Using non-negative matrix factorization for the identification of daily patterns of particulate air pollution in Beijing during 2004–2008

Abstract

<p><strong>Abstract.</strong> Increasing traffic density and a changing car fleet on the one hand as well as various reduction measures on the other hand may influence the composition of the particle population and, hence, the health risks for residents of megacities like Beijing. A suitable tool for identification and quantification of source group-related particle exposure compositions is desirable in order to derive optimal adaptation and reduction strategies and therefore, is presented in this paper. <br></br> Particle number concentrations have been measured in high time- and space-resolution at an urban background monitoring site in Beijing, China, during 2004–2008. In this study a new pattern recognition procedure based on non-negative matrix factorization (NMF) was introduced to extract characteristic diurnal air pollution patterns of particle number and volume size distributions for the study period. Initialization and weighting strategies for NMF applications were carefully considered and a scaling procedure for ranking of obtained patterns was implemented. In order to account for varying particle sizes in the full diameter range [3 nm; 10 μm] two separate NMF applications (a) for diurnal particle number concentration data (NMF-N) and (b) volume concentration data (NMF-V) have been performed. <br></br> Five particle number concentration-related NMF-N factors were assigned to patterns mainly describing the development of ultrafine (particle diameter <i>D</i><sub>p</sub> < 100 nm instead of <i>D</i><sub>P</sub>) as well as fine particles (<i>D</i><sub>p</sub> < 2.5 μm), since absolute number concentrations are highest in these diameter ranges. The factors are classified into primary and secondary sources. Primary sources mostly involved anthropogenic emission sources such as traffic emissions or emissions of nearby industrial plants, whereas secondary sources involved new particle formation and accumulation (particle growth) processes. For the NMF-V application the five extracted factors mainly described coarse particle (2.5 μm < <i>D</i><sub>p</sub> < 10 μm) variations, generated by processes like dust storm events. Because particle volume depends on particle diameter in a cubic manner, larger particles are emphasized in the latter application. <br></br> In order to gain insight in the day-by-day varying source-associated composition of the particle burden non-negative linear combinations of individual source-associated pollution patterns were used to approximate the original particle data. Consequently, this NMF-based procedure provides a reasonable numerical-statistical tool for the description of daily patterns of particle pollution, source identification and reconstruction of daily patterns by summarizing weighted factors.</p>

Air quality measurements in the western Eagle Ford Shale

“…NMF models have been developed for a variety of applications, including image decomposition (Lee and Seung, 1999), metagenomic analysis (Brunet et al, 2004), and atmospheric pollutant source attribution (e.g. Thiem et al, 2012). The NMF models work as follows:…”

Integrating Mobile and Fixed-Site Black Carbon Measurements to Bridge Spatiotemporal Gaps in Urban Air Quality

“…Matrix factorization techniques (e.g., principal component analysis [PCA], non-negative matrix factorization [NMF]) are widely used in air quality research. − These techniques exploit coherence across spatial, temporal, or chemical variability to uncover common pollution sources or meteorological conditions concurrently impacting these dimensions. Matrix factorizations are routinely applied to pollutant - time systems to leverage covariance among multiple pollutant time series to apportion each pollutant to one or more underlying contributing factors or sources represented by a set of covarying pollutants .…”

“…Matrix factorizations are routinely applied to pollutant - time systems to leverage covariance among multiple pollutant time series to apportion each pollutant to one or more underlying contributing factors or sources represented by a set of covarying pollutants. 49 The same techniques can also be extended to pollutant - location 53 and location - time 50 systems.…”