Statistical Assessment of Spatial Autocorrelation on Air Quality in Bengaluru, India

Abstract

Background The term “air quality” refers to how clean or unhealthy the air is. Polluted air may be harmful to human health as well as the health of the ecosystem, thus it’s crucial to monitor air quality. Through previous studies it has explained the significance of outdoor air quality using spatial techniques - Geographical information System (GIS). Objective Using Spatial autocorrelation techniques - GIS and Moran’s I to analyze the spatial patterns for air quality in Bengaluru, Karnataka, India, 2015. Methods Ten years of air quality data for Bengaluru from 1995 to 2015. The temperature, annual rainfall and $$CO_{2}$$ content in air was considered. Each 198 wards of Bengaluru were geocoded in an R Project tool with spdep library. The hedonic regression model was applied to estimate the GIS based distribution of air contaminants for each ward. To investigate the geographical connection between temperature, rise and yearly rainfall, spatial auto-correlation method (Global Moran’s I) was used. The primary reason and the importance of spatial autocorrelation are that statistics rely on observations being independent of one another. Results At the Central Business District (CBD) of Bengaluru, concentrations of CO2 showed statistical significance spatially clustered. The p-value is 0.74 and Moran I standard deviation is −0.65 whereas in monte-Carlo simulation of Moran I the statistic -s 0.006 with p value 0.77. It examines positive and negative correlation for the normal distribution due to air quality. Finally, autocorrelation is visualized in a GIS map, then the assumptions of observations independent of one another are broken. Conclusion These results help to inform that the air pollution problem is increasing temperature and annual heavy rainfall in Bengaluru, 1995–2015. Interesting Spatial patterns were observed using spatial autocorrelation modeling and GIS in Bengaluru.