Application of Exploratory and Spatial Data Analysis (EDA-SDA) to the Investigation of Metal Contamination in Groundwater from Electric Arc Furnace Slag and Dust

Ribeiro, Leonardo Guimarães; Bacellar, Luís Almeida Prado

Application of Exploratory and Spatial Data Analysis (EDA-SDA) to the Investigation of Metal Contamination in Groundwater from Electric Arc Furnace Slag and Dust

2015

Ribeiro, Leonardo Guimarães | Bacellar, Luís Almeida Prado

Electric arc furnace (EAF) dust and slag, materials which contain high metals in their composition, were improperly disposed in an industrial steel mill site between 1963 and 1999. Previous environmental investigations identified anomalous concentrations of metals in local groundwater but failed to relate these abnormalities to the disposed material or to natural geochemical processes. Aiming to identify the origin of such abnormalities, exploratory and spatial data analysis (EDA-SDA) method was applied on a hydrogeochemical data set obtained through 5 sampling campaigns in 32 groundwater monitoring wells installed upstream and downstream of the area impacted by the steel mill activities. Boxplot class-based and Eh vs. pH maps of physicochemical log-transformed data identified that wells located under the influence of EAF slag deposits in topographic hollows had lower Eh potential and increased electrical conductivity and pH, when compared to wells in the topographical nose of the surveyed area. Metal distribution maps showed that Al, Ca, K, Mg, Na, and Sr were consistently higher in topographic hollows while concentrations of Co, Cu, Cr, and Li were higher near the former steel-making plant, located in the topographical nose. Ba, Fe, Mn, and Zn, important indicators of EAF slag and dust, were observed in both topographic settings. Variable clustering was able to capture the relations among metals and thus validate the log-normalized data structure to be used into wells clustering. Clustering through the mclust algorithm carried out for two and three clusters allowed the distinction among localities that received an input of metals from dust or slag and those not influenced by either residue. This paper demonstrates that EDA-SDA is an effective method to identify areas under the influence of contamination from industrial activities from areas not affected by anthropogenic contamination.

Показать больше [+]