Dear Dr. Rao,
I am director of the Air Quality Modeling Center within the U.S. Environmental Protection Agency (EPA) Office of Transportation and Air Quality. The U.S. EPA has concerns that the paper by Ritner et al. published in volume 63, issue 6, is critical of certain aspects of the MOVES emissions model. The agency was not consulted by the authors during the initial research and was not included in JA&WMA’s prepublication peer review panel. The U.S. EPA believes that several key statements in the paper have the potential to cause unnecessary confusion among our stakeholder groups.
The following is a brief statement addressing the agency's concerns.
EPA would like to emphasize that the issues pointed out in the paper relating to MOVES generating “erratic EFs when longer links are broken into smaller sublinks” are largely theoretical and do not affect current PM or CO hot-spot analyses. EPA provides technical recommendations concerning how practitioners should define links given a variety of traffic activity data, including local drive cycles. EPA believes that entering a project-specific drive cycle, even a very short cycle, will better represent local conditions than relying on the MOVES default drive cycles (average speed option) for intersection projects. Additionally, the issues identified in the paper are limited to the use of local drive cycles in the MOVES project scale only; the county and national scales are unaffected. Furthermore, emissions generated using the MOVES project scale with default drive cycles (i.e., the average speed option) or project-specific operating mode distributions will not be affected by link length definitions.
The MOVES model was designed to quantify emissions at multiple geographic scales: national, county, and project scale. At all scales of analysis, a fundamental design feature within the model is the mapping of drive cycles to appropriate operating mode distributions. This allows the use of detailed emission rate information internally defined by operating mode bin. The process relies on an algorithm to convert second-by-second speed trajectories into vehicle specific power (VSP), such that each second is assigned an operating mode representing acceleration, cruise, deceleration, or braking, depending on the speed time series (e.g., the acceleration is the difference between that second (n), and previous second (n – 1)). A consequence of these calculations at the project scale is that the first second of a drive cycle will have an unknown acceleration rate, given that the previous second is also unknown. A design decision was made to assign zero acceleration to the first second. Over longer drive cycles, this choice has an insignificant impact on emissions. However, as is highlighted in the paper, dividing intersections into many “sublinks” with accompanying short drive cycles will result in lower emission rates than if the same intersection is described in terms of longer links. When drive cycles are very short, the consequence of MOVES assigning zero acceleration to the first second has an impact on the subsequent operating mode distribution (and resulting emission rates). This effect is increased when links defined as exclusively acceleration and deceleration are further subdivided, as was done in this research. Thus, the authors are seeing lower emission rates as the acceleration and deceleration links are divided into smaller segments.
EPA may consider improving the VSP algorithms to better accommodate very short drive cycles in a future version of MOVES. Any questions on the MOVES model or supporting guidance documents should be directed to [email protected].
Sincerely,
Edward Nam, Ph.D.
Director, Air Quality Modeling Center
Office of Transportation and Air Quality
U.S. Environmental Protection Agency