Immission levels and identification of carbon monoxide sources in a latin American megacity: use of bivariate polar plots and k-means clustering

Abstract

The identification of the main air pollution sources becomes important in cities with rapid economic growth of developing countries in order to effectively prevent and control atmospheric pollution. In this research, carbon monoxide (CO) concentrations recorded during the period 2015–2019 at four air quality monitoring stations in the Metropolitan Area of Lima-Callao (MALC) were evaluated with the aim of determining the immission levels and identifying the main emission sources of this pollutant. Bivariate polar plots and the k-means algorithm were used to identify and classify areas with similar pollution characteristics, and the conditional bivariate probability function was used to identify potential zones where the largest contributions (≥ 75%) to air pollution in the study area originate. The average CO concentration ranged from 579.3 ± 3.8 µg/m³ to 1490.9 ± 4.2 µg/m³. On average, the hourly variation showed two peaks, one in the morning (0700–0900 h LT) and another in the evening (1900–2100 h LT). On a weekly basis, the lowest CO concentrations were observed on Sundays. The results showed that the major contributions of CO come from gaseous emissions of vehicular traffic of 2- and 3-wheeled motorcycles, cars, combi vans and station wagons on the main avenues and highways using gasoline/gasohol as the main fuel, followed by diesel, LPG and NGV. The trends and patterns identified in this study show the vehicular flow on these major roads and fuel type as the main sources of CO pollution. A complementary analysis of black carbon concentrations revealed that forest fires and regional CO transport would influence the concentrations of this pollutant in the MALC. These findings contribute to the understanding of decision-makers in establishing strategies for improving air quality in metropolitan megacities such as the MALC.