Interim Report (D1) – v0.30 Service Contract on Improved Tools for Assessing NO2 Exposure (EU ambient air quality policy) Task 1 Interim Report: Review of Existing Methodologies

Bino Maiheu, Martin Lloyd Williams, Heather Anne Walton, Stijn Janssen, Lisa Blyth, Nele Velderman, Wouter Lefebvre, Marlies Vanhulzel, Sean David Beevers

Research output: Book/ReportCommissioned reportpeer-review


Introduction and Background
Health impacts attributable to elevated concentrations of NO2 in the ambient air are of increasing societal concern: the European Environment Agency (EEA) estimates this to be in the order of more than 70.000 premature deaths across the EU-28 in the year 2012 alone (EEA, 2015). However, the current European Commission's Clean Air Policy Package does not include the health impacts of NO2 exposure mainly because of the current lack of a robust methodology or tool for the assessment of NO2 exposure, and the use of appropriate dose-response relationships.
DG ENV has commissioned VITO (BE) and King's College London (UK) to propose methods and tools that are coherent with exposure metrics used when deriving the appropriate concentration-response relationships and compatible with currently used integrated assessment modelling (IAM) tools in the EU. The ultimate objective of the contract is to develop a (or refine an existing) module for improved NO2 exposure calculation for health impact assessment and cost benefit analysis.
This report is a key deliverable in the project and provides an overview and analysis of the main methods and tools currently used for assessing human exposure to NO2. One of its main purposes was to serve as input for an expert consultation workshop which took place at the WHO offices in Bonn on the 17th May 2016. This version has been updated to include the key recommendations and conclusions from that workshop.
NO2 Health Impact Assessment
A recent review of health effects by WHO in support of the European Commission 2013 Clean Air Policy Package, REVIHAAP (World Health Organization, 2013a) has concluded that evidence for the effects of long-term exposures to NO2 independent of those of PM has now strengthened. A subsequent exercise, HRAPIE, (World Health Organization, 2013b) recommended concentration-response functions (CRFs) relating mortality outcomes to long-term (annual mean) exposure to NO2 to be used in sensitivity studies. These recommendations have been considered along with the literature that HRAPIE reviewed, by COMEAP in the UK who provided some interim recommendations in July and December1 2015 (Committee on the Medical Effects of Air Pollutants (COMEAP), 2015). The COMEAP committee is due to produce a full report in the early Autumn of 2016. The COMEAP report will include newer studies that would likely be considered in any update of the HRAPIE recommendations but different expert groups may make different choices as to which studies to include in their meta-analyses.
The HRAPIE and COMEAP interim statement recommendations differ significantly. The HRAPIE relative risk coefficient was 1.055 per 10 μg/m3 annual mean NO2, for all-cause mortality, to be reduced by 33% to allow for the potential overlap with PM2.5. The relative risk should be applied only to concentrations over 20 μg/m3, based on confidence intervals widening below this level in a study in Rome and a similar result in one age group in a study in Norway. The interim recommendations from COMEAP in July 2015 were a relative risk of 1.025 per 10 μg/m3 annual mean NO2, for all-cause mortality, with a 33% reduction to allow for potential overlap with PM2.5. Unlike HRAPIE, the COMEAP recommendation for a cut-off was to suggest that calculations be carried out with a zero cut-off and one with the lowest concentration used in the epidemiological studies, the former based on absence of evidence for a threshold and the latter following the principle of using just the range of the data. More studies going to lower concentrations have been published since the HRAPIE

report. Both sets of CRFs are applied to people aged above 30 in the mortality calculations. CHAPTER 2 of this report provides an in-depth discussion on this topic.
NO2 Exposure Assessment
Different methods for NO2 concentration assessment have been used to date. The methodologies can in general be classified as follows:
1. regional scale Eulerian chemical transport models (CTM)
2. urban scale dispersion modelling (Gaussian/Lagrangian models), with/without street canyon parametrization
3. obstacle resolving dispersion modelling using computational fluid dynamics (CFD)
4. usage of operational in-situ measurements and dedicated campaigns using passive samplers
5. land-use regression (LUR) models, especially in epidemiological studies.
A recent review, performed within the framework of FAIRMODE (Denby, 2011) discusses those different NO2 modelling approaches in more detail. Often a hybrid multi-scale approach is chosen where for example Gaussian dispersion models (with/without street canyon parametrisations) describing the urban scale concentration pattern are combined with Eulerian CTMs which provide low resolution background concentration estimates (e.g. ROADMOD, UKIAM – BRUTAL, US-EPA APEX).
Further it is observed that in an IAM approach, the scale of the application or the run-time requirements may impose the need for simplified modelling approaches. Here dispersion-kernel methods are commonly used in which annual averaged dispersion patterns (kernels) are pre-computed for a set of generic conditions (road type, orientation, …).
In addition, the concentration responses to changes in emissions may be parametrised using linear or non-linear source-receptor relations. Such relationships are derived from sensitivity runs using the full model, from which a concentration change Δ퐶푖 in location i is parameterised as a function of the change in emissions Δ퐸푗 at nearby locations j.
Some effort has furthermore been spent to estimate NO2 road side increments on top of urban background concentrations. Those increments are parameterised and calibrated against historical measurement data (e.g. the GAINS methodology) or detailed model simulation (e.g. the TRANSPHORM approach) and are specifically designed to be used in IAM frameworks.
A full description of the methodologies and their evaluation is given in CHAPTER 3. The matrix in Table 7 summarizes the key advantages/disadvantages of the methods discussed in the context of integrated NO2 health impact assessment.
The Key Open Questions and the Way Forward
Based on the evaluation of the existing methods for NO2 concentration and related health impact assessment, we identified a number of key questions which point toward the essential elements in an updated NO2 exposure assessment scheme. In CHAPTER 5 we present these key questions and the recommendations from the experts on how to address them. In consideration of these recommendations we present a way forward which as a first step requires the implementation of some sensitivity studies to assess the importance of different aspects of the final exposure metrics
Original languageEnglish
PublisherEuropean Commission
Commissioning bodyEC European Commission
Number of pages108
Publication statusPublished - 2016


Dive into the research topics of 'Interim Report (D1) – v0.30 Service Contract on Improved Tools for Assessing NO2 Exposure (EU ambient air quality policy) Task 1 Interim Report: Review of Existing Methodologies'. Together they form a unique fingerprint.

Cite this