Skip to main content

What a noise impact assessment has to contain

Environmental Assessment · SoundPLAN Asia · · 8 min read

A noise impact assessment answers two questions for a regulator: what will the noise level be at the people nearest to this project, and is that acceptable under the rules that apply here. Everything in the report exists to make those two answers checkable by somebody who was not there. Reports get returned not because the modelling was wrong but because a reviewer could not follow how a number was reached.

The structure that survives review

  1. Scope and criteria. Which regulation or project standard applies, which indicators it uses, which limits, and at what times of day. Settle this first, because it determines everything downstream. A report that models beautifully against the wrong criterion is worthless.
  2. Baseline. Measured existing conditions at representative positions, with dates, durations, instrument details, calibration, and weather. State what the measured levels were actually controlled by - if the baseline position was dominated by a road, say so.
  3. Source inventory. Every source that will exist, with its sound power level and the provenance of that figure: measured, vendor data sheet, or a published reference. This is the single most-audited table in the document.
  4. Construction phase prediction. Modelled per stage, not averaged across the programme.
  5. Operational phase prediction. The steady-state condition, plus any abnormal condition that is reasonably foreseeable.
  6. Assessment against criteria at each receptor, for each relevant period.
  7. Mitigation, with predicted performance and a residual assessment after it is applied.
  8. Uncertainty. Stated, not implied.

Construction and operation are different assessments

They are routinely merged, and merging them hides the impact that residents actually complain about. Construction noise is temporary, its sources move as the works progress, its hours are usually restricted, and it frequently has impulsive character from piling or breaking. Operational noise is continuous, fixed in position, and assessed against a different limit - often a stricter one at night.

Model construction in phases - earthworks, foundations, structure, fit-out - each with its own plant list and working hours. A single "construction" scenario using the average plant complement will understate the worst weeks and overstate the quiet ones, and the worst weeks are the ones that generate the complaints and the enforcement.

Choosing receptors properly

Three errors recur:

  • Ground floor only. On a multi-storey facade the controlling receiver is often an upper floor, where a barrier or an intervening building no longer screens the source. Assess the floors, not the address.
  • Existing receptors only. If there is consented development between the project and the nearest current housing, the assessment needs to consider it. Consent status is a planning question worth asking explicitly.
  • Dwellings only. Schools, hospitals, places of worship and, in some jurisdictions, designated quiet areas often carry their own criteria and their own sensitive hours. A school matters during the day; a hospital matters at night.

The parts that get skipped

Low frequency and tonal content. Most criteria are A-weighted, and A-weighting substantially discounts low frequency. A plant can comply on an A-weighted basis and still produce the rumble that drives complaints. If a source has a dominant tone or strong low-frequency content, the assessment should say so and apply whatever character penalty the applicable criterion provides, because the propagation calculation will not add one by itself.

Cumulative effects. A project assessed in isolation can be compliant while the combined level from it and the neighbouring facility is not. Where other industrial sources exist nearby, the assessment needs to address the combination, even if the criterion is written for the project alone.

Uncertainty. A predicted level carries a spread - from source data provenance, ground assumptions, meteorology and receiver geometry. A report that states 54.8 dB at a receiver against a 55 dB limit, with no uncertainty statement, is claiming a precision the method does not support. Stating the spread is not a weakness; it is what makes the prediction credible, and it lets the reviewer see whether the margin is real.

Mitigation, in the order that actually works

  1. Layout. Distance and orientation are free at design stage and unobtainable later. Where the noisiest plant sits relative to the boundary decides more than any barrier will.
  2. Source selection. Specifying a quieter unit in the purchase order costs a fraction of enclosing a louder one after commissioning.
  3. Screening. Barriers, bunds, or positioning a building to do the screening. Effective only where the line of sight is genuinely broken, and limited at low frequency.
  4. Enclosure and attenuation. Acoustic enclosures, silencers, lagging. Reliable and well understood - and the most expensive route, which is why it should be the last one considered, not the first.

A noise impact assessment is ultimately a traceability exercise. If every number in it can be followed back to a measurement, a data sheet or a stated assumption, it will survive review. See how SoundPLAN is used for EIA noise chapters, or tell us which criterion your project must meet.

Environmental impact assessment · ISO 9613-2 in practice · Construction noise

All articles