Abstract
Climate impacts on the pollen season are well-described however less is known on how frequently atmospheric transport influences the start of the pollen season. Based on long-term phenological flowering and airborne pollen data (1987-2017) for six stations and seven taxa across Bavaria, Germany, we studied changes in the pollen season, compared pollen and flowering season start dates to determine pollen sources, and analyzed the likelihood of pollen transport by HYSPLIT back trajectories. Species advanced their pollen season more in early spring (e.g., Corylus and Alnus by up to 2 days yr(-1)) than in mid spring (Betula, Fraxinus, Pinus); Poaceae and Artemisia exhibited mixed trends in summer. Annual pollen sums mainly increased for Corylus and decreased for Poaceae and Artemisia. Start of pollen season trends largely deviated from flowering trends, especially for Corylus and Alnus. Transport phenomena, which rely on comparisons between flowering and pollen dates, were determined for 2005-2015 at three stations. Pre-season pollen was a common phenomenon: airborne pollen was predominantly observed earlier than flowering (median 17 days) and in general, in 63% of the cases (except for Artemisia and Poaceae, and the alpine location) the pollen sources were non-local (transported). In 54% (35%) of these cases, back trajectories confirmed (partly confirmed) the pre-season transport, only in 11% of the cases transport modeling failed to explain the records. Even within the main pollen season, 70% of pollen season start dates were linked to transport. At the alpine station, non-local pollen sources (both from outside Bavaria as well as Bavarian lowlands) predominated, in only 13% of these cases transport could not be confirmed by back trajectories. This prominent role of pollen transport has important implications for the length, the timing, and the severity of the allergenic pollen season, indicating only a weak dependency on flowering of local pollen sources.