Research Document - 2012/078
Predicting organic enrichment under marine finfish farms in southwestern New Brunswick, Bay of Fundy: Comparisons of model predictions with results from spatially-intensive sediment sulfide sampling
By B.D. Chang, F.H. Page, R.J. Losier, and E.P. McCurdy
The DEPOMOD (version 2) model was used to predict the rate of organic enrichment of sediments under five operating salmon farms in the southwestern New Brunswick (SWNB) portion of the Bay of Fundy. The model predictions were compared to observed sediment sulfide concentrations at the same farms. DEPOMOD incorporates current speed and direction, feed rates per cage, and bathymetry to predict the spatial distribution of carbon deposition rates on the seafloor under fish farms. Using DEPOMOD, higher current speeds resulted in predictions of larger areas of moderate impact, but smaller areas of highest impact and lower maximum deposition rates. Except where current speeds were low, DEPOMOD’s resuspension module appeared to overestimate the amount of particle transport caused by resuspension, and, therefore, underestimated the spatial extent and intensity of seafloor impacts; this suggests that the threshold current speed for resuspension (approximately 9.5 cm s-1) may be inappropriate for SWNB. Comparisons among sites between DEPOMOD predictions (with resuspension off) of the extent of seafloor area with elevated impacts and observed seafloor areas with elevated sulfide concentrations produced variable results. Comparisons among sites between DEPOMOD predictions of the maximum carbon deposition rates (with resuspension off) and the maximum observed sediment sulfide concentrations showed no relationship. The relationships between sediment sulfide concentration at individual sample locations and DEPOMOD predicted carbon deposition rates (with resuspension off) at the same locations showed wide variability at all study sites. However, where DEPOMOD predicted low deposition rates, sulfide concentrations were usually low, but where DEPOMOD predicted high deposition rates, sulfide concentrations ranged from low to high. A simple model, based on an average feed pellet sinking rate, median mid-depth current speeds, and the average site depth, was also tested to predict the spatial extent of impacted seafloor (but not the intensity or precise spatial distribution of impacts). The spatial extent of impacted seafloor predicted by the simple model showed relatively good agreement with the spatial extent of impacted seafloor predicted by DEPOMOD (with resuspension off). Possible sources of uncertainty in the DEPOMOD predictions are discussed.
- Date modified: