Marine scientists rely on a sprawling toolkit that spans every layer of the ocean, from satellites orbiting hundreds of kilometers overhead to sensors buried in deep-sea sediment. No single instrument can capture the complexity of the marine environment, so researchers combine dozens of technologies, each tailored to a different depth, scale, or type of observation. Some of these tools have been in continuous use for decades, while others, like environmental DNA sampling and AI-driven image analysis, have only recently matured enough for routine fieldwork.
Satellites and Remote Sensing
Observing the ocean starts, counterintuitively, from space. Earth-observing satellites carry sensors that measure sea surface temperature, ocean color (a proxy for chlorophyll and phytoplankton concentration), sea level height, wave patterns, and water turbidity across enormous areas simultaneously. The Moderate-resolution Imaging Spectroradiometer (MODIS) aboard NASA’s Aqua satellite, for example, has provided daily global ocean-color imagery for nearly two decades. Researchers have used MODIS data processed through cloud-computing platforms to track long-term changes in chlorophyll content, sea surface temperature, and turbidity in specific regions over periods of fifteen years or more.1Regional Studies in Marine Science. Monitoring of sea surface temperature, chlorophyll, and turbidity in Tunisian waters from 2005 to 2020 using MODIS imagery and the Google Earth Engine Satellite altimeters, which bounce radar pulses off the sea surface, map ocean circulation patterns and detect eddies that would be invisible from a ship. The limitation is depth: satellites see only the surface or the first few meters of water. Everything below requires instruments that go into the ocean itself.
Research Vessels
For all the advances in autonomous and remote technology, crewed research ships remain the backbone of deep-ocean science. They serve as large, stable platforms capable of deploying heavy equipment, collecting water and sediment samples from the surface to the seafloor, and running full-scale laboratories on board for immediate processing and analysis of biological, chemical, and biogeochemical data.2Limnology and Oceanography. The essential role of large research vessels in marine ecosystem observations and ocean sustainability A standard oceanographic cruise might lower a rosette of water-sampling bottles fitted with a conductivity-temperature-depth (CTD) profiler, tow plankton nets, deploy and recover autonomous vehicles, and carry out trawl surveys all in a single expedition. Ships also act as communication and power hubs for instruments that cannot operate independently for long periods. No robotic platform yet matches a vessel’s ability to adapt plans in real time when something unexpected turns up in the data.
Autonomous Profiling Floats
The Argo program is one of the most ambitious sustained observation efforts in ocean science. It consists of a global array of more than 3,000 free-drifting profiling floats that sink to a preset depth, drift with the current for about ten days, then rise to the surface while measuring temperature and salinity along the way. Once at the surface, each float transmits its data via satellite before sinking again.3Journal of Atmospheric and Oceanic Technology. “Deep-Arvor”: A New Profiling Float to Extend the Argo Observations Down to 4000-m Depth This cycle repeats for years, and the cumulative dataset has become essential for tracking ocean heat content, salinity changes, and large-scale circulation patterns.
Standard Argo floats profile the upper 2,000 meters. A newer generation, such as the Deep-Arvor float developed by the French oceanographic institute Ifremer, extends that range down to 4,000 meters, carrying sensors for temperature, salinity, and optionally dissolved oxygen.3Journal of Atmospheric and Oceanic Technology. “Deep-Arvor”: A New Profiling Float to Extend the Argo Observations Down to 4000-m Depth Reaching the deep ocean matters because a large share of heat absorbed by the planet ends up stored below 2,000 meters, and conventional floats simply miss it.
Surface Drifters
While Argo floats profile vertically, surface drifters track horizontal movement. These are small buoys equipped with satellite transmitters and underwater sails (called drogues) that keep them moving with the current rather than being blown by wind. Thousands of drifter trajectories have been used to map global surface circulation, identify convergence zones where debris accumulates, and study how ocean currents connect distant coastlines. Researchers have used the global drifter dataset to model the long-term fate of marine debris, revealing five major aggregation sites in the subtropical oceans maintained by converging surface currents.4PubMed. Pathways of marine debris derived from trajectories of Lagrangian drifters Beyond pollution studies, drifter data feed into weather forecasting, search-and-rescue planning, and calibration of satellite-derived current maps.
Underwater Gliders
An underwater glider looks a bit like a torpedo with stubby wings. It has no propeller; instead, it changes its buoyancy to glide downward, then shifts ballast to glide back up, tracing a sawtooth path through the water while sensors record temperature, salinity, dissolved oxygen, chlorophyll fluorescence, and sometimes acoustic signals along the way. Because gliders use so little energy, a single unit can operate for weeks or months on a single battery charge, covering hundreds of kilometers.
Repeated glider surveys in the Japan Sea, for instance, revealed the detailed structure of ocean fronts and currents that would have required multiple ship cruises to map at comparable resolution. Those surveys produced new estimates of the volume of water transported by major currents in the region, combining glider data with satellite altimetry and ship-based profiles to build a multiscale picture of circulation.5Journal of Marine Systems. Observations of oceanic fronts and water-mass properties in the central Japan Sea: Repeated surveys from an underwater glider Gliders are also increasingly used to carry hydrophones for passive acoustic monitoring, essentially turning them into mobile listening stations for whale calls and ship noise.
Remotely Operated Vehicles
Remotely operated vehicles (ROVs) are tethered robots controlled by a pilot on a surface vessel through a cable that supplies power, sends commands, and relays live video. They range from small, portable units the size of a suitcase to massive work-class machines weighing several tonnes. The smaller ones carry cameras and simple sampling tools; the larger ones have hydraulic manipulator arms, suction samplers, and enough thrust to operate in strong currents at depths well beyond what human divers can reach.
ROVs originally built for the offshore oil and gas industry have proven valuable for deep-sea biology. Industrial work-class vehicles have been used to collect taxonomic specimens alongside ecological observations, to document feeding behavior of deep-water fish species, and to carry out manipulative experiments on the seafloor, such as placing and recovering settlement panels or caging organisms for exclusion studies.6Zoologica Scripta. Using existing industrial remotely operated vehicles for deep‐sea science Their main drawback is the tether: ROVs can only travel as far as their cable allows, and they require a ship overhead to operate.
Cabled Seafloor Observatories and Deep-Sea Landers
Some questions demand that instruments stay in one place on the seafloor for months or years, recording conditions continuously. Cabled seafloor observatories address this by running fiber-optic cables from shore to a network of instrument nodes on the seabed. These cables supply electrical power and high-bandwidth data transmission, so sensors can operate indefinitely without battery or data-storage limits. The European Multidisciplinary Seafloor Observatory (EMSO) network, for example, connects instrument arrays at geologically and biologically active sites across Europe’s seafloor, supporting everything from seismometers to camera systems and chemical sensors.7Nuclear Instruments and Methods in Physics Research Section A. EMSO: European multidisciplinary seafloor observatory
Where running a cable is impractical, researchers use free-falling benthic landers. These are autonomous platforms that sink to the seafloor under their own weight, carry out pre-programmed measurements, and then release ballast to float back to the surface for recovery. A benthic chamber lander, for instance, seals a small patch of sediment and the overlying water inside a chamber, then monitors how concentrations of dissolved compounds change over time, revealing the rates at which the seafloor takes up or releases chemicals through biological and geochemical processes.8National Oceanography Centre. Landers and Benthic Instruments Landers have operated at depths exceeding 10,000 meters in ocean trenches, places no cabled observatory yet reaches.
Acoustic Tools
Sound travels far more efficiently than light in seawater, so acoustics underpin many marine observation methods. On the active side, multibeam echosounders emit fan-shaped pulses of sound and record the echoes to map the seafloor at high resolution. Over the past two decades, scientists have increasingly used the water-column data from these systems, not just the bottom echoes, to study fish schools, zooplankton layers, gas seeps from the seafloor, and even hydrothermal vents.9Estuarine, Coastal and Shelf Science. A review of oceanographic applications of water column data from multibeam echosounders Single-beam and split-beam echosounders are also widely used to estimate fish abundance during stock-assessment surveys.
Passive acoustic monitoring (PAM) takes the opposite approach: instead of sending sound out, you listen. Hydrophones deployed on moorings, gliders, or seafloor platforms record the sounds made by marine mammals, fish, invertebrates, and human activities like shipping and construction. Detecting endangered whale species in real time is a high-priority application. Researchers designing PAM systems for North Atlantic right whales in the Gulf of St. Lawrence found that hydrophone arrays using beamforming, a technique that focuses the listening direction, achieved effective detection ranges roughly fifteen times greater than a single hydrophone, because beamforming suppresses noise from nearby ships while amplifying the whale’s call.10Applied Acoustics. Optimizing passive acoustic systems for marine mammal detection and localization: Application to real-time monitoring north Atlantic right whales in Gulf of St. Lawrence Real-time PAM feeds can trigger vessel speed restrictions or route changes to reduce the risk of ship strikes.
Underwater Video Systems
Cameras are among the simplest conceptual tools in the marine scientist’s arsenal, but the engineering challenges of working underwater, in darkness, under crushing pressure, and often in turbid or biologically fouled conditions, make them anything but simple in practice. Baited remote underwater stereo-video systems (stereo-BRUVs) are a popular, cost-effective, and non-invasive approach for surveying bottom-dwelling fish communities.11Methods in Ecology and Evolution. A field and video annotation guide for baited remote underwater stereo‐video surveys of demersal fish assemblages A pair of cameras is mounted at a fixed separation on a frame, with a bait canister in the field of view. The stereo setup lets analysts measure the body length of individual fish from the footage, not just count them. Studies optimizing these systems in estuarine habitats have recorded dozens of species, including many of commercial importance, with soak times as short as thirty minutes yielding results comparable to longer deployments.12Journal of Experimental Marine Biology and Ecology. Optimisation of baited remote underwater video sampling designs for estuarine fish assemblages
Time-lapse camera systems mounted on fixed rigs or observatory nodes take the concept further, capturing images at regular intervals over weeks or months to track changes in sessile organisms like corals and sponges. The bottleneck with all camera-based methods is the sheer volume of imagery generated. A single deployment can produce tens of thousands of images, which is where machine learning steps in (covered below).
Aerial Drones
Above the water, uncrewed aerial systems (drones) have become a routine tool for studying marine megafauna and coastal habitats. Drones can hover directly above a whale, photograph it from a consistent altitude, and produce measurements of body length and width that, combined with body-volume modeling, yield estimates of body condition and even mass. Researchers have used this photogrammetry approach across multiple North Atlantic humpback whale foraging grounds to compare body condition between populations.13Frontiers in Marine Science. Drone-based photogrammetry reveals differences in humpback whale body condition and mass across North Atlantic foraging grounds The same technique has been applied to sperm whales, where drone-derived body volume combined with tissue density data from biologging tags allowed researchers to estimate total body lipid stores in free-ranging animals.14PubMed Central. Estimating Total Body Lipid Store of Free-Ranging Whales In Vivo Using Drone Photogrammetry and Biologging Tags
Beyond whales, drones are used to survey seal haul-outs, nesting seabird colonies, coral reef flats, and coastal erosion. Their advantage over crewed aircraft is cost: a single drone flight can be planned and executed in minutes by a two-person field team, whereas manned aerial surveys require extensive logistics. Their limitation is endurance. Most consumer-grade drones fly for only twenty to thirty minutes on a charge, and wind, rain, and salt spray can ground them quickly.
Animal Bio-Logging
Some of the hardest places to observe in the ocean are wherever animals actually go: beneath sea ice, inside deep canyons, through murky estuaries. Attaching miniaturized sensors directly to marine animals turns them into living oceanographic platforms. Early bio-loggers recorded little more than dive depth and time, but modern tags can measure water temperature, salinity, light levels, acceleration, magnetic heading, heart rate, and even the animal’s jaw movements to detect feeding events.15PubMed. Biologging and Biotelemetry: Tools for Understanding the Lives and Environments of Marine Animals Elephant seals carrying satellite-linked CTD tags have profiled waters under Antarctic ice that no float or glider can reach. Sea turtles fitted with GPS and dive recorders have mapped oceanic migration corridors across entire ocean basins. The data serve a dual purpose: they reveal animal behavior and simultaneously fill gaps in the oceanographic record.
Environmental DNA
Every organism sheds DNA into its surroundings through skin cells, mucus, feces, and reproductive material. Environmental DNA (eDNA) metabarcoding extracts and sequences this genetic material directly from water samples to identify which species are present, without needing to see, catch, or even disturb any animal. The method has transformed biodiversity monitoring because it is non-invasive, cost-effective, and capable of detecting rare, cryptic, and invasive species that traditional net-and-trap surveys often miss.16PubMed Central. Environmental DNA Metabarcoding in Marine Ecosystems: Global Advances, Methodological Challenges, and Applications in the MENA Region
Early proof-of-concept work showed that small seawater samples could yield eDNA from fifteen or more fish species, including commercially important ones and species rarely recorded by conventional methods. When compared head-to-head with nine traditional survey techniques, eDNA covered fish diversity as well as or better than any single conventional method.17PLoS ONE. Detection of a Diverse Marine Fish Fauna Using Environmental DNA from Seawater Samples The technique even picked up a rare vagrant species and several bird species whose DNA had washed into the sea. Limitations remain: eDNA degrades quickly in warm or UV-exposed water, it cannot tell you how many individuals are present with much precision, and contamination during sampling is an ever-present risk. Still, for rapid biodiversity screening, nothing else comes close to its efficiency.
In-Situ Chemical Sensors and Plankton Monitors
Understanding ocean chemistry on fine spatial and temporal scales requires sensors that can ride along on floats, gliders, moorings, or ROVs. Optical chemical sensors called optodes measure dissolved oxygen, pH, and carbon dioxide concentration in seawater by shining light onto a sensing material and reading the fluorescence response. A compact optode system with interchangeable sensor caps can switch between measuring oxygen, pH, and CO₂ using the same logger and battery, which reduces hardware costs and makes it practical to add chemical measurements to platforms originally designed only for temperature and salinity.18Limnology and Oceanography: Methods. A versatile optode system for oxygen, carbon dioxide, and pH measurements in seawater with integrated battery and logger These measurements are crucial for tracking ocean acidification and deoxygenation, two consequences of climate change that threaten marine ecosystems globally.
For plankton, the organisms at the base of the marine food web, scientists have used the Continuous Plankton Recorder (CPR) since the 1930s. The CPR is a mechanical device towed behind ships of opportunity (cargo vessels and ferries on regular routes) that continuously filters plankton onto a moving band of silk, which is later analyzed in the lab. It gives excellent taxonomic resolution but relatively coarse spatial sampling. Newer imaging instruments like the Plankton Imager (Pi-10) photograph plankton in situ as water flows through an optical chamber, capturing fine-scale variation in abundance and body size that the CPR misses. A comparison of the two instruments in the North Sea found that the Pi-10 recorded copepod abundances roughly 36 times higher than the CPR, likely because the recorder’s mesh lets smaller individuals pass through. The two tools captured different but complementary views: the CPR revealed clear shifts in community composition with latitude, while the Pi-10 tracked fine-scale spatial variation in total abundance and size.19ICES Journal of Marine Science. The Plankton Imager and Continuous Plankton Recorder observations provide complementary insights into copepod dynamics across spatial scales
AI and Automated Image Analysis
The explosion of camera-equipped platforms, from ROVs to stereo-BRUVs to time-lapse rigs, has created a data bottleneck. Manually reviewing tens of thousands of images to identify and measure organisms is slow and expensive. Machine learning, and specifically deep learning, is rapidly filling that gap. Marine scientists now use remote underwater imagery analyzed by deep learning models to classify fish species in their natural habitats, unlocking information about behavior and abundance that would take far too long to extract by hand.20Fish and Fisheries. Computer vision and deep learning for fish classification in underwater habitats: A survey
Interactive machine-learning tools are making this accessible to researchers who are not computer scientists. One open-source tool, RootPainter, was used to detect and measure the surface area of a cold-water sponge species across a dataset of more than 18,000 time-lapse images with high accuracy, achieving analysis speeds six to sixteen times faster than manual annotation thanks to transfer learning.21Royal Society Open Science. New interactive machine learning tool for marine image analysis Benchmarking studies have tested over a dozen deep-learning architectures on underwater species classification; the best-performing model in one recent evaluation reached about 91% top-1 accuracy and nearly 99% top-3 accuracy across twenty marine species photographed under difficult lighting and turbidity conditions.22arXiv. AQUA20: A Benchmark Dataset for Underwater Species Classification under Challenging Conditions The field is moving fast, and the trend is toward integrating these algorithms directly onto platforms so that an ROV or glider can identify organisms in real time, adjusting its sampling or flight path accordingly.
How These Tools Work Together
What makes modern ocean observation powerful is not any single instrument but the way these tools layer on top of one another. A satellite detects an anomalous bloom of phytoplankton in a coastal region. A glider is redirected to transect the bloom, profiling temperature, salinity, chlorophyll, and dissolved oxygen along the way. An Argo float drifting nearby provides context on the deeper water column beneath the bloom. An eDNA sample collected from a research vessel at the bloom’s edge reveals which fish species are aggregating there. A stereo-BRUV deployed on the seabed below captures video of the fish assemblage, and a deep-learning model processes the footage overnight. A drone overhead photographs a pod of whales feeding at the bloom’s surface, and a bio-logging tag on one of those whales records its diving behavior in relation to the prey layer detected by the echosounder.
Each instrument fills in a piece that the others cannot see. Satellites lack depth. Floats lack spatial resolution. Ships lack persistence. Gliders lack speed. ROVs lack range. Drones lack the ability to go underwater. eDNA lacks the ability to count individuals. Cameras produce more data than humans can analyze without AI. The trend across marine science is toward integrated observing systems that combine as many of these perspectives as possible, fusing their data streams into a more complete picture of an ocean that remains, even now, vastly under-observed.