I am evaluating Bring Back 6th alternatives for Our Streets using an enhanced version of the Metropolitan Council travel demand model. In this report, I evaluate the existing model. In subsequent work, I will develop the enhanced model and do the alternatives analyses.
Vehicle mobility is a key metric in roadway alternatives analyses. Given the relative lack of congestion in this corridor, the September 2023 Olson Memorial Highway Multimodal Study Purpose and Need Statement (“P&NS”) prepared by MnDOT lists vehicle mobility only as a secondary need. Even though vehicle mobility is given as a secondary need, it is possible that it will still have a large weight in evaluation because vehicle mobility metrics are ubiquitous in alternatives analyses and can appear to be more objective compared to some of the other metrics that are harder to quantify.
The vehicle mobility metrics are taken from computer models. For the metrics to be credible, it is necessary that the model match metrics for existing traffic conditions. As documented in this report, the current Met Council model does not. Problems in Olson Memorial Highway modeling include:
- large errors in traffic volumes,
- large errors in traffic speeds, and
- modeled speeds being overly sensitive to roadway capacity.
An old adage in computer modeling is: “All models are wrong; some models are useful.” Unfortunately, in its current state, the regional model is not useful for evaluating Olson Memorial Highway alternatives.
These types of model issues are present throughout the United States. Increased computer power has led to improved modeling in some domains, including weather forecasting, but it has not improved the accuracy of transportation forecasting significantly. The focus over the past 20 years has been on activity-based models (“ABMs”), including the Met Council model. It was assumed that simulating the travel patterns of individuals within individual households, including simulating future virtual individuals and households, would improve model accuracy. However, ABMs have not addressed the models’ fundamental issues in matching speeds and traffic volumes. In addition, ABMs have hindered progress in these areas because the added complexity makes the models very cumbersome to update or even to apply. The Met Council model is built on household activity survey data from 2010, a full decade before the start of the pandemic that has had lasting impacts on behavior.
The enhanced model will keep the ABM structure but focus on better matching of existing speeds and traffic volumes.
Even with an improved model, it is best to think of the model as one seat at the table in dialog with other types of data and community input. The horizon year for planning major road projects in the region is 2050. Any statement like: “In 2050, Alternative A will result in time savings of 1.37 minutes” should be met with extreme skepticism. No model will ever be that accurate. An improved model will be more useful than the current one, but it will not be able to predict the future in 2050 precisely. There are too many uncertainties about the effects of demographic changes, energy prices, climate regulation, telecommuting and distance learning, virtual reality, artificial intelligence, and other changes.
Although it will be incapable of making precise predictions for 2050, the enhanced model will be useful in a general comparison of alternatives. The model can also be used to test whether the general ranking of alternatives is sensitive to key assumptions about the future. How much the model should be weighted relative to other input will depend on how much the model can be improved, but the model should never override other input.
It is also important not to focus too much on the metrics that can be quantified. The P&NS documents that current average speeds in the corridor are considerably higher than the posted speed. Nevertheless, the Olson Memorial Highway Multimodal Study Evaluation Criteria report (January 2024) includes “corridor travel time” as a metric with an assumption greater than a 20% decrease is “good,” greater than a 20% increase is “poor,” and lesser changes are “fair.” The travel time metric is problematic in this corridor because calming traffic will be required to meet the primary walkability and bikeability needs. Therefore, any alternative that meets the primary needs will likely rate “poor” on corridor travel time. The decision process should focus on the primary purposes and needs, even if the metrics are qualitative rather than quantitative.
Read the full report.
Bring Back 6th
Once called the “Beale Street of Minneapolis,” the old Near-Northside was an integrated Black and Jewish community that was destroyed in 1939 for the construction of Olson Memorial Highway, a low-trafficked highway “to nowhere” that has been polluting the neighborhood ever since.