Background
In Belgium, ~10% of people are sensitive to birch pollen and ~15% to grass pollen, significantly affecting their quality of life. Mitigation measures can alleviate pollen-induced symptoms, but these require early warning systems providing timely information on forthcoming pollen episodes. The performance of such systems can be enhanced by assimilating pollen observations from strategically well-placed monitoring stations. Before assimilating observations into forecasting models, the network quality and coverage of pollen monitoring stations must be evaluated.
Materials & Methods
We explored the network quality (i) and coverage (ii) of the five current pollen monitoring stations in Belgium using the methodology developed by Sofiev et al. (STOTEN, 2023).
1. Network Quality (i): Daily observations of grass and birch pollen were spatially interpolated using radial-based functions applied to the five monitoring sites in Belgium. The interpolated values were compared to a reference dataset containing spatio-temporal distributions of daily surface airborne birch and grass pollen levels from the SILAM model (System for Integrated modeLling of Atmospheric coMposition), as implemented in the operational early warning system of the Royal Meteorological Institute of Belgium. The Root-Mean-Square-Error (RMSE) quantified the network's ability to reproduce the concentration field over the region of interest.
2. Network Coverage (ii): A footprint-based analysis was performed at each monitoring station by running SILAM in backward mode (three days), determining the trajectory of birch and grass pollen observed at the network stations. Nine pollen seasons (2013–2021) were analyzed using ECMWF ERA5 meteorology.
Results
The network performs well with respect to the spatio-temporal reference dataset for birch pollen derived from SILAM. Over 78% of daily RMSE values for birch pollen are below their mean value, with more than 99% achieved during the 2015 season but only ~44% for 2020. For grass pollen, the network quality is 80%, with values ranging between 63% (2021) and 97% (2017).
The footprint analysis shows that the average coverage of monitoring stations for birch pollen is good, although large differences exist across the 2013–2021 seasons, likely due to significant inter-seasonal variations in birch pollen production. For grass pollen, the average coverage is better, with lower inter-seasonal variation.
Conclusions
Initial results indicate that the monitoring network for airborne birch and grass pollen in Belgium is of good quality and sufficient coverage. However, the performance varies with the applied time scale. At smaller time scales (e.g., days or hours), network configurations degrade faster compared to larger time steps (e.g., weeks, months, or seasons).
