MOVES-Matrix for High-Performance Emission Rate Model Applications

Guensler, Randall; Liu, Haobing; Xu, Xiaodan; Lu, Hongyu; Rodgers, Michael · 2018 · ROSA P / National Center for Sustainable Transportation (NCST) (UTC)

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the computational inefficiency of the U.S. Environmental Protection Agency’s MOtor Vehicle Emission Simulator (MOVES), the regulatory standard for estimating vehicle emissions. While MOVES provides high-resolution, modal emission rates based on instantaneous speed and acceleration, its complex interface and slow processing speed make it impractical for large-scale, dynamic transportation analyses. Running MOVES for every link in a large network with varying fleet compositions and meteorological conditions requires millions of individual model runs, a process that can take years on standard hardware. To resolve this bottleneck while maintaining regulatory compliance, the authors developed MOVES-Matrix, a high-performance lookup system that pre-calculates emission rates for all relevant input variable combinations. The methodology involves running the MOVES model iteratively across a comprehensive set of input parameters, including calendar years, fuel types, inspection/maintenance programs, temperature, humidity, vehicle source types, model years, and operating conditions. The researchers utilized the Partnership for an Advanced Computing Environment (PACE) high-performance computing cluster to execute these runs. For the Atlanta region, the team performed 146,853 MOVES runs to generate on-road emission rate matrices, resulting in a database containing over 90 billion emission rates. The system was later expanded to MOVES-Matrix 2.0 to include off-network emissions such as engine starts, evaporative losses, and truck hoteling. The algorithm allows users to query specific subsets of this multidimensional array and weight them by local vehicle activity data to derive composite emission rates, replicating the exact calculation logic of MOVES without launching the model repeatedly. The study demonstrates that MOVES-Matrix achieves a 200-fold increase in speed compared to the batch mode of the MOVES graphical user interface while producing identical emission results. Verification tests, including an Atlanta case study using regional travel demand model data, confirmed that the matrix approach yields exactly the same emissions outputs as direct MOVES runs for both on-road and off-network sources. The generation of the matrix itself is efficient, taking approximately four days on the PACE cluster for standard temperature intervals. The system supports various input modes, including operating mode distributions, average speed/facility type, and second-by-second driving schedules, enabling seamless integration with traffic simulation models and big data sources. The significance of this work lies in enabling high-resolution, real-time emissions modeling for large, dynamic transportation systems. By decoupling the computationally intensive emission rate calculations from the application phase, MOVES-Matrix facilitates the assessment of environmental impacts for complex networks, near-road dispersion modeling, and emission inventory development. This approach allows researchers and planners to leverage high-resolution vehicle activity data from sources like smartphone GPS and traffic simulations without being constrained by the computational limits of the regulatory model, thereby advancing the precision and scope of transportation sustainability research.

Key finding

MOVES-Matrix emission rate generation is 200 times faster than using the batch mode of the MOVES graphic user interface and predicts exactly the same emissions results.

Methodology

modeling

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• exposure measurement