Advanced Source Apportionment for the Assessment of Exposure to Exhaust and Non-Exhaust PM in Health Effects Study
DAVID GREEN, Max Priestman, Anja Tremper, Ian (Gang) Chen, Klea Katsouyanni, Ian Mudway, James Scales, Hajar Hajmohammadi, Helen Wood, Christopher Griffiths, Imperial College London
Abstract Number: 58
Working Group: Source Apportionment
Abstract
Non-exhaust emissions currently make up a greater proportion of vehicle emissions by mass than exhaust emissions, with the increasing weight of vehicle fleets due to electrification and increasing uptake of larger vehicles it is likely that these will continue to be an increasing health challenge. Here we utilise the high time resolution aerosol measurement instrumentation available at the supersites in London (plus an additional mobile laboratory), coupled to advanced source apportionment techniques to provide source-specific exposure measurements for a randomised cross-over semi-experimental health impact study. It investigated the short-term respiratory health impacts in non-smoking adults with mild-moderate asthma during and after sequential standardised exercise exposures to three contrasting air quality environments, comprising:
(1) A busy road location characterized by stop-go traffic to enhance emissions from brake wear (Marylebone Road)
(2) High speed continuous traffic, to enhance Tyre and Road Wear Particle (TRWP) emissions (White City Roadside)
(3) An urban background location away from nearby traffic sources (Honor Oak Park)
Elemental composition was measured at 1 hourly time resolution using ED-XRF, to maximise size and chemical composition information, PM2.5 and PM10 were collected on alternate hours using a switching valve and intermediate hours are interpolated. An Aerosol Chemical Speciation Monitor (ACSM) measured the non-refractory composition of aerosol and black carbon will be measured using an Aethalometer.
Positive Matrix Factorisation (PMF) was applied to the PM10 and PM2.5 XACT data and ACSM data from all three stations using the Source Finder software (SoFi Pro, https://datalystica.com/) to provide hourly source factor time series and factor profiles for the duration of the exposures. For the ACSM data, this approach has been used in many recent studies to provide high-quality source information using priori information and a bootstrap resampling approach following standardized protocols and is now widely used in source apportionment. PMF of ACSM data yielded typical regional and urban sources of organic aerosol in PM1 and PM2.5. PMF was applied to XACT PM2.5 and PM10-PM2.5 data in single a multi-site analysis to provide consistent factor solutions. A range of regional and local sources were identified, this was achieved by first isolating the roadside increment to provide factor profiles which were then used to constrain PMF for the whole dataset. The analysis of these factors focuses on the non-exhaust and the contrasts between the different locations, vehicle fleet composition and speed.
This work was supported by the US Health Effects Institute funded under RFA 21-1 Quantifying Real-World Impacts of Non-Tailpipe Particulate Matter Emissions, the NERC Traffic Grant (NE/1007806/1) and the NERC OSCA Grant (NE/T001909/2).