Estimating fault slip rates in the Cascadia region of North America using joint geologic-geodetic block modeling Richard Styron, Tiegan Hobbs, Zach Lifton, Nick Harrichhausen, Murray Journeay [email protected] Seismological Society of America April 2021 

Overview • OR, WA, BC: Complex tectonics, incompletely known faults – Slip rate data lacking throughout – Very few known active faults in BC • New fault mapping and block modeling to characterize faults regionally • Results may inform targeted investigations, consider for hazard models 

Cascadia/ NW US • Distributed deformation (extensional + dextral) in US Cordillera • Transpressional deformation in US PNW • Cascadia subduction sinistral dextral reverse normal strike-slip unknown GEM Global Active Fault DB 

Cascadia/ W. Canada • Concentrated deformation on Queen Charlotte transform • Transpression in AK • Thrusting in Mackenzies (not shown) • What (if anything) is going on in BC? sinistral dextral reverse normal strike-slip unknown GEM Global Active Fault DB 

Cryptic Faults • BC has many old fault systems – Clear in topography, optimally (?) oriented – Activity levels unknown – Minor geodetic motion Vancouver Vancouver Island 

Cryptic Faults Kamloops 

Block models to characterize faults: Why? • Block models provide context that (independent) geologic fault studies do not – Integration of different data sets, types that may be sparse – Kinematic consistency, fault linkages – Deformation budget • Can be thought of as fault network models 

Block models to characterize faults: Data integration • Block models can integrate many types of deformation data – Geologic slip rates: neotectonic/geomorphic, x-sec balancing, … – GNSS, InSAR, VLBI, optical surveys, … – Paleomagnetic data – Paleoseismology (discrete earthquakes)? (not done yet AFAIK) • 2 data points from *any* combo of data types can fully characterize deformation all along block boundary (and faults) • Have to assume deformation rates constant through time (data duration) or explicitly model temporal changes 

Block models to characterize faults: Kinematic consistency • Block rotation is correct representation of rigid* body motion on Earth’s surface • Block boundary (fault) rates and horizontal kinematics completely specified by block rotation pole+rate (Euler vector) • Consistent integration of multiple block / fault linkages (“path integral constraint”) *main assumption here (though internal deformation can be accommodated) 

Block models to characterize faults: Deformation budget • Geodesy provides important bounds on budget of regional deformation • Deformation can be locally inaccurate but has to sum to far- field strain • Works against strain underestimation (missing faults) and persistent bias in geologic slip rates (i.e. pre-2007 Tibet) Bormann et al., 2016 

Block modeling: Faults to PSHA with new GEM methods • Block model yields relative motions of blocks -> kinematics and slip rates of block-bounding faults • Mapped faults used both as block boundaries, and as seismic sources in PSHA (same geometries) • Other seismic source types (e.g., smoothed instrumental seismicity) account for distributed seismicity in PSHA model • Internal/ off-fault strain could be incorporated as well 

Block model resolution • Blocks and faults must be mapped at ‘fault source’ resolution (e.g., 1:100 k) • Geologist’s intuition helpful for connecting faults, understanding strain partitioning Bormann et al., 2016 

Example: China • Test case and first use of software and workflow • 966 faults, 163 Quaternary geologic slip rates, 3364 GNSS velocities • 20,000 km fault mapping, 259 blocks Styron, in prep 

China block model Styron, in prep 

• Slip rates estimated for ~1000 faults (dip slip and strike-slip), from large, complex orogen • Broadly compatible, though mostly lower, than geologic estimates – Early geol. too high – Block model against strain localization? 10 mm/yr 1 mm/yr 0.1 mm/yr Styron, in prep China example: Slip rate results 

Seismic Hazard from Block Model • Faults and background seismicity used to make PSHA model • Hazard in west dominated by fast- slipping faults • Hazard in east dominated by distributed seismicity, hard to see slow faults Thomas Chartier (GEM) 

Block model to PSHA caveats • Data sparsity and quality – Block model helps w/ sparse data, but problems remain • Non-fault block boundaries: – Seismically active? – Distributed, smaller magnitude quakes, or just as large? • Fault branches / splays? • Temporal changes in regional deformation – Geodetic strain transients – kyr-scale earthquake clusters? 

‘Cascadia’ faults, blocks • Current model stretches from Alaska, NWT south to California, Nevada • Fault mapping, block geometry very tentative in places – Work in Progress 

GNSS catalogs • Good data coverage in US, sparse data in Canadian Cordillera • MIDAS, Global Strain Rate Model catalogs considered – More available • May need more processing (co/post-seismic effects, common mode corrections) Blue: Kreemer et al. 2014 Purple: Blewitt et al. 2016 

OR / WA • Mapping in eastern WA, OR and Puget Sound region well constrained – Previous work – Desert • Oregon Cascades, Coast Ranges, NW WA less clear geol. slip rate (USGS database) 

ID / MT • Basin and Range normal faults mapped, lots of slip rates • Some guesswork in central Idaho, S. Montana • Will test hypotheses for deformation in and around Snake River Plain geol. slip rate (USGS database) 

S BC map • Initial mapping of some major/obvious faults • Coast Ranges not yet mapped • Provisional block geometry (some placeholders) • Lots of ground to cover 

C BC map • Potential for more rapid deformation in Yukon, NWT • Little detailed work so far 

Subduction Interface • Subduction interface represented by triangular mesh – Two current models • Slip (+ locking) will be resolved on each triangle in block inversion • Inversion currently unstable, may need regularization Graham et al. 2018 Slab 2 

Conclusions: Cascadia block modeling • Integrated model of NW US faults, Canadian faults, Cascadia subduction interface – First of its scale and resolution (?) • Model can identify structures for more in-depth characterization (i.e., paleoseismology) • Test possible impact of faults on seismic hazard and risk (particularly in Canada) 

Conclusions: General • Block models provide coherent framework to integrate geologic and geophysical data types • Block / fault network model provides context to characterize poorly known structures • With higher-resolution fault + block mapping, block model results used directly as input for PSHA model 

Please attribute to the GEM Foundation with a link to www.globalquakemodel.org Except where otherwise noted, this work is licensed under https://creativecommons.org/licenses/by-nc-nd/3.0/