The Olympic Mountains of Washington state (USA) represent
the aerially exposed accretionary wedge of the Cascadia Subduction Zone and
are thought to be in flux steady state, whereby the mass outflux (denudation)
and influx (tectonic accretion) into the mountain range are balanced. We use
a multi-method approach to investigate how temporal variations in the influx
and outflux could affect previous interpretations of flux steady state. This
includes the analysis of published and new thermochronometric ages for (U–Th)
In total, we present 61 new AHe, ZHe, AFT, and ZFT thermochronometric ages
from 21 new samples. AHe ages are generally young (< 4 Ma), and, in
some samples, AFT ages (5–8 Ma) overlap ZHe ages (7–9 Ma) within
uncertainties. Thermo-kinematic modeling shows that exhumation rates are
temporally variable, with rates decreasing from > 2 to
< 0.3 km Myr
The assumption of a balance between opposing processes has allowed geoscientists to use proxy measurements (like denudation rates) to constrain difficult-to-measure variables like rock uplift. This has given rise to the concept of steady-state landscapes or mountain ranges. Likewise, a steady state (i.e., a mass balance) is commonly one of the boundary conditions in modeling studies investigating the evolution and dynamics of orogens in response to changes of other boundary conditions like climate or tectonic fluctuations (e.g., Batt et al., 2001; Stolar et al., 2007; Whipple and Meade, 2006; Willett, 1999). Two main types of steady state are often used to interpret mountain-building processes (e.g., Willett and Brandon, 2002): (1) topographic steady state, in which the topography is invariant because rock uplift and horizontal motion of material are balanced by denudation, and (2) flux steady state, in which the material influx (by accretion of sediment and rock) is balanced by the material outflux (by denudation) from a mountain range. The assumption of steadiness is both spatial- and timescale-dependent so that for a given timescale, steadiness might only be achieved on a large, orogen-wide spatial scale due to the spatial averaging of single processes acting on a small scale (e.g., catchment-wide sediment discharge vs. orogen-wide sediment discharge). Furthermore, a possible perturbation of steady state is sensitive to the timescale it takes for orogens to respond to variations in crustal deformation or a change in climate. If the timescales required for a change in the influx and outflux are significantly different from each other, a deviation from steady state is likely.
Likewise, studies from different orogens worldwide suggest strong variations in denudation and exhumation on million-year timescales. These variations can be linked to changes in tectonic conditions (e.g., Adams et al., 2015; Lease et al., 2016), the internal dynamics of drainage basins (e.g., Willett et al., 2014; Yanites et al., 2013), changes in the magnitude of precipitation (e.g., Lease and Ehlers, 2013; Whipple, 2009), or the onset of glaciation (e.g., Berger et al., 2008; Bernard et al., 2016; Ehlers et al., 2006; Glotzbach et al., 2013; Gulick et al., 2015; Herman et al., 2013; Herman and Brandon, 2015; Lease et al., 2016; Thomson et al., 2010, 2013; Valla et al., 2011; Yanites and Ehlers, 2012).
Based on thermo-kinematic modeling of thermochronometric cooling ages, the Olympic Mountains, USA (Fig. 1a), have been proposed to be in flux steady state since ca. 14 Ma (Batt et al., 2001; Brandon et al., 1998). The approach of these studies was to assume flux steady state along a two-dimensional profile across the Olympic Peninsula as a precondition in order to derive the kinematics of the model from the balance between accretionary influx (governed by the thickness of accreted sediment and plate convergence rate) and denudational outflux (as set by exhumation rates). Because the cooling ages can successfully be modeled with the kinematics used, the mountain range is then interpreted to be in flux steady state. However, possible temporal variations in parameters like sediment thickness, plate convergence rate, or exhumation rates were not considered in these studies. Likewise, the impact of Plio–Pleistocene glaciation on the flux steady-state hypothesis has not been considered yet, although the range was extensively incised by glaciers (Adams and Ehlers, 2017; Montgomery, 2002; Montgomery and Greenberg, 2000; Porter, 1964) and experienced significant changes in climate conditions over the past 3 Myr (Mutz et al., 2018). Numerical modeling studies investigated the mechanics of the wedge by either considering fluvial erosion (Stolar et al., 2007) or glacial erosion (Tomkin and Roe, 2007). A significant response of the orogenic wedge to glaciation was suggested (Tomkin and Roe, 2007) and recent studies proposed that exhumation rates in the Olympic Mountains increased due to Plio–Pleistocene glacial erosion (Herman et al., 2013; Michel et al., 2018). Resulting high sedimentation rates during the Quaternary increased the sediment thickness on the oceanic plate and seem to have caused a change in the deformational style of the offshore part of the wedge (Adam et al., 2004).
In this study, we test the hypothesis of flux steady state in the Olympic
Mountains considering variations in both the material influx and outflux.
First, we test the temporal steadiness of exhumation rates from bedrock
cooling histories with a 1-D thermo-kinematic model, capitalizing on new
samples that have been dated with three to four thermochronometers (apatite
and zircon (U–Th)
At present, the Juan de Fuca Plate subducts obliquely with respect to the
overriding North American Plate (Fig. 1a) at 34 mm yr
The accretionary wedge of the subduction zone is exposed onshore within the
Olympic Mountains (Fig. 1a) and is composed of Eocene–Miocene flysch
(Brandon et al., 1998; Tabor and Cady, 1978). This part of the mountain range
is known as the Olympic Structural Complex (Brandon et al., 1998) and is
separated from the surrounding Coast Range Terrane by the Hurricane Ridge
thrust fault (HRF; Fig. 1c), a major discontinuity traceable in seismic
surveys (e.g., Clowes et al., 1987; Calvert et al., 2011). Minor sedimentary
rocks of Eocene age (Eddy et al., 2017; Tabor and Cady, 1978) are contained
within the Coast Range Terrane besides the predominant
Plio–Pleistocene glaciation has strongly influenced the present-day
appearance of the Olympic Mountains (Fig. 1b). During its maximum extent at
Map of new and previously published thermochronometric ages within
the Olympic Mountains for
Within the Olympic Mountains, there is an extensive dataset of thermochronometric
cooling ages from bedrock samples (Figs. 1b and 2) for AHe (Batt et
al., 2001; Michel et al., 2018), AFT (Brandon et al., 1998), ZHe (Michel et
al., 2018), and ZFT (Brandon and Vance, 1992; Stewart and Brandon, 2004).
These thermochronometer systems record cooling through a temperature range of
In the Olympics, the youngest published reset AHe ages (
Based on thermo-kinematic modeling, Michel et al. (2018) attributed the observed AHe and ZHe age pattern to an ellipse-shaped exhumation pattern (with highest exhumation rates in the central, high-topography part of the mountain range; Fig. 2e), as predicted for a mountain range situated in an orogenic syntaxis setting (Bendick and Ehlers, 2014). Here, a bend in the subducted slab creates a mechanical stiffening, which in turn leads to rapid and focused exhumation at the surface (Bendick and Ehlers, 2014). High uplift rates in the central, high-topography part of the mountain range are also corroborated by topographic analyses (Adams and Ehlers, 2017) and denudation rates based on cosmogenic nuclides (Adams and Ehlers, 2018). Furthermore, modeling of particularly young AHe ages (< 2.5 Ma) suggests that exhumation rates increased significantly by 50 %–150 % due to Plio–Pleistocene glacial erosion (Michel et al., 2018).
Data for the ocean drill cores shown in Fig. 3.
For core ODP 888, information is taken from Westbrook et al. (1994), for ODP 1027 from Su et al. (2000), and for DSDP 174 from
Kulm et al. (1973). Sedimentation rates are calculated in this study using the
reported thicknesses and age constraints.
Data constraining the sediment thickness on the Juan de Fuca Plate before the incorporation of sediment into the accretionary wedge are summarized in Fig. 3. Three boreholes were drilled into the blanketing sediments of the Juan de Fuca Plate during deep-sea drilling projects (ODP 888, ODP 1027, and DSDP 174; Fig. 3 and Table 1) and provide estimates of the sediment thickness and age constraints. The sediment thickness at the deformation front of the subduction zone has been estimated from three seismic studies (Adam et al., 2004; Booth-Rea et al., 2008; Han et al., 2016).
Map of the Cascadia Subduction Zone showing the age of the oceanic crust (Wilson, 1993) and sediment thickness, estimated from the sediment cores of ocean drilling programs (holes ODP 888, OPD 1027, and DSDP 174) and seismic studies (Adam et al., 2004; Booth-Rea et al., 2008; Han et al., 2016). The amount of Quaternary sediment material estimated from cores is also included (Kulm et al., 1973; Su et al., 2000; Westbrook et al., 1994); more information about the drill cores is provided in Table 1. The locations of two major submarine fans (Nitinat Fan and Astoria Fan) are indicated by the dotted pattern. The Fraser and Columbia rivers are the main modern sediment sources for the Nitinat and Astoria fans, respectively. White dashed lines indicate the position of cross sections presented in this study (see also Fig. 7).
Most of the sediment is contained within two deep-sea sediment fans with different sediment sources. Today, sediment sources for the Nitinat Fan (offshore of Vancouver Island and the Olympic Mountains) include detritus from Vancouver Island, the Olympic Mountains, and material delivered by the Fraser River system (Fig. 3), which drains large parts of the Canadian Cordillera including the British Columbian Coast Mountains (Carpentier et al., 2014; Kiyokawa and Yokoyama, 2009). The Astoria Fan offshore of the Oregon coast is mostly fed by the Columbia River and is sourced by a large area in the interior of the USA (Fig. 3).
The total sediment thickness varies between 2600 and 3500 m at the deformation
front and decreases rapidly to 600 or 900 m approximately 100 km away from
the deformation front. At the locations of ODP 1027 and DSDP 174, up to
50 %–70 % of the total sediment thickness is made up of Quaternary deposits, and
sedimentation rates more than doubled during the Quaternary (from 80–110 to 250–270 m Myr
We use a multi-method approach to assess flux steady state in the Olympic Mountains. This includes thermochronometric dating, thermo-kinematic modeling of cooling ages to obtain exhumation rates, and independent estimates of accretionary influx and denudational outflux. We calculate the influx based on constraints of the incoming sediment thickness and plate convergence rate and the outflux based on spatial constraints of exhumation rates within the Olympic Mountains. The procedure for each method is outlined below.
Coordinates, elevations, and thermochronometric cooling ages for samples considered in this study.
Samples in italics are used for 1-D thermo-kinematic
modeling. Results from single-grain analyses for AHe and ZHe are reported
in Tables S1 and S2, respectively. Further details for AFT and ZFT dating
can be found in Table 3, and single-grain analyses for apatite and zircon
are reported in Tables S3 and S4, respectively. 1 SD indicates 1 standard
deviation, nd: not determined.
Our strategy with thermochronometric dating was (1) to obtain samples that
are multi-dated with up to four thermochronometer systems (because these are
particularly sensitive to reveal variations in exhumation rate) and (2) to
collect samples within vertical profiles in order to obtain estimates of the
exhumation rate at the site of the respective profile. Therefore, we dated
several literature samples with additional thermochronometer systems (Table 2) and we also present 19 new bedrock samples from vertical profiles
(Fig. 4, Table 2) and two additional bedrock samples (OP1528 and OP1556; Fig. 2,
Table 2) collected at an elevation of
Standard mineral separation techniques (sieving, magnetic and gravimetric
separation) were used to obtain apatite and zircon separates from crushed
rock samples. For AHe and ZHe dating mineral grains were handpicked and
dated in the thermochronometry lab of the University of Tübingen,
following the dating protocol of Stübner et al. (2016). The Ft correction
for apatite (Farley, 2002) and zircon (Hourigan et al., 2005) is applied to
the measured amount of helium. The (U–Th)
Results from fission-track dating.
For AFT and ZFT, 20 grains per
sample were dated in a first step and it was checked whether the sample passes the
Fission-track dating of apatite and zircon was performed using an external
detector and
To interpret cooling histories recorded by our thermochronometers as exhumation histories, we used a modified version of the thermo-kinematic model Pecube (Braun, 2003), which contains a built-in Monte Carlo approach to resolve temporal variations in exhumation histories (Adams et al., 2015; Thiede and Ehlers, 2013). The model allows for the exploration of possible exhumation histories for a particular sample by varying exhumation rates through time at defined time steps. The accuracy of a particular exhumation rate history is estimated by comparing modeled with observed cooling ages. More age constraints, and hence thermochronometer systems, lead to better-resolved modeled exhumation histories. Therefore, although we report 21 new thermochronometric ages, we only used the seven samples that have age constraints from AHe, AFT, ZHe, and ZFT in our modeling efforts (OP1513, OP1517, OP1533, OP1539, OP1551, OP1573, OP1582; Table 2).
List of parameters used for the Pecube modeling.
The thermophysical parameters chosen for the modeling are typical values
reported for the sandstones of the Olympic Mountains (Table 4). We performed
a sensitivity analysis in order to find the most suitable time step for our
simulations and the results of that analysis can be found in the
Supplement (Sect. S2). Based on the analysis, a time step
interval of 1 Myr seems to be most appropriate to use, given the range of
our thermochronometry ages and their respective uncertainties. During
further modeling, we initiated the models at 20 Ma and used the time step
interval of 1 Myr with a maximum testable exhumation rate of 6 km Myr
Results from thermo-kinematic Monte Carlo modeling for the
seven considered samples (OP1513, OP1517, OP1533, OP1539, OP1551, OP1573,
OP1582). The location of each sample within the Olympic Peninsula is shown,
together with the respective elevation (Elev). The entire range of
exhumation rates from the number of accepted model runs (
For our purpose, we focus on exploring temporal variations in exhumation rates and therefore use a 1-D model, whereby each sample is modeled independently from each other. In a 1-D model, heat transport and the movement of particles are only considered in the vertical dimension within a column of rock, ignoring topography. This mode of modeling was selected because it allowed us to efficiently perform thousands of simulations quickly in order to cover a large range of possible exhumation rates. The high number of exhumation histories accurately predicts our observed cooling ages and allows for a robust statistical assessment of the best-fitting exhumation history. Previous publications addressing exhumation histories in other orogens have also highlighted the fact that 1-D models are often sufficient to explain most of the signal recorded in thermochronometric systems (e.g., Adams et al., 2015; Thiede and Ehlers, 2013). In the Olympic Mountains, Michel et al. (2018) argued that exhumation histories for the thermochronometer systems considered here can also be explained well by vertical velocity paths. Because the spatial resolution of our seven considered samples is poor and they are all from the interior part of the mountain range (Fig. 4), we cannot further resolve the exhumation rates outside this area, making a 3-D model very difficult to validate. Therefore, we limit our interpretations to the better-resolved exhumation histories from the 1-D model and focus on the primary temporal changes, rather than paleotopography, or specific differences in the exhumation rates between samples.
Five of the seven considered samples are from the same elevation range (400–580 m), but two samples are from higher elevations (1360 and 1500 m; Fig. 5). Large differences in elevation between the samples can impact the direct comparison between them (e.g., it can affect how changes in exhumation rate are recorded from location to location). However, we are not able to correct for this circumstance (by using an age–elevation relationship) and therefore try to consider this complication when interpreting our exhumation rate histories from the different samples.
Constraints used for our quantitative accretionary influx and
denudational outflux calculations.
To assess the flux steady-state hypothesis of the Olympic Mountains, we need independent estimates of the material influx and outflux over time. For this, we focus on the time period since 14 Ma, which corresponds to the proposed establishment of flux steady state (Batt et al., 2001; Brandon et al., 1998). Flux steady state requires that the material influx into the wedge equates to the amount of accreted material removed from the subducting slab. We assessed the amount of accreted sediment (material influx) with two approaches. First, we calculated the amount of sediment incorporated into the accretionary wedge at the deformation front (Fig. 6a) during the 14 Myr period. Second, we compared this amount of “expected” accreted sediment with the observed amount of sediment residing in the accretionary wedge along two cross sections. The material outflux from the mountain range is estimated using results from thermo-kinematic modeling by equating modeled exhumation with denudation, which can then be integrated spatially and over the 14 Myr period.
Previous flux steady-state analyses in the Olympic Mountains were performed in two dimensions along a profile crossing the Olympic Peninsula. However, exhumation rates within the Olympic Mountains are known to vary spatially (Brandon et al., 1998; Michel et al., 2018). This suggests that the outflux is spatially variable, depending on the location within the mountain range. Hence, we performed our flux analysis in three dimensions and the resulting geometries are summarized in Fig. 6. The influx is calculated along the length of the deformation front, and for the calculation of the outflux we considered almost the entire area of the Olympic Peninsula.
We used a similar approach as Batt et al. (2001) to calculate the
accretionary influx, but used a three-dimensional geometry and additionally
considered temporal variations in the variables used. Assuming all sediments
resting on the subducting oceanic crust are incorporated into the
accretionary wedge, the volume of accreted sediment (
Results from influx and outflux calculations using a three-dimensional geometry.
The entire procedure for calculating the influx and outflux is described in
Sect. 3.3. The influx volumes are reported as ranges because minimum and
maximum convergence rates (Fig. 6b) have been obtained from the plate
reconstruction model of Doubrovine and Tarduno (2008).
The variable with the greatest uncertainty in this calculation is the sediment thickness back in time that has now been subducted below the Olympic Mountains. As discussed above (Sect. 2.3), the present-day sediment thickness of 2.5 km is the product of increased offshore sedimentation during the Quaternary, and the pre-Quaternary sediment thickness is difficult to determine. Following the approach described in the Supplement (Sect. S3.1), we estimated a pre-Quaternary sediment thickness of 1.5 km. In total, we calculated three different sediment volumes based on different sediment thicknesses (Table 5). Assuming a thickness of 1.5 and 2.5 km for the 14 Myr period yields a minimum and maximum value for the accreted sediment volume, respectively, representing a sediment volume unaffected by Quaternary sedimentation (1.5 km) and a volume for a likely too-high sediment thickness, using the modern thickness (2.5 km). Alternatively, we considered an increase in sediment thickness from 1.5 to 2.5 km at 2 Ma, which likely yields the geologically most meaningful volume.
The porosity of the sediment stack depends on the thickness and decreases
with increasing overburden. According to Yuan et al. (1994), the porosity at
depth
Because the dip direction of the present-day deformation front is
72
We estimated the actual volume of sediment currently residing in the accretionary wedge along two cross sections, which are approximately 50 km apart (Profile 1 and 2 in Fig. 7). The lower boundary of the accretionary wedge is the top of the subducting oceanic plate, which is constrained from the Slab 1.0 model (Hayes et al., 2012; McCrory et al., 2012). The upper boundary is defined by the present-day topography and bathymetry (from 10 and 500 m resolution digital elevation models, respectively) and the Hurricane Ridge Fault (HRF). At the surface, the location of the HRF is adopted from a geologic map (Tabor and Cady, 1978) and below the surface we use information provided by a seismic study at depths of 22 and 34 km (Calvert et al., 2011). The uncertainty related to the position of the HRF (error bars at HRF nodes in Fig. 7) was propagated to estimate an uncertainty for the calculated sediment volumes. A further explanation of this approach is given in the Supplement (Sect. S3.2). Because the location of the HRF is not resolved at greater depths, we truncate the area considered for volume calculation at 34 km of depth. Finally, the calculated volume is corrected for the porosity of the sediment stack. Davis and Hyndman (1989) use porosities of 4 %–10 % for sediments contained within the accretionary wedge offshore of Vancouver Island. Hence, we use an average porosity of 6 % in our correction.
Sediment volumes calculated along two cross sections spanning the Olympic Peninsula (Profile 1 and 2, the vertical exaggeration is 2; see inset for location). For an explanation of the procedure used, see the text. The reported uncertainties for the volume are based on the uncertainties in the position of the Hurricane Ridge Fault (indicated with error bars at the respective symbol). Numbers in the inset correspond to the following: (1) position of Profile 1, (2) position of Profile 2, and (3) position of profile by Davis and Hyndman (1989) referred to in the text.
In the absence of extensional faults, denudation acts as the prime mechanism for exhumation in the Olympic Mountains. Therefore, exhumation can be equated with denudation and the denudational outflux from the range can be obtained from the spatial and temporal integration of exhumation rates.
The exhumation histories presented in this paper (Fig. 5) are well-suited to resolve temporal variations in exhumation and hence provide qualitative information about variations in the denudational outflux. The low spatial density of the seven considered samples prohibits a quantitative assessment of the denudational outflux. To overcome this problem, we reverted to the pattern and exhumation rates suggested by Michel et al. (2018), providing good spatial coverage of almost the entire Olympic Peninsula (Fig. 6d). The total amount of exhumation, which is used for calculating the outflux and corresponds to the temporal integration of the exhumation rates, is similar within uncertainty in both datasets. For example, the modeled exhumation rate is sufficient to explain the ZHe age of 10.2 Ma for sample OP1513 in both studies (Michel et al., 2018, and this study).
Our outflux calculations are based on the spatial integration of the entire exhumation rate pattern displayed in Fig. 6d, which is then temporally integrated over the 14 Myr period. Additional to a constant exhumation scenario, we also considered an increase in exhumation rates, which is related to an increase in erosion due to Plio–Pleistocene glaciation of the Olympic Mountains (Michel et al., 2018). In Table 5, we report the denuded volumes for the case of constant exhumation rates and for the two possible increase scenarios suggested by Michel et al. (2018), equating to a 50 % increase in rates occurring at 3 Ma or a 150 % increase in rates occurring at 2 Ma. In order to account for the porosity of the denuded rocks, we corrected the denuded volumes by a porosity of 6 %, the same value we applied in the estimation of the volumes in the sedimentary cross sections
Along the Mt. Olympus elevation transect (Fig. 4b), AHe ages (1.9–3.7 Ma)
overlap each other within sample error (except for the uppermost
sample). ZHe ages (4.8–8.5 Ma) show a similar behavior (with the exception
of the lowermost sample; Fig. 4b). AFT ages for two samples are 5.1 and
6.2 Ma, and the obtained ZFT ages of this transect are all unreset. Within
the Mt. Anderson transect (Fig. 4c), AHe ages (1.5–3.9 Ma) increase with
elevation up to 1400 m and decrease between 1400 and 2100 m.
ZHe ages vary between 6.5 and 8.9 Ma and one sample at
Clear spatial patterns for the multi-dated thermochronometer samples are observable (compare Figs. 2 and 4). AHe ages are reset (apart from one sample in the northeast of the mountain range) and decrease towards the center of the mountain range, where very young ages (< 2.5 Ma) can be found. Seven fully reset AFT samples (5.0–7.8 Ma) are confined to the center of the range (samples OP1513, OP1517, OP1533, OP1539, OP1551, OP1573, OP1582), overlapping the area of reset ZHe samples. The remaining eight AFT samples are unreset (Table 3 and Fig. 4). Two samples at the north and east coast (OP1502 and OP1510) have the youngest age peaks at 26 Ma (comprising 29 % of the dates) and 36 Ma (35 %), respectively. Samples from the western part of the mountain range (OP1521, OP1522, OP1527, OP1528, OP1531) have younger age peaks of 5–16 Ma (comprising 20 %–76 % of the dates). Furthermore, the youngest age peak of these samples decreases in age towards the area of fully reset AFT samples.
We also collected samples (OP1527 and OP1528) close to locations with
the youngest AFT ages of Brandon et al. (1998), which were reported as
incompletely reset samples (with youngest peak ages of 3.9 and 2.3 Ma). In
the original publication, only a small number of grains were dated (
ZHe ages constrain an area of reset ages (4.8–10.2 Ma) in the central, high-topography portion of the mountain range (light grey shaded area in Fig. 4a). Five of these samples have AFT (5.1–7.8 Ma) and ZHe (4.8–8.9 Ma) ages that overlap within sample errors, implying rapid cooling (and hence fast exhumation) through both systems' closure isotherms. AHe ages of these samples are younger (1.7–3.9 Ma) and do not overlap AFT ages, indicating that exhumation rates decreased after cooling below the AFT closure isotherm.
Of the seven samples dated with the ZFT method, only sample OP1539 has a fully reset age (12.6 Ma). Together with data from Brandon and Vance (1992) and Stewart and Brandon (2004), this confines reset ZFT samples to a very small area east-southeast of Mt. Olympus, encompassing the headwaters of the Elwha and Quinault rivers (area outlined with a red dashed line in Fig. 4a).
Between 13 000 and 17 800 simulations provide a good fit to the data for each
of the seven samples used in the thermo-kinematic modeling (Fig. 5). As
expected, the four samples (OP1533, OP1539, OP1551, OP1582; Fig. 5) with
overlapping AFT and ZHe ages require fast exhumation rates of > 3 km Myr
The calculated volumes of the accretionary influx depend strongly on the
incoming sediment thickness (Table 5). With our three-dimensional
geometry (Fig. 6a) volumes vary among
In the following, the implications of the above described observations will be discussed in order to assess the flux steady-state balance between accretionary influx and denudational outflux within the Olympic Mountains. To do that, it is pivotal to have an understanding of both temporal and spatial variations in exhumation of the Olympic Mountains. First, we elaborate on results from thermochronometric dating, including the applicability of age–elevation relationships to reconstruct exhumation rates in the Olympic Mountains (Sect. 5.1). Second, we analyze the general pattern of exhumation based on the spatial distribution of cooling ages (Sect. 5.2). Third, we link thermochronometric cooling ages with thermo-kinematic modeling, which reveals the temporal evolution of exhumation rates (Sect. 5.3). Fourth, we discuss the outcome of our qualitative and quantitative assessment of flux steady state in the Olympic Mountains (Sect. 5.4). Finally, in Sect. 5.5, we elaborate on the limitations of the different approaches.
The cooling ages of samples collected from a quasi-vertical elevation profile (e.g., Fitzgerald et al., 1993; Reiners et al., 2003) can be analyzed by looking at the age–elevation relationship. Often, the purpose is to determine an apparent exhumation rate by fitting a line through the data points when ages are positively correlated with elevation. However, the prerequisite for this approach is that, over the lateral extent of the sampled transect, there is no significant gradient in exhumation rates. This is not necessarily given in the Olympic Mountains (Michel et al., 2018; see also Fig. 2e) and the new data represent this complication (Fig. 4b–d).
At Mt. Olympus, the AHe and ZHe age–elevation relationships do show a
positive correlation, suggesting fast exhumation rates of
A well-constrained spatial pattern of exhumation is needed for calculating
the denudational outflux. Looking at the spatial distribution of
thermochronometric cooling ages provides qualitative information about the
pattern of exhumation. In general, the distribution of thermochronometric
ages indicates that in the Olympic Mountains the magnitude of exhumation
increases from the coast to the center. As discussed above, areas belonging
to the Coast Range Terrane (close to the coast or the Blue Mountain area,
where unreset AHe ages can be found; Fig. 2a) correspond to the structurally
highest parts within the range (Fig. 1c) and were not sufficiently reheated
to reset the AHe system. Assuming a geothermal gradient typical for the
Cascadia Subduction Zone of
The aerial exposure of the accretionary wedge (the Olympic Structural Complex; Fig. 1c) records exhumation from greater depths. Here, all samples yield reset AHe ages, requiring a minimum exhumation depth of 2–3 km. In the center of the mountain range (encompassing the headwaters of the Hoh, Queets, Quinault, and Elwha rivers; Fig. 1b) the area of reset AFT ages approximately overlaps the area of reset ZHe ages (Fig. 4a), requiring deeper exhumation compared to the coastal part of the Olympic Structural Complex.
The area east-southeast of Mt. Olympus (corresponding to the area of reset
ZFT samples; Fig. 4a) has been exhumed from the greatest depths within the
Olympic Mountains. For an average ZFT closure temperature of
In summary, the central, high-topography part of the mountain range corresponds to the most deeply exhumed part. This corroborates the exhumation rate pattern (Fig. 2e) suggested by Michel et al. (2018), the pattern of denudation rates based on cosmogenic nuclide dating (Adams and Ehlers, 2018), and results from topographic analysis (Adams and Ehlers, 2017), which all suggest that most of the exhumation–denudation occurs at this location. Hence, we use this pattern for the calculation of the denudational outflux.
Summary of volcanic activity, tectonic and climatic events, and convergence rate and angle at the Cascadia Subduction Zone in comparison with our interpreted exhumation rates for the past 25 Myr. Exhumation rates are limited to the time interval covered by our thermochronometric ages (0–11 Ma). The curve depicts the interpreted evolution of exhumation rates based on the modeling results shown in Fig. 5 (see text for details). Volcanic activity after du Bray and John (2011). Tectonic and climatic events are (1) the start of exhumation of the Olympic Mountains (Brandon et al., 1998), (2) the onset of uplift of the Oregon Coast Range (McNeill et al., 2000), (3) rotation in the stress field (Priest, 1990), (4) faster uplift in the Oregon Coast Range (McNeill et al., 2000), (5) Pacific-wide plate reorganization (Wilson, 2002), (6) the onset of North American glaciation (Haug et al., 2005), (7) the onset of glaciation within the Olympic Mountains (Easterbrook, 1986), and (8) a change in the deformational style of the offshore accretionary wedge (Flueh et al., 1998). Convergence rate and angle from Doubrovine and Tarduno (2008).
Our new thermo-kinematic modeling revealed temporal variations in exhumation
rates in the Olympic Mountains (Fig. 5). The decrease in exhumation rates at
5–7 Ma can be readily explained by the reduction in plate convergence rate
and the change in convergence direction (Fig. 8). A Pacific-wide
reorganization of plate movement at 5.9 Ma has been suggested (Wilson, 2002),
and rapid uplift of the Oregon Coast Range at 6–7.5 Ma with a subsequent
cessation in uplift has also been attributed to variations in the plate
subduction parameters (McNeill et al., 2000). Furthermore, the volcanic
record of the Cascadia Subduction Zone shows temporal variations, and the
strongest volcanic activity lasted from 25 until 18 Ma (du Bray and John,
2011). A period of volcanic quiescence, lasting from 17 until 8 Ma, was
then followed by increased activity starting at
In contrast, the increase in exhumation rates at
Taken together, these observations indicate that temporal variations in exhumation rates within the Olympic Mountains are subject to changes in both the tectonic and climatic conditions (as summarized in Fig. 8). The implication of these variations should be considered for flux steady-state assessment.
Several variables that affect both the accretionary influx and the
denudational outflux show temporal variations. Exhumation rates decrease at
5–7 Ma and increase at
It follows that, qualitatively, both influx and outflux vary through time and are heavily influenced by the Plio–Pleistocene glaciation, which increased denudation rates and offshore sedimentation rates. However, we cannot quantitatively constrain whether variations in the influx and outflux on these short timescales (2–3 Myr) balance each other (and the system would still be in a flux steady state). Interestingly, measured denudation rates based on cosmogenic nuclide dating (temporally integrating over the Holocene) suggest that modern denudation rates have not been significantly influenced by Plio–Pleistocene glaciation, but are mostly driven by tectonic rock uplift (Adams and Ehlers, 2018). The Holocene accretionary influx, however, is still affected by the increased sediment thickness since the onset of glaciation. Hence, the current accretionary influx seems to exceed the denudational outflux in the Olympic Mountains.
Here, we discuss the quantitative assessment of influx and outflux for the
last 14 Myr (Table 5), the time since the Olympic Mountains are presumed
to be in flux steady state (Batt et al., 2001; Brandon et al., 1998). In our
geometry (Fig. 6a and d), we calculate the accretionary influx over a
distance along the deformation front and the spatial exhumation rate pattern
is integrated to infer the denudational outflux (Fig. 6d). Assuming an
increase in sediment thickness at 2 Ma yields an accretionary volume
(
Sediment volumes integrated along the cross sections (Fig. 7) also provide an
interesting perspective on the accretionary influx in the Olympic Mountains.
These volumes are not directly comparable with the influx–outflux volumes
discussed above (calculated from 14–0 Ma) because the sediment contained
within the cross sections (Fig. 7) records accretion since the
In summary, the assessment of flux steady state in the Olympic Mountains is nontrivial and several scenarios are possible. From a qualitative viewpoint, flux steady state is probably not achieved on short timescales (a few Myr) because the thickness of incoming sediment, plate subduction velocity, and exhumation rates show strong temporal variations on timescales of 2–3 Myr. From a quantitative viewpoint, influx and outflux volumes equate to each other over longer timescales (i.e., 14 Myr) if influx and outflux are considered in three dimensions.
In the sections above, we discussed exhumation in the Olympic Mountains and the results from our flux calculations. In the following section, we want to elaborate on possible restrictions or limitations in our approaches.
With our 1-D modeling, we revealed strong temporal variations in exhumation
rates (Fig. 5) related to variations in both tectonic and climatic
conditions (Fig. 8). However, two of our modeled samples (OP1513 and
OP1517) do not display the decrease in exhumation rates at
Regarding our flux analysis, we based our calculations on the volume of
accreted sediment within a certain time (governed by the sediment thickness
and the plate convergence rate) and the amount of denuded material (governed
by the exhumation rates). As we mentioned in Sect. 3.3.1, a variable with
great uncertainty is the sediment thickness over time, which has now been
subducted below the Olympic Mountains. In the Supplement
(Sect. S3.1) we outline our approach for assessing the pre-Quaternary
sediment thickness, which is used in our calculations. Although the reported
1.5 km sediment thickness seems to be a plausible value, we note that this
value is afflicted with uncertainties and might have been higher.
Nonetheless, our proposed balance between influx and outflux is still
tenable if the pre-Quaternary sediment thickness deviated from the assumed
1.5 km. In other words, we suggested an influx volume of 75–
During our influx calculations, we did not distinguish between different modes of accretion, such as frontal accretion or underplating. Batt et al. (2001) concluded that most accretion occurs at the front of the wedge. However, a recent seismic study showed that sedimentary underplating is taking place below the Olympic Mountains (Calvert et al., 2011). For our approach, the mechanism of accretion does not matter because we are only interested in whether mass is balanced over the entire wedge and not at a specific point. As indicated, this is a limitation of our approach and might lead to an overestimation of the actual influx volume because we do not account for the amount of sediment transported towards the mantle.
Flux steady state implies that the outflux from and influx into a mountain range balance each other. An inherent assumption is often that the material removed from a mountain range (the outflux) again enters the mountain range via the influx, which consists of the denuded material from the same source. So in the case of an accretionary wedge, this implies that sediment is recycled and the system behaves as a closed system. As we described in Sect. 2.3 of the paper, the sediment currently entering the accretionary wedge of the Cascadia Subduction Zone is a mixture of sediment from different source regions (e.g., Olympic Mountains, Vancouver Island, Canadian Cordillera, and in the case of the Astoria fan, the interior USA; Fig. 3). With the increased detrital input from the Cordilleran Ice Sheet from outside the Olympic Mountains, this effect became particularly pronounced since the onset of Plio–Pleistocene glaciation. Hence, our influx–outflux calculations for the Olympic Mountains do not represent a closed system, and the influx into the Olympic Mountains is solely controlled by the outflux out of the system. However, our calculations indicate that on long timescales (i.e., over 14 Myr) flux steady state is attained, which might seem surprising given that the sediment thickness is governed by contributions from different source regions. We suspect that processes during sediment deposition, like redistribution by turbidity currents and redeposition in more proximal parts of the Juan de Fuca Plate, play an important role in the final sediment budget. As a consequence, the amount of sediment denuded from the Olympic Peninsula in a given time period (the outflux) is dispersed as it enters the ocean so that for the same time period only a fraction of the sediment thickness (governing the influx) is composed of material originating from the Olympic Peninsula.
Variations in the geometry or extent of the accretionary wedge were also not
included in our flux analysis. Since the onset of subduction at the Cascadia
Subduction Zone with the present geometry at
As we pointed out in Sect. 5.4.2, flux steady state is obtained by using a three-dimensional geometry. However, we only considered the deformation front perpendicular velocity component for our influx calculations. The different sediment volumes contained in the reported cross sections (Fig. 7) could indicate that on long timescales additional velocity components must be considered. We can only speculate that margin-parallel transport, which is a contentious topic at the Cascadia Subduction Zone (e.g., Batt et al., 2001; McCrory, 1996; Wang, 1996), also contributes to the accretionary influx. Present-day GPS velocities corroborate this hypothesis, indicating northward movement of coastal areas south of the Olympic Mountains (e.g., McCaffrey et al., 2013; Wells and McCaffrey, 2013).
To summarize, several parameters, like the location of faults within the orogenic wedge, the sediment source region, the temporal evolution of the wedge geometry, and margin-parallel transport, are difficult to constrain from current observations. Although we emphasized that not all of these parameters affect our flux analysis, further knowledge of these will refine the current understanding of steady state in the Olympic Mountains.
Our new dataset of multi-dated thermochronometer bedrock samples together with thermo-kinematic modeling suggests that several mechanisms contribute to the evolution of the Olympic Mountains. Modeling of the observed AHe, AFT, ZHe, and ZFT ages shows that variations in both tectonic and climatic conditions result in temporal variations in exhumation rates. We revealed a hitherto unnoticed response of exhumation to the tectonic signal (a reduction in plate convergence rate causing a drop in exhumation rates), which can also be observed in other parts of the Cascadia Subduction Zone. Plio–Pleistocene glaciation of the Olympic Mountains led to increased denudation, resulting in increased exhumation rates.
Our approach of assessing flux steady state in the Olympic Mountains by estimating the material influx and outflux independently from each other is promising, but yields ambiguous results. The observed temporal variations in exhumation rate require a variation in the denudational outflux. Likewise, the accretionary influx is also temporally variable because the plate subduction velocity and incoming sediment thickness are variable through time. Qualitatively, this suggests that flux steady state is perturbed on short timescales by variations in the tectonic or climatic conditions. Our quantitative calculations of the influx and outflux show that flux steady state may be achievable over long timescales (i.e., 14 Myr). Contrary to a previous flux steady-state analysis in the Olympic Mountains, our calculated influx and outflux volumes only balance each other if a three-dimensional geometry is considered.
This study demonstrates the timescale (10
All thermochronometry data used in this article are freely available and are either presented in the main paper or can be found in the Supplement. The exhumation histories derived from the modeling are displayed in Fig. 5, and the underlying data and scripts used to produce the figure are available from the authors.
In Sect. 3.3 we performed our flux analysis in three dimensions due to the
spatially variable exhumation rates (Fig. 6d). In the following, we also
calculate the influx and outflux using a two-dimensional geometry so that
our calculations can be compared to those from Batt et al. (2001). Here, the
accretionary influx occurs at a single location at the deformation front,
and the sediment volume (
The outflux calculations are not based on the integration of the entire
exhumation pattern in Fig. 6d, but rates are only integrated along the
white line in Fig. 6. These integrals yield values of 68 km
A comparison of two-dimensional influx and outflux shows that the
accretionary influx (
The supplement related to this article is available online at:
LM, BA, and TE performed the fieldwork. Thermochronometric dating was done by CG, LM, and SF. LM performed the flux calculations with contributions from CG. The thermo-kinematic modeling was done by LM with help from BA and CG. LM drafted the initial versions of the paper and figures, and all authors commented on and contributed to the final version of the article.
The authors declare that they have no conflict of interest.
This work was funded by a European Research Council (ERC) Consolidator Grant (615703) to Todd Ehlers. During fieldwork, we had invaluable help and assistance from Holger Sprengel, William Baccus, Jerry Freilich, Roger Hofmann, and the Olympic National Park rangers. We acknowledge Matthias Nettesheim for sharing the code used for the evaluation of the tectonic plate reconstruction model and the help of Willi Kappler during Pecube modeling. We thank Associate Editor David Lundbek Egholm for editorial handling of the paper. The comments by Phillipe Steer and one anonymous referee helped to improve and clarify this paper. Edited by: David Lundbek Egholm Reviewed by: Philippe Steer and one anonymous referee