Fishermen lowers sea scallops into the water. Credit: NOAA Fisheries.
Milford Lab's GoPro Aquaculture Project
This project uses GoPro camera footage to understand interactions between shellfish aquaculture gear and fish communities.
Overview
Project Goals
To determine:
- Do oyster aquaculture cages provide habitat similar to that of naturally occurring rock reef environments?
- Do cage densities associated with shellfish farming attract fish differently than single cages placed in areas with little natural structure?
- Are different styles of oyster aquaculture cages providing different habitat services to the local fish community?
- What sizes and life stages of fish use oyster cages and boulders as habitat?
Project Description
Shellfish aquaculture can increase food production, create economic opportunities in coastal areas, and enhance natural harvests. To foster successful, responsible, and sustainable aquaculture for the future, we need to understand how aquaculture gear may interact with marine ecosystems. Off-bottom tiered oyster cages are growing in popularity as a method for culturing large numbers of oysters on a small footprint. These cages create complex structures that may attract fish and other animals seeking food sources, shelter, and refuge from water currents or protection from predators. Because of the cages' structure, they may function like naturally occurring rock reefs, providing beneficial habitat to ecologically and economically important fish and invertebrates.
Informing Decision Makers
Understanding if and how oyster cages function like naturally occurring habitats may help with regulatory and permitting processes when siting new shellfish farms. Regulatory challenges have been identified as one of the major bottlenecks to shellfish aquaculture expansion in the United States. This project is a close partnership between the NOAA Northeast Fisheries Science Center and Greater Atlantic Regional Fisheries Office. This partnership ensures that the results from this study will inform the regulatory review process. Our findings may support a more comprehensive review process for shellfish aquaculture projects, which considers environmental benefits as well as impact. Next steps include:
- Identify methods for, and barriers to, considering ecosystem services provided by gear within existing aquaculture management and permitting frameworks.
- Develop practical tools for regulators so they can use our project data for aquaculture management and permitting.
- Conduct an economic valuation of the ecosystem services provided by shellfish aquaculture gear.
![(Left) An oyster cage containing mesh bags full of oysters on the back of a boat. Two scientists are affixing GoPro cameras to the corners of the cage. (Right) A metal oyster cage on a boat deck with three shelves holding six mesh bags. Two GoPro cameras are affixed to corners of the cage.]()
(Left) A stacked tray oyster cage. (Right) A shelf and bag oyster cage.
Comparing Aquaculture Gear to Natural Habitat
Naturally occurring rock reefs create habitat complexity in an often otherwise flat and featureless landscape. Rock reefs add vertical height in the water column, crevices and shading for hiding, and surfaces on which epibenthic organisms can colonize and grow. Complex habitats like these often have greater density and diversity of species than a less complex habitat like a flat seafloor. Oyster cages offer similar features, and may be providing benefits similar to those of rock reef habitats. While few studies have investigated aquaculture gear as habitat, extensive anecdotal evidence from commercial shellfish growers suggests that both fish and invertebrates may be using the gear as habitat.
Comparing Oyster Cultivation Methods
Commercial oyster growers use a variety of cultivation methods, and differences in gear structure may influence how farms interact with the local fish community. We examined two commonly-used styles of oyster bottom cage, to see if structural differences affected fish abundance, community composition, and behavior. Shelf and bag oyster cages have three open shelves with bags of oysters, which creates a variety of spaces for fish to use. Stacked tray cages have three enclosed trays of oysters with smaller-sized openings, which may restrict the size of fish that are able to access the cage interior.
Size Structure of Fish Community around Cages and Boulders
We collected video with action cameras in stereo configuration to measure the body lengths of fish. This will allow us to document the size structure of the fish community associated with cages and boulders.
This research is being funded by the Northeast Fisheries Science Center and NOAA's Office of Aquaculture.
Study Methods
Study Site Locations
Our study sites are all located in Long Island Sound. We studied oyster cages at a high density oyster cage farm, single cages at low density on sand and shell seafloor, and boulders on a rock reef, in partnership with the Charles Island Oyster Farm.
The comparison of cage styles was conducted on three oyster cage farms:
- Charles Island Oyster Farm, Milford, Connecticut
- Copps Island Oyster Farm, Norwalk, Connecticut
- Hummock Island Oyster Farm, Westport, Connecticut
Image
To collect video for fish size measurements, we placed cameras in stereo configuration on cages at three shellfish farms and boulders at three rock reefs in central and eastern Long Island Sound.
- Charles Island Oyster Farms, Milford, Connecticut
- Indian River Shellfish in Clinton, Connecticut
- Noank Cooperative, Noank, Connecticut
Image
![The top left is a map of 3 field sites along the Connecticut shoreline from west to east labelled A, B and C. A (top right) is a map of the Connecticut shoreline showing a cage site and a boulder site in Milford, Connecticut. B (bottom left) is a map showing a cage site and a boulder site in Clinton, Connecticut. C (bottom right) is a map showing a cage site and a boulder site in Noank, Connecticut.]()
Map of 2022 field sites in Milford (A), Clinton (B), and Noank (C), Connecticut.
Methods
We used a series of GoPro cameras to record activity around the structures and document the abundance and behavior of fish around shelf and bag style oyster cages and naturally occurring rock reef habitat. Using cameras as opposed to diver observations reduced the amount of human interference with fish activity. During all camera deployments near cages or boulders, timers were used to record video for 8 minute intervals every hour from 7 a.m. to 7 p.m over a complete 12-hour tidal cycle. This enabled us to sample fish abundance and activity through changing environmental conditions over the course of a single day.
Cameras on Oyster Cages
We outfitted oyster cages with two cameras:
- one mounted to record activity along the horizontal surface across the top of the cage, like a periscope, and
- one mounted to hang down from the corner of the cage to record activity along two sides of the cage and the area where the cage rests on the seafloor.
Cages were brought onboard a boat, cameras were mounted on them, and then the cages were redeployed. Video recording started the following day and we retrieved cameras two days later. The oyster cages themselves stayed in the water for the entire field season.
Cameras on Natural Habitat
Image
To record activity at naturally occurring rock reef habitat, we constructed four stands with dual camera mounts. We then placed the platforms near boulders within the rock reef site where they remained for the entire field season. Divers then put the cameras on the stands and retrieved the cameras two days later. The two cameras provided a field of view similar to that provided by our oyster cage camera system: one looked across the top of the boulder and one looked down the side of the boulder to where it met the seafloor.
Cage Placement at Milford Shellfish Farm
To understand if oyster cage density affects fish abundance and behavior, we placed four cages about 50 meters apart at the inshore site adjacent to the oyster farm, and four cages about 90 meters apart at the inshore mixed-sand/shell (“Shell Bottom” in the map) habitat. We used the same recording approach as used for the cage/reef comparison.
Cage Placement Across Multiple Oyster Farms
To compare fish interactions with different styles of oyster cages, we placed cage pairs at three Connecticut shellfish farms. Two cages of each type were deployed at shellfish farms in Milford, Norwalk, and Westport, to compare fish interactions between cage styles at multiple farms across a broader geographic area. We collected video using timers as previously described.
Seasonality
Generally we deploy cameras during the active oyster growing season in Long Island Sound, from June to September. During the fall and winter of 2019–2020, we also deployed cameras monthly from October until February in Milford, to capture seasonal differences in fish activity around the gear and natural habitat.
Measuring Habitat Quality and Potential for Enhanced Fish Production
To compare the habitat quality provided by shellfish aquaculture gear to that of natural habitat, we collected video in 2022 using GoPro cameras in stereo configuration mounted on cages at shellfish farms and near boulders on natural rock reefs at sites in Clinton, Milford and Noank, CT. We will measure the body length of fish recorded in the videos to determine fish population size structure and to assess life history stages of fish associated with oyster cages and boulders. These data will promote a better understanding of the ecosystem services provided to fish by aquaculture gear and natural structure.
In 2023, we will use fish traps to collect juvenile black sea bass near oyster cages and boulders to measure their body length and weight. Using these methods, we can estimate the physiological condition of fish and compare the relative quality of oyster cages and boulders as juvenile fish habitat.
Collecting Environmental Data
During camera deployments, we also collected environmental data. We used data loggers to record water temperature and light intensity every 15 minutes throughout every deployment. Light is an important factor affecting visibility and video quality and it may influence fish behavior. Because changing tidal currents may also influence how fish interact with cages and boulders within rock reefs, we used tilt current meters to collect information on speed and direction of water currents. We took measurements once every minute throughout each deployment. We attached a current meter and temperature/light data logger to one cage at each field site and to a spare camera stand at the rock reef. Using handheld probes, we collected additional environmental data (such as salinity and dissolved oxygen) during trips to and from each site.Image
Describing Fish Communities using Environmental DNA (eDNA)
When fish and other animals swim, they shed scales, tissue, and waste, leaving traces of DNA in the water. Although this DNA is diluted, it can be extracted and analyzed. Each time we retrieved our GoPro cameras, we collected water for environmental DNA (eDNA) analysis. The eDNA method we used combined DNA-based species identification with high throughput next generation sequencing, or sequencing millions of copies of DNA in the environment simultaneously, to characterize the fish communities associated with oyster cages and rock reefs. Using eDNA we detected fish in the area surrounding the cages and rock boulders, even when they were not captured on the video. This may help create a more complete picture of fish community composition than video alone. Compared to capture-based methods, eDNA analysis is a comprehensive, rapid, non-invasive, and cost-effective method.
Artificial Intelligence
Analysis of underwater video by human coders can be time-consuming. We are often asked if we can use artificial intelligence to automate video analysis. In our case, it is challenging because visibility is often limited in Long Island Sound. While using AI to automate documentation of fish behavior is still a way off, there have been strides in using AI to count fish and identify species in video.
To find out more, we have collaborators who are working on AI programs. If we can use AI to automate the process of identifying and counting fish in video, we will be able to process large quantities of video more quickly and efficiently. Our AI collaborators are:
- Griffith University Coastal and Marine Research Centre / FishID
- MIT Sea Grant machine learning applications to fisheries monitoring project
- Microsoft, NOAA Northwest Fisheries Science Center, and the University of Washington's Fish identification model for analysis of underwater video
Fish Behavior
Observing fish behavior around aquaculture gear and natural structures can help us better understand how they provide habitat. Our team documented the behavior of two temperate-reef species, cunner and scup, in video we collected in 2018 on oyster cages at a shellfish farm and on boulders at a rock reef in Milford, Connecticut.
We observed small-bodied cunner feeding, taking shelter, escaping from predators, and acting territorial on cages and boulders. Scup occurred as single individuals or in large groups or schools, and commonly fed on colonizing plants and animals living on the aquaculture gear and boulders. Our team was excited to see an adult scup spawning adjacent to an oyster cage, right in front of our camera!
By counting the behaviors of individual fish on cages and boulders in the video we collected in 2018, we will be able to compare the habitat services provided by the shellfish farm and the rock reef. Our behavioral observations to date suggest that cages create habitat for fish similar to that provided by natural boulder reefs.
Results
We have collected and analyzed more than 1,600 hours of underwater video for fish abundance and community composition from the 2017–2019 and 2021-2022 field seasons. Because of challenging visibility in the water column and the need for nuanced fish behavior analysis to understand habitat use, technicians perform all video analysis.
To date, we've observed 21 species of fish in our videos associating and interacting with oyster cages. The four most common species observed in video have been:
- Black sea bass
- Cunner
- Scup
- Tautog
Image
Less-commonly observed species in videos include:
- Atlantic silverside
- Banded rudderfish
- Bluefish
- Butterfish
- Conger eel
- Crevalle jack
- Menhaden
- Naked goby
- Red hake
- Smallmouth flounder
- Striped bass
- Striped killifish
- Striped sea robin
- Summer flounder
- Weakfish
- Windowpane flounder
- Yellow jack
We’ve also observed a few fish species that have fallen out of cages during camera deployment/retrieval, but were not yet observed on video, including:
- Oyster toadfish
- Rock gunnel
Identifying fish to species level and counting fish are only the first steps in our video analysis. We are also analyzing fish behavior in the video we have collected.
We've partnered with two fish ecologists, Peter Auster from the University of Connecticut and the Mystic Aquarium, and Christian Conroy from the University of New Haven, to identify and quantify the different ways that fish are using oyster cages. So far we've seen fish feeding on the colonizing organisms that grow on the cages, little fish escaping from bigger fish by darting inside the cage itself, female fish retreating inside the cage to escape male fish of the same species, territorial behavior, courtship, and fish spawning.
Methods and data collected during the 2017 oyster growing season were published in Aquaculture Environment Interactions. Our analysis of eDNA samples collected during summer 2017 was published in Frontiers in Marine Science.
Contact Information
Northeast Fisheries Science Center Project/Principal Investigators:
Greater Atlantic Regional Fisheries Office Principal Investigators and Project Advisors
NOAA Fisheries Project Team:
Paul Clark, Bill DeFrancesco, Mark Dixon, Erick Estela, Zach Gordon, Tyler Houck, Peter Hudson, Kristen Jabanoski, Yuan Liu, Gillian Phillips, Matthew Poach, Jerry Prezioso, Dylan Redman, Ian Robbins, and Barry Smith.
Collaborators and Partners
Academic Partners
- Northeastern University
- Jon Grabowski
- Kelsey Schultz
- Rutgers University
- Daphne Munroe
- Jenny Shinn
- University of Connecticut/Mystic Aquarium
- Peter Auster
- University of New Haven
- Christian Conroy
- Adam Armbruster
NOAA
- NOAA Corps (Calandria DeCastro, Keith Golden)
- NOAA Northwest Fisheries Science Center (Beth Sanderson)
- NOAA Alaska Fisheries Science Center (Bridget Ferriss)
Nonprofit Partners
- The Nature Conservancy (Steve Kirk)
Students and Interns
- Eileen Bates (Bowdoin College)
- Bobbi Bevacqua (Western Washington University)
- Grace Cajski (Yale University)
- Jack Jerrild (Mitchell College)
- Deaven Maull (Washington College)
- Max Mauro (University of Maine)
- Sam Pletcher (Yale University)
- Brandon Rose (Eckerd College)
- Ryan Rubino (Texas A & M University)
State of Connecticut Aquaculture Division
- David Carey
- Kristin DeRosia-Banick
- Shannon Kelly
Shellfish Growers
- Atlantic Clam Farms (John Pinkowski)
- Charles Island Oyster Farm (Charles Viens)
- Copps Island Oysters (Jimmy Bloom, Rachel Precious)
- Fishers Island Oysters
- G & B Shellfish Farm (Gary Salce)
- Hummock Island Oyster Company (Jeff Northrup)
- Indian River Shellfish (Mike Gilman)
- Noank Aquaculture Cooperative (Jim Markow, Steve Plant)
- Robert Granfield
- Stella Mar Oysters (Steve Schafer)
- Stonington Farms Shellfish (Beth and Kris Simonds)
U.S. Coast Guard
- Office of Search and Rescue (Arthur Allen)
Citizen Science
We've created a citizen science guide to help growers capture high quality underwater footage of their aquaculture gear to monitor the ecosystem interactions:Related Projects
Rutgers University received funding from the Northeastern Regional Aquaculture Center to investigate fish and invertebrate interactions with oyster aquaculture gear and natural habitats using GoPro cameras in New Jersey. Their project engages oyster growers in Barnegat Bay that use floating oyster bags and oyster bottom cages. Natural habitats included a marsh edge and bare sandy bottom. Scientists have posted video clips from the project on the lab’s YouTube channel, and you can find more project information at the lab’s website.
The Nature Conservancy and Northeastern University have partnered to investigate how oyster farming gear provides habitat for other marine species using GoPro cameras in Massachusetts. They are working with commercial growers in Duxbury and Cotuit who use shelf-and-bag bottom cages, bottom trays, and OysterGro gear on both intertidal and subtidal farms. Natural habitat and control sites include constructed oyster reef, rocky reef, and bare sediment. The Nature Conservancy produced a video about this local project and how it contributes to the organization’s global aquaculture program.
We have shared our video collected in Milford with partners at the University of New Haven, who are analyzing videos of black sea bass and quantifying territorial and station-keeping behaviors around oyster cages and rock reefs.Learn More
- Eyes Underwater: Complementary Tools Can Determine How Fish Use Oyster Aquaculture Gear
- CERF Video: A Day of Fieldwork for the GoPro Project
- Capturing Life on Shellfish Farms (Story Map)
- NOAA Fisheries Lab Helps Shellfish Growers Become Citizen Scientists
- Tracking Marine Life with Invisible Clues: eDNA Enhances Ecosystem Monitoring
- Oyster Aquaculture Provides Habitat for Fish and Invertebrates - American Fisheries Society
- Window to an Underwater World: Milford scientists evaluate aquaculture oyster cages as habitat using GoPro cameras
- NOAA Scientists Monitoring Connecticut Aquaculture (CT Department of Agriculture)
- Aw Shucks: Scientists Use GoPro to Study Oyster Habitats
- GoPro cameras let scientists know how fish act near oyster cages
- Collaborative research with the Northwest Fisheries Science Center (Outreach Handout)
- Documenting the Ecosystem Services Provided by Oyster Aquaculture Gear (Outreach Handout)
Resources for Kids and Teachers:
Fishermen lowers sea scallops into the water. Credit: NOAA Fisheries.
