Particulate Matter Prediction Model (PM2.5) in Bogota city

Abstract

Air pollution has been widely studied due to the great threat it poses to both the environment and human health. Mainly, the big cities have been affected by the problem of air pollution due to various factors, such as industrialization and overpopulation. The latter have a direct implication in the atmosphere by increasing greenhouse gas emissions and smog, which contributes to the depletion of the ozone layer. Air pollution brings serious implications for human health, directly affecting the respiratory system and generating consequences such as throat irritation, worsening conditions such as asthma, bronchitis and even permanent lung damage. However, considering the great importance of detecting polluting agents, such as particulate matter (PM), it is possible to propose alternatives to predict their appearance in a certain location. In this study, we propose the use of computational tools and multiple linear regression to predict the behavior of PM2.5 particulate matter from the Kennedy locality, Bogotá. This prediction was made with the implementation of an algorithm, which is based on historical data of particulate matter suspended in the air, Ozone, NO, CO, wind direction and speed during the year 2018. The analyzes carried out were compared with real data, with the purpose of determining the veracity of the predictions made by the algorithm. The results revealed the level of PM2.5 is a consequence of CO emissions, which is related to vehicle traffic, since they are the largest producers of this gas. In addition, the prediction of PM2.5 values ​​showed a mean error of 5.6 PM2.5, which means that the model achieves an accurate prediction of air quality. Considering these results, this study represents a promising alternative for the prediction of particulate matter, in such a way that action routes can be established to mitigate the damage caused by these pollutants in the environment.