Abstract
Lab Collaboration Project (LCP) for Marine Energy: Quantifying Collision Risk for Fish and Turbines Final Technical Report (Task 10)
A persistent environmental concern for the widespread deployment of tidal turbines is the potential for fish and marine mammals to collide with rotating blades (Copping et al. 2016, Copping and Hemery 2020). This is a consequence of well-documented bird and bat mortalities around wind turbines (Smallwood 2007, Thompson et al. 2017), as well as fish mortality at conventional hydropower dams (Pracheil et al. 2016) and tidal barrages (Dadswell and Rulifson 1994). However, unlike hydropower dams or barrages, tidal turbines do not involve structures that channel all flow through the turbines. Similarly, while functionally similar to wind turbines, tidal turbines often operate at lower relative velocities and, depending on the end-use application, may be significantly smaller than utility-scale wind turbines. Both of these factors reduce the likelihood and severity of collision, but the knowledge base on this topic remains limited.
The objective of this task was to add to that knowledge base in four areas:
- To collect data on fish interactions with an operating tidal turbine;
- To contextualize these interactions with the changes the turbine makes to the physical environment – specifically, the proximate flow disturbance and radiated noise that are a consequence of energy harvesting;
- To interpret the behavior of fish interacting with the turbine and how this could increase or mitigate collision risk; and
- To employ models for these interactions that could be used in a predictive manner at other locations.
The project was initiated in 2017 as a collaboration between Pacific Northwest National Laboratory (PNNL) and PMEC researchers at the University of Washington (UW) and University of Alaska Fairbanks (UAF). Fish interactions were to be monitored around a pair of cross-flow turbines with a rated power output of 1 kW that were to be deployed in Sequim Bay, WA at PNNL’s Marine & Coastal Research Laboratory (MCRL). Because of uncertainties about the ability of optical or active acoustic sensors to detect and track individual fish targets in close proximity to the turbine rotor (Cotter and Polagye 2020), PNNL planned to implant JSATS tags (McMichael et al. 2010) in a representative fish species that would allow individual fish to be tracked with high precision.
The project deviated from this initial plan in several ways.
First, during the initial phase of the project, PNNL tagged 100 juvenile sablefish and released them upstream of an Adaptable Monitoring Package (AMP, Polagye et al. 2020) deployed at the intended turbine location at MCRL. However, only one of these fish definitively entered the field of view for the AMP’s imaging sonars and none entered the camera field of view. Because of this, it was determined that an infeasibly large number of fish would need to be tagged for a reasonable sample size of released fish to interact with the turbine rotor. Consequently, this activity was removed from the project scope.
Second, the turbine deployment was substantially delayed relative to the initial timeline for multiple reasons. The turbine was a prototype cross-flow device on a gravity lander being developed by UW with parallel support from the Department of Defense’s Naval Facilities and Engineering Systems Command (NAVFAC). The engineering development of this system took substantially longer than anticipated, such that the first system test with a fully submersible power take-off unit did not occur until summer 2020, two years after the initial deployment target at MCRL. In addition, NEPA and permitting processes took substantially longer than anticipated. Key issues involved a delay of nearly two years for one resource agency’s Endangered Species Act consultation and protracted legal discussions over relatively minor points between UW, PNNL, and the Department of Natural Resources on the terms of the seabed lease for turbine deployment. Because of this, a decision was made in early 2022 to conduct the data collection aspect of the project around an endurance test for the turbine in Agate Pass, WA with the turbine deployed from a moored vessel.
Third, because of the project delays, UAF was unable to identify staff or students to conduct the behavioral evaluation of data from the Agate Pass deployment. Consequently, this portion of the scope was transferred to researchers at Oregon State University (OSU).
Despite these challenges and changes, the task was successful in meeting many of its objectives. Important outcomes included:
- Demonstrated automated trajectory tracking of small fish targets using stereo-optical cameras and machine learning;
- An increased understanding of the capabilities and limitations of optical and acoustic systems for tracking fish around tidal turbines;
- A novel method for hypothesizing potential radiated noise from a deployed turbine and successful employment of acoustic localization of radiated noise in a tidal channel;
- An understanding of the extent and magnitude of flow field disturbances around cross-flow turbines;
- A framework for quantifying collision risk using physical attributes of marine animal motion relative to a turbine; and
- Statistical and agent-based evaluations of encounter and collision risk, assessing sensitivity to key parameters.
In addition, several of the approaches developed under this project were employed during the eventual turbine deployment at MCRL.
This report is broken down into five sections, each of which describes a functional subtask:
- Task 10.1: Fish Interaction with a Turbine: Field data collection of fish trajectories, including the baseline tracking of fish implanted with JSATS tags, cooperative target testing during turbine shakedown tests, and development of automatic detection and tracking capabilities for optical data streams
- Task 10.1: Acoustic Characterization: Acoustic measurements around the turbine deployment in Agate Pass, contextualized by close range measurements of the turbine being motored in a dockside setting
- Task 10.2: Velocity Field Characterization: Hydrodynamic disturbances around a laboratory-scale model of the field turbine, measured using Particle Image Velocimetry
- Task 10.3: Behavioral Evaluation: Development and preliminary application of a model framework that can assign a collision risk based on quantitative metrics derived from
- Task 10.4: Collision and Encounter Risk Modeling: Development and application of statistical and agent-based simulation to predict the likelihood of collision risk