Street-scale air quality dispersion modelling

Tools to setup, run and analyse the open source USEPA AERMOD dispersion model.

The aermodr R package is a set of tools to enable an AERMOD dispersion model to be created, run and analysed.

Package features

Meteorology derived for the site itself. The pipeline builds AERMOD-ready surface and profile meteorology directly from the ERA5 reanalysis, bypassing AERMET entirely. Boundary layer parameters — friction velocity, sensible heat flux, mixing heights, Monin–Obukhov length — are computed for the model domain’s own location, for any modelling year, anywhere in the world.

Surface roughness. The model is split into tiles, and the surface roughness derived from national LiDAR elevation data — 1 m resolution across England, 0.5 m AHN data across the Netherlands. Building and vegetation heights within the domain feed a morphometric (Macdonald) calculation of roughness length and displacement height, resolved by wind sector, so the model sees the actual urban form around the site: the terraces, the tower block, the open park upwind.

Tiled model runs that scale. AERMOD run times grow rapidly with sources, receptors, and hours. Large domains are split into self-contained tiles that run in parallel — and because each tile carries its own locally derived roughness and meteorology, the tiled model captures spatially varying surface characteristics that a single conventional AERMOD run cannot represent. Faster and more physically realistic.

Emissions with real temporal structure. Road sources are built from OpenStreetMap network geometry, overlaid on LIDAR to calculate gradients and derive segments, with hourly emission profiles (AERMOD HOUREMIS) so the diurnal and weekly rhythm of traffic is carried through to modelled concentrations.

How it works

flowchart LR
  A[ERA5 reanalysis] --> D[AERMOD-ready\nmeteorology]
  B[National LiDAR\nDSM/DTM] --> E[Sector-resolved\nroughness & terrain]
  C[OSM roads +\ngridded emissions] --> F[Hourly area\nsources]
  D --> G[Tiled AERMOD runs]
  E --> G
  F --> G
  G --> H[Evaluation against\nmonitoring data]
  H --> I[Interactive maps,\ntables & reports]

Scripted in R making studies reproducible and transparent: change the modelling year, the emissions scenario, or the domain boundary, and the entire chain — meteorology, roughness, sources, model runs, evaluation, report — regenerates consistently.

What you receive

Deliverables are tailored to the study, but a typical package includes:

  • Modelled concentration fields and receptor-level results for the pollutants and averaging periods that matter to your assessment (annual mean NO2, hourly exceedances, PM10/PM2.5).
  • Model evaluation against local monitoring — automatic comparison with reference-grade monitoring networks and diffusion tube data where available, so the model’s performance is demonstrated, not asserted.
  • Interactive maps and contour plots alongside publication-quality figures.
  • A written report (interactive HTML, project website and/or PDF) documenting inputs, methods, assumptions, and results — suitable for planning submissions, scheme appraisals, or internal decision-making.
  • On request: the full model input decks and configuration, so your team or a reviewer can re-run and scrutinise everything.

Typical applications

The approach suits assessments at the street-to-neighbourhood scale where urban form and local traffic dominate: development and planning applications, road scheme and junction changes, low-traffic neighbourhoods and cycle infrastructure, monitoring network design (where should the next site go?), and source apportionment questions — how much of what’s measured at a kerbside site is the road in front of it? Where a wider context is needed, local dispersion results can be combined with regional chemistry-transport model output to separate local increments from background.

Coverage

The pipeline is currently operational for England (Environment Agency LiDAR, OSGB) and the Netherlands (AHN LiDAR, RD New), with ERA5 meteorology available globally — extending the LiDAR roughness component to other territories with national elevation datasets is straightforward.

NoteBackground

Active Analytics builds on professional experience in atmospheric dispersion and air quality modelling, spanning street-scale Gaussian modelling (AERMOD, ADMS). The same evaluation tooling used in research — openair, meteorological normalisation, model–observation statistics.

Start a conversation

Every site is different, and the honest first step is a short scoping discussion: what question the modelling needs to answer, what data already exists, and whether dispersion modelling is even the right tool for it.

Get in touch to discuss your project, or read more about other services.

Back to top