Assessing pollen extreme events over a Mediterranean site: Role of local surface meteorology

The presence of very high pollen levels in the atmosphere is associated with a strong impact on health and a worsening of symptoms in people who already have a respiratory disease. However, there is no specification on the aerobiological, environmental and meteorological factors that allow for characterizing a pollen event as of great magnitude due to the significant impact it can cause on the population and the environment. This work proposes criteria to typify the levels of atmospheric pollen as an extreme pollen event (EPE), and aims to determine the meteorological variables that can affect the presence and permanence of high pollen concentrations over a period of time. To address this goal, the quasi-climatological pollen dataset recorded in Granada (Southeastern Spain) during the period 1992-2019, has been used. On the daily accumulated pollen concentrations, the 95th, 97th and 99th percentiles were calculated. Spearman's correlation between the pollen con-centration exceeding the proposed thresholds (C (> P95), C (> P97), C (> P99)) and surface meteorological variables recorded during up to five days before the event were established in order to identify the meteorological conditions that might affect the EPEs. As for the number of days with values higher than the established percentiles, it has been seen that in the case of total pollen and Olea, Cupressaceae and Pinus, there is a robust monotonically ascending trend throughout the study period. Regarding meteorological variables, relative humidity and 24-h accumulated precipitation are shown as the two most influential variables up to three days before the event, although temperatures, visibility and wind direction also show a correlation with some pollen types. The criteria proposed in this work allow us for classifying high levels of pollen as an EPE, and lay the foundations of these extreme events in a context of climate change in which they will become more frequent.