A roof held up by a fluid that is leaving
Underneath most large volcanoes sits a magma chamber — a body of molten and partially molten rock a few kilometres down, whose internal pressure helps support the weight of the rock roof above it. That support is not free: it depends on the chamber staying full and pressurised. When a large eruption rapidly drains a significant fraction of the chamber's volume, or when magma migrates sideways into a lateral intrusion instead of erupting upward, the roof loses the pressure that was holding it up and is left spanning an increasingly large, increasingly unsupported void.
Why the roof fails along a ring, not a random crack
Rock is strong in compression but comparatively weak in shear, so as the unsupported span grows, stress concentrates not by bending the whole roof downward uniformly but by localising into a narrow, roughly cylindrical shear zone that traces the outline of the magma chamber below — the ring fault. Once that ring fault forms and starts slipping, the block of roof rock it encloses, the future caldera floor, can subside as a relatively coherent unit, dropping into the space vacated by the erupted or withdrawn magma while the surrounding rock outside the ring stays largely undisturbed.
roof stability roughly tracked by: support = P_chamber * A_roof
load = weight of overlying rock
as P_chamber falls (magma erupted or withdrawn), support < load
→ shear stress concentrates at the chamber's rim (future ring fault)
→ once shear strength there is exceeded, the roof block slips downward
Piston versus trapdoor collapse
If the ring fault forms as a complete, roughly vertical cylinder and the enclosed block drops more or less uniformly, the result is a piston-style collapse — a broadly circular caldera subsiding as a single coherent unit, seen in classic large calderas like Crater Lake (Mount Mazama) and Krakatoa. If instead the magma chamber is asymmetric, or the ring fault only fully develops on one side, the roof block can hinge downward like a trapdoor, tilting on the side where the fault is weakest or least developed — a trapdoor-style collapse, which produces a caldera that is visibly deeper or more displaced on one side than the other.
Sudden versus incremental collapse
Not every caldera forms in one dramatic event. Some, like Crater Lake roughly 7,700 years ago, collapse essentially in a single climactic eruption over hours to days as an enormous volume of magma is expelled almost at once. Others subside gradually — Kilauea's summit caldera in Hawai'i has grown through repeated episodes of magma withdrawal and incremental slip on its ring faults over years, each individual step modest but the cumulative effect, over enough episodes, building a caldera basin all the same.
Reading the aftermath
Once collapse is complete, the resulting basin, often several kilometres to tens of kilometres across, frequently outlives the eruption that made it by thousands of years, filling with water to form a lake (Crater Lake), hosting later resurgent volcanism as fresh magma reinflates the floor (Yellowstone, Campi Flegrei), or simply persisting as a broad depression that reveals, in cross-section, the size and depth of the magma chamber that used to be underneath — which is exactly why volcanologists study old calderas to estimate how large and how shallow a modern volcano's plumbing might be.
Frequently asked questions
Is a caldera the same thing as a crater?
No. A crater is a comparatively small depression built by explosive excavation or minor collapse around a vent, typically less than a kilometre or two across. A caldera forms by roof collapse into a drained magma chamber and is usually many kilometres across, sometimes tens of kilometres.
Does a caldera always form during a single catastrophic event?
Not always. Some calderas do collapse in one violent, climactic eruption, but others subside gradually over years to decades through repeated smaller episodes of magma withdrawal and incremental slip on the ring fault, without ever producing one single dramatic collapse event.
Why does the roof fail along a ring rather than just sinking evenly?
A roof of rock is far stronger in compression than in shear, so as the void beneath it grows, stress concentrates into a narrow, roughly cylindrical shear zone around the edge of the chamber rather than being spread evenly. Failure along that ring fault is mechanically cheaper than bending the whole roof.
Try it live
Everything above runs in your browser — open Caldera Formation Simulator and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.
▶ Open Caldera Formation Simulator simulation