Where the wind blows

Air quality monitoring in densely populated urban areas faces a range of often compounding challenges. A growing number of cities can boast of air quality sensor networks capable of providing a minimum standard of monitoring. This minimum standard amounts to a ‘just enough’ approach, just enough data, just enough of an idea of air quality to suggest that air quality is above or below set guidelines.

But just enough isn’t enough at all. Citizens can’t take action to prevent their exposure to toxic levels of air quality using annual maps that only provide a rough outline of the problem. Precise route mapping of cleaner air travel routes, re-routing of vehicles and hyperlocal urban design interventions require granular data, street by street data reflecting the range of variables in urban microclimates is not yet available.

Using London as the example, this is the problem AirPublic is focusing on. King’s College London Environmental Research Group maintains over 120 monitoring stations across the London area. These provide a good foundation to base policy upon at a meta-level and from which to issue air quality alerts during pollution episodes, and to understand that we have serious air quality issues in London. But when it comes to implementing applications such as those above in order to reduce the effects of pollution this network alone is not sufficient.

Between monitoring stations, to make up the difference, air quality is modelled according to environmental conditions and a complex range of variables such as temperature and humidity. However, this modelling does not account for hyperlocal urban microclimates, such as the skyscraper effect or street canyon effect. Let’s take a closer look at the last of those.

The street canyon forms a key part of the urban microclimate. Cities, like forests and jungles, have their own urban canopy areas. This layer, from the surface to the roof-level, has a distinct climate which is dominated by micro-scale influences. The airflow and thermal characteristics of these street microclimates are important, they help us explain potential problems around wind-buffeting, heat stress, driving rain and concentrations of air pollutants.

The cooling effect of shadows cast by tall buildings, the shifting of sunlight from east to west throughout the day, heat emissions from surrounding buildings and waste systems, these everyday phenomena shape local microclimates, changing the strength, direction and pattern of convection currents that determine the airflow regime and windspeed. As a result of these compounding effects, localised climates are notoriously difficult to model.

London’s KCL ERG sensor network, a sparse array of highly accurate monitoring stations, currently relies on data modelling to infer air quality in areas not covered by sensors. Without a greater coverage of sensors, without building in an understanding of the city’s street canyons, skyscrapers etc, a sparse static network cannot create a richer and more accurate picture of air quality in a city. As a result of missing key criteria needed to make accurate assumptions about environmental variables, maps are produced and relied upon that do not necessarily reflect the reality of air quality in a city.

We can view an example in Battersea by comparing data modelled by the KCL network, here (data recorded over an hourly period), here (annual NO2 map) and data collected by hand using passive diffusion tubes, here (over the course of a month), mapped by Mapping for Change. Here you can see the spread of NO2 monitoring sites in the Battersea area relating to the maps above. It’s worth noting that whilst the seasons and dates differ in the examples, these demonstrate the lack of preciseness or granularity outside of the immediate vicinity of the KCL sensors.

It’s worth reflecting that the King’s College London Environmental Research Group’s sensor network is of immense value to the city and provides us with the foundations for us to build from and improve upon. Whilst it doesn’t provide such a total and uncompromising view of air quality in the city, it continues to provide decision makers and the community with solid actionable knowledge on pollution throughout London, when and where the hotspots are and when an air pollution alert must be issued by the Mayor’s office. KCL’s dedicated team have been invaluable to the development of the AirPublic sensor platform, providing testing facilities and much needed advice and expertise to our team.

Fundamentally, highly accurate monitoring stations, such as those run by King’s College London, act as a core backbone, a baseline for air quality monitoring. New technologies can now allow an ecosystem of smaller, cheaper monitors, such as AirPublic’s own, to fill in the gaps and complement the existing network, providing a more dynamic dataset to allow for smart city applications.

Whereas King’s network currently only provides hourly readings in order to make adjustments to the annual pollution map of the city, the AirPublic network of 100–1000s of mobile sensors will provide more granular readings, minutely instead of hourly and weekly averages instead of annual, providing the necessary detailed information for micro-urban interventions, traffic routing systems and other potential applications of these datasets.

Further reading:

Introduction to Street Canyon & Wind Flow Regimes

Whiston-Spirn, A., 1986, Air Quality at the Street Level: Strategies for Urban Design