In Belgium, ~ 10% of the people is estimated to suffer from allergies due to pollen emitted by the birch family trees. Timely information on forthcoming pollen exposure episodes using a forecasting system can allow patients to take preventive measures. To date, the only available information on pollen concentrations in Belgium comes from five stations that monitor daily airborne birch pollen concentrations, but real-time and detailed spatial information is lacking. Pollen transport models can both quantify and forecast the spatial and temporal distribution of airborne birch pollen concentrations if accurate and updated maps of birch pollen emission sources are available and if the large inter-seasonal variability of birch pollen is considered. Here we show that the SILAM model driven by ECMWF ERA5 meteorological data is able to determine airborne birch pollen levels using updated maps of areal fractions of birch trees, as compared to the pollen observations of the monitoring stations in Belgium. Forest inventory data of the Flemish and Walloon regions were used to update the default MACCIII birch map. Spaceborne MODIS vegetation activity combined with an updated birch fraction map and updated start and end dates of the birch pollen season were integrated into SILAM. The correlation (R2) between SILAM modelled and observed time series of daily birch pollen levels of 50 birch pollen seasons increased up to ~ 50%. The slopes of the linear correlation increased on average with ~ 60%. Finally, SILAM is able to capture the threshold of 80 pollen grains m−3 exposure from the observations.