Improvement of phenological models by means of time series of EO (Earth Observation) data for numerical pollen forecast

Allergic rhinitis has become a health problem in the Western countries. Timely and reliable warning before an increase of the atmospheric pollen concentration means a substantial support for physicians and allergy suffers. Recently developed numerical pollen forecast models have become a means to support the pollen forecast service, which still require refinement. This project aims at improving the phenological modeling in numerical pollen forecast procedures by assimilating Earth-Observation based temperature and vegetation information into the phenology model.

The expectation of improving the accuracy of the simulated 2 m temperature fields through a data fusion with satellite data has partly been achieved. Results show that the improvement is more evident for low elevation areas, while in high elevation areas the use of satellite data reduces the accuracy of the modelled temperature data. The correlation between the remotely sensed MODIS (Moderate Resolution Imaging Spectroradiometer) Land Surface Phenology and Ground Phenology collected via eye observation turned out to be insufficient in order to be useful for numerical pollen modelling.