Prediction of airborne pollen concentrations by artificial neural network and their relationship with meteorological parameters and air pollutants

After the early rainfall in the autumn of 2013, respiratory syndromes spread in the Khuzestan province of Iran with the most severity in Ahvaz. There have been recurring outbreaks in recent years. Considering that pollen-derived airborne allergens are regarded as key aeroallergens and the main cause of allergic rhinitis and asthma, this work aimed to forecast total pollen concentration in Ahvaz through an artificial neural network (ANN), followed by evaluating the pollen spatial distribution across the city and the association between pollen concentrations and environmental parameters. The utilized ANN in this work included an input layer with 13 parameters, a hidden layer of five neurons, and an output layer. Data were classified into training, validation, and testing sets. The ANN was implemented with 70% and 80% of data for training. The value of the correlation coefficient for the data validation of these two networks was 0.89 and 0.92, respectively. The results also indicated that despite the difference in the mean concentration of the pollens in various areas of Ahvaz, this difference was not statistically significant (P > 0.05). Furthermore, there was a negative correlation between the concentration of total pollen and relative humidity, precipitation, and air pressure. However, it had a positive correlation with temperature. Consequently, considering the logistical challenges of monitoring bioaerosols in the air, the ANN approach could predict total pollen concentrations. Therefore, in addition to measurements, the ANN technique can be a good tool to enable authorities to mitigate the impact of airborne pollen on people.