Deep Learning Modelling of Trace Metal Bioaccumulation in Coastal Food Webs

Abstract

Coastal ecosystems represent some of the most productive ecosystems where food webs comprise of planktons, benthic animals, fishes and animals from higher trophic levels. Unfortunately, anthropogenic activities such as industrial discharge, agricultural runoff and urbanization are causing more and more traces of metals in coastal ecosystems. Persistent contaminants such as mercury, lead, cadmium and arsenic can be accumulated by organisms living in aquatic ecosystems and pass on through the food chain through processes of bioaccumulation and biomagnification. Consequently, they pose dangers to the health of marine organisms and marine ecosystem integrity. This study provides a deep learning-based modelling technique for evaluating the patterns of metal bioaccumulation in coastal food webs using environmental and biological factors. In this study, the model incorporates factors such as water quality factors, sediments' characteristics, contamination of primary producers, reactions of lower trophic organisms and fishes' accumulation of metals to investigate the relationship between environmental factors and metal transport. The developed model reveals complicated ecological relationships that arise due to metal concentration in various trophic levels and estimates possible patterns of metal accumulation in the coastal organisms. As the findings revealed, there is a clear pattern of bioaccumulation that is increasing with rising trophic groups, which means higher risk of contamination for organisms occupying higher positions in food chains. The estimated values of metal accumulation progressively increased from phytoplankton to predators, which confirms the impact of dietary accumulation and biomagnification process. The presented deep learning method ensures enhanced ability to analyze multidimensional ecological data and reveal interactions between ecological stressors and biological responses. The developed approach will be efficient for coastal pollution evaluation and identification of vulnerable organisms and contamination routes in marine ecosystems. This research demonstrates the significance of AI-based ecological modeling techniques to serve as a useful instrument for biodiversity conservation, pollution monitoring, and sustainable coastal ecosystems management.