Seasonal outdoor PM10 changes based on the spatial local climate zone distribution

Abstract

Air pollution changes in urban and non-urban areas depend highly on the seasons and winds. However, it is scant to evaluate the effects of seasonality on air pollution, such as particulate matter (PM) using remote sensing data in Iran. Therefore, investigating the impacts of seasonal changes on PM₁₀ is imperative to mitigate its adverse effect. Local Climate Zone (LCZ) is a new approach in classification of urban land use and climate zones to estimate seasonal PM₁₀ changes in urban regions. In this article, seasonal PM₁₀ distribution changes were evaluated in terms of seasonality and spatial LCZ distribution in Tehran city. Machine learning and Random Forest algorithm were used to classify LCZs and Saraswat algorithm was used for evaluating spatial PM₁₀ distribution. The results showed that seasonality could significantly affect PM₁₀ levels in Tehran region. PM₁₀ levels in autumn and winter are much higher than that in spring and summer. There was the highest PM₁₀ level due to a low average precipitation in autumn while the lowest levels in summer. It is also found that the summer-autumn change caused substantial increases in all LCZs except for LCZ G of large water area. The largest percentage of increases in Tehran city was related to change of summer to autumn (93.9 %) while the largest decrease was in winter to spring (84.6 %). It was also found that PM₁₀ level changes more in the urban LCZs than in the non-urban LCZs.