Predicting essential and hazardous element concentrations in marine fish from the Northeast Atlantic Ocean: A Bayesian approach.

Micronutrient deficiency or 'hidden hunger' is of growing importance regionally and globally. Marine fish have the potential to mitigate hidden hunger although certain contaminants they often contain may also pose a health risk. Understanding biological and environmental drivers behind essential and hazardous element concentrations is therefore important to develop evidence-based advice for adaptive management strategies. We use Bayesian models to predict concentrations of ten essential and two hazardous elements in fillets of 14 marine fish species in the Northeast Atlantic Ocean. Data from 15,709 individuals of six lean, five semi-fatty, and three fatty species were included. Fish length, fat content, ocean basin, sea temperature and salinity were used as predictor variables. We found good model predictability and identified some important trends in driver effects. Fish length was the most important driver of element concentrations for most species with a negative effect for calcium, copper, manganese, and arsenic, and a positive effect for mercury, suggesting that smaller individuals may be a safer and better source of essential elements. Ocean basin was also an important driver in most cases. For concentrations of selenium, zinc, and mercury, effect sizes of ocean basins increased from north to south for several species. Fat content exhibited a small negative effect on concentrations of calcium, iron, and mercury, and a small positive effect on phosphorus and arsenic concentrations in many species. Temperature showed a small negative effect on zinc concentration for most species, while the effect of salinity varied among species without an apparent trend. This is the first multi-species and multi-element study to investigate drivers of element concentrations in marine fish at a large spatial scale using a Bayesian approach. The robust model predictability indicates the models' potential to further understand nutrient yield dynamics from fisheries, thereby empowering the implementation of informed strategies against hidden hunger.