Along the Copalis River on Washington’s outer coast, dead western red cedars stand where a living forest once grew. Their trunks rise from a tidal marsh, while living woodland occupies higher ground nearby. The name “ghost forest” describes those surviving remains. The mystery was how a substantial forest came to occupy ground repeatedly reached by salt water, and whether its death recorded a sudden disaster rather than a slow change in the shoreline.
The explanation emerged by combining evidence that could not supply the same answers on its own. Buried soils showed a change in land level. Sand recorded powerful flooding. Tree rings narrowed the death of selected cedars to a particular winter. Historical documents in Japan supplied a tsunami date that could be tested against the American evidence. Together they reconstructed a great Cascadia earthquake in 1700, centuries before an instrumental earthquake network could have recorded it.
A drowned forest rather than drifting timber
The standing trees matter because they remain rooted. They are not logs washed ashore and later mistaken for an ancient woodland. Western red cedar is especially durable, allowing parts of the forest to survive long after the environmental change that killed it. Below coastal marshes, researchers also found former land surfaces covered by younger tidal sediment. The combination preserves both vegetation and the ground on which it lived.
The principal source for this account is the second edition of The Orphan Tsunami of 1700, a US Geological Survey monograph by Brian Atwater and colleagues. Its photographs and discussion of Copalis distinguish the local ghost forest from comparable evidence at other estuaries. That distinction matters: a regional reconstruction draws on several sites, and observations from Willapa Bay or the Columbia River should not silently become observations made at Copalis.
One early interpretation illustrates the problem. James Graham Cooper described drowned cedars at Shoalwater Bay, now Willapa Bay, after visiting in 1853–54. He proposed gradual sinking into quicksand. This was a regional observation, not a nineteenth-century Copalis earthquake report. The later investigation had to establish whether slow processes could account for the buried coastal landscapes.
Why a sudden fall in land level fitted better
A forest can disappear as a shoreline changes gradually, so dead trees alone do not establish an earthquake. The monograph sets out additional tests. Some affected spruce trees retained broad outer rings immediately before death, indicating that they had been growing vigorously. That pattern was harder to reconcile with a long decline under increasingly unsuitable conditions. Delicate plant remains beneath tidal mud also indicated rapid burial rather than centuries of gradual replacement.
There was a modern comparison. In Alaska’s 1964 earthquake, ground at Portage dropped and ordinary tides subsequently reached former forest land. Saltwater flooding killed trees and new sediment covered earlier soils. Researchers could compare the known consequences of a witnessed earthquake with the older Cascadia remains. Portage was an analogue for the process; it was not evidence that every dead coastal forest necessarily has the same origin.
At Copalis and other relevant sites, the explanation became an abrupt lowering of the coast. Ground formerly high enough to support forest entered the tidal zone. Salt water then affected the roots repeatedly. The enduring environmental change, rather than simply a wave striking a tree once, explains why the forest could not continue growing in its former position.
The sand beneath the tidal mud
Sandy deposits supplied another part of the reconstruction. The monograph describes coastal sand sheets that thin inland, contain marine microscopic remains and may preserve evidence of successive flooding pulses. Their position above former soils and below later tidal mud links a major inundation with the change in land level. Sediment relationships provide a sequence of events rather than an isolated dramatic-looking layer.
A storm or a tsunami from a distant earthquake might carry sand inland. By itself, however, either proposal would leave the associated permanent lowering of the local coast unexplained. The value of the Cascadia interpretation lies in its ability to account for both inundation and subsidence. This is a stronger argument than treating the presence of sand as a unique signature that could have only one possible source.
The roots that preserved the missing year
Dating the trees was more difficult than counting rings visible on a weathered trunk. Outer wood can decay or be stripped away. An apparently final ring may therefore predate the tree’s actual death. Early samples collected in 1987 did not necessarily preserve the critical outside edge. Researchers needed surviving wood that retained the growth sequence right up to the bark.
In 1996 they excavated bark-bearing roots attached to already studied trunks. Patterns of wider and narrower rings could be matched against living trees whose annual sequence was known. This cross-dating connected the dead specimens with a calendar, rather than relying on a simple estimate of their age. The evidence came from several Washington estuaries, including the Copalis River area.
The result needs its original qualification. Of eight trees discussed in the monograph, seven completed their 1699 growth and died before the next growing season began in May 1700. One survived until 1708. It is inaccurate to turn that finding into eight identical death dates, or to imply that every surviving trunk at Copalis was individually dated. The selected trees bracketed a season; they did not themselves identify 26 January.
The research also corrected an earlier inference. Atwater and David Yamaguchi initially interpreted evidence as indicating death during a growing season. Gordon Jacoby’s later comparison with trees affected by the Alaska earthquake showed how survival after immersion could complicate that reading. A revised seasonal interpretation was part of the investigation, not a detail to remove from an otherwise tidy discovery story.
Why Japanese records mattered
A tsunami can reach a distant coast without an earthquake being felt there. Japanese records described damaging flooding in January 1700 without a local earthquake to explain it. That made the event an “orphan” tsunami: its effects were documented, but its source was not locally apparent. Researchers compared those accounts with potential Pacific sources and with the developing Cascadia evidence.
The North American tree-ring interval fitted the Japanese event. Tsunami travel calculations and historical times supplied a more precise association with 26 January 1700 on the American side of the Pacific. This division of evidence is essential. Japanese documents did not describe the Copalis trees, and the trees did not preserve an exact date or hour. Their independent constraints became persuasive when assessed together.
The monograph’s preferred reconstruction was an earthquake around magnitude 9, with a modelled range of 8.7–9.2. Those values are estimates derived from geological evidence and tsunami modelling, not instrument readings made in 1700. The same caution applies to a rupture extending roughly 1,000 kilometres: it is a reconstruction of an unobserved fault movement.
What remains under discussion
In a 2021 modelling study, Diego Melgar examined whether several earthquakes could explain some evidence usually combined into one full-margin rupture. His models left a rupture sequence plausible without disproving a single great event. The distinction concerns the regional earthquake reconstruction; it does not turn saltwater-killed trees into evidence of a haunting.
The appropriate outcome for the Copalis natural mystery is explained, with those limits stated. Abrupt subsidence and the resulting tidal inundation account for the forest’s death. Dating and tsunami evidence associate it with the 1700 Cascadia disaster. Exact rupture extent, earthquake size and the possibility of a wider sequence require more cautious language than the basic explanation of why trees now stand dead in a marsh.
Location and present-day access
The map pin is an approximate Copalis Beach locality reference, rounded from the US Census Gazetteer. It is not the position of a sampled cedar, a launch ramp or an authorised landing. The case concerns the nearby Copalis River estuary; a community reference avoids giving a false impression of a surveyed incident point.
The Washington Geological Survey’s access guidance describes viewing by water only, using a kayak, canoe or paddle board, and identifies public river access from small State Route 109 pull-offs. It warns about changing tidal currents and private property. That guidance does not confer permission to walk through the forest or collect specimens. Nearby park access and river conditions should be checked separately before a visit.
Sources and further reading
Atwater and colleagues, The Orphan Tsunami of 1700, second edition (2015), especially printed pages 14–19 and 92–99: modern subsidence comparison, Copalis photographs, sand deposits, tree-ring tests and tsunami reconstruction. The text is public domain; the volume also identifies separately protected illustrations.
Diego Melgar (2021), Was the January 26th, 1700 Cascadia Earthquake Part of a Rupture Sequence?: later modelling and the limits of a single-rupture interpretation.
Washington Geological Survey, Copalis Ghost Forest: regional interpretation and water-access guidance. USGS, Dead Red Cedar Forest: the photograph’s location, creator and explicit public-domain designation.
Related UKPX case: the Racetrack Playa sailing stones, another natural mystery investigated through physical evidence.
Visit Copalis Ghost Forest (water viewing)
Washington DNR describes water-only viewing by kayak, canoe or paddle board, with public river access from small SR109 pull-offs. Tidal currents change; do not trespass or assume permission to land among the trees. The locality pin is not a launch point.
Official visitor information → · Access information checked 2026-11-13. Check current opening and booking details before travelling.
Share an experience here →Experiences at Copalis Ghost Forest (water viewing)
First-person accounts shared by visitors and reviewed before publication. They are separate from the researched case above.
No visitor reports have been published here yet.