AAAR 36th Annual Conference October 16 - October 20, 2017 Raleigh Convention Center Raleigh, North Carolina, USA
Abstract View
Evaluation of pH Biases in Chemical Transport Models and Their Role on Nitrate Substitution
PETROS VASILAKOS, Armistead G. Russell, Athanasios Nenes, Georgia Institute of Technology
Abstract Number: 532 Working Group: Regional and Global Air Quality and Climate Modeling
Abstract Throughout the decade of 2001 to 2011 strict regulations on SO2, coupled with increasing ammonia concentrations, have been hypothesized to lead to increasingly alkaline aerosol, and the subsequent replacement of sulfate with nitrate. Increased alkalinity also has important implications for the formation of Secondary Organic Aerosol (SOA), since many production pathways are contingent on low pH conditions. Long-term observations from the SEARCH network however do not validate this hypothesis (Blanchard et al. 2012). On the contrary, aerosol has been found to remain strongly acidic, while nitrate is remaining relatively constant with a modest decrease in concentration over the SE US. Chemical transport models (CTMs) used for future policy-making are seldom evaluated on their ability to predict pH, and given the important role that aerosol acidity plays on nitrate partitioning and SOA formation, predictive biases can lead to incorrect estimations of aerosol composition.
In order to investigate existing biases of CTMs and the “nitrate substitution paradox”, we hypothesize that models overpredict aerosol pH, with an increasing positive bias as SO2 emissions are decreased in the simulations. We test this hypothesis with the Community Multiscale Air Quality (CMAQ) model, by evaluating the model pH trends over the US between 2001 and 2011, as well as the sources of potential biases, such as the relative humidity (RH), temperature (T) and the concentration of crustal elements, and comparing them to the thermodynamic analysis of ambient data from the SEARCH network throughout all season and sites, with the ISOROPIA II model. While CMAQ predictions of RH, T and PM2.5 pH during 2001 and 2011 compare favorably with observations from SEARCH sites, there is a clear decadal positive trend, with pH over the Eastern US increasing by 1 unit within the decade, something that is not observed in the data. In addition, non-volatile cations are identified as the reason for this inconsistent trend, since they are internally mixed in modelled PM2.5 at concentrations that increasingly influence aerosol pH as sulfate levels decrease. This resulting bias tends to be higher during the late afternoon through early morning for all the sites, and coincides with the times where the maximum of cation concentrations and RH are simulated.
We estimate that a predictive positive pH bias of 1 unit can result in a positive aerosol nitrate bias of 1-2 μg m-3. Possible pH biases, especially in future prognostic studies, can reaffirm the otherwise incorrect expectation of “nitrate substitution”. Therefore, evaluation of simulated aerosol pH against observations is a vital, but neglected, aspect of model evaluation for robust emissions policy.