Date of Award

Summer 8-31-2024

Document Type

Thesis: Open Access

Degree Name

MS Marine Science

Department

Environmental and Ocean Sciences

Committee Chair

Ann Bowles, Ph.D.

Committee Member

Steven Searcy, Ph.D.

Committee Member

Paula Sylvia, M.S

Abstract

Aquaculture, the farming of water-based organisms, is one of the fastest growing industries in the food production sector globally. Aquaculture farms, regardless of type, are categorized as either coastal or offshore marine aquaculture and are generalized into three categories: fed (e.g., finfish), unfed (e.g., filter-feeding bivalves), and autotrophic (e.g. kelp). Viewed as an agricultural industry, marine aquaculture is tightly regulated, particularly in the U.S. Despite this, an area of concern for the public and management alike is the potential for marine life interactions with aquaculture gear. Even the best-run farms represent a change in the physical and biological characteristics of their local environment. To assess potential effects, regulatory authorities have primarily relied on peer-reviewed, scientific data and publications on marine life interactions; however, to date, research on presence and behavior of marine life immediately outside of and in the near vicinity of marine aquaculture farms is still limited, particularly on interactions with gear over time. A majority of the current literature on close-range monitoring has been done by human by-eye observations. However, there is a need for low-cost, long-term, and often unattended monitoring sufficient to address the potential effects of marine aquafarms that has been highlighted by both agency reviewers and industry organizations. The goal of the present study was to study the feasibility of using affordable, off-the-shelf video monitoring devices to quantify the diversity, local abundance, and behavior of marine life around a net pen used to grow marine finfish to a size suitable for release for fishery enhancement purposes. Focusing on organisms close to the grow-out pens, my methods were designed to observe marine life in two states; in the presence of the farm pen structure with growing fish compared to its absence. As a second point of comparison, I also observed marine life at an adjacent open water control site in both states.

A series of four cameras, two in-air, and two underwater were tested. In air, a CamDo Solar UpBlink Construction Camera (CamDo) and Brinno BCC2000 Timelapse Camera (Brinno) were used; the latter representing a consumer-grade, off-the shelf model. Under water, a SubAqua Underwater Timelapse Camera (SubAqua) and a pair of Insta360 One X2 and X3 model cameras were used. The Insta360 cameras were chosen to test their 360° recording capability, as 360° video has yet to be applied to monitoring marine aquaculture. Usability in the field, the quality of the data that could be collected, and the human time-cost of analyzing footage was compared between both pairs of devices.

Deployments were conducted at the Hubbs-SeaWorld Research Institute (HSWRI) Leon Raymond Hubbard Jr. Marine Fish Hatchery. I observed white seabass (Atractoscion nobilis) grow-out pens located in Agua Hedionda Lagoon and at a control plot in the lagoon ~50m-75m south of the farm. Recordings were made in two states: 1) when both the farmed fish and the netting that contained them were absent from the farm structure, and 2) when farmed fish and the netting were present. This was done to test effects of the combined farm structure and presence of farmed fish on counts and species composition of local organisms; focusing on species likely to be detectable on video (fish, birds, and marine mammals). Organism counts and species richness (quantified using the Shannon-Weiner index) were compared between the test and control sites during both farmed fish presence and absence. An inventory of the species observed and behaviors in the presence of the farm structure were also reported.

The analysis of the in-air Brinno footage was the most efficient. CamDo analysis took 25.6% longer despite having only approximately a third of the total video collected with the Brinno. This was primarily a function of significant programming issues with the CamDo system. The Insta360 cameras’ 360° FOV was approximately five times greater than the 65°-70° FOV of the SubAqua. In addition, the light gathering capacity of the SubAqua, despite being designed for scientific applications, was limited due to a lack of light gathering capacity in the absence of an attached light source and automatic color balance to compensate for underwater light spectrum. As a result, the Brinno and Insta360 cameras were more suitable for future application in monitoring marine aquaculture.

A total of 28 species were observed both in-air (n=7) and underwater (n=21) during deployments. The underwater species included 16 bony fishes, four fishes classed as Chondrichthyes, and one marine mammal. Of the 21 underwater species, 19 were present at the control site. Eleven were observed during fish absence and 13 were observed during fish presence. At the net pens, 12 of the 21 underwater species were observed; eight during fish absence and 12 during fish presence.

Total organism counts varied significantly across both site and state. When farmed fish were present, fish counts were significantly greater at the test site compared to the control in either state. Post-hoc statistics showed that counts made at the net pen during fish presence were significantly greater than those at the control site during either state. However, counts were not significantly different at the test site in the two states. Counts increased and decreased over time when fish were introduced or released from the net pen, which likely affected count means and variability. There was no significant difference in species richness or relative abundance of fish between the test and control sites while the farmed fish were absent. However, species richness differed significantly when farmed fish were present vs. absent, with a significantly lower Shannon-Weiner index when fish were present. The number of species was also lower.

These findings supported the hypothesis that the farmed fish and netting structure presence significantly affected local organism counts and species composition. The effects of farmed fish presence were localized as similar differences were not observed in the control area, 50m-75m from the net pen. My observations also provided evidence that fish occupied the water column near the net pen as would be expected based on their guild.

The data I was able to obtain during this feasibility study show the efficacy of affordable, off-the-shelf video devices for monitoring species presence, counts, and trends at aquaculture facilities. Future research should continue to monitor fish movements and guild interactions associated with farm activities to understand the ecological footprint of farms. I also recommend that future work should expand the methods tested here to improve video recording of behavior. The deployment method I used was designed to maximize counts and species identification, which made behavior observations difficult. However, independent deployments designed to obtain detailed observations of behavior could provide accurate measures of swimming behavior, contacts with structures/gear, interactions among species, and more. With these data, the growing industry could evaluate concerns about harmful interactions quantitatively and create better gear (netting, lines, rope, cages, etc.) that could both withstand animal interaction and minimize any adverse effects.

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a CC BY License.

Download

Available for download on Saturday, November 22, 2025

Share

COinS