Trend patterns of nitrogen dioxide: OMI measurements and Machine Learning to understand the global anthropogenic emissions

Abstract

Abstract Tropospheric dioxide nitrogen is one of the criteria pollutants considered a toxic gas that contributes to climate change and affects public health. Anthropogenic activities are the primary NO2 sources affecting the planetary ecosystems. The Ozone Monitoring Instrument (OMI) on board AURA Missions is one of the most robust projects contributing to the NO2 investigation. A methodological coupling based on spatial analysis, clustering, machine learning, and statistical validation helped to analyze the OMI satellite data and its interactions with socioeconomic factors. Spatial contrasts show differences between continental and marine domains, highlighting the influence of coastal urban centers on the near marine areas; differences between hemispheres and latitudinal changes overall in the Pacific and Atlantic oceans; contrasts between urban and rural areas in continents; and outstanding regions by their high NO2 emissions. The trend analysis outlined regional contrasts that contribute to understanding the impact of economic activities and environmental policy implementation. From 2005 to 2021, trend patterns characterization established the framework to correlate the population size and the GDP of more than 250 developed urban centers worldwide. Prominent maximum NO2 densities between 2011 and 2013 stand out among the four trend patterns, outlining a point inflection (peak component) in the trend direction of several regions. The correlation, including all cities, followed a significant moderate relationship (R=0.573, p~0.000) where the population explained 33.7% of the productivity. However, the correlations by subgroups considering trend pattern classification indicated significant moderate to strong relationships for almost all trend types (R from 0.689 to 0.814, p~0.000), where the population explains 47.5 to 66.2% of the productivity. These results partially show the direct cause-effect relationship between the high NO2 emissions and development levels in urban centers. Conversely, the wide scattering in such correlations suggests the gradual and positive effects of Environmental policies in favor of better air quality, different from the sudden decrease in NO2 densities caused by confinement and preventive measures against COVID-19.