Spatial and seasonal association study between \(PM_{2.5}\) and related contributing factors in India

Global environmental pollution and rapid climate change have become a serious matter of concern. Remarkable spatial and seasonal variations have been observed due to rapid industrialization, urbanization, different festive occasions, etc. Among all the existing pollutants, the fine airborne particles \(\varvec{PM}_{\varvec{2.5}}\) ${\mathit{PM}}_{\mathbf{2.5}}$ (with aerodynamic equivalent diameter \(\varvec{\le 2.5\mu m}\) $\mathbf{\le }\mathbf{2.5}\mathit{\mu }\mathit{m}$ ) and \(\varvec{PM}_{\varvec{10}}\) ${\mathit{PM}}_{\mathbf{10}}$ (with aerodynamic equivalent diameter \(\varvec{\le 10\mu m}\) $\mathbf{\le }\mathbf{10}\mathit{\mu }\mathit{m}$ ) are associated with chronic diseases. This leads to carry out the study regarding the varying relationship between \(\varvec{PM}_{\varvec{2.5}}\) ${\mathit{PM}}_{\mathbf{2.5}}$ and other associated factors so that its concentration level might be under control. Existing literature has explored the geographical association between the pollutants and a few other important factors. To address this problem, the present study aims to explore the wide spatio-temporal relationships between the particulate matter ( \(\varvec{PM}_{\varvec{2.5}}\) ${\mathit{PM}}_{\mathbf{2.5}}$ ) with the other associated factors (e.g., socio-demographic, meteorological factors, and air pollutants). For this analysis, the geographically weighted regression (GWR) model with different kernels (viz. Gaussian and Bisquare kernels) and the ordinary least squares (OLS) model have been carried out to analyze the same from the perspective of the four major seasons (i.e., autumn, winter, summer, and monsoon) in different districts of India. It may be inferred from the results that the local model (i.e., GWR model with Bisquare kernel) captures the spatial heterogeneity in a better way and their performances have been compared in terms of \(\varvec{R}^{\varvec{2}}\) ${\mathit{R}}^{\mathbf{2}}$ values ( \(\varvec{>0.99}\) $\mathbf{>}\mathbf{0.99}$ in all cases) and corrected Akaike information criterion ( \(\varvec{AIC}_{\varvec{c}}\) ${\mathit{AIC}}_{\mathit{c}}$ ) (maximum value \(\varvec{-618.69}\) $\mathbf{-}\mathbf{618.69}$ and minimum value \(\varvec{-896.88}\) $\mathbf{-}\mathbf{896.88}$ ). It has been revealed that there is a strong negative impact between forest coverage and PM pollution in northern India during the major seasons. The same has been found in Delhi, Haryana, and a few districts of Rajasthan during the 1-year cycle (October 2022–September 2023). It has also been found that PM concentration levels become high over the specified period with the temperature drop in Delhi, Uttar Pradesh, etc. Moreover, a strong positive association is visible in PM pollution level with the total population.