Atmospheric Pollution Cause Analysis: Source Apportionment, Testing Methods, and Applications

Atmospheric pollution cause analysis is a systematic scientific process that identifies, quantifies, and attributes airborne pollutants to their natural and anthropogenic origins through chemical characterization, source apportionment modeling, and dispersion simulation. Accredited third-party laboratories perform comprehensive sampling and analysis of particulate matter, gaseous pollutants, and trace contaminants in accordance with recognized standards from ISO, EN, ASTM, US EPA, and GB systems. This technical article explains the principal causes of atmospheric pollution, the samples and parameters involved, the analytical and modeling methods applied, and the practical applications of the results.

Technical Overview

Atmospheric pollution cause analysis is a multidisciplinary evaluation that determines why air quality degradation occurs in a given region by combining emission source identification, chemical composition analysis, receptor-based source apportionment, and meteorological assessment. Unlike routine compliance monitoring, which only measures pollutant concentrations against thresholds such as GB 3095, cause analysis answers which sources contribute, how much each source contributes, and why pollution episodes form under specific weather conditions.

The analysis typically addresses the six criteria pollutants — fine particulate matter (PM2.5), inhalable particulate matter (PM10), sulfur dioxide (SO2), nitrogen oxides (NOx), carbon monoxide (CO), and ozone (O3) — together with volatile organic compounds (VOCs), ammonia (NH3), and toxic trace species. Causes are grouped into emission-related factors, atmospheric transformation processes, meteorological conditions, and regional transport effects.

Emission-Related Causes

• Fossil fuel combustion: coal-fired power generation, industrial boilers, and residential heating emit SO2, NOx, primary particles, and carbonaceous aerosol.
• Industrial processes: steel, cement, coke, chemical, and petrochemical production release dust, heavy metals, VOCs, and acid gases.
• Motor vehicle emissions: exhaust from gasoline and diesel vehicles contributes NOx, VOCs, CO, and ultrafine particles, while brake and tire wear adds road dust.
• Fugitive dust: construction activities, road resuspension, bare soil, and bulk material handling generate mineral dust rich in crustal elements.
• Agricultural and rural sources: biomass burning, open straw burning, livestock operations, and fertilizer use emit NH3, potassium-rich smoke, and organic aerosol precursors.
• Natural sources: wind-blown dust storms, volcanic eruptions, wildfires, sea salt aerosol, and biogenic VOC emissions.

Secondary Pollution Formation

A large fraction of PM2.5 in many polluted regions is secondary aerosol formed in the atmosphere rather than emitted directly. Gaseous precursors such as SO2, NOx, NH3, and VOCs undergo oxidation to form sulfate, nitrate, ammonium salts, and secondary organic aerosol (SOA). Ground-level ozone is entirely secondary, produced by photochemical reactions between NOx and VOCs under sunlight. Cause analysis therefore quantifies both primary emissions and the sensitivity of secondary formation to precursor reductions.

Meteorological and Transport Causes

Unfavorable meteorology frequently triggers heavy pollution episodes. Temperature inversions, low wind speeds, shallow boundary layers, and high humidity suppress vertical mixing, trap pollutants near the surface, and accelerate aqueous-phase sulfate formation. Regional transport under prevailing winds can carry pollutants hundreds of kilometers, meaning a receptor site may be affected by sources far beyond its administrative boundary. Backward trajectory analysis and sensitivity modeling are used to separate local contributions from regional transport.

Testing Samples

Robust cause analysis requires paired sampling of both the atmosphere (receptor samples) and the candidate sources (source profile samples) so that chemical fingerprints can be matched statistically.

Ambient Receptor Samples

• PM2.5 and PM10 collected on quartz and Teflon membrane filters using gravimetric samplers or medium- and high-volume samplers.
• Time-resolved filter samples, typically 2-hour to 24-hour intervals, to capture episode evolution.
• Whole-air VOC samples in evacuated canisters and aldehyde samples on DNPH cartridges.
• Real-time online data streams: continuous PM monitors, trace gas analyzers, aerosol chemical speciation monitors (ACSM/AMS), and black carbon aethalometers.
• Deposition samples such as precipitation and dustfall for long-term trend evaluation.

Source Profile Samples

• Surface dust samples: road dust, construction dust, soil dust, and urban surface deposits collected by sweeping or resuspension methods.
• Coal combustion samples: fly ash and bottom ash from power plants, industrial boilers, and residential stoves of different coal types.
• Vehicle emission samples: chassis dynamometer exhaust, tunnel aerosol, and roadside samples representing gasoline, diesel, and liquefied petroleum gas fleets.
• Industrial emission samples: stack particulate matter and process VOCs from representative local facilities.
• Biomass and open burning samples: straw, firewood, and domestic refuse combustion aerosol.
• Sea salt and crustal reference materials for coastal and arid regions.

Testing Items

The testing program quantifies the chemical constituents that allow pollutants to be traced back to their origins.

Chemical Composition of Particulate Matter

• Water-soluble inorganic ions: sulfate (SO42-), nitrate (NO3-), ammonium (NH4+), chloride (Cl-), potassium (K+), sodium (Na+), calcium (Ca2+), and magnesium (Mg2+).
• Carbonaceous components: organic carbon (OC) and elemental carbon (EC), with OC/EC ratio evaluation and secondary organic carbon estimation.
• Elemental analysis: more than 20 elements including Si, Al, Fe, Ca, and Ti as crustal markers; As, Se, Pb, Cd, Cr, and Hg as coal and industrial markers; V and Ni as oil combustion markers; Zn and Cu as traffic and industrial markers.
• Organic molecular tracers: levoglucosan (biomass burning), hopanes and steranes (vehicle lubricating oil), cholesterol (cooking), polycyclic aromatic hydrocarbons (combustion), and isoprene or terpene oxidation products (biogenic secondary organic aerosol).

Gaseous Pollutants and Precursors

• Concentrations and diurnal profiles of SO2, NO2/NOx, CO, O3, and NH3.
• VOC speciation covering PAMS and TO-15 target lists (alkanes, alkenes, aromatics) plus oxygenated VOCs and aldehydes, with ozone formation potential (OFP) calculation.

Physical Characterization

• Particle size distribution across ultrafine, accumulation, and coarse modes.
• Single-particle morphology and elemental composition by scanning electron microscopy with energy-dispersive spectroscopy (SEM-EDS).
• Light-absorbing properties: black carbon and brown carbon, with aerosol light extinction reconstruction for visibility studies.

Testing Methods

Sampling and analysis follow validated national and international methods, and source contributions are computed with receptor and dispersion models operated by accredited third-party laboratories.

Sampling and Laboratory Analytical Techniques

• Gravimetric mass determination of PM10 and PM2.5 following HJ 618, HJ 656, EN 12341, and EN 14907.
• Ion chromatography (IC) for water-soluble ions in accordance with HJ 800 and related standard methods.
• Thermal-optical transmittance and reflectance protocols (NIOSH 5040, IMPROVE_A) for OC and EC determination.
• X-ray fluorescence (XRF) per US EPA Compendium Method IO-3.3 and inductively coupled plasma mass spectrometry (ICP-MS) per EPA IO-3.5 for elemental analysis.
• Gas chromatography-mass spectrometry (GC-MS) for VOCs and organic tracers following HJ 644, HJ 759, EPA TO-15, and EPA TO-13A; high-performance liquid chromatography (HPLC) for aldehydes per EPA TO-11A.
• Stationary source (stack) sampling of particulate and gaseous emissions according to ISO 9096 and applicable national emission standards.

Receptor Models for Source Apportionment

Receptor models resolve the mixture measured at a sampling site into statistically estimated source contributions:

• Positive Matrix Factorization (PMF, EPA PMF 5.x): factorizes species-by-sample matrices with uncertainty weighting, typically requiring 60 or more valid samples.
• Chemical Mass Balance (CMB): fits measured species to locally measured source profiles; well suited when representative local profiles exist.
• UNMIX and other multivariate methods used as cross-validation tools.
• Second-order verification through tracer ratios, wind-direction dependence, and conditional probability function (CPF) analysis.

Dispersion and Chemical Transport Modeling

• Lagrangian and Gaussian models such as CALPUFF and AERMOD for source-receptor relationships around industrial facilities.
• Three-dimensional chemical transport models such as WRF-CMAQ and WRF-Chem for regional sensitivity analysis and control-scenario simulation.
• HYSPLIT backward trajectory clustering with concentration-weighted trajectory (CWT) and potential source contribution function (PSCF) analysis to identify regional transport pathways.
• Emission inventory compilation and processing, including tagged tracer experiments for source-oriented attribution.

Quality Assurance and Quality Control

Credible results depend on strict QA/QC procedures: field and laboratory blanks, duplicate sampling, standard reference materials, balance-room temperature and humidity control per HJ 656, chain-of-custody documentation, and measurement uncertainty propagated into the receptor model. Method detection limits and inter-laboratory comparison results are reported with every dataset.

Applications

Cause analysis converts measurement data into decisions, with typical applications spanning regulatory management, industrial compliance, episode response, and scientific research.

Air Quality Management and Policy Design

Annual or episode-based PM2.5 and O3 source apportionment reports guide local authorities in prioritizing coal replacement, VOC reduction, dust suppression, or traffic restrictions, and in evaluating the effectiveness of clean air action plans.

Industrial Compliance and Emission Tracing

Industrial parks and their surrounding sensitive receptors are monitored to verify compliance with emission permits, identify fugitive leaks, and trace elevated metals or VOCs to specific processes, enabling corrective action before enforcement escalation.

Episode Diagnosis and Early Warning

During haze or ozone episodes, high-frequency speciation data combined with meteorological analysis determine whether an event is driven by local accumulation, hygroscopic secondary growth, regional transport, or biomass burning, improving emergency response and public health advisories.

Environmental Impact Assessment, Research, and Litigation Support

Source apportionment evidence supports environmental impact assessments for new projects, academic research on formation mechanisms, and environmental disputes where liability for contamination must be attributed to specific emission categories.

FAQ

What is the difference between primary and secondary pollutants?

Primary pollutants, such as dust from construction sites or elemental carbon from diesel exhaust, are emitted directly in particle or gas form. Secondary pollutants, including sulfate, nitrate, secondary organic aerosol, and ozone, are formed in the atmosphere through chemical reactions among SO2, NOx, NH3, and VOC precursors. Control strategies differ accordingly: primary sources are addressed with end-of-pipe measures, while secondary pollution requires coordinated precursor reduction.

How is the contribution of each pollution source calculated?

Chemical fingerprints measured in source samples, for example potassium and levoglucosan for biomass burning or aluminum and silicon for crustal dust, are statistically matched against the composition of ambient samples using receptor models such as PMF or CMB, yielding percentage contributions by mass with estimated uncertainties.

How many samples are needed for a valid study?

For annual apportionment, sampling across all four seasons with at least 60 to 120 valid 24-hour PM2.5 samples is generally recommended, supplemented by representative local source profiles. Episode studies may use hourly or multi-hour online data instead of, or combined with, filter sampling.

How long does a complete cause analysis project take?

A full one-year campaign including four seasonal sampling rounds, laboratory analysis, modeling, and reporting typically requires 12 to 16 months. Rapid episode diagnostic reports can be delivered within two to four weeks using online monitoring data.

Which standards apply to sampling and analysis?

Commonly applied references include GB 3095 for ambient air quality assessment, HJ 618 and HJ 656 for particulate sampling, HJ 644 and HJ 759 for VOC sampling and analysis, EN 12341 and EN 14907 for reference gravimetric methods, US EPA Compendium IO-3 and TO series methods, and ISO 9096 for stationary source particulate measurement.

Can cause analysis identify a single responsible facility?

Apportionment reliably quantifies category-level contributions such as coal combustion, vehicles, dust, industry, and biomass burning. Attribution to an individual facility generally requires dispersion modeling, tracer correlation, and facility-specific profile comparison performed under a supervised evidence protocol suitable for legal proceedings.