# Automated seafloor search reveals 73 hidden volcanic calderas

> A study in Communications Earth & Environment reports 73 previously unknown volcanic calderas on the ocean floor, found through a semi-automated search of global seafloor maps. The work, led by volcanologist Andrea Verolino of Paris-Saclay University, expands a sparse global record of...

Canonical URL: https://www.argo.net/automated-seafloor-search-reveals-73-hidden-volcanic-calderas/
Byline: Paris-Saclay University
Published: 2026-07-16T06:35:02+00:00
Categories: Earth, News

![Bathymetry of a known caldera, Niuatahi, in the Tongan archipelago](https://www.argo.net/wp-content/uploads/2026/07/seafloor_caldera_bathymetry.jpg)

A study in [Communications Earth & Environment](https://www.nature.com/articles/s43247-026-03779-3) reports 73 previously unknown volcanic calderas on the ocean floor, found through a semi-automated search of global seafloor maps. The work, led by volcanologist **Andrea Verolino** of **Paris-Saclay University**, expands a sparse global record of large submarine volcanic structures.

The discovery matters because **submarine calderas** can mark some of the ocean's most powerful volcanic systems. These huge depressions form when magma drains from beneath a volcano and the surface above collapses inward. Underwater, they can be difficult to see, hard to visit and easy to miss in uneven seafloor data.

Before this survey, fewer than 30 submarine calderas had been documented worldwide. The new work points to many more candidates across the ocean basins and gives scientists a reproducible way to search for them as better maps become available.

## A hidden map of submarine volcanoes

Seventy-three newly identified caldera candidates give researchers a broader view of volcanic architecture beneath the sea. The study examined global bathymetric data, which records the shape of the seafloor, then used automated detection and expert review to separate likely calderas from look-alike depressions.

Calderas are among the largest features a volcano can leave behind. On land, they can be mapped with aircraft, satellites, field teams and rock samples. In the deep ocean, even a vast collapsed crater may sit far below ships and satellites, hidden by kilometers of water.

The paper describes the dataset as one that "fills a major observational gap." That short phrase captures the central problem. Earth's oceans cover most of the planet, yet scientists still have limited high-resolution views of much of the seafloor.

The global map also helps place the new candidates beside previously documented calderas. That comparison gives scientists a way to ask bigger questions about where these features form and which settings deserve more attention.

## How the algorithm searched the seafloor

The team adapted a detection approach originally used to spot crater-like shapes on Mars. For this study, the method was applied to the **General Bathymetric Chart of the Oceans**, a global dataset that provides broad coverage of seafloor topography.

The algorithm searched for depressions with shapes that could resemble calderas. A submarine caldera often appears as a broad hollow with a rim or surrounding volcanic structure. From above, it can look like a missing bite from a seamount or volcanic cone.

Automated detection gave the researchers a first pass across the global ocean floor. That matters because the seafloor is too large for manual inspection alone. A computer can flag possible structures at scale, then researchers can apply geological judgment.

After the automated search, the team used filters, manual inspection and statistical validation. The process combined machine efficiency with expert review, which is important for a dataset filled with ridges, basins, faults, landslide scars and other features that can mimic volcanic collapse.

## From 87,435 signals to 78 likely calderas

The first sweep produced 87,435 possible formations. That enormous number shows how messy global seafloor detection can be. Many circular or bowl-like features can appear in bathymetric maps, especially when the underlying data have variable resolution.

The researchers then narrowed the list through a series of steps. Filters removed poor fits. Experts reviewed the remaining candidates. **Principal Component Analysis**, a statistical tool for finding patterns in complex datasets, helped validate the classification.

The final list contained 78 likely submarine calderas. Five were already recognized as submarine calderas, which gave the team a useful check on the method. The remaining 73 were classified as **previously undocumented submarine calderas**.

This result could more than triple the number of known submarine calderas if future surveys confirm the candidates. The paper presents the findings as a framework as well as a dataset. That means the search can improve as higher-resolution bathymetric maps become available.

The study also keeps uncertainty in view. These are likely calderas identified from seafloor shape and supporting analysis. Direct exploration, detailed mapping, rock sampling and monitoring would help confirm their origins and assess their current state.

## Why submarine calderas matter

Submarine volcanoes account for a major share of Earth's volcanic activity. Many erupt quietly along mid-ocean ridges, where tectonic plates pull apart and magma rises to create new seafloor. Calderas point to a more explosive part of the underwater volcanic story.

A large underwater caldera can form after a major eruption drains a magma chamber. As support is lost below, the roof of the volcanic system collapses. The result is a broad depression that may remain visible in bathymetric maps long after the eruption has ended.

Some submarine calderas can be associated with hazardous events. Explosive eruptions can disturb the ocean, generate tsunamis, send pressure waves through the atmosphere and release steam and ash. The 2022 **Hunga Tonga-Hunga Haʻapai** eruption showed how an underwater volcanic system can have effects far beyond its island setting.

Knowing where calderas are located helps scientists decide which places deserve closer mapping and monitoring. A global inventory can guide research cruises, hazard studies and seafloor observatories. It can also help planners think more carefully about coastal communities, shipping routes and undersea infrastructure.

The new study does not report that the 73 newly identified candidates are active. Their importance comes from location, shape and context. A caldera can preserve a record of past volcanic behavior and that record can help researchers understand future risk.

## Where the new calderas were found

The candidates appear across several tectonic settings. Eight were found at **mid-ocean ridges**, where plates move apart. These settings are among the most volcanically active places on Earth, although much of the activity takes place far below the surface.

Nine candidates were identified in **volcanic arcs**. These arcs form where one tectonic plate sinks beneath another. Water and heat help generate magma, which can feed chains of volcanoes on islands, continental margins and the seafloor.

The largest group appeared in interior tectonic settings. The study identified 61 candidates away from plate boundaries, including areas associated with seamount chains. That result is striking because it suggests many caldera-forming systems may sit in places that receive less attention than classic plate-boundary volcanoes.

This distribution gives researchers new targets for comparing volcanic processes. Calderas near ridges, arcs and interior ocean settings may form under different magma supply conditions. Their shapes and depths may also reveal how eruptions and collapses unfold beneath water.

Because the search used globally available bathymetry, it also highlights the limits of current ocean mapping. Better maps could sharpen known candidates, remove uncertain ones and reveal additional structures now hidden by coarse resolution.

## Seven sites scientists want to explore next

The researchers highlighted seven newly identified calderas as especially interesting targets for future work. Their locations, water depths and shapes could make them useful for studying **submarine volcanic hazards** and caldera formation.

Future exploration could involve research vessels, sonar mapping, remotely operated vehicles and seafloor sampling. High-resolution bathymetry would be a first step for many sites. From there, scientists could look for volcanic deposits, hydrothermal activity, deformation, or other clues to each system's history.

Direct observations would also help distinguish older volcanic structures from systems with signs of recent activity. Submarine volcanoes can be quiet for long periods and the ocean hides many signals that would be obvious on land. That makes targeted follow-up especially valuable.

The broader value of the study is its repeatable method. As more detailed seafloor surveys are collected, the framework can be run again and refined. That could turn today's first global pass into a growing inventory of the ocean's hidden volcanic structures.

For now, the new dataset gives Earth scientists a sharper starting point. Beneath the waves, the planet's volcanic surface is still being mapped, one hidden caldera at a time.