Outdoor air pollution from particulate matter (PM) is a worldwide issue, affecting society in terms of not only public health impacts, but also of economic impacts such as increased health costs, production losses, and lost tourism revenue. These impacts are often disproportionally suffered by marginalized populations, and even in regions where air pollution has decreased with time, the same groups experiencing higher exposures a few decades ago are still overexposed today. Studying the scale and severity of outdoor particulate air pollution, its interaction with meteorology and climate, its exposure effects on human physiology, and its association with socioeconomic inequalities, is critical to improving our understanding of pollution impacts on visibility, economy, public health, and environmental justice. However, data scarcity has often precluded a proper characterization of the full spatio-temporal extent of pollution dynamics and their effects on different populations. The work presented in this dissertation addresses the challenges that data scarcity presents to the investigation of air pollution dynamics and effects through different approaches, in order to (i) characterize spatio-temporal patterns and main drivers of PM pollution, (i) create strategies for targeted uncertainty reduction of air quality health impact assessments, and (iii) orient uncertainty reduction strategies towards the benefit of previously overlooked populations. First, we investigate the emission and meteorological variables key to explaining spatio-temporal variability in coarse PM pollution levels, using a parsimonious statistical model and PM observations from a dense monitoring network in Malaysia. Second, we develop a method to compare the effects of information gain in air pollution and epidemiological data, by using the metric of information entropy, to identify the most efficient pathway to reduce uncertainty in estimates of air pollution-associated health risks. Lastly, we propose an expansion of the information entropy method for the study of socioeconomic disparities in the correlations between air pollution levels, epidemiological effects, and mortality assessment uncertainties, highlighting the important influence of minority representation in the uncertainty reduction of air pollution health assessments. The methods developed in this work are highly generalizable, allowing for their application to a wide variety of pollution-effects scenarios beyond the particular case studies presented.