Air dispersion modelling is only as reliable as the weather data driving it. In India's diverse terrain — from coastal industrial clusters to inland plateaus and Himalayan valleys — the meteorology feeding a model can make the difference between a defensible Environmental Clearance submission and a rejected one. Getting the right surface and upper-air data, processing it correctly, and proving it represents your site is the foundation of every credible study.
Meteorological data for air dispersion modelling is the set of hourly surface observations (wind speed, wind direction, temperature, cloud cover) and upper-air soundings (vertical temperature and wind profiles) that drive models like AERMOD and CALPUFF. Processed through AERMET, this data defines atmospheric stability, mixing height, and turbulence — the parameters that govern how pollutants disperse.
What is meteorological data in air dispersion modelling?
Meteorological data describes the state of the atmosphere hour by hour. It tells the model how fast and in which direction the wind carries a plume, how turbulent the air is, and how high pollutants can mix before they are trapped.
Two categories of data are essential. Surface data captures conditions near ground level, while upper-air data describes the vertical structure of the atmosphere. Both are required to characterise the planetary boundary layer, the lowest part of the atmosphere where emissions actually disperse.
Without accurate meteorology, even a perfectly specified emission source produces meaningless concentration predictions. This is why regulators and consultants treat met data quality as a first-order concern.
Why does meteorological data matter for air quality modelling?
Dispersion is a physical process governed by wind and turbulence. The same stack emitting the same pollutant will produce very different ground-level concentrations under calm night-time conditions versus a windy afternoon.
Regulatory decisions hinge on predicted maximum concentrations. If the model uses unrepresentative weather, it may under-predict impacts and understate risk, or over-predict and impose unnecessary cost. Both outcomes weaken an EIA report.
For submissions to the CPCB, MoEFCC, and State Pollution Control Boards (SPCBs) under the EIA Notification 2006, reviewers routinely scrutinise the meteorological basis. A well-documented, site-representative dataset strengthens the credibility of the entire assessment.
Key reasons met data is decisive:
- It determines atmospheric stability class and mixing height.
- It defines the wind rose that shapes the direction of maximum impact.
- It controls plume rise, downwash, and dispersion coefficients.
- It underpins compliance checks against National Ambient Air Quality Standards (NAAQS).
What are surface and upper-air data?
These are the two pillars of any modelling meteorology dataset, and both are mandatory for AERMOD through its preprocessor AERMET.
Surface meteorological data
Surface data is measured at or near ground level, typically at a standard height of 10 metres for wind. It includes hourly wind speed, wind direction, dry-bulb temperature, relative humidity, cloud cover, and station pressure. Cloud cover is particularly important because it drives the estimation of solar radiation and, in turn, daytime turbulence.
On-site data from a dedicated meteorological station is the gold standard because it directly reflects local conditions. Where a full year of quality on-site data is unavailable, nearby India Meteorological Department (IMD) stations or prognostic data become the fallback.
Upper-air meteorological data
Upper-air data comes from radiosonde soundings that measure temperature, pressure, humidity, and wind as a balloon rises through the atmosphere. These profiles establish the morning and afternoon mixing heights and the potential temperature gradient that AERMET needs.
In India, upper-air soundings are launched from a limited network of IMD stations, so most sites are not close to one. This gap is frequently filled using prognostic meteorological models such as MM5 or the Weather Research and Forecasting (WRF) model.
Where does meteorological data come from in India?
Consultants draw on several sources, each with strengths and limitations. The right choice depends on site location, terrain, data availability, and the regulatory expectation for the project.
| Data source | Type | Strengths | Limitations |
|---|---|---|---|
| On-site met station | Surface | Most representative; site-specific | Requires a full year; instrument upkeep |
| IMD observatory | Surface / upper-air | Long records; officially recognised | Sparse network; may be distant |
| MM5 / WRF prognostic | Surface + upper-air | Fills spatial gaps; complete profiles | Model bias; needs validation |
| Reanalysis datasets | Surface + upper-air | Global coverage; gridded | Coarse resolution; smooths local effects |
On-site monitoring is often established during the EIA baseline study, giving projects a season or a full year of directly measured data. For upper-air inputs, WRF-derived profiles are now widely accepted when observed soundings are unavailable, provided the model configuration and validation are documented.
How is meteorological data processed with AERMET?
AERMET is the meteorological preprocessor for AERMOD. It ingests raw surface and upper-air data along with site surface characteristics and outputs the two files AERMOD needs: a surface file and a profile file.
The processing follows a defined sequence:
- Extract and quality-check the raw surface observations for gaps, spikes, and calm periods.
- Extract the upper-air soundings, or generate equivalent profiles from WRF/MM5 output.
- Assign surface characteristics — albedo, Bowen ratio, and surface roughness — for the area around the site, usually by wind-direction sector and season.
- Merge the datasets and compute boundary-layer parameters such as friction velocity, Monin-Obukhov length, and mixing height.
- Output the AERMOD-ready surface and profile files.
Surface characteristics deserve special attention. Albedo, Bowen ratio, and roughness length are derived from land use around the site, and getting these wrong can distort turbulence estimates. GIS-based land-use analysis makes this step defensible and reproducible.
Related tools complete the modelling chain. AERMAP processes terrain elevation for receptors and sources, while BPIP-PRIME handles building downwash where nearby structures influence the plume.
When do you need on-site data versus modelled data?
The decision balances representativeness against feasibility and timeline.
Use on-site measured data when:
- The terrain is complex and nearby stations do not capture local wind patterns.
- The project is large or sensitive and demands the strongest defensibility.
- The EIA schedule allows a full season or year of monitoring.
Use IMD or prognostic (WRF/MM5) data when:
- No representative surface station exists within a reasonable distance.
- Upper-air soundings are needed but no radiosonde station is nearby.
- Rapid screening or a supplementary comparison run is required.
In practice, many robust studies combine sources: on-site surface data paired with WRF-derived upper-air profiles, all reconciled inside AERMET. The guiding principle is representativeness, not the label of the source.
How do you assess whether met data is representative?
Representativeness is the single most challenged aspect of modelling meteorology. It asks a simple question: does this weather data actually describe the atmosphere over your site?
Assessment considers several factors:
- Spatial proximity — distance between the data station and the site.
- Terrain similarity — whether elevation, valleys, and slopes match.
- Land-use similarity — coastal versus inland, urban versus rural roughness.
- Temporal coverage — ideally a full, recent, and continuous year.
- Data completeness — a high percentage of valid hours, with calms handled correctly.
A useful cross-check is comparing the wind rose from the candidate dataset against known local climatology and any short-term on-site measurements. Large discrepancies in prevailing direction or calm frequency signal a representativeness problem that must be resolved before modelling proceeds.
How much meteorological data is required?
Regulatory practice for refined dispersion modelling generally calls for at least one full year of hourly, quality-assured meteorological data. A complete year captures the seasonal shift in wind patterns — for example, the reversal of monsoon flows — that a single season would miss.
Data completeness thresholds matter too. A dataset with excessive missing hours or unexplained gaps invites reviewer questions, so consultants aim for high valid-hour recovery and document how calms and missing periods were treated.
Where multiple years are available, selecting a recent and meteorologically typical year strengthens the study further, and multi-year runs may be justified for high-stakes projects.
Common pitfalls to avoid
Even experienced teams encounter recurring issues that undermine model credibility. Awareness of these prevents costly rework during regulatory review.
- Using a distant station whose terrain and land use differ from the site.
- Mismatched surface characteristics that do not reflect actual land cover.
- Ignoring calm-wind hours, which can dominate low-level accumulation.
- Combining surface and upper-air data from very different locations without justification.
- Undocumented WRF/MM5 configuration, leaving prognostic data unverifiable.
- Treating a partial season as if it were a full year.
How DSTechnoverse Helps
DSTechnoverse provides end-to-end air dispersion modelling services for industries and consultants across India, with meteorological data handling as a core strength. Our team approaches met data the way regulators do — with a relentless focus on representativeness and traceability.
Our process begins with a data sourcing and representativeness assessment. We evaluate on-site monitoring, the nearest IMD observatories, and prognostic options, then recommend the most defensible combination for your site and terrain. Where observed upper-air data is unavailable, we generate and validate profiles using WRF/MM5 modelling.
We process every dataset through AERMET, deriving surface characteristics — albedo, Bowen ratio, and roughness — from GIS-based land-use analysis for each wind sector and season. Terrain is handled with AERMAP, building downwash with BPIP-PRIME, and dispersion is modelled in AERMOD for near-field assessments or CALPUFF for long-range and complex-terrain scenarios.
Clients receive a detailed, review-ready report that includes:
- The meteorological data source, period, and completeness statistics.
- A representativeness justification with wind roses and land-use context.
- Full AERMET processing documentation and surface characteristic tables.
- Model configuration, receptor grid, and terrain treatment.
- Predicted ground-level concentrations compared against NAAQS.
- A clear methodology narrative aligned with CPCB, MoEFCC, and SPCB expectations under the EIA Notification 2006.
Whether you are preparing an EIA report, seeking Environmental Clearance, or supporting a Consent to Establish/Operate application, we deliver modelling that stands up to scrutiny. Contact DSTechnoverse to discuss your air dispersion modelling needs and get a meteorological approach tailored to your site.
Frequently Asked Questions
What meteorological parameters does AERMOD need?
AERMOD requires hourly surface data (wind speed, wind direction, temperature, and cloud cover), upper-air temperature and wind profiles, and site surface characteristics. AERMET combines these to compute boundary-layer parameters like mixing height, friction velocity, and the Monin-Obukhov length that AERMOD uses to disperse the plume.
Can I use WRF data instead of on-site measurements?
Yes. When representative observed data is unavailable — especially for upper-air profiles, given India's sparse radiosonde network — WRF or MM5 prognostic data is widely accepted. The key is documenting the model configuration and validating the output against any available local observations so the dataset is defensible during regulatory review.
How much meteorological data is needed for an EIA study?
Refined dispersion modelling generally requires at least one full year of continuous, quality-assured hourly data. A complete year captures seasonal wind reversals, including monsoon patterns, that a single season would miss. High data completeness and proper treatment of calm and missing hours are equally important.
What makes meteorological data representative of my site?
Representativeness depends on spatial proximity, terrain similarity, land-use similarity, recent and continuous temporal coverage, and high data completeness. Comparing the dataset's wind rose against local climatology is a practical check. If prevailing direction or calm frequency differs sharply, the data likely is not representative.
Why is upper-air data important if my sources are near the ground?
Upper-air soundings define mixing height and the vertical temperature structure, which control how high pollutants can disperse before being trapped. Even ground-level and low-stack emissions are strongly affected by the depth of the mixing layer, so accurate upper-air input is essential for realistic concentration predictions.
Which is better, AERMOD or CALPUFF, for Indian projects?
It depends on the scenario. AERMOD suits near-field, steady-state assessments over relatively simple terrain and is the common choice for most industrial EIA studies. CALPUFF is preferred for long-range transport, complex terrain, calm-wind conditions, and situations where wind fields vary significantly across the modelling domain.