Cucapah Earthquake of Baja California in Mexico
Total Page:16
File Type:pdf, Size:1020Kb
LETTERS PUBLISHED ONLINE: 31 JULY 2011 | DOI: 10.1038/NGEO1213 Superficial simplicity of the 2010 El Mayor–Cucapah earthquake of Baja California in Mexico Shengji Wei1*, Eric Fielding2, Sebastien Leprince1, Anthony Sladen1,3, Jean-Philippe Avouac1, Don Helmberger1, Egill Hauksson1, Risheng Chu1, Mark Simons1, Kenneth Hudnut1,4, Thomas Herring5 and Richard Briggs6 The geometry of faults is usually thought to be more 118° W 116° W 114° W complicated at the surface than at depth and to control the GCMT Salton Sea SAF 34° N initiation, propagation and arrest of seismic ruptures1–6. The fault system that runs from southern California into Mexico is This study California a simple strike-slip boundary: the west side of California and 46 Arizona Pacific platem Sonora 32° N Mexico moves northwards with respect to the east. However, 33° N m ¬1 the Mw 7.2 2010 El Mayor–Cucapah earthquake on this fault CaliforniaGulf of system produced a pattern of seismic waves that indicates a Elsinore f 7 F4 far more complex source than slip on a planar strike-slip fault . ault Here we use geodetic, remote-sensing and seismological data iver to reconstruct the fault geometry and history of slip during this CucapahSierra fault lley ° earthquake. We find that the earthquake produced a straight 32.5 N F2 Mexicali va USA Latitude 120-km-long fault trace that cut through the Cucapah mountain ColoradoMexico r range and across the Colorado River delta. However, at depth, Sierra El Mayor f the fault is made up of two different segments connected by F1 ◦ Laguna Salad a small extensional fault. Both segments strike N130 E, but 10 cm Indiviso fault dip in opposite directions. The earthquake was initiated on the km 32° N F3 ault connecting extensional fault and 15 s later ruptured the two a 01020 main segments with dominantly strike-slip motion. We show that complexities in the fault geometry at depth explain well Gulf of the complex pattern of radiated seismic waves. We conclude California that the location and detailed characteristics of the earthquake 116° W 115° W could not have been anticipated on the basis of observations of Longitude surface geology alone. The El Mayor–Cucapah earthquake that occurred on 4 April Figure 1 j Setting of the El Mayor–Cucapah earthquake. The inset shows 2010 produced extensive liquefaction in the Colorado River delta historical M > 6:5 earthquakes (red circles) and a simplified plate boundary area and in the Mexicali and Imperial valleys, and numerous (red line). Yellow and white circles denote surface ruptures determined rockfalls occurred in the Sierra Cucapah. This is the largest from correlation of optical and SAR images. Circles show seismicity earthquake to have struck the southern California and northern (M > 2:5) 4 months before (red) and 11 days after (blue) the earthquake, 18 Baja California, Mexico area since the Mw 7.3 Landers earthquake relocated with the double-difference method . Left top corner, moment of 1992 (ref. 8). The GCMT (global centroid moment tensor) of the tensor derived from this study and GCMT. Labelled rectangles show the mainshock reveals a double-couple component corresponding to fault geometry used in the inversion. The focal mechanism derived from the 19 a scalar moment of 7:28 × 10 N m (Mw 7.17), with a significant first 15 s of teleseismic P-waves is shown in orange, with the epicentre non-double-couple (CLVD) component (2:4 × 1019 N m; Fig. 1; reported as the red star. Arrows show horizontal coseismic displacements ref. 7). The mainshock occurred where the system of continental measured at PBO GPS stations (data in white, with 95% confidence parallel right-lateral strike-slip faults including the San Andreas, ellipses, and synthetic in red). SAF, San Andreas fault. San Jacinto and Elsinore faults connect with a system of transform faults and active spreading centres in the Gulf of California to the area was the Laguna Salada fault, a right-lateral normal oblique fault 9 south (Fig. 1 and Supplementary Fig. S1). This fault system forms bounding the Sierra Cucapah to the west. It accommodated a Mw 7.1 the plate boundary in southern California, where the Pacific plate earthquake in 1892 (ref. 10; Supplementary Fig. S1). moves northwestwards with respect to North America at about We synthesize the earthquake data using modern methods 46 mm yr−1 (Fig. 1, inset). The main active fault recognized in the in seismology, tectonic geodesy and remote sensing (global 1Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, California 91125, USA, 2Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109, USA, 3Geoazur, Observatoire de la Côte d’Azur, Université de Nice–Sophia Antipolis, CNRS, IRD, Valbonne, 06103 Nice Cedex 2, France, 4United States Geological Survey, Pasadena, California 91106, USA, 5Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA, 6United States Geological Survey, Golden, Colarado 80401, USA. *e-mail: [email protected]. NATURE GEOSCIENCE j ADVANCE ONLINE PUBLICATION j www.nature.com/naturegeoscience 1 © 2011 Macmillan Publishers Limited. All rights reserved. LETTERS NATURE GEOSCIENCE DOI: 10.1038/NGEO1213 a a 5 Southeast Northwest SAR SAR azimuth offset data SPOT offset data 4 3 32.5° N 2 1 Along-strike offset (m) 0 ¬10 0 10 20 30 40 50 60 70 80 90 100 110 120 Distance along strike (km) Colorado River b 0 20 40 60 80 100 120 Latitude SPOT 0 32 22 32 32 22 10 32 22 32 32 Strike = 335° 22 N Depth (km) ° F4 2 Strike = 132 F3 F2 Strike = 312° 10 32° N (m) 0 Depth (km) 0 100 200 300 400 500 600 F1 Strike = 355° Slip (cm) ¬2 20 km c 0 20 40 60 80 100 120 S 0 115.5° W 115° W 10 Longitude Depth (km) b d ) F4 ¬1 4 F4 F3 N m s F2 18 2 F1 32.5° N 10 × F2 Moment rate ( 0 01020304050 Time (s) Figure 3 j Slip distribution and rupture history. a, Surface slip measured F1 from serial profiles across the fault trace based on the SAR (blue) and SPOT (red) images. b, Cumulative slip (arrows show slip vectors, and Latitude colour coding shows amplitude) and isochrons of the seismic rupture. The rupture times are given relative to the onset of slip at the epicentre. Note N F3 the large displacements around 17 and 27 s on F2. c, Comparison of 2 relocated aftershocks projected on the faults with slip distribution 32° N (m) (smoothed). d, Source time function showing the time evolution of released 0 moment rate. The contribution of each fault segment is colour coded. ¬2 S 20 km mainshock did not rupture the Laguna Salada fault but rather two other faults, the Borrego and Pescadores faults within the 115.5° W 115° W Sierra Cucapah, which had been mapped but not recognized Longitude to be active10 (Supplementary Fig. S1). Our data also reveal a major strike-slip segment that extends from the epicentre to Figure 2 j North–south surface displacements measured from subpixel the southeast, across the Colorado River delta. This part of correlation of optical and SAR images. a, Map of near-field coseismic the mainshock rupture, which has been verified in the field ground displacement measured from subpixel correlation of SAR amplitude and named the Indiviso fault11, occurred along basement faults images25 and optical SPOT images26,27 acquired before and after the beneath the sedimentary deposits of the Colorado River. Thus, the earthquake. The SPOT measurements are over-printing the SAR azimuth southeast surface trace does not coincide with previously identified offsets. b, Prediction from the preferred model. The colour scale is blue active faults having obvious geomorphic expression, such as the (respectively, red) for southward (respectively, northward) displacements. Laguna Salada fault or the Cañada David detachment. Rather, the The red star shows the location of the epicentre. The black dots delineate earthquake ruptured along a complex set of existing, less active the fault trace as determined from the SAR and optical image correlations. faults, illustrating the ongoing process by which the slip along The thin black rectangles show the idealized fault segments used in the Elsinore fault connects to the transform plate boundary in the modelling. the Gulf of California. Overall, the location and focal mechanism of the earthquake are positioning systems (GPS), interferometric synthetic aperture consistent with right-lateral slip along the right-lateral transform radar (InSAR), subpixel correlation of optical satellite images, plate boundary fault system (Fig. 1). However, the large non- and synthetic aperture radar (SAR)). See the Supplementary double-couple component of the moment tensor indicates a Information for details on data and methods. The remote-sensing substantial component of normal faulting. The modelling of the data reveal an almost linear and continuous fault trace extending first 15 s of the teleseismic waveforms (Supplementary Fig. S2) over about 120 km from the northern tip of the Sierra Cucapah indicates that the earthquake actually initiated as a normal event to the Gulf of California, with right-lateral slip of about 2 m (Fig. 1). This observation, together with the clear asymmetry of on average (Fig. 1). These data indicate that the 4 April 2010 surface strain seen from the correlation of the optical (SPOT) 2 NATURE GEOSCIENCE j ADVANCE ONLINE PUBLICATION j www.nature.com/naturegeoscience © 2011 Macmillan Publishers Limited. All rights reserved. NATURE GEOSCIENCE DOI: 10.1038/NGEO1213 LETTERS Table 1 j Moment tensor solutions of GCMT and this study. Mrr Mtt Mpp Mrt Mrp Mtp M0 MDC MCLVD Unit (N m) GCMT −2:49 −5:94 8.43 0.56 −0:14 −0:86 7.28 2.84 1019 Static inversion −1:42 −6:42 7.84 0.11 −1:05 −1:55 9.91 7.50 2.11 1019 Joint inversion −2:03 −6:07 8.10 −0:23 −1:28 −1:62 9.90 7.36 2.14 1019 The moment tensor solution for the finite-fault slip models is obtained by summing up the contributions of each subfault.