Book excerpt: Atmospheric aerosol is known to have adverse effects on visibility, population health, and climate. The sources and processes that contribute to atmospheric aerosol can be studied through a variety of aerosol chemical speciation techniques, and results from these analyses can be analyzed through factor analytic receptor modeling to identify origins and ultimately design effective mitigation strategies. While receptor modeling has been applied to nearly every type of aerosol speciation measurement, there has been little discussion in regards to how inputs, in terms of different aspects of measurement, relate to the answers that receptor modeling can provide in terms of air quality management and atmospheric processes. In this thesis, the factor analytic technique Positive Matrix Factorization (PMF) was applied to four sets of aerosol chemical speciation measurements of varied time resolution and chemical breadth, from the receptor region of greater Windsor, Ontario. At the outset of this research, receptor modeling was dominated by the use of long-term, low time-resolution integrated filter measurements. This research involved the extension of receptor modeling towards more advanced, state-of-the-art sub-hourly time-resolution aerosol measurements, including trace element measurements from a semi-continuous elements in aerosol sampler (SEAS), bulk aerosol mass spectra from an Aerodyne Aerosol Mass Spectrometer (AMS), and single particle mass spectra from a TSI Aerosol Time-of-Flight Mass Spectrometer (ATOFMS). Receptor modeling of measurements from these techniques each yielded insights specific to the method, time of year and duration of sampling. Taken together, receptor modeling results from the high time-resolution measurement methods provided a far more detailed understanding of the variety of the sources and processes that contribute to aerosol composition and concentration in greater Windsor, as compared to long-term, low time-resolution integrated filter measurements alone. In total, 14 different factors were identified. Receptor modeling of integrated filter measurements established that secondary aerosol dominated greater Windsor's PM2.5 mass (e.g., Secondary Sulphate and Nitrate contributed 37 and 24% respectively to Windsor's average PM2.5 of 12ug m-3); other local, and local-to-regional anthropogenic factors, such as Traffic, Steel Mills, and Oil Combustion emissions, also contributed (18%, 4%, and 2% by mass respectively). However, receptor modeling using high time-resolution measurements provided more detailed perspectives into aerosol origins. For instance, receptor modeling of AMS and ATOFMS measurements highlighted that secondary nitrate, identified in past studies in the region as mostly regional, is actually a superposition of regional and local nitrate, with local nitrate forming overnight in both winter and summer. Single particle mass spectral measurements confirmed as hypothesized, that regional summertime secondary sulphate and nitrate aerosol consists of a homogeneous external mixture of particles that are internally mixed with sulphate, nitrate, organic and elemental carbon. Traffic aerosol was shown to contain fresh, hydrocarbon-like organic aerosol, as well as black carbon and trace elements such as Ba and Fe, but showed a wide range of average concentrations (0.94 - 2.18ug m-3), specific to the circumstances of measurement. An integrated analysis across all four studies highlighted the strengths and limitations of the various chemical speciation techniques and receptor modeling approaches used in each study. For instance, receptor modeling of SEAS measurements that included refractory elements provided a source-based perspective useful for primary source identification, while receptor modeling of non-refractory combined organic and inorganic mass spectra from the AMS provided a more process-based perspective useful for characterizing the degree of aerosol chemical processing. Receptor modeling results of ATOFMS single particle measurements provided the most detailed representation of factors, highlighting both the internal and external mixing states of particle-types. An overarching conclusion emerged from this work, in that the degree of factor deconvolution in factor analytic receptor modeling is inter-related with the variability in chemical composition of the aerosol being measured, the properties of measurement, and modeling parameters. With a priori consideration given to this conclusion and more intentional study design, receptor modeling studies can lead to improved factor resolution, and ultimately more reliable and useful results.