064 The building of the bavarian electronic pollen information network - ePIN

Pollen monitoring is mostly performed manually. There is, however, a need for rapid reporting of pollen counts in addition to the alleviation of the workload of manual operation. We investigated the different steps needed for building an automatic pollen monitoring network in Bavaria, Germany. First, we tested the performance of the automatic pollen monitor BAA500, by comparing its data with data from a Hirst-type pollen trap at the same location. Next, during 2015, we built a new manual Bavarian pollen monitoring network with 27 Hirst-type pollen traps, including all closed but reopened operating traps from the German Polleninformation Network (PID). The selection of monitoring locations was based on: 1. Coverage of Bavarian land surface and population, but also 2. trying to represent the different environmental and climatic areas and 3. selecting the best positions to increase the predictability of pollen amounts based on forecasting models. Traps were located under optimal local monitoring conditions. The most informative locations were selected by applying redundancy and clustering analysis. Only 8 pollen traps are needed to cover most of variability in pollen dynamics. We found that three large groups of stations could be differentiated: 1. An extensive central cluster agglomerating most of the stations 2. a cluster agglomerating the stations located at the coldest areas of Bavaria and 3. a cluster of the Northern part of Bavaria. At the same time, we found the existence of eight sub-clusters related with different environmental conditions. We are now replacing these 8 manual stations by automatic pollen monitors and incorporating our network into local law to guarantee continuity.