Detrital-zircon fission-track ages for the ‘‘Hoh Formation’’:
Implications for late Cenozoic evolution of the
Cascadia subduction wedge
Richard J. Stewart†
Department of Earth and Space Sciences, University of Washington, Seattle, Washington 98195-1310, USA
Mark T. Brandon‡
Department of Geology and Geophysics, Kline Geology Laboratory, Yale University, P.O. Box 208109, New Haven,
Connecticut 06520-8109, USA
ABSTRACT
We report new fission-track (FT) ages for
detrital zircons for 34 sandstone samples
and 2 volcanic ash beds from the ‘‘Hoh
Formation,’’ exposed along the western side
of the Olympic Mountains of western
Washington State. The ‘‘Hoh Formation’’
is now formally known as the coastal unit
of the Olympic Structural Complex, or
Coastal OSC for short. About 35 zircons
were dated per sample. The fission-track
grain-age (FTGA) distributions are all
strongly discordant; grain ages range from
10 to older than 100 Ma. Low vitrinitereflectance values, short etch times for the
zircons, and a broad range of grain ages
indicate that the zircon FT ages are unreset
and thus preserve information about cooling events in the source region for these
sedimentary rocks. Five areas were sampled
repeatedly and yield similar FTGA distributions, demonstrating that sampling errors are not a problem. We show that almost
all of the samples contain a well-defined
young component that was probably derived from a contemporaneous active volcanic source, presumably the adjacent Cascadia arc. Binomial peak-fitting was used to
estimate the FT minimum age, which is the
age of the youngest concordant fraction of
zircon FT grain ages in a FTGA distribution. In most cases, minimum ages are similar to fossil ages where available. This result supports our contention that zircon FT
minimum ages from volcaniclastic sandE-mail: [email protected].
E-mail: [email protected].
stones commonly can be used as a proxy for
depositional age.
Our zircon FT minimum ages indicate
that the Coastal OSC is made up mainly of
lower Miocene (ca. 24 to 16 Ma) sedimentary rocks. We use these age data, together
with other geologic constraints, to reconstruct a tectonic history. Sedimentary rocks
of the Coastal OSC were derived from a
mixed-source region that included an active
volcanic arc and also older units, including
Cretaceous metamorphic rocks, probably
located in the Omineca crystalline belt in
the Canadian Rockies. The upper part of
the Clallam Formation, located on the
northern side of the Olympic Peninsula, appears to be a remnant of the sedimentary
system that fed the Coastal OSC. The sediments that formed the Coastal OSC were
initially deposited seaward of the Cascadia
trench, at water depths of .2000 m. This
debris was deposited seaward of the Cascadia trench, at water depths of .2000 m,
and subsequently accreted beneath the
frontal 50 to 100 km of the wedge. Owing
to continued accretion at the front of the
wedge, and erosion of the forearc high in
back of the wedge, these lower Miocene
sediments were moved rearward within the
Cascadia subduction wedge. A simple relationship based on the cross-sectional area
of the wedge and a steady accretion flux indicates that it would have taken ;22 m.y.
for the Coastal OSC to reach its present
position 140 km landward of the toe of the
wedge. This estimate is in good agreement
with the unit’s early Miocene age.
†
‡
Keywords: Cascadia subduction wedge,
Olympic Mountains, Coastal OSC, Hoh
Formation, zircon fission-track dating.
INTRODUCTION
The Olympic Mountains of northwest
Washington State (Fig. 1) mark the first part
of the Cascadia forearc to emerge above sea
level, starting at ca. 15 Ma (Brandon and
Vance, 1992; Brandon et al., 1998). Uplift and
erosion of the forearc high in this area provide
a deep window into the subduction wedge. Tabor and Cady (1978a, 1978b) used the informal name Olympic core for these rocks. Brandon and Vance (1992) suggested Olympic
subduction complex. To be consistent with
Stratigraphic Code (Salvador, 1994), we designate here the formal name Olympic Structural Complex (OSC). This name is a direct
replacement for the previous informal names.
Following Tabor and Cady (1978a, 1978b),
the OSC refers to the imbricated assemblage
of turbidite sandstone, siltstone, and lesser igneous rocks that structurally underlie the Crescent terrane and the Calawah and Hurricane
Ridge faults, and their lateral equivalents
(Figs. 1 and 2). These rocks are mainly Eocene to middle Miocene in age, but note there
is a small enigmatic slice of Mesozoic rocks
exposed in the northwest corner of the Olympic Peninsula. The OSC could be extended to
include all equivalent rocks within the Cascadia subduction wedge, but this is generally
not done given that the Olympic Mountains
represent the only exposure of the Cascadia
wedge north of the California border.
Tabor and Cady (1978a, 1978b) mapped
five informal lithic assemblages within the
OSC. Brandon and Vance (1992) reorganized
GSA Bulletin; January/February 2004; v. 116; no. 1/2; p. 60–75; DOI 10.1130/B22101.1; 13 figures; 3 tables; Data Repository item 2004023.
60
For permission to copy, contact [email protected]
q 2004 Geological Society of America
‘‘HOH FORMATION’’ AND LATE CENOZOIC EVOLUTION OF CASCADIA SUBDUCTION WEDGE
Figure 1. Index map, showing accreted rocks of the Cascadia subduction wedge, which
includes on-land exposures in the Olympic Mountains (Olympic Structural Complex) and
in northern California (False Cape and King Range terranes of the Coastal belt of the
Franciscan Complex). The accreted rocks are structurally overlain by the Coast Range
terrane. Solid, northeast-pointing arrow indicates convergence velocity of the Juan de
Fuca plate relative to North America at the latitude of the Olympic Mountains (DeMets
et al., 1990; DeMets and Dixon, 1999).
these units into three larger units. The Upper
unit is equivalent to the Needles-Graywolf and
Elwha assemblages of Tabor and Cady
(1978a). It contains mainly Eocene turbidites
and minor pillow basalt, and looks as if it
were derived by structural repetition from the
overlying Crescent terrane, which represents a
large thrust-bounded structural lid above the
OSC. The Lower unit is equivalent to the
Western Olympic and Grand Valley assemblages of Tabor and Cady (1978a). It appears
to be composed mainly of Oligocene and Miocene turbidites, derived by accretion from the
subducting Juan de Fuca plate and its sedimentary cover. Basalts are very rare in this
unit.
We focus here on the Coastal unit, formerly
called the Hoh Formation (Weaver, 1916,
1937), the Hoh rock assemblage (Rau, 1975,
1979), and the Hoh lithic assemblage (Tabor
and Cady, 1978a). It consists mainly of Miocene turbidites and mélange, with subordinate
exotic blocks of pillow basalt. The Coastal
unit was also probably derived by subduction
accretion, and may be a more westerly continuation of the Lower unit. This possibility was
first suggested by Brandon and Vance (1992)
and is supported by the new ages presented in
this paper. Following past practice (Brandon
and Vance, 1992; Brandon et al., 1998), we
use the following abbreviated names in our
discussion below: Upper OSC, Lower OSC,
and Coastal OSC.
The Coastal OSC has attracted considerable
attention over the years, in part because of its
spectacular exposure in wave-cut outcrops
along the Pacific Coast of the Olympic Peninsula (Fig. 1). Chaotic mudstone-rich mélanges, locally present in these exposures,
have been the subject of debate since first described by Arnold (1905) and Lupton (1914).
The mechanisms and timing of mélange formation in the Hoh remain controversial and
poorly resolved (Rau and Grocock, 1974; Rau,
1975, 1979; Snavely et al., 1993; Orange,
1990; Orange et al., 1993). This issue is significant because the Hoh represents some of
the youngest on-land exposures of accreted
rocks in the Cascadia subduction wedge.
In this paper, we present new fission-track
(FT) ages for detrital zircons from sandstones
of the Coastal OSC. This approach was first
applied in the Olympics by Brandon and
Vance (1992) in their study of the age and
origin of the Upper OSC and Lower OSC.
Age control is commonly sparse in subduction-complex rocks. The FT method provides
a useful approach in that one can date individual zircon grains. If a sandstone has remained relatively cool after deposition
(,;200 8C), then the grain ages represent the
age of cooling for rocks in the source region
from which the zircons were derived. Brandon
and Vance (1992) presented preliminary evidence indicating that the youngest zircon FT
grain ages were derived from contemporaneous volcanic sources in the active Cascade arc.
They proposed that the pooled age of that
young group of grain ages, called the FT minimum age, was a useful proxy for the depositional age of the sandstone.
We have three objectives here. The Coastal
OSC has many localities well dated by benthic
foraminifera. Thus, an important first objective is to compare zircon FT minimum ages
with fossil ages to see if the minimum ages
are a good proxy for depositional age. The
second objective is to use the overall distribution of zircon FT grain ages to identify the
source of sediments for the Coastal OSC. We
use this information to assess previous suggestions that rocks now within the coastal unit
are offset, by oblique convergence, from an
original accretionary setting in the southern
part of the Cascadia margin (e.g., Davis and
Hyndman, 1989; Aalto et al., 1995, 1998).
The third objective is to use depositional ages
to interpret the displacement history of the
Coastal OSC within the Cascadia wedge, from
an initial site of accretion at the front of the
wedge at a water depth .2000 m to its present
exposures along the Washington coast.
TECTONIC OVERVIEW
The Cascadia subduction wedge (Fig. 1) is
a doubly vergent wedge (Willett et al., 1993)
formed by 35 m.y. of subduction of the Juan
de Fuca plate beneath the Cascadia margin
(Brandon et al., 1998; Willett, 1999; Beaumont et al., 1999; Pazzaglia and Brandon,
2001). The active wedge is ;200 to 250 km
wide, bounded to the west by the Cascadia
trench and to the east by a forearc low, marked
by the Willamette Valley, Puget Sound, and
Georgia Straits. The forearc high, which corresponds to the Oregon-Washington Coast
Geological Society of America Bulletin, January/February 2004
61
STEWART and BRANDON
Figure 2. Tectonic map of the Olympic Structural Complex, showing locations of geologic
features mentioned in text. Figure adapted from Tabor and Cady (1978a, 1978b), Brandon
and Vance (1992), and Snavely and Wells (1996). OSC—Olympic Structural Complex; FT
PRZ—zircon fission-track partial retention zone. Cross section A–A9 is shown in Figure
12.
Ranges and the Insular Ranges of Vancouver
Island, marks the transition between the proside and retroside of this doubly vergent
system.
Accretion over the past 35 m.y. has allowed
the wedge to grow to its present large size
(e.g., Rau, 1973, 1975; Tabor and Cady,
1978a, 1978b; Dickinson and Seely, 1979;
Brandon and Vance, 1992; Brandon et al.,
1998). The seaward deformation front of the
wedge is currently ;140 km west of the present coastline of the Olympic Peninsula, and
seismic data indicate that the 2–3 km thickness of sediment carried into the subduction
zone on the Juan de Fuca plate gets thickened
to 20 km beneath the Washington coast and
;35 km beneath the central part of the Olympic Mountains (Clowes et al., 1987; Brandon
et al., 1998; Parsons et al., 1998; Pazzaglia
and Brandon, 2001). The modern convergence
rate along this part of the Cascadia subduction
zone is 36 mm/yr (DeMets et al., 1990;
DeMets and Dixon, 1999), and virtually all
sediment on the Juan de Fuca plate appears to
be incorporated by accretion into the Cascadia
subduction wedge (Davis and Hyndman,
62
1989; Davis et al., 1990; Pazzaglia and Brandon, 2001; Batt et al., 2001).
Along most of its length, the Cascadia subduction wedge includes a relatively coherent
tectonic sequence, called by some the ‘‘Coast
Range terrane’’ (or ‘‘Siletz terrane’’ by others), made up of early Eocene oceanic crust
and overlying marine sedimentary rocks. The
base of the Coast Range terrane is exposed in
the Olympics and is defined by a major thrust
fault, locally called the Hurricane Ridge fault
or Calawah fault (Fig. 2). The Coast Range
terrane can be viewed as a structural lid that
overlies the accretionary part of the wedge
(Clowes et al., 1987; Brandon et al., 1998).
We consider the lid to be part of the active
wedge, because it shows clear evidence of uplift and broad folding associated with the
maintenance of critical taper on the two sides
of the doubly vergent wedge (see discussion
in Pazzaglia and Brandon, 2001).
The focus here is on the Coastal OSC,
which is the youngest and westernmost unit
exposed in the Olympic Mountains (Fig. 2). It
consists mainly of a monotonous sequence of
sandstones and shales, typically well bedded
but locally chaotic, with rare interspersed lenses and blocks of volcanic rock. Clastic rocks
are mostly turbidites, and volcanic rocks are
geochemically similar to the thick basaltic
basement in the overlying Coast Range terrane
(Applegate and Brandon, 1989). Microfossils
in the Coastal OSC are overwhelmingly early
Miocene in age and record deposition in
lower-bathyal water depths, between ;2000
and 4000 m (Rau, 1975, 1979; Ingle, 1980).
There is a small number of younger localities
that indicates middle or late Miocene ages
(seven reported in Rau, 1975) and a few others
that indicate Eocene ages (four reported in
Rau, 1979). Definitive Oligocene localities
have not been recognized.
The more eastern parts of the Olympic subduction complex commonly show prehnitepumpellyite facies metamorphism and a
spaced pressure-solution cleavage (Tabor and
Cady, 1978a, 1978b; Brandon and Calderwood, 1990). In contrast, clastic rocks of the
Coastal OSC are typically unmetamorphosed,
although zeolites are locally present (Stewart,
1974). Basaltic blocks do show static
greenschist assemblages, probably related to
seafloor metamorphism, which would have
occurred prior to incorporation of the blocks
into the mélange units. Cleavage fabrics are
also notably absent, expect for scaly fabrics
locally developed in mélange units.
The Coastal OSC shows clear evidence of
northeast-southwest shortening, in the direction of convergence. Bedding generally strikes
to the northwest-southeast and is commonly
steep and locally overturned. Where recognized, folds are typically overturned to the
southwest. The section shows widespread evidence of repetition by thrust faults (Rau,
1975, 1979; Orange, 1990).
Mélange and broken formation (sensu Hsü,
1968) are expected, given the subduction-zone
setting, but these deformational styles are observed in less than ;20% of the Coastal OSC.
In fact, mélange is even less common (,10%)
in the other parts of the subduction complex
(Tabor et al., 1970; Tabor and Cady, 1978a).
These observations contrast with the large
fraction of mélange found in Miocene parts of
the Cascadia wedge exposed in the Franciscan
Complex in northern California (see Fig. 1;
King Range and False Cape terranes of
McLaughlin et al. [1982, 1994]; Aalto et al.
[1995]).
Local bodies of mélange in the Coastal
OSC have a matrix of intensely ‘‘scaly’’ siltstone with blocks of sandstone, conglomerate,
and volcanic rock (Weissenborn and Snavely,
1968; Rau, 1970, 1975, 1979; Rau and Grocock, 1974; Orange, 1990; Orange et al.,
Geological Society of America Bulletin, January/February 2004
‘‘HOH FORMATION’’ AND LATE CENOZOIC EVOLUTION OF CASCADIA SUBDUCTION WEDGE
1993). Published maps generally lump rare
exposures of broken formation together with
the more widespread coherent stratigraphic
units. Rau (1979) and Snavely and Kvenvolden (1989) recognized local areas where the
more coherent parts of the Coastal OSC appear to be resting depositionally on broken
formation. Modern subduction wedges are
commonly mantled by small ‘‘trench-slope
basins’’ (Moore and Karig, 1976) that accumulate in local depressions in the surface of
the wedge. The coherent sequences in the
Coastal OSC may represent trench-slope basins, deposited after the underlying section
was accreted.
The other alternative is that the chaotic
units record mass wasting. Mass wasting is
now widely recognized at sediment-rich subduction zones, like Cascadia (e.g., Goldfinger
et al., 2000). Mass-wasting deposits commonly reach the trench (e.g., Davis and Hyndman,
1989), where they become depositionally interfingered with trench-basin strata. Thus,
stratigraphic relationships are currently not
sufficient to judge whether the coherent units
in the Coastal OSC were deposited seaward or
landward of the toe of the wedge.
The paleobathymetric evidence provided by
benthic foraminifera provide a key constraint
in that they record deposition in water deeper
than ;2000 m. Rau (1975, 1979) recognized
that deep-water benthic foraminifera were
very common in the Coastal OSC, but he preferred a conservative estimate of .200 m for
water depth. One of the reasons is that the
foraminiferal assemblages were a mixture of
both deep- and shallow-water taxa. Downslope transport is now recognized as the reason for mixed depth provenance of benthic foraminifera in turbidite sequences. Thus,
paleodepth is now defined by using the deepest-water taxa in the assemblage (Ingle, 1980).
We have applied the paleobathymetry biofacies of Ingle (1980) to define paleodepth by
using the benthic foraminifera reports by Rau
(1975, 1979) for the Coastal OSC. The conclusion is that the unit was deposited in the
lower-bathyal biofacies, indicating water
depths of 2000 to 4000 m. The diagnostic taxa
are Cibicides pseudoungerianus evolutus
(Cushman and Hobson) and Gyroidina soldanii (d’Orbigny), which are very common in
all of Rau’s samples. Less common diagnostic
taxa are Melonis pompilioides (Fichtel and
Moll) (referred to by Rau, 1979, by an older
name: Nonion pompilioides), and Plectofrondicularia californica (Cushman and Stewart).
At present, the surface of the Cascadia
trench basin seaward of the wedge is at a water depth between 2000 to 2500 m. The rela-
tively shallow depth is due to the young age
of the Juan de Fuca plate, which is ca. 8 Ma
at the trench. The age of the plate at the trench
has remained relatively constant at ;8 m.y.
over the past ;35 m.y. (Wilson, 1988), so the
paleodepth of the trench basin would not have
been much different than the modern. Thus,
we conclude that most of the Coastal OSC
was probably deposited in the trench basin,
seaward of the Cascadia wedge. We recognize,
however, that some of the subduction complex
may represent slope basins, deposited after accretion. Any slope-basin sequences must have
been deposited soon after accretion in order to
account for the foraminiferal evidence for initial deposition at depths of .2000 m. Thus,
one can consider the age of the Coastal OSC
as defining when sedimentary materials first
were added to the front of the wedge. This
evidence is used in the Discussion to constrain
transport times through the wedge.
We have already noted the similarities between the Coastal OSC and correlative accretionary complexes of the False Cape and King
Range terranes, exposed on the Pacific Coast
in northern California (Fig. 1). These units
represent the only other part of the Cascadia
subduction wedge where upper Cenozoic accreted rocks are subaerially exposed. The
False Cape terrane is composed mainly of
scaly argillite and sandstone, and was deformed in early Miocene time (Aalto et al.,
1995). The King Range terrane, which contains the youngest rocks of the Coastal belt of
the Franciscan Complex, was accreted at the
front of the wedge in the middle Miocene, at
ca. 15 Ma (McLaughlin et al., 1982, 1994).
The King Range terrane is similar to the
Coastal OSC in that both include broken formations and mélanges, a general absence of
metamorphism, and middle Miocene turbidites that may be accreted sediments initially
deposited in the trench (McLaughlin et al.,
1982). An important difference is that the
King Range terrane, and perhaps the False
Cape terrane as well, appear to have been
strongly affected by northward migration of
the Mendocino triple junction, which may be
largely responsible for driving the present-day
uplift and rapid erosion of the wedge in the
northern California area (Dumitru, 1991).
SAMPLING AND DATING METHODS
Zircon FT analyses for 1264 zircons from
36 samples, all medium- to coarse-grained
sandstones, are shown on Tables 1, 2, and 3.
Maps showing the distribution of sample localities and zircon FT data are available from
the Data Repository.1 Probability density plots
for selected examples are shown on Figure 5.
Peak-fitting results (Tables 1, 2, and 3) are
shown compared to the IUGS time scale of
Remane (2000) on Figures 6 and 7.
Zircon separates were prepared by standard
techniques using the external-detector method
(Wagner and van den Haute, 1992). Both ends
of the irradiation package contained fluence
monitors, using the SRM 612 U-enriched
glass standard, and a mount of Fish Canyon
Tuff zircons. Samples were irradiated with
thermal neutrons in the TRIGA reactor at
Oregon State University (a well-thermalized
reactor) using a nominal fluence of 1.0 3 1015
neutrons/cm2. Following irradiation, the mica
detectors were etched in 40 percent hydrofluoric acid for 18 minutes. The effective
monitor track density was determined for each
sample by using the sample position in the
irradiation package to interpolate the densities
measured in the two fluence monitors. Fission
tracks were counted using oil immersion and
a Zeiss microscope at 12503. Based on 8
analyses, the Zeta value (Hurford and Green,
1983) for R. Stewart is 331.03 6 6.66 (6 1
standard error). Four samples were dated by J.
Garver (1990, personal commun.) using methods similar to ours (Garver et al., 1999). His
Zeta is 320.67 6 7.79.
The analyses are organized into three
groups. Table 1 consists of 25 samples from
the coherent well-stratified turbidite sequences
in the Coastal OSC. Five areas, shown as
Groups A–E on Table 1 and Figures 3 and 4,
were densely sampled, with samples ,1 km
apart, to check reproducibility of our results.
Table 2 includes ten samples collected from
sandstone blocks in Coastal OSC mélange. Table 3 is included for comparative purposes,
and consists of samples from adjacent parts of
both the Upper and Lower OSC. Five of these
samples are from Brandon and Vance (1992),
recalculated using BINOMFIT.
EVIDENCE AGAINST THERMAL
RESETTING
Our interpretation of these zircon FT ages
depends on whether the ages have been thermally reset. We are particularly concerned because the sediments that formed the Coastal
OSC were probably first deposited outboard
of the subduction zone, on a young, hot Juan
de Fuca plate.
1
GSA Data Repository item 2004023, list describing the location and geologic setting for new
samples dated, is available on the Web at http://
www.geosociety.org/pubs/ft2004.htm. Requests
may also be sent to [email protected].
Geological Society of America Bulletin, January/February 2004
63
STEWART and BRANDON
TABLE 1. DETRITAL-ZIRCON FT AGES FROM COHERENTLY BEDDED SANDSTONES, COASTAL OSC
Lab number
Nt
Minimum-age peak
Age
Single samples
168†
36
50
108†‡
†
43
41
†
42
183
†
13
162
154
51
†
151
50
27
143†
†
170
37
195
18
Group A
42a†‡
41
27
42b†
COMBINED
68
(continued for fourth peak)
Group B
155†
49
28
148†
COMBINED
77
Group C
171
28
157
10
COMBINED
38
Group D
†
150
23
48
142†
†
50
156
†
51
149
COMBINED
172
Group E
182
35
163
46
185
50
186
18
184
14
COMBINED
163
Old peaks
95% Conf. Int.
%
Age
95% Conf. Int.
%
Age
95% Conf. Int.
%
11.0
13.9
16.0
16.2
18.3
18.6
19.9
21.2
24.4
25.9
22.9
22.5
21.4
21.4
22.8
22.0
23.0
21.6
22.4
25.0
13.9
13.0
11.5
11.5
13.3
12.2
13.6
11.8
12.7
16.2
10.6
11.9
84.2
70.1
83.4
40.1
24.1
85.0
69.2
55.7
21.6
29.3
54.4
50.8
33.1
33.8
44.7
55.6
56.1
52.1
22.4
23.8
210.9
27.4
212.8
26.0
28.2
210.5
28.1
213.3
12.7
14.4
113.6
18.7
120.9
17.3
110.1
113.0
19.5
117.9
45.3
24.8
15.8
29.9
16.6
19.5
45.0
15.0
30.8
44.3
53.6
50.9
–
–
–
66.5
–
–
–
–
25.9
26.0
–
–
–
25.9
–
–
–
–
16.6
16.8
–
–
–
16.4
–
–
–
–
44.1
28.3
–
–
–
40.4
–
–
–
–
14.2
25.4
14.3
22.1
25.1
22.0
12.5
16.3
12.3
19.1
45.5
12.2
45.6
46.9
29.9
24.3
29.9
24.0
14.4
112.5
14.6
75.8
54.5
32.3
174.4
–
51.1
176.2
254.1
–
25.4
254.5
177.9
–
16.0
178.5
5.2
–
52.5
3.1
21.7
21.9
21.8
21.8
22.4
21.6
11.9
12.7
11.7
46.3
47.3
46.6
65.1
62.6
64.4
24.6
26.7
24.3
14.9
17.4
14.6
53.7
52.7
53.4
–
–
–
–
–
–
–
–
–
–
–
–
21.6
23.9
22.4
22.8
23.7
22.3
13.2
14.4
12.5
70.5
80.1
74.6
52.2
58.6
54.1
210.6
218.7
29.4
113.3
127.5
111.4
29.5
19.9
25.4
–
–
–
–
–
–
–
–
–
–
–
–
18.5
19.1
19.4
20.8
15.1
21.8
21.5
21.5
21.6
23.5
12.0
11.6
11.6
11.8
14.6
70.0
76.7
67.5
64.5
12.4
41.6
49.3
43.8
54.3
20.7
25.8
26.7
26.4
25.7
22.2
16.8
17.7
17.5
16.4
12.5
30.0
23.3
23.7
35.5
57.9
–
–
82.5
–
48.3
–
–
218.4
–
23.8
–
–
123.6
–
14.1
–
–
8.8
–
27.6
18.4
20.5
20.7
23.6
32.3
20.9
23.7
22.3
22.6
23.5
24.3
21.7
14.6
12.5
13.0
14.1
14.9
11.8
27.1
48.1
49.3
65.8
78.9
43.4
41.1
48.5
52.2
57.5
90.9
44.5
25.8
25.1
25.3
213.3
225.8
24.6
16.8
15.7
15.9
117.4
135.9
15.1
63.1
51.9
50.7
34.2
21.1
45.8
86.2
–
–
–
–
72.6
228.2
–
–
–
–
217.0
141.7
–
–
–
–
122.1
3.4
–
–
–
–
10.8
Notes: Peak ages and 95% confidence interval were estimated by using the binomial-fit method (Galbraith and Green, 1990). Nt—total number of dated grains; % 5
percent of total number of dated grains in an individual peak. COMBINED—data for all samples in group. Bold laboratory numbers correspond to samples illustrated in
Figure 5.
†
Samples used for comparison of minimum FT ages with fossil ages.
‡
Dated by J.I. Garver, otherwise dated by R.J. Stewart.
TABLE 2. DETRITAL-ZIRCON FT AGES FROM SANDSTONE BLOCKS IN MÉLANGE, COASTAL OLC
Lab
number
Nt
Single samples
117†
45
37
46
176
15
†
50
118
140
50
175
41
178
17
173
18
191
37
152
23
Minimum-age peak
Age
14.6
17.7
19.9
20.4
22.0
25.9
32.3
35.4
36.2
37.5
95% Conf. Int.
23.2
23.9
25.1
23.9
22.4
24.0
27.0
26.2
211.6
212.3
14.1
15.0
16.9
14.9
12.7
14.8
19.0
17.5
117.1
118.3
Old peaks
%
Age
6.4
21.6
66.2
11.3
41.2
46.1
40.3
76.6
25.8
44.0
33.9
37.7
45.4
38.9
57.5
52.9
68.7
73.3
54.8
52.8
95% Conf. Int.
28.6
24.0
215.4
27.6
24.7
27.8
212.2
224.3
29.9
214.0
111.5
14.5
123.3
19.4
15.1
19.1
114.8
136.3
112.1
119.0
%
Age
95% Conf. Int.
%
23.4
78.4
33.8
30.8
58.8
53.9
59.7
23.4
74.2
56.0
50.0
–
26.2
–
17.1
–
70.2
–
55.3
–
–
–
–
–
–
26.6
–
–
–
–
–
–
17.5
–
–
–
–
–
–
57.9
–
–
–
–
–
–
Notes: See Table 1.
Two observations indicate that resetting did
not occur. The first comes from vitrinite reflectance data, which range from 0.48% to
2.08% (Snavely and Kvenvolden, 1989;
Orange and Underwood, 1995); most values
are between 0.5% and 1.29%. Values .1.0%
are commonly associated with fault zones
where rocks of differing thermal maturation
64
may have been juxtaposed. Arne and Zentilli
(1994) showed that reflectance values of 0.7%
to 0.9% are diagnostic of thermal histories
needed to totally reset apatite FT ages. This
prediction is consistent with common partial
resetting of apatite FT ages in the Coastal
OSC (Brandon et al., 1998). In comparison,
resetting of zircon FT ages appears to be as-
sociated with vitrinite reflectance values of
.;6% (Green et al., 1996; http://
www.geotrack.com.au/zfta.htm; Kamp, 2001).
Thus, we would not expect any resetting of
the zircon FT ages in the Coastal OSC.
The second observation is based on experience with partial resetting of zircon FT ages
in the deeply exhumed core of the Olympic
Geological Society of America Bulletin, January/February 2004
‘‘HOH FORMATION’’ AND LATE CENOZOIC EVOLUTION OF CASCADIA SUBDUCTION WEDGE
TABLE 3. DETRITAL ZIRCON FT AGES FROM ADJACENT PARTS OF THE OLYMPIC SUBDUCTION COMPLEX
Lab
number
Lower OSC
ZD6
ZD50
53
Upper OSC
ZD22
ZD38
ZD44
Nt
Minimum-age peak
Age
95% Conf. Int.
Old peaks
%
Age
95% Conf. Int.
%
Age
95% Conf. Int.
%
50
25
51
18.5
26.5
19.2
21.9
23.8
22.1
12.1
14.4
12.3
27.3
15.1
65.8
43.8
46.4
50.9
23.9
24.4
25.2
14.2
14.9
15.8
37.1
42.5
34.2
67.4
76.8
–
26.1
26.8
–
16.7
17.4
–
31.6
42.4
–
50
61
25
30.9
43.1
47.8
22.0
23.9
23.8
12.2
14.3
14.1
49.7
27.1
60.5
44.9
71.0
71.1
23.9
26.2
29.5
14.2
16.8
111.0
37.3
56.9
31.5
69.5
108.3
146.6
28.2
222.8
229.2
19.3
128.8
136.3
13.0
12.7
8.0
Notes: Samples with ‘‘ZD’’ prefix are from Brandon and Vance (1992) and are recalculated here by using binomial peak-fitting method. See Table 1 for other details.
Figure 3. Simplified geologic map of the Coastal OSC after Rau (1975, 1979), showing
sample locations by laboratory number (Tables 1 to 3). The Coastal OSC is divided into
coherent turbidite sequences (light gray) and mélange sequences (dark gray). White circles—
sandstone samples from coherent sequences of the Coastal OSC; gray circles—sandstone
samples from blocks in mélange. Italic labels mark groups A to E, which represent areas
where a coherent sequence was sampled more than once. The Upper OSC and Lower
OSC (Fig. 2) are undifferentiated on this map but lie to the east of the thrust fault bounding the east side of the Coastal OSC. Comparative samples from those units are reported
in Table 3 and marked on the map with white squares. ZD6 and ZD50 are from the
Lower OSC, and ZD38, ZD44, and ZD22 are from the Upper OSC.
Mountains, in the vicinity of Mount Olympus
(Fig. 2). In that area, resetting is marked by a
reduction in the range of grain ages and by an
increase in the amount of time needed to etch
fission tracks. However, temperatures in that
area failed to bring the zircons to a common
reset age (i.e., the ‘‘reset’’ samples all failed
the x2 test), indicating that zircons had heterogeneous annealing properties during resetting.
Grain-to-grain variations in radiation damage are the most likely cause for this heterogeneous annealing. Most of the radiation damage in zircon is produced not by fission decay
of U, but by alpha decay of U and Th. Alpha
decay produces much smaller tracks, but occurs at a much higher rate relative to fission
decay. Experimental work shows that the time
needed to etch and reveal fission tracks increases as the amount of alpha damage decreases and that alpha damage and fission
tracks in zircon start to anneal at similar temperatures (Tagami et al., 1990). Furthermore,
the annealing temperature for fission tracks is
known to increase with decreasing alpha damage (Kasuya and Naeser, 1988).
From these observations, Brandon and
Vance (1992) suggested the following guidelines for detecting regions in a study area
where partial resetting may have occurred: (1)
The sample shows a marked increase in etch
time compared to other samples in the study
area. The actual change in etch time depends
on the procedures used and also on the time
since thermal resetting. (2) The sample shows
a systematic younging of old grain ages. Zircons with old detrital FT ages will tend to
have the greatest amount of alpha damage and
thus should show the largest reduction in FT
ages. In turn, the youngest fraction of detritalzircon FT grain ages should be the most resistant to partial resetting, especially if the detrital FT ages are close in age to the heating
event.
All of our samples show a large range in
grain ages and normal etch times (4–6 h). As
a result, we are confident that the measured
FT ages have not changed since deposition
and thus provide an accurate record of predepositional cooling events in the source
region.
PEAK FITTING AND MINIMUM AGES
The FTGA distributions for our samples
typically show a large span in grain ages, from
ca. 10 Ma to older than 100 Ma. The probability for the x2 test (Galbraith, 1981) is effectively zero for 34 of the samples and 9%
and 15% for samples 152 and 191, respectively. This result indicates that the variance
Geological Society of America Bulletin, January/February 2004
65
STEWART and BRANDON
Figure 4. Map, showing zircon FT minimum ages from Tables 1 to 3. See Figure 3 for
details.
in grain ages is much larger than expected,
given analytical error alone.
Each FTGA distribution was treated as a
mixture of components or peaks (Brandon,
1992, 1996). In formal terms, a mixed distribution is considered to be a mixture of a finite
set of component distributions, or components. Such components will commonly appear as bumps or peaks in a histogram or
probability density plot, but this need not be
the case given that overlapping component
distributions might appear as a single broad
bump in the density plot. Even so, the term
‘‘peak’’ is often used informally to describe
the components in a distribution.
Galbraith and Laslett (1993) introduced the
term ‘‘minimum age,’’ which can be loosely
viewed as the pooled age of the largest fraction of young concordant grain ages in the
FTGA distribution. They proposed a twocomponent mixture model for estimating the
minimum age. Binomial peak fitting can also
be used to estimate the minimum age, while
at the same time providing information about
older components in the distribution. The
main requirement, however, is that peak fitting
must be coupled with a test to find the maxi-
66
mum number of significant components in the
distribution, as discussed subsequently.
It is useful to ask why the binomial distribution is important for this problem. The measured data are the counts of spontaneous and
induced tracks for each grain, designated as ri
and si with the index indicating the ith grain
in a sample of i 5 1 to Nt dated grains (where
Nt 5 total number of dated grains). The variables ri and si are Poisson distributed, but
they can be converted into a single binomialdistributed variable by the transformation ri/(ri
1 si) (Galbraith and Green, 1990). This variable can be approximated by a Gaussian distribution when values of ri and si are large,
which is commonly the case for zircons, given
their relatively high U content, which implies
a high track density. However, as the cooling
age and U content decrease, ri and si become
small. Thus, the Gaussian approximation
tends to break down for young cooling ages,
especially for apatites, which commonly have
low U contents. In these cases, decomposition
methods based on the Gaussian distribution
(e.g., GAUSSFIT by Brandon, 1992; MIX by
Sambridge and Compston, 1994) will perform
poorly. In contrast, the binomial peak-fitting
method will provide unbiased estimates of
peak ages for track counts of any size.
Binomial peak fitting is based on the
maximum-likelihood method, which means
that the best-fit solution is determined directly
by comparing the distribution of the grain data
to a predicted mixed binomial distribution.
This approach is a significant improvement
over the more common least-squares method,
which requires more restrictive assumptions
about the distribution of the residuals.
A problem with all of the peak-fitting methods is that they require an initial guess of the
number and ages of the peaks in the distribution. The number of peaks could be as large
as the number of grains in the FTGA distribution. So we are left to ask, how many peaks
are significant? It is also important to start
with an initial guess that is not too far away
from the best-fit solution; otherwise the calculation may fail to find the best-fit solution.
We have used a Windows program BINOMFIT, written by M. Brandon, to find bestfit components for the FT data presented here.
The current version of BINOMFIT does an
iterative search of peak ages and number of
peaks to find the best-fit set of significant
components in the mixed distribution. The
program has an automated version of the Ftest approach outlined in Brandon (1992). The
program considers a large set of trial solutions. The trick is to organize the search and
to apply appropriate tests to find the best
solution.
The quality of fit for each of the trial solutions is scored by using the x2 statistic, in a
manner similar to the conventional x2 test.
The strategy for the search is to iterate the
search to include a successively larger number
of peaks and to try a large number of initial
guesses during each iteration to ensure that an
optimal solution is found at each step. The
initial guesses for peak ages are generated by
using the probability density plot for the
FTGA distribution. The density plot is estimated by using the method in Brandon
(1996), and the first and second derivatives of
the plot are used to find bumps and humps in
the plot. The probability density at each candidate age is used to guess the potential size
of the peak. We have tried an alternative approach, using evenly spaced ages, but the
computation takes longer and does not provide
any obvious advantages.
The next step is to iterate through an increasing number of peaks. The first iteration
produces a single component age, with an age
identical to the pooled age and a x2 value
identical to that produced by the conventional
x2 test. The next iteration finds a best-fit two-
Geological Society of America Bulletin, January/February 2004
‘‘HOH FORMATION’’ AND LATE CENOZOIC EVOLUTION OF CASCADIA SUBDUCTION WEDGE
provement in fit is large compared with the
expected random variability associated with
measurements. In other words, we want to be
assured that the additional peak is fitting signal and not noise.
This question is nicely addressed by the F
test. To explain, consider two solutions for m
and m 1 1 peaks and the quality of those fits
as indicated by x2m and x2m11. The F statistic is
given by F 5 (m 1 1)(x2m 2 x2m11)/x2m. When
F is large, then the improvement in fit associated with the additional peak is considered
significant. The F distribution is used to assign a probability P(F), which is the probability that random variation alone could produce the observed F statistic. We consider
P(F) , ;5% to indicate that the improvement
in fit is significant. Thus, we can find the optimal number of significant peaks by adding
peaks until we get a value of P(F) . ;5%.
All of the minimum ages and older peak
ages reported here have been estimated by using this method (Tables 1–3). Uncertainties
are cited at the 95% confidence level. Note
that the uncertainties are asymmetric: the older interval is larger than the younger interval.
Those interested in testing our analysis can
download the BINOMFIT program and all of
our FT grain data at http://www.geology.
yale.edu/;brandon.
RESULTS
Well-stratified Turbidite Sequences
Figure 5. Probability density plots (with histograms) for representative FTGA distributions from the Coastal OSC. Thick lines show the probability density distribution, and
thin lines show the best-fit peaks, as reported in Table 1. The FT minimum age corresponds to the age of the youngest peak. Plots were constructed according to Brandon
(1996). Age is plotted on a logarithmic axis. The probability density scale is the same for
both the density plots and the histograms. Density units are given relative to Dz 5 0.1,
which corresponds to an interval on the age scale approximately equal to 10% of the age.
component solution by using all combinations
of initial peaks as initial guesses. The best solution is the one with the lowest x2 value. The
program then considers a three-component solution, then a four-component solution, and so
on.
With each iteration, the F-test is used to determine whether the introduction of an addi-
tional component has produced a significant
improvement in the x2 statistic (Brandon,
1992). In general, one will find that the x2
statistic gets smaller with the introduction of
a new component. In part, this effect is due to
the fact that the additional component provides the model with greater flexibility to fit
the data. We need to know whether the im-
For most samples in this group (Table 1),
BINOMFIT found just two significant peaks.
Probability density plots and histograms for
selected samples from the coherent sequences
are shown on Figure 5. Particularly interesting
are analyses 143 and 150 (Table 1, Fig. 5),
tuff samples that preserve abundant delicate
glass shards and display clear evidence of contamination by older zircons. FTGA distributions for these samples are so similar to those
from sandstones in the Coastal OSC that we
conclude the provenance for these sediments
must have included active volcanic sources.
The distribution of FT minimum ages for
samples from the coherent well-stratified turbidite sequences (Table 1) are tightly clustered
between 26 and 11 Ma (Fig. 6), essentially
identical to the early and middle(?) Miocene
age assignments of fossils from the Coastal
OSC (Rau, 1975, 1979). In fact, at 62 SE,
only one sample has a minimum age significantly different from early Miocene, an excursion that could have happened by chance
alone in this group of 25 dates. Combined distributions for grouped samples from the co-
Geological Society of America Bulletin, January/February 2004
67
STEWART and BRANDON
Figure 6. Plot of minimum ages and older peak ages for coherent sandstone sequences in the Coastal OSC (Table 1). Samples 168 to
195 are from individual localities. Black triangles and gray circles—minimum ages and older peak ages, respectively. Samples in groups
A to E are from localities where continuous sequences were repeatedly dated. Open triangles and circles—minimum ages and older
peak ages, respectively, for single samples in each group. Black triangles and gray circles—minimum ages and older peak ages, respectively, for the combined grain-age distributions, each of which contains all of the grain ages dated for samples in that group. Error bars
show the 95% confidence intervals. Abbreviations for stratigraphic units are after Remane (2000): LM—late Miocene, MM—middle
Miocene, EM—early Miocene, LO—late Oligocene, EO—early Oligocene, LE—late Eocene, ME—middle Eocene, EE—early Eocene,
LP—late Paleocene, EP—early Paleocene, LK—late Cretaceous.
herent sequences (Table 1, Fig. 6) demonstrate
that peak ages are reproducible from sample
to sample. The peak ages for the combined
distributions are generally similar to those for
the individual sample distributions.
Mélange Blocks in the Coastal OSC
Minimum ages from mélange blocks in the
Coastal OSC range from 39 to 15 Ma (Fig. 7,
Table 2). Although most of these minimum
ages overlap with early Miocene dates from
the coherent sequences, four dates are clearly
Eocene or Oligocene. Because the total number of dated grains in these samples is small,
68
with Nt ranging from 17 to 37 (Table 2), additional dating might resolve younger components. However, mélange rocks in the
Coastal OSC indeed contain Eocene fossil localities (Rau, 1975, 1979), and basalt blocks
in mélange were probably derived from the
Eocene Crescent Formation of the adjacent
Coast Range terrane (Applegate and Brandon,
1989). These data indicate we should not be
surprised to find pre-Miocene sandstone
blocks in mélange in the Coastal OSC.
Aalto et al. (1998) reported Cretaceous to Paleocene 40Ar/39Ar ages for detrital muscovite
from sandstone blocks in mélange in the Coastal
OSC. Zircon FT ages from the same location
(samples 140 and 175, Table 2) are as young as
early Miocene. The discrepancy between the
40
Ar/39Ar ages and zircon FT dates probably results from the different thermal histories for the
detrital muscovite derived from deeply exhumed
granitic and metamorphic rocks and for the
young zircons derived from contemporaneous
volcanic sources. In addition, the effective closure temperature for 40Ar/39Ar muscovite is about
1508 C higher than that for zircon FT.
Detection of Small Peaks
The combined distributions from the coherent sequences illustrate possible problems as-
Geological Society of America Bulletin, January/February 2004
‘‘HOH FORMATION’’ AND LATE CENOZOIC EVOLUTION OF CASCADIA SUBDUCTION WEDGE
occasionally (i.e., 1 time out of 20, given the
95 percent probability level used for our testing). Group E contains a similar example,
where the BINOMFIT solution for sample 184
does not resolve a 21 Ma peak, which is present in the other 4 samples in this group. Because Nt is small for sample 184 (14 grain
ages total), the 21 Ma peak may have been
missed.
Peak-fitting results for individual samples
in Group D indicate a single well-defined
young peak at ;19 Ma, but the combined distribution shows two poorly defined peaks (15
and 21 Ma). This result suggests that the
greater number of grain ages in the combined
distribution provided the ability to resolve two
peaks, whereas the smaller individual samples
can only resolve one peak in the same age
interval. However, the uncertainties for the
ages of the two young peaks are larger because there are fewer grains defining each of
the peaks. In fact, based on the uncertainties,
there is no significant difference between the
minimum ages for the individual samples and
that for the combined result. This result shows
how the estimated uncertainties can be used
to guide judgments about peak resolution.
Figure 7. Plot of FT zircon results for sandstone blocks in mélanges of the Coastal OSC
(Table 2) and for nearby dated localities in adjacent units of the Lower OSC and Upper
OSC (Table 3). Black triangles and gray circles—minimum ages and older peak ages,
respectively. Error bars show the 95% confidence intervals.
sociated with resolution and detection of the
minimum-age component. The binomial distribution (Press et al., 1986) was used to calculate the probability that a peak in an FTGA
distribution would contain a specific number
of grain ages. The result is illustrated in Figure 8, from which we can extract the probability of achieving a specific result. These
probabilities are functions of the true size of
a component and the total number of dated
grains in the sample, each of which influences
the probability for detecting the minimum
number of grains related to that component.
Using this relationship as a guide, we have
tried to date 50 or more grains in each of our
samples to ensure that major peaks are detected. We define a major peak as having a
true size .15% of the distribution. If we dated
a large number of samples, each with Nt .50,
then Figure 8 predicts that 95 percent of the
dated samples should have 5 or more grain
ages for each major peak.
This approach allows us to evaluate with
confidence the problems associated with dating a small number of grains. For example,
the minimum age of sample 42a in Group A
(Table 1) is 14 Ma, but sample 42b does not
have a comparable peak. The combined distribution indicates the 14 Ma component
makes up only ;12 percent of the total distribution, suggesting the absence of a 14 Ma
peak in sample 42b might be due to random
variation associated with sampling. More specifically, did we date enough grains in sample
42b to conclude with confidence that the 14
Ma component is absent?
Reference to the binomial distribution (Fig.
8) indicates if the 14 Ma component makes up
only ;12 percent of the parent population, the
total number of dated grains (Nt) must be $25
to ensure, at the 95% probability level, that
we would find at least one grain age from this
component. To ensure finding 5 or more grains
for this component requires that Nt $80. For
sample 42b, Nt 5 27, so we conclude that we
should have found at least one ‘‘14 Ma’’ zircon if the 14 Ma component was present in
this sample. However, we must use caution in
this interpretation; because of the probabilistic
nature of the sample, our prediction will fail
MINIMUM AGES: VALID PROXIES
FOR DEPOSITIONAL AGE?
Key to our study is the question: Is the FT
minimum age for a detrital-zircon FTGA distribution from an Olympic Structural Complex
sandstone a useful proxy for the depositional
age of the sandstone? We can test this interpretation by comparing FT minimum ages
with fossil ages from Rau (1975, 1979). Of
the samples from the coherent sequences (Table 1), 16 are in demonstrable stratigraphic
continuity with localities that have agediagnostic fossils; these samples were selected
with caution, taking into account the complex
stratigraphy and structure of the Coastal OSC.
For this test, we focus on the lag time (Fig.
9), defined as the FT minimum age minus the
fossil age. The null hypothesis is that the lag
time is equal to zero, and we seek examples
where this hypothesis fails. A Monte Carlo
numerical routine was used to determine the
uncertainty for the lag-time estimate, given the
specified uncertainties for the FT minimum
age and the fossil age. The estimated distribution for the FT minimum age is calculated
by converting the age and its uncertainty (Table 1) to a new variable z, which is known to
be Gaussian distributed (Galbraith, 1990;
Brandon, 1996). Gaussian deviates of z were
generated using the GASDEV program (Press
et al., 1986, p. 203) and then converted back
Geological Society of America Bulletin, January/February 2004
69
STEWART and BRANDON
Figure 8. Graph showing probabilities that a sample grain-age distribution will contain
at least one grain (gray contours) or at least five grains (black contours) from a component
of that distribution. The probabilities are a function of the true size (i.e. expected size) of
the component and the total number of grains dated. The calculated probabilities are
based on the binomial distribution. See text for further discussion.
to an age distribution. The result is a simulated distribution of replicate measurements of
the FT minimum age for the population of zircons represented by our sample.
The estimated distribution of depositional
age was generated by random selection from
a uniform distribution defined by the age
range indicated by the fossil assemblage (Rau,
1975, 1979; Remane, 2000). The replicated
minimum ages and fossil ages were used to
create a distribution of lag time made up of
10,000 replicate values. The distribution was
sorted and used to find the median and 95%
confidence interval for the estimated lag time
(50%, 2.5%, and 97.5%, respectively), which
are shown in Figure 9.
The overall conclusion is that zircon FT
minimum ages seem to do a good job as a
proxy for the depositional ages of sandstones
in the Coastal OSC. Three samples (108, 42a,
168) are unusual in that they have large negative lag times, which means that the FT minimum age is significantly younger than the depositional age. We have no way to judge the
cause of these anomalies. They might be due
to incorrect assignment of fossil ages to these
samples, to problems with the FT dating, or
to random variations in the results.
70
SOURCE FOR SANDSTONES OF
COASTAL OSC
A number of authors have suggested the
possibility of significant coast-parallel transport within the Cascadia margin (e.g., Engebretson et al., 1985; Wells and Heller, 1988;
England and Wells, 1991). A corollary to this
idea is that the sediments that formed the
Olympic subduction complex may have been
deposited much farther south, perhaps offshore of western Oregon or northern California, and then transported northward to the
Olympic Mountains (e.g., Palmer and Lingley,
1989; Davis and Hyndman, 1989; Aalto et al.,
1995). Much of the motivation for this idea
comes from plate reconstructions that predicted oblique convergence across the southern
end of the Cascadia subduction zone (Engebretson et al., 1985; DeMets et al., 1990).
These ideas have recently been thrown into
question by new geodetic data from McCaffrey et al. (2000) that show that the Cascadia arc and forearc are moving as a separate
plate, independent from the North American
plate. Their results indicate that convergence
at the modern subduction zone is nearly orthogonal along the entire length of the Cas-
cadia margin. An important issue is whether
there is any independent evidence for coastparallel transport of the Coastal OSC. We use
provenance information here to evaluate this
possibility.
Aalto et al. (1995) proposed that the upper
Oligocene–lower Miocene Weaverville Formation in the Klamath Mountains might be a
nonmarine remnant of the sedimentary system
that fed the Coastal OSC. Aalto et al. (1998)
evaluated this interpretation by using 40Ar/39Ar
ages for detrital muscovites. The Weaverville
Formation was shown to have been derived
from local bedrock sources in the Klamath
Mountains, whereas sandstone in some of the
mélange blocks of the Coastal OSC was attributed to sources in the Idaho batholith. Heller
et al. (1992) argued for a similar Idaho batholith source for detrital muscovite that occurs
in older units of the Olympic subduction complex. A link to the Idaho batholith might be
taken as evidence for a depositional site for
the Coastal OSC farther to the south. However, Brandon and Vance (1992) showed that
the distinctive muscovite-bearing source identified by Heller et al. (1992) and Aalto et al.
(1998) was part of a long belt of schists and
two-mica granites, extending from the Idaho
batholith northward into southern Canada as
part of the Omineca Crystalline belt. As a result, this provenance link is not very useful
for judging coast-parallel transport.
So we turn our attention here to sedimentary units within the Cascadia forearc region.
In particular, we consider whether any of the
more inboard sedimentary units that underlie
the Cascadia forearc might be upstream equivalents of the sedimentary rocks currently
found in the Coastal OSC. Such equivalents
have never been clearly identified. Candidate
units of the right age (early Miocene) include
the Astoria Formation of western Oregon and
southwestern Washington and the Blakley
Harbor and Clallam Formations, exposed
along the eastern and northern sides of the
Olympic Peninsula (Armentrout et al., 1988;
Brandon et al., 1998). The Clallam Formation
is equivalent to the Sooke Formation, exposed
on the southwest side of Vancouver Island.
Collectively, these two units outline the
Tofino-Makah basin (Fig. 2), which underlies
the western side of the Straits of Juan de Fuca
and the continental shelf west of Vancouver
Island (Snavely et al., 1980; Garver and Brandon, 1994).
So, what do we look for to assess a connection with these more inboard units? Sandstone compositions in the Coastal OSC are
quite variable, but most rocks are lithic arkoses or lithic wackes, with the lithic fragments
Geological Society of America Bulletin, January/February 2004
‘‘HOH FORMATION’’ AND LATE CENOZOIC EVOLUTION OF CASCADIA SUBDUCTION WEDGE
composed mainly of volcanic material (Fig.
10). Our zircon FT results also indicate that
the Coastal OSC received much sediment
from an active volcanic source, probably the
Cascade volcanic arc. In contrast, the Astoria
and Blakley Harbor Formation are dominated
by arkosic sediment, with little volcanic debris
(Rau, 1967; Neim et al., 1992). Most of the
Clallam Formation is also dominated by arkosic sediment (Anderson, 1985) and lacks
young zircon FT ages (Garver and Brandon,
1994). However, the stratigraphically youngest part of the Clallam is volcaniclastic rather
than arkosic (facies 5 of Anderson, 1985).
We propose that this upper part of the Clallam Formation (Fig. 2) is the upstream equivalent to the Coastal OSC. This unit is the type
locality of the Pillarian molluscan stage of the
Pacific Northwest biostratigraphic standard,
which has an age of 24–20 Ma (Armentrout,
1987), virtually identical to the depositional
age of the Coastal OSC. Petrographic data
from Stewart (1970) and Anderson (1985) indicate similar sedimentary compositions for
these two units (Fig. 10). Sandstones in the
upper Clallam contain abundant andesitic and
dacitic rock fragments—undoubtedly derived
from the Cascade arc—as well as minor chert
detritus and rare metasedimentary and metabasalt detritus. These constituents are typical
of the Coastal OSC sandstones. Neither the
upper Clallam nor the Coastal OSC contain
high-grade metamorphic detritus.
The Clallam Formation and associated
Tofino-Makah basin are tied stratigraphically
to Vancouver Island and can be considered a
stable reference within the interior of the Cascadia forearc. Thus, our proposed correlation
suggests that the Coastal OSC formed in place
relative to this part of the Cascadia forearc. In
fact, it is useful to note that during the early
Miocene, basinal sequences within the Cascadia forearc were overwhelmed by arkosic
sediment, mainly derived from east of the arc
(Neim et al., 1992). We were unable to find
any basinal sequence, other than the upper
Clallam, that has the volcaniclastic sandstone
composition characteristic of the Coastal
OSC.
DEFORMATION AND ACCRETION OF
THE COASTAL OLYMPIC
STRUCTUAL COMPLEX
Most of the Coastal OSC consists of steeply
east-dipping strata, with sedimentary structures indicating younging to the east (Rau,
1975, 1979; Tabor and Cady, 1978a). The strata are undoubtedly repeated by thrust faults;
otherwise the Coastal OSC would have a
Figure 9. Lag times for samples where the depositional age is constrained by fossils. Lag
time is defined as the FT minimum age minus the depositional age. Point and error bars
show the expected value (median) and the 95% confidence interval for that estimate. For
most samples, the lag time does not appear to be significantly different from zero, which
is consistent with the interpretation that the FT minimum age is a good proxy for depositional age. See text for details.
Figure 10. Triangular diagrams, showing modal compositions of sandstones from the
Coastal OSC. Also shown are data for sandstones from facies 5 of the Clallam Formation,
which are considered to be an upstream remnant of the sedimentary system that fed the
Structural Complex. Data for the coherent sandstones are from Stewart (1970), and data
for sandstone blocks in the mélanges are from Aalto et al. (1998). The data for the Clallam
Formation are from Anderson (1985). QFL diagram: Q—total quartzose fragments, F—
total feldspars, and L—total rock fragments. Qp-Lv-Ls diagram: Qp—total polycrystalline
quartz, Lv—total volcanic rock fragments, and Ls—total sedimentary rock fragments.
Geological Society of America Bulletin, January/February 2004
71
STEWART and BRANDON
Figure 11. Schematic cross sections of the Cascadia wedge. (A) Steep imbricate structure
as proposed by Rau (1975, 1979) and Tabor and Cady (1978a, 1978b), and (B) domal
imbricate structure as proposed by Brandon and Calderwood (1990) and Brandon and
Vance (1992). Abbreviations refer to tectonic units exposed in the Olympic Mountains:
U—Upper OSC, L—Lower OSC, and C—Coastal OSC.
stratigraphic thickness of .20 km. On the basis of these kinds of arguments, Rau (1973,
1980) and Tabor and Cady (1978a, 1978b)
proposed a steeply imbricated structure for the
entire Olympics, as schematically shown in
Figure 11A. This interpretation predicts that
depositional ages should get systematically
younger from east to west. This relationship
does hold at the scale of the entire Olympic
subduction complex. Accreted rocks are latest
Eocene in age in the eastern Olympics (Upper
OSC), early Miocene in the western Olympics
(Lower and Coastal OSC), and Quaternary at
the modern toe of the wedge.
Our dating here was motivated, in part, to
see whether this pattern of younging was present at the local scale in the western side of the
Olympic Peninsula. Brandon and Vance
(1992) proposed that the Lower OSC was a
more deeply exhumed equivalent of the Coastal OSC. This interpretation was based on three
unreset zircon FTGA samples from the Lower
OSC with minimum ages ranging from 26 to
18 Ma. Two of these FT minimum ages are
shown here in Table 3 and Figures 3, 4, and
7. Our new results for the Coastal OSC demonstrate that a 50-km-wide area, extending
from the Pacific Coast to Mount Olympus, is
underlain by accreted lower Miocene sedimentary rocks (Fig. 4). There is subtle evidence of younging across this region from ca.
19 Ma at Mount Olympus to ca. 14 Ma at the
coast (Fig. 4).
We can make a rough prediction for the expected amount of younging in accreted sedimentary rocks if the structure were steeply
dipping as suggested in Figure 11A. It should
be approximately equal to the distance across
the region divided by the horizontal material
velocity relative to the wedge front. Pazzaglia
and Brandon (2001) estimated that at the
coast, this velocity has been steady at 3 km/
m.y. Thus, a steep imbricate structure should
show ;17 m.y. of younging across the 50 km
distance from the Pacific Coast to Mount
Olympus.
An alternative interpretation (Fig. 11B) is
that the Olympics have a more domal structure (Brandon and Vance, 1992; Brandon et
al., 1998). The structural lid is thought to have
originally extended to the west coast and perhaps farther offshore, as is the case for southern Vancouver Island and also for western
Oregon and southwestern Washington (Brandon and Calderwood, 1990). Early emergence
of the Olympics at ca. 15 Ma, relative to the
rest of the Cascadia forearc high, allowed for
uplift and deep erosion of the Olympics, removing the lid and exposing the underlying
subduction complex (Brandon and Calderwood, 1990; Brandon et al., 1998).
The more domal structure accounts for the
broad expanse of accreted lower Miocene sedimentary rocks. Figure 12 shows schematically how material might have moved through
the wedge. This interpretation builds on the
Figure 12. Inferred displacement path for lower Miocene sedimentary rocks accreted into the Cascadia subduction wedge. Rectangular
gray boxes mark successive locations of the Coastal OSC as it progressed from deposition at the deformation front to its current position
where it is exposed in the western Olympic Mountains. The location of cross-section A–A9 is in Figure 2.
72
Geological Society of America Bulletin, January/February 2004
‘‘HOH FORMATION’’ AND LATE CENOZOIC EVOLUTION OF CASCADIA SUBDUCTION WEDGE
conclusion that in the Olympics sector of the
Cascadia margin, the wedge probably reached
its present size early in its evolution, at ca. 15
Ma, and that accretion occurred primarily at
the front of the wedge (Brandon et al., 1998;
Pazzaglia and Brandon, 2001; Batt et al.,
2001). Thus, we envision that the Coastal
OSC was first accreted beneath the front 50 to
100 km of the wedge, but then moved rearward within the wedge because of further accretion at the front of the wedge and erosion
at the back of the wedge. As previously noted,
the most western exposures of the Coastal
OSC have never been deeply buried and exhumed. Thus, those rocks are shown in Figure
12 as moving along a path near the surface of
the wedge. Conversely, the more eastern lower
Miocene rocks of the Lower OSC followed a
deeper path through the wedge (Fig. 11),
reaching maximum depths of as much as 13
km, but ultimately rising to the surface near
Mount Olympus (Batt et al., 2001).
We can test this idea by comparing the age
of the Coastal OSC to the amount of time
needed to reach the coast within a wedge that
maintained a steady taper. This calculation is
relatively easy because the frontal part of the
wedge is not eroding and has probably maintained the same critical taper throughout much
of its evolution. In this case, the average horizontal velocity ū(x) at a distance x from the
front of the wedge is related to the accretionary flux Fa by
ū(x) 5
Fa
h(x)
(1)
Figure 13. Transport time needed to move from the site of initial accretion at the front
of the wedge to a location rearward in the wedge, according to equation 2 in the text.
where A(x) is the cross-sectional area of the
wedge between x and the front of the wedge.
Thermal-kinematic modeling by Batt et al.
(2001) indicates a long-term average accretionary flux of 58 km2/m.y. into the Olympics
sector of the Cascadia wedge. The profile in
Figure 11 was used to calculate A(x). Equation
2 was then used to estimate the transport time
through the wedge (Fig. 13). The prediction is
that material accreted at the deformation front
of the trench would reach the West Coast in
22 m.y., which is in excellent agreement with
our estimate of 24 to 16 Ma for deposition of
the Coastal OSC in the trench.
CONCLUSIONS
where h(x) is the thickness of the wedge at x
(see Dahlen and Barr [1989] and Pazzaglia
and Brandon [2001] for details about this calculation). The only assumptions are that the
wedge seaward of the coast has maintained a
constant taper and that the accretionary flux
has been steady over the time frame of interest. In other words, the full wedge does not
have to be at steady state for equation 1 to
hold. The accretionary flux is specified in
terms of solid rock mass, which is consistent
with the fact that most of the porosity is lost
within the frontal 10 km of the wedge (Davis
and Hyndman, 1989).
We have to integrate the velocity along the
particle path to get t(x), the transit time from
the trench to a position x in the wedge. The
integration gives
t(x) 5
A(x)
Fa
(2)
FT grain ages from detrital zircons in sandstones of the Coastal OSC indicate a mixed
source of detrital zircons, derived from the
Cascade volcanic arc and from older basement
rocks lying to the east of the arc. The youngest
components make up, on average, ;40% to
50% of a typical zircon FTGA distribution.
Comparison with fossil ages indicates that the
FT minimum age, which is the pooled age of
the youngest component, is a good proxy for
the depositional age of sandstones from the
Coastal OSC. This result is consistent with
other evidence indicating that the volcanic
sediment was derived from contemporaneous
active volcanoes in the arc. The zircon FT
minimum ages indicate that the Coastal OSC
is mainly early Miocene in age, which is in
close agreement with fossil ages.
Basinal units in the Cascadia forearc are
mainly arkosic, which is in contrast to the
more volcaniclastic composition of the Coast-
al OSC. We have found only one coeval unit,
within the upper part of the Clallam Formation on the north side of the Olympic Peninsula, that has a similar sedimentary composition. This provenance correlation argues
against large-scale coast-parallel transport of
the Coastal OSC after accretion. Instead,
transport appears to have mainly occurred to
the northeast, in association with marginnormal shortening of the Cascadia wedge, in
the direction of plate convergence.
We present a simple model that predicts the
average transport time for frontally accreted
materials moving through a wedge with a
steady taper. The model predicts that frontally
accreted sedimentary rocks should take ;22
m.y. to reach the present position of the Coastal OSC, located ;140 km landward of the
front of the wedge. This estimate assumes a
steady taper like the modern taper of the
wedge and a steady accretionary flux like the
modern. The predicted transport time is in
close agreement with the early Miocene age
of the Coastal OSC and thus is consistent with
other evidence for a flux steady state in the
Olympics sector of the Cascadia wedge (Brandon et al., 1998; Pazzaglia and Brandon, 2001;
Batt et al., 2001).
ACKNOWLEDGMENTS
This work would not have been possible without
the mapping, often under difficult conditions, by
Weldon Rau, Rowland Tabor, Parke D. Snavely, Jr.,
and Wally Cady. John Garver generously supplied
data from samples he collected on the Pacific Coast.
The assistance of J.A. Vance in fission-track analysis was invaluable. This study was supported in
part by grants from the Washington State Department of Natural Resources (to Stewart) and from
Geological Society of America Bulletin, January/February 2004
73
STEWART and BRANDON
the American Chemical Society Petroleum Research
Fund (35160-AC2) and the National Science Foundation (208371) (to Brandon).
REFERENCES CITED
Aalto, K.R., McLaughlin, R.J., Carver, G.A., Barron, J.A.,
Sliter, W.V., and McDougall, K., 1995, Uplifted Neogene margin, southernmost Cascadia–Mendocino triple junction region, California: Tectonics, v. 14,
p. 1104–1116.
Aalto, K.R., Sharp, W.D., and Renne, P.R., 1998, 40Ar/39Ar
dating of detrital micas from Oligocene–Pleistocene
sandstones of the Olympic Peninsula, Klamath Mountains, and northern California Coast Ranges: Provenance and paleodrainage patterns: Canadian Journal of
Earth Sciences, v. 35, p. 735–745.
Anderson, K.S., 1985, Sedimentology, sedimentary petrology, and tectonic setting of the lower Miocene Clallam Formation, northern Olympic Peninsula, Washington [M.S. thesis]: Bellingham, Washington,
Western Washington University, 135 p.
Applegate, J.D.R., and Brandon, M.T., 1989, An upper
plate origin for basalt blocks in the Cenozoic subduction complex of the Olympic Mountains, northwest
Washington State: Geological Society of America Abstracts with Programs, v. 21, no. 5, p. 51.
Armentrout, J.M., 1987, Cenozoic stratigraphy,
unconformity-bounded sequences, and tectonic history of southwestern Washington: Washington Division of Geology and Earth Resources Bulletin 77,
p. 291–320.
Armentrout, J.M., Heintz, L.F., Hull, D.A., Beaulieu, J.D.,
and Rau, W.W., 1988, Northwest region correlation
chart: Correlation of stratigraphic units of North
America (COSUNA): Tulsa, Oklahoma, American Association of Petroleum Geologists.
Arne, D., and Zentilli, M., 1994, Apatite fission track thermochronology integrated with vitrinite reflectance—A
review, in Mukhopadhyay, P.K., and Dow, W.G., eds.,
Vitrinite reflectance as a maturity parameter—
Applications and limitations: American Chemical Society Symposium Series 570, p. 249–268.
Arnold, R., 1905, Geological reconnaissance of the coast
of the Olympic Peninsula, Washington: Geological
Society of America Bulletin, v. 17, p. 451–468.
Batt, G.E., Brandon, M.T., Farley, K.A., and Roden-Tice,
M., 2001, Tectonic synthesis of the Olympic Mountains segment of the Cascadia wedge, using twodimensional thermal and kinematic modeling of thermochronological ages: Journal of Geophysical
Research, v. 106, p. 26,731–26,746.
Beaumont, C., Ellis, S., and Pfiffner, A., 1999, Dynamics
of sediment subduction-accretion at convergent margins: Short-term modes, long-term deformation, and
tectonic implications: Journal of Geophysical Research, v. 104, p. 17,573–17,602.
Brandon, M.T., 1992, Decomposition of fission-track grainage distributions: American Journal of Science,
v. 292, p. 535–564.
Brandon, M.T., 1996, Probability density plots for fissiontrack age distributions: Radiation Measurements,
v. 26, p. 663–676.
Brandon, M.T., and Calderwood, A.R., 1990, High-pressure
metamorphism and uplift of the Olympic subduction
complex: Geology, v. 18, p. 1252–1255.
Brandon, M.T., and Vance, J.A., 1992, Tectonic evolution
of the Cenozoic Olympic subduction complex, Washington State, as deduced from fission track ages for
detrital zircons: American Journal of Science, v. 292,
p. 565–636.
Brandon, M.T., Roden-Tice, M.K., and Garver, J.I., 1998,
Late Cenozoic exhumation of the Cascadia accretionary wedge in the Olympic Mountains, northwest
Washington State: Geological Society of America Bulletin, v. 110, p. 985–1009.
Clowes, R.M., Brandon, M.T., Green, A.G., Yorath, C.J.,
Sutherland Brown, A., Kanasewich, E.R., and Spencer, C., 1987, LITHOPROBE—Southern Vancouver
Island: Cenozoic subduction complex imaged by deep
74
seismic reflections: Canadian Journal of Earth Sciences, v. 24, p. 31–51.
Dahlen, F.A., and Barr, T.D., 1989, Brittle frictional mountain building: 1. Deformation and mechanical energy
budget: Journal of Geophysical Research, v. 94,
p. 3906–3922.
Davis, E.E., and Hyndman, R.D., 1989, Accretion and recent deformation of sediment along the northern Cascadia subduction zone: Geological Society of America
Bulletin, v. 101, p. 1465–1480.
Davis, E.E., Hyndman, R.D., and Villinger, H., 1990, Rates
of fluid expulsion across the northern Cascadia accretionary prism: Constraints from new heat flow and
multichannel seismic reflection data: Journal of Geophysical Research, v. 95, p. 8869–8889.
DeMets, C., and Dixon, T., 1999, New kinematic models
for Pacific–North America motion from 3 Ma to present: I. Evidence for steady motion and biases in the
NUVEL-1A model: Geophysical Research Letters,
v. 26, no. 13, p. 1921–1924.
DeMets, C., Gordon, R.G., Argus, D.F., and Stein, S., 1990,
Current plate motions: Geophysical Journal International, v. 101, p. 425–478.
Dickinson, W.R., and Seely, D.R., 1979, Structure and stratigraphy of forearc regions: American Association of
Petroleum Geologists Bulletin, v. 63, p. 2–31.
Dumitru, T.A., 1991, Major Quaternary uplift along the
northernmost San Andreas fault, King Range, northwest California: Geology, v. 19, p. 526–529.
Engebretson, D.C., Gordon, R.G., and Cox, A., 1985, Relative motions between oceanic and continental plates
in the Pacific Basin: Geological Society of America
Special Paper 206, 59 p.
England, P., and Wells, R.E., 1991, Neogene rotations and
quasicontinuous deformation of the Pacific Northwest
continental margin: Geology, v. 19, p. 978–981.
Galbraith, R.F., 1981, On statistical models for fission-track
counts: Journal of Mathematical Geology, v. 13,
p. 471–478.
Galbraith, R.F., 1990, The radial plot: Graphical assessment
of spread in ages: Nuclear Tracks and Radiation Measurements, v. 17, p. 207–214.
Galbraith, R.F., and Green, P.F., 1990, Estimating the component ages in a finite mixture: Nuclear Tracks and
Radiation Measurements, v. 17, p. 197–206.
Galbraith, R.F., and Laslett, G., 1993, Statistical models for
mixed fission track ages: Nuclear Tracks and Radiation Measurements, v. 21, p. 459–470.
Garver, J.I., and Brandon, M.T., 1994, Erosional denudation
of the British Columbia Coast Ranges as determined
from fission-track ages of detrital zircon from the Tofino basin, Olympic Peninsula, Washington: Geological Society of America Bulletin, v. 106,
p. 1398–1412.
Garver, J.I., Brandon, M.T., Roden-Tice, M., and Kamp,
P.J.J., 1999, Exhumation history of orogenic highlands
determined by detrital fission-track thermochronology,
in Ring, U., Brandon, M.T., Lister, G.S., and Willett,
S.D., eds., Exhumation processes: Normal faulting,
ductile flow, and erosion: Geological Society [London] Special Publication 154, p. 283–304.
Goldfinger, C., Kulm, L.D., McNeill, L.C., and Watts, P.,
2000, Super-scale failure of the southern Oregon Cascadia margin: Pure and Applied Geophysics, v. 157,
p. 1189–1226.
Green, P.F., Hegarty, K.A., Duddy, I.R., Foland, S.S., and
Gorbachev, V., 1996, Geological constraints on fission
track annealing in zircon [abs.]: Ghent, Belgium, Proceedings of International Workshop on Fission Track
Dating, August 1996.
Heller, P.L., Tabor, R.W., O’Neil, J.R., Pevear, D.R., Shafiqullah, M., and Winslow, N.S., 1992, Isotopic provenance of Paleogene sandstones from the accretionary
core of the Olympic Mountains, Washington: Geological Society of America Bulletin, v. 104, p. 140–153.
Hsü, K.J., 1968, Principles of mélanges and their bearing
on the Franciscan-Knoxville paradox: Geological Society of America Bulletin, v. 79, p. 1063–1074.
Hurford, A.J., and Green, P.F., 1983, The zeta age calibration of fission-track dating: Chemical Geology (Isotope Geoscience Section), v. 1, p. 285–317.
Ingle, J.C., Jr., 1980, Cenozoic paleobathymetry and de-
positional history of selected sequences within the
southern California Continental Borderland: Cushman
Foundation for Foraminiferal Research Special Publication 19, p. 163–195.
Kamp, P.J.J., 2001, Possible Jurassic age for part of Rakaia
terrane: Implications for the tectonic development of
the Torlesse accretionary prism: New Zealand Journal
of Geology and Geophysics, v. 44, p. 185–203.
Kasuya, M., and Naeser, C.W., 1988, The effect of adamage on fission-track annealing in zircon: Nuclear Tracks and Radiation Measurements, v. 14,
p. 477–480.
Lupton, C.T., 1914, Oil and gas in the western part of the
Olympic Peninsula, Washington: U.S. Geological Survey Bulletin 581-B, p. 23–81.
McCaffrey, R., Long, M.D., Goldfinger, C., Zwick, P., Nabelek, J., and Smith, C., 2000, Rotation and plate locking at the southern Cascadia subduction zone: Geophysical Research Letters, v. 27, p. 3117–3120.
McLaughlin, R.J., Kling, S.A., Poore, R.Z., McDougall, K.,
and Buetner, E.C., 1982, Post–middle Miocene accretion of Franciscan rocks, northwestern California:
Geological Society of America Bulletin, v. 93,
p. 595–605.
McLaughlin, R.J., Sliter, W.V., Frederiksen, N.O., Harbert,
W.P., and McCulloch, D.S., 1994, Plate motions recorded in tectonostratigraphic terranes of the Franciscan Complex and evolution of the Mendocino triple
junction, northwestern California: U.S. Geological
Survey Bulletin 1997, 60 p.
Moore, G.F., and Karig, D.E., 1976, Development of sedimentary basins on the lower trench slope: Geology,
v. 4, p. 693–697.
Neim, A.R., MacLeod, N.S., Snavely, P.D., Jr., Huggins,
D., Fortier, J.D., Meyer, J., Seeling, A., and Neim,
W.A., 1992, Onshore-offshore geologic cross section,
northern Oregon coast range to continental slope:
Portland, Oregon Department of Geology and Mineral
Industries Special Paper 26, 10 p.
Orange, D.L., 1990, Criteria helpful in recognizing shearzone and diapiric mélanges: Examples from the Hoh
accretionary complex, Olympic Peninsula, Washington: Geological Society of America Bulletin, v. 102,
p. 935–951.
Orange, D.L., and Underwood, M.B., 1995, Patterns of
thermal maturity as diagnostic criteria for interpretation of mélanges: Geology, v. 23, p. 1144–1148.
Orange, D.L., Geddes, D.S., and Moore, J.C., 1993, Structural and fluid evolution of a young accretionary complex: The Hoh rock assemblage of the western Olympic Peninsula, Washington: Geological Society of
America Bulletin, v. 105, p. 1053–1075.
Palmer, S.P., and Lingley, W.S., Jr., 1989, An assessment
of the oil and gas potential of the Washington outer
continental shelf: Seattle, Washington, University of
Washington Sea Grant Program, 83 p.
Parsons, T., Trehu, A.M., Luetgert, J.H., Miller, K., Kilbride, F., Wells, R.E., Fisher, M.A., Flueh, E., ten
Brink, U.S., and Christensen, N.I., 1998, A new view
into the Cascadia subduction zone and volcanic arc:
Implications for earthquake hazards along the Washington margin: Geology, v. 26, p. 199–202.
Pazzaglia, F.J., and Brandon, M.T., 2001, A fluvial record
of long-term steady-state uplift and erosion across the
Cascadia forearc high, western Washington State:
American Journal of Science, v. 301, p. 385–431.
Press, W.H., Flannery, B.P., Teukolsky, S.A., and Vetterling,
W.T., 1986, Numerical recipes: Cambridge, England,
Cambridge University Press, 818 p.
Rau, W.W., 1967, Geology of the Wynoochee Valley Quadrangle, Grays Harbor County, Washington: Washington Division of Mines and Geology Bulletin 56, 51 p.
Rau, W.W., 1970, Foraminifera, stratigraphy, and paleoecology of the Quinault Formation, Point Grenville–
Raft River coastal area, Washington: Washington Division of Mines and Geology Bulletin 62, 41 p.
Rau, W.W., 1973, Geology of the Washington coast between Point Grenville and the Hoh River: Washington
Division of Geology and Earth Resources Bulletin 66,
58 p.
Rau, W.W., 1975, Geologic map of the Destruction Island
and Taholah Quadrangles, Washington: Washington
Geological Society of America Bulletin, January/February 2004
‘‘HOH FORMATION’’ AND LATE CENOZOIC EVOLUTION OF CASCADIA SUBDUCTION WEDGE
Division of Geology and Earth Resources, Map GM13, scale 1:62,500, 1 plate.
Rau, W.W., 1979, Geologic map in the vicinity of the lower
Bogachiel and Hoh River valleys, and the Washington
coast: Washington Division of Geology and Earth Resources, Map GM-24, scale 1:62,500, 1 plate.
Rau, W.W., 1980, Washington coastal geology between the
Hoh and the Quillayute Rivers: Washington Division
of Geology and Earth Resources Bulletin 72, 57 p.
Rau, W.W., and Grocock, G.R., 1974, Piercement structure
outcrops along the Washington coast: Washington Division of Geology and Earth Resources Information
Circular 51, 7 p.
Remane, J., 2000, International stratigraphic chart: Geneva,
Switzerland, UNESCO and International Union of
Geological Sciences.
Salvador, A., editor, 1994, International Stratigraphic Guide—
A guide to stratigraphic classification, terminology,
and procedure, International Subcommission on Stratigraphic Classification, 2nd Edition: The International
Union of Geological Sciences and Geological Society
of America, 214 p.
Sambridge, M.S., and Compston, W., 1994, Mixture modeling of multi-component data sets with application to
ion-probe zircon ages: Earth and Planetary Science
Letters, v. 128, p. 373–390.
Snavely, P.D. Jr., and Kvenvolden, K.A., 1989, Preliminary
evaluation of the petroleum potential of the Tertiary
accretionary terrane, west side of the Olympic Peninsula Washington: A. Geology and hydrocarbon
potential: U.S. Geological Survey Bulletin 1892,
p. 1–17.
Snavely, P.D., Jr., and Wells, R.E., 1996, Cenozoic evolution of the continental margin of Oregon and Washington, in Rogers, A., Walsh, T., and Kockelman, W.,
eds., Assessing earthquake hazards and reducing risk
in the Pacific Northwest: U.S. Geological Survey Professional Paper 1560, p. 161–181.
Snavely, P.D., Jr., Neim, A.R., MacLeod, N.S., Pearl, J.E.,
and Rau, W.W., 1980, Makah Formation—A deepmarginal-basin sequence of late Eocene and Oligocene
age in the northwestern Olympic Peninsula, Washington: U.S. Geological Survey Professional Paper 1162B, 28 p.
Snavely, P.D., Jr., MacLeod, N.S., and Neim, A.R., 1993,
Geologic map of the Cape Flattery, Clallam Bay, Ozette Lake, and Lake Pleasant quadrangles, northwestern Olympic Peninsula, Washington: U.S. Geological
Survey Miscellaneous Investigations Map I-1946,
scale 1:48,000, 1 plate.
Stewart, R.J., 1970, Petrology, metamorphism and structural relations of graywackes in the western Olympic
Peninsula, Washington [Ph.D. thesis]: Stanford, California, Stanford University, 107 p.
Stewart, R.J., 1974, Zeolite facies metamorphism of sandstone in the western Olympic Peninsula, Washington:
Geological Society of America Bulletin, v. 85,
p. 1139–1142.
Tabor, R.W., and Cady, W.M., 1978a, Geologic map of the
Olympic Peninsula, Washington: U.S. Geological Survey Miscellaneous Investigations Map I-994, scale 1:
125,000, 2 sheets.
Tabor, R.W., and Cady, W.M., 1978b, Structure of the
Olympic Mountains, Washington—Analysis of a subduction zone: U.S. Geological Survey Professional
Paper 1033, 38 p.
Tabor, R.W., Cady, W.M., and Yeats, R.S., 1970, Broken
formations and thrust faulting in the northeastern
Olympic Mountains, Washington: Geological Society
of America Abstracts with Programs, v. 2, p. 152.
Tagami, T., Ito, H., and Nishimura, S., 1990, Thermal annealing characteristics of spontaneous fission tracks in
zircon: Chemical Geology (Isotope Geoscience Section), v. 80, p. 159–169.
Wagner, G.S., and Van den Haute, P., 1992, Fission-track
dating: London, Kluwer Academic Publishers, 285 p.
Weaver, C.E., 1916, The Tertiary formations of western
Washington: Washington Geological Survey Bulletin
13, 327 p.
Weaver, C.E., 1937, Tertiary stratigraphy of western Washington and northeastern Oregon: Washington University Publications in Geology, v. 4, 266 p.
Weissenborn, A.E., and Snavely, P.D., 1968, Summary report on the geology and mineral resources of Flattery
Rocks, Quillayute Needles, and Copalis National
Wildlife Refuges, Washington: U.S. Geological Survey Bulletin 1260-F, p. F1–F16.
Wells, R.E., and Heller, P.L., 1988, The relative contribution
of accretion, shear, and extension to Cenozoic tectonic
rotation in the Pacific Northwest: Geological Society
of America Bulletin, v. 100, p. 325–338.
Willett, S., 1999, Orogeny and orography: The effects of
erosion on the structure of mountain belts: Journal of
Geophysical Research, v. 104, p. 28,957–28,981.
Willett, S., Beaumont, C., and Fullsack, P., 1993, Mechanical models for the tectonics of doubly vergent compressional orogens: Geology, v. 21, p. 371–374.
Wilson, D.S., 1988, Tectonic history of the Juan de Fuca
Ridge over the last 40 million years: Journal of Geophysical Research, v. 93, p. 11,863–11,876.
MANUSCRIPT RECEIVED BY THE SOCIETY 23 JULY 2001
REVISED MANUSCRIPT RECEIVED 20 MARCH 2003
MANUSCRIPT ACCEPTED 31 MARCH 2003
Printed in the USA
Geological Society of America Bulletin, January/February 2004
75