Timing air quality monitoring and annualising the data

A diffusion tube survey run for four weeks tells you almost nothing about compliance with an annual mean objective. The nitrogen dioxide objective is 40 micrograms per cubic metre averaged over a calendar year, and a short survey captures a particular slice of weather, traffic and season. Before that number means anything it has to be bias adjusted and then annualised, and both steps need data you do not generate yourself.

Small businesses meet this when a council asks for baseline monitoring to support a change of use or an extraction scheme, and the usual reaction is to commission the shortest survey that sounds respectable. It is the wrong instinct, because a survey too short to annualise defensibly is money spent twice.

Why a short survey needs adjustment

Two separate corrections apply, and they do different jobs.

Bias adjustment corrects for the systematic difference between what a passive diffusion tube reports and what a reference analyser would have measured at the same place. Tubes from different laboratories, prepared by different methods, over-read or under-read by amounts that are measurable and are published.

Annualisation corrects for the period sampled. Nitrogen dioxide concentrations vary through the year, generally higher in winter under stable conditions and lower in summer. A survey run over January and February will overstate the annual mean. One run in July and August will understate it.

Apply both, in that order, and say in the report that you did. An assessment that presents raw tube results against the 40 microgram objective will be sent back by any environmental health officer who has read the national guidance.

The monitoring duration usually expected

A full calendar year of monitoring is the position the local air quality management guidance works from, because it needs no annualisation at all.

Where that is not practical, the survey should be long enough for annualisation to be meaningful, and short surveys of one or two monthly exposure periods are weak evidence. Six months or more spanning both heating and non-heating seasons is a far stronger position, and some councils will specify a minimum in pre-application correspondence. Ask them before you start, in writing.

Data capture matters as much as duration. The guidance works to a 75 per cent data capture target, and a tube that is lost, stolen or damaged takes a whole monthly period with it. Deploy duplicate or triplicate tubes at the locations that matter, because the cost of an extra tube is trivial against the cost of a gap.

Bias adjustment and where the factors come from

Two routes to a bias adjustment factor, and both are acceptable.

The national factor, taken from the published bias adjustment spreadsheet maintained for local authority monitoring, which collates co-location studies across the country by laboratory and preparation method. Use the factor for the correct laboratory, the correct preparation method and the correct year, and state all three.

A local co-location study, where tubes are exposed alongside a reference analyser at a nearby continuous monitoring station for the duration of the survey. This is more accurate and more expensive, and it is the better option where the result is likely to be contested.

State the factor applied, its source, its year and the number of studies behind it. A bias factor quoted without provenance is the single most common weakness in small-scheme air quality submissions.

Annualisation against continuous stations

Annualisation uses long-running continuous monitoring stations to work out how the survey period compared with the year as a whole.

The method takes the ratio between the annual mean and the period mean at each reference station, then applies the average of those ratios to your survey result. The national guidance expects the ratios to be derived from more than one station, typically at least three, chosen to be reasonably close to the site and of comparable site type.

Choosing the stations is where judgement enters. A roadside station and an urban background station behave differently through the year, and a rural background site tells you little about a town centre. State which stations were used, their site classification, their distance from the survey location and their data capture for the year in question.

Report the unadjusted result, the bias adjusted result and the annualised result separately. Anyone checking the work should be able to follow each step.

Seasonal variation and what it does to the answer

The seasonal pattern in nitrogen dioxide is real and it is large enough to change a conclusion.

Winter brings higher domestic and commercial combustion, more stable atmospheric conditions that trap pollutants near the ground, and lower mixing heights. Summer brings the opposite, along with more photochemical activity that affects the balance between nitrogen oxide and nitrogen dioxide.

Traffic patterns add their own seasonality. School term times raise weekday morning flows. August and the Christmas fortnight depress them, which is why traffic count methodology excludes those periods, as traffic counts in neutral months and why August fails sets out. The same reasoning applies to air quality monitoring undertaken to support a traffic-related assessment, and parking surveys follow a parallel rule set out in parking surveys and the timing the methodology requires.

If the survey has to be short, span the seasons rather than running consecutive months in one part of the year.

Programme lead times for a baseline

Work back from the submission date, because this is a long-lead item that people treat as a short one.

Allow time to agree the monitoring locations with the environmental health officer, which is worth doing in writing and takes a couple of weeks. Then tube supply and deployment. Then the monitoring period itself, in monthly exposure periods aligned to the national calendar where possible. Then laboratory turnaround, typically two to four weeks after each period. Then analysis and reporting.

A six month survey therefore occupies the better part of eight months from first contact to final report. For a business fitting out premises, that is the item that decides the programme, not the drawings.

Where the timescale cannot be met, say so and offer an alternative. Modelling using published background maps, verified against the nearest available monitoring, is often acceptable for a smaller proposal. Agree it with the officer first rather than presenting it as a fait accompli.

Presenting the data in the assessment

Set the monitoring section out so that each step can be checked.

Include a location plan showing tube positions with grid references, a photograph of each location, the height above ground and the distance from the kerb and from the nearest relevant receptor. Then a table of monthly results with exposure start and end dates, data capture, the bias factor applied, the annualisation factor and the final annualised mean. Then a short statement of uncertainty.

Say where the tubes sit relative to relevant exposure. A tube on a lamp post beside a busy junction is not measuring at a location where anyone is exposed for an annual period, and the objective applies where relevant exposure occurs. Getting this wrong produces alarming numbers that mean nothing.

The monitoring then feeds the wider air quality assessment, and where the proposal involves cooking extraction the odour and emission points need their own treatment in a ventilation and extraction statement.

One honest limitation to state in the report. An annualised estimate from a short survey carries real uncertainty, and quoting it to two decimal places implies a precision the method does not have. Round sensibly and say what the uncertainty is.

This is general guidance rather than legal advice. Check the position with your local planning authority or a qualified professional before relying on it.