Linking airborne pollen concentrations and meteorological conditions in Greece

Atmospheric pollen concentrations exhibit strong temporal variability driven by plant
phenology and affected by meteorological conditions, yet the strength and consistency of
pollen-meteorology relations depend on region and pollen taxon. In this study, we examine
the relationships between daily airborne pollen concentrations and meteorological
parameters across three regions in Greece, i.e. Athens, Thessaloniki and Finokalia (Crete)
using multiple meteorological datasets and correlation metrics. Pollen observations were
obtained from aerobiological monitoring stations, while meteorological information was
derived from ERA5 reanalysis and nearby meteorological stations. Additionally, for Athens
and Thessaloniki, the Climpact dataset with meteorological parameters was used.
Correlation analyses were performed for selected pollen taxa, grouped pollen categories and
total pollen concentrations using Pearson, Spearman, and Kendall correlation coefficients.
Lagged correlations were also examined for delayed influences on pollen variability, and
analyses were conducted with both including and excluding zero-pollen days.
The results indicate marked regional and taxon-specific variability in pollen–meteorology
relationships. Temperature and relative humidity exhibit the strongest associations with
pollen concentrations, with correlation values ρ reaching -0,75 and 0,57 respectively. Non-
parametric Spearman correlation coefficient provides more stable relationships compared to
Pearson correlation, particularly for taxa with highly skewed distributions. Across data
sources (ERA5, station observations, and climpact datasets), correlation estimates are
generally comparable, suggesting that the pollen–meteorology relationships are robust to
the choice of meteorological dataset.
Overall, the results demonstrate that airborne pollen concentrations are systematically
related to meteorological conditions, with the strength and structure of correlations
depending primarily on region, pollen taxon, and the statistical approach applied.