Geodaysit 2023

Mohammad Mehrabi


Air Quality Monitoring and Prediction in Ukraine During War Crisis Using Copernicus Data and Machine Learning
Marco Scaioni, Mohammad Mehrabi, Mattia Previtali

In late February 2022, the invasion of Russia in the Ukrainian territory started. As is known, air is one of the most affected components of the environment during such exceptional circumstances. The changes in the pattern of civilian and industrial activities may cause the variation of air quality in terms of different pollutants. Hence, conducting proper air quality assessment can be of great importance in the war-affected areas. The pivotal objective of this research is to present an overview of air quality monitoring and air pollution prediction carried out for Ukrainian territory. Utilizing the Copernicus Sentinel-5P TROPOMI observations, the emissions of ozone (O3), nitrogen dioxide (NO2), formaldehyde (HCHO), and carbon monoxide (CO) in Kiev, Kharkiv, Donetsk, Kherson, and Lviv are monitored during 2022. The relevant records are compared to the same business-as-usual (BAU) periods in 2019 and 2021 to detect significant changes. Visual interpretations supported by statistical analysis proved that the ongoing war has significant impacts on the concentration of pollutants throughout Ukraine. Following this, a hybrid machine learning model is developed to predict the concentration of a well-known air quality indicator called particulate matter 2.5 (PM2.5). The prediction results indicated a reliable accuracy of the proposed methodology, as well as its superiority over benchmark models. In short, this research shows promising application of state-of-the-art technologies inducing remote sensing and artificial intelligence for solving air quality problems in during exceptional events.

AIT Contribution
Sala Biblioteca @ PoliBa