In the Shasta Valley of northern California, hundreds of rounded hills rise from an otherwise flat plain. For decades, geologists puzzled over them. Some read them as small volcanic vents. Others saw glacial moraines left by ice that never quite reached this far.

The answer arrived in May 1980, when the north flank of Mount St. Helens collapsed. Within minutes, a debris avalanche carried roughly 2.5 cubic kilometres of rock down the North Fork Toutle River. It left behind a field of mounds strikingly similar to those at Shasta. The Shasta hills were not vents at all. They were the wreckage of an ancient collapse of Mount Shasta itself, one of the largest known landslides on Earth.

That moment of recognition captures something essential about landslide deposits. They are not just piles of rubble. Their internal structure, composition and age record how a slope failed, what pushed it over the edge, and how often it has happened before. Read carefully, they become a forecast written in the past tense.

The first clue lies in the surface. Debris avalanche deposits typically form a hummocky landscape: a chaotic field of mounds and closed depressions, sometimes stretching tens of kilometres from the source. Each hummock is often a piece of the original mountain. It was carried as a coherent block and set down as the moving mass spread and thinned.

Cut into those hummocks and the internal architecture tells more. Geologists distinguish a block facies, where large fragments of the parent edifice preserve their original layering, from a matrix facies, a mixed mass of crushed rock. Within blocks, a distinctive pattern called jigsaw fracturing appears: clasts shattered in place, their pieces still fitting together like a puzzle. This points to intense fracturing during transport but little internal mixing. The mass moved as a rapid granular flow rather than a turbulent, water-rich one.

Slumps look very different. Rotational failures move along a curved, spoon-shaped surface, so the displaced material tilts back toward the slope it came from. These back-rotated blocks, sometimes called toreva blocks, preserve bedding that dips into the hillside. A compressional bulge at the toe usually marks where movement stopped. Slow, coherent motion leaves coherent geometry.

Geometry also measures mobility. The ratio of fall height to runout length, known as the H/L ratio, is around 0.6 for small rockfalls but falls to 0.1 or less for very large avalanches. Albert Heim drew attention to this after the 1881 Elm disaster in Switzerland, and the mechanism behind it remains debated. A low ratio flags a mass that travelled far beyond what simple friction predicts. That is exactly the behaviour hazard planners most need to anticipate.

Takeaway

A deposit's structure is a recording of its motion: coherent blocks speak of sliding, shattered but unmixed clasts speak of rapid granular flow, and long runouts warn that size changes the rules.

Every slope failure has two kinds of cause: the slow preparation that weakens a slope and the final push that tips it over. Identifying that push from a deposit alone is difficult. Geologists therefore rely on context, composition and timing together.

Earthquake triggers often reveal themselves through synchronicity. When many landslides across a region yield the same age, a single shaking event is the most economical explanation. Lake and fjord sediments help here. Strong shaking can trigger simultaneous slides around a basin, leaving several turbidites or mass-transport deposits at the same stratigraphic level. Nearby liquefaction features, such as sand dykes, strengthen the case.

Rainfall-triggered failures leave a different signature. They tend to be shallower and more water-rich, and they cluster in time with wet periods. Deposits that grade into debris flows, with matrix-supported textures and signs of saturation, point toward hydrological triggers. When landslide ages cluster during known pluvial intervals in regional climate records, the link becomes compelling.

Volcanic collapses often carry the clearest fingerprints. Long-lived hydrothermal systems convert strong rock into weak clay minerals, and altered material is common in collapse deposits. When an avalanche contains fresh juvenile magma, such as the dacite fragments from the growing cryptodome at Mount St. Helens, intrusion is implicated directly. A horseshoe-shaped amphitheatre at the source completes the picture. Yet equifinality remains a constant caution: different triggers can produce similar deposits, so conclusions rest on converging lines of evidence rather than any single feature.

Takeaway

A trigger is rarely proven by one observation; it is inferred where timing, composition and setting all point the same way.

A single landslide tells you a slope can fail. Several dated landslides tell you how often. Establishing recurrence is where deposit studies shift from description to hazard science.

The toolkit is broad. Radiocarbon dating of wood or soil buried beneath a deposit gives a maximum age for the event, while organic material accumulating on top provides a minimum. Cosmogenic nuclide exposure dating measures isotopes such as beryllium-10 or chlorine-36 that build up in boulder surfaces exposed to cosmic rays, revealing when blocks came to rest. Tree rings in damaged or tilted trees can pin recent events to a single year, and volcanic ash layers offer precise regional time markers.

Offshore Norway, the Storegga Slide moved a vast volume of continental slope sediment about 8,150 years ago and sent a tsunami onto the coasts of Scotland. It sits within a longer sequence of older slides in the same area. Dating that sequence revealed failures broadly linked to glacial cycles, as rapid sediment loading built unstable slopes. Recurrence there follows a rhythm set by climate rather than a fixed clock.

But the record is always incomplete. Older deposits are eroded, buried or overprinted by younger ones, so the past tends to look quieter than it was. Small events vanish faster than large ones. Careful workers therefore treat recurrence estimates as minimum frequencies and attach honest uncertainties. A gap in the record may reflect lost evidence rather than genuine stability.

Takeaway

A landslide record is a survivor's archive; its silences are often erased chapters, so every recurrence estimate is best read as a minimum.

The hummocks of the Shasta Valley waited hundreds of thousands of years to be understood. Their meaning became clear only when a modern collapse supplied the key, a vivid demonstration of Lyell's principle that present processes illuminate past ones.

Together, deposit architecture, trigger evidence and dating turn landslide deposits into archives of slope behaviour. Structure reveals how material moved. Composition and timing reveal why. A sequence of ages reveals how often.

That knowledge reaches beyond Earth. Long-runout landslide deposits sprawl across Valles Marineris on Mars, testing our models where gravity and water behave differently. Every slope is, in a sense, between failures. The rocks simply tell us where it sits in that cycle.