Forecast model of allergenic hazard using trends of Poaceae airborne pollen over an urban area in SW Iberian Peninsula (Europe)

Cities are becoming bigger, being necessary the knowledge of associated natural hazards from organic and inorganic aerosols. This hazard could be included in the context of urban air pollution and climate change as environmental risk factors for allergy. Overall, grass pollens are the most important cause of pollinosis in Europe due to its high allergenicity and extensive distribution. The main objective of this work was to model daily average Poaceae airborne pollen concentrations from an urban area placed in a city in the SW of the Iberian Peninsula, taking into account the temporal distribution of five different meteorological variables from 23 years of continuous recording. This was achieved using a combination with the Shuffle Complex Evolution Metropolis Algorithm using as an optimisation function the root mean square error. Aerobiological sampling was conducted from 1993 to 2015 in Badajoz (SW Spain) using a 7-day Hirst-type volumetric sampler. The Poaceae Main Pollen Season lasted, on average, 89 days, ranging from 41 to 144 days, from April 17 to July 14. The model proposed to forecast airborne pollen concentrations is described by one equation composed of two terms. The first term represents the resilience of the pollen concentration trend in the air according to the average concentration of the previous 10 days, and the second term is obtained from considering the actual pollen concentration value, which is calculated based on the most representative meteorological variables multiplied by a fitting coefficient. The fit of the model was examined for a forecast horizon of 1, 7, 15 and 30 days. The R 2 values obtained were 0.70, 0.69, 0.62 and 0.57, respectively, which show a trend in decreasing order. These results confirm the suitability of the proposed model.