Air quality monitoring and dispersion modelling

In short

Air quality monitoring measures what is actually present in the ambient air or in a stack; dispersion modelling predicts what a given emission will do to ground-level concentrations downwind. Together they close the loop of air-pollution control: monitoring tells you whether the National Ambient Air Quality Standards (NAAQS) are met, modelling tells you how tall a stack must be, where to site an industry, and what an expansion will add. In India this is statutory work under the Air (Prevention and Control of Pollution) Act, 1981 and the Environment (Protection) Act, 1986 - every EIA report, consent-to-operate application and NCAP city action plan contains both a monitoring dataset and a model run. It is a standard GATE (Environmental Science and Engineering) scoring area - Gaussian plume, stability classes, effective stack height, AQI - and a routine interview topic for CPCB/SPCB, NEERI, EIA consultancies, cement, steel, refinery and thermal-power environment cells.

Explained

Why monitor, and under what law

• Objectives: check compliance with standards, establish a baseline before a project, find trends and sources, trigger episodic action (GRAP), validate models, and support epidemiological studies.
• Air Act, 1981: CPCB sets/advises standards, SPCBs enforce; the State Government, in consultation with the SPCB, declares air pollution control areas (section 19); consent and emission standards for industry.
• Environment (Protection) Rules, 1986: Schedule I gives industry-specific emission/effluent standards, Schedule VI general standards; stack monitoring and online monitors are consent conditions.
• NAAQS, notified 18 November 2009 under section 16(2)(h) of the Air Act: 12 parameters, one set for industrial/residential/rural areas and a stricter set for ecologically sensitive areas notified by the Centre. (CPCB has been reviewing/tightening these; verify the current notification before quoting values in a report.)
• National Clean Air Programme (NCAP), 2019: 131 non-attainment/million-plus cities, originally a 20-30% particulate reduction by 2024 against 2017, revised to 40% (or NAAQS compliance) by 2025-26; city action plans are funded partly through the 15th Finance Commission air-quality grants.
• Delhi-NCR: the Commission for Air Quality Management (CAQM), a statutory body since the CAQM Act, 2021, operates the Graded Response Action Plan (GRAP) - staged curbs (Stage I to IV) keyed to AQI bands.

Ambient monitoring: network design and sampling protocol

• Siting criteria: 3-10 m above ground (breathing-zone representativeness), away from obstructions by at least twice their height, 20 m from trees, no local chimney/kitchen/dusty road influence, unrestricted 270° airflow, secure power and access.
• Station types: background/upwind, source-oriented (industrial), traffic kerbside, residential/receptor, and downwind maximum-impact station chosen from a wind rose or a model run.
• Number of stations: depends on population, area, pollutant variability and objective; CPCB guidelines give population-based minima, and EIA baselines typically use 8-10 stations over one non-monsoon season (3 months).
• NAMP (National Air Quality Monitoring Programme): manual network run by SPCBs/NEERI, now over a thousand stations in 500+ cities (count changes - check the current CPCB figure). Core protocol: 24-hourly sampling twice a week, 104 observations per station per year; particulates over 24 h, gases as 4-hourly or 8-hourly averages within the 24 h.
• CAAQMS (Continuous Ambient Air Quality Monitoring Stations): real-time analysers reporting 15-minute/hourly data to the CPCB Central Control Room, feeding the live AQI. Several hundred are operating nationally and the number grows every year.
• Averaging and completeness: NAAQS annual values are the arithmetic mean of a minimum of 104 measurements taken twice a week, uniformly spread over the year; 24-h and 8-h values may be exceeded 2% of the time in a year but never on two consecutive days.
• Wind rose: polar plot of frequency of wind direction by speed class - decides the predominant downwind sector, hence siting and plant layout.

Measurement methods

• PM10 - respirable dust sampler (RDS): air at about 1.0-1.2 m³/min enters a cyclone that removes >10 µm particles; the PM10 fraction is collected on a 20×25 cm glass/quartz-fibre filter and weighed (gravimetric, after desiccation and conditioning).
• PM2.5: fine-particulate sampler with a WINS impactor or sharp-cut cyclone at 16.7 L/min on a 47 mm Teflon/quartz filter; gravimetric reference method.
• TSP/SPM - high volume sampler (HVS): total suspended particulate collected without size selection; largely replaced for compliance by PM10/PM2.5.
• Continuous particulate methods: beta-attenuation monitor (BAM) - absorption of beta rays by deposited mass; TEOM - frequency shift of a vibrating tapered element; optical/nephelometric sensors (useful for trends, needs correction against a reference, basis of low-cost sensor networks).
• Gaseous sampling: air is drawn through an impinger/bubbler with a specific absorbing reagent at about 0.5-1 L/min, then analysed in the laboratory:
  • SO₂ - modified West-Gaeke (pararosaniline) method, absorbed in sodium tetrachloromercurate, colorimetric at 560 nm.
  • NO₂ - modified Jacobs-Hochheiser / sodium-arsenite method, absorbed in alkaline solution, azo dye measured at 540 nm.
  • NH₃ - indophenol blue; O₃ - chemiluminescence or UV photometry; CO - NDIR; benzene/BTX - adsorbent tube with GC analysis; lead, As, Ni - filter digestion with AAS/ICP; BaP - PAH extraction with HPLC/GC-MS.
• Continuous analysers: SO₂ by UV fluorescence, NOx by chemiluminescence (with a molybdenum converter), O₃ by UV absorption, CO by NDIR/GFC, hydrocarbons by FID.
• Indian test methods are the IS 5182 series, "Methods for measurement of air pollution" (separate parts for SPM, SO₂, NO₂, CO, respirable PM, sampling planning); several parts have been revised or withdrawn, so quote the part number only after confirming the current edition. Stack monitoring follows CPCB's emission regulation/stack-sampling protocols and relevant IS methods.
• Stack (source) monitoring: isokinetic sampling through a port at a straight length of 8D downstream/2D upstream of a disturbance, using a pitot tube for velocity, thimble/filter for particulates, and gas conditioning for SO₂/NOx; results reported corrected to a reference oxygen/CO₂ and to normal conditions (Nm³, 25 °C, 1 atm, dry). Large sources carry CEMS (continuous emission monitoring systems) linked to the SPCB server.
• Quality assurance: flow calibration, filter conditioning and blank correction, span/zero checks, duplicate and collocated sampling, reporting detection limits, and data capture of at least 75-80% for an average to be valid.

National Air Quality Index (AQI)

• Launched 2014 (CPCB, "One Number - One Colour - One Description"), 8 pollutants: PM10, PM2.5, SO₂, NO₂, CO, O₃, NH₃, Pb.
• Six categories: Good 0-50, Satisfactory 51-100, Moderate 101-200, Poor 201-300, Very Poor 301-400, Severe 401-500 (above 500 is reported as Severe+ in GRAP practice).
• Averaging: 24 h for PM10, PM2.5, SO₂, NO₂, NH₃, Pb; 8 h maximum rolling for CO and O₃.
• A sub-index is computed for each pollutant by linear interpolation inside its breakpoint band; the AQI is the highest sub-index, and that pollutant is named the "prominent/responsible pollutant".
• Validity: at least three pollutants, one of which must be PM10 or PM2.5, otherwise the AQI is not reported.

Meteorology that controls dispersion

• Lapse rate: rate of temperature fall with height. Dry adiabatic lapse rate (DALR) ≈ 9.8 °C/km (about 1 °C per 100 m); the measured value is the environmental lapse rate (ELR).
• Stability: ELR > DALR → superadiabatic, unstable (strong mixing, looping plume); ELR = DALR → neutral (coning); ELR < DALR → subadiabatic, stable (fanning); ELR negative (temperature rising with height) → inversion, the worst case.
• Plume shapes: looping (unstable), coning (neutral), fanning (stable/inversion), lofting (inversion below, unstable above - best case, plume rises away), fumigation (inversion above, unstable below - plume trapped and brought down, highest ground concentrations, typical of morning inversion break-up), trapping (inversions above and below).
• Inversions: radiation (nocturnal, surface), subsidence (high pressure, elevated - behind Delhi's winter smog), frontal, and marine/advection inversions.
• Mixing height: depth of the layer available for mixing, from the morning temperature sounding; low mixing height (a few hundred metres) in winter mornings multiplies concentrations. Ventilation coefficient = mixing height × average wind speed (m²/s) - low values mean poor dispersion potential.
• Pasquill-Gifford stability classes A to F: A very unstable, D neutral, F very stable; assigned from surface wind speed plus incoming solar radiation (day) or cloud cover (night).
• Wind profile: speed increases with height; needed to convert a 10 m measurement to stack-top wind.

Dispersion modelling

• Idea: a plume from a continuous point source spreads by turbulent diffusion; the concentration distribution across and above the plume axis is taken as Gaussian (normal), with standard deviations σy (crosswind) and σz (vertical) that grow with downwind distance x and depend on stability.
• Assumptions (examinable): steady state, continuous emission, constant wind speed and direction, no chemical reaction or deposition, flat terrain, total mass conserved, complete reflection at the ground (image source), and no dispersion along the wind (advection dominates).
• σy, σz are read from Pasquill-Gifford curves or power-law fits σ = a·x^b per stability class - unstable classes give large σ, hence rapid dilution.
• Plume rise Δh: hot, fast exhaust rises above the stack because of momentum (velocity) and buoyancy (temperature difference). Effective stack height H = hs + Δh − downwash correction; Holland's and Briggs' equations are the examined estimates. Stack-tip downwash occurs when vs < 1.5u; building wake/cavity downwash is avoided by making the stack at least 2.5 times the height of nearby buildings.
• Maximum ground-level concentration occurs at the distance where σz = H/√2; dilution improves with the square of effective height - doubling H quarters the peak, which is why stack height is the cheapest first control and also why it only relocates, not removes, the pollution.
• Stack height rules (CPCB): for SO₂, H = 14 Q^0.3 (H in m, Q in kg/h of SO₂), with a minimum height prescribed for the category of plant; for DG sets, stack height = building height + 0.2√(kVA). Stack height alone is never a substitute for control equipment.
• Model families: Gaussian plume (ISCST3, the older regulatory workhorse; AERMOD, the current US-EPA and Indian regulatory preference with boundary-layer scaling and terrain handling); puff models (CALPUFF, INPUFF) for calm winds, long range and variable meteorology; box models for city-wide mass balance; Eulerian grid/photochemical models (CMAQ, WRF-Chem, CAMx) for secondary PM and ozone; Lagrangian trajectory models (HYSPLIT) for source apportionment and long-range transport; CFD for street canyons and building wakes.
• Model inputs: emission inventory (rate, stack height and diameter, exit velocity and temperature), hourly surface and upper-air meteorology, terrain elevation, land use and receptor grid. Outputs: isopleth maps of predicted short-term and annual increments, which are added to the measured baseline and compared with NAAQS.
• Indian practice: the EIA Notification, 2006 (under EPA 1986) requires one season of baseline AAQ data and a dispersion model run in the EIA/EMP chapter; CPCB/IITM's SAFAR and the Delhi Decision Support System provide operational forecasts and source apportionment.
• Limitations: Gaussian models are poor in calm winds (u < 1 m/s), complex terrain, coastal fumigation and for reactive/secondary pollutants; results are only as good as the emission inventory; model performance must be stated with validation against monitored data.

Key points

• NAAQS 2009 - 12 parameters, two area categories; annual means need ≥104 observations spread uniformly over the year.
• 2% exceedance rule - a 24-h/8-h standard may be exceeded 2% of days a year, never on two consecutive days.
• NAMP - manual network, 24-h sampling twice weekly; CAAQMS - real-time analysers feeding the live AQI.
• PM10 by RDS (cyclone + filter, ~1.1 m³/min), PM2.5 by impactor (16.7 L/min); gravimetric methods are the reference.
• BAM and TEOM - continuous particulate mass by beta attenuation and vibrating-element frequency shift.
• West-Gaeke - SO₂; sodium arsenite/Jacobs-Hochheiser - NO₂; NDIR - CO; chemiluminescence - NOx; UV photometry - O₃.
• Isokinetic sampling - stack sample must enter the nozzle at the same velocity as the gas stream, else particulate mass is biased.
• AQI - 8 pollutants, 6 bands 0-500, equals the highest sub-index; needs ≥3 pollutants including PM10 or PM2.5.
• DALR ≈ 1 °C/100 m - the yardstick for stability; superadiabatic = unstable, inversion = worst dispersion.
• Plume shapes - looping, coning, fanning, lofting (best), fumigation and trapping (worst).
• Mixing height × wind speed = ventilation coefficient - the city's dilution capacity.
• Pasquill classes A-F from wind speed with insolation or cloud cover; σy, σz grow with x and with instability.
• Gaussian plume - concentration ∝ Q/(u σy σz), with an image source for ground reflection.
• Effective stack height H = hs + Δh; peak ground concentration ∝ 1/H², so height dilutes but does not reduce mass.
• Max GLC at σz = H/√2, closer to the stack in unstable air, far away in stable air.
• CPCB SO₂ stack rule H = 14 Q^0.3; DG set stack = building height + 0.2√kVA.
• AERMOD/ISCST3 for regulatory point sources, CALPUFF for calms and long range, CMAQ/WRF-Chem for secondary pollutants.
• NCAP 2019 - 131 cities, 40% particulate reduction target by 2025-26; GRAP under CAQM for Delhi-NCR episodes.

Formulas and examples

NAAQS 2009 - frequently asked values (µg/m³ unless stated)

Example 1 - PM10 by gravimetry. An RDS runs at 1.2 m³/min for 24 h; filter mass rises from 3.4520 g to 3.6020 g.

• Volume sampled = 1.2 × 60 × 24 = 1,728 m³.
• Mass collected = 0.1500 g = 150,000 µg.
• C = 150,000/1,728 = 86.8 µg/m³ - below the 24-h PM10 standard of 100 µg/m³.

Example 2 - ground-level concentration under the plume. Q = 100 g/s of SO₂, u = 6 m/s at stack top, effective height H = 100 m. At x = 1 km, stability class C gives σy = 100 m, σz = 61 m. Find the centreline ground concentration.

• π u σy σz = 3.1416 × 6 × 100 × 61 = 114,982.
• Q/(π u σy σz) = 100/114,982 = 8.697×10⁻⁴ g/m³.
• exp(−H²/2σz²) = exp(−10,000/(2 × 3,721)) = exp(−1.3438) = 0.2608.
• C = 8.697×10⁻⁴ × 0.2608 = 2.268×10⁻⁴ g/m³ = 227 µg/m³.

Example 3 - plume rise and effective height. Stack height 60 m, d = 2 m, vs = 12 m/s, Ts = 420 K, Ta = 293 K, u = 5 m/s, P = 1,000 mb.

• Bracket: 2.68×10⁻³ × 1,000 × 2 × (127/420) = 5.36 × 0.30238 = 1.6208; add 1.5 → 3.1208.
• vs d/u = 12 × 2/5 = 4.8 m.
• Δh = 4.8 × 3.1208 = 14.98 ≈ 15 m, so H = 75 m.
• By Briggs: F = 9.81 × 12 × 4 × 127/(4 × 420) = 59,802/1,680 = 35.6 m⁴/s³; at x = 300 m, Δh = 1.6 × 35.6^1/3 × 300^2/3/5 = 1.6 × 3.291 × 44.80/5 = 47.2 m - Briggs gives a much larger rise for a strongly buoyant plume, which is why the model and formula used must always be stated.

Example 4 - maximum ground-level concentration. Q = 72 g/s, u = 5 m/s, H = 50 m, σz/σy = 0.5 at the critical distance.

• π e u H² = 3.1416 × 2.718 × 5 × 2,500 = 106,747.
• C_max = (2 × 72/106,747) × 0.5 = 1.349×10⁻³ × 0.5 = 6.75×10⁻⁴ g/m³ = 675 µg/m³.
• If the stack is made 100 m effective, C_max falls by 1/4 to about 169 µg/m³ - the inverse-square dependence on H.

Example 5 - AQI from a day's data. PM2.5 = 105, PM10 = 220, NO₂ = 95 µg/m³ (24-h averages).

• PM2.5 band 91-120 → index 201-300: I = 201 + (99/29)(105 − 91) = 201 + 47.8 = 249.
• PM10 band 101-250 → index 101-200: I = 101 + (99/149)(220 − 101) = 101 + 79.1 = 180.
• NO₂ band 81-180 → index 101-200: I = 101 + (99/99)(95 − 81) = 115.
• AQI = 249, "Poor", prominent pollutant PM2.5. (Three pollutants including PM2.5, so the AQI is valid.)

Example 6 - stability and stack height.

• Air is 26.0 °C at the surface and 23.5 °C at 300 m: ELR = 2.5/300 = 0.83 °C/100 m < 0.98 → subadiabatic, stable, fanning plume.
• A plant emits 500 kg SO₂/h: H = 14 × 500^0.3 = 14 × 6.454 = 90.4 m, so a 90 m stack (subject to the prescribed minimum for the category).
• Wind 3 m/s at 10 m, neutral p = 0.15: u at 90 m = 3 × (9)^0.15 = 3 × 1.3867 = 4.16 m/s - use this, not the 10 m value, in the plume equation.

Where it is used

• EIA and environmental consultancies (NEERI, TCE, Mott MacDonald, Ramky, ABC Techno Labs): one-season baseline AAQ monitoring at 8-10 sites plus AERMOD runs for every category-A project - cement, steel, refineries, power, ports, highways.
• Thermal power, cement, steel and refineries: CEMS on stacks, ambient stations at the plant boundary, stack-height design and periodic stack monitoring as consent conditions of the SPCB.
• Regulators: CPCB and SPCBs/PCCs run NAMP and CAAQMS, publish the daily AQI bulletin, and use models for non-attainment city action plans; CAQM operates GRAP in Delhi-NCR.
• Job roles: Environmental Engineer/Scientist, air-quality modeller, NABL-accredited laboratory analyst and sampling technician, EHS officer, CEMS/CAAQMS O&M engineer, Junior Research Fellow on air-quality projects.
• Research and forecasting: IITM Pune's SAFAR for Delhi, Mumbai, Pune and Ahmedabad; IIT source-apportionment studies; TERI and CSTEP emission inventories.
• Urban and industrial planning: siting of industry downwind of settlements using a wind rose, buffer-zone and green-belt width, and traffic/construction-dust management under city air action plans.

Interview questions

• Q: Why is the Gaussian distribution assumed for a plume? A: Turbulent eddies transport parcels randomly, so the ensemble-averaged concentration across the plume follows a normal distribution about the axis; it is an empirically validated simplification of the diffusion equation under steady conditions.
• Q: What does the second exponential term in the Gaussian plume equation represent? A: Reflection of the plume at the ground, modelled as an imaginary mirror source at −H, assuming no deposition or absorption.
• Q: Why does ground-level concentration fall as the square of effective stack height? A: The plume has to mix through a greater vertical depth before reaching the ground, and the exponential term with H²/2σz² combined with the critical σz condition makes C_max proportional to 1/H².
• Q: Is a taller stack a pollution control measure? A: It only disperses and relocates the same mass over a wider area, lowering local concentrations; it does not reduce emissions, so CPCB treats it as a dispersion aid, not a substitute for control equipment.
• Q: Which plume condition gives the highest ground-level concentration and why? A: Fumigation - an inversion above traps the plume while unstable air below brings it rapidly down, typically when a nocturnal inversion breaks up in the morning.
• Q: Difference between lofting and fanning? A: Lofting has a stable layer below and unstable air above, so the plume rises away from the ground (favourable); fanning occurs within an inversion, giving a thin ribbon that may touch down far downwind unchanged.
• Q: How is the Pasquill stability class determined in practice? A: From surface wind speed together with incoming solar radiation during the day or cloud cover/ceiling at night, read off the standard A-F table.
• Q: How is the National AQI computed? A: A sub-index is interpolated for each pollutant within its breakpoint band from its 24-h (or 8-h for CO and O₃) average, and the highest sub-index is the AQI, with that pollutant named as prominent; at least three pollutants including PM10 or PM2.5 are needed.
• Q: Why must stack sampling be isokinetic? A: If sampling velocity is lower than the gas velocity, inertia carries excess coarse particles into the nozzle and overstates concentration; if higher, coarse particles are excluded and it understates it.
• Q: PM10 and PM2.5 - what does the number mean? A: The aerodynamic diameter at which the sampler's inlet has 50% collection efficiency, i.e. 10 µm and 2.5 µm cut-points, not a sharp upper limit.
• Q: Why do Holland and Briggs give different plume rise for the same stack? A: Holland is an empirical momentum-plus-buoyancy fit with a conservative bias; Briggs scales with the two-thirds law from buoyancy flux and is generally preferred for hot buoyant plumes - the formula used must be reported.
• Q: Your model predicts compliance but monitoring shows exceedance. What is wrong? A: Most often an incomplete emission inventory (fugitive, area and line sources missed), unrepresentative meteorology, calm-wind or terrain conditions outside the model's validity, or secondary particulate formation the Gaussian model cannot represent.

Common mistakes

• Using the 10 m wind speed instead of the wind at effective stack height, and forgetting the power-law correction.
• Substituting the physical stack height for H and omitting plume rise - or adding plume rise twice.
• Mixing units: Q in kg/h with σ in m, or reporting g/m³ instead of µg/m³; concentrations must also be reported at a stated temperature and pressure (Nm³ for stack gas).
• Doubling the concentration for ground reflection at z = 0 and keeping both exponential terms - at the ground the two terms are already equal, so the correct ground-level form has π, not 2π, in the denominator.
• Using the maximum-concentration formula at an arbitrary distance instead of only at the critical distance where σz = H/√2.
• Taking the AQI as an average of sub-indices, or computing it from a single pollutant, or mixing 1-h data into a 24-h sub-index.
• Comparing a single grab sample with an annual NAAQS value, or an uncorrected low-cost sensor reading with a gravimetric standard, and ignoring the ≥104-observation and 75% data-capture requirements.

Related topics

Air pollution and control · Air pollution control equipment design: cyclones, ESP, bag filters and scrubbers · Noise pollution · Environmental impact assessment and environmental audit · Meteorology and micrometeorology · Environmental legislation and standards in India.