Tidal and riverine current energy show potential for supplying clean power to the grid, alternative methods for charging equipment or vessels at sea, and providing locally generated electricity to coastal and remote communities, with estimates of up to 7.8% of the electricity needs of the United States (see report here). That’s enough electricity to power approximately 30 million homes for a year! However, the overlap between tidal and riverine resources and habitats for fish, diving seabirds, and marine mammals create concern for regulators about the risk of collision between these animals and moving parts of the devices used to harness energy.
Collison risk is defined as the possibility of animals physically interacting with moving parts of a marine energy device. This environmental stressor applies to rotating parts of energy converters placed in areas with moderate to strong underwater currents, like tidal or riverine turbines, and dynamic devices like tidal kites that are often deployed in constricted areas. There is less concern of collision risk with wave energy technologies, because they have fewer submerged moving parts with collision potential and tend to be installed in more open-water environments. Encounters between a device and a marine animal could result from attraction to a structure and getting too close, or water currents causing an animal to be swept toward a device. Outcomes could be a near-miss event, a collision with or without injury, or a high-consequence contact leading to significant injury or mortality, which can be significant for endangered or threatened species.
Interactions depend on several variables including the device properties (e.g., rotor speed and turbine design), animal behavior, and animal densities at the location of the moving parts. Currently, collision risk remains a concern for marine energy permitting and regulatory decision makers and communities alike. As empirical research continues to be published, this stressor will become more informed allowing risk status determinations (such as low, medium, or high risk) to be made.
The U.S. Department of Energy Water Power Technologies Office (WPTO) recognizes the need for empirical research surrounding collision risk and the development of mitigation strategies to address it. The WPTO Triton Initiative works to reduce barriers to permitting of marine energy testing and installation through environmental monitoring research that can help inform decision-makers on potential environmental stressors associated with these systems. One of the chief concerns for regulators involves uncertainty around collision risk. Triton’s diverse research includes testing and developing technologies for monitoring collision risk, modeling encounter rates of animals at turbine sites, and innovating practical mitigation strategies for the industry.
Current collision risk research and understanding
Field and laboratory studies that have examined the presence and behavior of fish and other marine animals around marine energy devices indicate collisions are rare (Copping et al., 2021; Sparling et al., 2020). Avoidance describes behavior of an animal actively responding to and moving away from a turbine at a distance greater than 5 turbine diameters. Evasion explains when an animal changes its behavior to escape contact with a device within this 5-turbine diameter distance. An encounter is when an animal is in close proximity to a turbine (1–5 turbine diameters), while collision occurs when an animal comes in contact with the moving parts—typically a rotating blade—of the device. Interaction events of animals with turbines are difficult to observe because of the technical and practical challenges associated with in situ data collection. Future understanding and expectations of collision risk will rely on modeling and advancement of new technologies and field campaign methods.
Varying levels of fish and marine mammal avoidance have been observed from ongoing environmental monitoring at current marine energy sites and demonstrations. Studies assessing seal behavior around tidal turbines suggest that a proportion of seals encountering tidal turbines will exhibit behavioral responses resulting in avoidance of physical injury. In practice, these empirical data can inform collision risk models to predict the effects of tidal turbines on seals (see Hastie et al., 2018). With the exception of a few examples, U.S. studies in the collision risk space have predominantly focused on fish interactions with tidal and riverine turbines examining behavior, encounter rates, and the consequences of collision events.
To date, there have been no published instances of marine mammal or diving seabird collisions with turbines (Sparling et al., 2020), and few studies have observed fish collisions with operational turbines (e.g., Courtney et al. 2022). Still, knowledge around collision risk probabilities and consequences in diverse scenarios remains largely lacking but is vital for marine energy permitting and deployment. Development and refinement of research methodologies to improve understanding of spatial use in marine habitats around operating devices is necessary to improve our ability to observe the interactions of marine mammals, fish, and diving seabirds.
Read the Ocean Energy Systems (OES)-Environmental 2020 State of the Science chapter on collision risk here: https://tethys.pnnl.gov/publications/state-of-the-science-2020-chapter-3-collision
Challenges & solutions
Collision or near-miss events are hard to observe, presenting a unique set of challenges for researchers. Furthermore, observing animal movement and behavior in relation to an underwater object is difficult to study, and it is therefore challenging to collect observational data. This is largely due to the fact that studies must be conducted in dynamic and often low-visibility oceanic and riverine environments that devices are most suited to harness energy from. Underwater video cameras are not ideal for the job in murky, low-lit waters, so active acoustic technologies like imaging sonars—which send out very high-frequency sounds to visualize the scene in a given location—are being employed for these applications (read a recent review on these tools for studying collision risk). To observe a collision or near-miss event also requires constant monitoring, leading to large quantities of data in order to capture the moments during which an interaction could occur. Improvements in the way we collect, store, and analyze data are being pursued by researchers at Pacific Northwest National Laboratory and other research institutions to add efficiency and identify meaningful observations from the data collected during field and lab studies. Recent efforts are tackling the challenges associated with data collection and processing, including improving data analysis software and reducing time and costs without compromising accuracy.
Additionally, promising advancements are being made in the quantitative modeling space. Models take existing data and various known variables about a system and use computational tools to spit out possible scenarios based on the conditions and data inputs provided. Models include different parameters and can be repeatedly run to estimate possible scenarios and outcomes. Models have inherent uncertainty associated with them, but the more data inputs and variety of data a model has to work with, the better the outputs and lower the uncertainty. What these models can do is help better understand the likelihood of collision under different scenarios based on questions like: At what time of day is fish traffic the greatest at a marine energy site? When do different species of fish migrate? And at what depth in the water column are fish most likely to aggregate?
Despite the need for additional in situ field data, efforts to develop encounter-rate and collision risk models based on some of these questions are being leveraged to inform collision risk of fishes and marine mammals. Encounter-rate models estimate the chances that marine animals will enter the immediate vicinity (i.e., swept volume) of a turbine based on their spatial distribution, swimming behaviors, and ability to detect and avoid turbines at farther distances. Collision risk modeling can be used to understand, predict, and assess potential rates of fishes being struck by the moving parts of a turbine. In general, these models estimate the potential for collision using physical specifications of a turbine and characteristics of fishes and marine mammals such as body size, abundance, swimming direction, and ability to evade the turbine at close distances. To accomplish this, the models quantify how often the turbine parts and animals would be in the same place at the same time. The possibility of these meet-ups depends on the turbine size, shape, orientation, and rotation speed as well as where it’s deployed and how it moves. Further data collection is needed to bridge the gaps that still exist in modeling—those gaps primarily being quantitative inputs based on real observations. Specifically, more field-gathered data are needed on fish and seabird behavior, detection, and the outcomes of collisions—whether that be disorientation, injury, or no effect.
These models are constantly improved as more field studies are conducted and empirical data are gathered, making their predictions more accurate and reliable over time but nonetheless dependent on increased in-water studies.
Triton’s role in collision risk research
There is a large research portfolio in the United States committed to improving understanding of collision risk and environmental effects of marine energy. Triton dedicates several research efforts to advancing knowledge of collision risk through evaluating monitoring technologies, improving predictive models, and researching tangible mitigation solutions.
Triton Field Trials
The Triton Field Trials (TFiT) was a four-year effort that explored methods and instrumentation used to monitor the main environmental stressors related to marine energy devices—including collision risk. Through this project, Triton researchers performed field tests of various technologies at diverse marine energy deployment sites around the United States to inform recommendations on methods and instrumentation for monitoring collision risk. The TFiT collision risk team tested the most promising and cost-effective tools for observing fish behavior around turbines during field campaigns in New Hampshire and Alaska.
One of the most powerful tools for monitoring marine animal behavior in the presence of current energy converters is acoustic cameras. This is because they can identify and gather data on fish behavior under water despite the challenges caused by low visibility. The TFiT tests provided evidence that acoustic cameras are instruments fit for observing fish interactions with turbines but have limitations when fish are small and fast-moving. Through this campaign, the team also verified a need for more observational data, particularly about fish behavior near devices. The team published their findings in the Journal of Marine Science and Engineering in an article entitled “Capabilities of an Acoustic Camera to Inform Fish Collision Risk with Current Energy Converter Turbines.”
Probability of Encounter Model (PoEM)
Triton researchers are filling gaps in collision risk modeling through the Probability of Encounter Model—or PoEM—project. Encounter-rate estimates help inform the likelihood that an animal would approach a specific habitat where a marine energy device could be installed. This information helps determine what monitoring might be needed at current energy converter installations to estimate the risk they pose to marine animals and the environment. In collaboration with partners at Aquacoustics, LLC, the project team developed a prototype PoEM based on sonar data of salmon smolts migrating past a riverine turbine in Alaska. In the first phase of development, the model estimated the proportion of salmon smolt that would enter the volume of water directly upstream of the turbine, indicating potential interaction. Further assessment looked at light versus dark conditions and when the turbine was operating. In the next phase of the project, the team will work to identify and address spatial uncertainties of migrating salmon smolt and improve data collection methods with collaborator University of New Hampshire.
This data-driven development process is foundational to refining models informed by in situ observations that can then be applied to monitoring and mitigation efforts in the future. Improved understanding of these environmental concerns helps reduce barriers to installing and testing turbines and supports strategies to minimize fish collision risk for the marine energy industry.
Collision Risk Data Collection and Processing
Triton’s Collision Risk Data Collection and Processing research aims to reduce data accumulation during field deployments that observe animal interactions with turbines. Most data collection efforts use existing commercial technologies, like sonar, video cameras, and acoustic cameras, to observe animal interactions with devices in motion. Data collection often occurs over days and weeks, and much of it consists of footage when no animals are near a turbine. Removing or archiving these data is necessary to reduce time and cost associated with expensive human review efforts. The collision risk data collection research leverages commercially available sensors and created a way to only capture data of interest—like footage of when an animal is near a turbine. For example, the research explores how software that uses signal triggers from upstream echosounders to store downstream video or acoustic camera data when animals are present, can minimize data storage needs while reducing labor for analyses (read the report here).
The collision risk data-processing research has similar goals but aims to improve the processing efficiency of the large volumes of gathered data. The task aims to reduce acoustic camera datasets into subsets of interest (like when an animal is present) so analyses can focus on the most valuable and relevant data. This reduction is done after data collection using data-processing algorithms that remove times when animals are absent and archives those when animals are present, allowing developers and researchers to test and monitor their devices more effectively. This effort attacks the large data volume challenge by reducing data collected in the field with other sensor triggers or reducing data that is reviewed by removing non-event times with processing algorithms.
Integrated Collision Detection and Mitigation
While significant efforts are going toward better monitoring, understanding, and modeling of collision risk, mitigation solutions are an important piece of the puzzle when it comes to improving the safety of animals around marine energy devices. The Integrated Collision Detection and Mitigation (ICDM) project aims to see if strain gauge sensors integrated into turbines can be utilized to slow or stop blade movement when a marine mammal is detected in its vicinity. In partnership with the University of Washington, experiments to inform the study are being conducted at the university’s Alice C. Tyler flume—a laboratory instrument that simulates a tidal channel in a controlled environment. During flume studies, small marine mammal models of seals and whales are placed in the path of a small-scale cross-flow turbine to simulate collisions. These data are then used to better understand collision events and parse apart strain indicators. The goal is that these simulated lab studies will help build knowledge around collisions and the hydrodynamics of these events to inform tangible mitigation strategies that could reduce or mitigate the impact of collisions entirely.
ICDM is an innovative proof-of-concept laboratory study that will build knowledge around collision risk without putting any marine animal in harm’s way. This project is an example of fundamental science that has the potential to transform into applied science in the future, hopefully one day informing strategies to be used by developers in the industry to mitigate the possibility of injurious collisions.
Learn More:
Triton Collision Risk Projects
- The Probability of Encounter Model (PoEM)
- Collision Risk Data Collection and Data Processing
- Integrated Collision Detection and Mitigation
- Triton Field Trials – Collision Risk
Recent peer-reviewed publications
- Capabilities of an Acoustic Camera to Inform Fish Collision Risk with Current Energy Converter Turbines
- Observing Fish Interactions with Marine Energy Turbines using Acoustic Cameras
Other Resources
- Triton Talks Webinar: Collision Risk with Garrett Staines
- OES-Environmental Collision Risk Evidence Base
