Data mining indicates an association between ambient PM2.5 levels and wind speed in an urban environment (Education City, Doha, Qatar)

Abstract

Background: Elevated PM2.5 levels pose serious health hazards and are implicated in numerous acute and chronic conditions. Delineating the contributions of meteorological factors to PM2.5 levels is a daunting task, especially in confined or semiconfined urban spaces. This study aims to (1) characterize the influence of wind speed and direction on outdoor PM2.5 levels within a semiconfined urban environment, and (2) develop a simple and readily accessible data mining method for this purpose. The ultimate goal is to evaluate the extent to which PM2.5 correlations demonstrated in open spaces hold in semiconfined outdoor settings with irregular terrain. Methods: In this study, data mining techniques were applied to retrieve patterns pertaining to the effects of meteorological factors on PM2.5 levels. As a proof of concept, a feasible framework was developed to elucidate the associations between wind speed and direction and PM2.5 levels during May 2020 in Education City, Doha, Qatar. Results and Discussion: The results showed a modest negative correlation between wind speed and PM2.5 levels, at low to moderate, but not high, PM2.5 readings. Meanwhile, no correlation was detected between wind direction and PM2.5 levels. Conclusions: Limited by the geographical location, microenvironment, and duration of this study, it can be said with moderate statistical confidence that low PM2.5 readings are associated with high wind speeds. As a result, increasing wind speed may be beneficial at low to moderate PM2.5 levels. However, delineating a single contributing factor to high PM2.5 readings may prove infeasible. Moreover, an association with wind direction was not immediately obvious, possibly due to microenvironmental limitations. These findings underscore the applicability of data mining and the importance of microenvironmental factors in air quality research and mitigation.