Roughly a quarter of the global ocean floor has been mapped to what scientists consider modern resolution, meaning sonar instruments have directly measured the depth at a meaningful level of detail. That leaves about three-quarters of the seabed known only through indirect estimates, primarily satellite-derived gravity measurements that can detect large features but miss anything smaller than a few kilometers across. The gap between what we have surveyed and what remains unseen is enormous, and closing it involves challenges of cost, technology, and sheer physical scale that make deep-ocean mapping one of the most resource-intensive scientific endeavors on Earth.
What “Mapped” Actually Means
The answer to “how much has been mapped” depends entirely on what you count as a map. In one sense, we have a rough picture of the entire ocean floor. Satellite altimeters measure the height of the sea surface with extreme precision, and because the gravitational pull of underwater mountains, ridges, and trenches subtly warps the water above them, scientists can infer the shape of the seabed from those surface measurements. A landmark 1997 study combined satellite gravity data from the Geosat and ERS-1 spacecraft with ship depth soundings to produce a global digital bathymetric map with a horizontal resolution of roughly 1 to 12 kilometers.1Science. Global Sea Floor Topography from Satellite Altimetry and Ship Depth Soundings That approach gave us the first truly global view of ocean topography, and updated versions of it still form the backbone of products like GEBCO’s global grid.
But a resolution of several kilometers means you can see a massive seamount while completely missing a volcanic vent, a submarine landslide scar, or an abyssal hill a few hundred meters across. The gold standard for ocean-floor mapping is multibeam echosounder data, collected by ships or underwater vehicles that send sound pulses to the seabed and measure the return time across a wide swath. Multibeam can resolve features on the scale of tens of meters or less. When oceanographers and projects like Seabed 2030 report the percentage of the ocean that has been “mapped,” they are counting only areas where multibeam or equivalently detailed sonar data exist. By that measure, roughly 25 percent of the ocean floor qualifies, and the rest is filled in with satellite-estimated depths that look smooth and featureless compared to reality.
Why So Little Has Been Surveyed
The ocean covers about 361 million square kilometers, and a research vessel running multibeam sonar can only survey a relatively narrow strip of seabed on each pass. The deeper the water, the wider that strip becomes, but even in the deep ocean a single ship covers only a few kilometers of width at a time. One widely cited estimate puts the effort required to survey the entire ocean floor with modern multibeam at more than 200 ship-years, at a cost running into the billions of dollars.2Comptes Rendus. Géoscience. Bathymetry from space: Rationale and requirements for a new, high-resolution altimetric mission That is 200 years of a single ship running continuously, or a massive fleet operating for a decade or more. By comparison, the surface of Mars and the Moon have been mapped in far greater detail, largely because radar and cameras in orbit can image solid planetary surfaces much more easily than sound can penetrate kilometers of seawater.
Cost is only part of the problem. Much of the unmapped ocean lies in remote areas far from shipping lanes, research stations, and ports. The Southern Ocean around Antarctica, vast stretches of the central Pacific, and parts of the Arctic are among the least surveyed. Ships that do transit these regions are often on other missions and may not carry multibeam equipment, or may not share their data if they do. Weather, sea ice, and fuel logistics all limit how much time any vessel can spend surveying a given patch of deep water.
The Tools Used to Map the Seabed
Several technologies contribute to ocean-floor mapping, each suited to different depths, resolutions, and budgets.
- Multibeam echosounders: Mounted on ship hulls, these are the workhorse of deep-ocean surveying. They emit fan-shaped pulses of sound and record the returns across dozens or hundreds of beams, building a swath of depth measurements with each ping. Resolution depends on depth, but in deep water a hull-mounted system can typically resolve features on the order of 50 to 100 meters.
- Autonomous underwater vehicles (AUVs): For finer detail, AUVs can fly within tens of meters of the seabed, producing centimeter-to-meter-scale maps. Researchers have developed algorithms that allow AUVs to conduct near-bottom surveys autonomously in the deep sea, covering terrain that would be impractical for crewed submersibles to map systematically.3The International Journal of Robotics Research. Techniques for Deep Sea Near Bottom Survey Using an Autonomous Underwater Vehicle The trade-off is speed: an AUV covers a tiny area compared to a surface ship.
- Satellite altimetry: As described above, satellites measure sea-surface height to infer underwater topography. This gives complete global coverage but at low resolution, typically several kilometers at best.
- Satellite-derived bathymetry for shallow water: In coastal and nearshore zones where water is clear enough for light to penetrate, optical satellites like Sentinel-2 can estimate depth from the color of the water. One study comparing this method to high-resolution airborne lidar found a median absolute error of about half a meter for depths of 7 meters or less, with the satellite approach offering far greater spatial coverage than lidar because it is not limited by turbidity or flight constraints in the same way.4Coastal Engineering. On the use of Sentinel-2 satellites and lidar surveys for the change detection of shallow bathymetry: The case study of North Carolina inlets This technique is useless in the deep ocean, but it is increasingly valuable for monitoring coastal change.
None of these tools alone can solve the mapping problem. The deep ocean is too vast for AUVs to cover at high resolution, too deep for optical satellites to penetrate, and too detailed for satellite gravity to capture. The practical path forward involves combining all of them, along with creative approaches to data collection.
Crowd-Sourced Bathymetry and the Seabed 2030 Push
One of the more promising developments in recent years is the idea of turning the global commercial fleet into an ad hoc mapping force. Thousands of cargo ships, fishing vessels, and ferries cross the ocean every day carrying some form of depth-sounding equipment, from survey-grade multibeam arrays to simple fish-finders. If those vessels share their data, the cumulative coverage can be significant. The commercial charting company Olex AS runs what may be the largest such initiative, having compiled over 8.6 billion depth measurements from approximately 10,000 vessels worldwide.5PLoS ONE. Generating higher resolution regional seafloor maps from crowd-sourced bathymetry The quality of crowd-sourced data varies widely depending on the equipment and conditions, but researchers have shown it can meaningfully fill in gaps in existing maps, especially in well-trafficked areas like fishing grounds and shipping corridors.
The Nippon Foundation-GEBCO Seabed 2030 project, launched in 2017, set the ambitious goal of producing a complete map of the ocean floor by 2030. When the project started, less than 6 percent of the seabed had been mapped by multibeam sonar. By the early 2020s, that figure had climbed past 20 percent and continues to grow as new data are incorporated from research cruises, crowd-sourced contributions, and dedicated survey campaigns. Whether the 2030 deadline is realistic remains an open question. The rate of new coverage has accelerated, but the remaining unmapped areas are overwhelmingly in remote, deep, and logistically difficult parts of the ocean. Optimists point to falling costs for autonomous surface vehicles and AUVs; skeptics note that even with technological improvements, the sheer area left to cover is staggering.
What Incomplete Maps Actually Cost Us
The gaps in our ocean-floor knowledge are not just an abstract scientific concern. They have practical consequences across a surprising range of human activities.
Submarine cables carry more than 95 percent of intercontinental data traffic, and routing those cables safely requires detailed knowledge of the seabed. Engineers need to avoid steep slopes, unstable sediments, active faults, and features like submarine valleys and seamounts. In the East China Sea, for example, field investigations using multibeam sonar, side-scan sonar, and sediment sampling identified a catalog of hazards along planned cable routes, including scouring, sand waves, shallow gas pockets, landslide-prone slopes, and hard rock outcrops.6Geofluids. Analysis of Engineering and Geological Conditions of International Submarine Optical Fiber Cable Routing in the East China Sea Section Autonomous underwater vehicles are increasingly used to classify seafloor types and help determine safe cable corridors before installation begins.7IEEE Journal of Oceanic Engineering. Efficient Seafloor Classification and Submarine Cable Route Design Using an Autonomous Underwater Vehicle Without high-resolution maps, cable companies are essentially laying critical infrastructure through poorly charted terrain.
Deep-sea mining is another domain where mapping gaps create real problems. The manganese nodule fields in the abyssal Pacific, for instance, are a target for future mineral extraction, but nodule density varies dramatically over short distances depending on local topography and sediment conditions. Research using AUV-based acoustic and optical surveys has shown that meaningful differences in nodule abundance appear at scales of 10 to 100 meters, a level of detail far beyond what satellite-derived maps can provide.8Biogeosciences. Understanding Mn-nodule distribution and evaluation of related deep-sea mining impacts using AUV-based hydroacoustic and optical data Accurate resource assessment and environmental impact modeling both depend on terrain knowledge that simply does not exist for most of the deep ocean.
How Mapping Gaps Affect Ocean Science
Oceanographers studying deep-water circulation and mixing rely on bathymetric models to understand how tides interact with the seafloor. When currents flow over rough bottom terrain, they generate small-scale internal waves that help mix heat and nutrients vertically through the water column, a process that matters for global climate models. But current global bathymetry products do not resolve abyssal hills, the ubiquitous mid-ocean ridge features typically a few kilometers across, even though these hills drive a non-negligible amount of tidal energy conversion.9PubMed Central. Deep-ocean mixing driven by small-scale internal tides Researchers have to supplement their models with separate estimates of energy conversion at these unresolved features, introducing uncertainty that propagates into predictions of ocean circulation and, by extension, climate.
The implication is circular and frustrating: we need better bathymetry to build better ocean models, and we need better ocean models to understand how the climate system works, but mapping the features that matter most requires exactly the kind of expensive, time-consuming ship-based surveys that have so far covered only a fraction of the seafloor. Satellite gravity can tell us where a ridge system exists; it cannot tell us the fine-grained roughness that determines how much energy the tides lose to mixing there.
Geopolitical Stakes of Seabed Mapping
Ocean-floor mapping is not just a scientific enterprise. It is a legal and political one. Under the United Nations Convention on the Law of the Sea, coastal nations can claim an extended continental shelf beyond their standard 200-nautical-mile exclusive economic zone, but only if they provide specific kinds of scientific evidence about the shape and geology of the seabed. Bathymetric data are central to those claims, and in contested regions like the Arctic, where multiple countries have overlapping shelf claims, the quality of seafloor mapping directly shapes who controls what resources.10Political Geography. Technology and the construction of oceanic space: Bathymetry and the Arctic continental shelf dispute
This creates an interesting tension. Governments have strong incentives to map their own continental margins in detail, but some of that data has historically been classified or restricted, particularly when collected by military vessels. A longstanding debate in the United States, for instance, has centered on the confidential classification of multibeam data collected within the U.S. exclusive economic zone, with critics arguing that withholding such data hinders civilian research and international cooperation. The result is a patchwork: some of the best-mapped areas on Earth are continental shelves where national interests have justified the expense, while the deep abyssal plains far from any coast remain among the least known places on the planet.
Why We Know More About Planetary Surfaces Than Our Own Seabed
People are often startled to hear that we have better topographic maps of Mars than of Earth’s ocean floor, and the comparison is essentially accurate. The reason is straightforward: electromagnetic radiation, whether radar, laser, or visible light, travels freely through the vacuum of space and through thin planetary atmospheres. A single orbiting spacecraft can image an entire rocky planet or moon in a matter of months. Water blocks those signals. Below a few meters of depth, mapping the ocean requires sound, and sound requires a source close to the surface or, for high resolution, close to the bottom. You cannot map the deep ocean from orbit the way you can map the Martian highlands. Every additional meter of resolution requires getting closer, moving slower, or spending more money.
The comparison also highlights a funding disparity. Planetary mapping missions have attracted billions of dollars in government investment, driven partly by public fascination with space and partly by clear institutional mandates at agencies like NASA. Ocean mapping has no single agency champion of comparable scale, and its benefits, while substantial, are spread across industries and scientific disciplines in ways that make it harder to rally political support. Seabed 2030 and similar initiatives have raised the profile of ocean mapping considerably, but the total investment remains modest compared to what has been spent imaging other worlds.
Emerging Approaches and Uncrewed Vehicles
The most plausible path to full ocean-floor coverage involves reducing the cost per square kilometer of multibeam surveying, and that means taking humans off the ships wherever possible. Uncrewed surface vehicles, essentially autonomous boats equipped with multibeam sonars, can operate for weeks or months at a time without crew, significantly cutting the cost of fuel, provisions, and personnel. Several companies and research groups are already deploying these platforms on mapping missions, particularly in remote areas where sending a crewed research vessel would be prohibitively expensive.
AUVs are evolving in parallel, gaining longer endurance and better navigation capabilities that allow them to operate farther from support ships. For deep-sea applications where centimeter-scale resolution matters, such as mineral exploration, habitat assessment, or infrastructure inspection, AUVs remain indispensable. The challenge is scaling up: even the most capable AUV covers a tiny patch of seabed per deployment compared to a surface vessel. The likely future is a layered approach, with satellites providing the coarsest global view, uncrewed surface vessels filling in multibeam coverage over large areas, and AUVs deployed selectively where the highest resolution is needed.
Advances in data processing also matter. Machine learning techniques are being applied to extract more information from existing satellite gravity data, potentially improving the resolution of global models without collecting a single new sonar measurement. Crowd-sourced data from fishing and cargo fleets continue to grow in volume, and better algorithms for filtering and merging these noisy datasets could help turn a large quantity of mediocre soundings into a useful regional map. None of these shortcuts replaces systematic multibeam surveying, but together they could accelerate the timeline for achieving something close to complete coverage.