Identifying the cause of soil cadmium contamination with Monte Carlo mass balance modelling: a case study from Potosi, Bolivia

Abstract

The Chayanta river in Potosi, Bolivia is polluted by present and past mining activities in the districts of Siglo XX and Llallagua. The river water, which is enriched with cadmium (Cd), is used for irrigation in the Quila Quila and Asiruri valleys where the median soil Cd concentration is 20 mg kg(-1), well above the background value of < 0.5 mg kg(-1). The objective of this study was to predict current soil contamination using a retrospective mass balance. Monitoring data were collected on Cd concentrations in irrigation water, irrigation application rates, crop yield and composition, and soil properties including pore water composition. The mass balance was made assuming constant model parameters since the start of upstream mining operations about 85 years ago (1920). The parameter uncertainty was taken into account with a Monte Carlo analysis. The current annual Cd input by irrigation is 800 g ha(-1). The annual output of Cd through removal of the crop harvest and leaching is less than 10 g ha(-1). The predicted soil Cd concentrations after 85 years of contamination (geometric mean: 21.9 mg x kg(-1), 10th and 90th percentile 7.2-65.1 mg kg(-1) respectively) matched the distribution of observed values (geometric mean: 18.6 mg kg(-1); 10th and 90th percentile 4.7-65.9 mg kg(-1) respectively; n = 56). This study confirmed that irrigation water is the prime source of soil Cd enrichment in that area. The Monte Carlo analysis is a convenient way of including parameter uncertainty in mass balance modelling and of estimating spatial variability of the contamination.