Date of Completion


Embargo Period



Aerosols, Climate Change, Fine Particulate Matter, Air Quality Remote Sensing, Atmospheric Chemistry Modeling, Climate Extremes

Major Advisor

Chuanrong Zhang

Associate Advisor

Richard Anyah

Associate Advisor

Daniel L. Civco

Associate Advisor

Kenneth Foote

Associate Advisor

Nathaniel Trumbull

Field of Study



Doctor of Philosophy

Open Access

Open Access


A good understanding of trends, variations, and causes of atmospheric aerosols is vital to quantifying the role of air quality in climate change and health concerns and for informing relevant regulatory policies. This thesis presents four projects that exploit a range of observational data and modeling tools to characterize and interpret aerosol dynamics and its potential determinates.

An econometric model is implemented to identify aerosol variations, predictions, and the driving forces using six sites spreading across North America and East Asia during 2003–2015. Regional differences caused by impacts of climatology and land cover types are observed. Statistical validation of time series ARIMA models indicates the applicability and feasibility of ARIMA modeling. The reasonably-predicted AOD values could provide reliable estimates to inform better the decision-making for sustainable environmental management and the initiative of reforestation on emission sinks could have potential implications for climate change mitigation.

The time series analyses and modeling of aerosol variability are further investigated in a spatially continuous framework based on the valuable spatiotemporal dimension of the remote sensing data over the contiguous United States (U.S.) and China. By comparing variations and trends in these two countries, we attribute the large differences to the energy strategies, economic and urban development, and lifestyle activities. Areas most suitable for applying the model for prediction are those with high AOD quality, high completeness of AOD data, low-AOD values, and AOD time series with clear seasonal variations.

Then, a geographically weighted regression (GWR) model is employed to estimate PM2.5 concentration and analyze its relationships with AOD, meteorological variables, and nighttime light (NTL) data across the Northeastern United States in 2013. Improved GWR model performance is found for the warm season when applying the index that incorporates normalized difference vegetation index (NDVI) into NTL (17% and 7.26% better than GWR model without NDVI and NTL data and GWR model without NTL data, respectively). The spatial distribution of the estimated PM2.5 levels clearly reveals patterns consistent with those densely populated areas and high traffic areas.

Finally, an evaluation framework to identify both mean and extreme conditions of PM2.5 is proposed and applied to the WRF-CMAQ simulations over the contiguous U.S. for the period of 2001-2010. While the model exhibits satisfactory performance over the eastern U.S., PM2.5 mean variations and extreme trends in the western U.S. are not well represented partly due to the complex terrains and active fire activities. Moreover, the relationship between extreme PM2.5 pollution episodes and abnormal synoptic conditions is quantified. More extreme PM2.5 pollution episodes are expected in a warming climate, with rural stations and the western U.S. suffer the most.