RESEARCH ARTICLE
Combined seismic and borehole investigation of the deep
granite weathering structure—Santa Gracia Reserve case in
Chile
Rahmantara Trichandi
1,2
| Klaus Bauer
1
| Trond Ryberg
1
|
Dirk Scherler
1,3
| Klaus Bataille
4
| Charlotte M. Krawczyk
1,2
1
GFZ German Research Centre for
Geosciences, Potsdam, Germany
2
Technische Universität Berlin, Berlin,
Germany
3
Freie Universität Berlin, Berlin, Germany
4
University of Concepci
on, Concepci
on, Chile
Correspondence
Rahmantara Trichandi, GFZ German Research
Centre for Geosciences, Telegrafenberg,
14473 Potsdam, Germany.
Email: [email protected]e
Funding information
German Science Foundation (DFG) priority
research programme SPP-1803 ‘EarthShape:
Earth Surface Shaping by Biota’, Grant/Award
Number: KR 2073/5-1
Abstract
Imaging the critical zone at depth, where intact bedrock transforms into regolith, is
critical in understanding the interaction between geological and biological processes.
We acquired a 500 m-long near-surface seismic profile to investigate the weathering
structure in the Santa Gracia National Reserve, Chile, which is located in a granitic
environment in an arid climate. Data processing comprised the combination of two
seismic approaches: (1) body wave tomography and (2) multichannel analysis of sur-
face wave (MASW) with Bayesian inversion. This allowed us to derive P-wave and S-
wave velocity models down to 90 and 70 m depth, respectively. By calibrating the
seismic results with those from an 87 m-deep borehole that is crossed by the profile.
We identified the boundaries of saprolite, weathered bedrock, and bedrock. These
divisions are indicated in the seismic velocity variations and refer to weathering
effects at depth. The thereby determined weathering front in the borehole location
can be traced down to 30 m depth. The modelled lateral extent of the weathering
front, however, cannot be described by an established weathering front model. The
discrepancies suggest a more complex interaction between different aspects such as
precipitation and topography in controlling the weathering front depth.
KEYWORDS
borehole, geophysics, Bayesian inversion, body wave tomography, critical zone, geomorphology,
geophysics, Rayleigh wave, regolith, seismic survey, surface wave tomography, velocity
gradient, weathering front
1|INTRODUCTION
The interaction between geological, chemical, and biological processes
plays an important role in forming the Earth’s surface. For example,
tectonic uplift exposes bedrock at the surface and allows water- and
biota-assisted weathering processes to dissolve rock or transform its
structure and composition (Brantley et al., 2007). After disintegration,
erosion processes can then remove sediment and expose fresh bed-
rock from depth. The material that exists, and the collection of geolog-
ical, chemical, and biological interactions that occur, at the surface and
in the shallow subsurface, make up the so-called critical zone
(Figure 1). The vertical extent to which the bedrock is actively
transforming into weathered bedrock is called the weathering front.
Previous studies of weathering in different climates observed the
weathering front at different depths (Bazilevskaya et al., 2013;
Brantley et al., 2017; Hayes et al., 2020; Stierman & Healy, 1984;
Vázquez et al., 2016).
Different weathering processes create porosity and permeability,
which itself influences water flow through the rock and thus further
influences the dissolution of minerals from the bedrock (Graham
et al., 2010). To fully understand how the different processes interact,
it is crucial to know the critical zone structure, including the thickness
of the regolith layer. To investigate the critical zone structure, direct
measurement approaches are often used, based on artificial outcrops,
Received: 5 August 2021 Revised: 8 June 2022 Accepted: 18 July 2022
DOI: 10.1002/esp.5457
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.
© 2022 The Authors. Earth Surface Processes and Landforms published by John Wiley & Sons Ltd.
3302 Earth Surf. Process. Landforms. 2022;47:3302–3316.
wileyonlinelibrary.com/journal/esp
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
soil pits, or borehole cores. While such approaches may yield highly
accurate 1D data of the critical zone structure, they commonly pro-
vide only limited extent and depth penetration. In addition, while
borehole coring and logging can reach down to greater depth, it is typ-
ically relatively expensive. Therefore, 2D measurements are needed
to spatially sample the heterogeneity across a region (Callahan
et al., 2020).
Geophysical approaches such as the seismic method are often
used to provide 2D information on the subsurface without altering
the subsurface environment. Geophysical methods offer an indirect
investigation of rock properties in the subsurface by performing mea-
surements on the surface. Information on rock properties is generally
derived from P- and S-wave velocity data, as they are affected by the
rock density, porosity, and mineralogy (Befus et al., 2011; Holbrook
et al., 2019). In the critical zone, weathered rocks will show signifi-
cantly lower P- and S-wave velocities values compared to the
unweathered bedrock at depth, due to higher porosity, lower density,
or the transformation of minerals during the weathering processes.
Different types of rock layers can be identified based on the dif-
ferent P- and S-waves velocities. The P-wave velocity model is com-
monly produced using body wave tomography, in which arrival times
of P-waves are identified to infer the rock’s P-wave velocity
(e.g. Baumann-Wilke et al., 2012; Befus et al., 2011; Flinchum
et al., 2018). As it travels faster than any other waves, the P-wave
arrivals can easily be determined in the seismogram as the first
impulse recorded. However, this is not the case with S-wave arrivals,
as other wave types often superimpose the incoming S-waves. This
superimposition creates a problem in producing a shear wave velocity
model, especially for shallow targets. Therefore, shear wave velocity
models are increasingly built from surface wave data from active seis-
mic sources (Comina et al., 2017; Ivanov et al., 2006; Miller
et al., 1999; Park et al., 2005). Previous studies have also shown that
producing a shear wave velocity model from surface wave inversion
can help to image and identify weathering zones, by their low veloci-
ties (Keifer et al., 2019; Wang et al., 2019; Yaede et al., 2015).
The Chilean Santa Gracia Reserve is one of the four focus sites in
the EarthShape research priority programme (Dal Bo et al., 2019;
Oeser & Von Blanckenburg, 2020; Oeser et al., 2018) that studies the
effects of different weathering processes in different climate and
environments. The reserve is located in a granodioritic environment
with little precipitation and low vegetation cover. In this setting, we
expect a relatively shallow regolith depth due to the limited precipita-
tion (Braun et al., 2016; Hayes et al., 2020; Vázquez et al., 2016). Pre-
vious geophysical investigations using ground penetrating radar (GPR)
could not fully image the critical zone structure as the penetration
down to 3 m depth did not reach the bedrock (Dal Bo et al., 2019).
We thus aim to image the critical zone in Santa Gracia using body
wave tomography and multichannel analysis of surface waves
(MASW) methods. We use body wave tomography to create a P-wave
velocity model, and MASW, combined with transdimensional Bayesian
inversion, to produce an S-wave velocity model. Integrating the P-
and S-wave velocity models with existing borehole information
enables us to create a conceptual model of the weathering structure
in Santa Gracia Reserve. We discuss our results in conjunction with
existing borehole logging (Weckmann et al., 2020) and geochemical
analysis (Krone et al., 2021) to present an integrated interpretation of
the weathering structure. Finally, we compare our results with existing
established models of weathering front advances.
2|STUDY SITE AND GEOLOGIC SETTING
The Santa Gracia Reserve is located northeast of La Serena, Chile
(Figure 2a). It is located in a transitional area between the arid and
semi-arid climatic zones. The mean annual precipitation is 77 mm/
year, and the mean annual temperature is 15C (Karger et al., 2017).
The semi-arid climate in this area creates a relatively low vegetation
coverage of 30–40% (Bernhard et al., 2018; Oeser & Von
Blanckenburg, 2020), dominated by cacti and shrubs sustained by fog.
Regional geological data indicates coastal uplift rates of
0.2 0.1 mm/year (Kukowski & Oncken, 2006). The regional geologi-
cal map shows that the reserve is located in an intrusion (130–
110 Ma) with a younger intrusion (100–97 Ma) east of the study area
(modified from Gobierno de Chile Servicio Nacional de Geologia y
Mineria, 2003).
Geologically, the study area is composed of tonalities, diorites,
monzodiorites, granodiorites, and monzongranites (Gobierno de Chile
Servicio Nacional de Geologia y Mineria, 2003). Some hydrothermal
alteration can also be found west of the study area (Figure 2a). Soil pit
information from previous studies found that clastic rocks with a
2 mm grain size characterize the upper 0.3 m of the subsurface, while
below, more fractured blocks with 5–20 cm size prevail (Bernhard
et al., 2018; Oeser et al., 2018). Previous geophysical investigation
using GPR imaged the saprolite layer down to 3 m depth (Dal Bo
et al., 2019). In Figure 2b, we present a shaded topography map with
the seismic profile (red line) and borehole location plotted. The study
area is located in a relatively flat area, as shown in the seismic profile
topography in Figure 2d, with topographic variation of 15 m
at most.
FIGURE 1 Conceptual illustration of the critical zone structure in
an eroding landscape. Soil, saprolite, and weathered bedrock comprise
the regolith, which rests on top of the unweathered bedrock. Erosion
at the surface allows the weathering front to advance to greater
depth with time
TRICHANDI ET AL.3303
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
As part of the EarthShape research programme, a borehole was
drilled in 2019, which is located along the seismic profile. Core analy-
sis suggests different zones in the subsurface down to 87 m depth
(Krone et al., 2021). Geophysical logging provides P- and S-wave sonic
logs and borehole televiewer data (Weckmann et al., 2020). Unfortu-
nately, the P- and S-wave sonic logs provided data from 17 to 87.5 m
depth with the upper 17 m of the logging data missing, due to acquisi-
tion problems. This part of the borehole is crucial as it covers the
depth range where we expected to have the strongest weathering.
2.1 |Seismic data acquisition
In January 2020, active seismic acquisition was conducted in the
Santa Gracia National Reserve, Chile. A 500 m-long seismic profile
was surveyed along a dirt road to image the deep weathering struc-
ture around the borehole location (Figure 2b). The elevation varies
between 610 and 625 m along the line (Figure 2d). The acquisition
parameters were designed to allow for the application of body wave
tomography and surface wave analysis. To provide sufficient energy,
seismic waves were generated using a 40 kg accelerated weight drop
source (Figure 2c). Altogether 135 shot locations were distributed
along the profile with shot spacings of 4 m. The shots were recorded
by a constant receiver spread of 90 channels deployed with 6 m spac-
ings along the entire line. For each receiver point, we plant a three-
component geophone with a 4.5–150 Hz response which is con-
nected to an autonomous CUBE data logger. The connected CUBEs
then saved all the data transmitted from the geophone, which will
then be individually extracted from each CUBE. The source and
receiver locations along the line were projected onto a straight line
using the approach of Zelt (1999). Source–receiver offsets remain
unchanged in this procedure. Maximum source–receiver offsets vary
between 250 and 500 m.
3|METHODOLOGY
3.1 |Body wave tomography
Seismic tomography using body waves is a well-established method
for imaging near-surface weathering (e.g. Befus et al., 2011; Flinchum
et al., 2018; Holbrook et al., 2014; Leone et al., 2020). Similar to the
latter mentioned authors, we use first-arrival P-wave travel time
tomography to determine the P-wave velocity structure. Noteworthy
is that no clear S-wave arrivals could be identified in the weight drop-
generated data to run a complementary S-wave travel time tomogra-
phy since the weight drop source is excited mainly by P-waves and
surface waves. For this reason, instead, the surface waves were used
to image S-wave velocities.
3.2 |Inversion of travel time data
First-arrival travel times were first picked in a receiver gather seismo-
gram, where for a given receiver, the trace recordings from all shots
are plotted over the source–receiver offset. An example is shown in
Figure 3a. Receiver gathers were chosen instead of shot gathers
because of the slightly denser trace spacing (4 m in receiver gathers
compared to 6 m in shot gathers), enhancing the continuity of the
first-arrival phase. Bandpass filtering (Butterworth, 20–100 Hz) was
applied to improve the quality of signals at larger distances. Approxi-
mately 8500 travel times were picked in all available receiver gathers.
FIGURE 2 Overview of the study area: (a) regional geological map showing the dominance of granitic rock in the study area (after Gobierno
de Chile Servicio Nacional de Geologia y Mineria, 2003); (b) shaded topography map with the locations of the seismic profile (red line) and
borehole (blue dot); (c) photo of the seismic survey showing the seismic weight drop source used; and (d) topography of the seismic profile also
showing the borehole’s location (blue dot) [Color figure can be viewed at wileyonlinelibrary.com]
3304 TRICHANDI ET AL.
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Additional quality control was undertaken by inspecting the picked
travel times within the alternative shot-gathered presentation of the
data and picks. Based on the comparison of reverse observations, we
estimate travel time uncertainties of about 2.5 ms. A starting model
for the tomographic inversion was developed using the RAYINVR
package (Zelt & Smith, 1992). As a result, a 2D layered velocity struc-
ture was established. The root mean square (RMS) misfit between
observed and calculated travel times was reduced during the
RAYINVR inverse modelling down to about 10 ms.
The subsequent tomographic inversion was carried out using the
SIMUL2000 package (Thurber & Eberhart-Phillips, 1999). This itera-
tive method is based on ray tracing and a damped least-squares inver-
sion algorithm. Although SIMUL2000 was initially developed for
crustal studies using mainly local earthquake data, it can also be
applied to controlled-source data sets and near-surface investigations
(e.g. Baumann-Wilke et al., 2012). The grid-based tomographic
starting model was generated from the 2D layered RAYINVR model
using horizontal and vertical node spacings of 10 and 3 m, respec-
tively. Then, iterative inversion was carried out until the RMS misfit
reached values on the order of magnitude of the estimated picking
uncertainties. This approach is essential to avoid over-fitting of
erroneous data and related potential introduction of artifacts. An
example of individual data fitting after the tomographic inversion is
shown in Figure 3c. A generally good agreement between observed
(Figure 3c) and calculated (Figure 3d) travel times for all picked
source–receiver pairs indicates that the tomographic model can
explain many details of the measured data. The final model has an
RMS misfit of 2.5 ms.
3.3 |Multichannel analysis of surface waves
MASW is commonly used to model near-surface S-wave velocity
(Konstantaki et al., 2015; Olona et al., 2010; Park et al., 1999; Yaede
et al., 2015). Using the MASW method, we extracted the frequency-
dependent phase velocity dispersion curves. The dispersion curves
were then inverted using a Bayesian approach to produce a 1D S-
wave velocity model for observation points along the profile. With
multiple 1D S-wave velocity models, we can build a pseudo-2D S-
wave velocity profile.
For the seismic data acquired with a weight drop source, we
expect the dominant surface wave to be the Rayleigh wave composed
FIGURE 3 Tomographic inversion
approach: (a) exemplary seismogram
section showing receiver-gathered traces
for all shots recorded by a given receiver,
red dots indicate the picked travel times
of the first-arrival P-wave; (b) observed
picks with error bars and calculated travel
times (black line) for one receiver gather,
grey lines indicate travel times for all
other receiver locations; (c) observed
travel times for all picks as a function of
source and receiver location;
(d) calculated travel times using the final
tomographic model [Color figure can be
viewed at wileyonlinelibrary.com]
TRICHANDI ET AL.3305
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
of compressional and vertical particle motion along the surface.
Therefore, these specific particle motions are best observed in the
vertical component of the seismic data. For this reason, we processed
the vertical component of the measured seismic data.
3.4 |Extraction of dispersion curves
To extract the Rayleigh wave’s phase velocity dispersion curves, we
follow the surface-wave dispersion inversion and profiling (SWIP)
described by Pasquet and Bodet (2017). Figure 4a shows an example
of how we window the seismic data for dispersion stacking for the
observation point (X
mid
) located at 205 m in our profile. The steps are
as follows:
1. Selection of traces. We first select the seismic data trace to be
included in our dispersion analysis by creating a receiver window
(L) around the mid-point (X
mid
). In this example, we used a symmet-
rical receiver window of 50 m, spanning from X
mid
–L/2 to X
mid
+-
L/2. Every trace from the receiver located inside this window will
be included for the dispersion analysis of the corresponding X
mid
.
In Figure 4a, the included receivers are represented by blue
inverted triangles.
2. Selection of shots. We then select the shots (FFID) to be included in
the dispersion analysis by determining a shot window (dS). For
each end of our receiver window, we included shots located inside
the range of X
mid
–L/2 to X
mid
–L/2 –dS and from X
mid
+L/2 to
X
mid
+L/2 +dS, respectively. For the example in Figure 4a,we
used a shot window of 30 m, and the red squares represent the
included shots.
3. Extraction of shot gathers. For the next step, we extracted the
selected shot gather using the receivers and shots included in
steps 1 and 2. An example of a shot gather from X
mid
=205 m and
FFID =254 is shown in Figure 4b. For this specific example, we
then collected all 11 shot gathers.
4. Frequency–phase velocity domain transformation. For each
extracted shot gather, we then transformed it to the equivalent
frequency–phase velocity domain to produce a dispersion image.
We performed this transformation by using the tools provided in
the GEOPSY package (Wathelet et al., 2020).
5. Selection and stacking of dispersion image. After the transformation
of each shot gather, we then select the dispersion image to be sta-
cked manually to increase the dispersion image quality. An exam-
ple of the stacked dispersion image of X
mid
=205 m is shown in
Figure 4c.
These processes are repeated for every X
mid
in our profile, which we
sample every 8 m so that we have sufficient changes in the shot and
receiver combination.
Finally, we picked the dispersion curve in the frequency–phase
velocity domain for each stacked dispersion image. An example of a
stacked dispersion image for X
mid
=205 m is shown in Figure 4c.In
this figure, the picked dispersion curve is marked by the blue line. The
first step of the dispersion curves picking process was to automati-
cally pick the maximum amplitude for each frequency. After the auto-
matic determination of the dispersion curves, we manually picked the
fundamental mode dispersion curve while also filtering the frequency
points which contained low amplitude as suggested in the SWIP
method (Pasquet & Bodet, 2017). With this workflow, we ensured
that the dispersion curve picking is consistent with the maximum
amplitude while also avoiding higher mode dispersion. The picking
procedure was then repeated for every stacked dispersion image. The
picked dispersion curves were then used for the 1D S-wave velocity
inversion.
3.5 |Bayesian inversion of dispersion curves
To reconstruct the 1D S-wave velocity from the dispersion curves, we
implement the transdimensional Bayesian inversion routine described
FIGURE 4 Example of Rayleigh wave’s phase velocity dispersion curve extraction process at mid-point (X
mid
) 205 m: (a) geometry used for
processing at X
mid
=205 m, red squares show the included shots and blue inverted triangles show the included receivers; (b) example of a single-
shot gather (time –offset) from a source located at 254 m using the geometry from (a); (c) stacked dispersion image for X
mid
=205 m using the
geometry from (a), blue lines represent the Rayleigh wave’s phase velocity dispersion curves of different modes [Color figure can be viewed at
wileyonlinelibrary.com]
3306 TRICHANDI ET AL.
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
by Bodin (2010). One of the advantages of using this algorithm is that
we require only minimal initial information of the real model, for
example, the number of layers and each layer’s thickness (Ryberg &
Haberland, 2019). These parameters will be derived during the inver-
sion by the data itself (data-driven). Additionally, with the trans-
dimensional inversion, we need not provide a fixed number of layers
as the inversion scheme will also look for the number of layers by
adding and reducing the number of layers in the model, according to
the data noise level.
The general workflow of the inversion steps (Figure 5) can be
summarized as follows. We used the dispersion curve as the input
data (d
obs
) and provided a range of values as prior information for the
forward modelling parameters. For the initial information, we set a
range of values for parameters required for the forward modelling:
(i) number of layers P(N); (ii) Voronoi cell location P(C); (iii) data noise P
(σ); (iv) P(V
s
) and P(V
p
) value of each Voronoi cell. A random initial
model m
0
is then generated. For nnumber of layers in m
0
, each layer
as defined by the Voronoi cell location will have V
s
and V
p,
which is
then used for the forward modelling of the dispersion curves. We
used the modelling tools from the GEOPSY package (Wathelet
et al., 2020). General steps for a chain of Bayesian inversion are as
follows:
1. Generate a random starting model (m
0
) using initial information P
(V
s
), P(N), P(C), P(V
p
/V
s
), and P(σ).
2. Generate a perturbed model m
1
by randomly performing one of
the following:
a. Perturb the data noise (σ) value.
b. Perturb the cell location of a random cell. Changes in the Voronoi
cell location will affect the layer’s thickness of the input model.
c. Perturb the V
s
value of a random cell.
d. Perturb the V
p
/V
s
value of a random cell and calculate V
p
accord-
ingly for the input model.
e. Delete a random cell.
f. Add a random cell according to the prior information provided by
the chain.
g. Calculate the acceptance probability (α) of moving from m
0
to m
1
based on the d
obs
, modelled dispersion curve of m
0
, and modelled
dispersion curve of m
1
.
h. Randomly accept or reject the move from m
0
to m
1
based on the
acceptance probability calculated in step 3. If accepted, m
0
is then
replaced by m
1
.
i. Repeat from step 2 until the predetermined number of iterations is
reached.
For the inversion, we perform multiple chains of Bayesian inversion
for a single input of data d
obs
. The final 1D model is then derived using
all the models from all chains.
As previously mentioned, the advantage of using the trans-
dimensional Bayesian inversion is that we need only minimum initial
information. For all the inversion routines, we provided uniformly dis-
tributed initial values on P(V
s
), P(N), P(C), P(σ), and P(V
p
). These are as
follows: 0.1–5.0 km/s for P(V
s
); 2–20 layers for P(N); 0–100 m for P
(C); 0.01–0.30 km/s for P(σ), and 0.3–8.0 km/s for P(V
p
). We also per-
form the inversion independent of any information either from the
available borehole data or the body wave tomography model.
4|RESULTS
4.1 |P-wave velocity model
The final P-wave velocity model for the Santa Gracia critical zone is
displayed in Figure 6a. We use a vertical exaggeration of 2:1 to
improve the clarity of the image. In general, the areas close to the sur-
face can be modelled with higher reliability than deeper parts. The
FIGURE 5 General workflow of the transdimensional Bayesian inversion. Input data d
obs
is the extracted Rayleigh wave dispersion curve. We
give minimum prior information on the expected probability of shear wave velocity value P(V
s
), the number of layers P(N), Voronoi cell’s location P
(C), data noise P(σ), and V
p
/V
s
ratio P(V
p
/V
s
). The information is then used to generate a starting model m
0
for a single chain. A randomly picked
parameter from m
0
is then perturbed to create a proposed model m
1
. Both m
0
and m
1
are then used to calculate the acceptance probability (α).
Based on α, the proposed model m
1
is then randomly accepted or rejected. If accepted, we take the proposed model m
1
as the initial model m
0
in
the next iteration. Multiple chains of these processes can be independently run, in which, from all the calculated chains, a final model is then
inferred. The model is then limited to the top of the halfspace layer inferred during the 1D inversion of the dispersion curve. The resulting final V
s
model is then used to create a pseudo-2D V
s
profile by performing linear interpolation and Gaussian smoothing across the profile
TRICHANDI ET AL.3307
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
model is muted at the edges and for deeper parts where no
information on velocity can be derived because of insufficient ray
coverage.
To first order, the velocity model indicates a layered structure
that runs, with some minor deviations, parallel to the surface topogra-
phy. The uppermost layer shows P-wave velocities 0.8 km/s near
the surface, increasing to 2–2.5 km/s around 20 m depth. This low-
velocity layer appears slightly thicker in regions of local topographic
lows, that is, at distances 50–110 and 320–400 m along the profile.
Velocities of 2.5–4.5 km/s occur at around 20–40 m depth. At greater
depth, we observe velocities reaching values of 4.5–6 km/s. In this
deeper layer, contour lines reveal an interesting feature, with lowered
velocities and a slight dip encountered around 65–75 m depth at the
location of the borehole (vertical line at 205 m distance). The subdivi-
sion into three layers is not very sharp and is rather transitional. To
enhance more subtle differences in the layered structures, we show
the vertical velocity gradient (e.g. Bauer et al., 2010) in Figure 6b. This
representation highlights the strong velocity increase with depth, par-
ticularly in the range of 10–30 m. The geological meaning of the layers
and dipping features will be discussed in the context of the borehole
data and the S-wave velocity structure from surface wave analysis.
4.2 |S-wave velocity model
The results of the MASW are presented in Figure 7. The S-wave
velocity profile is shown in Figure 7a and the vertical S-wave velocity
gradient is displayed in Figure 7b. We muted the bottom part of the
model, where the depth is primarily defined by a 1D halfspace of the
phase velocity modelling. The selection of the halfspace boundary
was done automatically by inferring the highest probability of the hal-
fspace during inversion. In addition to the halfspace muting, we also
present the standard deviation contour line of 0.2 km/s (the dashed
black line in Figure 7a) as recommended in the SWIP approach
(Pasquet & Bodet, 2017). While the standard deviation threshold pro-
vides a more robust and conservative assumption on the resolution,
we consider it is still useful to show the model down to the halfspace
boundary limit.
For the S-wave velocity profile, we observe values between 0.38
and 3.10 km/s. We see only slight lateral variations, especially in the
upper 30 m, as the horizontal trend of the S-wave velocity contour
lines is relatively parallel to the surface topography. Below 30 m
depth, we can see some horizontal variations, especially below topo-
graphic lows around 300–400 m distance and 50–150 m distance
(Figure 7). In these parts of the profile, we observe the deepening of
the S-wave velocity contour.
Vertically, we observe a thick, low S-wave velocity layer in the
upper part of the profile. This low-velocity feature goes down to
20 m. Below this thick, low-velocity zone, we observe a high
increase in velocity as indicated by the high-velocity gradient
(Figure 7b). Finally, we have the relatively constant high-velocity zone
of our profile underlying the low-velocity layer.
Like the vertical P-wave velocity gradient model (Figure 6b), the
vertical S-wave velocity gradient model in Figure 7b enhances subtle
details not apparent in the S-wave velocity model. A high vertical
velocity gradient shows an area with strongly varying material proper-
ties (heterogenous), while low values represent relatively homogenous
material. In the vertical velocity gradient model, we have a top layer
with a low gradient followed by a layer with a steep gradient. Similar
to the P-wave vertical velocity gradient model, the shape of the
strong gradient layer is also relatively parallel with the surface topog-
raphy. Underlain by a layer with a high vertical velocity gradient, we
have another layer with relatively low vertical velocity gradient
values.
FIGURE 6 P-wave velocity model at
Santa Gracia: (a) smoothed model resulting
from tomographic inversion of first-arrival
travel times; (b) vertical P-wave velocity
gradient illustrating changes of velocity
with depth. The vertical black line at 205 m
distance indicates the location of the
borehole. Models are shown with vertical
exaggeration (VE) of 2:1 [Color figure can
be viewed at wileyonlinelibrary.com]
3308 TRICHANDI ET AL.
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.3 |V
p
/V
s
model
With the availability of both P- and S-wave velocity information, it is
common to combine both data sets to produce a V
p
/V
s
profile
(e.g. Wadas et al., 2020). However, here we need to consider the dif-
ferent methods and approaches in determining the seismic velocity.
The MASW method created a pseudo-2D profile by combining multi-
ple 1D S-wave velocity profiles reconstructed by modelling
frequency-dependent phase velocity dispersion curves. In contrast,
the body wave tomography method reconstructed the 2D P-wave
velocity profile by modelling of ray paths travelling from source to
receiver (Pasquet et al., 2015). With the different approaches and sen-
sitivities, we therefore expect a certain degree of incompatibility that
can make non-typical V
p
/V
s
ratios difficult to interpret reliably. We
present the V
p
/V
s
ratio model in Figure 8. While the colour pattern
presents interesting isolated high V
p
/V
s
ratios at the top layer of the
topographic lows, we can see that the range of V
p
/V
s
ratios does not
present typical values (>1.73 for unconsolidated rock). A thorough V
p
/
V
s
interpretation requires information on rock petrology, such as
porosity, saturation, bulk, and shear modulus (Brantut & David, 2019),
which is unavailable for our data set. Because it is beyond the scope
of this study to include modelling of these petrophysical parameters,
we concentrate on the combined interpretation of the V
p
and V
s
model across the profile.
5|DISCUSSION
The P- and S-wave velocity information provided by both the body
wave tomography and MASW methods provides different perspec-
tives on the critical zone structure in Santa Gracia. Here, we focus on
the first-order information provided by these two approaches and
integrate them with data from an existing borehole (Krone
et al., 2021; Weckmann et al., 2020) to interpret the weathering zone
from the surface down to 90 m depth at the borehole location, and
then extend it to the entire profile with the obtained 2D seismic data.
5.1 |Borehole data confirmation
We integrate the P- and S-wave velocity models with existing bore-
hole information (Weckmann et al., 2020). The borehole is located at
distance 205 m from the profile (Figure 9). The borehole information
contains televiewer and sonic log data, which we can use to confirm
and correlate with the P- and S-wave velocity models.
The televiewer data in Figure 9a shows acoustic amplitudes
which are correlated with rock quality. Weathered and fractured rocks
return low acoustic televiewer amplitudes. In the uppermost
section of the televiewer profile in Figure 9a, we observe low ampli-
tudes between 7 and 30 m depth. Below 30 m depth, we mainly
observe high televiewer amplitude. We also encounter another layer
with low televiewer amplitude between 72 and 82 m depth in the
deeper part of the section. The variability in the televiewer amplitude
shows that in the borehole we encounter rocks with differing hard-
ness. The relatively lower amplitude in the upper 30 m shows that the
rock in this depth range is relatively weaker compared to the rocks
below 30 m depth. The low televiewer amplitudes between 72 and
82 m depth also indicate a relatively weakened rock.
The velocity comparison in Figure 9b shows the S-wave velocity
from the MASW (blue) and the sonic log (grey), as well as the P-wave
velocity from the body wave tomography (green) and its
corresponding sonic log (orange). Due to the weak rock encountered
in the upper 20 m below the surface, the sonic log could not be used
to determine seismic velocities at this shallow depth range. Using both
FIGURE 7 S-wave velocity model at
Santa Gracia: (a) smoothed 2D S-wave
velocity model derived from surface wave
tomography; (b) vertical S-wave velocity
gradient derived from (a). Models are
shown with vertical exaggeration (VE) of
2:1. Dashed black line in (a) represents the
0.2 km/s standard deviation boundary
from the generated models [Color figure
can be viewed at wileyonlinelibrary.com]
TRICHANDI ET AL.3309
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
the MASW and body wave tomography methods, we managed to
model the seismic velocity structure at the shallow part of the subsur-
face between 0 and 20 m depth. At the borehole location, both P-
wave velocities from the body wave tomography and S-wave veloci-
ties from the MASW model are in good agreement with the sonic log-
derived velocities (Figure 9b). However, as expected, both the MASW
and body wave tomography methods could not resolve the fine scale
revealed by the sonic log as the methods present different frequency
content. Sonic logs use a higher seismic wave frequency, and thus are
able to resolve the fine details, while the body and surface wave
methods represent lower frequency content and are limited by the
wavelength resolution.
In Figures 9c and e, we also present the P- and S-wave velocity as
coloured images. We use the same colour scale for both images, with
the P-wave colour scale adjusted as 1.8 times the S-wave colour scale.
Comparing Figures 9c and e, we can see that we have a consistent
feature of a low absolute P- and S-wave velocity value in the upper
16 m (blue to light green colour). Below 16 m depth, P- and S-wave
velocity values increase steeply as shown by the vertical velocity
gradient images in panels D and F. The relatively high vertical velocity
gradient (indicated by red colours) shows a clear transition between
the upper layer with relatively low P- and S-wave velocity and the rel-
atively higher P- and S-wave velocity layer below.
5.2 |Seismic expression of the weathering
structure
We identified three major layers from the seismic results. When cali-
brated by the borehole data, the layers are: saprolite, weathered bed-
rock, and bedrock. As shown in Figure 9, each layer has different
characteristics, especially in terms of seismic velocity. By combining
our geophysical results with the geochemical data from Krone et al.
(2021), we discuss the possible layer interpretation as summarized in
Figure 10. The interpreted saprolite and weathered bedrock layers
reach 30 m depth, which we interpret as the extent of the weathering
front. We use the layers identified at the borehole to produce a 2D
interpretation of the seismic profile.
FIGURE 8 V
p
/V
s
model produced using the
P-wave velocity model from body wave
tomography and the S-wave velocity model from
the MASW method. The different methods in
modelling the P- and S-wave velocity presents a
sensitivity problem which disables interpretation
of the V
p
/V
s
model [Color figure can be viewed at
wileyonlinelibrary.com]
FIGURE 9 Summary of the information at the borehole location: (a) normalized televiewer amplitude data; (b) seismic velocity data from P-
wave velocity from body wave tomography (green), S-wave velocity from MASW (blue), P-wave velocity from sonic log (yellow), and S-wave
velocity from sonic log (grey); (c) colour plot of P-wave velocity; (d) colour plot of P-wave vertical velocity gradient; (e) colour plot of S-wave
velocity; (f) colour plot of S-wave vertical velocity gradient; (g, h) Fe(III)/Fe total and porosity data. Dashed lines in (g), (h), and (i) shows the mean
value [Color figure can be viewed at wileyonlinelibrary.com]
3310 TRICHANDI ET AL.
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.3 |Saprolite
In a previous study on a similar granitic environment in Wyoming,
Flinchum et al. (2018) suggested using the depth of the borehole cas-
ing as an indirect indication of the boundary between saprolite and
weathered bedrock. Following this approach, the borehole casing
depth in the study site is located at 6 m depth and attributed to P-
and S-wave velocity of 0.8 and 0.6 km/s, respectively. These P- and
S-wave velocity values are considerably lower than the P-wave veloc-
ity of 1.2 km/s suggested by Flinchum et al. (2018). Inversely, when
we consider the P-wave velocity value of 1.2 km/s as the marker for
the bottom of the saprolite layer, we reach a depth of 10 m. On the
other hand, based on geochemical interpretation of samples taken
from the borehole core, Krone et al. (2021) identified the depth range
of 0.1–8.6 m as the saprolite layer, followed by a saprock layer as
shown in Figure 10b. Based on this interpretation, the bottom of the
saprolite layer in the borehole would correspond to a P-wave velocity
of 1.05 km/s and an S-wave velocity of 0.75 km/s.
We attribute the difference in the identified seismic velocity of
the saprolite layer between our work and the study by Flinchum et al.
(2018) to the different protolith encountered in Santa Gracia and the
Blair–Wallis critical zone (BWCZ) in Wyoming. In addition, the degree
of weathering is also likely to account for different seismic velocities
of the saprolite in both field sites. For example, the observed saprolite
S-wave velocity in Santa Gracia shows a value of 0.75 km/s at 8.6 m
depth and is relatively higher compared to the S-wave velocity of
0.6 km/s used to identify the saprolite layer in the BWCZ (Keifer
et al., 2019). The higher S-wave velocity found in our data is likely to
be attributed to the different weathering degrees of the saprolite.
Nevertheless, we consider the difference to be reasonable and the
layering visible in the P- and S-wave velocity images further
strengthens the interpretation of saprolite occurring down to 8.6 m
depth (Krone et al., 2021).
Another characteristic in the upper 8.6 m is the relatively con-
stant P- and S-wave velocity as also shown by the low vertical veloc-
ity gradient (Figures 8c–f). This could indicate a relatively
homogenous saprolite layer in terms of physical properties and com-
position. We expect that down to 8.6 m depth, the layer is highly
affected by surface-related weathering processes. However, since the
velocity gradient around the 8.6 m depth is also relatively low, we
consider the interface between the saprolite and the underlying
weathered bedrock to be more transitional and not of a sharp bound-
ary. The transition can be attributed to a decreasing weathering rate
with depth (Brantley et al., 2008).
5.4 |Weathered bedrock
We identified weathered bedrock from the bottom of the saprolite at
8.6 m to a depth of 30 m in the borehole location. The bottom of the
weathered bedrock layer was identified based on a P-wave velocity of
4.0 km/s (Flinchum et al., 2018), which will be discussed in detail in
the bedrock discussion. While Krone et al. (2021) interpreted the layer
between 8.6 and 34.3 m depth as a single unit, the televiewer, sonic
log, as well as the P- and S-wave velocity data (Figure 9) indicate
another significant change of lithology around 16 m depth. At this
depth range, we observe a P-wave velocity of 2.0 km/s and an S-wave
velocity of 1.36 km/s. Previous P-wave velocity studies on a weath-
ered granite environment in Spain chose this 2.0 km/s P-wave veloc-
ity to mark the boundary between saprolite and moderately
weathered bedrock (Begonha & Sequeira Braga, 2002; Olona
et al., 2010). The relatively high vertical velocity gradient of the P- and
S-wave in our data also coincides with this 2.0 km/s marker, which
indicates greater heterogeneity than indicated in the single zone iden-
tified by Krone et al. (2021). The high vertical velocity gradient of the
P- and S-wave could indicate a sharper lithology change, especially at
the peak between 16 and 20 m depth. Based on this observation, we
expect that the layer between 8.6 and 16 m depth underwent stron-
ger weathering compared to the layer between 16 and 34.3 m depth.
Intuitively, we tried to attribute the sudden increase in P- and S-
wave velocity shown by the high vertical velocity gradient between
16 and 20 m depth to closures of pores and fractures with increasing
depth. However, the porosity data in Figure 9i does not show signifi-
cant changes of porosity over this depth range. On the other hand, the
Fe (III)/Fe total information provided in Krone et al. (2021)showsarel-
atively high iron reduction in the upper 16 m (Figure 9h). The differing
features between the porosity and iron redox information could indi-
cate that the weathering processes down to 16 m depth can be
FIGURE 10 Critical zone structure in
the Santa Gracia reserve: (a) from seismic
and borehole data; (b) from geochemical
data; (c) conceptual process model. From
top to bottom, we identified saprolite and
weathered bedrock which forms the
regolith down to 30 m depth [Color figure
can be viewed at wileyonlinelibrary.com]
TRICHANDI ET AL.3311
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
attributed more to iron reduction by dissolution processes than to
increase of porosity. Additionally, reactions of Fe minerals can increase
strain in the rock and trigger weathering-induced micro-fracturing
(Behrens et al., 2015;Fletcheretal.,2006). It is, however, difficult to
prove this relation without fracture density data from the core.
For a depth range of 16–34.3 m, we expect different weathering
processes compared to the weathering process in the upper layer
(<16 m depth). Based on the observation of high chemical depletion
factor (CDF; Figure 9g), we relate the weathering of the bedrock layer
down to 16 m depth to chemical alteration due to water infiltrating
from precipitation at the surface. In contrast, the low iron reduction
and sharp increase in seismic velocities of the weathered bedrock
from 16 to 30 m depth suggest weathering that is unrelated to
surface-derived fluids. Instead, we expect the sharp increase of seis-
mic velocities to be due to the change of physical properties.
5.5 |Bedrock
The P-wave velocity data in Figure 9b show that velocities >4.0 km/s
occur at 30 m depth. Previous studies in BWCZ suggest that the
4.0 km/s P-wave velocity can be used as a marker to identify the top
of granite bedrock (Callahan et al., 2020; Flinchum et al., 2018). There-
fore, we identify the top of the bedrock at 30 m depth. This inter-
pretation is also supported by the observation of the high televiewer
amplitude below 30 m depth (Figure 9a).
The 4.0 km/s velocity marker at 30 m depth also occurs below a
relatively steep vertical velocity gradient, similar to what has been
observed in the BWCZ (Flinchum et al., 2018). At the same depth, and
also beneath a relatively steep S-wave velocity gradient, the S-wave
velocity shows a value of 2.5 km/s (Figures 9d and e). High P-wave
velocities extend down to 80 m and reach up to 5.0 km/s at 40 m.
The acoustic televiewer in Figure 9a also shows rather constant ampli-
tudes down to 75 m. We interpreted the layer with a decrease of
acoustic televiewer amplitudes and lowered sonic P- and S-wave
velocities between 72 and 82 m depth as a local zone of mechanical
weakness such as a fault zone. However, we lack the necessary infor-
mation for further interpretation of these features.
With the interpreted top of the bedrock at 30 m depth, we iden-
tify the weathering front at our borehole location to reach down to
30 m depth, which includes the saprolite and weathered bedrock layer
as the regolith, as shown in Figure 10a. Based on geochemical data,
Krone et al. (2021) argue for multiple weathering fronts, with the first
one located at 35 m depth. While the geochemical interpretation pro-
vided a different interpretation, the depth difference is not significant
(<5 m) and the P- and S-wave velocity still manages to provide a rela-
tively good match with the interpreted layer by Krone et al. (2021),
especially for the saprolite and weathered bedrock layer. The differing
interpretation can be attributed to how the geochemical interpretation
attributes the existence of chemical reactions such as oxidation as evi-
dence of weathering processes. However, the physical characteristics
of the bedrock with a P-wave velocity >4.0 km/s is attributed more to
the crystalline fresh bedrock, and chemical weathering in this lithology
would be limited to fracture surfaces (Flinchum et al., 2018).
The hypothesized major decrease in surface-related weathering
effect from 16 m depth (Figure 10c) requires further evidence. This
hypothesis is mainly driven by the iron reduction data shown in
Figure 9h and the high vertical velocity gradient shown in Figures 9d
and f. Nevertheless, the extent of the weathered bedrock from our
data at 30 m depth generally agrees with the interpreted Zone III from
the geochemical data interpretation at 34.6 m depth. As for the bot-
tom of the saprolite layer, we find a good agreement between the
geophysical data and geochemical data.
5.6 |Lateral extent
The interpretation of the weathering structure as calibrated at the
borehole is finally applied along the entire seismic profile (Figure 11a).
While the weathering front is identified around the 30 m depth at the
borehole location, laterally the weathering front shows variations rela-
tive to the topography. This variation is most pronounced for the rela-
tively thick saprolite and weathered bedrock layer 300–400 m profile
distance, as shown also in Figure 11b. While we can intuitively relate
the thickening of the regolith to alluvial accumulation in the valley
bottom, that is not the case in our seismic profile as an alluviated val-
ley will show a stronger relief—as shown in the northwestern part of
the seismic profile (Figure 2b). Therefore, we expect that any varying
weathering front depth across the profile is due to the different effect
of weathering processes.
The low televiewer amplitude between 72 and 82 m depth
(Figure 9a), as also observed in the geochemical analysis by Krone
et al. (2021), coincides with a slight velocity distortion shown in the
conceptual model (Figure 11a). We therefore speculate that the zone
of low televiewer amplitude could be related to a larger geological
structure, such as a fault. Such a structure could provide the pathway
for surface-derived water to infiltrate into the deeper subsurface and
enable chemical weathering at depth (Holbrook et al., 2019). How-
ever, further evidence is needed to support the hypothesis that the
identified feature is part of a larger structure. For example, should
geochemical analyses show indications of meteoric water in the frac-
ture zone, the structure may be linked to the surface.
To estimate the regolith residence time, we can divide the rego-
lith thickness (27 m) by an estimate of the denudation rate of
11 m/Myr, based on cosmogenic nuclides (Krone et al., 2021). This
results in a turnover time of the regolith of 2.5 Myr, which covers
the whole Quaternary period. Such a long time period suggests that
the processes involved in creating the weathering zone we observe in
Santa Gracia must be operating over geological time scales. Short-
term variations that may occur over orbital or glacial–interglacial time
scales are therefore less likely to affect the thickness of the
weathering zone significantly.
5.7 |Controlling processes
The identified structure of the critical zone in Santa Gracia is a result
of various controlling processes which affect the surface and subsur-
face characteristics and structures of the critical zone. To understand
these processes, we compare our conceptual model in Figure 11a to
various geomorphological models of the critical zone that make
explicit predictions about the depth of the weathering zone.
Traditional models of weathering advance suggest a ‘top-down’
control due to the infiltration of meteoric water into the bedrock,
3312 TRICHANDI ET AL.
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
which drives chemical reactions in the subsurface (Brantley &
White, 2009). The downward infiltration of meteoric water will con-
tinue as permeability allows and induce weathering processes until
the water reaches chemical equilibrium. Following this hypothesis and
assuming a homogenous bedrock across our profile, we expect a
weathering depth that is parallel to the surface topography. However,
the saprolite in our model appears to be thinner in the topographic
high, especially around the 120–360 m profile distance. Inversely, we
observed thicker saprolite at topographic lows. Based on this observa-
tion, and following the ‘top-down’hypothesis, we may conclude that
the different weathering depths are due to different bedrock charac-
teristics along the profile. For example, bedrock with higher fracture
density at the topographic low would allow for deeper penetrating
weathering (Brantley et al., 2017; Lodes et al., 2021). With the current
data set, however, we are unable to provide clear evidence of differ-
ent fracture density along the profile. Further seismic analysis which
considers seismic anisotropy and bedrock fracture density data could
help in testing this hypothesis in future works.
In contrast to the ‘top-down’control, Rempe and Dietrich (2014)
proposed a ‘bottom-up’control, in which the thickness of the weath-
ered zone is controlled by whether the bedrock can be drained. This
drainage is largely determined by the topographic profile on the sur-
face and the groundwater profile in the subsurface. The hypothesis
predicts a thickening of the weathered zone in topographic highs
(upslope) and a thinning of the weathered zone towards channels
(here, the topographic lows). This prediction is opposite to what we
observe in Santa Gracia National Reserve. The difference compared
to the model presented by Rempe and Dietrich (2014) could be
related to the aridity and the absence of perennial flow in our
study area.
Another model of regolith formation was presented by Braun
et al. (2016), who hypothesized that the depth of the weathering front
is limited by the ability of groundwater to transfer solutes before the
reacting fluids reach saturation. The model predicts steady-state rego-
lith thickness as a function of topography (length and slope), uplift
rate, erosion rate, weathering rate, hydraulic conductivity, and surface
transport coefficient. In an arid landscape, the model predicts a thicker
regolith at the top of a hill, similar to the bottom-up model of Rempe
and Dietrich (2014). On the other hand, where the precipitation rate
is high and/or the uplift rate is low, the model predicts a thicker rego-
lith under topographic lows. In Santa Gracia, which is located in an
arid area (Karger et al., 2017; Werner et al., 2018), we observe that
regolith is relatively thicker in the topographic low. Braun et al. (2016)
have also shown, however, that changes in precipitation can have
striking results on spatial variations in regolith thickness. In Santa
Gracia, where the estimated time scale of regolith formation covers
the entire Quaternary period, the present-day thickness of the rego-
lith could thus be the result of very different climates in the recent
geological past. In addition, Braun et al. (2016) have shown that in
eroding landscapes, changes in precipitation can have striking results
on spatial variations in regolith thickness. In Santa Gracia, where the
estimated time scale of regolith formation covers the entire Quater-
nary period, the present-day thickness of the regolith could thus be
the result of very different climates in the recent geological past.
In a transient experiment by Braun et al. (2016), in which the criti-
cal zone transitions from saturated to unsaturated conditions, the
downslope trend in regolith thickness switches trend and eventually
attains a relatively uniform thickness. Specifically, when precipitation
and uplift rates are low and the regolith is unsaturated, their model
predicts a regolith layer that is thinning towards the topographic high.
Based on this comparison, it is likely that the critical zone structure in
our study site is in a transient state and going through a continuous
process of desaturation of the regolith, which could be linked to the
aridification of the Atacama Desert.
In addition to the existing regolith thickness prediction models,
which consider the groundwater level, we also investigate a geomor-
phological model which predicts possible fractured bedrock. We first
compare our conceptual model with the topographic stress model
presented by Slim et al. (2014). The model used a two-dimensional
boundary element method to calculate the minimum cohesion
FIGURE 11 (a) Conceptual model of
the measured seismic profile from
combined P- and S-wave velocity
information. Black line shows the
borehole location, with the red line
indicating the relatively weak zone
detected in the televiewer data. The
conceptual model is shown with vertical
exaggeration (VE) of 2:1. (b) Estimated
regolith thickness across the profile with
observed thinning around topographic
high (150 m distance) and thickening
around topographic low (300–400 m
distance) [Color figure can be viewed at
wileyonlinelibrary.com]
TRICHANDI ET AL.3313
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
required to prevent shear failure as a proxy for potential fracturing.
The study shows that in valleys/topographic lows, the bedrock is
more likely to have shear fractures compared to bedrock at hills/
topographic highs. When we consider the thicker regolith in the topo-
graphic low at our study site, we can also assume that this behaviour
is due to the higher fracture density in the valley of our model. With
higher fracture density, the bedrock in this valley will be easier to
weather and thus to develop a deeper regolith. The study also shows
rapid decline of C
min
in the upper 15 m, which we speculate can be
related to a rapid increase of seismic velocity as shown in Figure 9b.
However, a more detailed analysis of the rock’s physics, especially in
the borehole location, is still needed to investigate this hypothesis.
Further modelling of the effect of topographic stress on bedrock
was also presented by St. Clair et al. (2015). In the presented failure
potential model, abundance of fracture in the bedrock is controlled by
the ratio between tectonic stress and gravitational stress (σ*). The
model predicts that in areas with low tectonic compression or widely
spaced ridges and valleys, the depth of the weathered rock will be rel-
atively uniform in space. Considering that the Chilean coastal cordil-
lera is a tectonically active area, where tectonic stresses are large
(>7 MPa around 36S; Luttrell et al., 2011), we estimate σ* > 0.75.
With σ* < 1.0, the model predicts that the weathered rocks across the
landscape will be parallel to the surface topography, which does not
match completely with our observation. One possible explanation for
this difference is that the observed model that we present is already a
result of complex interaction between different weathering processes,
which is not completely considered by the topographic stress
modelling.
6|CONCLUSION AND OUTLOOK
In this study, we seismically imaged the weathering structure in the
Santa Gracia Reserve in Chile using body wave tomography and the
MASW method. The resulting P- and S-wave velocity models agree
with available borehole information and, moreover, complement the
missing upper 20 m of the sonic log data and support the existing geo-
chemical interpretation. Across the surveyed seismic profile, we esti-
mate a mean weathering front depth of ca. 27 m and identified
saprolite thickness variations at minor topographic highs and lows.
The vertical velocity gradients of the P- and S-wave velocities provide
a complementary perspective on the lithological character at depth
and evidence additional subtle features within the different layers.
For example, a high-velocity gradient can be related to the rapid
changes in the rock’s physical properties or mineral composition. The
identified weak zone at depth in the borehole data can be related to a
possible fault, however, further investigation of such features is still
required, with both geophysical and geochemical approaches.
The seismic, borehole, and geochemical results allow us to derive
an integrated interpretation of the weathering structure. While we
found some differences between the geophysical and geochemical
interpretation, both data sets complement and strengthen one
another. Therefore, we suggest that geophysical investigation using
the seismic method with both body and surface waves will enhance
other near-surface investigations of the critical zone.
Comparison to different models of regolith evolution shows that
the conceptual model we have in Santa Gracia cannot be attributed to
any single specific model. While the study area has only limited pre-
cipitation and no water table, it is likely that the effect of the limited
precipitation still affects the advance of regolith due to water infiltra-
tion. It is important to note that regolith formation covers the entire
Quaternary period and thus integrates over different climate condi-
tions during this time period. On the other hand, the infiltration of
water into the subsurface requires a certain level of permeability,
which can be related to the opening of fractures in the bedrock as a
result of topographic and tectonic stresses. Therefore, the develop-
ment of a regolith evolution model, which couples both aspects, can
be important for future studies of the weathering zone.
ACKNOWLEDGEMENTS
We acknowledge support from the German Science Foundation
(DFG) priority research programme SPP-1803 ‘EarthShape: Earth Sur-
face Shaping by Biota’(Grant No. KR 2073/5-1 to Charlotte
M. Krawczyk). We thank CONAF and Park Rangers for the possibility
of working in the national parks, as well as providing access to the
research site. We thank Martin Krüger and our colleagues at the Uni-
versity of Concepci
on for their assistance during seismic data acquisi-
tion in Santa Gracia, Chile. We also thank Geophysical Instrument
Pool Potsdam (GIPP) for providing all the necessary equipment for the
seismic campaign. The reviews from two anonymous reviewers
greatly helped to improve the manuscript. Open Access funding
enabled and organized by Projekt DEAL.
DATA AVAILABILITY STATEMENT
The data that support the findings will be available in the GFZ Data
Repository following an embargo to allow for doctoral publication of
research findings.
ORCID
Rahmantara Trichandi https://orcid.org/0000-0002-4536-9202
Klaus Bauer https://orcid.org/0000-0002-7777-2653
Trond Ryberg https://orcid.org/0000-0001-7129-5596
Dirk Scherler https://orcid.org/0000-0003-3911-2803
Klaus Bataille https://orcid.org/0000-0001-6006-6747
Charlotte M. Krawczyk https://orcid.org/0000-0002-5505-6293
REFERENCES
Bauer, K., Moeck, I., Norden, B., Schulze, A., Weber, M. & Wirth, H. (2010)
Tomographic Pwave velocity and vertical velocity gradient structure
across the geothermal site Groß Schönebeck (NE German Basin):
Relationship to lithology, salt tectonics, and thermal regime. Journal
of Geophysical Research, 115(B8), B08312. Available from: https://
doi.org/10.1029/2009jb006895
Baumann-Wilke, M., Bauer, K., Schovsbo, N.H. & Stiller, M. (2012) P-wave
traveltime tomography for a seismic characterization of black shales
at shallow depth on Bornholm, Denmark. Geophysics, 77(5), EN53–
EN60. Available from: https://doi.org/10.1190/geo2011-0326.1
Bazilevskaya, E., Lebedeva, M., Pavich, M., Rother, G., Parkinson, D.Y.,
Cole, D. et al. (2013) Where fast weathering creates thin regolith and
slow weathering creates thick regolith. Earth Surface Processes and
Landforms, 38(8), 847–858. Available from: https://doi.org/10.1002/
esp.3369
Befus, K.M.M., Sheehan, A.F.F., Leopold, M., Anderson, S.P.P. &
Anderson, R.S.S. (2011) Seismic constraints on critical zone architec-
ture, Boulder Creek watershed, front range, Colorado. Vadose Zone
Journal, 10(3), 915–927. Available from: https://doi.org/10.2136/
vzj2010.0108
3314 TRICHANDI ET AL.
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Begonha, A. & Sequeira Braga, M. (2002) Weathering of the Oporto gran-
ite: Geotechnical and physical properties. Catena, 49(1–2), 57–76.
Available from: https://doi.org/10.1016/S0341-8162(02)00016-4
Behrens, R., Bouchez, J., Schuessler, J.A., Dultz, S., Hewawasam, T. & Von
Blanckenburg, F. (2015) Mineralogical transformations set slow
weathering rates in low-porosity metamorphic bedrock on mountain
slopes in a tropical climate. Chemical Geology, 411, 283–298. Avail-
able from: https://doi.org/10.1016/j.chemgeo.2015.07.008
Bernhard, N., Moskwa, L.M., Schmidt, K., Oeser, R.A., Aburto, F., Bader, M.
Y. et al. (2018) Pedogenic and microbial interrelations to regional cli-
mate and local topography: New insights from a climate gradient
(arid to humid) along the coastal cordillera of Chile. Catena, 170,
335–355. Available from: https://doi.org/10.1016/j.catena.2018.
06.018
Bodin, T. (2010) Transdimensional approaches to geophysical inverse prob-
lems. PhD thesis, Claude Bernard University Lyon 1. https://doi.org/
10.1021/jo00151a027
Brantley, S.L., Bandstra, J., Moore, J. & White, A.F. (2008) Modelling chem-
ical depletion profiles in regolith. Geoderma, 145(3–4), 494–504.
Available from: https://doi.org/10.1016/j.geoderma.2008.02.010
Brantley, S.L., Goldhaber, M.B. & Vala Ragnarsdottir, K. (2007) Crossing
disciplines and scales to understand the critical zone. Elements, 3(5),
307–314. Available from: https://doi.org/10.2113/gselements.3.
5.307
Brantley, S.L., Lebedeva, M.I., Balashov, V.N., Singha, K., Sullivan, P.L. &
Stinchcomb, G. (2017) Toward a conceptual model relating chemical
reaction fronts to water flow paths in hills. Geomorphology, 277,
100–117. Available from: https://doi.org/10.1016/j.geomorph.2016.
09.027
Brantley, S.L. & White, A.F. (2009) Approaches to modeling weathered
regolith. In: Oelkers, E.H. & Schott, J. (Eds.) Thermodynamics and
Kinetics of Water–Rock Interaction. Berlin: De Gruyter, pp. 435–484
https://doi.org/10.1515/9781501508462-012
Brantut, N. & David, E.C. (2019) Influence of fluids on V
P
/V
S
ratio: Increase
or decrease? Geophysical Journal International, 216(3), 2037–2043.
Available from: https://doi.org/10.1093/gji/ggy518
Braun, J., Mercier, J., Guillocheau, F. & Robin, C. (2016) A simple model for
regolith formation by chemical weathering. Journal of Geophysical
Research –Earth Surface, 121(11), 2140–2171. Available from:
https://doi.org/10.1002/2016JF003914
Callahan, R.P., Riebe, C.S., Pasquet, S., Ferrier, K.L., Grana, D., Sklar, L.S.
et al. (2020) Subsurface weathering revealed in hillslope-integrated
porosity distributions. Geophysical Research Letters, 47(15),
e2020GL088322. Available from: https://doi.org/10.1029/
2020gl088322
Comina, C., Krawczyk, C.M., Polom, U. & Socco, L.V. (2017) Integration of
SH seismic reflection and love-wave dispersion data for shear wave
velocity determination over quick clays. Geophysical Journal Interna-
tional, 210(3), 1922–1931. Available from: https://doi.org/10.1093/
gji/ggx276
Dal Bo, I., Klotzsche, A., Schaller, M., Ehlers, T.A., Kaufmann, M.S., Fuentes
Espoz, J.P. et al. (2019) Geophysical imaging of regolith in landscapes
along a climate and vegetation gradient in the Chilean coastal cordil-
lera. Catena, 180, 146–159. Available from: https://doi.org/10.
1016/j.catena.2019.04.023
Fletcher, R.C., Buss, H.L. & Brantley, S.L. (2006) A spheroidal weathering
model coupling porewater chemistry to soil thicknesses during
steady-state denudation. Earth and Planetary Science Letters, 244
(1–2), 444–457. Available from: https://doi.org/10.1016/j.epsl.2006.
01.055
Flinchum, B.A., Holbrook, W.S., Rempe, D., Moon, S., Riebe, C.S., Carr, B.J.
et al. (2018) Critical zone structure under a granite ridge inferred
from drilling and three-dimensional seismic refraction data. Journal of
Geophysical Research –Earth Surface, 123(6), 1317–1343. Available
from: https://doi.org/10.1029/2017JF004280
Gobierno de Chile Servicio Nacional de Geologia y Mineria. (2003) Mapa
Geologico de Chile: Version Digital.
Graham, R.C., Rossi, A.M. & Hubbert, K.R. (2010) Rock to regolith conver-
sion: Producing hospitable substrates for terrestrial ecosystems. GSA
Today, 20(2), 4–9. Available from: https://doi.org/10.1130/
GSAT57A.1
Hayes, N.R., Buss, H.L., Moore, O.W., Krám, P. & Pancost, R.D. (2020)
Controls on granitic weathering fronts in contrasting climates. Chemi-
cal Geology, 535(March 2019), 119450. Available from: https://doi.
org/10.1016/j.chemgeo.2019.119450
Holbrook, W.S., Marcon, V., Bacon, A.R., Brantley, S.L., Carr, B.J.,
Flinchum, B.A. et al. (2019) Links between physical and chemical
weathering inferred from a 65-m-deep borehole through Earth’s crit-
ical zone. Scientific Reports, 9(1), 4495–5005. Available from: https://
doi.org/10.1038/s41598-019-40819-9
Holbrook, W.S., Riebe, C.S., Elwaseif, M., Hayes, J.L., Basler-Reeder, K.,
Harry, D.L. et al. (2014) Geophysical constraints on deep weathering
and water storage potential in the Southern Sierra critical zone
observatory. Earth Surface Processes and Landforms, 39(3), 366–380.
Available from: https://doi.org/10.1002/esp.3502
Ivanov, J., Miller, R.D., Lacombe, P., Johnson, C.D. & Lane, J.W. (2006)
Delineating a shallow fault zone and dipping bed rock strata using
multichannel analysis of surface waves with a land streamer. Geo-
physics, 71(5), A39–A42. Available from: https://doi.org/10.1190/1.
2227521
Karger, D.N., Conrad, O., Böhner, J., Kawohl, T., Kreft, H., Soria-Auza, R.W.
et al. (2017) Climatologies at high resolution for the earth’s land sur-
face areas. Scientific Data,4,1–20. Available from: https://doi.org/
10.1038/sdata.2017.122
Keifer, I., Dueker, K. & Chen, P. (2019) Ambient Rayleigh wave field imag-
ing of the critical zone in a weathered granite terrain. Earth and Plan-
etary Science Letters, 510, 198–208. Available from: https://doi.org/
10.1016/j.epsl.2019.01.015
Konstantaki, L.A., Ghose, R., Draganov, D., Diaferia, G. & Heimovaara, T.
(2015) Characterization of a heterogeneous landfill using seismic and
electrical resistivity data. Geophysics, 80(1), EN13–EN25. Available
from: https://doi.org/10.1190/geo2014-0263.1
Krone, L.V., Hampl, F.J., Schwerdhelm, C., Bryce, C., Ganzert, L., Kitte, A.
et al. (2021) Deep weathering in the semi-arid coastal cordillera,
Chile. Scientific Reports, 11(1), 1–15. Available from: https://doi.org/
10.1038/s41598-021-90267-7
Kukowski, N. & Oncken, O. (2006) Subduction erosion –the “normal”
mode of fore-arc material transfer along the Chilean margin? In:
Oncken, O., & Kukowski, N. (Eds.) The Andes. Berlin: Springer,
pp. 217–236 https://doi.org/10.1007/978-3-540-48684-8_10
Leone, J.D., Holbrook, W.S., Riebe, C.S., Chorover, J., Ferré, T.P.A., Carr, B.
J. et al. (2020) Strong slope-aspect control of regolith thickness by
bedrock foliation. Earth Surface Processes and Landforms, 45(12),
2998–3010. Available from: https://doi.org/10.1002/esp.4947
Lodes, E., Scherler, D., Wittmann, H. & van Dongen, R. (2021) Patchy bed-
rock explained: Tectonic fracture control on landscape evolution pat-
terns in South-Central Chile. EGU General Assembly,19–30 April,
EGU2021--9231.
Luttrell, K.M., Tong, X., Sandwell, D.T., Brooks, B.A. & Bevis, M.G. (2011)
Estimates of stress drop and crustal tectonic stress from the
27 February 2010 Maule, Chile, earthquake: Implications for fault
strength. Journal of Geophysical Research: Solid Earth, 116(B11),
B11401. Available from: https://doi.org/10.1029/2011jb008509
Miller, R.D., Xia, J., Park, C.B. & Ivanov, J.M. (1999) Multichannel analysis
of surface waves to map bedrock. The Leading Edge, 18(12), 1392–
1396. Available from: https://doi.org/10.1190/1.1438226
Oeser, R.A., Stroncik, N., Moskwa, L.M., Bernhard, N., Schaller, M.,
Canessa, R. et al. (2018) Chemistry and microbiology of the critical
zone along a steep climate and vegetation gradient in the Chilean
coastal cordillera. Catena, 170, 183–203. Available from: https://doi.
org/10.1016/j.catena.2018.06.002
Oeser, R.A. & Von Blanckenburg, F. (2020) Do degree and rate of silicate
weathering depend on plant productivity? Biogeosciences, 17(19),
4883–4917. Available from: https://doi.org/10.5194/bg-17-4883-
2020
Olona, J., Pulgar, J.A., Fernández-Viejo, G., L
opez-Fernández, C. &
González-Cortina, J.M. (2010) Weathering variations in a granitic
massif and related geotechnical properties through seismic and
TRICHANDI ET AL.3315
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
electrical resistivity methods. Near Surface Geophysics, 8(6), 585–
599. Available from: https://doi.org/10.3997/1873-0604.2010043
Park, C.B., Miller, R.D., Ryden, N., Xia, J. & Ivanov, J. (2005) Combined use
of active and passive surface waves. Journal of Environmental and
Engineering Geophysics, 10(3), 323–334. Available from: https://doi.
org/10.2113/JEEG10.3.323
Park, C.B., Miller, R.D. & Xia, J. (1999) Multichannel analysis of surface
waves. Geophysics, 64(3), 800–808. Available from: https://doi.org/
10.1190/1.1444590
Pasquet, S. & Bodet, L. (2017) SWIP: An integrated workflow for
surface-wave dispersion inversion and profiling. Geophysics, 82(6),
WB47–WB61. Available from: https://doi.org/10.1190/GEO2016-
0625.1
Pasquet, S., Bodet, L., Dhemaied, A., Mouhri, A., Vitale, Q., Rejiba, F. et al.
(2015) Detecting different water table levels in a shallow aquifer
with combined P-, surface and SH-wave surveys: Insights from V
P
/V
S
or Poisson’s ratios. Journal of Applied Geophysics, 113, 38–50.
Available from: https://doi.org/10.1016/j.jappgeo.2014.12.005
Rempe, D.M. & Dietrich, W.E. (2014) A bottom-up control on fresh-
bedrock topography under landscapes. Proceedings of the National
Academy of Sciences of the United States of America, 111(18),
6576–6581. Available from: https://doi.org/10.1073/pnas.
1404763111
Ryberg, T. & Haberland, C. (2019) Bayesian simultaneous inversion for
local earthquake hypocentres and 1-D velocity structure using mini-
mum prior knowledge. Geophysical Journal International, 218(2), 840–
854. Available from: https://doi.org/10.1093/gji/ggz177
Slim, M., Perron, J.T., Martel, S.J. & Singha, K. (2014) Topographic stress
and rock fracture: a two-dimensional numerical model for arbitrary
topography and preliminary comparison with borehole observations.
Earth Surface Processes and Landforms, 40(4), 512–529. Available
from: https://doi.org/10.1002/esp.3646
St. Clair, J., Moon, S., Holbrook, W.S., Perron, J.T., Riebe, C.S., Martel, S.J.
et al. (2015) Geophysical imaging reveals topographic stress control
of bedrock weathering. Science, 350(6260), 534–538. Available from:
https://doi.org/10.1126/science.aab2210
Stierman, D.J. & Healy, J.H. (1984) A study of the depth of weathering and
its relationship to the mechanical properties of near-surface rocks in
the Mojave Desert. Pure and Applied Geophysics, 122(2–4), 425–439.
Available from: https://doi.org/10.1007/BF00874609
Thurber, C. & Eberhart-Phillips, D. (1999) Local earthquake tomography
with flexible gridding. Computers and Geosciences, 25(7), 809–818.
Available from: https://doi.org/10.1016/S0098-3004(99)00007-2
Vázquez, M., Ramírez, S., Morata, D., Reich, M., Braun, J.J. & Carretier, S.
(2016) Regolith production and chemical weathering of granitic rocks
in Central Chile. Chemical Geology, 446, 87–98. Available from:
https://doi.org/10.1016/j.chemgeo.2016.09.023
Wadas, S.H., Tschache, S., Polom, U. & Krawczyk, C.M. (2020) Ground
instability of sinkhole areas indicated by elastic moduli and seismic
attributes. Geophysical Journal International, 222(1), 289–304. Avail-
able from: https://doi.org/10.1093/gji/ggaa167
Wang, W., Chen, P., Keifer, I., Dueker, K., Lee, E.J., Mu, D. et al. (2019)
Weathering front under a granite ridge revealed through full-3D seis-
mic ambient-noise tomography. Earth and Planetary Science Letters,
509, 66–77. Available from: https://doi.org/10.1016/j.epsl.2018.
12.038
Wathelet, M., Chatelain, J.L., Cornou, C., Di Giulio, G., Guillier, B.,
Ohrnberger, M. et al. (2020) Geopsy: A user-friendly open-source
tool set for ambient vibration processing. Seismological Research Let-
ters, 91(3), 1878–1889. Available from: https://doi.org/10.1785/
0220190360
Weckmann, U., Bauer, K., Krawczyk, C., Kück, J., Übernickel, K. & von
Blanckenburg, F. (2020) Geophysical Borehole Logging Data from Santa
Gracia. Chile: GFZ Data Services. https://doi.org/10.5880/GFZ.2.7.
2020.001
Werner, C., Schmid, M., Ehlers, T.A., Fuentes-Espoz, J.P., Steinkamp, J.,
Forrest, M. et al. (2018) Effect of changing vegetation and precipita-
tion on denudation –Part 1: Predicted vegetation composition and
cover over the last 21 thousand years along the coastal cordillera of
Chile. Earth Surface Dynamics, 6(4), 829–858. Available from: https://
doi.org/10.5194/esurf-6-829-2018
Yaede, J.R., McBride, J.H., Nelson, S.T., Park, C.B., Flores, J.A., Turnbull, S.
J. et al. (2015) A geophysical strategy for measuring the thickness of
the critical zone developed over basalt lavas. Geosphere, 11(2), 514–
532. Available from: https://doi.org/10.1130/GES01142.1
Zelt, C.A. (1999) Modelling strategies and model assessment for wide-
angle seismic traveltime data. Geophysical Journal International,
139(1), 183–204. Available from: https://doi.org/10.1046/j.1365-
246X.1999.00934.x
Zelt, C.A. & Smith, R.B. (1992) Seismic traveltime inversion for 2-D crustal
velocity structure. Geophysical Journal International, 108(1), 16–34.
Available from: https://doi.org/10.1111/j.1365-246X.1992.
tb00836.x
How to cite this article: Trichandi, R., Bauer, K., Ryberg, T.,
Scherler, D., Bataille, K. & Krawczyk, C.M. (2022) Combined
seismic and borehole investigation of the deep granite
weathering structure—Santa Gracia Reserve case in Chile.
Earth Surface Processes and Landforms, 47(14), 3302–3316.
Available from: https://doi.org/10.1002/esp.5457
3316 TRICHANDI ET AL.
10969837, 2022, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.5457 by Technische Universitaet Berlin, Wiley Online Library on [09/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
