Adding more freeway capacity at the Rose Quarter will thousands of tons to the region’s greenhouse gas emissions

If you say you believe in science, and you take climate change seriously, you can’t support spending $800 million or more to widen a freeway.

SYNOPSIS:

• Wider freeways—including additional ramps and “auxiliary lanes”—induce additional car travel which increases greenhouse gas emissions.
• The I-5 Rose Quarter project will add approximately 33,000 vehicles per day to I-5 traffic, according to ODOT’s own estimates
• These 33,000 vehicles will directly add 56,000 daily vehicle miles of travel and indirectly add 178,000 daily vehicle miles of travel.
• Additional vehicle travel will directly produce between 8,000 tons of greenhouse gas emission per year; and with induced travel outside the project a total increase of 35,000 tons of greenhouse gas emissions per year
• The engineered right-of-way for the Rose Quarter project allows for eight standard freeway lanes, which would double freeway capacity in this area and further increase vehicle travel and greenhouse gas emissions.
• Claims that widening freeways will reduce greenhouse gas emissions by reducing crashes and idling have been disproven.

Additional Vehicle Miles of Travel and Greenhouse Gases, ODOT estimates

Currently, I-5 at the Rose Quarter carries about 122,000 vehicles per day.  With the Rose Quarter freeway widening project proposed by the Oregon Department of Transportation, we estimate that traffic will increase to 155,000 vehicles per day.  This represents an increase of 33,000 vehicles per day over current levels.

I-5 North Volumes Existing conditions 2016 v. with Freeway Widening
		Northbound	Southbound	Total	Implied ADT
Time Period		RQ Existing Conditions (2016)
AM Peak	8AM-9AM	2,146	5,133	7,279	122,000
PM Peak	5PM-6PM	3,360	3,639	6,999	122,000
		Widened I-5 RQ Conditions (2045)
AM Peak	8AM-9AM	4,680	5,176	9,856	148,945
PM Peak	5PM-6PM	4,707	5,070	9,777	161,385
Average					155,165
RQ Existing, “2016 Existing Conditions” “Mainline North of Going”
Existing Volumes from pages 333-340 of ODOT “Volume Tables”, dated 5-21-18

We’ve had to compute that estimate ourselves, because, as we have noted, ODOT has suppressed inclusion of average daily traffic figures (the most commonly used traffic volume statistic) from the project’s Environmental Assessment.  To compute average daily traffic from the hourly data in the EA, we have factored up hourly traffic to daily levels, at given the current relationship between peak and total daily travel. Peak hour travel accounts for about 14 percent of daily travel; we’ve used the reciprocal of this amount as our multiplier to calculate future ADT implied by ODOT projections.

Today, according to the Environmental Protection Agency, the average vehicle emits about 411 grams of greenhouse gases per vehicle mile traveled.

Incremental Greenhouse Gas Emissions, I-5 Rose Quarter Project
Line	Item	Direct	Indirect
1	grams per mile	411	411
2	vehicles per day	33,000	33,000
3	miles per vehicle	1.7	5.4
4	miles per day	56,100	178,200
5	grams per day	23,057,100	73,240,200
6	tons per day	23	73
7	tons per year	8,416	26,733
Notes
1	EPA estimate of greenhouse gases per vehicle mile traveled
2	ODOT estimate of increased traffic on I-5
3	Length of project (1.7 miles), average commute (7.1 miles)
4	Line 2 * Line 3
5	Line 1 * Line 2
6	Line 5 * 1,000,000
7	Line 6 * 365

Conservatively, we estimate that the additional 33,000 vehicles per day traveling on just the widened 1.7 mile segment of the I-5 Rose Quarter freeway will generate and additional 56,000 vehicle miles traveled per day, and in turn, that will produce about an additional 8,400 tons of greenhouse gases annually.

Moreover, we anticipate that widening the freeway in this location will induce additional automobile travel on roads connected to this section of freeway.  The 1.7 miles traveled on this segment of roadway is just a portion of typical trips. Given that the average commute trip in the Portland metropolitan area is 7.1 miles each way, we anticipate that the freeway widening will produce an additional 5.4 miles of travel elsewhere in the region, for a total of 178,00 additional vehicle miles traveled per day region wide, which in turn will produce an additional 27,800 tons of greenhouse gases per year.

Combining the direct and indirect effects of additional freeway capacity on travel, the Rose Quarter Freeway widening project is likely to increase Portland area greenhouse gas emissions by more than 35,000 tons per year.

ODOT did not analyze or model the effects of induced demand

While ODOT maintains that the I-5 Rose Quarter Freeway widening project will reduce congestion, that is because it has crafted a model which, by its construction, rules out the possibility of induced demand.  ODOT’s “static assignment model” has been shown to over-estimate traffic levels in base case situations, and understate traffic volumes in “build” scenarios, with the effect that they are systematically unable to accurately predict increased traffic due to induced demand.

The modeling has two related sources of bias:  First, it assumes that in the base case, travel patterns are not influenced by roadway congestion (i.e. that travelers don’t alter trip making behavior to avoid congestion).  These models also allow predicted traffic volumes to exceed the physical capacity of roadways, something that is simply impossible, but which again, leads to over-stating base case volumes.  Second, the models fail to predict that trip-making will respond to increases in capacity.

The EA makes no mention of induced demand, the phenomenon by which increases in highway capacity in urban areas generate additional travel that leads to a recurrence of congestion at even higher levels of traffic. (A text search of both the EA and its Traffic Technical Report show no mention of the word “induced”).

In all of its analyses, the EA uses a single set of assumptions about future land use and travel demand, including the distribution of jobs and population within the metropolitan area general, and within the Project Impact Area in particular. This analysis assumes that building (or not building) this additional freeway capacity will have no impact whatsoever on the pattern and intensity of traffic over the next two or more decades.

This approach has two effects, both of which subvert the analysis of environment impacts and which violate NEPA. In the “No-Build” scenario, levels of traffic are improperly inflated, producing much higher level estimates of congestion than will actually occur. In each of the “Build” alternatives, levels of traffic are systematically understated. This bias causes the EA to mischaracterize the relative merits of the build and no-build alternatives, and therefore violates NEPA.

The phenomenon of induced demand is so well-established in the academic literature that it is referred to as the “Fundamental Law of Road Congestion.”  Add as many un-priced lanes as you like in a dense, urban environment and that capacity will elicit additional trip-making that quickly fills new lanes to their previously congested levels. In the extreme, one ends up with Houston’s 23-lane Katy Freeway, successively widened at the cost of billions of dollars, but which now has even longer travel times than before its most recent widening.

These findings hold for the Rose Quarter Project as well. Key project staff have publicly conceded that the project will not produce significant improvements in regular, daily traffic congestion, which engineers refer to as “recurring congestion.”

Induced demand is firmly established science

It is well established in the scientific literature that increased roadway capacity generates additional vehicle travel. The definitive work by Duranton and Turner estimates that there is, in the long run, a unit elasticity of miles traveled with respect to road capacity, i.e. each 1 percent increase in road capacity generates a 1 percent increase in vehicle miles traveled:

This paper analyzes new data describing city-level traffic in the continental US between 1983 and 2003. Our estimates of the elasticity of MSA interstate highway VKT with respect to lane kilometers are 0.86 in OLS, 1.00 in first difference, and 1.03 with IV. Because our instruments provide a plausible source of exogenous variation, we regard 1.03 as the most defensible estimate. We take this as a confirmation of the “fundamental law of highway congestion” suggested by Downs (1962), where the extension of interstate highways is met with a proportional increase in traffic for US MSAs.

More recently, Hymel (2019) has independently reached a nearly identical conclusion.  His analysis concludes:

These findings offer persuasive evidence supporting the fundamental law of traffic congestion, and indicate that capacity expansion is not a viable long-term solution to urban traffic congestion. Across specifications of the dynamic model that controlled for endogenous lane-mileage and state fixed effects, the within-group estimator generated long-run induced demand elasticities ranging from 0.892 and 1.063, all with very small standard errors. . . . Furthermore, results from the dynamic model suggest that after five years, induced vehicle travel is expected to grow to 90% of its equilibrium level, quickly decreasing traffic speeds on the new roadway capacity.”

More comprehensive and independent reviews of the literature on induced demand have reached essentially the opposite conclusion from that asserted in the EA. These reviews include: Avin, U., R. Cervero, et al. (2007), Litman, (2007) and Williams-Derry, C. (2007), and Handy & Boarnet (2014). I

Whether development is consistent with local land use plans or not bears no necessary relationship to whether there is induced demand. Many different levels of development (from vacant to fully allowed density with variances) are possible under any local land use plan. Asserting that the level of development is “consistent” with land use plans is a straightforward evasion of the requirement to consider the impacts of induced demand. This is simply irrelevant to determining whether there may be impacts. Local land use plans only specify the maximum amount of development that may occur in the area influenced by the project. There is a wide range of possible levels and intensities of development that are possible under these land use plans, from no development to the full maximum allowed by law.

The fundamental law of road congestion is so well know that it has long been reflected in administrative guidance for the preparation of environmental reviews of road construction projects.  The Federal Highway Administration guidelines for preparing environmental impact statements clearly instruct the analysis of induced impacts: It specifically anticipates a different analysis for each alternative “substantial, foreseeable, induced development should be presented for each alternative”

1. Environmental Impact Statement (EIS) — FORMAT AND CONTENT

2. Environmental Consequences

 Land Use Impacts

This discussion should identify the current development trends and the State and/or local government plans and policies on land use and growth in the area which will be impacted by the proposed project.

These plans and policies are normally reflected in the area’s comprehensive development plan, and include land use, transportation, public facilities, housing, community services, and other areas.

The land use discussion should assess the consistency of the alternatives with the comprehensive development plans adopted for the area and (if applicable) other plans used in the development of the transportation plan required by Section 134. The secondary social, economic, and environmental impacts of any substantial, foreseeable, induced development should be presented for each alternative, including adverse effects on existing communities. Where possible, the distinction between planned and unplanned growth should be identified.

Federal Highway Administration, U.S. Department of Transportation, TECHNICAL ADVISORY: GUIDANCE FOR PREPARING AND PROCESSING ENVIRONMENTAL AND SECTION 4(F) DOCUMENTS, T 6640.8A
October 30, 1987 (http://www.fhwa.dot.gov/legsregs/directives/techadvs/T664008a.htm)

The FHWA has developed substantial technical resources to illustrate how induced demand can be estimated for projects such as the CRC. For example, DeCourla-Souza and Cohen document long-term demand elasticities of traffic with regard to travel time averaging -0.57 and ranging from -0.2 to -1.0. This means that in the long run, all other things being equal, a 10% reduction in travel time in a corridor would be associated with a 5.7% higher level of traffic. (Patrick DeCorla-Souza and Harry Cohen, Accounting For Induced Travel In Evaluation Of Urban Highway Expansion, 1998.) More recent estimates by Duranton and Turner (2011), and Hymel (2019)  put the long-term elasticity of traffic with respect to capacity at 1.0: an increase in capacity is exactly offset by an increase in travel.

A  review of transportation models used in estimating future demand and project benefits, including the type used in this process, concludes:

“Failure to account for indirect demand effects likely exaggerates the travel-time savings benefits of capacity expansion and ignores the potentially substantial land use shifts that might occur because of the marginal increase in accessibility provided.”
Avin, U., R. Cervero, et al. (2007). Forecasting Indirect Land Use Effects of Transportation Projects. Washington, DC, American Association of State Highway and Transportation Officials (AASHTO) Standing Committee on the Environment. (Page 5).

ODOT’s claims about GHG are false

ODOT has advanced two claims about the project’s potential for reducing greenhouse gases.  It has argued that the project will reduce the number of crashes in the corridor, and thereby lower the amount of greenhouse gases emitted when cars drive slowly. Similarly, but more generally, it has argued that by reducing congestion, the project will raise travel speeds, reduce idling and lower overall greenhouse gases.  Both of these claims have been disproven by independent research.

Non-recurring delay will not be reduced

In addition the Rose Quarter project has no demonstrable real-world evidence that the freeway widening will reduce delays associated with automobile crashes, so called “non-recurring congestion.” Just a few years ago, ODOT widened a nearby stretch of I-5 which carries mostly the same traffic, adding a travel lane and widening shoulders (just as it proposes to do at the Rose Quarter). ODOT’s own crash statistics show that the rate of crashes on this stretch of road not only did not decrease, but actually increased in the years following the freeway widening.

ODOT’s claims that additional lanes and wider shoulders will reduce crashes are based on it’s claim that it used a computer spreadsheet called ISAT to calculate probable crashes (Traffic Technical Report). However, the user manual for the ISATe model says that the model is not applicable to freeway segments that are controlled by ramp meters. (Ramp meters control the flow of traffic onto the roadway and reduce the likelihood of crashes associated with merging). This model is not a valid basis for predicting crashes or changes in the number of crashes because this segment of roadway includes ramp meters. See Bonneson, et al., 2012.

ODOT’s experience with I-5 suggests that widening one bottleneck at one point in the system only speeds and intensifies the process of traffic congestion at other bottlenecks in the system. For example, ODOT has made improvements to I-5 in the area north of Lombard Street, including the freeway widening project described in the previous paragraph). While this has removed some “bottlenecks” in some locations, it has funneled more vehicles, more rapidly into others, with the result that these locations become congested sooner, and actually lose capacity. The I-5 bridges now carry about 10 percent fewer vehicles in the afternoon peak hour than they did 10 and 20 years ago. (“Backfire: How widening freeways can make traffic congestion worse,” February 26, 2019, City Observatory Commentary). Similarly, an ODOT project to increase the capacity of the freeway interchange on I-5 at Woodburn also apparently has resulted in no reduction in crashes, and may actually be associated with an increase in more severe crashes (and attendant delays). See,  “Safety Last: What we’ve learned from ‘improving’ the I-5 freeway,” March 21, 2019, City Observatory Commentary).

Claims that less congestion will reduce idling and lower greenhouse gas emissions have been disproven

Claims that the project will result in less carbon emissions are based on the the discredited theory that smoothing traffic flow and reducing idling results in lower carbon emissions. That claim has been discredited by Bigazzi and Figgliozzi (2010), Williams-Derry (2007), Noland & Quddus (2006).

Also, experience has shown that carbon estimates prepared by the Oregon Department of Transportation are untrustworthy. In 2015, The Director of the Oregon Department of Transportation conceded publicly to the Legislature that ODOT had exaggerated by a factor of more than four the possible carbon emission reductions associated with certain transportation projects.

It doesn’t matter what you call the added lanes

And we don’t buy for a minute that it matters in any way that ODOT wants to call the additional lanes its building “auxiliary lanes”.  If the point is that the right hand lane on I-5 at the Rose Quarter is handling merging traffic, that is true whether the facility is 2 lanes in each direction or three.  If we apply ODOT’s logic and nomenclature to the current setup, the freeway now consists of one through lane and one auxiliary lane–and the proposed project would increase that to two through-lanes and one auxiliary lane. Using sophistry and shifting definitions doesn’t change the fact that this project adds lane miles of freeway. And more lane miles of freeway, as these calculators show, produces millions more miles of driving and thousands of tons more greenhouse gas emissions every year.

References:

Avin, U., R. Cervero, et al. (2007). Forecasting Indirect Land Use Effects of Transportation Projects. Washington, DC, American Association of State Highway and Transportation Officials (AASHTO) Standing Committee on the Environment.

Bonneson, J., Pratt, M., and Geedipally, S., (et al), Enhanced Interchange Safety Analysis Tool: User Manual, National Cooperative Highway Research Program, Project 17-45, Enhanced Safety Prediction Methodology and Analysis Tool for Freeways and Interchanges, May 2012.

Bigazzi, A. and Figliozzi, M., 2010, An Analysis of the Relative Efficiency of Freeway Congestion as an Emissions Reduction Strategy.

DeCorla-Souza, P. and H. Cohen (1998). Accounting For Induced Travel In Evaluation Of Urban Highway Expansion. Washington, Federal Highway Administration.

Duranton, G., & Turner, M. A. (2011). The fundamental law of road congestion: Evidence from US cities. American Economic Review, 101(6), 2616-52.

Federal Highway Administration, U.S. Department of Transportation, TECHNICAL ADVISORY: GUIDANCE FOR PREPARING AND PROCESSING ENVIRONMENTAL AND SECTION 4(F) DOCUMENTS, T 6640.8A
October 30, 1987 (http://www.fhwa.dot.gov/legsregs/directives/techadvs/T664008a.htm)

Handy, S., & Boarnet, M. G. (2014). Impact of Highway Capacity and Induced Travel on Passenger Vehicle Use and Greenhouse Gas Emissions. California Environmental Protection Agency, Air Resources Board.             https://www.arb.ca.gov/cc/sb375/policies/hwycapacity/highway_capacity_bkgd.pdf

Hymel, K. (2019). If you build it, they will drive: Measuring induced demand for vehicle travel in urban areas. Transport policy, 76, 57-66.

Kneebone, E., & Holmes, N. (2015). The growing distance between people and jobs in metropolitan America. Washington, DC: Brookings Institution, Metropolitan Policy Program.  https://www.brookings.edu/wp-content/uploads/2016/07/Srvy_JobsProximity.pdf

Litman, T. (2019). Generated Traffic and Induced Travel Implications for Transport Planning. Victoria, BC, Victoria Transport Policy Institute.

Marshall, N. L. (2018). Forecasting the impossible: The status quo of estimating traffic flows with static traffic assignment and the future of dynamic traffic assignment. Research in Transportation Business & Management. https://www.sciencedirect.com/science/article/pii/S2210539517301232?via%3Dihub

Noland, R. B., & Quddus, M. A. (2006). Flow improvements and vehicle emissions: effects of trip generation and emission control technology. Transportation Research Part D: Transport and Environment, 11(1), 1-14.

Parsons Brinckerhoff, Land Use-Transportation Literature Review for the I-5 Trade Corridor Regional Land Use Committee, September 17, 2001. Pages 4-5 http://nepa.fhwa.dot.gov/ReNEPA/ReNepa.nsf/All+Documents/CCECF4D789DB510E85256CE6006142A0/$FILE/land_use_literature_review.pdf

Williams-Derry, C. (2007). Increases in greenhouse-gas emissions from highway-widening projects. Seattle, Sightline Institute.