Environmental Engineering Reference
In-Depth Information
8.10 Long-Term One-Atmosphere CMAQ Modeling
in Central California: Model Performance Evaluation
James T. Kelly
1
, Kathleen Fahey
2
, and Ajith Kaduwela
1,3
1
Planning and Technical Support Division, Air Resources Board, California Environmental
Protection Agency, Sacramento, CA, USA
2
Science and Technology Team, Northeast States for Coordinated Air Use Management
(NESCAUM), Boston, MA, USA
3
Department of Land, Air, and Water Resources, University of California at Davis, CA, USA
Abstract
Techniques for evaluating model performance for long-term (i.e., annual)
air quality simulations are needed for use in regulatory modeling applications.
This study applies numerous evaluation techniques to results of a 14-month
simulation in California. Traditional evaluation approaches are shown to be useful
for defining model performance criteria and providing an overview of model
performance. Spectral decomposition methods and evaluations of spatial correlations
can isolate temporal and spatial scales with characteristic model performance.
Mechanistic and chemical-ratio evaluations indicate needs for model process
development. Future work will focus on combining the various evaluation tech-
niques into a single cohesive model evaluation framework.
Keywords
CRPAQS, long-term evaluation, Kolmogorov-Zurbenko filter
1. Introduction
As state regulatory bodies develop implementation plans to reach goals dictated by
the US EPA's Regional Haze Plan and National Ambient Air Quality Standards
(NAAQS) for PM
2.5
and ozone, a need exists for reliable chemical transport
modeling to assess effects of emissions controls on pollutant concentrations. The
focus of such modeling efforts has recently shifted from short-term pollution
episodes to long-term simulations that elucidate seasonal and regional differences
in pollutant concentrations. Long-term modeling is required to construct mutually
beneficial control strategies for pollutants whose concentrations become elevated
under different environmental conditions (e.g., PM
2.5
and ozone). In central
California, the annual PM
2.5
NAAQS is restrictive in the sense that attainment of
the 24-h standard does not ensure attainment of the annual standard. Thus annual
modeling is an imperative for this region. A suite of accepted model performance
