Olive cultivation is of great importance in Southern Europe but olive pollen is the leading cause of allergy in many regions where it is grown. The best preventive measure for allergic patients is to avoid exposure. Thus, aerobiological monitoring networks must supply realistic pollen classes for the different types of allergic pollen. Even though those pollen classes are defined, they do not necessarily fit local data. Altogether, they should use predictive models to assess flowering intensity in advance. In this study, the Olea pollen degree of exposure classes (OPDEC) are defined based on percentiles and a predictive model is suggested for Cartagena, Spain. 24year (1993-2016) Olea pollen counts series was used to characterize the Main Pollen Season (MPS). The aerobiological samples were processed following the methodology proposed by Hirst and developed by the Spanish Aerobiology Network. The aerobiological database was completed with the meteorological data supplied by AEMET (Spanish State Meteorological Agency). MPS evolution over time, and its relation with temperature and rainfall, has been analysed. The study showed an increase in MPS duration and the amount of Olea pollen grains collected both in MPS and the peak day. The OPDEC should fit local data to improve preventive measures. Based on the 24year series, the proposed OPDEC for Cartagena are: Low (≤10grains/m3), Medium (between 10 and 50grains/m3), High (between 51 and 100grains/m3) and Very High (≥100grains/m3). Olea pollen estimations in the MPS and in the peak day were obtained by means of three Regression Methods and climatic factors. The analysis reveals that the Bagging for Regression Trees (BRT) method is a good predictive alternative and stablishes the importance for each meteorological variable.
Keywords: Airborne pollen modeling; Bootstrap aggregating; Meteorological variables; Olea pollen degree of exposure classes; Olive pollen; Pollen forecasting.
Copyright © 2017 Elsevier B.V. All rights reserved.