Articles | Volume 10, issue 12
Geosci. Model Dev., 10, 4743–4758, 2017
Geosci. Model Dev., 10, 4743–4758, 2017

Development and technical paper 22 Dec 2017

Development and technical paper | 22 Dec 2017

A case study of aerosol data assimilation with the Community Multi-scale Air Quality Model over the contiguous United States using 3D-Var and optimal interpolation methods

Youhua Tang1,2, Mariusz Pagowski4,5, Tianfeng Chai1,2, Li Pan1,2, Pius Lee1, Barry Baker1,2, Rajesh Kumar6, Luca Delle Monache6, Daniel Tong1,2,3, and Hyun-Cheol Kim1,2 Youhua Tang et al.
  • 1NOAA Air Resources Laboratory, College Park, MD, USA
  • 2Cooperative Institute for Climate and Satellites, University of Maryland, College Park, MD, USA
  • 3Center for Spatial Information Science and Systems, George Mason University, Fairfax, VA, USA
  • 4NOAA Earth System Research Laboratory, Boulder, CO, USA
  • 5Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder, CO, USA
  • 6National Center for Atmospheric Research, Boulder, CO, USA

Abstract. This study applies the Gridpoint Statistical Interpolation (GSI) 3D-Var assimilation tool originally developed by the National Centers for Environmental Prediction (NCEP), to improve surface PM2.5 predictions over the contiguous United States (CONUS) by assimilating aerosol optical depth (AOD) and surface PM2.5 in version 5.1 of the Community Multi-scale Air Quality (CMAQ) modeling system. An optimal interpolation (OI) method implemented earlier (Tang et al., 2015) for the CMAQ modeling system is also tested for the same period (July 2011) over the same CONUS. Both GSI and OI methods assimilate surface PM2.5 observations at 00:00, 06:00, 12:00 and 18:00 UTC, and MODIS AOD at 18:00 UTC. The assimilations of observations using both GSI and OI generally help reduce the prediction biases and improve correlation between model predictions and observations. In the GSI experiments, assimilation of surface PM2.5 (particle matter with diameter < 2.5 µm) leads to stronger increments in surface PM2.5 compared to its MODIS AOD assimilation at the 550 nm wavelength. In contrast, we find a stronger OI impact of the MODIS AOD on surface aerosols at 18:00 UTC compared to the surface PM2.5 OI method. GSI produces smoother result and yields overall better correlation coefficient and root mean squared error (RMSE). It should be noted that the 3D-Var and OI methods used here have several big differences besides the data assimilation schemes. For instance, the OI uses relatively big model uncertainties, which helps yield smaller mean biases, but sometimes causes the RMSE to increase. We also examine and discuss the sensitivity of the assimilation experiments' results to the AOD forward operators.

Short summary
In order to evaluate the data assimilation tools for regional real-time PM2.5 forecasts, we applied a 3D-Var assimilation tool to adjust the aerosol initial condition by assimilating satellite-retrieved aerosol optical depth and surface PM2.5 observations for a regional air quality model, which is compared to another assimilation method, optimal interpolation. We discuss the pros and cons of these two assimilation methods based on the comparison of their 1-month four-cycles-per-day runs.