Noise mapping for smart cities: what the data is actually for
Urban Planning · SoundPLAN Asia · · 8 min read
Most published city noise maps are used as illustrations. The coloured contours appear in a strategy document, everyone agrees the ring road is loud, and nothing downstream changes. That is a waste of the most expensive part of the exercise, because a strategic noise map is not a picture of a city. It is a population exposure dataset, and the exposure numbers are what change decisions.
The quantities, and why there are two of them
Strategic mapping is normally reported as two indicators. Lden is a day-evening-night level that adds a 5 dB penalty to the evening period and a 10 dB penalty to the night, reflecting that the same physical level is more intrusive later. Lnight is the night period alone, and it is the one that correlates with sleep disturbance.
The reason both are reported is that they drive different interventions. A high Lden along a commercial corridor is an amenity and annoyance problem. A high Lnight in a residential block is a health problem, and it is usually the one that justifies capital spending.
What the map is built from
A strategic map is only as good as four input layers, and in practice one of them is always the weak link:
- Terrain and buildings. A digital terrain model plus building footprints with heights. Building heights matter more than people expect: they determine screening, facade reflection and which floors are exposed.
- Source data. For roads, annual average daily traffic split by vehicle category, speed and road surface. For rail, train types, speeds and track condition. For industry, sound power levels per plant. For aircraft, flight tracks and fleet mix.
- Population. Dwellings assigned to buildings, and people assigned to dwellings. This is the layer that converts a contour map into an exposure statistic, and it is the layer most often approximated.
- The calculation method. CNOSSOS-EU is the common method under the European Environmental Noise Directive, and it is increasingly the reference elsewhere because it is documented, harmonised and reproducible. Using a national legacy method instead is legitimate, but it makes comparison with other cities meaningless.
Exposure statistics are the output that matters
The deliverable that changes a budget is not the map. It is a table: how many people are exposed above each threshold, broken down by source. Once a city can say that a given number of residents experience Lnight above a stated value, and that most of that exposure comes from one corridor rather than being spread evenly, the conversation moves from aesthetics to prioritisation.
That framing also exposes the interventions that do not work. Resurfacing a road with a low-noise surface might reduce levels by a few decibels along its whole length but affect relatively few dwellings if the corridor is commercial. A single barrier beside a residential tower might affect far more people for less money. Without exposure counts, both look like reasonable ideas; with them, one is obviously better.
Where quiet areas fit
Protecting places that are still quiet is cheaper than fixing places that are not, and it is the part of the framework most often skipped. Identifying and designating quiet areas - urban parks, courtyards, campus interiors - gives a planning authority a defensible basis for refusing or conditioning development that would degrade them. It only works if the designation exists before the development application arrives.
The "smart" part, used honestly
The appeal of continuous monitoring is obvious: sensors across a city, live data, a dashboard. The limitation is equally obvious once stated. A sensor measures one point. A city has millions of receiver points, most of them on facades at heights no sensor occupies. Monitoring cannot replace calculation, and a dashboard of sensor readings is not a noise map.
What monitoring does well is validate and update the model. Measured data at representative points confirms that the traffic assumptions are right, catches sources the model missed, and provides a defensible calibration record. The productive architecture is a calculated map as the spatial backbone, with monitoring used to keep its inputs honest over time - not two parallel systems that disagree.
Making the map operational
A strategic map earns its cost when it is wired into decisions that happen anyway:
- Planning applications are screened against it, so a residential proposal on an exposed facade is identified before consent rather than after occupation.
- Road maintenance programmes are prioritised partly by exposure, so resurfacing happens first where it benefits most people.
- New infrastructure is assessed against the existing map, so cumulative exposure is visible rather than each scheme being judged in isolation.
- The map is rebuilt on a fixed cycle, so it stays a live dataset rather than a document with a date on it.
None of that requires new legislation. It requires the map to be a maintained dataset that other processes can query, which is a decision about how the work is commissioned, not about the acoustics.
Strategic noise mapping · Urban planning applications · Talk to us about a city-scale project