Current Research
Open-Source, Autonomous, Network-Ready Ocean Sensing Systems
Coastal stewardship and blue economy activities have heavily relied on person-intensive methods to quantify their effectiveness, which are difficult to scale across space and time. With the ongoing “Maker” Revolution, building ocean instruments is rapidly becoming cheaper and more accessible. Our focus is the development of open-source, non-invasive, and network-ready monitoring technologies to evaluate marine ecosystem health and the effectiveness of marine protected areas (MPAs).
We are currently working on several projects: (1) a drifting autosampler to measure net ecosystem metabolism, (2) an unmanned surface vehcile equipped with a profiling environmental sensor package and an autosampler as a rapid response toolkit for coastal communities to measure and map water quality in real-time after disasters, (3) a modular, autonomous water sampler to study coastal biogeochemistry, (4) a distrubted salinity array to detect freshwater runoff after estuary restoration, and (5) a multi-sensory platform that combines a camera, hydrophone and temperature sensor to monitor coastal biodiversity.
Previous open-source projects have included: (1) an optical imaging system to capture images in low-light aquatic habitats using only ambient light, (2) a low-cost (< $1,000) underwater passive acoustic recorder that can be calibrated, and (3) a rapid method to map the three-dimensional structure of kelp forests using an off-the-shelf “fish finder”.
Publications
Butler and Pagniello (2022). Emerging, Low-Cost Ocean Observing Technologies to Democratize Access to the Ocean. Pagniello et al. (2021). An Optical Imaging System for Capturing Images in Low-Light Aquatic Habitats Using Only Ambient Light.Statistical Signal Processing for Underwater Acoustics
Coastal acoustics are challenging due to complex propagation channels, highly variable environmental factors and high ambient noise levels, but sound provides a powerful, non-invasive window into biodiversity and human activity. Our focus is on statistical signal processing methods that exploit the spatial/temporal coherence of propagating waves to improve array performance for detection, classification, and localization of transient sounds in shallow, noisy, multipath environments.
We are currently working on two projects: (1) a generalized power-law detector to automatically find transient, low-frequency sounds in a wide range of ocean noise conditions, and (2) investigating the hypothesis that kelp forests act as an acoustic refuge for sound-producing fish.
Previous work has demonstrated the ability to conduct coastal surveys of sound-producing fish spawning aggregations with an autonomous surface vehicle and tracked individual fish within these aggregations near the border of an MPA using a novel localization algorithm that improves precision and accuracy to a few meters.
Publications
Butler, Pagniello et al. (2021). Diel and seasonal variability in kelp forest soundscapes off the Southern California coast. Pagniello et al. (2019). Mapping Fish Chorus Distributions in Southern California Using an Autonomous Wave Glider.Data Science and Machine Learning for Marine Ecology
Animal-borne electronic tags, passive acoustic recorders, and optical imaging systems now collect high-resolution observations of animal presence, abundance and behaviour across scales. However, integrating heterogeneous biological, physical, and environmental data remains a computational challenge. Our focus is on building scalable data science and machine learning workflows to detect and classify animals to a species level, extract behaviour through tracking and temporal modeling, and integrate animal-borne measurements into climate- and ecosystem-relevant datasets.
We are currently working on two projects: (1) an open and reproducible pipeline to extract ocean properties from animal-borne electronic tags and integrate them into global ocean observational datasets, and (2) a new type of profile classification model to identify water masses from temperature-depth profiles collected by electronically-tagged animals.
Previous work has developed algorithms to dientify individual dives from time series to shown that that Atlantic bluefin tuna exhibit spatially-varying dynamic vertical diving behaviours, with deep diving co-occurring with long-lived, quasi-stationary anticyclonic features (i.e., eddies or recirculation), and that electronically-tagged sharks have the potential to serve as submesoscale-resolving oceanographic platforms and fill spatial gaps in the ocean observing network, particularly in coastal and polar regions.
Publications
Andrzejaczek et al. (2024). Lunar cycle effects on pelagic predators and fisheries: insights into tuna, billfish, sharks, and rays Ferter*, Pagniello* et al. (2024). Atlantic bluefin tuna tagged off Norway show extensive annual migrations, high site-fidelity and dynamic behaviour in the Atlantic Ocean and Mediterranean Sea. Pagniello et al. (2024). Novel CTD tag establishes shark fins as ocean observing platforms. Pagniello et al. (2023). Tagging of Atlantic bluefin tuna off Ireland reveals use of distinct oceanographic hotspots.Other Ongoing Projects
In the MILLO Group, we love ocean data and challenging problems! Team members work on a wide variety of projects. Projects include:
- the influence of mesoscale features on deep scattering layers,
- genome-wide population structure of commercially-important fisheries,
- sensitivity of ecological conclusions to length-to-weight constants,
- impact of policy changes on livelihoods in small island state communities,
- rescue and digitization of historic oceanographic datasets, and
- size-structured biomass of the zooplankton community.
We are always excited to explore new research directions, especially with complex, multivariable and FAIR datasets.