Abstract

LONDONO-CIRO, Libardo Antonio  and  CANON-BARRIGA, Julio Eduardo. Spatio-Temporal Characterization Methodology of PM2.5 in the Urban Area of the City of Medellín-Colombia. Rev.EIA.Esc.Ing.Antioq [online]. 2018, vol.15, n.30, pp.113-132. ISSN 1794-1237.  https://doi.org/10.24050/reia.v15i30.1217.

In this paper, we propose a methodology to characterize the spatiotemporal concentration of PM2.5 that integrates spatial interpolation, spatial statistics and econometrics. It is implemented with geographic information systems and monthly average concentration data of PM2.5, temperature and wind speed measured in eight monitoring sites during 2013-2014 in Medellín. First we make a multivariate spatial characterization of PM2.5 with Geographically Weighted Regressions (GWR) and then the temporal characterization with econometric algorithms. The model with explanatory variables temperature, spatial wind gradient, spatial domain of PM2.5 calculated with GWR and the concentration of PM2.5 with a lag of order one explains the variability of PM2.5 by 87%. The performance of the algorithms is calculated with cross validation. Finally, spatio-temporal characterization maps are adjusted with a spatial proximity model that uses distance to PM2.5 emission and mitigation sources.

Keywords : Air Pollution; Spatial Analysis; Geographically Weighted Regression; Autoregressive Distributed Lag Models.

        · abstract in Spanish | Portuguese     · text in Spanish     · Spanish ( pdf )