You are here
Comparison of Long-term Air Pollution Exposure assessment based on mobile monitoring, low-cost sensors, dispersion modelling and routine monitoring-based models (CLAIRE)
Hoek and colleagues will prepare maps of modeled annual average air pollution across the Netherlands, validate the maps using new measurements from over 100 sites, and evaluate the performance of several exposure models. The investigators will conduct cross-comparisons to evaluate how different exposure assessment methods compare in their ability to predict long-term pollutant concentrations, with a particular focus on spatial variability of pollutants.
Poster abstract for HEI Annual Conference 2023
Comparison of Long-Term Air Pollution Exposure Assessment Based on Mobile Monitoring, Low-Cost Sensors, Dispersion Modelling and Routine Monitoring-Based Models
Gerard Hoek1, Femke Bouma1, Kees Meliefste1, Ulrike Gehring1, Roel Vermeulen1, Kees de Hoogh2, Sjoerd van Ratingen3, Wouter Hendrickx3 Nicole Janssen3, Joost Wesseling3
1Institute for Risk Assessment Sciences, Utrecht University, the Netherlands.
2Swiss Tropical and Public Health Institute, Switzerland
3National Institute for Public Health and the Environment, the Netherlands
Background. Assessment of long-term exposure to traffic-related outdoor air pollution remains a major challenge for epidemiological studies. One challenge is the characterization of the spatial variation of the ambient concentrations of key traffic-related air pollutants including ultrafine particles (UFP), Black carbon (BC) and nitrogen dioxide (NO2). Epidemiological studies have used different approaches, including land use regression (LUR) models based upon mobile monitoring (UFP, BC), models based upon low-cost sensors (PM2.5, NO2), dispersion models and increasingly sophisticated hybrid models. Little information is available about the relative performance of these different methods, which may affect conclusions from epidemiological studies applying different approaches.
Methods. The project will compare seven exposure assessment methods, which differ in modeling approach (empirical LUR, deterministic dispersion and hybrid models) and monitoring data used (low-cost sensors, mobile/routine monitoring). For all empirical models we will test three model development algorithms: supervised linear regression, random forest and LASSO. We will also combine empirical and deterministic models using data fusion / assimilation.
The predictions of the models will be compared at 20,000 residential addresses across the Netherlands and tested on newly collected and existing external validation data. Epidemiological analyses in three cohort studies will be conducted to compare health risk estimates of the different exposure assessment methods. The studies include a national administrative cohort, a cohort with detailed lifestyle data study and a mature birth cohort in which we will assess mortality, cardiovascular disease incidence and lung function / asthma respectively.
Results. Data collection for the low-cost NO2, PM2.5 and PM10 sensor network of 98 sites started in September 2021 and is nearly finished. Especially for NO2 calibration was challenging. Preliminary analyses show large variation of average NO2 concentrations with expected differences in rush hour peaks between week and weekend days.
Data collection for the new external validation measurement campaign also started in September 2021 and is nearly finished. Preliminary analysis of UFP concentrations shows large spatial variation across the 100 sites.
Exposure models based on pre-existing data have been developed. Preliminary comparisons show that the model predictions are moderately (R 0.7-0.8) for comparisons across methods to highly correlated (R 0.8 – 0.9) for comparisons across algorithms.
Analysis of UFP-related mortality in the national administrative cohort has been completed for our national UFP models. The results show significant associations between annual average UFP exposure and natural cause and lung cancer mortality (hazard ratio 1.012 and 1.038, respectively per IQR range of exposure) which remained significant in two-pollutant models.
Conclusions. No strong conclusions can yet be drawn about the comparison of the exposure models. The interpret from the preliminary analyses is that most model predictions are sufficiently moderately correlated that differences in predictive performance and effect estimates in epidemiological analyses are possible.
