Regionalized PM2.5 Community Multiscale Air Quality model performance evaluation across a continuous spatiotemporal...

Area of Science:

• Environmental Science and Engineering
• Atmospheric Chemistry and Physics
• Computational Modeling

Background:

• The Community Multiscale Air Quality (CMAQ) model is crucial for understanding air pollutant sources, concentrations, and regulatory compliance.
• Evaluating CMAQ model performance requires substantial resources and effective methodologies.
• Existing methods often provide a generalized view of model performance across large domains.

Purpose of the Study:

• To introduce and evaluate the Regionalized Air quality Model Performance (RAMP) method for assessing CMAQ model performance.
• To explore novel visualization and error evaluation techniques for daily Particulate Matter ≤ 2.5 micrometers (PM2.5) concentrations.
• To demonstrate the spatial and temporal heterogeneity of CMAQ model performance.

Main Methods:

• Developed the RAMP method for non-homogenous, non-linear, and non-homoscedastic model performance evaluation at each CMAQ grid.
• Applied RAMP to a well-documented 2001 regulatory episode for PM2.5 concentrations across the continental United States.
• Compared RAMP's error correction performance against other model evaluation methods.

Main Results:

• CMAQ model performance was found to be non-homogeneous across space and time.
• The RAMP correction of systematic errors resulted in a 22.1% reduction in Mean Square Error compared to a constant domain-wide correction.
• RAMP accurately reproduced simulated performance with a correlation of r = 76.1%, indicating that most CMAQ error is random, not systematic.

Conclusions:

• The RAMP method offers a superior approach to evaluating and correcting systematic errors in air quality models like CMAQ.
• High systematic and random error areas are collocated, suggesting shared underlying sources.
• Addressing systematic errors using RAMP can concurrently mitigate random errors, improving overall model accuracy for PM2.5.