The processes that form, transport and recycle the reactive forms of nitrogen in the atmosphere are important in global air quality, atmospheric nutrient deposition, and for controlling the long term oxidation capacity of the atmosphere. The NOx gases (NOx = NO + NO2) are also important greenhouse gases. The ultimate goal being pursued in this modeling study is the understanding of past (polar) atmospheric composition, and through a better accounting of the variability of contemporary atmospheric reactive nitrogen compounds as may be discerned from measurements of nitrate concentrations retained in snow, firn and ice cores on the Antarctic continent. The approach will use a inverse chemical transfer model (adjoint GEOS-Chem) to test the sensitivity of nitrate deposition to Antarctic ice from various possible sources. By integrating a snowpack radiative transfer subroutine into the chemical transport model, comparisons of calculated with observed fluxes of snowpack NOx at some observed inland and coastal Antarctic sites are to be used to parameterize photdenitrification rates. An additional question to be investigated with the developed modeling tools is what fraction of the nitrate concentration undergoing photolysis at the snow surface is recycled back to the atmosphere, or else preserved in the accreting ice.
Nitrogen oxides NOx (NOx ≡ NO + NO2) are associated with important atmospheric environmental issues such as aerosol concentrations and the oxidation capacity of atmosphere. Given their importance for atmospheric chemistry, there is interest in understanding past variability of atmospheric NOx. Ice cores from polar regions can be utilized to reconstruct past variability from chronological records of chemical species spanning more than 120,000 years. However, interpretation of these ice cores is hindered by understanding the factors influencing the record of nitrate, which forms from photochemical conversion of NOx to nitric acid (HNO3) and nitrate aerosol (NO3-). Several factors have been proposed as influencing the Antarctic nitrate concentrations, from stratospheric intrusions to solar activity to transport of nitrate and nitrate precursors from species emitted into the troposphere, though it has yet to be conclusively determined which process dominates. To lean more about the mechanisms governing nitrate in Antarctica, we use a model of atmospheric transport and chemistry (GEOS-Chem) to simulate the formation and transport of nitrate to Antarctica. We then apply the adjoint of this model to attribute the amount of simulated nitrate in Antarctica to various sources. Adjoint modeling provides a unique perspective in that the amount of nitrate in the model in Antarctica can be partitioned into influences from individual surface emissions, production and loss rate from the stratosphere, and the chemical reactions governing the transportation of nitrate from NOx. This helps us understand the processes which control levels of nitrate in Antarctica. We found that surface emissions alone can explain in part the seasonality of observed nitrate in Antarctica. Of all surface emissions, NOx emissions from anthropogenic sources in South America, South Africa, Australia, and New Zealand have the largest influence on HNO3 (and hence total nitrate) levels in most seasons. The impact of ammonia (NH3) emissions on HNO3 over Antarctica was investigated due to its impacts on the partitioning of HNO3 to aerosol nitrate. In the season with high total nitrate (May – September in the model), nitrate had a significant positive sensitivity with respect to NH3 emissions, indicating that transport of nitrate itself was driving its Antarctic burden. Production of NOx and HNO3 at higher altitudes in the atmosphere (the stratosphere) from natural processes contributes comparable amounts. Our model estimates are likely a lower bound on the importance of stratospheric sources, as contributions to the Antarctic surface from the settling of frozen nitrate on stratospheric clouds (PSCs) is thought to be an important factor which was not yet included in our modeling analysis.
