Source apportionment of airborne particulates through receptor modeling: Indian scenario

Highlights

• Large heterogeneity prevails in aerosol spatial distribution.

• Basis of selections of tracers for specific source categories has improved.

• Particulate source profile are largely incomparable and sometime confusing.

• Limited use of organic molecular markers and gas-to-particle conversion in SA.

• Finer particulates source profiles are extremely limited.

Abstract

Airborne particulate chemistry mostly governed by associated sources and apportionment of specific sources is extremely essential to delineate explicit control strategies. The present submission initially deals with the publications (1980s–2010s) of Indian origin which report regional heterogeneities of particulate concentrations with reference to associated species. Such meta-analyses clearly indicate the presence of reservoir of both primary and secondary aerosols in different geographical regions. Further, identification of specific signatory molecules for individual source category was also evaluated in terms of their scientific merit and repeatability. Source signatures mostly resemble international profile while, in selected cases lack appropriateness. In India, source apportionment (SA) of airborne particulates was initiated way back in 1985 through factor analysis, however, principal component analysis (PCA) shares a major proportion of applications (34%) followed by enrichment factor (EF, 27%), chemical mass balance (CMB, 15%) and positive matrix factorization (PMF, 9%). Mainstream SA analyses identify earth crust and road dust resuspensions (traced by Al, Ca, Fe, Na and Mg) as a principal source (6–73%) followed by vehicular emissions (traced by Fe, Cu, Pb, Cr, Ni, Mn, Ba and Zn; 5–65%), industrial emissions (traced by Co, Cr, Zn, V, Ni, Mn, Cd; 0–60%), fuel combustion (traced by K, NH₄⁺, SO₄⁻, As, Te, S, Mn; 4–42%), marine aerosols (traced by Na, Mg, K; 0–15%) and biomass/refuse burning (traced by Cd, V, K, Cr, As, TC, Na, K, NH₄⁺, NO₃⁻, OC; 1–42%). In most of the cases, temporal variations of individual source contribution for a specific geographic region exhibit radical heterogeneity possibly due to unscientific orientation of individual tracers for specific source and well exaggerated by methodological weakness, inappropriate sample size, implications of secondary aerosols and inadequate emission inventories. Conclusively, a number of challenging issues and specific recommendations have been included which need to be considered for a scientific apportionment of particulate sources in different geographical regions of India.

Introduction

Airborne particulates are distinguished as multi-component mixtures originated from a wide range of sources, subsequently evolved through several microphysical processes like nucleation, coagulation and condensation before ultimately scavenging off either through wet or dry deposition (Kumar et al., 2015). Association of airborne particulates with human health, crop yield, regional circulation systems, climate dynamics and many other sectors of earth's system is well established in numerous scientific literatures (Ramanathan and Feng, 2009). Implications of airborne particulate are typically regional as predominant particulate species have comparably lower residence time in contrast to long-lived greenhouse gases. However, trans-boundary movement of particulates amplifies the quantum of impacts to greater distances (Ramanathan and Feng, 2009, Banerjee et al., 2011a, Kumar et al., 2015). Particulate diversity in Indian subcontinent is extremely diverse and complex which requires methodical understanding in terms of composition, morphology and mixing state, size distribution and chemical evolution. Among others, particulate morphology and chemical heterogeneities principally regulate its optical properties, volatilities, interactions and phase transformations in different ways and therefore, efforts have been made over the years to study atmospheric particulates as a function of size, shape and chemical characteristics. Particulate morphology and chemical characteristics are essentially functions of associated sources and therefore, it is extremely essential to be characterized for proper estimation. Source apportionment of airborne particulates essentially quantifies the contribution of individual sources to particulate loading based on source and receptor characteristics and in certain cases, with nature of pollutants. Such may be accomplished either through analyzing specific tracers in bulk filter analysis or by numerical/statistical analysis of specific parameter with prevailing meteorological variables or by coupling emission inventory information with dispersion models. In that way, three principal SA techniques are (a) chemical transport models based upon pollutant composition driven by meteorological variables (Banerjee et al., 2011a, Belis et al., 2013), (b) receptor-oriented models based on analysis of chemical data acquired at receptor sites (Balachandran et al., 2000, Srivastava et al., 2008), and (c) emission inventories and dispersion models (Laupsa et al., 2009, Banerjee et al., 2011b).

A number of SA studies for different atmospheric pollutants with some degree of certainty are available in India with majority of studies being conducted using receptor models (RMs) based on monitored particulate concentrations and their source profile. Receptor models are broadly classified into microscopic and chemical methods based on the nature of particle characteristics opted for simulation. Microscopic RM analyzes the morphological features of airborne particulate through optical and scanning electron microscopes which are efficient enough to characterize aerosols in mixing states (Pipal et al., 2011). However, applicability of microscopic RM limits in large-scale as it mostly yields qualitative products (Pant and Harrison, 2012) and confines in identifying inorganic compounds (Shi et al., 2008). In contrast, chemical RM utilizes the chemical composition of airborne particles for identification and apportionment of specific sources. The foundation of all receptor based models is mass conservation which may simply be explained by Eq. 1:

$$C=S\cdot F+e$$

where, C is the particulate concentration profile in the receptor site, S denotes source contribution which needs to be measured, F is the particulate source profile and e denotes error between the measured and predicted concentrations. However, mass balance equation itself based on certain assumptions is critical for specific RM techniques than for others (Belis et al., 2013). In most of RMs, the system is considered as quasi-stationary representing statistically insignificant variations in source profile with time. Additionally, evolved particulate species are required to be chemically inactive during its aerial transport. For a proper SA, these assumptions are essentially required to be satisfied, except which definite source profiling may not be achieved or RMs may under/overestimate the individual contribution of particular source (Belis et al., 2013). Additionally, the concept of RMs is based on proper identifications and quantification of specific signatory molecules which virtually establish missing links between sources and receptors. Chemical signatures of specific sources are used to be extremely sensitive and as these signatures gradually evolve with time, may undergo chemical phase transformations and eventually be masked, which critically limits its applications as a tracer. Therefore, selection of the specific RMs for SA studies is extremely important as only few RMs can tolerate deviations of pre-identified assumptions (Watson et al., 2008, Belis et al., 2013). These RMs especially consider selective losses of particular tracers caused by phase transformations and ultimately re-introduce it in the analysis as an error-input.

The present submission only synthesizes the results archived in SA research articles and scientific reports published in the Indian context and available in abstract and citation database of peer-reviewed literature. Particulate sources are summarized and quantitatively evaluated to reduce uncertainties in identifying specific airborne particulate sources. A total of 90 research articles originated in a span of 1985–2014 and available in citation database of peer-reviewed literature were scrutinized. Interestingly, SA studies in India initiated way back in 1985, however, researchers were only aspired for particulate SA studies from 2005 onwards which share 80% (72) of the total scientific publications (90) (Fig. 1a). Until 1990s, there were only 9 publications (10% of total) available typically originated using EF (42%) and FA (33%), while only 3 instances were there when advanced RMs like PCA (17%) and CMB (8%) were in use (Fig. 1b). Unavailability of regional source profile may possibly restrict the use of advanced RM. Since 2000, SA of airborne particulate established itself as a principal research domain (90% of publications) to atmospheric scientists possibly due to raising concern of aerosol induced adverse health impacts and more precisely its association with regional climate change. Since 2010–14, a total of 81 publications were found involving SA of particulates involving many advanced RMs likewise EF (34%), FA (3%), PCA (36%), PMF (15%), UNMIX (3%) and CMB (10%). However, in most of the cases SA was only performed for PM₁₀ and SPM and only in few instances, PM_2.5 was considered as a particulate metric. For the entire SA study, most preferred particulate metrics were PM₁₀ (41%) followed by PM_2.5 (26%) and SPM (22%). The trend is well comparable to that of European SA case studies (Belis et al., 2013) where significant proportions of SA have been carried out for PM₁₀ (56%) followed by PM_2.5 (37%) and SPM (1%). For both instances, SA for ultrafine particulates was relatively less (India: 10% and Europe: 6%) signifying specific requirement of regional case studies emphasizing on ultrafine particulates, which are supposed to be predominately anthropogenic and carcinogenic in nature. In India, considerable amount of particulate SA was carried out from 2010 and 5 years of research resulted to staggering 44 publications (49% of total). Fig. 2 indicates the geographical distribution of SA studies that have been conducted in India. It is evident that most of SA studies have been performed in and around Delhi, Mumbai, Chennai and Kolkata with some contributions from Hyderabad, Tirupati, Durg, Kanpur, Agra and Chandigarh.

A comprehensive analysis on airborne particulate loading and its spatial distribution has also been extensively reviewed. Such meta-analysis clearly indicates the presence of substantial particulate loading especially in the context of Indo-Gangetic Plain (IGP) which requires adequate source segregation for effective implementation of control strategies. Emphases were made to isolate individual species that were characteristics of particulate composition of a specific location. Assignment of specific tracers to precise source categories was also evaluated. Conclusively, a comprehensive review has been made for particulate SA studies in the Indian perspective yielding a quantitative estimation of most relevant particulate sources and there temporal variations.