Sample design considerations for a post-deposit monitoring program for pesticides and drugs discharged from salmon net-pen farming operations

This document is part of a CSAS process in support of the development of a post-deposit monitoring program for drug and pesticide use at Canadian marine finfish farms. This report focusses on designing sampling programs in relation to the discharge of chemicals associated with bath pesticides and in-feed drugs used in marine net-pen aquaculture farming operations in Canada. Selection of a post-deposit sample design should follow a structured and systematic approach that includes clearly stated objectives, decision rules, decision tolerances, sampling constraints, spatial and temporal coordinates of potential sampling locations, and sampling methodologies. Sample designs that are based on probabilistic (statistical) principles are preferred to judgement-based designs. For in-feed medication deposits, bottom sampling should include several phases. The first phase is to identify and map the boundaries of bottom types in the area of interest. The purpose of the second phase is to detect the location and intensity of discharge deposits using a sampling design and sampling methodologies that are appropriate given the knowledge gained in phase one. The purpose of the third phase, if needed, is to refine the characterization of the detected deposits and to monitor temporal change in the characteristics of the deposition (area, concentration). Phase one designs should be grid based, phase two designs should be stratified random designs with random grid or random sample allocation within strata, and phase three designs should be finer scale random grids or random sample allocations within focused areas of interest. This approach recognises that there are inherent uncertainties in discharge properties (location, time, duration, intensity, frequency) and estimates of discharge transport, dispersal, deposition and redistribution; it also helps minimize bias introduced by judgement. The probabilistic approaches enable statistical inferences to be made and trade-offs between precision of sample statistics and cost effectiveness to be evaluated in relation to tolerance criteria. This is particularly important in aquaculture post-deposit monitoring when practicalities generally limit sampling efforts to relatively low sample sizes which results in a significant risk of underestimating the area and intensity of deposits and low precision in estimates of in-situ discharge concentrations. General suggestions for bottom sampling methodologies include the use of bottom sediment samplers with low bottom and sample disturbance characteristics (corers preferred) and visual imagery for hard bottoms. For pesticides, samples of the bath water should be taken just prior to discharge. Due to the constantly changing nature of the pesticide discharge cloud, use of a probabilistic sampling design is not practical. General suggestions for sampling methodologies include the use of a visible tracer introduced into the pesticide bath water prior to discharge. Routine monitoring of bath pesticide post-discharges is probably not feasible; however, targeted monitoring should be occasionally undertaken to help improve models. A minimum sampling effort could involve water samples being taken over time at horizontal and vertical locations that are indicated by the tracer to be areas of high pesticide concentration. Imagery coupled with additional in-situ sampling of tracer and pesticide concentrations can be used to produce calibrated estimates of discharge areas. When the tracer indicates contact with the seabed, a focused random gridded sampling effort for chemical concentration or impact could be undertaken.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.