You are here: Home / BUMBA / WP5


High resolution air quality modelling and validation


In order to provide accurate NO2 exposure estimates and the concomitant health risks, high resolution air quality modelling is needed. The RIO interpolation model currently used for the operational near real time assessment of air pollutant concentrations and related exposure assessment at IRCEL-CELINE has a spatial resolution of (only) 4km x 4km. An air quality modelling chain with sufficiently high resolution of up to 25m x 25m was already developed by VITO-RMA and IRCEL-CELINE, but two major problems remain concerning this model chain, making it unsuitable for operational use at this moment.

(1) A thorough validation of the high resolution air quality model results with measurements has been impossible until today, because of the lack of suitable data to validate the model with. Currently, the validation is generally carried out for a few points where ground based observations from the air quality measurement networks are available. However, a validation for a whole area is currently impossible. APEX now offers, for the first time, a possibility to provide us with high-resolution NO2 concentration maps that can be used for the model validation. Hence, a comparison between APEX derived NO2 maps and the results of air quality models, and additional reference ground-based in situ and remote sensing observations, would provide extremely valuable information both for the validation of model results, but also for the assessment of the APEX NO2 product itself.

(2) Secondly, the developed model chain is not suited to use as an operational near real time exposure assessment model because of the large amount of necessary CPU time. In order to be useful as a near real time exposure assessment tool, the execution of the calculations has to be accelerated to a sufficiently high speed. This issue is discussed further under WP6


Validation of the high resolution air quality model

The local-scale model requires two main input parameters: meteorological conditions and local-scale emissions. Both have an important influence on the ground concentrations, the air quality modelling and the conversion of the APEX vertical columns to ground concentrations. Firstly, the wind speed and direction determine the speed and direction of the dispersion of plumes.Apart from them, the boundary layer height is of crucial importance when modelling near-surface atmospheric processes. The boundary layer is the stable atmospheric layer immediately adjacent to the surface. 

In addition to the analysis of the meteorological input, we performed a quick scan of the emissions in the neighbourhood of Antwerp focussing mostly on an analysis of the height of the emissions and their relation to the location of the source. The domain for Antwerp is hereto divided in three parts, one containing the cit centre, one the port of Antwerp and the last one the “rural” region around the city centre. The situation in Brussels is completely different. There are almost no industrial point sources in the domain, and most emissions are related to traffic and thus occur in the lowest level for the entire domain. Finally, for Liege, the situation is a bit particular. Although there are a lot of point sources in the domain related to industrial plants, these are not well documented in the E-PRTR point source database, which is currently the best available dataset for the Walloon region. Hence, the emission data set is very incomplete and mostly contains sources close to the surface related to traffic.

The modelled surface concentrations were validated using the results of the airborne measurement campaigns and the additional ground-based in-situ measurements. The focus hereby lies on comparing (the spatial pattern of) the surface concentrations, as this quantity is related closely to the human exposure and thus provides the basis for the high resolution model chain for exposure estimates and risk assessment.