Fig 1.
Turbine and monitoring equipment.
(a) The Turbine Lander prepared for deployment. (b) The Adaptable Monitoring Package instrument head with sensors annotated. Photo credit: Abigale Snortland.
Fig 2.
The background outlines central North America with the test site at the inlet to Sequim Bay, Washington (USA) highlighted by the red dot (upper right corner). (a) The bathymetry (MLLW) of Sequim Bay. The box at the constriction at the north end of the bay highlights the inlet, which is shown in (b). (b) The bathymetry of the inlet to Sequim Bay showing the locations of the PNNL Marine and Coastal Research Laboratory (MCRL) dock and the deployment location of the Turbine Lander and the Adaptable Monitoring Package (AMP).
Fig 3.
Timeline summarizing data collection.
The three data subsets that were analyzed are indicated. Periods with a dark border include nighttime data, while periods without a dark border were only collected during daylight hours. Periods of data in white were not analyzed, and asterisks (*) indicate periods where the AMP was offline for up to 5 hours for system maintenance. The dashed portion of the timeline after 11 February indicates the period when the turbine was deployed, but not operating. Note that the timeline is not to scale.
Table 1.
Data summary. Summary of the data contained in each subset, including the total number of hours contained in each sampling window, the total number of hours in the sampling window when the turbine was operating, the total number of hours of recorded data at at least 1 Hz, and the total number of hours of data recorded at a frame rate of at least 1 Hz while the turbine was rotating.
Table 2.
Summary of all individual fish, fish schools, birds, and seals detected in data subsets 1, 2, and 3. Note that some events in data subset 2 are also contained in data subset 3.
Fig 4.
Distribution of seal and bird events.
Distribution of events containing seals and seabirds in each data subset with respect to hour of day (local time), flow velocity, and tidal elevation (colored bars). Gray bars in the background of each histogram indicate the distribution of each environmental covariate throughout all images that were recorded and reviewed. Note that the y-axis scale differs between classes and data subsets.
Fig 5.
Representative timeseries of events.
Number of events associated with each class per hour for a representative week of data. Events from both data subset 2 (continuous acquisition) and data subset 3 (detected events) are shown. Hours between sunset and sunrise are indicated in gray, and the tidal elevation is shown for reference. The darker regions of the tidal elevation time series indicate periods when the turbine was rotating.
Fig 6.
Examples of individual fish when the turbine was rotating.
(a) Small fish drifting, head down, between AMP and rotor. (b) Large, unidentified fish swimming above rotor. (c) Flatfish drifting above moving rotor. (d) Perch-like species drifting with current between AMP and rotor. (e) Fish swimming within moving rotor. (f) Fish swimming away from rotor after collision.
Fig 7.
Example of a collision between a fish and the turbine blades.
The yellow arrows approximate the direction the fish was moving in a given frame, and times are presented relative to the first frame. The inset in the middle frame shows a larger view of the collision seen on the blade on the right-hand side of the image.
Fig 8.
Examples of evasion events while the turbine was operational.
In the images, the white arrows point to the fish and the yellow arrows approximate the direction of travel in optical images. Times are presented relative to the first frame. (Example 1) An evasion event with a single small fish. The fish approaches the rotor before turning and diving down away from the rotor. (Example 2) Two fish evading the turbine rotor using different tactics. Fish 1 swims against the current and moves out of the field of view. Fish 2, swimming with the current, executes a turn and swims away from the rotor. (Example 3) A seal pursuing a fish. The fish swims into and out of the rotor, evading a collision with the moving rotor in the process. As the seal approaches the moving rotor it stops pursuing the fish, slows down, and swims away from the rotor.
Fig 9.
(a–b) A small school of forage fish near the AMP. (b) A fish school near the rotor. Most fish swim up or around the backside of the rotor although several swim between the rotor and the camera. (c) A fish school near the surface while the turbine was stationary. (d) A fish school evading the turbine rotor at night while the turbine was rotating. Red circles highlight the individual fish.
Fig 10.
(a–c) Cormorants diving, swimming to the surface, and presumably foraging in the vicinity of the AMP. (d-f) Pigeon guillemots swimming, picking at the rotor, and interacting with the AMP.
Fig 11.
a) Seal swimming behind the moving rotor at night. b) Seal approaching the moving rotor in the wake during the day. c) A seal diving towards the seabed. d) A seal entering the rotor’s swept area and bending around the shaft. In both (c) and (d) the rotor was stationary.