CMAQ + SPECIATE

Fine-particle composition sensitivity analysis using EPA SPECIATE, S2S, and CMAQ v5.5.

Project Overview

This project was designed to develop an end-to-end understanding of the EPA SPECIATE database and the Community Multiscale Air Quality (CMAQ) modeling workflow. I selected the SPECIATE profile for gas-fired combined-cycle and cogeneration plants, converted it into CMAQ-compatible model species with the Species-to-Species (S2S) Tool, configured CMAQ v5.5 on an AWS EC2 instance, and conducted a controlled chemical-composition sensitivity experiment.

Research Question

If PT_EGU fine-particle emissions retain the same source-by-source and hour-by-hour mass but use the EPA SPECIATE profile for gas-fired combined-cycle and cogeneration plants, how does CMAQ's predicted aerosol composition change?

Workflow

  1. Review SPECIATE and CMAQ documentation, terminology, tutorials, and benchmark workflows.
  2. Select SPECIATE v5.4 profile 5671 for gas-fired combined-cycle and cogeneration plants.
  3. Use the S2S Tool and PM-CR2 mapping rules to translate the measured profile into 19 CMAQ-compatible CRACMM2 particulate species.
  4. Compile and run the CMAQ v5.5 EPA benchmark in a Linux environment on AWS EC2.
  5. Apply the new composition to all PT_EGU fine-particle records and compare the matched experiment with the original benchmark.

Model Configuration

ModelCMAQ v5.5
Emissions profileEPA SPECIATE v5.4, profile 5671
Profile processingS2S Tool with PM-CR2 mapping rules
ChemistryCRACMM2
DepositionSTAGE
Domain2018_12NE3, 100 × 105 cells at 12 km resolution
SimulationJuly 1, 2018; 24 hourly records; 35 vertical layers
ComputingUbuntu Linux, OpenMPI, NetCDF, I/O API, and AWS EC2

Controlled Composition Experiment

The experiment changed only the chemical composition of PT_EGU fine-particle emissions. Total fine-particle mass, hourly emission patterns, source locations, stack information, meteorology, land-surface data, and initial and boundary conditions were held constant. Both inventories contained approximately 378.9 metric tonnes per day of fine-particle emissions; coarse-particle emissions were not modified.

Stacked bars comparing the chemical composition of PT_EGU fine-particle emissions in the EPA benchmark and the gas-fired composition experiment
The total PT_EGU fine-particle emission mass remained approximately 378.9 metric tonnes per day, while its chemical composition was redistributed using the gas-fired combined-cycle and cogeneration profile.

Total PM₂.₅ Response

Domain-average total PM₂.₅ remained essentially unchanged, but individual grid cells showed small spatial responses to the revised composition. Daily-average experiment-minus-baseline differences ranged from approximately −0.133 to +0.019 µg/m³, demonstrating that mass-preserving emissions changes can still alter modeled atmospheric concentrations locally.

Maps comparing baseline and experimental total surface PM2.5 with their spatial difference
Daily-average total surface PM₂.₅ for the original benchmark, the composition experiment, and the experiment-minus-baseline difference on July 1, 2018.

Organic-Aerosol Response

The five selected primary-organic aerosol groups increased by approximately 20% in the combined domain-and-hour mean. Their spatial response was much more visible than the change in total PM₂.₅ because the experiment directly redistributed the emitted fine-particle composition toward these organic groups.

Maps comparing baseline and experimental selected organic-aerosol concentrations with their spatial difference
Daily-average surface concentrations of the five selected organic-aerosol groups for the baseline, composition experiment, and experiment-minus-baseline difference.

Key Interpretation