progress report on 2014 scec proposal: embedded …...iurp0dqwohwudfwlrq :86bu *36 global ref....
TRANSCRIPT
Progress report on 2014 SCEC proposal:Embedded geodynamic stress model of the southern
California crust
Thorsten W. Becker and Tom Parson
March 25, 2015
ObjectiveSCEC IV has committed to the development of a Community Stress Model (CSM), and suchan effort speaks to core problems of earthquake mechanics; how faults are loaded, and how thestress changes due to individual ruptures affect the overall stress state of the system. In order tounderstand fault system evolution, it is important to have forward models of the background stressstate of the crust, and there are numerous unresolved issues as to the degree of typical heterogeneityof the stress field, or its amplitude.
We proposed to impose the mantle tractions from global mantle circulation computations withregional resolution of ⇠ 20 km (Ghosh et al., 2013; Becker et al., 2014) on a California scale,crustal model with lateral heterogeneities and ⇠ 5 km resolution (Parsons, 2006). The goal isto understand the effects of heterogeneous rheology on vertical force transmission and the likelybackground stress state in southern California. This speaks directly to the 2014 RFP section whichseeks “geodynamic models that explore the coupling of side, gravity, and basal loading to observedgeodetic strain-rates and co-seismically imaged stress”.
ProgressAs proposed, we extracted velocities and tractions from the mantle flow models of Ghosh et al.(2013) and Becker et al. (2014), and visualized the corresponding fields. We analyzed similaritiesand differences between the models and settled on one particular case from Becker et al. (2014) touse as reference. Parsons modified the finite element model of Parsons (2006) to allow prescribingbasal and side traction or velocity boundary conditions, and after some experimentation settled onusing velocity boundary conditions for the stress modeling.
The ANSYS software model includes settling under gravity first, and then velocities fromthe mantle flow model are applied afterwards. Topography, variable thermal gradients, and 3Dcrustal structure/composition are included, as in Parsons (2006). Mantle flow loading is applied
for 100 kyr right now, introducing some arbitrariness in the mantle flow model contribution’simportance.
Figure 1 shows preliminary results. We interpret those to imply that the general approach ofcoupling produces reasonable results, but we are still in the process of analyzing the robustness ofthe results. As implemented, it appears that the topographic and crustal contribution to the totaldifferential stresses is dominant and mantle contributions minor. However, the vertical displace-ment rates, and the resulting horizontal displacement rates over places of mantle upwelling anddownwelling (Figure 1) are significant.
We will continue to explore the coupling parameters, and will strive to construct a “best” com-bined model. The stresses from this model, as well as their components (e.g. only crustal loading,only mantle) will then be submitted to the CSM. We expect further insights into the origin of thecrustal stress field along the plate boundary.
Additional work on intraplate seismicityBecker also used funding from this grant to explore the role of temporal changes in vertical stress-ing rates from mantle flow. Seismicity in the western United States (U.S.) away from the plateboundary is clustered along a roughly north-south trending, meandering “intermountain” belt (Her-rmann et al., 2011, Figure 2). This zone coincides with a transition from thin, actively deformingto thicker, less tectonically active crust and lithosphere, and such structural gradients have beeninvoked to explain seismicity localization (Lowry and Smith, 1995; Levander and Miller, 2012).However, the actual cause of seismicity remains unclear.
Using a comprehensive correlation analysis, we show that results from improved mantle flowmodels which reveal a relationship between rate-change of vertical normal stress from mantleflow (“uplift rates”) and seismicity. This correlation is much stronger than with any of the otherobservables examined (Figure 3).
These findings suggest that active mantle upwellings provide a major, and gravitational poten-tial energy variations a minor, driving force for seismogenic intraplate deformation. Convectiveupwellings are modulated by lithospheric strength heterogeneity to localize seismicity. Our resultson deformation processes are consistent with findings from other mobile belts (Faccenna et al.,2014), and imply that mantle flow plays a significant and quantifiable role in shaping the topogra-phy, tectonics, and seismic hazard within intraplate settings.
A manuscript on this work has been submitted.
2
-800
N1��Ý:
0.0 0.001 0.002 0.003
Rate MPa/yr
-600
-400
-200
0
200
400
600
800
1000
1200
-550 -350 -150 50 250 450 650 850
'LVWDQFH��NP��IURP���Ý1
'LVWDQFH��NP��IURP����Ý:-350 -150 50 250 450 650 850
&UXVWDO�VWUHVVLQJ�UDWH�IURP�PDQWOH�WUDFWLRQ��:86BU�� 7RWDO�GLIIHUHQWLDO�6WUHVV��:86BU��
-350 -150 50 250 450 650 850
'LIIHUHQWLDO�VWUHVV��JUDYLW\�RQO\���:86BU��
Horizontal displacement rate
IURP�0DQWOH�WUDFWLRQ��:86BU��Vertical displacement rate
IURP�0DQWOH�WUDFWLRQ��:86BU��
*36�global ref.
frame
-800
-600
-400
-200
0
200
400
600
800
1000
1200
-550 -350 -150 50 250 450 650 850
'LVWDQFH��NP��IURP���Ý1
'LVWDQFH��NP��IURP����Ý:
8=�PP�\U
-10 0.0 10
-800
-600
-400
-200
0
200
400
600
800
1000
1200
-550 -350 -150 50 250 450 650 850
'LVWDQFH��NP��IURP���Ý1
'LVWDQFH��NP��IURP����Ý:
N1��Ý:
-4000 -2000 0 2000
Elevation (m)
10 mm/yr
0 350 ����
6WUHVV�03D
Figure 1: Preliminary analysis of mantle and topographic (gravitational potential energy) contri-butions to crustal stress in California. Left plots show the stressing rate (top) and vertical dis-placement rate (bottom) from the mantle flow component. Top center and right compare the totalstress with the component that arises from topography loading only, and bottom right shows thehorizontal displacement rates along with an overview map of GPS velocities for the study region.Oblique Mercator projection (see inset map) with fault system superimposed for reference.
3
-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚
0
2
4topo [km]
-0.50
-0.25
0.00
0.25
0.50εm/max(εi)
-0.50
-0.25
0.00
0.25
0.50εm/max(εi)
YS
cBR
CP
SNP
a)-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚
earthquake distribution
b)
0.0
0.2
0.4
0.6φ(quake)
-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚
normal strain-rates
c)
-0.50-0.250.000.250.50
εm [10-15 s-1]
-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚
GPE
d)
-4
-2
0
2
4[TN/m]
-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚
dynamic topography
e)
-2
-1
0
1
2
δzdyn [km]
-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚-125˚ -120˚ -115˚ -110˚ -105˚
30˚
35˚
40˚
45˚
50˚
mantle uplift rates
f)
-1.50
-0.75
0.00
0.75
1.50δvup [mm/yr]
Figure 2: a) Topography, provinces (orange lines) and major geographic features (SNP: SnakeRiver Plain, cBR: central Basin and Range, CP: Colorado Plateau), and focal mechanisms (depth 50 km, colored by the mean horizontal strain divided by maximum, absolute horizontal eigenstrain; blue/red indicate extension/compression, respectively). Gray and black outline beach ballsare from the gCMT (Ekstrom et al., 2014) and SLU (Herrmann et al., 2011) catalogs, respectively(accessed 12/2014). b) Merged earthquake catalog (NCEDC, 2014; Engdahl et al., 1998) alongwith smoothed seismicity density function, f (arbitrary units), as used for correlation analysis.Dark shading in b)-f) indicates regions that are considered to be associated with plate boundarydeformation and are excluded from the analysis. c) Mean, horizontal normal strain-rates fromgeodesy Kreemer et al. (2014) (extension positive). d) Gravitational potential energy estimatebased on integration of our preferred crustal model Becker et al. (2014). e) Dynamic topographyinferred from mantle flow model induced vertical stresses. f) Rate of change of topography frommantle flow (“uplift”) (i.e. temporal change of e)).
4
-0.2 0.0 0.2 0.4 0.6
correlation
stru
ctur
est
ruct
ure
grad
ient
sm
odel
GPS
Moho depthLAB depthelastic thicknessGPEGPEc
Moho depth gradientLAB depth gradientTe gradientGPE gradientGPEc gradientmantle flow stressdynamic topographymantle uplift ratesnormal strain-ratesshear strain-rates
-0.2 0.0 0.2 0.4 0.6
correlation
Figure 3: Regional, linear correlation between seismicity density (Figure 2b), and structural mod-els (Moho (Lowry and Perez-Gussinye, 2011) and LAB (Levander and Miller, 2012) depth, elasticthickness (Lowry and Perez-Gussinye, 2011), and two kinds of GPE models, regular (Figure 2d)and compensated, GPEc), gradients thereof, and geodynamic model predictions (cf. Figs. 2e andf), and geodetic models (Kreemer et al., 2014) (cf. Figure 2c). Correlation is obtained by even-areasampling of regions away from the plate boundary zone (light shading in Figure 2). Gray verticalrange indicates inferred 99.7% significance (±3 standard deviations from random Gaussian fields).
5
ReferencesBecker, T. W., C. Faccenna, E. D. Humphreys, A. R. Lowry, and M. S. Miller (2014), Static and
dynamic support of western U.S. topography, Earth Planet. Sci. Lett., 402, 234–246.
Ekstrom, G., M. Nettles, and A. M. Dziewonski (2014), Global CMT web page, Available onlineat www.globalcmt.org, accessed 12/2014.
Engdahl, E. R., R. D. van der Hilst, and R. Buland (1998), Global teleseismic earthquake relocationwith improved travel times and procedures for depth determination, Bull. Seismol. Soc. Am., 88,722–743.
Faccenna, C., T. W. Becker, M. S. Miller, E. Serpelloni, and S. D. Willett (2014), Isostasy, dynamictopography, and the elevation of the Apennines of Italy, Earth Planet. Sci. Lett., 407, 163–174.
Ghosh, A., T. W. Becker, and E. D. Humphreys (2013), Dynamics of the North American continent,Geophys. J. Int., 194, 651–669.
Herrmann, R. B., H. Benz, and C. J. Ammon (2011), Monitoring the earthquake process in NorthAmerica, Bull. Seismol. Soc. Am., 101, 2609–2625, catalog available online at http://www.eas.slu.edu/eqc/eqc mt/MECH.NA/MECHFIG/mech.html, accessed 12/2014.
Kreemer, C., G. Blewitt, and E. C. Klein (2014), A geodetic plate motion and Global Strain RateModel, Geochem., Geophys., Geosys., 15, doi:10.1002/2014GC005407.
Levander, A., and M. S. Miller (2012), Evolutionary aspects of the lithosphere discontinu-ity structure in the western U.S., Geochem., Geophys., Geosys., 13(Q0AK07), doi:10.1029/2012GC004056.
Lowry, A. R., and M. Perez-Gussinye (2011), The role of crustal quartz in controlling Cordillerandeformation, Nature, 471, 353–357.
Lowry, A. R., and R. B. Smith (1995), Strength and rheology of the western U.S. Cordillera, J.Geophys. Res., 100, 17,947–17,963.
NCEDC (2014), Northern California Earthquake Data Center, Dataset. UC Berkeley SeismologicalLaboratory, doi:10.7932/NCEDC.
Parsons, T. (2006), Tectonic stressing in California modeled from GPS observations, J. Geophys.Res., 111, doi:10.1029/2005JB003946.
6