Correlation and principal component regression analysis for studying air quality and meteorological elements in Wuhan, China

Abstract

The states of human health and the environment are closely related to air quality (AQ). Further, in addition to pollutant emissions, meteorological elements are key factors that affect AQ. In this study, variations in the air quality index (AQI) were analyzed; the relationships between AQI and meteorological elements during 2013–2017 in Wuhan were examined, using the Pearson correlation coefficient (PCC) technique. Meanwhile, the principal component regression (PCR) technique was used to predict the daily AQI based on the previous day's AQI and five meteorological variables. An obvious seasonal pattern was seen in the AQI variations. Temperature, relative humidity, precipitation, and wind speed were negatively correlated with AQI, while atmospheric pressure was positively correlated with AQI for the entire study period. The performance of the PCR model was evaluated using different statistical indicators. The results showed that the PCR model could predict the daily AQI effectively at all six monitoring stations with determination coefficient (R2) values of 0.549, 0.563, 0.561, 0.596, 0.534, and 0.602, respectively.