Predicting pollutant concentrations in rivers exposed to alluvial gold mining in Mazowe Catchment, Zimbabwe

Alluvial gold mining poses a threat to river water quality, yet methods for assessing its impacts on water quality remain laborious and costly. Water quality modelling offers an alternative, which is cost-effective. In this study, we assess the effect of alluvial gold mining on surface water quality at three sites located along the Mazowe and Mwenje rivers in northern Zimbabwe using a simple exponential decay model. The model predictions were regressed against measured in-situ Total Suspended Solids (TSS) and turbidity. Strong linear relationships were observed between the measured and predicted pollutant concentrations. Specifically, the results indicate that TSS concentrations can be predicted in the catchment, with high accuracy (R2 values of 82%, 88% and 93% and root mean square errors (RMSEs) of 5, 2 and 3 mg/l across the three monitoring sites, respectively). Similarly, comparison between predicted and measured values river turbidity equation showed high prediction accuracies for the three sites. The results showed that water turbidity could be predicted at R2 = 77%, 88% and 93% at sites 1, 2 and 3, with correspondingly low RMSEs of 4.5, 1.4 and 2.7 NTU, respectively. Overall, these results provide an alternative approach to assess the impact of artisanal mining on physical water quality in a tropical environment.