Modelling and predicting ecosystem exposure to in-feed drugs discharged from marine fish farm operations : an Initial perspective

Abstract

This document focuses on modelling in relation to the discharge of active ingredients associated with in-feed drugs used in marine net-pen aquaculture farming operations in Canada. The document includes an overview of the context and associated conceptual processes to be modelled, specific modelling challenges, a review of modelling efforts to date, and a description of some simple models for potential use. In general, modelling for in-feed drug discharges and depositions is in an early stage of development. Few models have been developed specifically to predict the benthic deposition of in-feed drugs and these range in complexity. Model results are sensitive to input parameters, including treatment details, hydrographic conditions, drug partitioning specifics, and sinking rates and timing of discharges. Many of these parameters are poorly understood, difficult to measure, and hence, not well quantified. Thus, determining the quantification of uncertainties and sensitivities of model results remain challenging. Many models for predicting the deposition of organic waste produced by net-pen fish farms have been developed. Although similarities exist between the underlying assumptions of these models and those for in-feed drugs, their adaptation to use for modelling the deposition of in feed drugs is not necessarily simple or straightforward. Of particular importance is the inclusion of drug partitioning specifics which is necessary in order to correctly model in-feed drug deposition dynamics; simple conversion factors between organic waste deposition and in-feed drug deposition are likely not a suitable approach as the ratio between carbon and drug in the released feces varies with time. The objectives of modelling must be specified before a model is selected and assessed for its adequacy and sufficiency. Once models have been selected and/or developed, models must be validated before being used. In general, existing models of in-feed drug deposition have not been extensively calibrated or validated. Of the few validations that have been done, the literature suggests that, regardless of complexity, existing models give, at best, an order of magnitude estimate of seabed drug concentrations. Despite the uncertainties surrounding model precision and validity, models can be useful for regulatory decision support. Model selection depends on the decision maker’s objectives. Precision and accuracy of the model cannot be estimated until the chosen model is verified and validated. Existing validation studies indicate that available models are only able to provide order of magnitude estimates of in-feed drug depositions. At this time, simple models may be sufficient for decision support. This sentiment may change as science better characterizes model inputs and processes, and conducts more validation studies.