Melanie M. Carolan
Master's Thesis Defense
Oceanography
Thursday, July 31, 2025, 1:00pm
Chase 130
Abstract
The difficulty of sampling shallow, complex, and hard-bottomed coastal and marine environments has led to fisheries monitoring gaps in the nearshore despite their recognized status as Essential Fish Habitat and their vulnerability to habitat modification and degradation. To address this, MaxN densities of commercially and ecologically important marine species were collected using Remote Underwater Video Systems (RUVS). In chapter one, video observations were combined with bathymetric data, seafloor terrain, substrate type, and biogenic habitat variables. Observed guilds included American lobsters, Atlantic pollock, cunner, seals, and several guilds of flatfish and crabs. In the prediction of guild densities, Random Forest models, which incorporate a suite of predictor variables, outperformed the simpler IPDW models that only accounted for local variability in guild densities. Habitat features emerged as the number one predictor for 45% of guilds, and geographic position and seafloor characteristics each served as key predictors for 27% of guilds. The percentage cover of kelp bed observed in the video deployments emerged as the single most influential predictor for cunner densities, which is the most common middle trophic level forage fish in the nearshore Gulf of Maine. Analysis techniques that quantified guild overlaps, such as percentage overlap of spatial distributions and Spearman rank correlation coefficients calculated from predicted mean densities, were found to serve as powerful spatial complements to traditional multivariate cluster and indicator species analyses. Favorable overlaps between predator and prey species were demonstrated between cunner and crabs, cunner and seabirds, crabs and lobster, and pollock and seals. The identification of spatial and temporal overlaps between seals and pollock also highlighted bycatch risks between a commercial fish species and marine mammals known to be historic victims of bycatch. Overall, this study summarizes current fine-scale distributions of guilds through the creation of several interpolation and prediction maps, illuminates guild-specific vulnerabilities to terrain modification and habitat degradation through the discussion of the relative importance of terrain and habitat variables for guild densities, and identifies bycatch risks as well as overlaps in the distributions of predator and prey species through the exploration of guild co-occurrences. These findings are relevant to fishery conservation and management research focused upon the density, interdependence, and life history characteristics of nearshore commercially and ecologically important species.
Bio
Melanie Carolan obtained her Bachelor of Arts in Earth Science with a minor in Mathematics at Vassar College. She served as the Fisheries Science Undergraduate Researcher at Shoals Marine Laboratory during the summer of 2022, and then as the Education and Outreach Intern at the Loon Preservation Committee during the summer of 2023. These experiences helped shape her interest in oceanography and aquatic science in general. Melanie is particularly interested in the interactions between geomorphology and the distributions of marine species. She is also interested in science education, data visualization, natural disasters, and humanity's use of natural resources.
