Cluster analysis of back trajectories and atmospheric pollen in Catalonia, NE Spain

We propose in this study the use of back-trajectories and their analysis in clusters as a new tool in analyzing the provenance of 5 airborne pollen types measured in Bellaterra (Catalonia, NE Spain) during a ten-year period (2000-2009). The pollen types selected for a first analysis are: Ambrosia, Betula, Fagus anddeciduous and perennial Quercus.For each day in the period under study we calculated the 96-hours three-dimensional back-trajectories at the heights of 500, 1500 and 3000 m.a.s.l. at 12 GMT, using HYSPLIT-4 of NOAA. The trajectories were clustered into groups. Mean daily pollen concentrations were then used to establish the contribution of each of the selected pollen taxa to the different cluster group. The number of groups was determined depending on the significant differences found between the number of clusters and the pollen type, as a result of anANOVA and Tukey test.First results showed that the analysis at 1500 m.a.s.l. represents better the pollen transport and that the air-mass esat 1500 m.a.s.l. reaching NE Spain during the study period can be grouped into 6 clusters for Fagus and 8 clusters for Betula.For Betula and Fagus the clusters containing the highest percentage of pollen at 1500 m.a.s.l. corresponded to medium distant back-trajectories coming from France and Central Europe. These results are in agreement with previous studies in the group using source-receptor models. For deciduous and perennial Quercus and Ambrosia the results showed no significant differences between clusters at any of the tree studied heights. Contrary to the clusters at 1500 m, the heights of 500 and 3000 m showed no statistically significant differences between groups, comparing the 5 pollen types and the number of clusters at each height respectively.