The effect of wood burning on particulate matter concentrations in flanders, Belgium

Abstract

Since 2010, new estimations of the particulate matter emissions in Flanders have been made by using a tier-II approach. By means of a survey, the quantity of the wood by households, buildings (services sector), industry and agricultural sector consumption in Flanders has been estimated. A survey is necessary as not all consumed wood is purchased, e.g. some of the wood is sourced locally or waste wood is used. These surveys also provided data on the installation stock in Flanders, such as the type (fireplaces, pellet stoves, …) and age of the heating installation. Besides the wood consumption and use patterns, emission factors of particulate matter per type of wood and per type and age of heating installation were also updated based on international, recent literature on emission factors. Both new estimations resulted in an actualization of the particulate matter emissions of the residential wood burning in Flanders. Overall, the estimations were a factor 13 higher than the old estimations. This factor 13 is due to a factor 4 increase in the average emission factor and a factor of 3.4 in the wood consumption. As a result, heating in residential sector is estimated to be the most important emission source for primary particulate matter (37 % of all primary emissions) in Flanders. To confirm these results a comparison was carried out with the estimates of the contribution of wood burning to ambient PM10 concentrations based on measurements of levoglucosan. It was shown that about 10 % of the particulate matter concentrations in winter were due to wood burning, while in summer this amounted only 2–3 %. In the next step, a dispersion model has been used in order to check the consistency of both results. It is shown that the increased emissions are consistent with the measured particulate matter concentrations due to wood burning. Indeed, the increase in emissions by a factor 13 is needed to understand the high levoglucosan concentrations in ambient air in Flanders. Finally, the large changes that have been made to the Belgian SNAP-2 emission inventory due to the results of this study, raise questions about the homogeneity (both spatial and temporal) of the European emission inventory for this sector. Adjustments to this sector could possibly decrease the modelled PM-gap quite substantially.