Author(s): Basmah Bushra; Leyla Bazneh; Lipika Deka; Paul Wood; Diganta Das

Linked Author(s): Basmah Bushra

Keywords: Metal pollution; Aquatic ecosystems; Time series analysis; Multiple regression analysis; Artificial neural network

Abstract: The long-term effects of heavy metal contamination in aquatic ecosystems are a major environmental concern internationally due to their potential persistence in the environment and bioaccumulation in aquatic organisms and the food chain. Despite significant historic and current issues associated with metal pollution, there have been comparatively few contemporary studies examining the long-term legacy and effect of heavy metals on aquatic ecosystems in post-industrial areas. This paper examines a series of lakes within the Attenborough Nature Reserve, UK as a model system to study the effects of heavy metal contamination on the aquatic environment. Attenborough Nature Reserve is a protected area in Nottinghamshire, UK located near the confluence of the River Erewash and River Trent. A 15-year hydrological and meteorological dataset was analysed to identify the dominant trends and statistical correlations which may have driven historical pollution levels within the lakes. A predictive model was developed for forecasting the heavy metal concentrations of the nature reserve based on physico-chemical, biological and meteorological parameters by combining complementary methods; namely, time series analysis, multiple regression analysis and artificial neural network (ANN) analysis. Correlation analysis indicated that electrical conductivity values (Pearson correlation coefficient, r = 0.22) and river flow volume (r = 0.23) were significantly associated with metal concentrations. Further, time series analysis showed that the lakes directly connected to the inflowing river received a greater pollution load than the unconnected lakes, and that copper (Cu) and zinc (Zn) were the metals of greatest concern. Electrical conductivity, total suspended solid (TSS), flow volume and precipitation were the best predictors in regression models (explaining 13.4% of the variance for copper and 10.1% of the variance for zinc). Dissolved oxygen (DO) and TSS were the best predictors for the developed ANNs. Overall, the results suggest that the accuracy of ANN models was the highest among the methods tested. The successful selection and application of the most appropriate technique is important for rapid assessment because high metal concentrations in aquatic ecosystems and bioaccumulation may ultimately lead to significant human health implications via transfer within food chains. Field sampling and laboratory experiments to characterise the metal concentrations in river water, sediment and the potential bioaccumulation within invertebrates are currently in progress. Future research will explore the bioaccumulation by specific target organisms to establish theoretical model (s) to determine the bioaccumulation kinetics. The results of this research will subsequently be used to identify and test potential remediation techniques to mitigate the effects of heavy metal pollution.

DOI: https://doi.org/10.3850/IAHR-39WC2521711920221418

Year: 2022