Urban airborne releases are difficult to model as the city itself is not a flat land surface. Urban structure, including streets, intersections, traffic and local meteorology, will play a major part in determining how material released to the atmosphere will disperse.
So the question in an emergency is not only what has been released but where the material will be, how fast will it disperse, where high concentrations could occur and which locations could be threatened for a short period.
This is why the subsequent DAPPLE-HO project focused on short-range dispersion in the urban canopy and its relevance for emergency planning. It built on previous DAPPLE work but placed more weight on short-term releases, repeat experiments, urban canopy flow and model evaluation.
Why Are Urban Airborne Releases Hard to Predict?
Airborne material does not simply disperse equally in all directions from a release point. It is modified by the urban built structure from the instant it joins a flow.
For example, it can be directed along a street or drawn away down a street, lifted over the roof of a building or trapped within a street canyon, or it can disperse more erratically in the area around an intersection.
All this will depend on where the release point is, the wind direction, the pattern of buildings, the local flow and turbulence in and around the release point, and the nature of the wider street pattern.
This is not too dissimilar to another DAPPLE topic: why it is hard to predict street-level air pollution. In this case, however, the timescale is tighter, and the impact of poor decision-making is probably greater and more immediate.
Emergency planning has to be informed by evidence about the likely short-range behaviour of materials. This is not only based on assumptions about the likely scale of the area affected or the direction of a general flow at regional scale.
Short-range dispersion is important because the decision-making needs are defined within a certain proximity to the source and over a short time span. For both distance and time scales, dispersion can vary greatly, in some cases with minor changes in flow conditions.
For example, a slight variation in wind direction may shift the main plume of dispersed material onto a neighbouring street. A corner of a building may give rise to complex flows. An intersection may promote complicated patterns of mixing. Roof-level data, for example, is a useful guide to local flow conditions, but it may not adequately reflect those at street level.
In this respect, the DAPPLE-HO project treats short-range dispersion in the urban canopy as a defined topic. It sought to characterise wind conditions and dispersion conditions at street level with respect to the prevailing meteorology, and then assess the performance of urban dispersion models for those conditions.
This is relevant to emergency response because it helps identify whether a model or emergency plan that adequately covers the larger scale may not adequately describe local features within an urban environment.
As such, it could be argued that the most important scale for any exposure assessment within a city is the scale of individual streets, entrances, road crossings, transport stations or enclosed public spaces such as malls, airports and underground tunnels.
The Urban Canopy Layer
The urban canopy is the area of the atmosphere occupied by buildings, streets and roughness elements in an urban environment. Within this layer, flows are complicated by walls, roofs, corners, gaps and street canyons.
A canopy may delay dispersion and redirect the flow or cause material to linger in an area longer than expected. Sharp variability may be seen from one street side to another. Material may enter a side street when not expected to, based on prevailing wind direction. A courtyard or other enclosed space may be affected differently than a street.
This is not only a minor detail. These variations directly affect exposure estimates, interpretation of sensor readings and understanding of advice given in an emergency.
The Source Is Only the Tip of the Problem
In an emergency release, much of the attention is focused on the source: what was released, where it was released from, when it started and how long it was released. These are indeed the questions that need the most immediate answers, but they are only part of the dispersion problem.
If the release is at a single point in time, along a route, or over a longer time period, that changes the problem. A very short release may behave differently from a longer one. The same release in the middle of a street canyon may behave differently from one near a junction.
A major part of the DAPPLE research programme looks at these questions through fieldwork, tracer experiments and modelling work. The article on what tracer experiments reveal about pollution movement in cities outlines the benefit of knowing the release and using it to understand the paths that material may take in an urban area.
In an emergency, this is useful because it indicates that a release may follow specific paths based on how it interacts with the current or future urban design, providing a quantifiable method to evaluate an otherwise invisible process.
Field Measurements and Repeatability
Field measurements provide information on real conditions, including actual geometry, traffic, real winds and real background variability.
Field measurements cannot easily be repeated with the exact same conditions due to changing wind conditions, traffic variations, changing background concentrations or even just different wind directions. This adds to the reality of the field measurements, but it makes interpretation of the data difficult.
The DAPPLE “How” page provides more information about how fieldwork, wind tunnel and computational simulations work together in understanding the flow of air, pedestrians and vehicles, and the dispersion of pollutants at street and neighbourhood scales.
In an emergency response setting, these different measurements are important. Field measurements provide insight into the complexity that may be encountered by emergency responders, while controlled experiments are useful for understanding which physical processes result in which observed phenomena.
Wind Tunnel Modelling
Wind tunnel results provide a method of studying the urban area through modelling with repeatable conditions. In a short time, multiple experiments may be tested to see the different results of a given change.
This is particularly useful in the study of emergency-related dispersion, as short-term releases are very variable in their outcome. In other words, one field release may indicate one possible outcome. With wind tunnel modelling, multiple experiments have the ability to provide information on different possible outcomes under controlled conditions.
A wind tunnel cannot recreate all of the conditions found in the urban street. It strips away the complexity of background pollution, traffic, heat and pedestrian flows, allowing researchers to probe the effects of building design, location of pollution sources and the orientation of the wind on the dispersion of materials.
This is why wind tunnel evidence remains a valuable tool for validating models and for developing an understanding of why concentrations at certain locations might be expected to be higher than elsewhere.
Models Must Be Tested, Not Assumed to Be Good
Urban dispersion models are important for emergency planning because there are rarely enough measurements from incidents to give any certainty on what is happening or what might happen. Models allow us to say where material might go, how it might disperse, and what concentrations can be expected in different areas.
But models should not necessarily be believed because they are complicated. Models need to be checked against evidence from field observations, tracer releases and controlled wind tunnel experiments.
Urban dispersion model evaluation was a key part of DAPPLE-HO, reflecting a very practical issue: emergency response models need to be fast and, therefore, relatively simple, but they also need to be believable.
Some models might not be quick enough for an emergency, or they might be too simplistic to account for local conditions, or they might be highly complex but still produce an uncertain result because their inputs were uncertain.
We need, therefore, not ask whether modelling should be part of an urban emergency response, but rather how modelling results can be tested, understood and applied with a clear awareness of uncertainties and limitations.
Monitoring During an Incident
Monitoring can also provide evidence of what is happening, but urban monitoring is challenging: any sensor or monitoring point can only measure the conditions at that location. Nearby locations in an urban street network could have very different concentrations.
The result is a challenge for emergency responders to interpret: a low reading at one location does not necessarily mean that other nearby locations are unaffected, and a high reading could be representative of a localised feature, a plume path or a temporary concentration maximum.
The DYCE project dealt with a related issue: how to deploy and redeploy limited sensor assets in the event of a release of chemicals in an open-air urban or industrial environment.
Its emphasis on dynamic deployment of sensor networks is pertinent to monitoring in an urban emergency situation, in that the value of a sensor network depends not only on the number of sensors, but where the sensors are located so they can measure useful information.
A good monitoring plan must recognise the heterogeneity of the city, and should support decision-making rather than just collecting data.
From Data to Decisions
Emergency response planning must be done under uncertainty. Information can be incomplete, changing rapidly or difficult to interpret. That does not make evidence less important. It makes designing the right evidence more important.
The best evidence base includes the following:
- field observations of the urban environment;
- tracer experiments that measure how material moves;
- wind tunnel experiments that give controlled repeatable measurements;
- numerical models that calculate conditions where it is not possible to measure directly;
- smart monitoring strategies capable of reacting to local conditions.
Each of these has strengths and weaknesses, but together they form a more reliable basis for judgement than any one on its own.
This is one of the important lessons of the DAPPLE approach to urban dispersion: dispersion research is not just a matter of drawing maps of concentration in the city. Ultimately, this is about understanding air movement in the built environment where people live, work and travel, and how this information can inform safer and more informed decision-making.
Planning for the Emergency Response
A lot of valuable emergency response work can take place before an emergency occurs, because once a release has already taken place, it may not be possible to gather data and understand local context quickly enough.
Pre-planning should allow for identification of those parts of the local environment, such as local street network, building morphology and meteorological conditions, that are likely to give rise to the greatest difficulties with dispersion.
It should also allow for evaluation of the applicability of different models for different types of decision-making, identification of monitoring needs and an understanding of where uncertainties might be greatest.
Of course, no emergency scenario can be fully predicted in advance, especially in an urban environment where conditions change quickly. But emergency planners can improve the quality of questions asked once an emergency occurs if they take time to prepare.
Implications for Urban Policy
There is no point in having separate emergency response and urban air quality policies; they must be linked because both rely on a good understanding of short-range dispersion, human exposure pathways and whether monitoring data is representative of what might happen in the rest of a city or neighbourhood.
In terms of urban air quality policy, this has several implications:
- Policy should not rely simply on broad assumptions about how air disperses in cities. Short-term releases are strongly influenced by local urban morphology, particularly if short-range dispersion occurs.
- Model validation and verification should be a part of preparedness. Any emergency response model should be validated against relevant empirical evidence before it is used during an emergency.
- Monitoring networks for urban air quality should be designed and validated with complex environments in mind. Sensor network placement, redeployment and data interpretation are key aspects in determining whether monitoring data can be used effectively to inform policy and decision-making.
- Emergency planning for an airborne release should consider the conditions that people are likely to experience at the scale of the street. This means not only the wider city or neighbourhood, but also pavements, street junctions, street entrances and exits, bus stops, transport tunnels and street canyons.
- Research evidence must continue to be translated into emergency response policy and operational frameworks. At the same time, policy should remain cognizant of uncertainties in urban dispersion modelling and prediction. The goal is not to achieve perfect prediction, but to achieve better decision-making.
The Continuing Requirement for Urban Dispersion Research
Emergency response planning for airborne releases is not solved; it remains an ongoing challenge because the urban environment is so much more complex than simple open environments. The interaction between buildings, vehicles, weather systems and people creates an environmental complexity that does not lend itself to single, universal rules.
The DAPPLE-HO project helped define this challenge by exploring short-range urban air pollutant dispersion, urban canopy layer flow, repeat experiments to assess repeatability and model evaluation.
DYCE helped explore the possibility of using networked air quality sensor data to understand how chemical releases disperse in complex outdoor environments. Combined with DAPPLE field campaigns, this series of work illustrates the need for emergency response planning to be founded on multiple types of empirical evidence: field data, tracer release data, wind tunnel experiments, numerical modelling and smart sensor network deployment.
The key implication for urban air quality policy is that airborne releases in cities must be planned for and considered at the scale at which they occur and are experienced: the scale not only of the whole city or city-region, but of individual streets, road junctions, building facades, air gaps, street canyons and urban spaces in which people may encounter releases.
Improvements in emergency planning depend on improvements in our understanding of the physical principles of urban air pollutant dispersion and human exposure.
That understanding must be technically sound, but it should also be tied to how urban planners, emergency responders, policy-makers and the public interact with the built environment.
