The Journal of Ocean Technology (2019)

Abstract

Harbour porpoise echolocation activity within and outside of the FORCE tidal turbine demonstration area in Minas Passage of the upper Bay of Fundy, Nova Scotia, has been monitored by up to eight bottom-moored (30-80 m depth) C-POD click train detectors across 1,210 days between May 2011 and May 2018. Detections occurred on almost every day (98.9% of days), averaging 86.5% across C-PODs, at a median of seven detection positive minutes per day and for a maximum of 44 minutes per day, noting that porpoise detection probabilities using buoy-mounted C-PODs should be considered minimum estimates. No confirmed dolphin click detections were documented. A statistical (Generalized Additive Model using GEE, or GAMGEE) approach with a logit link function confirmed porpoise presence varied significantly by time of year (peak period mid-June and a lower secondary peak early November), by tidal current speed and tidal height (preference for 0-2.5 m/s ebb tides), by time of day (higher click activity at night), and across the lunar cycle. Seasonal and diel variability may be partly driven by prey accessibility. C-POD monitoring performance (termed Percent Time Lost) varied spatially and temporally. The Percent Time Lost was associated with non-biological sediment movement-related clicks causing the C-PODĚ•s maximum click detection memory buffer to be exceeded, notably during very high (e.g., spring tide) tidal speeds. The effects of these covariates were controlled for within the subsequent GAM-GEE modelling process to assess the effects of the installation and 130 days of operations of a 16 m diameter 2 MW OpenHydro tidal turbine. Importantly, despite the winter turbine deployment period, the C-POD data highlighted that porpoises were not excluded by turbine installation and operations over the mid-field study area, with click detections occurring on 98.5% of days, averaging 82% across C-PODs, with a median of three minutes per day. However, GAM-GEE modelling identified a significant decrease in porpoise click activity when the turbine was operational; this effect varied spatially, with a significant decrease in porpoise activity at the two sites closest (200-230 m) to the operating turbine berth of the five C-POD monitoring sites assessed. C-PODs at two sites that were more than a kilometre away did not show a significant turbine effect, while a third site, at 1,690 m from the turbine and in deeper water to the south, showed an increase in porpoise click detections. A longer time series of turbine deployment is believed required before robust conclusions can be drawn on turbine effects. Nevertheless, it was clear that after turbine removal, porpoise click activity rates returned to pre-installation baseline rates.