Contributions of natural and anthropogenic sources to ambient ammonia in the Athabasca Oil Sands and north-western Canada

Atmospheric ammonia (NH3) is a short-lived pollutant that plays an important role in aerosol chemistry and nitrogen deposition. Dominant NH3 emissions are from agriculture and forest fires, both of which are increasing globally. Even remote regions with relatively low ambient NH3 concentrations, such as northern Alberta and Saskatchewan in northern Canada, may be of interest because of industrial oil sands emissions and a sensitive ecological system. A previous attempt to model NH3 in the region showed a substantial negative bias compared to satellite and aircraft observations. Known missing sources of NH3 in the model were re-emission of NH3 from plants and soils (bidirectional flux) and forest fire emissions, but the relative impact of these sources on NH3 concentrations was unknown. Here we have used a research version of the high-resolution air quality forecasting model, GEM-MACH, to quantify the relative impacts of semi-natural (bidirectional flux of NH3 and forest fire emissions) and direct anthropogenic (oil sand operations, combustion of fossil fuels, and agriculture) sources on ammonia volume mixing ratios, both at the surface and aloft, with a focus on the Athabasca Oil Sands region during a measurement-intensive campaign in the summer of 2013. The addition of fires and bidirectional flux to GEM-MACH has improved the model bias, slope, and correlation coefficients relative to ground, aircraft, and satellite NH3 measurements significantly. By running the GEM-MACH-Bidi model in three configurations and calculating their differences, we find that averaged over Alberta and Saskatchewan during this time period an average of 23.1% of surface NH3 came from direct anthropogenic sources, 56.6% (or 1.24 ppbv) from bidirectional flux (re-emission from plants and soils), and 20.3% (or 0.42 ppbv) from forest fires. In the NH3 total column, an average of 19.5% came from direct anthropogenic sources, 50.0% from bidirectional flux, and 30.5% from forest fires. The addition of bidirectional flux and fire emissions caused the overall average net deposition of NHx across the domain to be increased by 24.5 %. Note that forest fires are very episodic and their contributions will vary significantly for different time periods and regions. This study is the first use of the bidirectional flux scheme in GEM-MACH, which could be generalized for other volatile or semi-volatile species. It is also the first time CrIS (Cross-track Infrared Sounder) satellite observations of NH3 have been used for model evaluation, and the first use of fire emissions in GEM-MACH at 2.5 km resolution.