Assessment of Spatial Analysis Methods in Mapping of Air Pollution in Mashhad

Abstract

Background: This study aimed to compare and evaluate the spatial and statistical models to predict PM2.5 concentrations at ground level and at the macro scale in Mashhad. Methods: To investigate the status of air pollution in the metropolis of Mashhad air, three interpolating models including Ordinary Kriging (OK), Universal Kriging (UK) and inverse distance weighting (IDW) were used. Root Mean Square Error (RMSE) and correlation coefficient (R2) were employed to compare three models and choose the best one. As well as to select the most optimal conditions for the implementation of both OK and UK, used from Standardized RMSE. Results: The results showed that the highest monthly average of PM2.5 was belonged to September and “Sakhteman” station (95.1 μg/m3). Also, the lowest monthly average pollution had happened in "Torogh" station, in November (15.5 μg/m3). According to the data, the OK had the lowest RMSE (10.601) compared to the UK and IDW. Lower RMSE represents lower error between the predicted and measured values. So, OK model selected as better one in interpolation. Also, Judging by correlation coefficient (R2), the highest correlation belonged to OK compared to other two models. UK model showed a greater standard error of predicts than OK. The greatest standard errors of prediction were related to areas that have more distance from air pollution monitoring stations. Conclusion: it should be noted that the production and use of geo-referenced maps could quickly provide spatial analyses, and because it can be combined with GIS, the user is able to investigate the influence the various concentrations of contaminants.