Preparations and discussions addressed the potential contributions and needs of the relevant disciplines, with a primary workshop goal of identifying commonalities among them. Coordinators were designated for each discipline, as follows:
a. The offshore – Coordinator: Watt
b. Land-level changes - Coordinator: Witter
c. Tsunamis – Coordinator: Nelson
d. Landslides – Coordinator: Perkins
e. Paleo-ground failure (liquefaction, lacustrine slumps) – Coordinator: Sherrod
f. Analog modern earthquakes – Coordinator: Frankel
g. Models of subduction zone earthquake cycle – Coordinator: Pollitz
~40 attendees participated, from offices in WHOI, Santa Cruz, Menlo Park, Golden, Seattle, and Anchorage. In addition to presentations and discussions, ~10 participants shared posters.
2. Workshop Intro
We began with introductions by each participant, and then summaries of the goals of the project and workshop (described in the project proposal). [Sherrod and Gomberg]
3. Disciplinary Presentations about outstanding questions pertaining to recurrence, location and magnitude of past mega-thrust earthquakes in Cascadia, focusing on existing data sources, and major data gaps.
3.1 The Offshore-Structure [Janet Watt]
3.1.1 Outstanding issues and questions:
1) Increasingly, we’ve learned of the importance of splay faults as generators of destructive tsunamis (e.g., they were responsible for most fatalities in the 1964 earthquake). In Cascadia we don’t know much about splay fault connectivity.
2) Is there manifestation of rupture segmentation in the morphology and structure of the overriding plate?
3) We need to connect onshore and offshore fault characterizations. For example, the two Q-fault databases (built by USGS focused onshore and Goldfinger et al. offshore) differ significantly.
4) What are the relationships between margin structure, fluid concentration, and patterns of seismicity?
5) What is the distribution/age of submarine landslides along the Cascadia margin? How does it inform paleoseismic records? The bathymetric data to constrain this are old and inadequate; acquisition of modern multi-beam data would provide 15 m resolution, which in a study off Willapa channel revealed previously interpreted landslides to be Missoula flood deposits.
6) Do we know the sources, volumes and pathways of sediment delivered to the shelf and slope during the Pleistocene/Holocene. How do these change temporally and spatially?
1) High-resolution multi-beam bathymetry (highest priority).
2) Integrated (nested) crustal imagery, possible using high-res multi-channel seismic (MCS) and ultra-high-resolution Chirp surveys.
3) Targeted sub-seafloor sampling to establish the stratigraphic framework.
3.1.3 Current resources, examples.
1) Seamless bathymetry for the entire Cascadia margin with 40-50 m resolution has been acquired by OSU with support from NSF.
2) The Oregon state waters mapping program is 54% complete.
3) Existing 2D MCS (air-gun) Cascadia open-access seismic transects (COAST) cover most of the margin.
4) Experience shows multichannel seismic reflection and CHIRP data permit ‘acoustic trenching’.
3.2 The Offshore-Turbidites [Jason Chaytor]
3.2.1 Goldfinger et al. have acquired ~80 cores between 1999-2015 and others exist, so that the total inventory is ~150 cores. These are all archived at OSU’s and are accessible. Many are in bad shape, simply because they degrade with age. A sensible strategy would involve relooking at a small subset of these (rather than collect new data or look at the entire suite).
3.2.2 Jason notes that the fingerprint of a turbidite is not diagnostic of what initiates, because the physical process that generates it is the same whether along a passive or active margin. He shows an example of turbidite profiles that appear to correlate, yet one is from the Atlantic and the other from the Pacific. Extensive work in the canyons in the Atlantic shows they have numerous flows in them, that appear to occur spontaneously.
3.3 Onshore-Land level changes [Witter, also see presentation]
3.3.1 The coastal record constrains slip variations along strike. Each data point along the coast represents a huge amount of work. Some future directions:
1) Compare coastal subsidence estimates using a variety of methods at a variety of sites (calibrated in Alaska where there have been great earthquakes)
2) Examine coastal evidence to more clearly define the triggering mechanism(s) of T2’s sandy turbidity current.
3) Exploit evidence thresholds for onshore/offshore records to better define sources that triggered muddy turbidity currents.
4) Extend the paleogeodetic record back in time (e.g., Wang et al.).
5) Use modern dating methods to revise earthquake chronologies (e.g., OSL dating, Bayesian age models).
6) Map long-term shoreline change that may follow a future event.
3.3.2 Research questions and considerations (see presentation also):
1) How much did the land subside during past Cascadia earthquakes?
2) What is the threshold size for coseismic subsidence to register in coastal geology?
3) Relative sea-level change of tens of cm can be detected in microfossils, but assigning a seismic origin requires multiple lines of evidence.
4) What do offshore/onshore discrepancies reveal about rupture size, location, and endpoints? Onshore data implies that some earthquakes triggered mud turbidites, but coastal evidence is spotty for rupture around the time of T2.
5) Long or piecemeal ruptures?
6) What barriers to rupture are evident in coastal geology and are they persistent through time?
7) What do offshore/onshore discrepancies reveal about rupture size, location and endpoints?
8) Turbidites are more sensitive to shaking, land level changes to slip. T2 (pre-1700 full-margin rupture inferred from turbidite record) has only spotty evidence in coastal evidence. Thus, there may be reasons why they don’t exist that are informative.
3.4 Tsunamis [Nelson, also see presentation]
3.4.1 Big science questions
1) How high and far inland have tsunamis from the greatest (>M8.6) Cascadia earthquakes extended, how often have they occurred, and have the frequency of such events changed over time?
2) What are the characteristics and probabilities of tsunamis accompanying smaller great earthquakes (<M8.6)?
3) How does local coastal bathymetry, as well as different models of megathrust rupture, affect tsunami inundation at different kinds of sites?
4) How do locations and characteristics of source earthquakes for prehistoric tsunamis, obtained through inverse simulation modeling, compare with paleogeodetically determined models for the same earthquakes?
3.4.2 Opportunities and challenges:
1) Much can be learned from mapping freshwater sites (relative to the offshore, they have better preserved deposits, are nearer limit of inundation, have more datable materials, sediment rates are relatively uniform).
2) Tsunami deposits may be used to calibrate rupture models. Tsunami volumes have been used to estimate slip magnitudes. CT scans of deposits from cores reveal lots more detail than previously possible.
3) Dating now permits both maximum and minimum dates rather than previous maximum-only dates. It is also possible now to work with much smaller amounts of material, even single-grain methods are being used (OSL dates). Shannon Mahan (USGS) does OSL dating – she is very over-subscribed. Within USGS generally all the dating facilities are way over-subscribed. OSL is one of the few ways to date more recent events (e.g., T2 = possible penultimate Cascadia event).
4) Most sites of subsidence do not have tsunami deposits – very sensitive to conditions for deposition and preservation. Usually look for sands, so there also needs to be a source of sand.
5) Other lakes exist besides Bradley that likely contain useful evidence.
3.5 Landslides [Perkins, also see presentation]
1) Very little data exist correlating known subduction zone earthquakes and landslides; e.g. moment vs landslide volume compilations include mostly data from crustal events.
2) The current big challenge is how to add temporal information to spatial distributions of landslides.
3) Current approach being developed correlates landslide roughness to age. It is still unclear how to best quantify roughness.
4) Dating is a big challenge. A few places exist where precise dates might be possible (e.g. where landslide dams killed trees).
3.5.2 Major goals:
1) to be able to estimate recurrence interval from the spatial distribution of coseismic landslides.
2) to quantify site effects and material property controls on stability. Joe Wartmann (UW) has an inventory from Tohoku useful for this purpose. Material properties also affect distribution and roughness. Topographic ground motion amplification also may affect susceptibility.
3.6 Modern Earthquakes [Frankel, also see presentation]
Studies of the Maule and Tohoku earthquakes illustrate the importance of M8 sub-events in great subduction zone earthquakes. These sub-events tend to radiate most of the high frequencies and could give rise to along-strike variations in land-level changes, failures. Simulated Cascadia ruptures that contain sub-events produces irregular pattern of subsidence onshore. Some evidence suggests sub-events may correlate with structural features (e.g as evident in gravity maps). Some important questions pertain to how stationary sub-events are and whether other measurable features permit us to identify them in advance of an earthquake.
3.7 Site affects and submarine slope failure susceptibility [Gomberg]
Seismic data from the Cascadia Initiative experiment offer the potential to test the hypothesis that along-strike variations in turbidite abundance results from variations in site affects that make shaking stronger and/or longer duration in southern Cascadia. Preliminary work clearly indicates that the prism sediments significantly amplify and prolong shaking, relative to motions west of the deformation front.
3.8 Modeling [Pollitz, also see presentation]
The spatial distribution of plate-interface locking likely controls the slip distribution during single and multiple earthquakes. Various types of earthquake cycle models that predict locking patterns exist (block, viscoelastic, dynamic rupture, simulators). Locking models have been generated for Cascadia, but are very poorly constrained offshore. Preliminary calculations indicate that significant improvement would be achieved with offshore GPS measurements.
3.9 Crustal faults [Blakely, see presentation]
4. Database discussion
1) Determining if and how to connect faults across the shoreline should be a key target of the database. Industry data stop well offshore, but potential fields cross the coast. The sparsity of faults onshore in the current USGS Q-faults database reflects both the fact that is hasn’t been updated with new info and that there are lots of areas that haven’t been mapped adequately. An updated fault database should definitely go at least as far inland as the forearc.
2) Many of the sources of relevant data exist in other public databases, and it makes most sense just to point to these and not duplicate them.
3) We should investigate the USGS Community for Data Integration as an avenue for advice and assistance with data integration.
4) The database should be kept highly focused to ensure quality control.
5) A database of Cascadia offshore data compiled in 1984 that is not digital but very complete, and probably worth looking at (Sam Johnson).
4.2 Database components
1) Onshore/offshore fault characteristics
2) Potential field data
4) Cores and onshore boreholes
7) Velocity models (particularly offshore)
8) Diatom sample locations for transfer function studies
9) GPR data
5. Discipline group reports
5.1 Offshore – Santa Cruz & WHOI [Watt]
The CMGP has a Cascadia Recurrence Projecty for FY17. Participants include J. Watt, A. East, D. Brothers, J. Chaytor, M. Walton (post-doc), K. Coble, R. Sliter, J. Conrad, G. Barth, M. Malkowski (post-doc).
5.1.1 FY17 and FY18 Project Goals:
1) Compile/evaluate/synthesize existing multibeam bathymetry, seismic, potential field, and core data.
2) Build synergy with EHP in Seattle (Sherrod and Bennett) and UW (E. Roland) –onshore/offshore Seattle Fault earthquake history/recurrence (“low hanging fruit”).
3) Identify FY18 geophysical survey areas (one each in northern and southern Cascadia).
4) Collect multibeam and MCS sparker data in FY18
5.1.2 FY17 Tasks - Turbidites:
1) Analyze existing geophysical logs from Goldfinger cores.
2) Develop testable hypotheses for evaluating turbidite record in regards to megathrust earthquake recurrence.
3) Assess potential impact of canyon wall mass wasting/landsliding on the turbidite record.
4) Identify targets for detailed geophysical data.
5.1.3 FY17 Tasks - Seismic data:
1) Compile and evaluate existing potential field and seismic data (primarily 2D MCS air-gun data).
2) Reprocess key 2D MCS air-gun seismic lines/surveys along the margin.
3) Contribute to development of broader Project onshore/offshore fault database.
4) Preliminary map of forearc structure, identify probable Quaternary structures. (Definitive classification of Quaternary activity will require high-res seismic data, such as MCS sparker and/or chirp data.)
5.1.4 Synergies: Would be able to identify submarine landslide sources. Would work with the UW to capitalize on use of Thompson, and possibly NOAA too.
5.2 Crustal faults [Staisch]
End goal is to develop a 3-D community fault model that includes the megathrust and onshore and offshore faults, focused on the forearc from Cape Mendocino to Vancouver Island.
5.2.1 Tasks include
1) Compilation of gravity and magnetic data, onshore and offshore seismic, seismicity, offshore cores and onshore boreholes, quaternary active faults.
2) Work requires collaborations with USGS coastal marine, geophysicists, paleoseismologists, academic institutions, Canadians.
3) Focused studies of locations of potential high-frequency sub-event sources, and four areas of onshore offshore structural complexity:
• Heceta Bank
• Nehalem Bank
• Greys Harbor
• Juan de Fuca straights
5.2.2 Deliverables by FY18
1) Preliminary community fault model.
2) input into RSQSIM (earthquake simulator) of Cascadia.
3) Preliminary community velocity model.
4) input into improved ground motion simulations.
5) Improved PacNW contribution to Q-faults database.
6) Potential high-frequency sub-event sources, based on geophysical and geological data.
5.3 Landslides [Perkins]
5.3.1 FY17 Tasks
1) Find and date high-potential co-seismic landslides
a. Participate in Cascadia landslide workshop (includes field component, June 2017)
b. Explore use of dendrochronology and paleo-limnology as dating tools for large slides with dead-stand forests resulting from slide displacement.
c. Compare above dates with isotopic geochronometers.
2) Create new deep-seated landslide inventories for Cascadia using suite of LiDAR data, existing inventories (e.g., DOGAMI), and novel detection techniques. Contribute to the Project database.
3) Derive a relative landslide chronology using morphologic features measured from LiDAR data (in collaboration with UO, UW, DOGAMI).
4) Test inventories versus predicted spatial distributions for co-seismic or storm-driven failures.
5.3.2 Synergies with
1) academic (UO, UW) and state agencies (WA DNR, DOGAMI) operating in concert to understand landslide linkages to subduction zone earthquakes,
2) work with paleo-seismology group on dating strategies for landslide occurrence and re-activation,
3) offshore/onshore records to direct studies and/or corroborate earthquake record,
4) crustal fault group.
5.4 Tsunami & land-level changes [Nelson, Witter]
5.4.1 Personnel and collaborators
1) Principal Investigators: Alan Nelson (EHP Golden), Guy Gelfenbaum (CMG Santa Cruz), Rob Witter (EHP Anchorage), SeanPaul La Selle (CMG Santa Cruz), Bruce Jaffe (CMG Santa Cruz), Shannon Mahan (CGGSC Denver), Scott Bennett (EHP Seattle), Bruce Richmond (CMG Santa Cruz)
2) Non-USGS collaborating PIs (NSF and other funding): Simon Engelhart (University Rhode Island), Tina Dura (Rutgers University, Humboldt State University), Jason Padgett (University Rhode Island), Isabel Hong (Rutgers University), Ben Horton (Rutgers University), Andrea Hawkes (University North Carolina), Kelin Wang (University of Victoria; Geological Survey of Canada), Harvey Kelsey (Humboldt State University), Yuki Sawai (Geological Survey of Japan), Eileen Hemphill-Haley (Humboldt State University), Bre MacInnes (Central Washington University)
3) Others: DOGAMI, Washington DNR and Emergency Management
1) Leverage ongoing analyses of stratigraphic, microfossil, and age data already collected and partly synthesized at Siletz Bay in central Oregon, Nehalem River in northern Oregon, and the Siuslaw River in southern Oregon. Application of modest additional resources (OE and staff time) will allow us to complete sample analyses, finalize interpretations, and publish papers summarizing the earthquake and tsunami history of these estuaries.
2) Leverage ongoing field campaigns directed at deciphering the earthquake and tsunami estuarine stratigraphy of eastern Willapa Bay in southern Washington, Cannon Beach in northern Oregon, the Salmon River in central Oregon, and Humboldt Bay in northern California. Application of additional resources (OE and staff time) will allow us to complete fieldwork in these areas and begin analyzing and modeling the earthquake and tsunami history of these estuaries. The utility of tsunami volume modeling to estimate earthquake magnitudes will be tested at Cannon Beach and Salmon River.
3) Develop a long-term plan to revitalize studies of Holocene tsunami inundation in coastal lakes, particularly at sites where lacustrine histories of tsunami inundation can be compared with adjacent tidal wetland histories of coseismic land-level change.
5.4.3 Questions to be addressed
All our research is directed towards understanding how great megathrust earthquakes at Cascadia, as expressed through coastal land-level changes and tsunami inundation, have varied in magnitude, extent, and recurrence over the past 3000-7000 years. More specific questions include:
1) How do the frequency and locations of smaller great (~M8.2-8.6) earthquakes differ from those of the largest (~M8.6-9.2) earthquakes?
2) What is the threshold of creation and preservation of tsunami/subsidence evidence at different types of coastal sites over different time spans?
3) Is the recurrence of smaller and larger earthquakes on various segments consistent over thousands of years, or do they change over time?
4) How high and far inland have tsunamis from the greatest (>M8.6) earthquakes extended, how often have they occurred, and have the frequency of such events changed over time?
5) What are the characteristics and recurrence of tsunamis accompanying smaller great earthquakes (<M8.6)?
6) How does local coastal bathymetry, as well as different models of megathrust rupture, affect tsunami inundation at a bay site versus an open coastal site?
7) How do locations and characteristics of source earthquakes for prehistoric tsunamis obtained through iterative simulation modeling compare with paleogeodetically determined models for the same earthquakes?
8) Can a more detailed history of subsidence or tsunamis, which might compare more completely with earthquake histories developed from interpretations of the offshore turbidite record, be developed for some time intervals along some parts of the subduction zone?
9) To what extent, if at all, do regional structures along the margin limit the extent of megathrust earthquake ruptures?
5.4.4 Two year deliverables
1) Two or more synthesis papers describing megathrust earthquake/tsunami history at Siletz Bay, Nehalem River, and the Siuslaw River, and the implications of these histories for earthquake and tsunami recurrence and magnitude along the Oregon portion of the subduction zone. For example, comparisons among histories for Nehalem River, Siletz Bay, and Siuslaw River could directly address proposed segment rupture boundaries and zones of low slip among the sites. The paper about Siuslaw River stratigraphy will directly address the limits of resolution of identifying coseismic subsidence in tidal wetland sequences, that is whether or not turbidite events not yet identified onshore might be recognized. The papers might also discuss the ability of the coastal record to record evidence of slip during hypothetical megathrust earthquake subevents. A final in progress paper will summarize an analysis of several hundred old and new radiocarbon ages for evidence of earthquake subsidence and tsunamis, which yields more precise ages for earthquakes of the past 2000 years and suggests that more earthquakes have been dated, at least in northern Oregon, than previously recognized.
2) Completion of field campaigns directed at deciphering the earthquake and tsunami stratigraphy of eastern Willapa Bay, Cannon Beach, Salmon River, and Humboldt Bay, with ongoing data analysis and synthesis, will similarly lead to papers addressing the above and related questions for these additional parts of the subduction zone. In particular, the Salmon River and Cannon Beach studies should show how successful tsunami deposit volume studies are in estimating prehistoric source earthquake magnitudes over the past few thousands of years.
3) A coordinated plan to revitalize coastal research focused on identifying and dating tsunami deposits in lakes, and coseismic subsidence contacts in nearby tidal wetlands (paired study sites). The application of new methods of dating, micropaleontology, and modeling at such paired sites will lead to more accurate and precise histories of the magnitude and recurrence of great megathrust earthquakes and their accompanying tsunamis. A modern coring platform, suitable for sampling Cascadia coastal lakes, will be designed and constructed.
5.4.5 Synergies: The tsunami modeling proposed in this research plan will greatly benefit from plans to acquire high-resolution bathymetry along the Cascadia margin, especially in shallow coastal areas. And evidence from coastal earthquake and tsunami histories may be able to test increasingly sophisticated models of megathrust rupture, such as those incorporating megathrust earthquake subevents and splay faulting on the shelf.
1) Existing resources: Existing FY17 OE and salary for the Golden Cascadia project and for the Cascadia task of CMG tsunami studies. Fieldwork and travel ($8k) and 14C dating ($12k) already listed in the budget for FY17.
2) Needed additional resources: $50k for diatom and foraminiferal studies of coseismic subsidence and coastal environmental change. Collaboration with Tina Dura, beginning an NSF Fellowship at Humboldt State this winter, is critical for the success of earthquake extent and magnitude studies using diatoms. $30k for development and wide application of OSL dating of tsunami deposits.
5.5 Modeling [Harris, Pollitz, Frankel]
5.5.1 Modeling efforts will use geologic data as constraints to predict ground motions and geodetic deformation to possibly eliminate some scenarios.
5.5.2 Ultimately simulations for M~8 earthquakes would be useful for explorations of what sort of deformation they predict regarding the landslide distribution onshore and offshore.
6. Going forward
6.1 Powell Center Proposal
6.1.1 A Powell Center project might provide needed resources that would benefit some of the disciplinary work in this project. PC Projects typically sponsor working groups of 6-20 people (need to be diverse) and last a minimum of 2 years. Funding is granted for two 1-week meetings.
6.1.2 Due date for FY18, Jan. 31 2017.
6.1.3 Lydia Staisch, Janet Watt, Jonathan Perkins, and Rob Witter have agreed to be PIs and will explore submitting a proposal.
6.2 Potential ways to involve outside collaborations
6.3 Next steps
6.4 Potential collaborators (please do not share this with others yet)
USGS – Uri tenBrink, Paul Bedrosian, Eric Geist, Tom Parson, Kenny Ryan, Pat McCrory, Bill Schulz, Kate Allstadt, Roland vonHuene, Bill Stephenson, Dave Scholl, Stephanie Ross, Jonathan Godt, Nate Wood, Tom Pratt, Andy Lamb, Diane Moore, Evelyn Roeloffs, Barbara Bekins, Ben Brooks
University of Oregon -Toomey, Roering, Weldon, Dan Gavin (interested in landslide work, geophysics)
OSU – Ann Trehu, Chris Goldfinger, Chris Goldfinger, Ann Morey, Eric Kirby, Andrew Meigs, Ben Mason, Michael Olson (turbidites, geomorphology)
Humboldt State – paleoseismology, subsidence (Kelsey, Eileen Hempel-Haileys, Jay Patton)
Rutgers – Ben Horton, Tina Dura (sea level studies)
University of Rhode Island – Andrea Hawkes (sea level studies)
Cornell – Katie Kerenen, Geof Abers
UW – Alison Duvall, Joe Wartmann, Paul Johnson, Susan Hautala, Emily Roland, M9 participants
USB – Doug Wilson
Boise State – Lee Liberty
Smith College – Jack Loveless
Portland State University – Adam Booth, Ashley Streig, Rob McCaffrey, Curt Peterson
UC Riverside (Jim Dieterich, Keith Richards-Dinger) – RSQSIM
UVic – Kristin Morrel
UBC – Michael Bostock
WWU – Elizabeth Schermer, Colin Amos
CWU – Brie McInnes, Tim Melbourne, Lisa Ely
Boston University – Christene Rogalla
Berkeley – Richard Allen, Bill Deiterich, tomography, geomorphology
WHOI, Lamont, Scripps – offshore work, gravity
Wyoming – Steve Holbrook
UTIG – Sean Gulick, Nathan Bangs
Lamont – Donna Shillington, Kerry Key
Wash DOT, O DOT, California DOT
Cal Geological Survey – Rick Wilson
DOGAMI – Bill Burns, Ian Madin, Yumei Wang
WA DNR - Tim Walsh, Stephen Slaughter, landslides, tsunamis
NOAA – office of coastal services (Sam Johnson knows)
Olympic National Marine Sanctuary - habitat studies/bathymetry
OSU/NOAA – (Bill Chadwick)
PGC – Kelin Wang, John Cassidy, Roy Hyndemann, John Clague
UNAVCO – Megan Miller
GS Japan – Yuki Sawai