Seismicity and structure of the Orozco transform fault from ocean bottom seismic observations

In this thesis, seismic waves generated by sourcesranging from 2.7 kg shots of TNT to magnitude 5 earthquakesare studied in order to determine the seismic activity andcrustal structure of the Orozco transform fault. Most of thedata were collected by a network of 29 ocean bottomseismometers (OBS) and hydrophones (OBH) which were deployedas part of project ROSE (Rivera Ocean Seismic Experiment).Additional information is provided by magnetic anomaly andbathymetric data collected during and prior to ROSE and byteleseismic earthquakes recorded by the WWSSN (WorldwideSeismic Station Network).In Chapter II, the tectonic setting, bathymetry andteleseismic history of the Orozco Fracture Zone aresummarized. Covering an area of 90 x 90 km which includesridges and troughs trending both parallel and perpendicularto the present spreading direction (approximately east-west),the bathymetry of the transform portion of the fracture zonedoes not resemble that of other transform faults which havebeen studied in detail. A detailed study of one of thelargest teleseismic earthquakes (mb=5.1) indicates rightlateral strike-slip faulting with a strike parallel to thepresent spreading direction and a focal depth of less than 5km. The moment sum from teleseismic earthquakes suggests anaverage fault width of at most a few kilometers. Because theteleseismic earthquake locations are too imprecise to definethe present plate boundary and the magnetic anomaly data aretoo sparse to resolve the recent tectonic history, morequestions are raised than are answered by the results in thischapter. These questions provide the focus for the study ofthe ROSE data.Chapter III contains an examination of the transferfunction between seafloor motion and data recorded by the MITOBS. The response of the recording system is determined andthe coupling of the OBS to the seafloor during tests at twonearshore sites is analysed. Applying these results to theROSE data, we conclude that the ground motion in the absence of the instrument can be adequately determined for at leastone of the MIT OBS deployed during ROSE.Hypocentral parameters for 70 earthquakes, calculatedfor an assumed laterally homogeneous velocity structure whichwas adapted from the results of several refraction surveys inthe area, are presented in Chapter IV. Because of the largenumber of stations in the ROSE network, the epicentrallocations, focal depths and source mechanisms are determinedwith a precision unprecedented in marine microseismic work.Relative to the assumed model, most horizontal errors areless than ±1 km; vertical errors are somewhat larger. Allepicenters are within the transform region of the OrozcoFracture Zone. About half of the epicenters define a narrowline of activity parallel to the spreading direction andsituated along a deep topographic trough which forms thenorthern boundary of the transform zone (region 1). Mostwell determined depths are very shallow (<4km) and noshallowing of activity is observed as the rise-transformintersection is approached. In fact, the deepest depths(4-10km) are for earthquakes within 10 km of theintersection; these apparent depth differences are supportedby the waveforms recorded a t the MIT OBS. First motionpolarities for all but two of the earthquakes in region 1 arecompatible with right lateral strike-slip faulting along anearly vertical plane striking parallel to the spreadingdirect ion. Another zone of activity is observed in thecentral part of the transform (region 2). The apparenthorizontal and vertical distribution of activity is morescattered than for the first group and the first motionradiation patterns of these events do not appear to becompatible with any known fault mechanism. No difference canbe resolved between the stress drops or b values in the two regions.In Chapter V, lateral variations in the crustalstructure within the transform region are determined and theeffect of these structures on the results of the previouschapter is evaluated. Several data sources provideinformation on different aspects of the crustal structure.Incident angles and azimuths of body waves from shots andearthquakes measured at one of the MIT OSS show systematicdeflections from the angles expected for a laterallyhomogeneous structure. The effect of various factors on theobserved angles and azimuths is discussed and it is concludedthat at least some of the deflection reflects regionallateral velocity heterogeneity. Structures which can explainthe observations are found by tracing rays through threedimensional velocity grids. High velocities are inferred atupper mantle depths beneath a shallow, north-south trendingridge to the west of the OBS, suggesting that the crust underthe ridge is no thicker, and perhaps thinner, than thesurrounding crust. Observations from sources in region 2suggest the presence of a low velocity zone in the centraltransform between the sources and the receiver. That thepresence of such a body provides answers to several of thequestions raised in Chapter IV about the hypocenters andmechanisms of earthquakes in region 2 is circumstantialevidence supporting this model. These proposed structures donot significantly affect the hypocenters and fault planesolutions for sources in region 1. The crustal velocitystructure beneath the north-south trending ridges in thecentral transform and outside of the transform zone is determined by travel time and amplitude modeling of the datafrom several lines of small shots recorded at WHOI OBH.Outside of the transform zone, a velocity-depth structuretypical of oceanic crust throughout the world oceans is foundfrom three unreversed profiles: a 1 to 2 km thick layer inwhich the velocity increases from about 3 to 6.7 km/secoverlies a 4 to 4.5 km thick layer with a nearly constantvelocity of 6.8 km/sec. A reversed profile over one of thenorth-south trending ridges, on the other hand, indicates ananomalous velocity structure with a gradient of 0.5 sec-1throughout most of the crust ( from 5.25 km/sec to 7.15km/sec over 3.5 km). A decrease in the gradient at the baseof the crust to about 0.1 sec-1 and a thin, higher gradientlayer in the upper few hundred meters are also required tofit the travel time and amplitude data. A total crustalthickness of about 5.4 km is obtained. An upper mantlevelocity of 8.0 to 8.13 km/sec throughout much of thetransform zone is determined from travel times of large shotsof TNT recorded at MIT and WHOI instruments. "Relocations" ofthe large shots relative to the velocity model assumed inChapter IV support the conclusion from the ray tracing thatresults from region 2 may be systematically biased because oflateral velocity heterogeneity whereas results from region 1are not affected.In the last chapter, the results on crustal structureand seismicity are combined in order to define the presentplate boundary and to speculate on the history of the presentconfiguration.

Thisresearch was supported by the Office of Naval Research, undercontracts N00014-75-C-0291 and N00014-80-C-0273

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy and the Woods Hole Oceanographic Institution February 1982