ORCID

https://orcid.org/0009-0004-5038-4281

Degree

Bachelor of Science (B.S.)

Major(s)

Wildlife, Fisheries and Aquaculture

Document Type

Immediate Campus-Only Restricted Access

Abstract

Seafloor seeps are locations where gas is discharged from marine sediments into the ocean. Methane released from seeps in the deep sea drives a wide range of interconnected biogeochemical processes, resulting in unique chemosynthetic ecosystems with relatively high biodiversity. Due to their remote location, these ecosystems are understudied, and there is a lack of information on benthic species distribution, environmental preferences, and habitat associations. We developed a spatial distribution model for the seep foundation species Bathymodiolus childressi mussels, and Paragorgia spp. corals at the deep-sea Veatch Canyon seep site (Depth: ~1410 m). A presence only maximum entropy (MAXENT) model was used to predict the probability of the presence of mussels and deep-sea corals with acoustic data used as most of the explanatory variables. The predictive model was based upon high-resolution (~1m2) multibeam sonar observations of the seafloor bathymetry and acoustic reflectivity collected with an autonomous underwater vehicle. Validation of the model was conducted with a database of mussel and coral location data generated from mosaiced seafloor imagery collected with the same AUV. The accuracy of the model in predicting species presence was evaluated through ten-fold cross validation. The model predicted the presence of mussels in 1.82% of the study area and had an accuracy of 73.22%. The model predicted the presence of corals in 10.93% of the study area and had an accuracy of 73.97%. Explanatory variables such as distance from seeps, seafloor depth, and acoustic backscatter intensity including its statistical derivatives- like standard deviation, maximum backscatter and minimum backscatter- were strongly related to probability of mussel presence and probability of coral presence. This interdisciplinary approach to species distribution modeling demonstrates the potential value of incorporation of less commonly applied acoustic data sources in deep-sea benthic ecological modeling.

DOI

https://doi.org/10.54718/SQGL2911

Date Defended

4-22-2025

Funding Source

NOAA, NSF, ORED, Shackouls Honors College

Thesis Director

Adam Skarke

Second Committee Member

Garrett Street

Third Committee Member

David Hoffman

Download

Share

COinS