Three-Dimensional Geologic Map of the HaywardFault Zone, San Francisco Bay Region, CaliforniaBy G.A. Phelps, R.W. Graymer, R.C. Jachens, D.A. Ponce, R.W. Simpson, and C.M. WentworthPamphlet to accompanyScientific Investigations Map 3045U.S. Department of the InteriorU.S. Geological Survey1

ContentsAbstract. 3Scope and Purpose of this report . 31. Earthquake studies. . 32. Ground water studies. 33. Mineral hazards studies. 44. Education and scientific inquiry. 4Introduction. 4Geologic Background . 6Mesozoic terrane complexes . 6Paleogene unroofing and overlap . 8Neogene transpression and development of the San Andreas Fault system . 8Methods . 9Hayward Fault. 12East Hills Area . 12Cenozoic sedimentary deposits and tertiary volcanic deposits . 13San Leandro gabbro of Ponce and others (2003). 14Geology west of the Hayward Fault. 14Description of Map Features . 15Description of Map Units. 16Description of Map Structures . 17Hayward Fault. 17Calaveras Fault. 18Rodgers Creek Fault. 18Palomares-Miller Creek-Moraga-Pinole Fault System. 18Chabot Fault . 19Other San Leandro gabbro-bounding faults. 19Cretaceous/Tertiary (KT) faults east of the Hayward Fault. 19San Pablo Bay Fault. 19Cenozoic (Ce) faults east of the Hayward Fault . 20Mesozoic (Mw) faults west of the Hayward Fault . 20Data . 20Acknowledgments . 24References Cited . 24Tables1. Density-depth function for Cenozoic basin deposits and Cenozoic volcanic rocks .142. Description of data and algorithms .21Figures1. Generalized map of the major faults of the Hayward Fault zone .52. Generalized terrane map of the Hayward Fault zone (modified from .73 .Isostaticresidualgravityanomalymapofthestudyarea. .104. dix I. Data Files .31Appendix II. Introduction to the Basic EarthVision 3D Viewer Tools .322

THREE-DIMENSIONAL GEOLOGIC MAP OF THE HAYWARD FAULT ZONE, SAN FRANCISCO BAYREGION, CALIFORNIABy G.A. Phelps, R.W. Graymer, R.C. Jachens, D.A. Ponce, R.W. Simpson, and C.M. WentworthAbstractA three-dimensional (3D) geologic map of the Hayward Fault zone was created by integrating theresults from geologic mapping, potential field geophysics, and seismology investigations. The mapvolume is 100 km long, 20 km wide, and extends to a depth of 12 km below sea level. The mapvolume is oriented northwest and is approximately bisected by the Hayward Fault. The complexgeologic structure of the region makes it difficult to trace many geologic units into the subsurface.Therefore, the map units are generalized from 1:24000-scale geologic maps. Descriptions ofgeologic units and structures are offered, along with a discussion of the methods used to map themand incorporate them into the 3D geologic map. The map spatial database and associated viewingsoftware are provided. Elements of the map, such as individual fault surfaces, are also provided ina non-proprietary format so that the user can access the map via open-source software. The sheetaccompanying this manuscript shows views taken from the 3D geologic map for the user to access.The 3D geologic map is designed as a multi-purpose resource for further geologic investigationsand process modeling.Scope and Purpose of this reportThe purpose of this new map is to compile the best available data on: 1) identity and subsurfacedistribution of bedrock units, 2) sedimentary deposits, and 3) geologic structures at a regional scale,and to integrate the data into a 3D map. The 3D nature of the map allows the modeled geology tobe incorporated into physically based process models that are used to predict the behavior ofgeologic systems. The 3D map also offers interesting new possibilities for visualizing complexgeologic relations with the help of increasingly powerful computer software. Potential applicationsof the 3D map include:1. Earthquake studies.The intensity of earthquake shaking has been shown to depend strongly on the distribution ofgeologic units. For example, the intensity increases within sedimentary basins. Simulations of theshaking produced by the great 1906 San Francisco, California earthquake were based on a 3Dvelocity model made by assigning seismic velocities to the various rock units in a 3D geologic mapof the San Francisco Bay region (see, for example, Hartzell and others, 2006; McPhee and others,2007). Accurate seismic velocity models are required to precisely locate earthquake epicenters,which can be used to map fault location, geometry, and activity. A superior velocity model couldbe developed using the 3D geologic map that shows the regional distribution of geologic materials.2. Ground water studies.In urban areas, such as the San Francisco Bay region, ground water contamination from surfacewater pollution is a major problem (Howard, 1997). Because the flow of groundwater pollutants isgoverned by the porosity and permeability of materials, and these materials can often be related to3

geologic units described in 3D geologic maps, contaminant flow and transport modeling can beimproved by incorporating the subsurface distribution of the geologic units.3. Mineral hazards studies.Certain naturally occurring minerals can be hazardous to human health and the environment whendisturbed by human development. Because certain minerals, such as asbestos and serpentinite, areassociated with specific geologic units, geologic maps that show the subsurface distribution ofthose units can be used to regionally delineate the potential hazard.4. Education and scientific inquiry.A 3D geologic map allows the user to explore the shapes of the geologic units and structures, inisolation or as an integrated whole. Geologic features can be viewed from different angles,individually or in aggregate, and slices through the map can easily be constructed. This allows theuser to grasp complex spatial relationships quickly and easily. Thus, 3D geologic maps are anexcellent teaching aid and interpretive tool. This 3D map can be used by the academic communityand by the public to understand the geology of the Hayward Fault at a regional scale.Uses of this map database are limited by compilation scale and content, therefore the map is notintended for site-specific studies. Refer to the fault maps of the California Geological Survey( for site-specific studies of fault activity.IntroductionThe Hayward Fault (fig. 1) is one of three major faults of the San Andreas Fault System that hasproduced large (M 6) historic earthquakes in the San Francisco Bay area; the other major faults arethe San Andreas and Calaveras Faults. The Hayward Fault roughly bisects the San Francisco Bayarea, a densely populated region of about seven million people. A severe earthquake, approximatelyM7 (U.S. Geological Survey, 2007), was generated by this fault zone in 1868, and a number ofother major events have been captured in the paleoseismic record over the last 2,000 years(Lienkaemper and others, 2003). The Hayward-Rodgers Creek Fault zone (fig. 1) is considered tobe the most likely source of a major earthquake in the San Francisco Bay area, with a probability of27 percent for an earthquake M6.7 over the next 30 years (Working Group on CaliforniaEarthquake Probabilities, 2003). The Hayward Fault has the highest potential earthquake loss in thearea due to the high concentration of manmade structures and lifelines in the region.The purpose of this report is to describe the geology of the Hayward Fault zone and surroundingarea and to provide views of the fault zone in three dimensions. The 3D geologic map is based oninterpretations of geologic and geophysical data, modeling algorithms, and expert-driven decisionsthat were used to construct a geometric model of the inferred geology. The 3D map extendsapproximately 10 km on either side of the Hayward Fault, to a depth of 12 km.4

RodgerekresC tulFaSears PointlayToultFaHaywSan PabloBayaroPinltaudFPointPinoleleConcealed faultMajor faults of the Hayward Fault zoneultFaraMogaultFaN0BerkeleyMarinHeadlands010 milesllereerCCalaveraywHaArea of mapultFaFOLICAFremontIARN34 NardSanFranciscoArea ofFigure 1ltHayward0 100 200 kmults FaaukFrooult42 NybFatonot-SabrasChcoSanLeandro116 W120 mealoisncran F Bayk-PSaSanFrancisco38 20 kmMiOakland124 W10Los AngelesSan JoseFigure 1. Generalized map of the major faults of the Hayward Fault zone (modified from Graymer andothers, 2005).5

The San Francisco Bay region includes nine counties surrounding San Francisco Bay and threemajor metropolitan areas: San Francisco, Oakland, and San Jose. The Hayward Fault demarcatesthe eastern side of the topographic San Francisco Bay plain; the East Bay hills extend along the eastside of the fault (fig. 1). Within the map area, the metropolitan division of Oakland-FremontHayward houses a population of approximately 2.5 million people (U.S. Census Bureau, 2006).The map digital files are described in Appendix I. Instructions for viewing and querying the 3Dgeologic map are described in Appendix II.Geologic BackgroundMesozoic terrane complexesA tectonostratigraphic terrane is a fault-bounded body of rock that has a geologic history thatdiffers from those of adjacent rock bodies (Jones and others, 1983), implying some amount ofdisplacement from its original position in order to juxtapose dissimilar rock bodies. Associatedterranes (that is, terranes that share some common aspects of their geologic history) are groupedinto terrane complexes. The Mesozoic rocks in the study area are composed of two terranecomplexes (Blake and others, 1984; 2000; Graymer and others, 1996), the Franciscan Complex andthe Great Valley complex (fig. 2). Each terrane complex comprises several related terranes (Blakeand others, 1984).The Great Valley complex is composed of at least five terranes, two of which crop out in the studyarea (the Healdsburg terrane and the Del Puerto terrane). Each terrane in the Great Valley complexis made up of Middle-Late Jurassic ophiolitic rocks (called collectively the Coast Range ophiolite),with or without Late Jurassic silicic volcanic rocks, and with one exception, is depositionallyoverlain by Late Jurassic and Cretaceous sedimentary strata (called cumulatively the Great Valleysequence). Although the origin of the Coast Range ophiolite remains controversial (see Dickinsonand others [1996] for a summary of three different theories of Coast Range ophiolite origin), weherein accept the origin outlined in Graymer (2005) and Graymer and others (2006); the CoastRange ophiolite formed in the forearc of an oceanic island arc outboard of the North Americancontinental margin, and the island arc and the forearc-related ophiolite were subsequently accretedto North America during the late Kimmeridgian Nevadan Orogeny (approximately 151-153 Ma).The earliest Great Valley sequence strata vary between terranes, but the bulk of Great Valleysequence sediments are thought to derive from a continental margin arc associated with the SierraNevada batholith (Ojakangas, 1968; Dickinson, 1970).The Franciscan Complex is composed of many terranes, plus mélange (a chaotic mix of rocksderived from multiple terranes). In the study area, four coherent terranes (Yolla Bolly, AlcatrazIsland, Marin Headlands, and Novato Quarry terranes), as well as mélange, crop out (Blake andothers 1984; Graymer, 2000). Each coherent terrane in the Franciscan Complex is made up of oneor more types of seafloor rocks (mid-ocean ridge basalt, seamount basalt, pelagic limestone and(or) chert, turbidite-derived sandstone and shale) accreted and amalgamated in a subduction zone(Ernst, 1969). The emplacement in the subduction zone has resulted in a low T/high Pmetamorphic overprint on some Franciscan Complex terranes ranging from incipient developmentof low T/high P minerals to complete metamorphism to glaucophane schist. Many FranciscanComplex terranes, however, have undergone only low-grade metamorphism, evidently nottraveling deep into the subduction zone before being accreted to that overlying crust. For most of6

Cenozoic OverlapCzsCzsCzvCzvCenozoic sedimentary rocksCenozoic volcanic rocksGREAT VALLEY COMPLEXfsrCzvHealdsburg terraneSedimentary rocks (Great Valleysequence)Coast Range ophiolite and keratophyreCoast Range ophiolite, San Leandrogabbro memberghsghsfyb CzvghoghoghogghogghsfnFRANCISCAN COMPLEXfnfnfaifaifmhfmhfybfybfsrfsr( )CzsfnUndifferentiated Great Valley complexCoast Range ophioliteguoguoPointPinolefsrDel Puerto terraneSedimentary rocks (Great Valleysequence)gdsgdsSan PabloBayfsrmixedfsrNovato Quarry terraneAlcatraz Island terraneMarin Headlands terraneYolla Bolly terraneMélangeparantheses indicate unit is coveredConcealed faultMajor faults of the Hayward Fault ghooiscncran F Bay116 W120 42 N1010 milesghog(fai)SanLeandro(guo)gdsgdsHayward0 100 200 km20 km0SaSanFrancisco124 WNghs(fmh)Area of mapghsSanFrancisco(guo)CA38 (fyb)LIRNCzsIA34 NfsrFremontFOArea ofFigure 2CzsfsrfybLos AngelesSan JoseFigure 2. Generalized terrane map of the Hayward Fault zone (modified from Graymer and others, 2005).7

the history of Franciscan accretion, the terranes have accumulated in the subduction zone’s westernmargin of the North American Plate (a process that may be continuing today at the active Cascadiasubduction margin to the north). However, Graymer (2005) showed that the age of the oldestFranciscan Complex metamorphic rocks suggest that these rocks were subducted into the trenchassociated with the island arc outboard of the North American subduction margin, and were carriedto North America along with the island arc and forearc as described above.The subduction emplacement of the Franciscan Complex means that Franciscan Complex terraneswere originally all structurally under the Great Valley complex (although this position has beenlocally reversed by later tectonics as explained below). This juxtaposition allowed some admixtureof materials of the two terrane complexes. Specifically, serpentinite from the base of the CoastRange ophiolite (Great Valley complex) is found entrained in the Franciscan Complex mélange,whereas high-grade metamorphic blocks (Franciscan Complex) are found entrained in serpentinitematrix mélange that is associated with the Coast Range ophiolite. Previous workers have includedserpentinite as a component of the Franciscan Complex, but the absence of middle-lower oceaniccrust rocks in coherent Franciscan Complex terranes, as well as the plate tectonic model fromFranciscan Complex terrane emplacement, suggests that the base of the downgoing (FranciscanComplex bearing) slab was entirely subducted, and that all serpentinite in the California CoastRanges was originally part of the upper plate (Coast Range ophiolite). The emplacement ofFranciscan Complex terranes beneath the Great Valley complex also implies that FranciscanComplex rocks exist at depth beneath exposed Great Valley complex rocks at least as far east as theGreat Valley margin.Paleogene unroofing and overlapAlthough subduction and Franciscan Complex accretion at the western margin of the NorthAmerican Plate continued into Neogene time, based on the presence of Miocene marine fossils insome Franciscan Complex terranes several hundred kilometers north of the study area (McLaughlinand others, 1996), accretion of the bulk of Franciscan Complex terranes was probably complete byabout the end of the Cretaceous Period. Tectonic and(or) erosional unroofing of the FranciscanComplex terranes, as evidenced by the presence of Franciscan Complex detritus in Tertiary strataoverlying Great Valley Sequence rocks, began in Paleocene time (Berkland, 1973) and waswidespread by Eocene time (based both on the presence of Franciscan Complex detritus in Eocenestrata and the presence of Eocene strata unconformably overlying Franciscan Complex rocks).Therefore, although certainly modified by later tectonics, the basic amalgamation of the two terranecomplexes (Franciscan and Great Valley) was complete by Tertiary time. As a result, for this 3Dmap we treat all Tertiary strata in the San Francisco Bay region (except Tertiary FranciscanComplex terranes in northwestern Sonoma County) as an overlap sequence (that is, strata formedafter amalgamation of separate terranes and deposited over both of the older units), even though theoldest units that are actually observed to be deposited on both terrane complexes are middleMiocene marine strata in the East Bay hills, the hills just east of the cities of Berkeley, Oakland,San Leandro, and Hayward (fig. 1) (Hall, 1958; Graymer and others, 1996).Neogene transpression and development of the San Andreas Fault systemStarting about 30 Ma, the subduction zone at the western North American plate margin began toconvert to the present transpressional margin during the formation and migration of two triplejunctions, the Mendecino Triple Junction and the Baja Triple Junction (McKenzie and Morgan,1969; Atwater, 1970; Atwater and Stock, 1988). In northern California this conversion resulted in8

three major structural developments: the San Andreas Fault, a series of volcanic fields, andwidespread compressive deformation. First, the largely strike-slip San Andreas Fault System wasformed as the northwestward migrating Mendocino Triple Junction passed northern California.This system is made up of multiple faults, which jointly have accomodated at least 450 km of rightlateral offset over the past 23 m.y. (Matthews, 1976; Clark and others, 1984). In much ofCalifornia about two-thirds of strike-slip offset has been taken up by the San Andreas Fault itself,but in the San Francisco Bay area, the fault system is complex and some strands of the presentlyactive San Andreas Fault have taken up as little as 28 km of right-lateral slip (Cummings, 1968;Jachens and Zoback, 2000). About 175 km of the total strike-slip offset on the San Andreas Faultsystem has been taken up in the past 12 Ma by faults in the East Bay (Jones and Curtis, 1991;McLaughlin and others, 1996; Jachens and others, 1998), including about 100 km of offset on theHayward Fault and about 50 km of offset on the Miller Creek-Moraga-Pinole Fault systems in thestudy area (Graymer and others, 2002a).Second, in the wake of the migrating triple junction a northward-younging series of volcanic fieldswas developed (Fox and others, 1985). Many of these volcanic fields formed at or near the largelystrike-slip faults of the San Andreas Fault System, and as a result parts of these volcanic fields havebeen widely separated; offset of correlated volcanic fields provide one of the primary controls onthe amount and timing of fault offset.Third, the late Neogene has been a time of widespread compressive deformation throughout theCoast Ranges (Suppe, 1977; Namson and Davis, 1984; Miller, 1999). Some of the compressionhas resulted from restraining bends and stepovers in the largely strike-slip faults (for exampleAndrews and others, 1993), but there is also a regional component of fault-normal compression asevidenced by 1) the uniformly compressed folded and reversely faulted Neogene strata in theregion (Page, 1982), 2) a reverse-offset component of fault motion on some of the San AndreasFault system, and 3) basement wedge-faulting (Wentworth and others, 1984; Unruh and Moores,1992).The results of the combined compressional and transform deformation in the San Francisco Bayregion include two important developments. First, fault offset has juxtaposed structural blocks ofdiffering stratigraphy (miniature terranes termed stratigraphic assemblages by Jones and Curtis,1991). Second, during the Neogene, the present north-northwest trending ridges developed in theCoast Ranges (Dibblee, 1966; Vanderhurst and others, 1979; Jones and others, 1994; Graymer,1995). The Berkeley-Oakland Hills in the study area were probably uplifted in Quaternary time,based on the presence of Pleistocene alluvial gravels near the ridge tops in the Oakland Hills, thehills just east of the city of Oakland (fig. 1) (Graymer, 2000).MethodsThe methods used to construct the 3D geologic map combined geologic mapping with geophysicalmapping techniques. Geologic mapping techniques included the mapping of surface geology andinferring the sub-surface geology using down-dip projections and the surface geologic mappatterns. Geophysical techniques were used to investigate the primarily unobservable subsurface.Geophysical techniques included seismic reflection and refraction surveys, tomography, andpotential field surveys. Potential field surveys, especially gravity and magnetic data, which offeredbroad areal coverage, were used extensively to define geologic structures in this model (figs. 3 and4). Several mapping techniques were applied in a complimentary fashion to maximize theeffectiveness of each piece of evidence for different portions of the model; five areas in particular,the Hayward Fault, the east hills area (the hills just east of the cities of Berkeley, Oakland,9

122 30'0"W122 15'0"W122 0'0"W121 45'0"W38 15'0"N38 15'0"NSan PabloBay basin38 0'0"N38 0'0"N3D map boundarycoastline37 45'0"N37 45'0"NCalaveras F.Hayward F.Rodgers Creek F.mGal31 - 36San Leandrobasin26 - 3021 - 2516 - 2011 - 156 - 101-5-4 - 0-9 - -5-14 - -10-19 - -1537 30'0"N37 30'0"N-24 - -20-29 - -25-34 - -30-39 - -35-44 - -40-49 - -45Evergreenbasin-54 - -50-59 - -55-67 - -60122 30'0"W122 15'0"W00122 0'0"W2551015121 45'0"W50 Kilometers2025 MilesFigure 3. Isostatic residual gravity anomaly map of the study area. Data fromRoberts and Jachens (1993) and unpublished gravity data.10

122 30'0"W122 15'0"W122 0'0"W121 45'0"W38 15'0"N38 15'0"NTv38 0'0"N38 0'0"N3D map boundarycoastlineRodgers Creek F.Calaveras F.JgbHayward F.37 45'0"NnT37 45'0"N261 - 280241 - 260221 - 240201 - 220181 - 200161 - 180141 - 160121 - 140101 - 12081 - 10061 - 8041 - 6021 - 401 - 2037 30'0"N37 30'0"N-19 - 0-39 - -20-59 - -40-79 - -60-99 - -80-119 - -100-139 - -120-159 - -140-482 - -160122 30'0"W122 15'0"W00122 0'0"W2551015121 45'0"W50 Kilometers2025 MilesFigure 4. Magnetic anomaly map of the study area. Labels indicate the magneticanomalies that helped define the volcanic rocks of uncertain age (Tv) and the rocksof the Coast Range ophiolite (Jgb). Data from U.S. Geological Survey (1999).11

San Leandro, and Hayward, fig. 1), the Quaternary deposits and Tertiary volcanic deposits, the SanLeandro gabbro of Ponce and others (2003), and the Mesozoic terranes west of the Hayward Fault,represent the full spectrum of mapping techniques used to make the 3D geologic map.Hayward FaultThe Hayward Fault is represented in the map as a single surface, ignoring the possibility ofmultiple active strands in a finite width fault zone. This surface was created using two data sets:(1) the mapped trace of the creeping Hayward Fault (Lienkaemper, 1992), and (2) the locations ofhypocenters of earthquakes from 1984-2000 relocated using the double-difference technique(Waldhauser and Ellsworth, 2000; Ponce and others, 2004). Software tools were used to project thefault trace onto the earth’s surface, as defined by USGS 30-meter digital elevation models. Usingmap view slices spaced 1.25 km apart in depth, and cross-section slices perpendicular to the faultspaced 1.25 km apart along the fault, a set of curves, representing the fault trace in each profile,was constructed and adjusted by eye, honoring insofar as possible the surface fault trace data andthe hypocenter data. Simpson and others (2003) show a set of sections at 2.5 km spacing, givingsome impression of the hypocenter density along the fault. In places with few or no earthquakes, itwas necessary to interpolate and extrapolate using the nearest clusters of events. When faced withambiguities and complexities, our choice was to opt for a simple, smooth surface consistent withmost hypocenters. In places where there were hints of multiple active strands, the surface wasmade to pass through either the better defined strand or the strand that resulted in a morecontinuous overall surface. Adjustments were made in an iterative fashion, alternating betweenmap views and cross-section views, so that interpolations and extrapolations in both horizontal andvertical directions would be consistent. The curves were then sampled at 1.25 km intervals, togenerate a set of xyz points. The final set of xyz points was then imported into EarthVision (Dynamic Graphics Inc.) to create the Hayward Fault surface using interpolation routines providedin the software. Additional control points were added as required, based on geologic reasoning, toremove artifacts created by the interpolation algorithm in areas of poor data control.East Hills AreaEast of the Hayward Fault, pre-Quaternary strata are exposed by the east-side-up component ofmotion within the fault zone. The strata are deformed; tight, complex folds and many minor faultsresult in a complicated ge

The Hayward Fault (fig. 1) is one of three major faults of the San Andreas Fault System that has produced large (M 6) historic earthquakes in the San Francisco Bay area; the other major faults are the San Andreas and Calaveras Faults. The Hayward Fault roughly bisects the San Francisco Bay