A novel application of PM10 back trajectory industrial plume modelling to characterise the vulnerabilities of the exposed population: Implications for public health emergency planning

Introduction
Globally, uncontrolled open fires are of increasing concern because of significant increases in atmospheric pollution impacting human health and societal resilience. Since 2009, the UK Air Quality Cell (AQC) function has provided health protection advice during the incident response, but its remit has stopped there. The AQC records show that waste management sites are the most common cause of significant fire incidents. Near these sites, communities were impacted by the plume from the fires.

AQC’s monitoring ability has limitations: (1) monitoring teams are located outside of the main plume to protect their occupational health, and (2) the monitoring instrument (OSIRIS) has a ceiling of 6,528 µg m-3 which is reached at several sites. Time-series back-trajectory-modelled plumes can overcome these shortcomings. Spatial analysis correlating the plume with small-scale (output area) population characteristics indicates the numbers potentially exposed, as well as the likely US EPA AQI. It also provides a population vulnerability profile that considers pre-existing social, economic, and health indicators linked to land use, including the age of properties. In the context of the standard advice to shelter, property age can indicate the ability of a property to protect residents.

The default advice in response is for an exposed population to shelter in place (“go in – stay in – tune in”) but the question remains: how effective this advice is against significant air pollution levels. In layman’s terms, how ‘leaky’ a property is but this requires an understanding of the properties affected an AQC event.


Method
A tyre fire in Mexborough in 2010 was used as a typical example of a significant major air pollution or ‘Air Quality Cell’ event. The ADMS-STAR modelling software generated 22 1-hour plumes that were spatially averaged to present a single plume that represents a single 24-hour exposure for usual residents.

A ‘reasonable worst case scenario’ (RWCS), as opposed to the absolute worst case, was developed using a novel spatial weighted AQI. Each Output Area (OA) (c. 300 usual residents) often had >1 AQI reported across the footprint. Within each OA, the area (m2) for each AQI reported was identified and compared with the total area of the OA. Weighted by the proportion of a single AQI was identified for each OA. RWCS is an approach used in resilience planning that constructs an appropriate risk based scenario that is plausible and challenging, but not the absolute worst case.

ArcGIS was used to determine the intersection of the averaged plume with population data to calculate: (a) total population exposed, (b) predict health impact using 24-hour US EPA Air Quality Index (AQI), (c) concentrations and AQI at vulnerable locations, (d) a population vulnerability profile, and (e) a profile for land-use with a focus on the age and nature of the residential accommodation exposed.

Results
It was seen that the total number of usual residents exposed to AQIs ‘Unhealthy for Sensitive Groups’ and worse was indicated as 30,023. Of which, 8,119 are to have been exposed to the ‘Hazardous’ and ‘Above the AQI’ AQIs. The majority, 7,856, were residents in Mexborough and Conisbrough. Sensitive receptors (schools, hospitals, care homes): 17 were exposed to ‘Unhealthy for Sensitive Groups’ and higher; 9 were ‘Hazardous’ AQI.

Relative to England’s median for vulnerability factors, we report increased representation of young people (0 to 4 (+3.90%), 5 to 7 (+2.0%), 16 – 17 (+6.35%), 18 to 19 (+17.35%)); older (65-74 (+0.45%), 85 to 89 (+3.90%)); and self-reported pre-existing health conditions (fair (+27.50%), bad (+37.40%), very bad health (+30.15%).

The work to determine the property type and age is to be completed. It is noted that the fire was located about 125 m from terraced properties in Mexborough. This analysis will identify (a) types of property exposed and (b) their age which can be used as a base for the quality of the accommodation.

Conclusion
The indicated number of residents exposed to unhealthy and hazardous AQIs, coupled with the vulnerability profile developed, provides an evidence base for future public health response and preparedness for major incident fires, including the level of clinical and social support that should be provided post-event. There are also implications for how waste materials are processed, how close to populations, etc. Additionally, this has implications for managing risks to communities under the Civil Contingencies Act 2004 including the need to consider micro-variations in vulnerabilities as part of emergency planning and preparedness. More broadly, in responding to the incident, the health services (NHS) can be better prepared to support exposed populations.