Change detection using CUSUM and modified CUSUM method in air pollutant concentrations at traffic site in Delhi

Understanding the changes in the air pollution of an area due to implementation of control strategies is important as it helps in making further action plans. Time series analysis provides ways to interpret the effect of any policy changes. In this study, the applicability of the CUSUM method for change detection in air pollutant concentrations in Delhi is investigated. The method detects any shift from mean of the process. Delhi has undergone major policy changes during the past few years. Change of fuel in vehicles to compressed natural gas (CNG) is one amongst them. The data observed at a traffic site in Delhi for nitrogen dioxide (NO₂), carbon monoxide (CO) and particulate matter (with size less than 10 micron-PM10) concentrations is used to carry out the analysis. Increase in NO₂ concentration and decrease in CO concentration levels is observed using CUSUM method. The choice of base period does not affect much for these two pollutants but for PM10 concentration, however its role is crucial. In order to counter any variability shifts, the CUSUM method is further modified to account for the change in the variance of the time series. Modified CUSUM method indicated similar nature of variability in NO₂ and PM10, whereas CO variability has decreased significantly after CNG implementation.