DAY 1 (Nov. 6, 2012) Technical Presentations

Transcription

1 Presentation Sam Johnson Jeanne Hardebeck DAY 1 (Nov. 6, 2012) Technical Presentations Key points/action items : Consider vertical rates as transitory and localized Review Mann publications and consider sidewall rip-out process. o There is a rip-out that is being translated; need to ask Vicki, Janet about defining limits. Consider asymptotic fault branching geometries (Casmalia, Shoreline, Los Osos) for dynamic rupture modeling Consider changes to lateral slip rate along strike at fault intersections (e.g., addition to south from Piedras Blancas, subtraction to south from Los Osos). OADC algorithm plus consideration of first motions constrains preferred Hosgri fault dip to range p. 1

2 Janet Watt from 76 to 89 degrees east. No dip less than 70 is acceptable fit. Shoreline fault dip degrees southwest. : Jeanne can send map of where the OADC plane is located and where it projects to surface. Jeanne indicates up-dip projection is not preferred location for source-site distance. Reconcile/compare Johnson and Watt fault trends with Estero bay seismicity Track progress by Jeanne to find an improved way to consider microseismicity patterns using focal mechanism data. Hosgri may be vertical to steeply east dipping Langenheim: gravity data indicate east dip to 3 km depth Shoreline fault bends to north as it approaches Hosgri p. 2

3 Hans AbramsonWard Doug Hamilton Action Items Test San Luis Bay fault location and dip along magnetic low offshore of Rattlesnake. Consider likely challenges to getting a well constrained model with a blind fault. Alternative northern ends to Hosgri fault source hampered by sparse data; lack of study (Comment emphasized by S. Johnson) Point Buchon fault is the preferred northward extension of Shoreline fault Evaluate possible soft segment boundaries along Hosgri-San Gregorio-San Andreas faults. Keep track of S. Johnson plan to evaluate high-resolution seismicreflection data along Big Sur coast. Prefers NE dipping thrust faulting as principal driver of Irish Hills uplift p. 3

4 Marcia McLaren Considers likely surface trace across San Luis Obispo Bay, blind and impinging on Shoreline, then emergent at N40W/Pt. Buchon East fault and out along Islay Shelf. Considers very recent development of Shoreline fault and resulting truncation of westward continuation of San Luis Bay thrust. Considers rock shelf out to Hosgri to have been uplifted by San Luis Bay thrust, only stopped recently. Consider fault location, geometry and slip rate uncertainty Consider tests for fault location, uplift rate boundary predictions, Consider limitations of using microseismicity locations to define faults (e.g., Hauksson 2010 paper, Hardebeck fault style approach in SSC-2 presentation) Patterns of seismicity beneath Irish Hills do not constrain preferred p. 4

5 Bill Lettis Presentation Russ Graymer fault planes Consider limitations of microseismicity to define faults (e.g., Hauksson paper, Hardebeck SSC-2 presentation) Key models (1) transpressional shear, or (2) NE-directed crustal shortening, inform characterization of fault geometry and slip rate Test models with stress/strain directions inferred from GPS velocities and focal mechanism inversions DAY 2 (Nov. 7, 2012) Technical Presentations Key points/action items Magnetic and gravity constraints critical for viable cross-section solutions 2 episodes of deformation; Q faulting on San Miguelito per faulted Squire p. 5

6 Consider Edna and San Miguelito Favors current shortening as seen in finite strain Test with new mapping; seismic data Test with marine terraces for vertical deformation Test with velocity field Consider forward modeling with new data Vicki Langenheim Lithosphere publication Southern end of Hosgri has undisrupted gravity low; gravity does not favor continuation of Hosgri to south Given Ma initiation of fault (J. Clark) long-term rates higher than present but same pattern Data cannot show/explain km drop in slip rate past Pt. Buchon. Difficult to reconcile strike-slip p. 6

7 deficit by shortening; lacks good alternative Consider Lithosphere paper; implications for tectonic model Chris Sorlien Consider publications; block models Consider transfer of slip to west onto Southwest Channel-Ferrelo fault? Consider alternative concepts for slip stepping west onto Santa Lucia bank; Develop tests for hypothesis Obtain data from Chris / locations of subsidence interpreted from Sta Maria Kathryn Hanson Slip rate for San Gregorio may be 3 to 4.5 mm/yr (between Frijoles and Seal Cove sites) p. 7

8 San Simeon terraces best defined by Osos terraces 0.7 to 2.6; 200 ky Consider uncertainties of basal gravel of airport offset 1.2 to 2.9; Ka Consider Piedras Blancas onland strand for activity evaluation Upper limit of 6 mm/yr considered too high for geologic data what additional data could be gained to keep it? Lower limit of 0.5 mm/yr considered too low not enough data to support losing 0.5 mm/yr by time we get to plant but deferring judgment until more studies performed Explore uncertainty in onset, shear direction of Piedras Blancas shortening Tim Dawson UCERF3 description of how geodetic modeling and geologic p. 8

9 rates are considered Consider datasets and approaches; be weary of direct use of results Chuck DeMets Jessica Murray We can trust that he did it right Pacific Plate motion uncertainties from islands off Mexco; what is start of Pacific plate margin 3 mm/yr approximate for Hosgri Get data from Mexican islands; is pacific plate deforming from thermal contraction. May subtract 1.5 mm/yr? Consider maximum limit perhaps by looking at total west of Rinconada? Challenge to use GPS to resolve closely spaced faults mm/yr range for Hosgri p. 9

10 Preferred model? Residuals suggest NE-SW shortening (fits most models) Evaluate 1857? More modeling? What is preferred? Recommends FEM Peter Bird 2 mm/yr dextral Hosgri+Rinconada+West Huasna; 1.2 Hosgri alone; result dependent on assumed Euler pole 2-3 mm/yr shortening; result prefers shortening for faults east of Hosgri. Refine the FE grid Modify fault dips Test several compilations of GPS data Test Euler Poles Reconsider geologic prior rates p. 10

11 Refine Geometry Optimize weights Vary mu (rigidity); concern about criteria Hans AbramsonWard Steve Thompson Resolves two platforms; interprets different uplift rates Uplift rate boundary coincides with San Luis Bay fault zone; crosses Shoreline fault Uncertainty within model? Epistemic? Slip rate on Shoreline uncertainties Vertical rate of zero Seismicity rate a minimum? Rattlesnake hypothesis a minimum for fault (due to continuation of South segment) Shoreline fault may not show a lot of accumulated slip comment from John Stamatakos. p. 11

12 Bill Lettis Presentation Jeanne Hardebeck Alternative models must fit firstorder observations of coastal terrace uplift rates, localized basins, trench data Slip rate ranges consider observations for vertical and lateral Incorporate Edna Valley results Day 3 (Nov. 8, 2012) Technical Presentations Key points/action items Defined segment, multi-segment fault ruptures as specific predictions of rupture endpoints. Critique of PGE (2011) Shoreline Report segmentation of Shoreline, Hosgri, and lack of Hosgri-Shoreline linked rupture. Historical examples Jeanne considers GI methodology to be mature; incorporating constraints is not (Agreement by M. Page). p. 12

13 David Schwartz Segmentation is not useful as simplifying assumption ACTION ITEMS Consider effects of segmentation hot spots and anti-hot spots Consider grand inversion-type approach for a sub-set of area; simplified fault locations away from site M. Page considers main uncertainty is epistemic uncertainty going into the model Glenn UCERF3 can fit global average; need to consider agreement at site. Consider linked rupture of Hosgri and Shoreline; Hosgri and others; backwards rupture as well Segmentation is real Review of Static, Dynamic criteria Wasatch and Denali examples reviewed Consider Denali-Totschunda intersection; fault branch with sub- p. 13

14 Craig depolo equal slip rates ACTION Consider getting more MRE timing, slip data Review paleoseismic data on SAFN, SG, and SS faults Consider uncertainty in actual segmentation point Considers behavioral information, slip rate information most critical. Geometry plays a role. Morgan Page: Regarding slip rate change considerations, consider looking at frequency of linked ruptures with Hosgri from perspective of Shoreline and Los Osos slip rates. Consider primary, or core fault segment concept (Carpenter et al., 2012) Consider length scales of separate fault zones; details about rupture ends are less significant Consider Biasi, Wesnousky papers p. 14

15 Glenn Biasi Consider multiple rupture scenarios Consider single-event displacements Consider importance of paleoseismic data. Emphasized value of Wesnousky s (2006) maps Consider UCERF3 appendices expanding on Wesnousky (2006) plus rupture endpoint data. 75% anchored at least one end; 33% at both ends. Implementation of this as a constraint in UCERF3 is non-trivial, but may be used as a check; Glenn suggests it needs to be explored. Morgan Page considered as possible improbability constraint; not required Size of step may not matter up to maximum (~5 km; data being compiled by student at UNR) Consider dynamic rupture informing question. ~5 km limit may scale with p. 15

16 Discussion seismogenic thickness Strike-slip faults don t jump faultto-fault very often (18%), but preferentially jump to other strikeslip faults; less frequently jump SS to reverse. Consider whether affected by rupture direction. Stress Matters Consider insights from historical reverse faults Segments may anticipate multievent ruptures Consider Coulomb for geometric compatibility; need rupture direction and amplitude assumptions Consider Lozos et al. (2011) dynamic rupture results. Consider scales of simulations considered WILLS: RE: UCERF2-style: For A faults only, not B. Morgan uncertainty is healthy with lack of data. Steve W. Consider aspect ratio p. 16

17 Norm Abrahamson (Recurrence rate) Tom Rockwell (Recurrence rate) with widths of Jackson counterpoint with understanding of width Thompson: Concern that quality of mapping of historic ruptures plus retrospective lens may suggests some stepovers may have thrugoing faults that would be discoverable with more detailed mapping; i.e., limitations or uncertainties in defining a step predictively. Consider incorporating uncertainty now so we can be within uncertainty bounds later reject it later. Consider not only assuming Poisson. Presenting data from San Jacinto fault n= and 0.7 COV typical of CA studied faults. Rare 0.2 (Turkey, not applicable due to lacks networks of faults). p. 17

18 Dave Jackson (recurrence rate) Morgan Page Glenn low COV of southern Alpine fault; basically supporting Tom s result. Recommends high COV for DCPP Curse of Retrospective confirmation and its applicability Consider Kagan and Jackson, 1991 GJI Always give randomness its respect. Evidence for and against Characteristic behavior of faults GR Hypothesis constant B; rate may change through time Consider Page et al. (2011). Observations on southern SAF o PE opinion these are changes in rate; informed by ETAS simulations ETAS simulations suggest we are likely to underestimate long-term rate Insights from U3 GR implied on faults; doesn t work p. 18

19 Norm Abrahamson well fit state-wide. Not enough budget for background. If it is correct, need 20-30% increase in seismicity rate (historic rate too low?) or lower b-value Slip COV of characteristic branch ~0.5 median; mean higher for GR Branch Consider exploring ETAS to inform us about Characteristic MFD Multi-fault ruptures needed for buldge problem Speculates (insight, not study result) that order of magnitude changes in a-value between a snapshot versus mean (for a fault; including aftershocks); consider Landers. Consider Hecker et al as insight to deviation from GR at high magnitudes. Data fit Y&C, exponential models that limit Mmax; don t favor larger Mmax with GR. p. 19

20 Dave Jackson Paleoseimologists: Address the Consider bias in the observations Dave J. comment; directed at geotypes: Consider site effects as affecting low COV observation ; near-surface phenomenon Stochastic model Big earthquakes prefer faults, but some occur off of known faults. Blend faults and smoothed seismicity Consider geodesy, paleoseismology Consider applicability for sitespecific studies, calibrate parameters as appropriate. Consider site data. Probed, challenged about larger earthquake possibly occurring off of faults. Consideration: really big earthquakes on faults are rare; really big earthquakes not on recognized fault still rarer. Examples: New Zealand; El Mayor- p. 20

21 Bruce Shaw Tom Hanks Cucapah. Review M-A relations considered for UCERF Shaw (2009) considers Third regime saturation of width effect; constant stress drop. Consider surface slip information and an alternative slip-length scaling for slip rate balancing. Allows for possibility of deeper slip. Considers relation as appropriate for dip-slip faults Consider adding epistemic uncertainty; Ellsworth B should not be precluded? Uncertainty in W important consideration; bias in W estimates from pre-event seismicity. Norm concern: Want to know it works at a median value; can add standard deviations. Don t want models that don t fit the median values. Notes slip data suggest better p. 21

22 behavior (x3) at large magnitudes; slip data suggest more variability at shorter lengths (x10). Does not favor Ellsworth B for magnitude 5-7 range. Review Consider Stirling and Goded report for GEMS; probe Mark for his rationale for selection of relations Consider whether use of Ellsworth B should be correlated with higher stress drops in GMC Consider Diablo model 1.25 Log A + 3.3, A > 537 km^2 Recommends need for updated dataset Dave Jackson consider Kagan s work at rupture zones of global M>7 events; linear fit between length to cube root of Moment p. 22