An IFDM modelling study has been conducted to determine the optimal siting of air quality monitoring stations around five oil refineries. The purpose of this immission monitoring network is specified in environmental legislation. The most appropriate output parameter of the IFDM model for this study is the 98th percentile of the moving 24 h averages, P-98,P- m24h. Modelling for the optimal siting of air quality monitoring stations turned out to be different from modelling for permit granting. Of interest is not the peak value of a relevant immission parameter, but the places where this parameter is most likely to have a higher peak value compared with neighbouring places. For this study, modelling has been done for eleven years of hourly meteorological data. The eleven yearly immission fields obtained by modelling are synthesized using the minimum, median and maximum values of the P-98,P-m24h values obtained for each receptor point. This synthesis of the modelling results was complemented with an analysis of the available immission data for the region, so that the resulting air quality monitoring network is expected to monitor the impact of all emissions, not only those emissions that are in the emission inventory used for modelling.