Date of Award
8-2026
Degree Type
Masters Thesis
Degree Name
Master of Science (MS)
School
Ocean Science and Engineering
Committee Chair
Dr. Christopher Hayes
Committee Chair School
Ocean Science and Engineering
Committee Member 2
Dr. Chelsea Pederson
Committee Member 2 School
Ocean Science and Engineering
Committee Member 3
Dr. Taylor Lee
Abstract
Seafloor sediment characterization is essential for understanding sediment physical and geoacoustic behavior, yet direct measurements of bulk density and P-wave velocity remain spatially limited due to the logistical demands of sediment coring and laboratory analysis. This study evaluates the potential for X-ray fluorescence (XRF)-derived elemental composition to serve as a quantitative proxy for predicting sediment physical and geoacoustic properties. The study location is the New England Shelf Break Area (NESBA), including the New England Mud Patch (NEMP). Sediment cores were analyzed using both bulk (including sediment, water, and organics) and ash (sediment only) samples to assess how sample preparation influences predictive performance of XRF data. Bulk density and P-wave velocity measurements obtained from multi-sensor core logging (MSCL) were compared with XRF-derived elemental concentrations using multivariate linear regression models to determine predictive skill. In addition, reduced-model approaches were evaluated to determine whether a smaller subset of elemental predictors could maintain predictive performance while reducing model complexity.
Results demonstrate strong relationships between elemental composition and both bulk density and P-wave velocity, with dry XRF consistently producing stronger and more coherent predictive relationships than wet XRF. Water content exhibited strong relationships with bulk density with useful, but weaker relationships with P-wave velocity, indicating that acoustic behavior is influenced by sediment composition and structure in addition to porosity. Reduced-model analyses further demonstrated that much of the predictive information contained within the full XRF datasets could be retained using substantially fewer predictors. In many cases, reduced Top 7 models matched or exceeded the performance of the full models, while a common six-element predictor set (LE, Mg, Al, Si, Ca, and Fe) maintained strong predictive capability across response variables, sample preparation methods, and validation approaches.
These findings demonstrate that XRF-derived elemental composition provides a viable framework for predicting sediment physical and geoacoustic properties in fine-grained continental shelf sediments. The results highlight the importance of sample preparation, identify key sedimentological controls governing bulk density and P-wave velocity, and demonstrate that simplified predictor frameworks can effectively capture much of the predictive information contained within larger geochemical datasets. This approach has the potential to expand geoacoustic datasets using archived sediment cores and improve seafloor characterization in regions where direct measurements are unavailable.
Copyright
Karina D. Ledezma, 2026
Recommended Citation
Ledezma, Karina D., "Predicting Bulk Density and P-wave Velocity Using XRF" (2026). Master's Theses. 1205.
https://aquila.usm.edu/masters_theses/1205
COinS