You are on page 1of 60

FINAL TECHNICAL REPORT

Project Title: Assessment of Seismogenic Sources between the Rodgers Creek and San Andreas Faults, Northwestern San Francisco Bay Region, Sonoma County, California Recipient: William Lettis & Associates, Inc. 1777 Botelho Drive, Suite 262 Walnut Creek, California 94596 (925) 256-6070 Principal Investigators: Christopher S. Hitchcock William Lettis & Associates, Inc., 1777 Botelho Dr., Suite 262, Walnut Creek, CA 94596 (phone: 925-256-6070; email: hitch@lettis.com) Keith I. Kelson William Lettis & Associates, Inc., 1777 Botelho Dr., Suite 262, Walnut Creek, CA 94596 (phone: 925-256-6070; email: kelson@lettis.com)

Program Elements: I

U. S. Geological Survey National Earthquake Hazards Reduction Program Award 1434-HQ-97-GR-03153

November 1998 Research supported by the U.S. Geological Survey (USGS), Department of the Interior, under USGS award number 1434-HQ-97-GR-03153. The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the U.S. Government.

ABSTRACT Detailed geomorphic mapping and analyses provide constraints on the distribution of late Quaternary strain in the region between the Rodgers Creek and San Andreas fault zones in the northern San Francisco Bay Area, south of the Russian River. The regionally extensive, 4-6 Ma Wilson Grove Formation provides a datum for identifying and characterizing locations of longterm tectonic deformation. Deposition of the Wilson Grove Formation occurred on a surface of low to moderate relief beveled across Franciscan basement rock, likely within a shallow marine embayment. The base of the Wilson Grove Formation is offset across the northwest-trending Americano Creek, Bloomfield, Joy Woods, Dunham, Tolay, Salmon Creek, and Burdell Mountain faults. Contours on the base of the Wilson Grove Formation document uplift and regional eastward tilting of the Sebastopol structural block, with maximum uplift within the western and central portions of the block. Uplift of the structural block is accommodated locally by discrete reverse faulting and associated anticlinal folds that extend between the northwestern end of Petaluma Valley and the mouth of the Russian River near Jenner, and along a narrow coastal zone east of the San Andreas fault. Localized tectonic uplift along late Cenozoic faults and associated folds is expressed geomorphically within the study area as two distinct series of northwest-trending hills that rise 30 to 150 m (100 to 500 feet) above accordant hilltops underlain by the Wilson Grove Formation. The linear hills appear to be the axes of folds underlain by and/or bounded by thrust or reverse faults. The easternmost of these linear hills, typified by the prominent English Hill located approximately 3 km (2 miles) north of Bloomfield, is underlain by resistant rock of the Franciscan Complex. The hills [and associated folds] are bounded on the southwest by the Tolay, Dunham Bloomfield, and Joy Woods faults. For convenience, we refer to this zone of crustal shortening as the Bloomfield fold-and-thrust belt. The second topographically distinct fold belt consists of low rugged hills 3 to 5 km (2 to 3 miles) wide that border the Sonoma Coast adjacent to the San Andreas fault. These hills, and associated deeply incised stream valleys and steep sea cliffs along the coast, are developed entirely within resistant Mesozoic rocks of the Franciscan Complex. Several southwest-flowing drainages cross the northwest-trending structural grain of the study area. Longitudinal profiles of Pleistocene and younger stream terraces within these drainages show localized warping and uplift coincident with the loci of deformation within the Wilson Grove Formation. Additionally, valley height to valley-floor width ratios reveal changes in stream valley morphology along Salmon, Americano, Stemple, and Walker creeks across the Bloomfield fault and near the San Andreas fault. The observed changes in morphology are consistent with a response of these fluvial systems to localized late Quaternary uplift. Similarly, profiles of marine terraces along the coast show progressive tilting of older terraces and warping of the entire terrace sequence consistent with localized folding adjacent to the San Andreas fault. These uplifts represent localized strain between the San Andreas and Rodgers Creek faults, and likely are the result of large-scale interactions between these two major faults. The results of this study document northeast-southwest directed shortening between the San Andreas and Rodgers Creek fault. Deformation of both the Wilson Grove Formation and younger Quaternary fluvial and marine terraces indicate progressive, continuing deformation along a series of anticlinal folds underlain by and/or bounded by Quaternary active reverse or thrust faults. Structural relief on the base of the Wilson Grove Formation suggests uplift rates between 0.03 to 0.15 m/ka. We obtain an average slip rate between approximately 0.07 + .03 m/ka and 0.13 + .04 m/ka for the Bloomfield fault. These rates are consistent with stream incision of uplifted Quaternary fluvial terraces. Marine terrace elevations along the coast suggest maximum uplift rates of approximately 0.14 m/ka for the Sonoma fold-and-thrust belt adjacent to the San Andreas fault.

ii

TABLE OF CONTENTS Section Page Abstract ........................................................................................................ii Table of Contents............................................................................................iv 1.0 INTRODUCTION...................................................................................1 1.1 Significance and Purpose......................................................................1 1.2 Summary of Results ...........................................................................1 1.3 Acknowledgements ............................................................................4 2.0 REGIONAL GEOLOGIC SETTING ............................................................5 2.1 Structural Setting...............................................................................5 2.1.1 Major strike-slip faults in the study region.......................................7 2.1.2 Major folds in the study region....................................................7 2.1.3 Major late Cenozoic thrust faults in the study region...........................8 2.2 Stratigraphy.....................................................................................8 2.2.1 Franciscan Complex................................................................8 2.2.2 Petaluma Formation ................................................................9 2.2.3 Wilson Grove ("Merced") Formation ............................................9 2.2.4 Glen Ellen Formation.............................................................10 2.2.5 Sand and Gravel of Cotati .......................................................10 2.3 Regional Geomorphic Setting...............................................................10 2.3.1 Mendocino Plateau.............................................................15 3.0 4.0 TECHNICAL APPROACH......................................................................18 RESULTS..........................................................................................23 4.1 Quaternary Stratigraphy .....................................................................23 4.1.1 Marine Terraces...................................................................23 4.1.2 Fluvial Terraces...................................................................28 4.1.3 Correlation and Age of Map Units..............................................28 4.2 Geomorphic Analyses .......................................................................32 4.2.1 Contours on the base of the Wilson Grove Formation.......................33 4.2.2 Fluvial Terrace Longitudinal Profiles..........................................36 4.2.2.1 Americano Creek...................................................36 4.2.2.2 Stemple Creek ......................................................36 4.2.3 Marine Terrace Profiles ..........................................................37 4.2.4 Fluvial Channel Profiles ........................................................37 4.2.4.1 Salmon Creek.......................................................40 4.2.4.2 Americano Creek...................................................40 4.2.4.3 Stemple Creek ......................................................40 4.2.4.4 Discussion...........................................................40 4.2.5 Stream Valley Morphology......................................................40

iii

TABLE OF CONTENTS (CONTINUED) Section Page 4.3 Stream Development and Landscape Evolution...........................................43 5.0 DISCUSSION .....................................................................................46 5.1 Rates of Quaternary Deformation ..........................................................46 5.2 Hazard Assessment of Seismogenic Sources.............................................47 6.0 7.0 CONCLUSIONS..................................................................................51 REFERENCES ....................................................................................52

List of Tables 1 2 List of aerial photography used in reconnaissance geologic mapping.......................17 Estimated ages and correlation of marine and fluvial terraces ................................32

List of Photographs 1 2 3 4 5 6 7 8 9 10 Photo of rolling hills developed within the Wilson Grove Formation, looking northeast from Sugarloaf Hill, above Dillon Beach, toward English Hill...................12 Photo looking northeast toward English Hill from Highway 1, showing typical topography associated with outcrops of rock of the Franciscan Complex..................12 Photo looking southwest showing the lower reach of Stemple incised across the Sonoma Coast fold-and-thrust belt along the coast.......................................................14 Photo looking southwest toward the Pacific Ocean showing the lower reach of Americano Creek incised across the Sonoma Coast fold-and-thrust belt....................14 Photo looking southeast, near Dillon Beach, of the exhumed erosional contact between the Wilson Grove Formation and underlying Franciscan Complex...............16 Photo of marine terrace deposits exposed at Dillon Beach ...................................27 Photo looking northeast across Estero Americano showing stream terraces east of the coastal hill zone ....................................................................................29 Photo looking east at exposure of bedding within the Wilson Grove Formation showing how bedrock bedding can mimic a fluvial terrace surface..........................29 Photo of fluvial terrace deposits near Valley Ford.............................................30 Photo of fluvial terrace deposits with closeup view of sub-rounded terrace gravels...............................................................................................30

List of Figures 1 2 3 4 Regional map of the San Andreas and Rodgers Creek faults and other major late Cenozoic structural features in the northwestern San Francisco Bay Area....................2 Map of major late Cenozoic structural features between the San Andreas and Rodgers Creek faults with major streams.........................................................3 Shaded relief map of major late Cenozoic structural features showing the outcrop limits of the Wilson Grove Formation...................................................6 Regional drainage pattern showing major stream and drainage basins, drainage divides, and major wind gaps....................................................................13
iv

5 6 7 8 9 10 11 12 13 14 15 16 17 18

Conceptual model of stream changes across an idealized uplift..............................20 Generalized photogeologic and reconnaissance map of marine terraces between Russian River and Bodega Bay..................................................................24 Detailed photogeologic and reconnaissance map of lower Americano Creek...............25 Detailed photogeologic and reconnaissance map of lower Stemple Creek..................26 Diagram showing tentative correlations and general uplift history of terraces along the Sonoma Coast between the Russian River and Dillon Beach.....................31 Map showing contours on the base of the 4-6 Ma Wilson Grove Formation...............34 Regional southwest-northeast cross-section A-A showing folding and faulting of the base of the Wilson Grove Formation. ...................................................35 Longitudinal profile of terraces along Stemple Creek .........................................38 Profile B-B of marine terraces between Russian River and Bodega Bay ..................39 Stream channel profile of Salmon creek shown on semilogarithmic plot ...................41 Stream channel profile of Americano and Stemple creeks shown on semilogarithmic plots..............................................................................42 Plots of the ratio of valley floor width to valley height for Salmon, Americano, and Stemple Creeks................................................................................44 Detailed photogeologic and reconnaissance map of the Bloomfield area, showing the Bloomfield and Americano Creek faults.........................................49 Structural cross-section C-C through the Bloomfield area, across the Americano Creek and Bloomfield faults ........................................................50

1.0 INTRODUCTION The San Andreas and the Rodgers Creek faults exhibit geologic, seismologic, and geodetic evidence of activity within the northern San Francisco Bay region. However, the style and amount of strain transfer between these major faults is poorly constrained as the origin and kinematic significance of contractional structures between the faults generally is undefined (Brocher and Furlong, 1994). The presence of potentially active thrust faults and folds (Pampeyan, 1979; Wagner and Bortugno, 1982), associated microseismicity (Wong, 1991), and deformed Quaternary geomorphic surfaces (Holway, 1914; Weaver, 1949; Higgins, 1952) provide evidence of Quaternary deformation within the region bounded by the San Andreas and Rodgers Creek fault zones. In addition, stream patterns reflect incision, diversion, and capture that may reflect regional and local late Quaternary uplift (Holway, 1914; Dickerson, 1922; Gealey, 1951; Higgins, 1952). 1.1 Significance and Purpose This study investigates the origin, deformation rates, and seismic potential of late Cenozoic faults and folds between the San Andreas and Rodgers Creek faults in the northwestern San Francisco Bay region, California (Figures 1 and 2). Currently, the seismic potential of these potentially active structures has not been fully incorporated into assessments of seismic hazards in the San Francisco Bay region. Information on the style and rates of deformation in this area, therefore, is important for two fundamental purposes: (1) to assess the origin of crustal shortening and nature of strain interaction between the San Andreas and Rodgers Creek faults; and (2) to improve our understanding of potential seismic sources in the north Bay area. To address these issues, we employ several geologic and geomorphic techniques to identify and characterize late Cenozoic structures in the region. Initially, we contoured the base of the 4-6 Ma Wilson Grove Formation, a regionally extensive unit that was deposited on a surface of low relief in a shallow marine embayment. Previous studies in the area used the outcrop pattern and orientation of bedding within the Wilson Grove Formation to map and characterize faults and folds in the region. Because the Rodgers Creek fault system likely developed within the last 2-4 Ma (Helley and Herd, 1977), the late Quaternary distribution of strain within the region may not be accurately reflected by deformation of the 4-6 Ma Wilson Grove Formation. Therefore, we have identified and evaluated other strain indicators in the late Quaternary geologic and geomorphic record within the region. This report provides longitudinal profiles of stream channels and geomorphic surfaces, and stream-channel gradient changes along stream reaches, that cross potentially active contractional structures identified via contouring of the Wilson Grove Formation. These data, combined with the application of correlative dating techniques, allow interpretation of the locations, styles, and approximate amounts of late Cenozoic deformation, and provide an initial characterization of potential seismic sources between the San Andreas and Rodgers Creek faults. 1.2 Summary of Results The major results of this study include the following: 1) A belt of late Cenozoic folds and active faults extends between the Rodgers Creek fault to the San Andreas fault, from northern Petaluma Valley to the mouth of Salmon Creek, south of the Russian River (Figure 2). This fold belt is expressed as a series of northwest-trending hills, eroded into resistant bedrock of the Franciscan Complex, that

N
124 39 123
a am ac t Ma Faul

122

n Sa s ea dr An
SA LI

K H ult OC Fa RT BL ek re NO sC er SA dg RO Ro g K A ur OC NT ldsb BL a SA He OL

AM

SE BA NI AN

ER

38

STUDY AREA (Figure 2)


BL OC K

OP ST

IC AN E AT PL
Co

CI PA FI C PL m /y r
N

nc ord

yw Ha

39 m

d ar u Fa

Ca lav

E AT
u Fa lt

s e ra
u Fa lt

lt

37
0 100 km

Figure 1. Regional map showing the study area, major faults in the northern San Francisco Bay area, and major structural blocks after Fox (1983).

rise 30 to 150 m (100 to 500 feet) above rounded accordant hilltops underlain by the Wilson Grove west of the Santa Rosa Valley. The folds are bounded on the southwest by the Tolay, Bloomfield, and Joy Woods faults. English Hill, the most prominent of these hills, has over 150 m (500 feet) of topographic relief and is capped by a thin uplifted remnant of Wilson Grove deposits that has been vertically offset a minimum of 183 m (600 feet) across the Bloomfield fault. Therefore, for convenience, we refer to this zone of crustal shortening as the Bloomfield fold-and-thrust belt. 2) A second belt of late Cenozoic folds is expressed as a narrow belt of low rugged hills 3 to 5 km (2 to 3 miles) wide bordering the Sonoma Coast adjacent to the San Andreas fault (Figure 2). These hills are developed within resistant bedrock of the Franciscan Complex. These folds locally warp the base of the Wilson Grove Formation, resulting in erosional stripping of deposits within the Wilson Grove Formation adjacent to the coast, and locally uplift and tilt marine terraces along the coast. For convenience, we refer to this zone of fault-parallel shortening as the Sonoma Coast fold-and-thrust belt. Profiles of marine terraces along the coast suggest uplift rates of approximately 0.14 mm/yr for the Sonoma fold-and-thrust belt. 3) Longitudinal profiles of Pleistocene and younger stream terraces show localized warping and uplift coincident with the two identified fold belts within the study area and with loci of deformation within the Wilson Grove Formation. Additionally, changes in stream valley morphology along major streams that cross the Bloomfield fold-and-thrust belt and the Sonoma Coast fold-and-thrust belt are consistent with a response of these fluvial systems to localized Quaternary uplift. 4) Development of the modern drainage systems in the region, as documented herein, is consistent with localized folding as part of regional uplift. Active folding and uplift appears to disrupt the flow patterns of stream developed on the Wilson Grove Formation. Specifically, major streams that cross the study area, with the exception of the Russian River, appear to have been beheaded from their ancestral drainage basins, likely located east of Santa Rosa Valley. This change in regional drainage patterns may reflect the post-Wilson Grove development of the ancestral Santa Rosa basin and initiation of faulting along the Healdsburg-Rodgers Creek fault. 5) Our geomorphic analyses provide evidence that the loci of Quaternary deformation roughly coincide with the loci of past Wilson Grove Formation Pliocene deformation. Documented offset of the base of the Wilson Grove Formation and localized uplift of fluvial terraces derived from longitudinal terrace profiles provide a basis for estimating preliminary uplift rates and maximum earthquakes associated with the thrust faults beneath the folds. Structural relief on the base of the Wilson Grove Formation suggests uplift rates between 0.02 to 0.2 m/ka. These rates are consistent with stream incision of uplifted Quaternary fluvial terraces. However, additional age dates on Quaternary stratigraphic units are necessary to constrain horizontal shortening rates and slip rates on these faults. 1.3 Acknowledgements Support for this research was provided to William Lettis & Associates, Inc., by a grant from the Department of Interior, U.S. Geological Survey (National Earthquake Hazards Reduction Program, contract 1434-HQ-97-GR-03153). The contents of this report do not necessarily represent the policy of the U.S. Geological Survey, however, and the endorsement of the federal government should not be assumed. We appreciate conversations with Robert McLaughlin, of the USGS, Jeffrey Unruh and William Lettis, of William Lettis & Associates, Inc., and review of this report by William Lettis. 4

2.0 REGIONAL GEOLOGIC SETTING 2.1 Structural Setting The location and characteristics of late Cenozoic deformation between the Rodgers Creek fault and the San Andreas fault, south of the Russian River, is poorly understood. Numerous steeply dipping, en-echelon reverse faults and associated folds trend about N60W between, and oblique to, the San Andreas and Rodgers Creek faults (Travis, 1952; Wagner and Bortugno, 1982). These potentially active faults (Wagner and Bortugno, 1982; Pampeyan, 1979), include the Americano Creek, Bloomfield, Dunham, Tolay, and Burdell Mountain faults (Figure 2). Contemporary microseismicity within the region, including seismic events that coincide with the mapped traces of the Americano Creek, Bloomfield, and Dunham faults, suggests that these structures have predominately reverse slip along northwest-striking fault planes (Wong, 1991). The oblique trend of the faults and folds suggests that crustal shortening in the region is driven by a left restraining transfer of strain between the San Andreas and Rodgers Creek faults. Much of the study area is within the Sebastopol structural block defined as a semirigid crustal block located between the Pacific and North America plates (Furlong et al., 1989; Figure 1). Movement between the tectonic plates is accommodated primarily by rightlateral, strike-slip displacement on the San Andreas fault, along the western boundary of the Sebastopol structural block, and on the Rodgers Creek and Healdsburg faults, separating the Sebastopol block and Santa Rosa structural block (Figures 1 and 2). Rightlateral slip of 6 to 10 mm/yr on the Rodgers Creek fault (Schwartz et al., 1992), must transfer northward onto the southern Healdsburg and Maacama faults or step northwestward, in part, onto the northern San Andreas fault. However, because the average dip of deposits within the 4-6 Ma Wilson Grove Formation in the Sebastopol block is substantially less than the average dip of Quaternary deposits within the Santa Rosa structural block to the east, Fox (1983) has argued that the Sebastopol block is located within a zone of reduced compressive stress along the plate boundary. The eastern margin of the Sebastopol block, as delineated by Fox (1983), corresponds roughly to the western margin of Santa Rosa Valley, west of the Rodgers Creek fault (Figure 2). This boundary encompasses the mapped northwestern end of the Tolay fault and several major northwest-trending folds that underlie and bound Santa Rosa Valley. Not surprisingly, the block boundary also closely coincides with the eastern margin of the exposed portion of the Wilson Grove Formation. The assumption that the Sebastopol block has experienced relatively minor deformation during the Quaternary primarily is based on sub-horizontal bedding in the Wilson Grove Formation (typically dips are less than 10, although locally dips can exceed 20 adjacent to faults; Bedrossian, 1969), and relatively gentle topography (Fox, 1983). However, the northwestern projection of the Tolay fault coincides with the southeastern end of the Bloomfield fold-and-thrust belt described in this report, suggesting that strain within the Santa Rosa block may be transferred across the inferred boundary between the two structural blocks. In addition, as discussed below, unique lithologic characteristics of the Wilson Grove Formation has profoundly influenced geomorphic development within the Sebastopol block, resulting in gentle, rolling topography upon regionally uplifted and tilted terrain (Figure 3). This uplift within the Sebastopol block appears to be primarily expressed as discrete block faulting (Johnson, 1934; Higgins, 1952), with localized folding occurring adjacent to mapped faults. Below we discuss the folds and reverse faults mapped within the Sebastopol block along with the other major faults and folds within the region.

He ald g ur sb

r dso Win e clin Syn

u fa lt

Forestville
Ca mp

Me ek er

Santa Rosa

S A N

e lin nc Sy

Jo y
Sa lm on C re

Ro

W oo ds

r dge

Freestone
k
fa
e

Sebastopol

O S A
Cr

k ee

fau
ul t

lt
Bl oo

t ul fa

Bodega Bodega Bay


Bo

ield Fa Bloomfield ul A m e ri c a t no

mf

un h

am
U D

Was ho e

L E

PA CI FI C OC EA

Cree

Fa ul t

nt ic lin e

Fa

Penngrove
Tola y

de ga Ba y

Two Rock Dillon Beach

t ul
Fau lt

Petaluma

N
0 0 5 km

Bur dell Mtn

5 mi

Figure 3. Shaded relief map showing major late Cenozoic structural features between the San Andreas and Rodgers Creek faults, with the limits of the Wilson Grove Formation shown in black.

fau lt

Sa n An dr ea s fa ul t

2.1.1 Major Strike-Slip Faults in the Study Region The San Andreas fault bounds the western margin of the study area. Where exposed on shore, the right-lateral San Andreas fault juxtaposes rocks of the Salinian block on the west with bedrock of the Mesozoic Franciscan Complex on the east. Paleoseismic studies at Olema, southwest of the study area, constrain the slip rate of the San Andreas fault to 243 mm/yr (Niemi and Hall, 1992). Geodetic surveys between 1972 and 1989 across the 115km wide Point Reyes network, which spans the San Andreas fault system (about 80 km wide at Olema) measure a right-lateral displacement rate of 313 mm/yr (Lisowski et al, 1991). Since the 1906 earthquake, the San Francisco segment of the fault from Point Arena to Woodside has been largely inactive with minimal associated seismicity. The Rodgers Creek fault is a right-lateral strike-slip fault within the regional San Andreas system that extends approximately 40 to 50 km from San Pablo Bay to the vicinity of Santa Rosa on the north. At its northern end, the Rodgers Creek fault may intersect, or stepover to, the southern end of the Healdsburg fault or the Maacama fault (Figure 1). Significant associated earthquakes include the ML 5.6 and 5.7 earthquakes that occurred at the northern end of the Rodgers Creek fault near Santa Rosa in October, 1969 (Wong and Bott, 1995). The Rodgers Creek and Healdsburg faults likely are part of an en-echelon fault system trending N35W, approximately parallel to the San Andreas fault to the west. Evidence for right-lateral displacement on the Rodgers Creek fault includes offset streams, shutter ridges, and sag ponds. 2.1.2 Major Folds in the Study Region Several major folds have been mapped within the study area, primarily based on detailed mapping of the Wilson Grove, Petaluma, and Glen Ellen Formations. In general, mapped folds trend northwest and occur in broad (5-10 km wide) belts (Figure 2). The largest of these folds underlies the northern end of Santa Rosa Valley. The Santa Rosa basin is an approximately northwest-trending syncline named the Windsor Syncline by Gealey (1951). The axis of the Windor Syncline extends from Rohnert Park northwest through Windsor toward Healdsburg. The Windsor syncline is mapped based on dips within the Glen Ellen Formation at the northern end of the valley. The syncline, as mapped, is broadly asymmetrical with gentle dips on the southwestern limb and steeper dips on the northeastern limb. The western margin of the Santa Rosa basin is bounded by a range of low hills formed by a series of northwest-trending folds identified, and mapped, by Weaver (1949), Gealey (1951), and Travis (1952). The Washoe Anticline was named by Weaver (1949) for a mapped structure in the Wilson Grove Formation and Sonoma Volcanics along Washoe Creek, west of Cotati. The anticline extends from the margin of the Petaluma Valley, south of Penngrove, northwest to Gossage Creek, a distance of 10 kilometers, and forms the divide between Santa Rosa Valley and Petaluma Valley to the south. The Washoe Anticline is well expressed by deformed bedding within the Petaluma Formation, and in particular by the outcrop pattern of the folded Tolay Volcanics exposed in the core of the anticline. Several unnamed, poorly defined faults and folds are mapped northwest of the Washoe Anticline (Gealey, 1951; Travis, 1952), along the western margin of Santa Rosa Valley. Numerous poorly defined, northwest-trending anticlines and synclines are mapped within the central portion of the study area, based primarily on bedding orientations and the mapped distribution of remnants of the Wilson Grove Formation. These folds include the Camp Meeker Syncline and other unnamed folds mapped by Travis (1952) north of Occidental (Figure 2). These folds, and unnamed folds identified within this study, are described in more detail in Section 4. 7

2.1.3 Major Late Cenozoic Thrust Faults in the Study Region Previous workers, including Johnson (1934), Weaver (1949), and Travis (1952), have mapped several northwest-trending faults in the study region primarily based on observed offset of units within the Wilson Grove and Petaluma Formations. These faults include the Dunham, Bloomfield, Joy Woods, Americano Creek, and Salmon Creek faults (Figures 2 and 3). Based on the topographic expression of these faults, localized drag folding adjacent to the faults, vertical separation of bedding, and fault exposures (Travis, 1952), these faults have been mapped as reverse or thrust faults. The northwest-trending Bloomfield, Dunham, and Joy Woods faults juxapose Franciscan Complex rocks on the northeast with the Wilson Grove Formation on the southwest. The Dunham fault has over 91 m (300 ft) of displacement of the basal Wilson Grove formation and the Bloomfield a minimum of over 183 m (600 ft). The Joy Woods fault is over 6 km (3.5 mi) long with maximum vertical displacement of about 91 m (300 ft). The Tolay fault is an approximately 40-km-long fault that extends from the mouth of Tolay Creek at San Pablo Bay northwestward (N50W) to the hills south of Sebastopol. The Tolay fault has been mapped as both a strike-slip fault and a thrust fault. Morse and Bailey (1935) estimated at least 1,400 m of reverse dip-slip displacement based on oil and gas well logs. Fault-related features along the Tolay fault include several offset streams and a large playa lake (Tolay Lake) that have been cited as evidence for Pleistocene or younger displacement (Armstrong, 1974). 2.2 Stratigraphy Geologic units exposed within the study area provide a record of sedimentation, erosion, and regional deformation. These units have been used by past researchers to map major faults and folds. In addition, the composition of geologic units exposed within the study area has directly influenced the geomorphic evolution of the region. Johnson (1934) provided detailed information on, and mapped the lateral extent of, the Wilson Grove Formation. Gealey (1951) mapped bedrock geology and Quaternary deposits of the Healdsburg 1:62,500-scale quadrangle. Travis (1952) mapped bedrock geology and surficial deposits of the adjacent Sebastopol 1:62,500-scale quadrangle. Higgins (1952) mapped fluvial terraces along the lower Russian River. Dickerson (1922) published bedrock geologic maps of the Santa Rosa-Petaluma-Point Reyes area. We used the bedrock mapping by Johnson (1934), Travis (1952), Gealey (1951), and Dickerson (1922) as a basis for contouring the base of the Wilson Grove Formation, constructing geologic cross sections, and for evaluating stratigraphic relationships. Below we provide general descriptions of the major geologic formations in the area. More detailed descriptions of individual stratigraphic units are provided in Gealey (1951), Travis (1952), and Fox (1983). Descriptions of younger, late Quaternary sediments mapped for this study are provided in Section 4.1. 2.2.1 Franciscan Complex Bedrock of the Franciscan Complex underlies much of the study area and forms most of the rugged highlands that border the coast and rise to the north of the Russian River. As defined by Berkland and others (1972), the Franciscan Complex consists of structurally juxtaposed bodies of various origins. Within the study area, the Franciscan Complex contains complexly folded and fault-bounded slivers of graywacke, shale, sandstone, ultramafic bedrock, and mafic volcanic rocks. Sandstone is the most common sedimentary rock type within the Franciscan Complex exposed in the study area. Where exposed, 8

sandstone within the Franciscan Complex typically is massive with little or no internal stratification (Johnson, 1934), making it difficult to trace folds and faults mapped within the Wilson Grove Formation into areas of the study area underlain by the Franciscan Complex. 2.2.2 Petaluma Formation The Petaluma Formation is exposed within the southeastern portion of the study area, typically as dissected uplands along the northeast margin of Petaluma Valley. The Petaluma Formation consists of deposits of clay, shale, and sandstone with occasional thick beds of diatomite (Morse and Bailey, 1935). The exposed portion of the Formation is at least 915 m (3000 feet) thick and lies in contact with Francican bedrock or is covered by the Wilson Grove Formation, as mapped by Weaver (1949). The upper part of the Petaluma Formation apparently interfingers with sediments of the lower part of the Wilson Grove Formation, thus causing confusion in mapping areas of the Petaluma Formation (Dickerson, 1922). As described below, Sarna-Wojcicki and others (1976) have correlated the Petaluma Formation to the Wilson Grove based on a similar 6 Ma tuff found in both Formations. 2.2.3 Wilson Grove ("Merced") Formation First mapped in detail as the Merced Formation by Johnson (1934), marine deposits of the Wilson Grove Formation form an undulating cover of generally uniform thickness that rests unconformably upon complexly folded and faulted Franciscan basement rock within the central part of the study area (Figure 2). The Wilson Grove Formation has a maximum exposed thickness of about 150 m (500 feet) with a gentle northeast dip that rarely exceeds 10 degrees. The Wilson Grove Formation consists of gray to buff, often very fine grained, and fossiliferous, sandstone with lenses of conglomerate and sandy shale. Based on petrographic analyses, Johnson (1934) noted that the marine Wilson Grove Formation likely was derived solely from reworking of the Franciscan Complex, with the exception of an interbedded tuff bed near the base of the Formation. Deposition of the Wilson Grove Formation appears to have occurred on a surface of low to moderate relief beveled across Franciscan basement rock (Travis, 1952), likely within a large shallow marine embayment (Johnson, 1934; Bedrossian, 1969). Marine fossils preserved in fossiliferous sandstone beds within the Formation are representative of a shallow-water fauna assemblage consistent with deposition in a nearshore or marine embayment areas with water depths ranging from 9 to 45 m (30 to 150 feet) (Bedrossian, 1974). Additionally, as noted by Johnson (1934), well-rounded pebbles within conglomerate beds within the Wilson Grove Formation are consistent with a high-energy, near-shore environment. The age of the Wilson Grove Formation is still a matter of considerable debate. Johnson (1934) called it Middle Pliocene, and Eaton (1943) placed it in the Lower Pleistocene. Based on over sixty species of predominately molluscan fauna recognized within the Wilson Grove Formation, Bedrossian (1974) assigned the Formation to the upper Pliocene, approximately 4 million years old. Sarna-Wojcicki and others (1976) correlated a 6 Ma tuff in the basal section of the Wilson Grove Formation with a similar tuff in the Petaluma Formation near Sears Point and a tuff near the base of the continental Tassajara Formation. We assume an age range for the Wilson Grove Formation between 4 to 6 Ma.

2.2.4 Glen Ellen Formation The Glen Ellen Formation, first identified by Weaver (1949), is exposed within the eastern portion of the study area, within northeastern Santa Rosa Valley. The Formation consists of at least 91 m (300 feet) of stratified, but poorly sorted, gravels and sands with interbedded conglomerates typically containing cobbles of andesite (Weaver, 1949). The Glen Ellen Formation is of continental origin, likely deposited as a sequence of coalesced alluvial fans and fluvial deposits, and contains clasts of both the Franciscan Complex and Sonoma Volcanics (Cardwell, 1958). Beds mapped as the Glen Ellen Formation near Glen Ellen contain very sparse pebbles of black obsidian apparently derived from the upper member of the Sonoma Volcanics (Fox, 1983). Obsidian is not found in otherwise similar but older gravels of the Petaluma Formation, and thus their presence in gravel is considered definitive of the Glen Ellen Formation. The Glen Ellen Formation is believed to be Pleistocene because of stratigraphic relations (Fox, 1983), however, no fossils or other age diagnostic materials have been identified that provide absolute age for the Formation. 2.2.5 Sand and Gravel of Cotati West and south of Cotati, stratified sand and gravel deposits are present, termed the sand and gravel of Cotati by Fox (1983). Dickerson (1922) described an old stream channel full of gravel with large fragments of petrified wood within an exposure 0.4 km (0.25 mi) northwest of Penngrove. During field reconnaissance, we found similar old stream gravel deposits but were unable to find the exposure described by Dickerson (1922), possibly because it has been obscured by recent development, or any petrified wood within similar exposures. However, as discussed below in Section 4.2.6, the east-west direction of stream flow in Pleistocene time inferred by Dickerson (1922) from this exposure, and in similar stream gravels exposed near Penngrove, may provide significant clues to the Pleistocene drainage pattern. We found similar gravels, likely correlative with the sand and gravel of Cotati within the prominent wind gap located between Lichau and Stemple Creeks. In Section 4.2.6 we discuss the possibility that this unit likely was originally laterally continuous across the study area. Vertebrate fossils collected from the "sands and gravels of Cotati" are stored under the name "Merced" Formation within the University of California Museum of Paleontology (pers. comm., Patricia Holroyd, UCMP, July, 1998), and apparently have not been studied. The fossils include bison, horse, deer, aurock (early cattle), camel, and two types of ground sloth similar to fossils collected in older stream gravels within the study area west of the Santa Rosa Valley. 2.3 Regional Geomorphic Setting The coastal region south of the Russian River consists of a broad coastal plain separated from the Sonoma Mountains to the east by elongate, northwest-trending valleys. These alluvial-filled valleys typically are oriented sub-perpendicular to the regional east-west drainage pattern, coincide with mapped synclines and/or major faults, and may represent areas of active subsidence. The largest of these valleys are Santa Rosa and Petaluma Valleys, which extend northward from the San Francisco Bay. Between Santa Rosa Valley and the Pacific Ocean to the west, is a coastal plain characterized by gently rolling hills, broad valleys, and rounded hilltops. However, these low-lying hills are surrounded on the north and south by rugged, often inaccessible, mountainous terrain. The northern portion of the study area, in particular, is characterized by rugged highland topography traversed by the Russian River. The Russian River, with a drainage area of over 3850 square km (1,485 square miles) in the California Coast Ranges northwest of the study area, flows within a deep sinuous canyon incised across resistant

10

bedrock of the Franciscan Complex before entering the Pacific Ocean at Jenner, northwest of Bodega Bay. South of the Russian River, topographic relief typically decreases and slopes are more subdued. An exception to the subdued topography is a narrow belt of low rugged hills 3 to 5 km (2 to 3 miles) wide bordering the coast, and several prominent hills that rise 30 to 150 m (100 to 500 feet) above regionally accordant hilltops northeast of Bloomfield. Higgins (1952) used the term Merced Basin for the central portion of the study area as he observed that the limits of the rolling topography closely corresponds to the areal limits of soft marine sands of the "Merced" Formation, or Wilson Grove Formation as renamed by Fox (1983). As illustrated in Figure 3, this area of low-lying hills is essentially restricted to the area of outcrop of the Wilson Grove Formation. North of Bodega, near Freestone and Occidental, where only remnants of the Wilson Grove Formation are preserved, elevations are greater, from 215 to 245 m (700 to 800 feet) and rugged highland topography is developed entirely within underlying Mesozoic rocks of the Franciscan Complex. Where rocks of the Franciscan Complex crop out within the central portion of the study area, as they do along the coast and in the bottom of the stream valleys of the larger drainages, slopes tend to be steeper and the hillsides much more rugged. The most prominent topographic feature in the region of gently rolling hills west of the Santa Rosa Valley is English Hill, located approximately 3 km (2 miles) north of Bloomfield (Photo 1). Bounded on the southwest by the Bloomfield fault, the steep hillslopes of English Hill, underlain by bedrock of the Franciscan Complex, exhibit the gullied appearance characteristic of that developed within the Franciscan Complex (Photo 2). Four major streams and their tributaries flow from east to west across the study area into the Pacific Ocean. These major streams are Salmon Creek, Americano Creek (called Estero Americano at the stream mouth where it enters the Pacific Ocean), Stemple Creek (called Estero San Antonio at the mouth), and Walker Creek (Figure 4). Of the four streams, Walker Creek is the only stream that lies south of the outcrop limits of the Wilson Grove Formation and, unlike the others that drain to the Pacific Ocean, flows into Tomales Bay. The central parts of these streams have developed broad, essentially flat valleys. However, the lower parts of these streams are deeply incised across the range of hills adjacent to the coast (Photos 3 and 4). These hills are formed within rocks of the Franciscan Complex and generally are at least 30 m (100 feet) higher than the hills developed on the Wilson Grove Formation to the east. The upstream reaches of these streams abruptly abut, or terminate against, northwest trending hills within the Bloomfield fold-and thrust belt, described in this report. Wind gaps are present along drainage divides separating the upper stream valleys and Santa Rosa Valley to the east. These wind gaps generally are aligned roughly perpendicular to the regional trend of faults and associated fold axes. The significance of these windgaps to the history of stream development within the region, and the associated landscape evolution within the study area, are discussed in more detail in Section 4.2.6. With the exceptions noted above, ridges and hilltops within the central part of the study area have a generally uniform elevation of 122 to 152 m (400 to 500 feet). Flat summit areas and accordant hilltops within this area, and north of the Russian River, have been cited by previous researchers (Lawson, 1894; Osmont, 1904; Holway, 1913, Dickerson, 1922; Fox, 1983) as evidence of a regional erosional surface, initially termed the Mendocino Plateau by Lawson (1894). The presence, or absence, of this inferred regional erosional surface, is of critical importance to evaluation of Cenozoic structures within the region because: (1) the surface could be used to evaluate regional post-Wilson 11

1)

English Hill

2)

Photograph 1. Photo of rolling hills developed within the Wilson Grove Formation, looking northeast from Sugarloaf Hill, above Dillon Beach, toward English Hill. Photograph 2. Photo looking northeast toward English Hill from Highway 1, showing typical topography associated with outcrops of rock of the Franciscan Complex.

3)

4)

Photograph 3. Photo looking southwest showing the incised lower reach of Stemple Creek across the zone of hills along the coast associated with the Sonoma Coast fold-and-thrust belt. Photograph 4. Photo looking southwest toward the Pacific Ocean showing the incised lower reach of Americano Creek across the Sonoma Coast fold-and-thrust belt.

Grove uplift; and, (2) as argued by Fox (1983), the presence of a surface with minimal vertical variation implies the lack of significant deformation since late-Pliocene time. Specifically, Fox (1983) cites the presence of this inferred surface as evidence, along with gentle (<10) dips in the Wilson Grove Formation, that minimal Quaternary deformation has occurred within the Sebastopol structural block. However, subsequent stream dissection of the inferred erosional surface, particularly deep incision (up to 300 m) by the Russian River, is consistent with substantial late Quaternary regional uplift. Below we discuss evidence cited in the literature for and against the "Mendocino Plateau" surface. As Higgins (1952) noted, the belief in a regional erosional surface of minimal relief as first suggested by Lawson (1894), has been widespread and more or less unquestioned among subsequent researchers. 2.3.1 Mendocino Plateau Previous researchers (Lawson, 1894; Osmont, 1905; Holway, 1914; Gealey, 1951; Fox, 1983) have noted that accordant rounded summits of the low hills and ridges in the central part of the Sebastopol area appear to define an undulating surface. This surface, first termed the "Mendocino Plateau" by Lawson (1894), has been described as a regionally extensive erosional surface inferred to extend northwestward from the Sebastopol region at least as far north as Fort Bragg. As defined by Fox (1983) the Mendocino Plateau cuts equally across the Franciscan basement and the Wilson Grove Formation and, north of Fort Ross, the Ohlson Ranch Formation of Higgins (1960). However, as originally argued by Lawson (1894), the Mendocino Plateau formed as the result of initial regional marine planation, followed by deposition and uplift of marine sediments, and subsequent planation of the uplifted sediments. As Higgins (1952) noted, this implies two distinct peneplains with the former once producing the basal surface of the Wilson Grove Formation and the younger one beveling the Wilson Grove Formation. However, Lawson (1894) presented no evidence to support a post-Wilson Grove erosion surface other than the general accordance of ridgecrests and hilltops across the region. Osmont (1905) cited evidence of phola-borings within Franciscan bedrock exposures on the road from Bodega Bay to the town of Bodega. Higgins (1952) was unable to confirm the presence of these borings although he noted two large outcrops of the basal Wilson Grove Formation near Bodega that he thought Osmont (1905) might have mistaken for remnants of the Mendocino Plateau. During field reconnaissance for this study we also were unable to locate these borings and found no evidence supporting a regionally extensive, post-Wilson Grove erosional surface. Holway (1914), and subsequent researchers, including Fox (1983), have primarily cited the presence of flat ridgetops and the findings of Lawson (1894) and Osmont (1905), to support the existence of the Mendocino Plateau surface. Within the study area, the rolling lowlands within the Wilson Grove Formation have tens of meters of relief between "accordant" summits and ridge crests within rocks of the Franciscan Complex to the north locally have over 300 meters of relief. Higgins (1952) noted that the upper surfaces typically coincide with the contact between the Wilson Grove Formation and Franciscan Complex, and thus appear lithologically controlled. For example, based on our field reconnaissance, several broad ridge-tops south of Stemple Creek are evidently resurrected remnants of the pre-Merced erosion surface cut into Franciscan rocks (Photo 5). Only lower topographic spurs within the study area appear to be possible post-Merced erosional surfaces. The absence of prominent sea stacks, sea cliffs, and old beach deposits were cited by Higgins (1952) as evidence that these lower surfaces may have formed by subaerial exposure rather than regional marine planation. We examined several of these lower surfaces for this study, and, where fluvial gravels are

15

Photograph 5. Photo looking southeast, near Dillon Beach, of an exhumed erosional contact between poorly consolidated marine sands within the Wilson Grove formation and underlying resistent beds.

preserved, have mapped these surfaces as older terrace remnants deposited by ancestral streams. The origin of the accordant hilltops within the study area thus appears to be the result of a combination of lithologic factors, primarily the contrast between erodible sands within the Wilson Grove Formation and well indurated rock of the Franciscan Complex, and exhumation of a Pliocene erosional surface. We agree with Higgins conclusion that the rolling surface in the area underlain by the Wilson Grove Formation likely was formed by subaerial erosion via ancestral coastal streams (Higgins, 1952). Additionally, locally exhumed, or "stripped", portions of the erosional surface cut on the Franciscan Complex prior to, or during, deposition of the basal Wilson Grove Formation are associated with flat ridge and hilltops within the region. Higgins (1960) noted that the Wilson Grove Formation, and the Ohlson Ranch Formation to the north, clearly have been uplifted and internally deformed as part of regional uplift. Most of the deformation within the Ohlson Ranch Formation, documented on the basis of detailed contouring of the base of the Ohlson Ranch deposits (Higgins, 1960), is associated with what he interpreted as discrete block faulting. Where stripped, these uplifted blocks typically have formed hills with flat tops (e.g., English Hill). It is remnants of this pre-Wilson Grove Pliocene landscape, or rather seascape, that we contour in the area south of the Russian River to help constrain regional tectonic deformation.

17

3.0 TECHNICAL APPROACH Quaternary mapping, field reconnaissance, and geomorphic techniques were utilized to assess the location, style, and, where possible, estimate rates of late Quaternary contractional deformation within the region between the Rodgers Creek and San Andreas faults. The initial phase of study involved stereoscopic interpretation of conventional aerial photographs. Air photo interpretation was used to identify and map Quaternary geomorphic surfaces, develop a Quaternary stratigraphic framework for the study area, and identify sites for field reconnaissance. Four vintages of aerial photographs, flown at various dates and various scales, were incorporated into the mapping (Table 1). Table 1. List of Aerial Photography Used in Reconnaissance Mapping. Date/Series Approx. Scale 6/7/42 1:20,000 COF series 5/4/61 CSH series 6/27/65 SON series 6/6/74 unnamed 1:22,000 Image Type Black-and-white Agency USDA USGS 7.5 Quadrangles Valley Ford, Sebastopol, Two Rock, Valley Ford, Cotati, Santa Rosa Duncan Mills, Bodega Head, Valley Ford, Tomales, Camp Meeker, Sebastopol Duncan Mills, Bodega Head, Valley Ford, Tomales, Camp Meeker, Sebastopol Coastal strip within Duncan Mills, Bodega Head, Valley Ford, Tomales

Black-and-white

USGS

1:12,000

Black-and-white

USGS

1:20,000

Color

USGS

Geomorphic features identified on the aerial photographs were transferred to 1:24,000scale topographic maps and supplemented by field reconnaissance. Field reconnaissance helped verify and refine geologic contacts and collection of information on deposits associated with mapped geomorphic surfaces in readily accessible exposures (e.g., stream banks and roadcuts). Quaternary surfaces identified on the aerial photographs were verified in the field, where possible, and correlated using geomorphic and stratigraphic position, relative degree of surface modification, physical continuity within drainage basins, and, to a limited extent, relative degree of soil-profile development. Fluvial terrace ages are generally constrained by vertebrate fossils collected by previous researchers and documented in the literature, and by correlation to marine terrace levels based on detailed mapping. Vertebrate fossils collected from terrace deposits within the study area include mammoth, mastodon, horse, and bison (Stirton, 1939). Unfortunately, as is abundantly documented in the literature (e.g. Johnson, 1934; Travis, 1952; and Higgins, 1952), exposures of Quaternary deposits associated with these mapped surfaces are rare. Additionally, where exposed, semi-consolidated sands and gravels of the Pleistocene stream terraces of interest are almost identical to deposits of the Wilson Grove Formation. This similarity is due, in large part, to these deposits typically being derived either from the Wilson Grove Formation itself or from similar sources (e.g. reworking of weathered rocks of the underlying Franciscan Complex). An important exception, however, is the presence of rare clasts of volcanic rock, typically andesite. As noted above, Johnson (1934) documented the absence of a significant volcanic component within 18

the Wilson Grove Formation. As discussed below in Section 4.2.6, the presence of volcanic clasts within the fluvial terrace gravels may provide information on the original source areas of the gravels, and thus, the original drainage areas of ancestral streams that once flowed across the region. We conducted geomorphic analyses of the mapped Quaternary surfaces to: (1) evaluate possible tectonic influences on the geomorphic development of the study area; (2) assess evidence for late Quaternary growth of folds; and (3) constrain uplift rates across folds and faults. Analyses performed for this study include construction of a structure contour map on the base of the Wilson Grove Formation, construction of longitudinal terrace and stream channel profiles, analysis of stream channel morphology using stream-gradient index, evaluation of transverse stream valley morphology across the study area, and assessment of the late Quaternary evolution of fluvial drainages across the region. Results from these techniques were integrated with our review of published literature and field reconnaissance to assess evidence for late Quaternary deformation. First, we contoured the base of the Wilson Grove Formation based on detailed bedrock mapping by Dickerson (1922), Johnson (1934), Weaver (1949), Gealey (1951), and Travis (1952), supplemented by our air photo interpretation and field reconnaissance. As described above, the Wilson Grove Formation was deposited within a shallow marine embayment upon an erosional surface with minimal initial vertical relief (Johnson, 1934; Bedrossian, 1969; Bedrossian, 1974). We plotted elevations of the contact between the Wilson Grove Formation and the underlying Franciscan Complex, where exposed, and used these points to contour the base of the Wilson Grove Formation across the study area. As part of our analysis we constructed regional cross-sections showing folding and faulting of the basal contact of the Wilson Grove Formation based on mapped exposures and, within Santa Rosa Valley, water and oil-test well data. Second, we examined the distribution of terraces along major drainages within the study area including the Russian River, Salmon Creek, Americano Creek, Stemple Creek, and Walker Creek. In addition, we examined the reach of Santa Rosa River incised across the floor of the Santa Rosa Valley. Of these drainages, we constructed longitudinal terrace profiles along the Russian River, Americano Creek, and Stemple Creek because these drainages contain distinct, well-preserved sequences of fluvial terraces and flow across contractional structures. Surficial geologic maps of terrace and other geomorphic surfaces constructed for this study were used to construct the longitudinal terrace profiles. Deviations in the profile from an initially graded terrace profile were used to identify loci of tectonic deformation and, when integrated with age constraints, provide a basis for estimating rates of Quaternary deformation. As part of our analysis of late Quaternary deformation, we also constructed longitudinal profiles of stream channels along major drainages to help identify the location and style of deformation. Field studies by Burnett and Schumm (1983), Merritts and Vincent (1989), Bullard and Lettis (1993), and Marple and Talwani (1993) show that changes in stream channel gradient may record tectonic uplift. Along streams that are unaffected by regional or localized uplift, stream gradient generally decreases with distance downstream, typically in an exponential or logarithmic manner producing a characteristic convex graded profile (Hack, 1957; Schumm, 1977). Longitudinal profiles of stream channels within homogeneous terrain typically are convex upward across an axis of uplift (e.g. Figure 5). Thus downstream variations in channel gradient that are not logarithmic or exponential, typically expressed as a local convexity or concavity in the channel profile, may be related to localized uplift or subsidence.

19

a) STREAM TERRACE FORMATION

INCREASED

SINUOSITY

INCISION

CHANNEL STRAGHTENING

b)

Terrace

c) Terrace Stream Channel Profile

Stream Channel convexity

Figure 5. Generalized conceptual model showing; (a) map view of possible stream changes across localized uplift, (b) three-dimensional model of stream incision and terrace formation across uplift and, (c) longitudinal stream profile showing possible variations in channel profile produced across localized uplift.

In addition, localized uplift or subsidence may change the stream channel gradient, upsetting the equilibrium between channel gradient and hydraulic properties of the stream (Ouchi, 1985). Localized disruption of these properties may be expressed as straightening of the channel on the upstream side of an uplift and incision across the central reach of the uplift with associated development of a series of fluvial terraces (Ouchi, 1985; Figure 5). Often these terraces are absent, or poorly preserved, elsewhere along the stream channel. These terraces can be diagnostic of localized uplift and useful in evaluating the amount of folding or faulting across the uplift. Downstream of an axis of uplift, stream sinuosity generally increases as the stream attempts to maintain a uniform stream gradient on the steepened slope. For this study, the stream-gradient index (Hack, 1973), that relates channel gradient to channel length is used to analyze the longitudinal stream profiles. As defined by Hack (1973), the stream-gradient index is equal to the product of the channel slope at a point and total stream length from the drainage divide to that point as shown: SL (stream gradient index) = H/L*L (1)

where H/L is the channel slope or gradient (H is the change in elevation of the stream reach and L is the length of the reach), and L is the total channel length from the stream divide. The stream-gradient index is a means of quantifying stream power, with higher index values representing stream reaches with greater available power to incise and transport sediment. The stream-gradient index is very sensitive to changes in channel gradient, which generally correspond to differences in bedrock strength, particle size, and load supplied to the stream. Where climatic factors and lithological controls on the fluvial system are effectively constant, anomalously high stream-gradient values typically are interpreted to indicate localized uplift (Keller and Pinter, 1996). Unfortunately, rock strength varies significantly across the study area, and this technique must be used with caution. Poorly consolidated marine Pliocene deposits of the Wilson Grove Formation underlie the central portion of the study area (Figure 3). Based on outcrop and topographic expression, the most resistant rocks in the study area are rocks of the Franciscan Complex. However, there is significant variability within the Franciscan Complex. For this reason, we present index data only as one line of evidence among several. Where presented, the stream-gradient index is calculated for stream reaches underlain by similar bedrock units and any variations between reaches are identified and discussed. Climatic and cultural changes are an additional source of error that may have obscured or, at worst, produced changes in the channel profiles and stream-gradient index that appear tectonically controlled. For the purpose of our geomorphic analyses, climatic influences are assumed minor. The late Pleistocene climate of the coastal areas of northern California was characterized by alternating glacial and interglacial periods but variations along the coast were of much less magnitude that of continental interior regions, based on vegetation changes (Johnson, 1977). However, early settlers and cattle grazing have had, and continue to have, profound impacts on erosion within the area. For example, a small tributary of Walker Creek was formerly the tidewater port for the town of Tomales about 1860. At that time, the tributary had enough water at high tide for schooners and a small steamer to bring supplies. The harbor was abandoned around 1875 because cultivation of the former grass-covered slopes had resulted in rapid erosion and associated filling of the channel (Holway, 1914). Where possible, we attempt to identify any climatic or cultural changes that may influence our interpretation. Stream valley morphology is used in the present study to describe changes in stream valley morphology in relation to inferred Quaternary contractional structures. Specifically, the ratio of valley height to valley-floor width provides a measure of geomorphic changes in 21

relation to localized Quaternary uplift. Bull and McFadden (1977) expressed transverse valley morphology as a ratio of valley-floor width to valley height (Vf ), that may be defined as: Vf = 2Vfw / ((Eld - Esc) + (Erd - Esc)) (2)

Where Vfw is the width of the valley floor, Eld and Erd are elevations of the left and right valley divides, and Esc is the elevation of the stream channel. For this study, the ratio was along Salmon, Americano, and Stemple Creeks and used to compare valley morphology across hills associated with uplift and between belts of hills. Channel incision typically is strongly influenced by a relative base-level fall due to tectonic uplift. Broad-floored valleys may be indicative of areas of subsidence or areas that are uplifting at relatively low rates. Narrow valleys typically reflect stream incision that commonly is associated with tectonic uplift. Finally, small-scale regional topographic maps (historic 1:62,500-scale Santa Rosa and Sebastopol quadrangle with 50 ft contours; and the 1:100,000-scale Napa quadrangle with 50 m contours) were used to assess the geomorphic evolution of the study area. During our limited field reconnaissance, we examined major stream and wind gaps carefully for remnants of deposits. Relative correlation of the windgaps and stream gaps is attempted but such correlations are only tentative, as we uncovered no datable material during our field investigations.

22

4.0 RESULTS 4.1 Quaternary Stratigraphy Thirteen Quaternary map units are present within the study area based on interpretation of aerial photographs and field reconnaissance. From oldest to youngest, map units include: (1) a sequence of five marine terraces along the coast between the mouth of the Russian River at Jenner and Tomales Bay; (2) older alluvial fan deposits (unit Qof) within Santa Rosa Valley; (3) isolated fluvial terrace remnants (unit Qt1) on the divide separating Santa Rosa Valley from streams to the west; and (5) a sequence of three to five fluvial terraces along the Russian River, Salmon Creek, Americano Creek, Stemple Creek, and Walker Creek (units Qt1 to Qt5). Of these mapped deposits, and associated surfaces, the marine terraces and fluvial terraces are the most important for our evaluation of tectonic deformation and are described in more detail below. 4.1.1 Marine Terraces The section of coastline between the Russian River and Bodega Bay contains some of the best-developed and preserved terraces along the entire Sonoma coast (Figure 6). As part of our mapping, we compiled mapping by Travis (1952), Bauer (1952), and unpublished reconnaissance mapping by K. Lajoie and J. Tinsley of the USGS. Bauer (1952) mapped five terrace levels along the coast between the Russian River and Salmon Creek. We conducted limited field reconnaissance to delineate and correlate the major terraces. Because of the rugged terrain and lack of access, with the exception of well-exposed terraces near Bodega Bay and Dillon Beach, we primarily relied upon aerial photographic interpretation to map discontinuous terraces along the coast south of Bodega Bay. Marine terraces located between Bodega Bay and Dillon Beach, and fluvial terraces along the lower reaches of Americano and Stemple Creeks, are shown on Figures 7 and 8. The lowest marine terrace (map unit Qmt5) is expressed as broad, well-preserved surfaces between the Russian River and Bodega Bay (Figure 6) and discontinuous remnants between Bodega Bay and Dillon Beach (Figure 7 and 8). The terrace wave-cut platform is veneered by up to one meter of marine deposits based on exposures within gullies along the coast (Photo 6). The terrace deposits consist of well-bedded sand and gravel with wellrounded pebbles that typically seldom exceed 4 cm in diameter. We infer that the lower marine terrace (Qmt5) is associated with the stage 5e (125 ka) sea level high stand, the highest late Pleistocene highstand formed when the sea level was at least 6 to 8 m higher than present. This correlation is based on the geomorphic similarity of the broad terraces to those formed during the stage 5e elsewhere along the California coast (Kennedy et al., 1982) and to broad surfaces associated with the Millerton Formation within Tomales Bay, south of the study area. Samples collected from the Millerton Formation within Tomales Bay that are associated with surfaces at roughly the same elevation as the lower marine terrace Qmt5, yielded a TL age of 13412 ka (G.W Berger cited in Grove and others, 1995). This age is consistent with a similar aminostratigraphic age that suggests that parts of the Olema Creek and Millerton Formations were deposited simultaneously during the 5e substage highstand of sea level (Grove and others,1995). Additionally, Dwyer and Borchardt (1994) correlated an exposure of the lower terrace at Bodega Bay, location shown on Figure 6, to oxygen isotope stage 5e based on soil development.

23

Marine terrace deposits

Franciscan bedrock

Photograph 6. Exposure of marine terace deposits (Qmt5) overlying wave-cut platform (substage 5e, 125 ka) at Dillon Beach.

Older, higher marine terrace surfaces (Qmt 4 through Qmt1) are associated with wave-cut platforms that are typically more discontinuous and poorly preserved than the lowest marine terrace. We found no exposures of bedded gravels associated with these surfaces although well-rounded pebbles are occasionally present. The slopes of these surfaces typically dip westward (seaward) at about three and 1/2 percent (Bauer, 1952) and the terrace deposits locally dip westward up to 20 degrees (Travis, 1952). 4.1.2 Fluvial Terraces Five terrace levels are present along the lower Russian River (Gealey, 1952; Higgins, 1952; this study). These terraces, typically preserved as isolated remnants within rugged and often inaccessible terrain, were first mapped in detail and correlated by Higgins (1952). During our field reconnaissance, we examined several of the localities described by Gealey (1952) in the northeastern portion of the study area and by Higgins (1952) along the lower Russian River, but found no datable materials. As noted by Higgins (1952) the Russian River terraces are cut, without exception, on rocks of the Franciscan Complex and the deposits can be differentiated from Franciscan conglomerate primarily by the presence of pebbles derived from the Sonoma Volcanics exposed to the east. South of the Russian River, a sequence of two to four distinct fluvial terraces are identified along Salmon, Americano, Stemple, and Walker Creeks. These terraces are most poorly developed, or preserved, along Salmon Creek and Walker Creek. The deposits underlying these terraces are designated, from oldest to youngest, mapping units Qt1 to Qt4. Unit Qt4 is relatively widespread, preserved as broad surfaces along the central portions of Americano and Stemple Creeks. Where exposed, unit Qt4 is preserved as thin (<0.5 m thick) fluvial gravels associated with strath (fluvially cut) surfaces. In areas of poor exposure, bedding within the Wilson Grove Formation and/or the contact between the basal Wilson Grove Formation and underlying rocks of the Franciscan Complex can appear similar to unit Qt4 (Photos 7 and 8). Units Qt2 and Qt3 locally are preserved as isolated, thin deposits of fluvial gravel and gravelly sand associated with strath surfaces cut either on bedrock of the Wilson Grove formation or on rocks of the Franciscan Complex (Photos 9 and 10). Unit Qt1 forms isolated flat surfaces on topographic spurs and lower hilltops and is preserved within prominent wind gaps. 4.1.3 Correlation and Age of Map Units The ages of marine terraces are estimated, as shown in Figure 9, by: (1) extrapolating the derived uplift rate (after Chappell, 1974) and; (2) correlating terraces with sea-level highstands (after Pillans, 1983; Lajoie, 1986). First, the elevation of the wave-cut platform of unit Qmt5, currently 242 m above sea level and assumed to be correlative with the stage 5e (125 ka) sea level high stand (approximately 6 m above current sea level), was used to obtain an uplift rate of about 0.14 mm/yr. Second, the elevations of the older terraces were plotted, along the slope of the derived uplift rate, versus inferred age. Finally, the elevations of the marine terraces above sea level were correlated to the sea level curve of Chappell (1983), as calibrated to oxygen isotope stages of Shackelton and Opdyke (1973). The slope of the correlation lines, equivalent to the derived uplift rate, are parallel as required by the method described by Pillans (1983) and Lajoie (1986). Fluvial terraces in the study area were identified and correlated primarily by geomorphic and stratigraphic position within the terrace sequence, relative degree of surface modification, presence of fossils, and correlation to marine terrace levels. Due to the lack of good exposures, and minimal thicknesses of the terrace gravels associated with strath surfaces, we were not able to evaluate relative soil profile development with confidence. 28

7)

AREA OF PHOTO 8

8)

Wilson Grove Formation outcrop

Photograph 7. Photo looking northeast across Estero Americano showing the broad alluvial valley and stream terraces east of the Sonoma Coast fold-andthrust belt. Photograph 8. Photo looking east at exposure of deposits within the Wilson Grove Formation showing how bedding can mimic a fluvial terrace surface.

9)

Terrace deposits

Wilson Grove formation

10) 10 a)

10 cm

Photograph 9. Photo of exposure of stream terrace deposits (Qt3) overlying the Wilson Grove Formation near Valley Ford. Photograph 10. Photo of stream terrace deposits in contact with Franciscan bedrock, south of bridge crossing of Stemple Creek on Highway 1. Photo 10a is a close-up of the deposits showing the sub-rounded terrace gravels (Qt3).

100
Elevation (in m) Qmt2 Qmt4 Qmt5 Qmt3 Qmt1

m/ka 0.14 plift = of u rate

0
e 10 SEA 13 12 11

100

200

300 400 Age (ka)

500

600

8
LEVEL

2
a

9 O STAGES
18

1 -100

-160 500 400 300 Age (ka) 200 100 0

Figure 9. Diagram showing tentative correlations and general uplift history of marine terraces along the Sonoma coast between the Russian River and Dillon Beach. The numerals (1 through 13) are oxygen isotope stages from Shackelton and Opdyke (1973) and the sea level curve is from Chappel (1983).

Therefore, ages of the fluvial terraces are constrained by correlation to marine terrace levels and vertebrate fossils previously identified within the fluvial terrace deposits. For example, a thin fluvial terrace remnant at an elevation of 30.5 m (100 feet), located 2.4 km (1.5 miles) N55E of Dillon Beach contained Rancholabrean fauna consisting of a mastodon molar, three mammoth molars, and part of a bison horn collected by R. A. Stirton (located on Figure 8 based on independent citations in Johnson, 1934; Higgins, 1952). Estimates for the beginning of the Rancholabrean range from about 10 ka to 550 ka. The earliest record of Bison in the conterminous United States, based on Bell (1998), is from Teichert Quarry near Sacramento where it was assigned an age of approximately 130 ka (Repenning and others, 1990). Therefore, we infer that this terrace, mapped as unit Qt3, likely is approximately 130 ka or older, but no older than 550 ka Additional Rancholabrean fossil localities previously identified along major drainages within the study area are used to correlate fluvial terraces across the region (Figure 4). Remains of the ground sloth Mylodon harlani were found in fluvial gravels (mapped as unit Qt2) approximately 11 km (7 miles) northwest of Petaluma, west of the main road connecting Stony Point with Cherry. Rancholabrean fossils found in fluvial gravels along Ebabias Creek (Section 18, Township 6N Range 9W), a tributary of Americano Creek, include Artiodactyla Bovidae (UCMP #V72107, specimen number 119351), and the mid humerus of Perissodactyla Equidae (UCMP #V72107, specimen number 119350). These gravels, and the fossils they contain, are used to correlate fluvial terrace unit Qt3, and the fluvial terrace sequence, across the study area. Approximate ages of the fluvial terrace sequence are estimated via correlation to the marine terrace sequence along the coast (Table 2). These correlations are based on detailed mapping of fluvial and marine terraces the lower stream reaches of Americano and Stemple Creeks adjacent to the coast (Figures 7 and 8). Two of the four major fluvial terraces (units Qt2 and Qt3) grade to marine terrace levels (Qtm3 and Qtm4 respectively). Correlation and estimated age of the oldest fluvial terrace (Qt1) is based primarily on relative position within the terrace sequence and is uncertain as no well-preserved remnants of this terrace were mapped at or near the coast. The age of the youngest terrace (Qt4) is tentatively estimated between 25 ka and 130 ka, based on relative elevation within the terrace sequence. Table 2. Estimated ages and correlation of marine and fluvial terraces. Marine Terrace Qmt1 Qmt2 Qmt3 Qmt4 Qmt5 Elevation above sea level (in m) 793 733 612 522 242 Estimated age (in ka) 516 472 385 320 125-130 Fluvial Terrace Qt1 Qt2 Qt3 Qt4 Elevation above modern stream channel (in m) 73 to 76 27 to 39.5 18.2 to 25.1 6 to 12.2 Estimated age (in ka) 385-472 320 130 25-130

4.2 Geomorphic Analyses Deformation within the Sebastopol structural block by past researchers almost exclusively has been evaluated based on bedding dips and mapped offset of units within the "Merced"

32

Formation of Johnson (1934), renamed the Wilson Grove Formation by Fox (1983). Because the average dip of the Wilson Grove Formation typically is less than 10, uplift of the region was thought to have occurred with minimal internal deformation of the block (Fox, 1983; Furlong et al., 1989). However, our structure contour map on the base of the Wilson Grove Formation documents up to 300 meters of uplift, and minimum vertical separation of at least 183 m (over 600 feet) across the Bloomfield fault (Figure 10). Below we discuss the deformation of the base of the Wilson Grove Formation, based on our structure map and a regional east-west cross-section showing inferred folding and faulting of the Wilson Grove Formation across the study area. Because the Rodgers Creek fault system likely developed within the last 2-4 Ma (Helley and Herd, 1977), the rate and distribution of late Quaternary strain within the region may not wholly reflected by deformation of the Wilson Grove Formation. In particular, the timing of the initiation of faulting and folding within the Wilson Grove is not well constrained. Therefore, we evaluate deformation of the Wilson Grove Formation in conjunction with other strain indicators in the late Quaternary geologic and geomorphic record. 4.2.1. Contours on the base of the Wilson Grove Formation Contoured elevations of the base of the Wilson Grove Formation show substantial regional uplift across the Sebastopol structural block, with maximum uplift within the central portion of the block roughly equidistant between the San Andreas and Rodgers Creek faults (Figure 10). Across the Sebastopol block, the base of the Wilson Grove Formation has an overall antiformal shape with maximum elevations of approximately 200 to 330 m (up to 1100 ft) above sea level. Bedding dips in the Wilson Grove Formation show a general northeastward tilting (8 to 12) of the formation between San Andreas fault and the Santa Rosa Valley. This deformation is consistent with uplift and tilting of the base of the Wilson Grove Formation across the Sebastopol block. Locally, the base of the Wilson Grove Formation is offset across the northwest-trending Americano Creek, Joy Woods, Bloomfield, Dunham, and Burdell Mountain faults (Figure 11). Contours on the base of the Wilson Grove Formation are consistent with anticlinal folding within the hanging walls of these inferred reverse faults. Specifically, a series of northwest stepping anticlines are mapped within the hanging walls of the Dunham, Bloomfield, and Joy Woods faults (Figure 10). These folds have exhumed the basal Wilson Grove Formation locally up to an elevation of 330 m (1100 ft). We call this series of folds and related faults, the Bloomfield fold-and-thrust belt. The contours on the base of the Wilson Grove Formation show that localized contraction within the fold-and-thrust belt progressively steps to the northwest from the Dunham and Bloomfield faults to the Joy Woods and related faults. The Wilson Grove Formation extends beneath Santa Rosa Valley where it is covered by younger deposits (Figure 11). The Wilson Grove has been identified in the subsurface beneath Santa Rosa Valley based on the presence of diagnostic fossiliferous sandstone within water and oil-test wells. For example, oil-test well 7/9-24ja, drilled in 1948, and located 4.2 km (2.6 miles) northeast of Sebastopol in Santa Rosa Valley penetrated marine deposits of the Wilson Grove Formation from a depth of 59 m (195 feet) to the bottom of the well at 632 m (2,075 feet) except for an interval of continental deposits between depths of 187 m and 251 m (615 to 826 feet) (Cardwell, 1958). Adjacent to the Sonoma coast, the basal Wilson Grove Formation has been uplifted over 150 m (500 ft) and locally stripped within a 3 to 5 km (2 to 3 miles) wide zone. Elevations of the base of the Wilson Grove decrease inland as the contact between the Wilson Grove 33

A
Southwest
1,000

A
Northeast Twg

Wilson Grove formation elevation (in feet)


Americano Creek Laguna de Santa Rosa Santa Rosa Valley

sea level
eld fault Bloomfi

Glen Ellen Formation Wilson Grove Formation Franciscan Complex Petaluma Formation

Franciscan Complex

-1,000

V.E. =10x

Figure 11. Southwest-northeast cross-section A-A from the Pacific Ocean to the eastern margin of Santa Rosa Valley.

Formation and underlying Franciscan Complex flatten to sub-horizontal. Inland of the Sonoma Coast fold-and-thrust belt, the basal contact is essentially horizontal and the Franciscan Complex is exposed by lateral incision of Americano and Stemple Creeks. These windows exposing the Franciscan Complex trend roughly southeast-northwest and coincide with the projection of Salmon Creek fault and unnamed folds mapped within the Franciscan Complex southwest of the Burdell Mountain fault by Weaver (1949). We interpret this pattern as possibly reflecting synclinal folding between the Bloomfield and Sonoma Coast fold-and-thrust belts. Interestingly, the inferred syncline roughly coincides with the submerged portion of Americano Creek, occupied by Estero Americano, possibly reflecting localized subsidence. 4.2.2 Fluvial Terrace Longitudinal Profiles Fluvial terrace profiles provide a means of assessing both regional deformation and localized deformation across specific folds and faults. Detailed longitudinal profiles of terraces along two of the four streams that cross the study area (Americano and Stemple Creeks) were constructed from 1:24,000-scale USGS topographic maps (Stemple Creek is shown in Figure 12). These drainages have at least one widespread, and relatively well preserved, terrace extending across transverse Cenozoic structures, including the Bloomfield fault. Salmon Creek, located north of the outcrop of Wilson Grove Formation, and Walker Creek, located south of the outcrop of the Wilson Grove Formation are both deeply incised into the Franciscan Complex. These streams do not have well-developed and/or well-preserved terrace sequences. 4.2.2.1 Americano Creek Americano Creek has its headwaters currently within the distinctive hills northeast of the Bloomfield fault. As described earlier, windgaps in the hills to the east indicate that the headwaters of the creek at one time extended into the Santa Rosa Valley. The creek flows southwest across the Dunham and Bloomfield faults before abruptly turning westward, flowing sub-parallel to the Americano Creek fault to the south. With the exception of incision across a block of uplifted Franciscan Complex within the hanging wall of the Bloomfield fault, Americano Creek flows within deposits of the Wilson Grove Formation between Bloomfield and Valley Ford. West of Valley Ford, Americano Creek is incised below the contact between the Wilson Grove Formation and underlying Franciscan Complex (Figure 7). Approximately 2 km west of Valley Ford, American Creek flows into Estero Americano, a broad estuary formed by late Holocene sea level rise, or tectonic subsidence, drowning the former mouth of the creek. Several terrace remnants are preserved within the rugged range of hills formed by the Sonoma Coast fold-and-thrust belt adjacent to the San Andreas fault. We identify a sequence of three distinct terraces and higher terrace remnants (units Qt4 through Qt2, and unit Qt1) along Americano Creek. The discontinuous nature of the terrace sequence along Americano Creek, presence of exhumed portions of the Wilson GroveFranciscan contact west of Valley Ford, and the fault-parallel orientation of the stream reach within the fault block between the Bloomfield and American Creek faults makes interpretation of the profile problematic. Specifically, the central reach of Americano Creek near Valley Ford contains numerous surfaces that appear to be terraces but likely are either developed on horizontal bedding within the Wilson Grove Formation and/or the contact between the Wilson Grove Formation and the underlying Franciscan Complex. 4.2.2.2 Stemple Creek Stemple Creek has its headwaters currently within the distinctive hills northeast of the Bloomfield fault. Stemple Creek flows southwest across the Dunham and Bloomfield

36

faults before turning westward. Within its central reach between Two Rock and the Sonoma Coast fold-and-thrust belt, Stemple Creek is incised into a broad bedrock floor. We identify a sequence of three distinct terraces (units Qt3, Qt2, and unit Qt1) along Stemple Creek (Figure 12). The longitudinal terrace profile shows maximum elevations of the fluvial terraces upstream of the Bloomfield fault (within the hanging wall of the fault), and across the Sonoma fold-and-thrust belt adjacent to the coast. Fluvial terrace Qt2 is at a maximum elevation of 39.5 m and fluvial terrace Qt3 at a maximum elevation of 18.3 m within the hanging wall of the Bloomfield fault. The maximum elevations above sea level for fluvial terrace Qt2 across the Sonoma fold-and-thrust belt are 36.5 m and 30 m for fluvial terrace Qt3. These areas of locally elevated terraces coincide with localized uplift of the base of the Wilson Grove formation (shown on Figure 12). We therefore conclude that the fluvial terrace profile shows Quaternary warping across the Bloomfield fold-and-thrust belt and the Sonoma Coast fold-and-thrust belt. 4.2.3 Marine Terrace Profile A coast-parallel profile of marine terraces between the Russian River and Bodega Bay provides evidence of general uplift along the coast (Figure 13). Based on the outcrop of the Wilson Grove Formation, maximum uplift and stripping of the Wilson Grove deposits has occurred roughly adjacent to the coast. Thus, marine terraces between Salmon Creek and Dillon Beach likely are located at, or near, the crest of the uplift along the inferred Sonoma Coast fold-and-thrust belt. As described above, and illustrated in Figure 9, our tentative correlation of these terraces suggests a relatively low uplift rate of approximately 0.14 mm/yr along the Sonoma Coast fold-and-thrust belt. South of Bodega Bay, the maximum uplift of the basal Wilson Grove Formation and, by inference the crest of the uplift associated with the Sonoma Coast fold-and-thrust belt, is located inland approximately 0.5 to 1 km (e.g. Figure 12). West of this uplift, based on inspection of aerial photographic mapping transferred to topographic maps, several marine terraces appear to be tilted westward toward the ocean. In addition, Travis (1952) noted that terrace deposits dip westward up to 20 degrees about 1.7 km (1 mi) north of Dillon Beach. Although we were unable to obtain access to the locality mentioned by Travis (1952), the observed westward tilting of the terrace deposits suggests that the marine terraces have been progressively tilted within the western flank of folds within the Sonoma fold-and-thrust belt. 4.2.4 Fluvial Channel Profiles Detailed longitudinal profiles of the four streams that cross the study area (Salmon, Americano, Stemple, and Walker Creeks) were constructed from 1:24,000-scale USGS topographic maps. For this study, average stream-gradient index values were calculated for stream reaches with similar gradients, which are associated with linear slopes on semilogarithmic stream profiles. Merritts and Vincent (1989) have shown that stream-gradient index values can be useful in distinguishing areas with high uplift rates from areas with low to intermediate rates. Specifically, in areas of rapid uplift, index values typically are high and stream profiles are convex on semi-logarithmic plots. Rock resistance is closely correlated to stream-gradient index values as index values typically increase across hard rock and decrease against soft rock (Hack, 1973). Salmon Creek and Walker Creek are both incised within rocks of the Franciscan Complex. However, Americano and Stemple Creeks are incised within both the Wilson Grove Formation and rocks of the Franciscan Complex. Therefore, for the purpose of this study,

37

AMERICANO CREEK FAULT

500 UNNAMED FAULT

LEGEND
400

Wi l

e rov nG so

1
H

Elevation (ft)

J J
H

300

Wilson Grov e Fm

UNNAMED FAULT

Qmt4 Qmt5

BLOOMFIELD FAULT

Qt2 Qt3 Qt4 modern stream channel

Grove F Wilson

Qt2
11

200

Qmt4
1 1

Qt2
H H 11

Wil so nG rov e
;; H

Qt3
J

J JJ J

JJ J

m
1 1 1 1 JJ J
JJ J

100

Qmt5
JJ J J

Qt3
M M
D

Qt4
D

HH D H H HH D D

J J
6

JJ J 6

J J J JJ J H

JJ JJ H H J J HH H H
J J JJ J

Wilson Grove Fm

H H H H HH

Qt4
50000

nnel modern stream cha

80000

70000

60000

40000

30000

20000

10000

Horizontal Distance (in feet)

Figure 12. Longitudinal terrace profile along Stemple Creek.

B northwest 400
elevation (in feet)
Qmt1
11 1 111

B southeast

Qmt2
3 3

300
Qmt3 Qmt3

Qmt2 Qmt3
D 33 33 33

200 100 0

Qmt4

Qmt4

Qmt5

Qmt5

10000

20000

30000

40000

50000

60000

70000

Horizontal distance (in feet)

Figure 13. Profile B-B of marine terraces between the Russian River and Bodega Bay (see Figure 6 for location)

the stream-gradient index primarily is used to identify anomalously high index values within stream reaches underlain by a particular rock type. 4.2.4.1 Salmon Creek The average stream-gradient index values for Salmon Creek range from 136 to 523 gradient-feet for eight stream segments and are illustrated on a semi-logarithmic stream profile (Figure 14). The profile of Salmon Creek exhibits a large convexity within the upper three stream reaches. The observed convexity likely reflects the recent stream capture of an unnamed tributary that formerly flowed into Dutch Creek toward the Russian River. This stream capture coincides with a prominent wind gap, located 1 km south of Occidental, that separates Dutch Creek and upper Salmon Creek (Figure 4). Although there is no apparent convexity across the fault blocks associated with the Joy Woods fault, the steep gradient of Salmon Creek is suggestive of active fluvial response to broad uplift. 4.2.4.2 Americano Creek The average stream-gradient index values for Americano Creek range from 60 to 165 gradient-feet for five stream segments (Figure 15a). Overall, the stream profile is convex upstream of the Bloomfield fault and essentially flat downstream. A pronounced stream convexity is present between the Dunham and Bloomfield faults. However, this convexity also coincides with incision of Americano Creek across resistant rock of the Franciscan Complex between these two faults. 4.2.4.3 Stemple Creek The average stream-gradient index values for Stemple Creek range from 39 to 202 gradientfeet for five stream segments (Figure 15b). A broad stream convexity is present within rocks of the Franciscan Complex upstream of the Bloomfield fault. This convexity coincides with the remarkably linear upper portion of Stemple Creek that is aligned with a prominent wind gap to the northeast (Figure 4). The windgap is located within the northeast-trending Washoe Anticline that separates Stemple Creek from Lichau Creek to the east. This linear upper portion of Stemple Creek likely represents an older stream reach beheaded from its headwaters by uplift along the Washoe Anticline. Currently the upper drainage basin of Stemple Creek is insufficient to explain the presence of this wellentrenched, broad stream valley. The observed convexity within the hanging-wall of the Bloomfield fault therefore may represent subsequent uplift and folding of this stream reach. 4.2.4.4 Discussion Systematic variations in the stream gradient index and anomalous index values generally correspond with stream reaches traversing uplifted fault blocks and folds within the Wilson Grove Formation. In general, the highest index values occur at, or slightly upstream of, mapped northeast-dipping faults. These variations in stream index values suggest that the streams have responded to late Quaternary fold growth by adjusting the gradients of their channels. However, these variations in the stream gradient index coincide with distinct changes in rock type traversed by the streams. Therefore, the observed pattern, although indicative of tectonic uplift, may represent the influence of lithological controls on the fluvial system. Thus, although the distribution of rock types is fault controlled (i.e. blocks of rock within Franciscan Complex are bounded by major faults), the stream channel variations in themselves may not provide as reliable an indicator of recent uplift as demonstrated by Merritts and Vincent (1989). 4.2.5 Stream Valley Morphology As noted previously in Section 2.3, in their middle reaches, Salmon, Americano, and Stemple creeks meander through fairly broad valleys. Valley height to valley-floor width ratios reveal morphologic changes in the valleys of Salmon, Americano, and Stemple 40

800
STREAM CAPTURE

700

136

600 Elevation (in feet)

184

500
JOY WOODS FAULT

400

523

300

257 140

200

254 180 190

100

0 100000

10000 Distance from stream divide (in feet)

1000

Figure 14. Stream channel profile of Salmon creek shown on semilogarithmic plot. Average stream-gradient index values are shown on the left of each stream reach.

BLOOMFIELD FAULT

a)

300 Elevation (in feet)

DUNHAM FAULT

69 143

200

60

100
88

165
W. G. Franciscan Wilson Grove (W. G.) Fm. Fm. Complex

0 100000

10000 Distance from stream divide (in feet)

1000

300 Elevation (in feet)

AMERICANO CREEK FAULT

BLOOMFIELD FAULT

b)

200

56 77

100
84

39
Franciscan Complex W. G. Fm. Franciscan Complex

0 100000

202

10000 Distance from stream divide (in feet)

1000

Figure 15. Stream channel profile of: (a) Americano Creek and, (b) Stemple Creek Average stream-gradient index values are shown on the left of each stream reach.

creeks across the Sebastopol structural block. The distribution and magnitude of the changes is similar for all three drainages. Valley height to valley floor width ratios are shown in Figure 16 for the valleys of Salmon, Americano, and Stemple Creeks. High ratios depict relatively broad valleys with low-lying adjacent hills and low ratios depict narrow, deeply incised valleys. The general pattern is of low ratios of valley floor width to valley height across the belt of folding associated with the Bloomfield, Dunham, and Joy Woods faults. These low ratios represent relatively narrow valley reaches where the creeks have incised across the Bloomfield fold-and-thrust belt. High ratios are present within the central reaches of these streams, coincident with meandering stream reaches and broad valleys. At the downstream ends of the broad valleys, low ratios represent stream reaches within steep-walled canyons incised across the Sonoma Coast fold-and-thrust belt adjacent to the coast. The observed pattern is consistent with the response of these fluvial systems to long-term tectonic deformation, characterized by localized uplift coincident with the Bloomfield foldand-thrust belt at the upstream ends of these streams, and the Sonoma Coast fold-andthrust belt at the downstream ends. Between these two belts of inferred uplift, Stemple Creek, Americano Creek, and to a lesser extent, Salmon Creek have formed broad, flat stream valleys. 4.3 Stream Development and Landscape Evolution The four major southwest-flowing streams located south of the Russian River appear to have been truncated or disrupted by localized tectonic deformation between the San Andreas and Rodgers Creek faults. As noted by Holway (1914) and Higgins (1952), these streams lack extensive drainage basins, are fairly straight and narrowly spaced, are on trend with prominent windgaps preserved within the hills bounding Santa Rosa Valley, and have very short tributaries (Figure 4). The drainage divides separating these streams from the Santa Rosa Valley coincides closely with the Bloomfield fold-and-thrust belt. Additionally, as described earlier, remnants of older stream gravels preserved along these streams contain volcanic clasts possibly derived originally from Sonoma Volcanics exposed east of Santa Rosa Valley. Taken together, these data suggest that these streams are the lower or, as Holway (1914) phrased it, "mature" reaches of what were once more regionally extensive streams that likely had drainage basins within the Sonoma Mountains east of Santa Rosa Valley. The value of this observation is that it may help provide general constrains on the development of the Santa Rosa Valley, and indirectly, constrain the timing of initiation of slip on the Rodgers Creek fault. An example of possible tectonic disruption of regional drainages is the northern end of Stemple Creek, which is aligned with a prominent wind gap that crosses the Washoe Creek anticline near the town of Cotati (Figure 4). The windgap is within a ridge that separates the Petaluma and Santa Rosa Valleys. Deposits of an abandoned stream channel that may have been a remnant of ancestral Stemple Creek (Higgins, 1952) are located approximately 0.4 km northwest of Penngrove, east of the windgap. The east-west direction of stream flow in Pleistocene time inferred by Dickerson (1922), based on flow indicators within these stream gravels exposed along Lichau Creek, strongly suggests that the ancestral PlioPleistocene drainage pattern. In addition, the southwest-flowing Lichau Creek appears to be diverted around Mecham Hill (Figure 4), suggesting probable late Quaternary uplift. Currently Lichau Creek is separated from the linear, northeast-southwest flowing upper reach of Stemple Creek by the Washoe Anticline, as mapped by Weaver (1949). The Washoe Anticline is bounded on the southwest by the Tolay fault. We found fluvial gravels, likely correlative with the deposits mapped by Dickerson (1922) and Higgins 43

Ratio of valley floor width

Sonoma Coast fold-and-thrust beltt

to valley height

10

0 0 2 4 6 8 10 12 14 16 Distance upstream from Pacific Ocean (in km) 18

b)
Ratio of valley floor width to valley height

25

20
Sonoma Coast fold-and-thrust beltt Bloomfield fault Dunham fault

15

10

0 0 2 4 6 8 10 12 14 16 18 Distance upstream from Pacific Ocean (in km) 20

c)
Ratio of valley floor width to valley height

30
Americano Creek fault

25

20
Sonoma Coast fold-and-thrust beltt

Joy Woods fault

a)

15

15

10

0 0 2 4 6 8 10 12 14 16 18 Distance upstream from Pacific Ocean (in km) 20 22 24

Figure 16. Plots of the ratio of valley floor width to valley height versus distance upstream from the Pacific Ocean for; (a) Salmon Creek, (b) Americano Creek and, (c) Stemple Creek.

Bloomfield fault

(1952), and possibly with the sand and gravel of Cotati, within the prominent northeastsouthwest trending wind gap located between Lichau and Stemple Creeks. Additional evidence that the growth of anticlines along the southwestern margin of the valley have beheaded the coastal streams from the upper reaches of their former headwaters. For example, Weaver (1949) mapped an unnamed anticline along the drainage divide between Walker and San Antonio Creeks, 6 km southwest of Petaluma (Figures 2 and 4). This anticline appears to have disrupted the regional drainage network. Previous researchers have suggested that an ancestral stream flowing either westward from Petaluma Valley (Holway, 1914; Dickerson, 1922), or eastward towards Petaluma Valley (Weaver, 1949), could not keep pace with deformation of the region. This ancestral stream valley is now occupied by San Antonio Creek, which flows eastward into Petaluma Valley, and Walker Creek, which flows westward to the Pacific Ocean (Figure 4). West-flowing tributaries enter the east-flowing San Antonio Creek at acute angles that suggest the direction of flow in San Antonio Creek has been reversed (Higgins, 1952). Thus, the ancestral stream probably originally flowed from east to west across the anticline, and has since been disrupted by folding. In conclusion, the Plio-Pleistocene coastal drainage pattern likely was defeated by development of the Santa Rosa Valley. If this conclusion is correct, the Santa Rosa Valley is a young structural and geomorphic feature, as first argued by Higgins (1952). Higgins (1952) has shown that north of the Santa Rosa Valley, the Russian River probably prograded its way to the west as the sea level dropped, maintaining its course across the uplift that developed subsequently. Ponding at Laguna de Santa Rosa at the end of Santa Rosa Creek (Figure 4) suggests that subsidence along the valley margin, and/or uplift west of the valley, may be continuing.

45

5.0 DISCUSSION Our geomorphic analyses provide evidence that the loci of Quaternary deformation roughly coincide with the loci of past Wilson Grove Formation Pliocene deformation. Evidence for Quaternary activity includes: (1) the evolution and pattern of Quaternary drainage systems; (2) anomalous stream channel profiles and stream-gradient index values; (3) changes in stream valley morphology across Cenozoic structures; and (4) warped Holocene stream terrace profiles. Below we discuss estimates of uplift rates derived from cross-sections and contouring of the basal Wilson Grove formation and from profiles of Quaternary geomorphic surfaces. We use these preliminary estimates to evaluate the seismic hazard of these potential seismogenic sources. 5.1 Rates of Quaternary Deformation We estimate rates of Quaternary deformation based on: (1) folding and offset of the base of the 4-6 Ma Wilson Grove Formation and, (2) analysis of marine and fluvial terrace profiles. The range of permissible uplift rates derived from the fluvial terrace profiles incorporates the assumption that the rate of stream incision is controlled by the rate of tectonic uplift. Within the study area, channel incision rates likely reflect a combination of factors, and incision rates can therefore be used only to constrain the maximum uplift rate. Therefore, given the reconnaissance nature of this study, and the lack of absolute ages for the fluvial deposits, these rates should be regarded as order-of-magnitude estimates only. Additional age dating studies are required to refine our estimated deformation rates. Based on our detailed mapping of the Bloomfield area (Figure 17), and bedrock mapping by Travis (1952), we have constructed a structural cross section across the Bloomfield and Americano Creek faults (Figure 18). We use this cross section to estimate a maximum vertical offset of 213 m (700 ft) of the base of the Wilson Grove Formation across the Bloomfield fault. Assuming that the base of the Wilson Grove formation is between 6 and 4 Ma, we obtain a vertical separation rate of 0.036 to 0.053 mm/yr. If the fault dips between 30 and 45 northeast, then the average slip rate is approximately 0.07 + .03 mm/yr. Uplift and internal deformation of the Wilson Grove Formation may not have begun until within the last 2-4 Ma, following development of the Rodgers Creek fault system. If we assume that faulting initiated between 2 and 4 Ma, we obtain vertical separation rates of 0.05 to 0.11 mm/yr and average slip rate of approximately 0.13 + .04 mm/yr. Assuming that stream incision rates are equivalent to, or less than, uplift rates, we can estimate rates of Quaternary uplift rates from the elevations of fluvial terraces above the modern stream channel. Based on elevations of 39.5 m for fluvial terrace Qt2 and 18.3 m for fluvial terrace Qt3 within the hanging wall of the Bloomfield fault along Stemple Creek (Figure 13), and assuming ages of 320 and 130 ka respectively (Table 2), we obtain an incision rate of 0.12-0.14 mm/yr. This rate is similar to that derived from the offset base of the Wilson Grove Formation across the Bloomfield fault, using the assumption that faulting initiated approximately 2 Ma. Based on our tentative correlation of marine terraces along the Sonoma coast, we have estimated a Quaternary uplift rate of approximately 0.14 mm/yr within the Sonoma Coast fold-and-thrust belt adjacent to the San Andreas fault. This uplift rate is similar to that derived from the maximum elevations above sea level of 36.5 m for fluvial terrace Qt2 and 30 m for fluvial terrace Qt3 along Stemple Creek (Figures 8 and 12). Assuming ages of

46

320 and 130 ka respectively (Table 2), we obtain a minimum Quaternary uplift rate of between 0.1-0.2 mm/yr, roughly equivalent to that derived from the marine terraces. 5.2 Hazard Assessment of Seismogenic Sources Our geomorphic mapping and geomorphic analyses support the interpretation that the two belts of contractional structures identified in the study region, the Bloomfield fold-andthrust belt and the Sonoma Coast fold-and-thrust belt, have accommodated localized Quaternary uplift. In this section, we estimate maximum earthquakes for potentially seismogenic thrust faults based on fault and/or fold length, inferred fault geometry, and the rates of late Cenozoic crustal shortening estimated above. Maximum earthquake estimates are based primarily on empirical relations between rupture area and earthquake magnitude (Wells and Coppersmith, 1995). For calculating rupture area, rupture width first is calculated based on the assumption that the aspect ratio for the rupture (i.e. the length: width ratio) will not be less than 1:1. This assumption is consistent with data from moderate to large magnitude earthquakes, which show that the aspect ratio for reverse and thrust earthquakes typically is 1:1 or greater (Nicol et al., 1996). Thus, for example, if the surface length of a mapped fault or associated anticline is approximately 12 km, we assume that the maximum rupture area is 12 km x 12 km (i.e., 144 km2 ). Estimated maximum earthquakes for major Cenozoic structures within the study area are based on fault and fold lengths derived from the results of our mapping, geomorphic, and geologic analyses. For example, the maximum length of the Bloomfield fault, the longest fault segment within the Bloomfield fold-and-thrust belt, is approximately 18 km. We assume that fault width is not likely to exceed 15 km. With these assumptions, we adopt a rupture area of 18 km x 15 km for this scenario, or a total potential rupture area of 270 km2 . Using the empirical relationships in Wells and Coppersmith (1995), we estimate a maximum earthquake of Mw 6.5 for this scenario. The Bloomfield fault is the best-expressed fault within the Bloomfield fold-and-thrust belt. Based on the map pattern and contours on the base of the Wilson Grove Formation (Figure 10), the Dunham fault is significantly shorter than the Bloomfield fault and may accommodate folding within the hanging wall of the Bloomfield fault. Other possible seismogenic sources include the northern end of the Tolay fault, the Americano Creek fault, and the Joy Woods fault. None of these structures, with the possible exception of the Tolay fault, is expected to generate an earthquake greater than that estimated for the Bloomfield fault of Mw 6.5. Based on the proximity of the folds identified within the Sonoma Coast fold-and-thrust belt to the San Andreas fault, it is likely that thrust or reverse faults associated with development of these folds intersect the San Andreas fault at shallow depths. We therefore believe it is unlikely that these structures are capable of independent events. The relatively low uplift rate and short (<10 km), discontinuous nature of these folds, suggest that growth of the folds may occur aseismically as the result of local strain partioning during large nearby earthquakes on the San Andreas fault (Lettis and Hanson, 1991). Contemporary microseismicity between the Rodgers Creek and San Andreas fault consists of small (<M 3.0) seismic events that coincide with the mapped traces of the Americano Creek, Bloomfield, and Dunham faults (Wong, 1991). These events are consistent with predominately reverse motion along northwest-striking fault planes. The absence of significant historical seismicity west of the Rodgers Creek-Healdsburg fault zone may be 47

the result of the 1906 earthquake that probably released most of the stored crustal strain in the area.

48

Wilson Grove formation

C
Southwest
800 600 400

C
Northeast

Wilson Grove formation minimum vertical offset = 600 (183 m) estimated maximum vertical offset = 700 (213 m)

elevation (in feet)

200 0 -200 -400 -600 -800 0

Wilson Grove formation estimated maximum vertical offset = 100 (30.5 m)

AME RICANO CREEK F AULT

LT IELD FAU BLOOMF

Franciscan Complex

Franciscan Complex

Franciscan Complex

V.E. = 20x
2000 4000

6000

8000

10000

12000

14000

16000

18000

20000

22000

24000

Horizontal Distance (in feet)

Figure 18. Structural cross-section C-C through the Bloomfield area across the Americano Creek and Bloomfield faults. See Figure 18 for cross-section location.

6.0 CONCLUSIONS Localized tectonic uplift above late Cenozoic faults and associated folds is expressed as two topographically distinct series of northwest-trending hills. The Bloomfield fold-and-thrust belt extends between the northwestern end of Petaluma Valley and the Russian River. These hills, including English Hill located approximately 3 km (2 miles) north of Bloomfield, consist of resistant rock of the Franciscan Complex that rise 30 to 150 m (100 to 500 feet) above the rounded accordant hilltops underlain by the Wilson Grove Formation. The folds are bounded on the southwest by the Tolay, Bloomfield, Dunham, and Joy Woods faults. The Sonoma Coast thrust-and-fold belt consists of low rugged hills 3 to 5 km (2 to 3 miles) that border the Sonoma Coast adjacent to the San Andreas fault. These hills, and associated deeply incised stream mouths and steep sea cliffs, are developed entirely upon resistant Mesozoic rocks of the Franciscan Complex. Profiles of marine terraces along the coast show progressive uplift of older terraces and warping of the entire terrace sequence consistent with localized folding adjacent to the San Andreas fault. Longitudinal profiles of Pleistocene and younger stream terraces show localized warping and uplift coincident with loci of deformation within the Wilson Grove Formation. Valley height to valley-floor width ratios reveal changes in stream valley morphology along Salmon, Americano, Stemple, and Walker creeks across the Bloomfield fault and near the San Andreas fault. The observed variations in stream valley morphology, coincident with anomalous variations in stream channel gradients, are consistent with response of these fluvial systems to localized late Quaternary uplift. These uplifts represent localized strain between the San Andreas and Rodgers Creek faults, and likely are the result of large-scale interactions between these two major faults.

51

7.0 REFERENCES Bauer, F.H., 1952, "Marine terraces between Salmon Creek and Stewarts Point": University of California thesis. Bedrossian, T., 1969, Paleoecological evaluation of the "Merced" Formation, Sonoma County, California: Master's thesis, University of California, Davis. Bedrossian, T., 1974, Fossils of the "Merced Formation, Sebastopol Region: California Geology, August 1974, p. 175-182. Bell, C.J., 1998, North American Quaternary land mammal ages and the biochronology of North American microtine rodentsin Sowers, J.M., Noller, J.S., and Lettis, W.R., eds.,
Dating and Earthquakes: Review of Quaternary Geochronology and its Application to Paleoseismology, NUREG/CR 5562, U.S. Nuclear Regulatory Commission, 6 chapters + apps.

Berkland, J.O., Raymond, L.A., Kramer, J.C., Moores, E.M., and ODay, Michael, 1972, What is Franciscan?: American Association of Petroleum Geologists Bulletin, v. 56, n. 12, p. 2295-2302. Brocher, T. M., and Furlong, K. P., 1994, Resolving the tectonic framework of the San Francisco Bay Area, California: A Report of the April 18-20, 1994 Workshop: USGS Open-File-Report 94-259. Bull, W.B., and McFadden, L.D., 1977, Tectonic geomorphology north and south of the Garlock fault, California: in Geomorphology in arid regions, Doehring, D.O., ed., Proceedings of Eighth annual Geomorphology Symposium, State University of New York at Binghamtion, p. 115-138. Bullard, T. F., and Lettis, W. R., 1993, Quaternary fold deformation associated with blind thrust faulting, Los Angeles Basin, California: Journal of Geophysical Research , Vol. 98, No. B5, p. 8349-8369. Burnett, A. W., and Schumm, S. A., 1983, Active tectonics and river response in Louisiana and Mississippi: Science, v. 222, p. 49-50. Cardwell, G.T., 1958, "Geology and groundwater in Santa Rosa and Petaluma Valleys: U.S. Geological Survey Water Supply Paper 1548. Chappell, J.M., 1974, Geology of coral terrace, Huon Peninsular, New Guinea; A study of Quaternary tectonic movements and sea-level changes: Geological Society of America Bulletin, v. 85, p. 553-570.and Dickerson, R. E., 1922, Tertiary and Quaternary history of the Petaluma, Point Reyes, and Santa Rosa quadrangles: California Academy Science Proceedings, 4th serial, Vol. 11, No. 19, p. 527-601. Dwyer, M. J, and Borchardt , G., 1994, Paleoseismicity and liquefaction potential of a Sangamon marine terrace near the San Andreas Fault, Sonoma County, California: in Prentice, C S., eds. U. S. Geological Survey Open-File Report OFR 94-0568, p. 59-61. Fox, K., 1983, Tectonic Setting of Late Miocene, Pliocene, and Pleistocene Rocks in Part of the Coast Ranges North of San Francisco, California: Geological Survey Professional Paper 1239, 33 p. Furlong, K. P., Hugo W. D., and Zandt, G., 1989, Geometry and evolution of the San Andreas fault zone in northern California, Journal of Geophysical Research, v. 94, p. 31000-3110. Gealey, W. K., 1951, Geology of the Healdsburg quadrangle, California: California Division of Mines and Geology Bulletin 161, 50 p. Grove, K., Colson, K., Binkin, M., Dull, R., and C. Garrison, 1995, Stratigraphy and structure of the late Pleistocene Olema Creek Formation, San Andreas fault zone north of San Francisco, California: in Sagines, E.M., Andersen, D.W., and Buising, A.B., eds, Recent Geologic Studies in the San Francisco Bay area: Pacific section SEPM, v. 76, p. 55-76.

52

Hack, J.T., 1973, Stream-profile analysis and stream-gradient index: U.S. Geological Survey Journal of Research, v. 1, p. 421-429. Helley, E. J., and Herd, D. G., 1977, Map showing faults with Quaternary displacement, northeastern San Francisco Bay Area region: U.S. Geological Survey Miscellaneous Field Studies Map MF-881, scale 1:125,000. Higgins, C. G., 1952, Lower course of the Russian River, California: University of California Publications in Geological Sciences, vol. 29, no. 5, p. 199-263. Higgins, C.G., 1960, Ohlson Ranch Formation, Pliocene, northwestern Sonoma county, California: University of California Publications in Geological Sciences, v. 36, n. 3, p. 199-231. Hill, D.P., Eaton, J.P., and Jones, L.M., 1990, Seismicity, 1980-86, in Wallace, R.E., ed., The San Andreas fault system, California: United States Geological Survey Professional Paper 1515, p. 115-151. Holway, R. S., 1913, The Russian River, a characteristic stream of the California Coast Ranges: University of California Publications in Geography, Vol. 1, No. 1, p. 1-60. Holway, R.S., 1914, Physiographically unfinished entrances to San Francisco Bay: University of California Publications in Geography, v. 1, n. 3, p. 81-126. Johnson, D.L., 1977, Late Pleistocene environments of coastal California: Evidence for an ice age refugium: Quaternary Research, v. 8, p. 154-179. Johnson, F. A., 1934, Geology of the Merced, Pliocene, Formation north of San Francisco Bay, California: Ph.D. Dissertation, University of California, Berkeley. Keller, E.A., 1977, Adjustment of drainage to bedrock in regions of contrasting tectonic framework: Geological Society of America Abstracts with Programs, v. 9, no. 7, p. 1046. Keller, E.A., and Pinter, N., 1996, Active tectonics, earthquakes, uplift, and landscape: Prentice-Hall, Inc. Kelson, K.I., Lettis, W.R., and Lisowski, M., 1992, Distribution of geologic slip and creep along faults in the San Francisco Bay region, in Borchardt, G., Hirschfeld, S.E., Lienkaemper, J.J., McClellen, P., Williams, P.L., and Wong, I.G., eds., Proceedings of the Second Conference on Earthquake Hazards in the Eastern San Francisco Bay Area: California Division of Mines and Geology Special Publication 113, p. 31-38. Kennedy, G.L., Lajoie, K.R., and Wehmiller, J.F., 1982, Aminostratigraphy and faunal correlations of late Quaternary marine terraces, Pacific Coast, USA: Nature, v. 299, p. 545-547. Lajoie, K.R., 1986, Coastal tectonics, in Usselman, T.M., ed., Studies in geophysics: Active tectonics: Washington, D.C., National Academy Press, p. 95-124. Lawson, A. C., 1894, The geomorphology of the coast of Northern California: University of California Publications in Geography, Vol. 1, No. 8, p. 241-271. Leopold, L.B., and Maddock, T., 1953, Hydraulic geometry of stream channels and some physiographic implications: U.S. Geological Survey Professional Paper 252. Lettis, W.R., and Hanson, K., 1991, Crustal strain partioning: implications for seismic hazard assessment in westen California: Geology, v. 19, p. 559-562. Lisowski, M., Savage, J.C., and Prescott, W.H., 1991, The velocity field along the San Andreas fault in central and southern California: Journal of Geophysical Research, v. 96, p. 8369-8389. Lisowski, M., and Savage, J.C., 1992, The velocity field in the San Francisco Bay Area and the inferred depth of creep on the Hayward fault, in Borchardt, G., Hirschfeld, S.E., Lienkaemper, J.J., McClellen, P., Williams, P.L., and Wong, I.G., eds., Proceedings of the Second Conference on Earthquake Hazards in the Eastern San Francisco Bay Area: California Division of Mines and Geology Special Publication 113, p. 39-44. Marple, R. T., and Talwani, 1993, Evidence of possible tectonic upwarping along the South Carolina coastal plain from an examination of river morphology and elevation data: Geology, v. 21, p. 651-654.

53

Merritts, D., and Vincent, K.R., 1989, Geomorphic response of coastal streams to low, intermediate, and high rates of uplift, Mendocino triple junction region, northern California: Geological Society of America Bulletin, v. 101, p. 1373-1388. Morse, R.R., and Bailey, T.L., 1935, Geological observations in the Petaluma district, California: Geological Society of America Bulletin, v. 46, p. 1437-1456. Muhs, D.R., and Szabo, B.J., 1982, Uranium-series age of the Eel Point terrace, San Clement Island, California: Geology, v. 10, p. 23-26. Nicol, A., Watterson, J., Walsh, J.J., and Childs, C., 1996, The shapes, major axis orientations and displacement patterns of fault surfaces: Journal of Structural Geology, v. 18, p. 235-248. Niemi, T.M., and Hall, N.T., 1992, Late Holocene slip rate and recurrence of great earthquakes on the San Andreas fault in northern California: Geology, v. 20, p. 195-198. Osmont, V.C., 1905, A geologic section of the Coast Ranges north of the Bay of San Francisco: California University, Department of Geological Science Bulletin, v. 46, p. 39-87. Ouchi, S., 1985, Response of alluvial rivers to slow active tectonic movement: Geological Society of America Bulletin, v. 96, p. 504-515. Pampeyan, E. H., 1979, Preliminary map showing recency of faulting in coastal north-central California: U. S. Geological Survey Map MF-1070, 1:250,000 scale. Pillans, B., 1983, Upper Quaternary marine terrace chronology and deformation, South Taranaki, New Zealand: Geology, v. 11, p. 292-297. Repenning, C.A., Fejfar, O., and Heinrich, W.D., 1990, Arvicolid rodent biochronolgy of the northern hemisphere, in Fejfar, O., and Henrich, W.D., eds., International symposium: evolution , phylogeny and biostratigraphy of arvicolids (Rodentia, Mammalia); Prague, Czeh, Geological Survey, p. 385-417. Sarna-Wojcicki, A.M., 1976, Correlation of late Cenozoic tuffs in the central Coast Ranges of California by means of trace element geochemistry: U.S. Geological Survey Professional Paper 972, 30 p. Schumm, S. A., 1977, The Fluvial System: John Wiley and Sons, New York, 338 p. Schwartz, D.P., Pantosti, D., Hecker, S., Okumura, K., Budding, K.E., and Powers, T., 1992, Late Holocene behavior and seismogenic potential of the Rodgers Creek fault zone, Sonoma County, California, in Borchardt, G., and others, eds., Proceedings of the Second Conference on Earthquake Hazards in the Eastern San Francisco Bay Area: California Division of Mines and Geology Special Publication 113, p. 393-398. Shackleton, N.J., and Opdyke, N.D., 1973, Oxygen isotope and paleomagnetic stratigraphy of equatorial Pacific core V28-238; Oxygen isotope temperatures and ice volumes on a 105 and 106 year scale: Quaternary Research, v.3, p. 39-55. Stirton, R.A., 1939, Cenozoic mammal remains from the San Francisco Bay Region: California University, Department of Geological Science Bulletin, v. 24, p. 339-410. Travis, R. B., 1952, Geology of the Sebastopol quadrangle, California: California Division of Mines and Geology Bulletin 162, 33 p. Wagner, D. L., and Bortoguno, E. J., 1982, Geologic map of the Santa Rosa quadrangle: California Division of Mines and Geology Regional Geologic Map Series, Map 2A, scale 1:250,000. Weaver, C. E., 1949, Geology of the Coast Ranges immediately north of the San Francisco Bay region: California: Geological Society of America Memoir 35: 242 p. Wells, D.L., and Coppersmith, K.J., 1994, New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement: Bulletin of the Seismological Society of America, V. 84, p. 974-1002. Williams, S.D.P., Svarc, J.L., Lisowski, M., and Prescott, W.H., 1994, GPS measured rates of deformation in the northern San Francisco Bay Region, California, 1990-1993: Geophysical Research Letters, v. 21, p. 1511-1514. Wong, I.G., and Bott, J.D.J, 1995, A new look back at the 1969 Santa Rosa, California, earthquake: Bulletin of the Seismological Society of American, v. 85, n. 1, p. 334-341. 54

Wong, I. G., 1991, Contemporary Seismicity, Active Faulting and Seismic Hazards of the Coast Ranges Between San Francisco Bay and Healdsburg, California: Journal of Geophysical Research, Vol. 96, No. B12, p. 19,891-19,904.

55

You might also like