Journal of the Bulgarian Geographical Society 48: 3-14, doi: 10.3897/jbgs.e99206
Modeling of arsenic dynamics in groundwater of а river floodplain contaminated with mine tailings: Ogosta River case, NW Bulgaria
expand article infoZvezdelina Marcheva, Tsvetan Kotsev, Assen Tchorbadjieff§, Velimira Stoyanova
‡ National Institute of Geophysics, Geodesy and Geography - Bulgarian Academy of Sciences, Sofia, Bulgaria§ Institute of Mathematics and Informatics - Bulgarian Academy of Sciences, Sofia, Bulgaria
Open Access

This study aims to reveal the arsenic dynamics in groundwater of а river floodplain contaminated with mine tailings under temperate climate conditions and natural river hydrodynamics. Arsenic concentrations were monitored in the primary morphological units of the floodplain in the upper stretch of the Ogosta River in NW Bulgaria. Iron, lead-silver, and gold mining heavily affected the river valley in the second half of the 20th century. We used groundwater monitoring data from 21 piezometers for the period 2016-2020. Based on the geochemical and geomorphological conditions in the valley, the piezometers were grouped into three clusters. Regression models were developed for each cluster and representative piezometers to predict arsenic concentrations. In the active floodplain, seasonal fluctuations in arsenic concentrations followed the river and groundwater regime. In this part of the valley floor, we determined two periods of elevated arsenic concentrations during the spring and autumn/winter seasons that coincide with high river water stages. Arsenic content in the groundwater of the higher floodplain was less dependent on the water level fluctuations but followed changes in redox potential, electrical conductivity, and water temperature. The obtained results showed the elaborated models as valuable tools for studying arsenic dynamics in alluvial aquifers of contaminated river floodplains. The suggested models could be coupled with groundwater monitoring systems to monitor arsenic concentrations and identify periods of the year with levels below and above threshold values.

alluvial aquifer contamination, generalized linear regression models, monitoring data