Introduction

This report compares different sources of Vehicle Miles Traveled (VMT) data readily available for counties in Minnesota’s metropolitan region on an annual scale.

The sources included are:

Each of these sources uses different methods to estimate VMT. To compare the different sources of VMT data, I have limited my analysis to the counties of Anoka, Carver, Dakota, Hennepin, Ramsey, Scott, and Washington in Minnesota.

Vehicle Miles Traveled

Vehicle Miles Traveled is a metric for the amount of travel for all vehicles in a geographic region over a given time, typically a one-year period.

Vehicle Miles Traveled are relevant in climate change mitigation discussions as they reflect the amount of driving that occurs in a community, and therefore, the greenhouse gas emissions.

Source A: Minnesota Department of Transportation

The Minnesota Department of Transportation (MNDOT) publishes Vehicle Miles Traveled estimates by county by different road types on an annual basis.

The methodology for the VMT Estimate from MNDOT can be found here: https://www.dot.state.mn.us/roadway/data/data-products.html

The main advantage of this approach is that it is based on actual observations of road transportation networks in the region. However, a disadvantage of network data on VMT is that it fails to represent the actual activity generated by a community, and therefore is less accurate in measuring greenhouse gas emissions at the community scale. For example, a small community that is intersected in its geographic boundary by an interstate highway might get an overestimate of emissions, without really being the source of such emissions.

From MNDOT’s website:

Minnesota Department of Transportation calculates Vehicle Miles Traveled by multiplying annual average daily traffic (AADT) by the centerline mileages of each roadway segment under consideration.

Annual Average Daily Traffic is the roadway estimates of total vehicles on a road segment on any given day of the year (all directions of travel). This represents the total number of vehicles per year divided by 365 and is developed using factors to adjust for season, day of the week, and vehicle type.

One mile of a single roadway, regardless of the number of lanes, is called a centerline mile. While the centerline mileage does not account for the number of lanes, lane mileage does. Lane mileage can be found by multiplying the continual driving lanes and centerline mileage. Temporary lanes like turn lanes are not counted nor are lanes on ramps or in auxiliary areas such as rest areas.

Notes

  • Traffic volume data does not always represent the current publication year; when traffic volumes are unavailable for individual sections of road in a given year, the traffic volumes from an earlier year are adjusted and applied to the current year’s VMT calculation.

Table 1

Source B: UrbanFootprint

UrbanFootprint(c) is a private technology firm that offers Vehicle Miles Traveled as one of multiple outputs available through their platform. Their platform takes parcel level information on land uses to model the travel demand of a community.

The methodology for the Transportation Analysis of Urbanfootprint can be found here: https://help.urbanfootprint.com/methodology-documentation/transportation-analysis#methodology

One disadvantage of Urbanfootprint is that it makes difficult for communities to understand how emissions change over time. Urbanfootprint is not a tool to understand historical trends. However, one advantage unique to Urbanfootprint is that their platform allows to create ‘scenarios’ based on different land use assumptions, which could be advantageous in trying to understand the connection between land use and transportation.

From UrbanFootprint’s website:

The core of the current Transportation Module, the MXD method, is based on a traditional four-step travel demand forecasting model. It has three key components⁴:

  • Trip generation
  • Trip distribution
  • Mode choice modeling

Table 2

Source C: National Renewable Energy Laboratory

The National Renewable Energy Laboratory offers an estimate of Vehicle Miles Traveled as part of their SLOPE (State and Local Planning for Energy) Tool.

The Methodology for the National Renewable Energy Laboratory estimate of Vehicle Miles Traveled from the TEMPO model (Transportation Energy & Mobility Pathway Options) can be found here:

https://www.nrel.gov/transportation/tempo-model.html

The NREL data is so far a projection based on a baseline year (2016) thus it performs poorly for recent years. However, it is the only platform to provide a forecast of transportation emissions.

From NREL’s website:

Projected VMT for light duty personal vehicles are modeled for each county through 2050 using NREL’s TEMPO: Transportation Energy & Mobility Pathway Options. The TEMPO model represents household-level travel needs, technology adoption, and vehicle use and estimates future VMT. The data shown here do not include fleets or medium- and heavy-duty vehicles.

Table 3

Source D: Google Environmental Insights Explorer

Google Environmental Insights Explorer is a tool that provides GHG metrics at the city and county levels for cities around the world.

Their estimate of vehicle miles traveled is derived from mobile device data, similar to Streetlight. However, Google has unique capabilities given that they are the company that provides the Google Maps service, which continues to be the #1 GPS navigation app for smartphones.

The Methodology for Google’s estimate of Vehicle Miles Traveled can be found here:

https://insights.sustainability.google/methodology#transportation-methodology

From Google Environmental Insights Explorer website:

To characterize trips taken within and across city boundaries, we applied privacy filters, aggregation and anonymization techniques, and inference models to data derived from Google’s proprietary location history data.

There are clear advantages from the data provided by Google 1) it is based on large sample size

Table 4

Source E: Streetlight Data

Streetlight is a private company that provides analytics related to transportation. They assert to be pioneers in the use of big data and machine learning to gain insights on transportation.

While Streetlight provides more than one method to estimate Vehicle Miles Traveled, I will focus on their origin-destination analysis, as this is the recommended approach to calculate greenhouse gas emissions.

The origin-destination analysis is limited to trips with origin and/or destination in the seven counties of the metropolitan area.

The methodology used to estimate Vehicle Miles Traveled can be found here: https://support.streetlightdata.com/hc/en-us/articles/360038900092-How-to-Get-Regional-VMT-using-an-O-D-Analysis

From Streetlight’s website:

The calculation then multiplies the average length of these trips by the number of trips to generate an estimated VMT for the region.

Table 5

Comparisson of Data Sources

Having presented how each of these sources are able to capture (or not) annualized trends of Vehicle Miles Traveled. The figure below shows how each of the sources compares next to each other.

Conclusion

All the data sources seem to show results that are relatively consistent with each other. The NREL dataset must be used with caution, as it fails to capture the annual trends in Vehicle Miles Traveled. Figure 5 shows that in smaller counties, the margin of error between the different sources is smaller. The margin of error gets bigger with larger and more populated counties.