Street-based studies provide the essential evidence required for urban air pollution research. Yet, in a real-world environment, performing controlled experiments is almost impossible. Traffic patterns fluctuate, wind directions shift, background pollution levels change, and that exact set of circumstances never repeats itself.
It is one reason why wind tunnel modelling retains its relevance. While wind tunnels do not recreate entire cities, they do enable researchers to isolate the underlying physical processes governing the movement of air and pollutants within building canyons, street junctions and along roads.
In the DAPPLE programme, wind tunnel work did not act as a replacement for street monitoring campaigns, but as a component of a larger evidence ecosystem that spanned field and tracer release experiments, exposure monitoring and computational modelling. The value of that approach is the union of realism with control.
Control and Street-Scale Dispersion
At street level, dispersion is heavily dependent upon the fine detail of the urban landscape. The precise configuration of the urban fabric — the presence of a building angle, the size of a gap between structures, the difference in roof height, or even a shift in source position — will change how a plume disperses.
In a real street environment, it is difficult to distinguish between those factors. If we observe a change in concentrations measured over successive hours, it is not clear whether that is due to wind, the quantity of emissions, background pollution levels, a change in source strength, or a combination of all of these. It is part of the reason why street-level air pollution is difficult to predict.
In contrast, a wind tunnel experiment allows the researcher to fix some of these factors. Wind direction is fixed. The flow speed can be repeated. The source location can be changed at will. Concentration measurements obtained from different experiments can then be compared directly.
This control does not remove complexity from the problem, but it does make complex processes more manageable.
What We Can Learn from Wind Tunnels
Wind tunnel modelling is particularly well suited for investigating the influence of urban form on air dispersion. A physical model can reveal the pattern of how material moves around a group of buildings, whether a street canyon retains or expels material and the extent to which a junction can promote mixing, or the reaction of a plume to a change in wind direction.
We can also gain information as to whether high local concentrations are a result of local recirculation, poor ventilation, source position or the surrounding street layout.
The DAPPLE “How” page describes the use of wind tunnel and computational simulations alongside field campaigns to characterise the flow of air, traffic and people at the street and neighbourhood scale, and the interplay with dispersion.
The point is that a wind tunnel experiment represents a simplified, controlled version of a real physical problem. This allows us to ask: if everything else stays the same, what happens when we change this? That is a very difficult question to answer in a full-scale campaign.
Repeatability and Short-Range Dispersion
Another reason why wind tunnel modelling remains valuable is its repeatable nature.
For short-range dispersion, a single release is not necessarily an adequate basis for assessing plume characteristics. We might expect the dispersion and local concentrations to vary significantly from one release to the next, even though the conditions looked similar. To get a handle on the statistics of such a process, a series of experiments is needed.
The DAPPLE-HO project addressed short-range urban dispersion, repeat experiments and model evaluation, with particular relevance for emergency response planning.
Within the urban fabric, it is impossible to conduct many repeat measurements before wind speed and direction or weather conditions change. In the controlled flow of a wind tunnel, one can conduct repeat experiments under nominally identical flow conditions. This then allows one to investigate the variability systematically.
This is important in the context of urban policy. The evidence to support decisions about exposure, emergency planning or model performance should not depend on single measurements, where the process itself has an inherent variability.
Wind Tunnels and Tracer Experiments
Tracer experiments are particularly well suited to wind tunnel work.
One has control of a tracer release source which is released from a well-defined source. Its concentration can be measured in the model at many points in the city. It can be released from street level, in a street canyon, at a junction, above roof level, below roof level or whatever other source locations are of interest. One may have different wind directions, but the source conditions remain constant and can be compared to one another.
Tracer releases in the field are one way of investigating how the material is transported through the urban environment. Tracer experiments in the wind tunnel provide a way of repeating the same release conditions and exploring the results in a more systematic way, providing more insight into the possible pathways the material could take.
The article on what tracer experiments reveal about pollution movement in cities explores this role in more detail. In terms of wind tunnel work, the key point is that tracer work allows a process which would be otherwise invisible to be measured under repeatable flow conditions.
This is useful in relation to investigating the dispersion from sources of ordinary pollution, but also when considering short-term releases and how a numerical model is performing.
The Limitations of Wind Tunnels
Of course, a wind tunnel is not the same as a full-scale city and these limitations should be acknowledged.
Traffic is represented in a more simple way. Human movement may be represented to a lesser extent. There can be issues with the reproduction of thermal effects. Pollution from other sources may not change as it would in the field, for example. The scale model should represent the most pertinent processes relevant to the problem without having to recreate everything that is part of the urban fabric.
Therefore, wind tunnel data should not be treated as the only measure of urban air pollution exposure in the same way field measurements could be taken to describe the urban environment. It is a controlled physical model which gives results that relate to the real urban environment in a limited way.
The value of the wind tunnel experiments is therefore to assist in understanding these processes. They can help answer the question of how the buildings are influencing the flow, how the retention of material is affected within the street canyon, or how the source position changes the plume and concentrations observed. But the results have to be interpreted within the wider context of the available field evidence.
This is why the relationship between field data and wind tunnel data is so important within urban dispersion research. Field data gives a more representative result. The wind tunnel has the control to allow the understanding of processes, but neither should stand alone.
The Use of Wind Tunnels in Model Evaluation
Wind tunnel experiments have an important role to play in the development and evaluation of numerical models.
A range of numerical models can be utilised to investigate dispersion in a variety of situations where measurement of the plume may not be possible. They are useful as a way to investigate scenarios, estimate exposure, support decision-making and aid emergency response.
But their outputs are also based on assumptions about emissions, meteorology, buildings and boundary conditions. Wind tunnel data provides a controlled evidence base by which to compare those models.
If a numerical model does not do a reasonable job of reproducing the main flow and dispersion patterns found in a controlled physical experiment, then the application of the model to a real city should be treated with a degree of caution.
So the relationship between physical modelling and digital modelling is not a competitive one. The former underpins the latter with test data, and the latter extends the former by enabling the investigation of scenarios that may not all be possible to test in a wind tunnel.
Between the two, researchers can not only get an idea of what a model predicts, but of whether it makes physical sense.
Model Validation: Why We Validate and Its Policy Significance
Model validation is often framed as a technical exercise, but it has wider implications for policy.
Urban dispersion models may be used to influence transport planning, building design, public health or emergency response, so if the outputs of such a model have not been tested against good evidence, it may be over-representing its own accuracy.
Wind tunnel experiments play their part in model validation by giving models the opportunity to work under known conditions. They can be asked to reproduce results for a specified wind direction, source position and urban geometry, then the outcomes can be compared to what a physical experiment tells us.
Obviously, this does not prove that a model will work well in the real city, but it can at least show whether there are significant flaws in how the model deals with some key processes, such as flow separation around buildings, recirculation in a street canyon, plume spread and concentration gradients.
Understanding the limitations of an urban dispersion model is useful for policy users because it helps to define what it can reasonably be used for. For example, a model might be good at a coarse-level screening exercise, but not at assessing short-term exposure, or vice versa. A model might work better in an open area than a dense urban canyon. Validation helps to distinguish between these cases.
Wind Tunnels and Emergency Planning
Urban wind tunnels can be useful for understanding emergency response.
The difficulty with planning for an airborne release during an incident is that you may have very limited time in which to get answers, and a lot of those answers will depend upon what is not known, or what is not known very precisely.
For example, an airborne release could move to a variety of different locations depending upon its physical characteristics and the weather conditions on the day, so you want to know where it may be in the future so that you may prioritise monitoring and evacuation.
Short-term dispersion is also likely to be much more difficult to predict in an urban area because of the variability of airflow in the street canyon, around junctions, and near the corners and roofs of buildings.
The post on emergency response planning for urban airborne releases sets out the difficulty that exists when dealing with such an event in a practical context, and shows that wind tunnel studies can assist by enabling the preparation and testing of response plans before an incident occurs.
This could involve testing a number of release scenarios, exploring how different source positions, wind directions and building geometries may result in different exposures, and generating data to test the effectiveness of fast-running operational models that may be used in real time.
Wind tunnel models will not be able to predict everything that an emergency response team will need to know or the scenarios they may face. But they can show how emergency models will perform, and give ideas of the types of response and monitoring strategy that will be necessary.
Wind Tunnels, Sensors and Monitoring Strategy
The results of wind tunnel experiments can also influence how monitoring should be carried out. For a sensor to be effective, it needs to be placed in the appropriate location relative to how the plume or pollution field behaves.
When a plume is likely to flow along a street, to be driven across a junction, or to hang down by the side of a building, then this is a possibility that monitoring locations should account for.
The article on how sensor networks can improve urban pollution monitoring says that more sensors do not mean better evidence. Placement, representativeness and interpretation are crucial.
Wind tunnel studies are useful because they show where concentration differences are likely to be found under different flow situations. This can help inform thinking about where to place sensors, how to interpret their readings and what sites may not be represented by a single fixed monitor.
In this sense, physical modelling is not just a source of evidence for model validation, but also evidence that can lead to better monitoring design.
Policy Relevance of Controlled Experiments
Controlled experiments have policy relevance because they help to separate process from coincidence.
In the field, a high concentration might be observed at a specific location, but there may be no obvious reason why this is so. It might be because of traffic, wind direction, a local building effect, poor ventilation or some transient event. Wind tunnel experiments can help to test what physical mechanisms are likely to be important.
This is useful for different policy areas:
- Urban planning, where building forms and street layout may affect ventilation and exposure.
- Transport policy, where traffic emissions interact with local dispersion conditions.
- Air quality monitoring, where fixed instruments may not be equally representative of surrounding streets.
- Emergency preparedness, where short-range transport and localised plume behaviour may be important.
- Model evaluation, where model predictions should be tested against both controlled and field evidence.
Wind tunnel modelling does not make judgement calls. It helps to improve the evidence that these judgement calls are based on.
Why Controlled Experiments Matter
The future of urban dispersion research will involve more detailed computational models, larger data sets, better sensors and better data on buildings and traffic. All of these developments are important, but they do not remove the need for controlled experiments.
As models become more detailed, they need more evidence to evaluate them against. As sensor networks become more extensive, they need physical interpretation. As policy becomes more localised, street-scale dispersion and the processes that control it become more important.
This is where wind tunnel modelling remains relevant. It provides repeatable evidence about physical processes that are difficult to isolate in the real city. It can help to test models, interpret field data, study tracer behaviour and support emergency preparedness.
The relevance of DAPPLE’s use of wind tunnel work lies in the fact that it saw controlled experiments as part of a combined evidence base. Field measurements, wind tunnel studies, numerical modelling and tracer experiments all provided different parts of an answer to the same problem.
Urban dispersion research needs to include that combination if it is to continue to deliver progress. Better prediction does not simply result from more data or from more complex models, but from clearer understanding of how air and pollutants move in the built environment.
Wind tunnel modelling is valuable because it can provide that understanding where evidence can be repeated, tested and compared.
