Forecasting airborne Platanus pollen in the Madrid region

This study analyses the atmospheric concentration of Platanus pollen in four stations in the Madrid region over a period of 10 years (1994–2003). Various statistical analyses (regression analysis and decision tree) were used to prepare a forecasting model for possible application as a preventive measure in pollinosis. The data comes from the PALINOCAM network and the samplers used were Hirst type (Burkard pollen trap). Platanus pollen is present in the atmosphere during a short period of time in spring, and the maximum concentrations are detected during the last two weeks of March and the first week of April. Regression analysis shows that the pollen concentration of the two previous days is the best predictive variable. The models obtained for the four stations analysed account for between 37 and 61% of the variation in pollen levels in the air. The decision trees show how the introduction of meteorological variables improves prediction for this pollen type.