Date of Award




Document Type


Degree Name

Doctor of Philosophy (PhD)


Department of Atmospheric and Environmental Sciences

Content Description

1 online resource (xii, 93 pages) : illustrations (chiefly color), color maps.

Dissertation/Thesis Chair

Sarah Lu

Committee Members

Scott Miller, Qilong Min, Liming Zhou


Aerosol, Air Quality, Machine Learning, Wildfire Smoke, Smoke, Wildfires, Air quality, Aerosols

Subject Categories

Atmospheric Sciences


In New York State (NYS), summertime long-range transported smoke plumes and high fine mode particulate matter (PM2.5) concentration during the smoke episodes have been reported. Since the frequency, intensity, and burned area of wildfires over Northern America have increased in recent years, the contribution of transported smoke aerosols to local air quality could become more significant. This study aimed to characterize the transport of smoke aerosols and quantify their impacts on the local air quality over NYS. A case study of a transported smoke event in mid-August of 2018 and multi-year analysis of the smoke cases during the summer seasons of 2012 – 2019 were conducted. Various analyses were applied, including 1) identification of smoke plumes using satellite measurements and aerosol reanalysis products, 2) quantification of the external smoke-associated PM2.5 fluxes and their contribution to local air quality, 3) investigation of the annual variation of smoke transport, and 4) characterization of the transport pattern of smoke aerosols using machine learning, in particular for the influence of the planetary boundary layer (PBL) evolution and atmospheric vertical mixing on smoke transport. Results showed that, on average, there were about 300 kg m-1 of column-integrated smoke-associated PM2.5 transported into NYS from the west, resulting in an increase of 5 μg m-3 in PM2.5 concentrations statewide on smoke days. About 30% of high PM2.5 cases in NYS, cases with PM2.5 concentrations exceeding 20 μg m-3, were associated with transported smoke aerosols. In addition, case study and artificial neural network (ANN) models revealed that synoptic subsidence, entrainment process, and turbulent mixing collectively contributed to the downward transport of smoke aerosols and the increases of surface PM2.5 concentrations. Future studies about the key factors which determine the fate of transported smoke aerosols, such as plume rise, large-scale circulation, aerosol physical/chemical processes, and PBL evolution over downwind regions, are recommended.