Making MOVES Move: Fast Emissions Estimates for Repeated Transportation Policy Scenario Analyses

Efforts to reduce US emissions from transportation rely greatly on the Environmental Protection Agency’s MOVES software to estimate pollutant emissions, a high-precision, computationally intensive tool with many required inputs. To compare numerous scenarios at scale, decision-makers need a faster, user-friendly, and credible emissions estimator. This study introduces MOVESLite, a framework for making fast, highly accurate predictions (with respect to MOVES) of emissions tailored to a user's county, with a limited, more user-friendly number of inputs. MOVESLite uses county- and pollutant-specific models trained on previous MOVES estimates to estimate transportation emissions immediately. While MOVES may take minutes-to-hours to estimate one county-year’s emissions, MOVESLite takes approximately 1 millisecond, with acceptable accuracy. We test MOVESLite performance, comparing thousands of possible models’ accuracy for random samples of US counties, and we implement a case study of public transportation planning in Tompkins County, NY. MOVESLite aims to help decision-makers compare and select effective climate action policies.