
Loading content ...
Speaker: Natalie Kukshtel
[Announcement (PDF)]
Speaker Affiliation: PhD Candidate, MIT-WHOI Joint Program, Woods Hole, MA
Date: Friday, February 20, 2026 at 2:00 p.m., in 3-376 and on Zoom
Abstract: Oceanography is an inherently interdisciplinary field, with many interconnected processes of varying spatiotemporal scales contributing to a dynamic ocean environment. Underwater acoustics is a valuable tool for studying these oceanographic processes, as environmental variability greatly influences acoustic propagation and scattering. The work in this dissertation follows an interdisciplinary approach to explore the multifaceted connections between acoustics and different branches of oceanography. An autonomous underwater vehicle (AUV) was used to study physical, biological, and geological oceanography and their joint acoustic effects as part of the New England Shelf Break Acoustics (NESBA) experiment. The AUV, a modified REMUS 600 equipped with an onboard 2.5-4.5 kHz transducer and towed hydrophone array, was deployed among a network of oceanographic and transceiver moorings. Acoustic signals transmitted and received throughout this network were used to analyze the physical links between environmental variability and acoustic propagation and scattering effects. These contributions further highlight the versatile role of AUVs in advancing ocean acoustic research.
In this work, the AUV was first localized using an acoustics-based multi-channel backpropagation approach, as accurate localization of the vehicle is crucial for contextualizing the AUV data. This process involved back-propagating acoustic wavefronts between the AUV source and mooring hydrophones, as well as signals transmitted from a ship-towed source to the AUV array. Analyses on physical oceanographic uncertainty and mooring tilt were performed to further improve the localization result and understand the influence of environmental uncertainty. Next, the same AUV acoustic dataset was used to explore mid-frequency 3D bathymetric reflection and scattering at a submarine landslide. Computational acoustic modeling, including ray tracing and parabolic equation models, were used to recreate the complex seafloor-interacting acoustic arrival patterns observed in the data, with the final data-model comparison showing evidence of 3D out-of-plane acoustic reflection and scattering. Finally, mid-frequency biological attenuation of a mesopelagic deep scattering layer (DSL) was investigated. A swimbladdered-fish scattering model was used to estimate DSL biological attenuation, which was then applied to a comparison of two acoustic propagation paths traveling through and avoiding the layer, respectively. The final results emphasize the joint influence of biological scattering and physical oceanographic uncertainty on sound propagation.
Mesoscale features and their related submesoscale structures can transport heat, freshwater, and biogeochemical tracers (e.g., phytoplankton, oxygen, and carbon) from the surface to the interior. These structures may grow, decay, and redistribute energy through several dynamical processes. The study examines the evolution of small (~20 km) mesoscale eddies and the associated energetic pathways using a four-dimensional, three-month-long autonomous glider survey in the Balearic Sea, Western Mediterranean. The combined glider fleet covered nearly 15,978 km, accumulated 704 glider days, and completed more than 4,837 dives to depths of 700 m, measuring physical and biogeochemical properties. We investigate how energy is redistributed during mesoscale eddy merging and eddy splitting and how these processes relate to changes in flow divergence, strain, and vertical velocities. Scale-dependent diagnostics, support the presence of enhanced upscale kinetic energy transfer during merging, whereas splitting events exhibit strain-dominated dynamics and diminished cross-scale energy transfer. During eddy merging, vertical velocities reach values of up to 20 m day⁻¹. However, the spatial extent of significant vertical motion decreases, consistent with weakened frontogenesis along the eddy periphery and a redistribution of kinetic energy within the merging eddy. Examination of the baroclinic conversion, quantified through the eddy vertical buoyancy flux ⟨w′b′⟩, suggests higher vertical energy conversion during eddy merging. This is localized within areas where strain and frontogenetic activity occur. Although peak values of vertical energy conversion remain elevated, the fraction of the domain contributing to significant conversion decreases, indicating increasing spatial localization of energetic processes during merging. During eddy splitting, vertical velocities are substantially reduced (less than 10 m day⁻¹) following a frontolytic event in the northern eddy. Eddy splitting reduces positive and negative divergence, as well as eddy kinetic energy. Vertical velocities decrease from 20 m day⁻¹ to 10 m day⁻¹. The related baroclinic conversion declines during splitting, suggesting a decrease in vertical exchange under frontolytic conditions. Analysis of baroclinic energy transfers indicates vertical and horizontal pathways, with vertical conversion having a key role during eddy merging and a decreasing role during eddy splitting. Eddy merging is characterized by enhanced upscale transfer of kinetic energy across mesoscale and submesoscale ranges, whereas eddy splitting shows strain-dominated dynamics and reduced cross-scale energy transfer. The observations indicate that mesoscale eddy merging serves as an efficient mechanism for energy redistribution, while eddy splitting promotes energetic decay and reorganizes the vertical exchange.
We illustrate the use of our Lagrangian flow map analyses to quantify submesoscale transports and non-advective dynamics. We utilize our flow map predictions to extract dynamical regions and coherent structures, classify submesoscale processes, and inform classical analyses. Our emphasis is on the use of spatiotemporal flow maps to help differentiate the advective transports from time-integrated non-advective transformations of water masses and submesoscale features. Results are presented for real-time sea experiments with autonomous sensing platforms and probabilistic modeling systems in diverse ocean regions and dynamical regimes. They include the Nova Scotia Shelf-Slope and New England Seamount Chain regions, the Gulf of Mexico, the Balearic and Alboran Seas, and the Southern California Bight. Our analyses highlight regions of higher shear and mixing, Lagrangian energy and buoyancy dissipation rates, frontogenesis and frontolysis zones, and strong vertical and helical-spiral motions, including filaments and internal waves.
In the winter 2022, a multidisciplinary experiment in the Balearic Sea (northwestern Mediterranean Sea) combined multiplatform in-situ observations with high-resolution numerical simulations to investigate the evolution of a mesoscale oceanic front. The study focuses on analyzing the energy transfer from the mesoscale front to submesoscale cyclonic eddies (SCEs) and understanding their impact on subduction processes from the ocean surface to the interior, using a numerical simulation with 650 m horizontal resolution. The frontal evolution exhibited two distinct phases: (i) an intensification phase driven by strain-induced frontogenesis, and (ii) a subsequent decay phase occurring under conditions favorable to overturning instabilities, triggered by a down-front wind event. These processes enhanced vertical velocities through an ageostrophic secondary circulation across the front, contributing to upper-ocean restratification. Following the wind event, the front decayed and fragmented into smaller-scale structures, leading to the formation of SCEs along its edges. The formation of SCEs was associated with the frontal decay, as well as with centrifugal and gravitational instabilities, which transferred energy from the mesoscale front to the SCEs. These eddies exhibited a three-dimensional helical-spiral recirculation pattern that facilitated the vertical transport of water parcels. Submesoscale eddy-induced frontogenesis drove subduction into the mixed layer, intensified by submesoscale instabilities and guided by downward-sloping isopycnal surfaces at the eddy periphery.