Human Impacts on Atmospheric Particulate Matter in Amazonia

Abstract

Human activities can change the concentration, chemical composition, and properties of atmospheric particulate matter (PM). A quantitative understanding of human impacts on PM is essential for accurate modeling of anthropogenic climate forcing and for regulating air quality. The tropical forests of Amazonia play a critical role in the regional and global climates. Urbanization and deforestation in the region have intensified in the last decades, and yet a quantitative knowledge on the impacts of these activities on PM is scarce. This thesis presents field measurements of submicron PM concentration, composition and optical properties in central Amazonia, with a focus on delineating the anthropogenic impact on the observed quantities. Emphasis is given to organic material, the dominant PM component in the region. The primary site of the studies was located 70 km to the west of Manaus, a city of over two million people in the heart of the Brazilian Amazon. Datasets were collected during the wet and dry seasons of 2014 as part of the GoAmazon2014/5 experiment. The main datasets were obtained by a high-resolution time-of-flight aerosol mass spectrometer (AMS), and additional data from a large suite of instrumentation were employed in the analyses. Positive-matrix factorization (PMF) of organic mass spectra allowed for the classification of the organic PM into its component classes. During the wet season, the effects of the Manaus urban plume were investigated. Fuzzy c-means clustering (FCM) was applied to concentrations of pollution indicators and identified four clusters, which represented distinct background and polluted conditions. When crossed with the AMS and PMF results, increases of up to 200% in PM mass concentrations for polluted compared to background conditions were identified, corresponding to average concentrations as high as 3 µg m-3 and as low as 1 µg m-3, respectively. The pollution-induced changes in the composition of secondary organic PM1 suggested (i) a shift in pathways of organic PM production from HO2- to NO-dominant, and (ii) an acceleration of the atmospheric processing of organic PM. Nitrogen oxides were indicated to play a critical role in both processes. In another study, the influences of the Manaus plume on the pathway of PM production from isoprene epoxydiols (IEPOX) were investigated in detail. The “IEPOX-SOA” factor served as a surrogate for IEPOX-derived PM. This factor had a bi-variate dependence on sulfate and NOy, which was reasoned through the chemical mechanism. For an approximately fixed sulfate concentration, a change in NOy concentrations from 0.5 to 2 ppb was associated with a decrease in IEPOX-SOA factor loadings of two- to three-fold. It was found that the Manaus pollution plume elevated concentrations of NOy more significantly than of sulfate over background. It was therefore concluded that the suppressing effects of increased NO often prevailed over the enhancing effects of increased sulfate on the production of IEPOX-derived PM in the region. During the dry season, the importance of biomass burning increased, and the final study focused on quantifying the anthropogenic contributions to light-absorbing organic PM (brown carbon, BrC). Optical properties were determined from aethalometer measurements. The relationships of the absorption coefficient of BrC at 370 nm with AMS-measured quantities suggested that BrC absorption was associated with less oxidized and nitrogen-containing organic compounds, and that atmospheric processing caused bleaching of the BrC. Estimates of the effective mass absorption efficiencies associated with each class of organic PM (i.e., PMF factor) were obtained through multi-variate linear regression. Together, urban and biomass burning emissions were estimated to contribute on average to 90% of the absorption by organic PM, while they contributed to about 40% of organic PM mass concentrations.