Evaluating commercial marine emissions and their role in air quality policy using observations and the CMAQ model

We investigate the representation of emissions from the largest (Class 3) commercial marine vessels (c3 Marine) within the Community Multiscale Air Quality (CMAQ) model. In present emissions inventories developed by the United States Environmental Protection Agency (EPA), c3 Marine emissions are divided into off-shore and nearshore files. Off-shore c3 Marine emissions are vertically distributed within the atmospheric column, reflecting stack-height and plume rise. Near-shore c3 Marine emissions, located close to the US shoreline, are erroneously assumed to occur only at the surface. We adjust the near-shore c3 Marine emissions inventory by vertically distributing these emissions to be consistent with the off-shore c3 Marine inventory. Additionally, we remove near-shore c3 Marine emissions that overlap with off-shore c3 Marine emissions within the EPA files. The CMAQ model generally overestimates surface ozone (O3) compared to Air Quality System (AQS) site observations, with the largest discrepancies occurring near coastal waterways. We compare modeled O3 from two CMAQ simulations for June, July, and August (JJA) 2011 to surface O3 observations from AQS sites to examine the efficacy of the c3 Marine emissions improvements. Model results at AQS sites show average maximum 8-hr surface O3 decreases up to ∼6.5 ppb along the Chesapeake Bay, and increases ∼3–4 ppb around Long Island Sound, when the adjusted c3 Marine emissions are used. Along with the c3 Marine emissions adjustments, we reduce on-road mobile NOX emissions by 50%, motivated by work from Anderson et al. 2014, and reduce the lifetime of the alkyl nitrate species group from ∼10 days to ∼1 day based on work by Canty et al. 2015, to develop the “c3 Science” model scenario. Simulations with these adjustments further improve model representation of the atmosphere. We calculate the ratio of column formaldehyde (HCHO) and tropospheric column nitrogen dioxide (NO2) using observations from the Ozone Monitoring Instrument and CMAQ model output to investigate the photochemical O3 production regime (VOC or NOX-limited) of the observed and modeled atmosphere. Compared to the baseline, the c3 Science model scenario more closely simulates the HCHO/NO2 ratio calculated from OMI data. Model simulations for JJA 2018 using the c3 Science scenario show a reduction of surface O3 by as much as ∼13 ppb for areas around the Chesapeake Bay and ∼2–3 ppb at locations in NY and CT downwind of New York City. These reductions are larger in 2018 than in 2011 due to a change in the photochemical O3 production regime in the Long Island Sound region and the projected decline of other (non-c3 Marine) sources of O3 precursors, highlighting the importance of proper representation of c3 Marine emissions in future modeling sce-