Constraining Ammonia in Air Quality Models with Remote Sensing Observations and Inverse Modeling

Ammonia is an important species in the atmosphere as it contributes to air pollution, climate change and environmental health. Ammonia emissions are known to be primarily from agricultural sources, however there is persistent uncertainty in the magnitudes and seasonal trends of these sources, as ammonia has not traditionally been routinely monitored. The first detection of boundary layer ammonia from space by the NASA Tropospheric Emissions Spectrometer (TES) satellite has provided an exciting new means of reducing this uncertainty. In this thesis, I explore how forward and inverse modeling can be used with satellite observations to constrain ammonia emissions. Model simulations are used to build and validate the TES ammonia retrieval product. TES retrievals are then used to characterize global ammonia distributions and model estimates. Correlations between ammonia and carbon monoxide, observed simultaneously by TES, provide additional insight into observed and modeled ammonia from biomass burning. Next, through inverse modeling, I show that ammonia emissions are broadly underestimated throughout the U.S., particularly in the West. Optimized model simulations capture the range and variability of in-situ observation in April and October, while estimates in July are biased high. To understand these adjustments, several aspects of the retrieval are considered, such as spatial and temporal sampling biases. These investigations lead to revisions of fundamental aspects of how ammonia emissions are modeled, such as the diurnal variability of livestock ammonia emissions. While this improves comparison to hourly in situ measurements in the SE U.S., ammonia concentrations decrease throughout the globe, up to 17 ppb in India and Southeastern China. Lastly, the bi-directional air-surface exchange of ammonia is implemented for the first time in a global model and its adjoint. Ammonia bi-directional exchange generally increases ammonia gross emissions (10.9%) and surface concentrations (up to 3.9 ppb) throughout the globe in July, except in India and Southeastern China. It decreases ammonia gross emissions in the Northern Hemisphere (e.g., 42.5% in April in China) and increases ammonia in the Southern Hemisphere in April and October. While bi-directional exchange is fundamentally a better representation of ammonia emissions from fertilizers, emissions from primary sources are still likely underestimated