Search for Evidence of Prehistoric Tsunamis and Great Earthquakes on Chirikof Island, Eastern Aleutians
Scientists from the U.S. Geological Survey (USGS) and the University of Pennsylvania (UPenn) visited Chirikof Island, Alaska, in August 2010 to document evidence of prehistoric tsunamis and land-level changes caused by large earthquakes on the Aleutian megathrust, the fault along which the Pacific plate is sliding beneath the North American plate, periodically generating "great" earthquakes of magnitude 8 or higher. Members of the field team were Alan Nelson, Rich Briggs, and Guy Gelfenbaum from the USGS and Simon Engelhart and Tina Dura from UPenn.
Computer models of how tsunamis move (or "propagate") across the ocean predict that rupture on the Aleutian megathrust between Kodiak Island and the Shumagin Islands will cause seafloor displacements that direct tsunamis toward the west coast of the United States, with the maximum wave energy centered on southern California. With support from the USGS Southern California Multihazards Demonstration Project, fieldwork was undertaken on Chirikof Island to (1) demonstrate the potential for identifying and dating prehistoric tsunamis on Chirikof Island and elsewhere in the eastern Aleutians, (2) investigate evidence for megathrust-related land-level changes on Chirikof Island, and (3) establish a scientific and logistical framework for future studies of prehistoric tsunamis and earthquakes in the Aleutian region.
Chirikof Island lies at the eastern edge of a segment of the Aleutian megathrust that ruptured during great earthquakes in 1788 and 1938. The magnitude 8.2 1938 earthquake did not produce a large tsunami, but fragmentary historical records suggest that the 1788 event generated a large tsunami and was presumably of greater magnitude than the 1938 event. Because it lies directly above the megathrust, Chirikof Island likely underwent land-level changes during past great earthquakes. Before this field study, no Quaternary geologic investigations, including studies of historical or prehistoric earthquakes and tsunamis, had been undertaken on the island, although archaeologists did reconnaissance work there in 1963 and 2005.
The field team met in Kodiak, Alaska, on July 28. Bad weather turned back two attempts to fly from Kodiak to the study area. Finally, the weather improved, and three members of the team landed on Chirikof Island on August 4; the other two arrived the next day. Our fieldwork focused on the southwest corner of the 11-by-17-km island, where the terrain includes areas, such as low-lying basins and estuaries, with a high potential for preserving sediment deposited by large tsunamis. Working from a base camp near a small lake on the island's southwest tip, we collected data at seven sites, four of which were studied in detail.
Observations included nearly 40 reconnaissance gouge cores collected along transects both parallel and perpendicular to the shore. A gouge corer with a 5-cm-diameter, semi-enclosed core barrel that was hand-driven into the ground was used to sample vertical sequences of sediment layers and peat (soil made almost entirely of organic matter accumulated in marshy areas) down to several meters below the surface. Depending on the accumulation rates of the peat and sediment layers, these cores could represent hundreds to thousands of years of geologic record to examine for evidence of past tsunamis. The field team also conducted geomorphic mapping and precision elevation surveys, using a Real Time Kinematic (RTK) Global Positioning System (GPS) receiver. The geomorphic mapping helped in the interpretation of past sedimentary environments and how they may have changed during large earthquakes, when land levels may have risen or fallen by several meters. A National Ocean Service (NOS) tidal benchmark tied into the RTK GPS survey provides a common datum for the maps and cores. Finally, four complete Russian cores (13 m total length) were collected and are undergoing detailed radiocarbon and microfossil analyses at USGS laboratories in Golden, Colorado, and UPenn laboratories in Philadelphia. The Russian corer collects 5-cm-diameter cores that—unlike those from the gouge corer—are not compressed or shortened during recovery.
Tsunami deposits consist mainly of sand, commonly with some fragments and thin layers of mud. The settings found to be most useful for identifying these sandy deposits were those with thick sequences of peat, in which the lighter colored sand deposits are relatively easy to spot. Possible tsunami deposits were observed and sampled at several sites on Chirikof Island. One site (TR), a broad valley 11 m above sea level, contained freshwater peat more than 5 m thick. At least two conspicuous sand layers, possibly of tsunami origin, are present in the peat, with a layer of tephra (volcanic ash) directly beneath the lower sand layer. At least one of the sand layers is normally graded (the sand grains vary gradually in size from coarser at the bottom to finer at the top), as are many tsunami deposits. To understand the spatial distribution of the sand layers, we described 10 reconnaissance cores and collected two Russian cores for dating, microfossil analysis, and lithologic analysis and description.
At a second site (RR) at a lower elevation, freshwater peat extends up a narrow valley 4 to 10 m above sea level. In the upper reaches of the valley, a stream and its tributaries have added land-derived sediment to the marsh sediment in the valley floor, and tsunami deposits are difficult to distinguish from the stream deposits. In the valley's lower reaches, thick sand deposits prevent deep coring and complicate interpretations of depositional history. Between these zones, however, are five to nine distinct sand beds, which were probably deposited by tsunamis or storm surges. We described eight reconnaissance cores and collected one Russian core here.
At a third site (FM), beneath what may have been a marsh fringing a former estuary, we found pairs of peat and mud layers capped by a bed of sand with an erosive base. This stratigraphy may record multiple cycles of uplift and subsidence related to earthquakes, followed by tsunamis or breaching of the beach berm (a low ridge of sediment built up on the beach by waves). Several reconnaissance cores were described and one Russian core collected here.
Because they are shaped by waves, beach berms are good indicators of the position of the shoreline. Along Chirikof Island's southwest coast, we mapped cobble berms that are now beyond the reach of the waves and serve as markers of changing shoreline positions, possibly associated with land-level changes during and between earthquakes. The relationships between relict shorelines and archaeological and historical sites should provide important information about late Holocene uplift of the area.
Mapping, topographic profiles, outcrop descriptions, and interpretation of remote-sensing imagery of the west coast of Chirikof Island indicate a probable eolian (windblown) origin for the landforms and thick sand layers mantling the older surfaces. The absence of peats or other soils hinders identification of tsunami deposits along most of the island's west coast.
We intended to leave Chirikof Island by floatplane on August 15, but bad weather delayed our departure until August 17, when all five scientists finally made it safely back to Kodiak. Gelfenbaum returned to the island 2 days later to retrieve gear that had been left behind. Most of our days on Chirikof Island were foggy, rainy, or windy.
As long as field parties plan for weather-related delays, fieldwork in the Aleutians can be quite rewarding: As we anticipated, freshwater marshes and coastal geomorphology on Chirikof Island preserve evidence of possible prehistoric tsunamis and megathrust-related land-level changes. The data we collected in 2010 will provide a baseline for further work in the region and a basis for comparison with emerging records of prehistoric earthquake-induced land-level changes and tsunamis on Kodiak Island, 125 km to the northeast.
Promising areas for future study include islands near the Aleutian Trench, such as Simeonof Island in the outer Shumagin Islands, Sanak Island (about 200 km west-southwest of the Shumagins), and, possibly, sites on the south-facing coast of the Alaska Peninsula. Constructing a history of great megathrust earthquakes and their accompanying tsunamis in this part of the Aleutian Arc will help us better understand the tectonic behavior of this plate boundary—and other, similar boundaries—as well as assess earthquake and tsunami hazards.
in this issue:
Prehistoric Tsunamis and Great Earthquakes