understanding of how travel patterns have changed during and after the COVID pandemic.

To request this data, contact Saima Masud.

 

 

Developing Regional Solutions
SEMCOG is a regional planning partnership of governmental units serving 4.8 million people in the seven-county region of Southeast Michigan striving to enhance the region's quality of life.