Air Pollution Source Contribution Rate Analysis: Source Apportionment Methods, Testing, and Applicat

Air pollution source contribution rate analysis, also known as source apportionment, quantitatively determines how much each emission source—such as motor vehicle exhaust, coal combustion, industrial processes, dust, and biomass burning—contributes to ambient particulate matter and gaseous pollutants. By combining comprehensive chemical speciation of PM2.5 and PM10 samples with receptor models such as Positive Matrix Factorization (PMF) and Chemical Mass Balance (CMB), this analysis provides the scientific foundation for targeted air quality management and emission control strategies.

Technical Overview

Source contribution rate analysis is a quantitative technique used to identify and measure the relative contributions of different pollution sources to ambient air quality. The approach is founded on the principle of mass conservation: the chemical composition measured at a receptor site represents the sum of chemically conserved species emitted by each individual source, weighted by its contribution. By collecting PM2.5, PM10, and volatile organic compound (VOC) samples at representative monitoring sites and resolving the resulting chemical data with statistical receptor models, the analysis produces a quantified breakdown of source contributions—for example, secondary aerosols 28%, vehicle exhaust 24%, fugitive dust 20%, coal combustion 16%, industrial emissions 9%, biomass burning 3%—each reported with associated uncertainty estimates.

This technique is widely used by environmental authorities, research institutions, and industrial enterprises to move beyond simple concentration monitoring and answer the critical question of where pollution actually comes from, enabling cost-effective, prioritized emission reduction decisions. Work is typically performed by accredited third-party laboratories following internationally recognized modeling guidelines and national technical specifications for source apportionment.

Principal Receptor Models

• Positive Matrix Factorization (PMF): decomposes the sample-by-species data matrix into factor profiles and factor contributions without requiring source profiles in advance.
• Chemical Mass Balance (CMB): solves effective-variance least-squares equations by fitting ambient species concentrations against measured local or library source profiles.
• UNMIX: a multivariate model using geometric concepts such as edge analysis to resolve source compositions and contributions.
• Absolute Principal Component Analysis / Multiple Linear Regression (APCS/MLR): combines principal component analysis with regression scaling to quantify source contributions.

Typical Source Categories Resolved

• Secondary sulfates, nitrates, and secondary organics
• Motor vehicle exhaust (gasoline and diesel)
• Coal combustion (power generation, heating, industrial boilers)
• Industrial processes (steel, cement, non-ferrous metallurgy, chemical production)
• Construction and road fugitive dust, soil dust
• Biomass burning and open burning of waste
• Sea salt and other crustal or regional transport sources

Testing Samples

Ambient Receptor Samples

PM2.5 and PM10 samples are collected simultaneously on quartz-fiber filters (for carbon and ion analysis) and Teflon membrane filters (for gravimetric mass and elemental analysis) using medium- or high-volume air samplers deployed at urban, suburban, industrial, traffic, and background sites. Sampling campaigns are designed to cover representative seasons and both pollution episodes and clean periods, with continuous monitoring of meteorological parameters such as temperature, humidity, wind speed, and wind direction to support data interpretation.

Source Profile Samples

To support Chemical Mass Balance modeling and factor identification, representative source samples are collected under controlled conditions, including resuspended road dust, construction dust, and surface soil; fly ash and raw coal of various types; vehicle exhaust from gasoline and diesel fleets; industrial stack emissions; and biomass burning smoke. These samples undergo the same chemical speciation as ambient samples to construct local source profiles.

VOC and Gaseous Samples

Whole-air samples collected in passivated canisters or adsorbent cartridges are analyzed for ozone precursors and toxic organics, enabling VOC source apportionment, ozone formation potential evaluation, and identification of solvent use, petrochemical, and vehicle-related sources.

Testing Items

Comprehensive chemical speciation of each sample typically includes the following components, forming the data matrix required for receptor modeling:

• Mass concentration: gravimetric determination of PM2.5 and PM10 under constant temperature and humidity conditions.
• Water-soluble ions: SO₄²⁻, NO₃⁻, Cl⁻, F⁻, NO₂⁻, NH₄⁺, Na⁺, K⁺, Mg²⁺, Ca²⁺.
• Carbon fractions: organic carbon (OC) and elemental carbon (EC) with eight thermal-optical carbon fractions using protocols such as IMPROVE_A or NIOSH 5040.
• Elements: typically 20–40 elements, including crustal markers (Al, Si, Ti, Ca, Fe, Mn), anthropogenic tracers (Cu, Zn, Pb, Cd, Cr, Ni, As, Se, V, Mo, Sb, Ba), and marine markers (Na, Cl).
• VOC species: alkanes, alkenes, aromatics, halogenated hydrocarbons, and oxygenated VOCs covering the PAMS and TO-15 target lists.
• Optional tracers: stable lead isotopes, radiocarbon (¹⁴C) for fossil vs. modern carbon split, and levoglucosan as a biomass burning marker.

Testing Methods

Analytical Techniques

• Gravimetric weighing of filters in a temperature- and humidity-controlled balance room, referencing methods such as EN 12341 and GB/T 15432.
• Ion chromatography for water-soluble inorganic and organic ions.
• Thermal-optical transmittance or reflectance analysis for OC/EC quantification.
• X-ray fluorescence (XRF) for non-destructive elemental analysis, and inductively coupled plasma mass spectrometry (ICP-MS) or ICP-OES following acid digestion for trace elements.
• Gas chromatography–mass spectrometry (GC-MS) with cryogenic preconcentration for VOC speciation.

Modeling and Data Analysis Procedure

Quality-assured concentration and uncertainty data are compiled into a sample-by-species matrix and processed through established receptor model software. In PMF, factor numbers are optimized by evaluating the objectivity function Q, signal-to-noise ratios of individual species, factor interpretability, and bootstrap and displacement stability diagnostics. In CMB, measured ambient species are fitted against source profiles, with performance statistics (R², chi-square, percentage mass explained) used to judge solution quality. Source identification is confirmed using diagnostic ratios (for example, OC/EC, NO₃⁻/SO₄²⁻, K⁺ for biomass, Si/Al for crustal matter) and by cross-checking results against local emission inventories and dispersion modeling.

Quality Assurance and Quality Control

Full quality assurance measures include field and laboratory blanks, duplicate and split samples, continuing calibration verification, analysis of certified reference materials, ion balance and anion-cation consistency checks, and reconstructed mass closure evaluation. Results from accredited third-party laboratories are reported with complete uncertainty analysis, ensuring defensibility for regulatory and decision-making purposes.

Applications

• Formulation and refinement of regional and urban air quality improvement action plans, providing prioritized emission reduction targets.
• Cause analysis of heavy pollution episodes, distinguishing local accumulation, regional transport, and secondary formation.
• Evaluation of the effectiveness of emission control policies, such as coal-to-gas conversion, vehicle restriction, and dust suppression programs.
• Environmental fingerprinting of industrial parks and identification of responsibility for cross-boundary pollution disputes.
• Scientific support for environmental impact assessments, air quality attainment planning, and environmental litigation.
• Research on secondary aerosol formation mechanisms and ozone control strategy development through VOC precursor apportionment.

FAQ

How many samples are required for reliable source apportionment?

For PMF modeling, a minimum of approximately 60–100 valid samples is generally recommended, with seasonal campaigns of 80–120 samples per period providing substantially improved robustness. A larger sample count reduces rotational ambiguity, strengthens temporal representativeness, and narrows reported uncertainty ranges.

What is the difference between PMF and CMB?

PMF does not require source profiles in advance; it mathematically resolves factor profiles and contributions from the ambient data matrix, which are then interpreted based on tracer species. CMB requires pre-measured local source profiles but can deliver source contributions for individual sampling days. The two methods are complementary and are frequently applied together for cross-validation.

How long does a complete source apportionment project take?

A typical project involves a sampling campaign spanning at least one full year or a minimum of four representative seasonal periods, followed by laboratory analysis and modeling, which generally requires several additional weeks. Rush programs focusing on specific pollution episodes can be compressed into one to two months.

What uncertainty is associated with the contribution results?

Depending on data quality and model stability, reported source contributions typically carry uncertainties of approximately 10–30%. Bootstrap resampling, displacement analysis, and sensitivity runs are used to quantify and disclose the stability of each resolved factor.

Can gaseous pollutants such as ozone also be apportioned?

Yes. Ozone is a secondary pollutant, so apportionment focuses on its VOC and NOx precursors, which are speciated by GC-MS and resolved with PMF or chemical transport models to rank precursor sources and guide ozone control strategies.

Do the results support regulatory decision-making?

When conducted by accredited third-party laboratories using validated sampling and analytical methods, documented modeling diagnostics, and complete quality assurance records, source apportionment results are widely accepted as the technical basis for air quality management plans, policy evaluation, and pollution responsibility determinations.