What are the direct impacts of fires on air quality and aerosol transport?

WRF-chem simulation of PM and Ozone from fires (UCLA)

We will use the WRF-Chem model - WRF (Skamarock et al., 2008) coupled with online chemistry and aerosols (Grell et al. 2005; Fast et al. 2006) to simulate the air quality effects of Southern California wildfires. Our focus will be on PM, particularly PM2.5 - the dominant size range for wildfire-generated particles (Reid et al. 2005), and ozone in the LA Basin. We will leverage off two ongoing projects involving Co- Li, funded by NOAA and the Califronia Air Resource Board (ARB), to constrain trace gas emissions in the LA Basin using ground-based remote sensing observations and WRF-Chem. The same domain for WRF climate reconstruction/prediction (section 3.4) will be used for WRF-Chem simulations. For the WRF- Chem simulations, the meteorological fields will be nudged toward the WRF reconstructed/predicated fields. The GEOS- Chem model (Bey et al. 2001) will be used to provide chemical boundary conditions for the outer domain. We will use the CBM-Z photochemical mechanism (Zaveri and Peters 1999) among the gas-phase chemistry options in WRF-Chem, from EPA AirData. which includes emissions or oxidation of biogenic monoterpenes, known to be a significant source of secondary organic aerosols (SOAs), as well as more efficient SOA formation from anthropogenic VOCs. The MOSAIC sectional aerosol scheme (Fast et al. 2006) will be chosen to predicts not only aerosol mass but also size and number information for organic carbon, elemental carbon, sulfate, nitrate, ammonium, and other (unspecified) inorganics, among other species. We will use the EPA's 2005 National Emissions Inventory (NEI), with updates for California from the ARB. Fire emissions’ impacts are currently included in some air quality forecasts but only as averages of historical fire events to represent typical fire emissions (Hu et al. 2008; references therein). Air quality models generally show good statistical skills for ozone prediction but have difficulties predicting PM (McKeen et al., 2009). Accurate prediction of PM concentrations is challenging because of the poor understanding of the sources and formation of PM (Goldstein et al. 2007; Kanakidou et al. 2005). In particular, SOA, often the single largest component of fine PM in the LA Basin (Stone et al. 2009; Docherty et al. 2008), is generally underestimated in photochemical models in urban environments (McKeen et al. 2009; Hu et al. 2008), due likely in part to an underestimate of the anthropogenic SOAs (de Gouw and Jimenez 2009). We will explore a simple parameterization of anthropogenic SOAs recently proposed by Jose Jimenez (Univ. Colorado – Boulder; personal communication). We expect the inclusion of wildfire emissions, of which PM2.5 with mostly organic aerosols dominates, to improve PM predictions. The standard WRF-Chem model does not include wildfire emissions. We will add the wildfire emissions derived as part of the proposed study (section 3.1.4). Additionally, smoke injection height has been shown to be critical for accurately predicting surface PM2.5 concentrations near the source regions (Stein et al. 2008). We will explore using MISR-derived smoke plume height (Val Martin et al. 2009) to prescribe wildfire emissions vertically, as we have demonstrated for boreal forest fire emissions (Chen et al. 2009).