Investigating emissions and evolution of reactive nitrogen in western U.S. wildfire smoke plumes
Date
2020
Authors
Lindaas, Jakob, author
Fischer, Emily, advisor
Ravishankara, A. R., committee member
Collett, Jeffrey, Jr., committee member
Jathar, Shantanu, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Wildfires are an important source of reactive nitrogen (Nr) to the atmosphere. Large wildfires are becoming more frequent in the western U.S. and smoke from large western U.S. wildfires is becoming a proportionately larger driver of poor air quality in certain U.S. regions. Nr in smoke contributes to the production of ozone (O3), the formation of secondary inorganic and organic aerosol, and nitrogen deposition to downwind ecosystems, with attendant negative impacts on human and ecosystem health and important consequences for the earth's radiative budget. Smoke from wildfires is difficult to sample though, since it is hard to predict exactly where fires will start and which fires will grow larger across the western U.S. Scientific aircraft are able to sample smoke where and when it occurs, making them ideal platforms from which to gather observations of gases and particles in smoke. This dissertation presents results from three analyses investigating the emissions and evolution of reactive nitrogen in wildfire smoke using data from the Western wildfires Experiment on Cloud chemistry, Aerosol absorption, and Nitrogen (WE-CAN). The WE-CAN field intensive sampled fresh smoke from 23 identified large wildfires in the western U.S. during August and September 2018. Pseudo-Lagrangian ("lawn-mower pattern") sampling was accomplished for about half of these fires, meaning that the plane was able to intercept the same smoke multiple times as it was transported downwind. Additionally, mixed and older smoke from indeterminate sources was sampled in the free troposphere as well as in the California Central Valley where it mixed with anthropogenic emissions. These data represent a large increase in the number of western wildfire smoke plumes intercepted by research aircraft in a systematic fashion. First, I present a general overview of Nr emissions from 16 wildfires using a large cross section of the suite of chemical and physical measurements made onboard the National Science Foundation/National Center for Atmospheric Research (NSF/NCAR) C-130 aircraft. I find that reduced N compounds generally make up more than half of the total measured Nr (∑Nr) (39 - 80%, median = 66%). This was not necessarily expected since we sampled plumes from fires at their most active time of day, with assumed burning conditions that favored emission of oxidized forms of Nr. I observe evidence of rapid chemistry in the minutes between emission and the first sampling periods in all fires via significant abundances of peroxyacetic nitric anhydride (PAN) even in these youngest plume samples. I find evidence for the influence of both combustion conditions and fuel N content influencing the ratio of the sum of measured ammonia (NH3) plus particulate ammonium (pNH4) and the sum of oxidized nitrogen (∑NHx/∑NOy). Finally, estimated emission factors (EF) from these fires can be compared with previous literature for similar fuel types, with NH3 EFs similar to or higher than previous lab and field observations, and NOx EFs generally lower than previous estimates. Next I explore the evolution of NH3 in fresh and aged smoke. Focusing on 8 pseudo-Lagrangian sampled plumes, I observe e-folding loss timescales for NH3 with respect to gas-particle partitioning on the order of minutes to hours, similar to previous estimates in fresh smoke. I find empirical evidence for the association of NH3 and nitric acid (HNO3) to form ammonium nitrate (NH4NO3), though not all plumes contain conditions favorable to NH4NO3 formation. Fresh, dense plumes injected at higher altitudes (and lower temperatures) are more likely to favor NH4NO3 formation, a conclusion consistent with previous model simulations in the literature. Measured organic acid ions also suggest the presence of NH4-organic salts in the plumes sampled. Finally, I also use observations collected in medium and old aged smoke to find additional evidence for the formation of NH4NO3 in plumes injected higher in the atmosphere with larger oxidized N to NH3 ratios. Lastly, I investigate the production of PAN and peroxypropionic nitric anhydride (PPN) in the same set of pseudo-Lagrangian sampled plumes. Increases in dilution-corrected PAN and PPN mixing ratios in all plumes suggest that PAN and PPN are produced in all smoke plumes sampled, with the rate of increase similar to the handful of previous observations. I then use a simple observation-based model to determine the dominant precursors of PAN and PPN in fresh smoke plumes. From the model I infer that acetaldehyde is the dominant immediate PAN precursor in large western wildfire smoke plumes, with biacetyl also serving as an important precursor. To my knowledge this is the first time these conclusions have been drawn for smoke from an observational framework and it suggests that biacetyl should be included in models of smoke plume evolution. With respect to PPN, I find that propanal is an important immediate PPN precursor in fresh wildfire smoke. Unexpectedly, I also find that at least one other immediate PPN precursor is likely needed to explain PPN production, and suggest that this precursor may be ethylglyoxal. However, very few in situ measurements of ethylglyoxal exist to test this hypothesis.
Description
Rights Access
Subject
nitrogen oxides
reactive nitrogen
ammonia
wildfire smoke
PAN