Deepsea Cores Offer New Clues To Earthquake Cycles

Subduction zones are areas where two plates of the earth's crust meet, with one plate bending and sliding down under the other. The bottom plate extends downward and eventually melts, with some of the molten lava rising to erupt at a distance inland. This process builds volcanic mountain chains such as the Cascades of the Pacific Northwest. In another typical scenario, the two plates of a subduction zone may periodically slip violently, causing some of the greatest earthquakes in the earth's crust.

Paleoseismologists are a lot like detectives, searching painstakingly for geologic clues about events to which there are no living witnesses. For a region stretching 1,000 km along the Pacific northwest coast, U. S. Geological Survey (USGS) and university coastal paleoseismologists have examined remains of trees whose roots died suddenly after being plunged into salt water as a result of coastal land areas dropping down during earthquakes.

Those roots dated back 300 years, coinciding with ages of beach sand layers deposited far inland by earthquake sea waves. These so-called tsunamis surged up estuaries to form thin sand layers in inland bay and coastal lake mud layers. Japanese scientists also have traced a historic tsunami event on their coast, originating with a great earthquake in the Cascadia subduction zone 300 years ago (Jan. 1, 1700).

Marine geologist Hans Nelson (Texas A&M University's College of Geosciences) and paleoseismologist Chris Goldfinger (Oregon State University's College of Oceanic and Atmospheric Sciences) are building on the idea of Canadian scientist John Adams that shaking from Cascadia subduction zone earthquakes should generate sandy density currents. Such turbidity currents flow down deep-sea channels on the continental margin and deposit turbidite sand layers that record paleoseismic events.

"We are now coring into the deep-sea bottom of these channels to collect turbidite layer evidence of past seismic events," Nelson said. "Cascadia is an excellent area to develop this deep-sea paleoseismic record because 7,600 years ago, when Oregon's Mt. Mazama erupted, forming Crater Lake, 100 times more volcanic ash than that of Mt. St. Helens' 1980 eruption was deposited in the Columbia River drainage area."

During his Ph.D. studies in the late 1960's, Nelson originally discovered that this ash was washed into turbidite sand layers of the Cascadia, Astoria and Rogue channels off the Pacific coast. Now this turbidite layer with Mazama ash is a key bed that can be identified in cores from these areas.

"We have found that 12 turbidite bed layers occur for 600 km along the Cascadia margin, above the first bed with Mazama ash," Nelson said. "The best explanation is that great earthquakes on the subduction zone have triggered these events about every 600 years since the Mt. Mazama eruption 7600 years ago. Our deep-sea paleoseismic record now has verified the coastal record, extended it reliably back in time for thousands of years more and shown that the Pacific Northwest faces the hazard of future great earthquakes. "

It's a complicated, expensive and labor-intensive process to develop a reliable deep-sea paleoseismic record and requires significant funding from USGS and the National Science Foundation (NSF).

"We need to be sure of where we are coring in complex channel pathways," Nelson said. "To do that, we're using GIS database technology to plot channel bathymetric information. We collect our data at sea, process it in 30 minutes, incorporate it with the existing data, then use a computerized virtual map to 'fly' through the channels and locate the next place to core."

"Using 20 computer and coring experts, GIS/GPS equipment and a ship with dynamic positioning capabilities, we find the most reliable location to core and then do much of the core laboratory analysis at sea to be sure we have the most accurate turbidite event record at each channel location, then correlate our cores at sea to develop a paleoseismic record," he continued. "Next, we spend several years at our university laboratories to make a high-resolution age analysis for each paleoseismic event to determine as accurately as possible the time variance between each of them.

"The bottom line we want to know for earthquake planning is the minimum time between earthquake events over thousands of years. If we can eventually establish such information, then earthquake hazard planners would know, for example, if we probably have at least 100 years before the possibility of the next great earthquake in the Cascadia subduction zone."

Next summer Nelson, Goldfinger and their team will be coring in multiple channels from Cape Mendocino to San Francisco Bay.

"Our deep-sea record in this new area help us further establish methods that can be used elsewhere in the world to trace the history of earth's active crust and provide more accurate estimates of the minimum time period between great earthquakes on the San Andreas fault in the highly populated northern California region."