Environmental Engineering Reference
In-Depth Information
1.14 Comparing Models/Methods for Estimating
Multi-pollutant Fine-Scale Air Quality Concentrations
Karen Wesson
1
, Kirk Baker
1
, Uarporn Nopmongcol
2
, Greg Yarwood
2
,
Tanarit Sakulyanontvittaya
2
, Madeleine Strum
1
, James Thurman
1
, Louise
Camalier
1
, Darrell Ensley
3
, Brian Timin
1
, Sharon Phillips
1
, and Tyler Fox
1
1
U.S. Environmental Protection Agency, RTP, NC, USA
2
ENVIRON International Corporation, Novato, CA, USA
3
Computer Science Corporation, RTP, NC, USA
Abstract
Photochemical models are commonly used in regulatory and policy
assessments to estimate pollutant concentrations and deposition of both inert and
chemically reactive pollutants over large spatial scales. These models are generally
run for horizontal grid resolutions of 36 and 12 km. However, several recent
assessments have revealed the need for air quality predictions at resolutions finer
than 12 km to resolve important local-scale gradients in pollutant concentrations.
Given this need, we are undertaking a study to investigate several methods that
can be used to obtain local-scale air quality concentrations. This study looks at the
application and evaluation of a variety of models including CMAQ, CAMx and
AERMOD. In addition, we will also evaluate the use of a new method called the
Multiplicative Approach to the Hybrid Method (MAHM), which combines CMAQ
and AERMOD predicted concentrations to generate local-scale air quality predic-
tions. To do this, these models/methods are applied at ≤ 4 km resolution for both a
winter and summer month in the same local area: Detroit, MI. The study looks at
model/method performance of PM
2.5
, O
3
, and several toxic pollutants by comparing
modeled versus ambient measured concentrations. Resources for implementation
of each model/method are also evaluated.
1. Modeling Applications
All model runs were performed for a summer month (July) and a winter month
(January) in 2002. The multi-pollutant versions of CMAQ v4.6.1i (Byun et al.,
1999) and CAMx v4.5 (Environ, 2008) were both run for three domains centered
on the Detroit urban core: a large Midwest domain at 12 km horizontal resolution
(114 × 117 grid cells), a 4 km domain (36 × 45 grid cells) and 1 km domain (72 ×
108 grid cells). AERMOD was run for PM
2.5
and several toxics with a receptor
do-main that was slightly smaller than the 1 km photochemical domain and was
also centered on the Detroit urban core. The emissions data for these runs were
