Prognosis of metal concentrations in sediments and water of Paraopeba River following the collapse of B1 tailings dam in Brumadinho (Minas Gerais, Brazil)

Abstract

In January 25, 2019, the B1 dam of Córrego do Feijão mine located in Brumadinho municipality (Minas Gerais, Brazil) collapsed and injected nearly 2.8 Mm3 of iron (Fe)- and manganese (Mn)-rich tailings in the Paraopeba River. This study assessed the contribution of tailings to the contamination of sediments and water by those metals. The dataset was built through daily to weekly samplings executed in the two years following the event, at 27 sites located along the Paraopeba plus 9 sites located at the confluence of main tributaries. The results evidenced a distinct contribution in the sectors “Anomalous” (8.6–63.3 km downstream from the dam) and “Natural” (115.8–341.6 km). The “Anomalous” sector presented large Fe/Al (12.2 ± 6.4) and Mn/Al (0.33 ± 0.19) ratios in sediments, thus being rich in tailings, while the “Natural” sector presented small ratios (2.4 ± 1.0; 0.06 ± 0.03) comparable to the natural sediments. A 500–700 m3/s stream flow discharge in the Paraopeba caused pronounced drops to the Fe/Al and Mn/Al ratios in the “Anomalous” sector, attributed to the mixture of contaminated sediments from the main water course with uncontaminated sediments injected by the tributaries during the event. Non-linear regressions showed Fe/Al and Mn/Al declines in the “Anomalous” sector, related with tailings mobilization downstream. The concentrations of Fe and Al in the sediments correlated positively with the corresponding concentrations in the Paraopeba water, conditioned by raising discharge rates and variations in the water pH. The contribution of tailings to the Fe correlation was demonstrated. No direct relation was established between the Mn concentrations in water and stream discharge, because manganese is associated with fine particles in the tailings that are mobilized to the water column even under low flows. The preliminary results of Seasonal Autoregressive Integrated Moving Average models predicted the return of Paraopeba to a pre-collapse condition in 7–11 years.