Long-term trend of PM2.5 over five Indian megacities using a new statistical approach

PM(2.5 )is one of the major contributors to air pollution in India, resulting in poor air quality and human health issues. The current study explores the applicability of TTAinterfaceTrendAnalysis (an 'R' language-based Graphical User Interface package) in estimating the annual and month-wise trends in PM2.5 over five Indian megacities (New Delhi, Kolkata, Mumbai, Hyderabad, and Chennai). The package offers a one-stop solution for various statistical analyses such as data imputation, outlier detection, trend analysis, data smoothing, diagnostic tests, etc. The trends are estimated using seven (2014-2020) years of PM2.5 data. All estimated month-wise trends of PM2.5 over the study cities were found to be statistically non-significant. Annual trends (both with and without applying data smoothing techniques) were found to be statistically significant with magnitudes ranging between - 0.27 and - 5.2 mu g m(-3) y(-1) (negative sign indicates a declining trend). Autocorrelation analysis and normality tests are also conducted using the inbuilt options of TTAinterfaceTrendAnalysis. Results obtained are thoroughly discussed and compared with those of similar contemporary studies. To our knowledge, the current study is the first to apply TTAinterfaceTrendAnalysis for quantifying long-term trends in PM2.5 . The sensitivity of the trend estimates due to the inclusion of COVID-19 lockdown PM2.5 data was also examined. The net impact of lockdown on the PM2.5 long-term trend ranged between - 0.13 and 2.29 mu g m(-3) y(-1) across the study cities.

Automated summary

This study examines the applicability of TTAinterfaceTrendAnalysis in estimating the annual and month-wise trends in PM2.5 over five Indian megacities. The results show statistically non-significant month-wise trends, but statistically significant annual trends. The study also considers the inclusion of COVID-19 lockdown PM2.5 data and examines its impact on the long-term trend.