WEBVTT
Kind: captions
Language: en-US

00:00:00.392 --> 00:00:04.208
Hi, I’m Chris DuRoss with
the USGS in Golden, Colorado.

00:00:04.240 --> 00:00:08.000
I’m excited to share new research
on a portable optically stimulated

00:00:08.000 --> 00:00:11.253
luminescence age map
of a paleoseismic exposure.

00:00:11.253 --> 00:00:14.800
I’d like to acknowledge my co-authors,
Ryan Gold with the USGS in Golden,

00:00:14.800 --> 00:00:18.320
Colorado; Harrison Gray with the
USGS Luminescence Dating

00:00:18.320 --> 00:00:22.000
Laboratory in Denver, Colorado;
and Sylvia Nicovich with the

00:00:22.000 --> 00:00:25.416
Bureau of Reclamation
in Denver, Colorado.

00:00:25.440 --> 00:00:29.736
Okay, so, as those who have
worked with paleoseismic data know

00:00:29.760 --> 00:00:32.216
geochronological
data can be messy.

00:00:32.240 --> 00:00:35.680
Noisy age data can typically
relate to the depositional

00:00:35.680 --> 00:00:39.096
and geomorphic setting
of the trench or exposure.

00:00:39.120 --> 00:00:43.176
Mechanisms, either known or unknown,
of sediment transport and deposition.

00:00:43.200 --> 00:00:44.800
Age inheritance,
such as recycled

00:00:44.800 --> 00:00:48.696
charcoal, and post-depositional
modification, including burrowing.

00:00:48.720 --> 00:00:51.440
These factors lead to a poor
understanding of the contextual

00:00:51.440 --> 00:00:54.456
uncertainties of samples
and the resulting ages,

00:00:54.480 --> 00:00:58.000
inconsistent methods of constructing
Bayesian models, and over- and

00:00:58.000 --> 00:01:00.696
under-constrained models
of earthquake timing.

00:01:00.720 --> 00:01:04.320
This is illustrated in the conceptual
figure here, which plots earthquake

00:01:04.320 --> 00:01:07.496
timing uncertainty versus
the number of ages.

00:01:07.520 --> 00:01:10.160
This is simply meant to illustrate
that models with limited ages

00:01:10.160 --> 00:01:13.736
do a poor job of refining
earthquake timing uncertainties.

00:01:13.760 --> 00:01:16.400
For example, at some stages,
timing uncertainty may actually

00:01:16.400 --> 00:01:20.216
increase with an increase
in the number of samples.

00:01:21.120 --> 00:01:25.200
So considering these processes and
uncertainties, how do we develop

00:01:25.200 --> 00:01:30.160
sampling strategies for geochronology
that maximize the accuracy and,

00:01:30.160 --> 00:01:33.976
critically, the reproducibility of
paleoseismic rupture histories.

00:01:34.000 --> 00:01:37.040
In this presentation, I’ll argue
that portable optically stimulated

00:01:37.040 --> 00:01:40.240
luminescence,
or portable OSL, can help.

00:01:40.240 --> 00:01:43.840
I’ll go through the background of
the method, results from a field site,

00:01:43.840 --> 00:01:46.456
and end with some
practical applications.

00:01:46.480 --> 00:01:49.256
Okay, so back to
our conceptual diagram.

00:01:49.280 --> 00:01:52.880
This is where we want to be,
highlighted in yellow, constructing

00:01:52.880 --> 00:01:57.736
Bayesian models of earthquake timing
based on abundant age control.

00:01:57.760 --> 00:02:01.680
So what is portable OSL?
Portable OSL is the determination of

00:02:01.680 --> 00:02:06.240
bulk luminescence as measured in a
portable reader from small samples.

00:02:06.240 --> 00:02:09.040
The measurement is made without
the usual procedures used for

00:02:09.040 --> 00:02:12.400
traditional OSL, for example
separating quartz grains

00:02:12.400 --> 00:02:17.256
of a uniform grain size, and thus
does not yield absolute ages.

00:02:17.280 --> 00:02:21.440
However, with constant lithology,
dose rate across the exposure,

00:02:21.440 --> 00:02:24.080
and some pretreatment,
bulk luminescence can reveal

00:02:24.080 --> 00:02:28.696
relative age and also show zones
of partially bleached sediment.

00:02:28.720 --> 00:02:32.000
We applied these methods to a natural
exposure of the Wasatch Fault at

00:02:32.000 --> 00:02:35.760
the Deep Creek site in central Utah.
The figure in the upper right shows

00:02:35.760 --> 00:02:38.960
a slope shade map of the site with
the Wasatch Fault scarp on a

00:02:38.960 --> 00:02:43.360
Holocene alluvial fan surface marked
by black arrows. The red arrow marks

00:02:43.360 --> 00:02:46.456
a natural exposure of the fault
shown here in the photograph.

00:02:46.480 --> 00:02:49.360
At the site, several meters
of exposed alluvial fan gravel

00:02:49.360 --> 00:02:53.120
have been vertically displaced
about 1.8 meters by a single

00:02:53.120 --> 00:02:56.456
prehistoric surface-rupturing
earthquake.

00:02:56.480 --> 00:02:59.600
We studied the Wasatch Fault
exposure hidden behind the gray trap

00:02:59.600 --> 00:03:02.613
here and highlighted
by the white box.

00:03:03.280 --> 00:03:06.000
The exposure consisted
of faulted alluvial fan gravel

00:03:06.000 --> 00:03:09.096
with a 30-centimeter-thick
paleosol.

00:03:09.120 --> 00:03:12.560
The paleosol is faulted and buried
by about a 1.5-meter-thick

00:03:12.560 --> 00:03:16.160
scarp drive colluvial wedge.
The colluvial wedge consists of

00:03:16.160 --> 00:03:20.616
moderate- to steeply dipping silt,
sand, and gravel interbeds,

00:03:20.640 --> 00:03:22.936
shown there with
the dashed lines.

00:03:22.960 --> 00:03:26.240
Our principle goal was to test whether
we could generate an age map for

00:03:26.240 --> 00:03:31.256
the exposure using portal OSL.
So related questions included,

00:03:31.280 --> 00:03:35.440
are we able to resolve spatial
trends in the portable OSL data,

00:03:35.440 --> 00:03:39.496
how do sedimentary and pedogenic
processes affect the results,

00:03:39.520 --> 00:03:43.416
and do the data help us
understand earthquake timing.

00:03:43.440 --> 00:03:47.840
To help make sense of how portable
OSL data relate to age, we also extracted

00:03:47.840 --> 00:03:52.240
samples for radiocarbon and traditional
OSL dating, with the results shown here

00:03:52.240 --> 00:03:57.680
in thousands of years before present.
In total, 23 radiocarbon ages shown

00:03:57.680 --> 00:04:01.440
here by the black pluses,
and 11 traditional OSL ages shown by

00:04:01.440 --> 00:04:04.640
the white pluses, constrain the
timing of sediment deposition

00:04:04.640 --> 00:04:08.800
and soil formation at the site.
These results suggest that alluvial fan

00:04:08.800 --> 00:04:13.040
deposits at the site are as old as about
13,000 to 7,000 years in the footwall

00:04:13.040 --> 00:04:15.656
and 8,000 to 5,000 years
in the hanging wall.

00:04:15.680 --> 00:04:18.560
Paleosol formation occurred
between around 4,000 years and

00:04:18.560 --> 00:04:21.920
shortly after around 1,000 years.
With some exceptions that

00:04:21.920 --> 00:04:24.320
I’ll discuss in later slides,
colluvial wedge sediments

00:04:24.320 --> 00:04:27.736
were deposited mostly
in the past 1,000 years.

00:04:27.760 --> 00:04:31.360
With then extracted 342 samples
for portable OSL, shown here

00:04:31.360 --> 00:04:35.200
by the white circles. The samples
were arranged in a grid across the site

00:04:35.200 --> 00:04:39.416
without regard to the stratigraphic
contexts or faults.

00:04:39.440 --> 00:04:43.760
Just a few notes about these samples.
We collected small 3- to 4-centimeter-

00:04:43.760 --> 00:04:48.136
diameter samples using film
canisters in darkroom conditions.

00:04:48.160 --> 00:04:50.480
The samples were spaced
10 centimeters vertically and

00:04:50.480 --> 00:04:55.200
about 15 centimeters horizontally.
We sieved the sediment on-site,

00:04:55.200 --> 00:04:58.000
and off-site weighed similar-size
aliquots before measuring

00:04:58.000 --> 00:05:01.920
in the portable OSL reader.
We normalized the photon counts

00:05:01.920 --> 00:05:05.360
by sample weight because
of the extreme range in low to high

00:05:05.360 --> 00:05:09.256
OSL results, plot the
log of the photon counts.

00:05:09.280 --> 00:05:12.320
The portable OSL results are
colored here by luminescence.

00:05:12.320 --> 00:05:15.576
In this and later slides,
the color scheme will be the same,

00:05:15.600 --> 00:05:18.320
gradient from yellow to blue,
with yellow indicating the

00:05:18.320 --> 00:05:21.496
greatest luminescence signal
and blue the least.

00:05:21.520 --> 00:05:26.136
These results correspond well with the
radiocarbon and traditional OSL ages,

00:05:26.160 --> 00:05:30.720
thus we used a simple linear regression
shown here at the lower left to relate

00:05:30.720 --> 00:05:34.719
photon counts from the
portable OSL reader to age.

00:05:35.360 --> 00:05:39.176
Here’s a surface interpolation based
on the portable OSL results.

00:05:39.200 --> 00:05:41.440
Again, the color grades
from yellow to blue,

00:05:41.440 --> 00:05:45.096
showing decreasing
OSL signal and thus age.

00:05:45.120 --> 00:05:48.400
Here the radiocarbon and traditional
ages are also shown and colored in the

00:05:48.400 --> 00:05:51.920
same way as the portable OSL map,
and you can see the general agreement

00:05:51.920 --> 00:05:56.616
across the alluvial fan, paleosol,
and colluvial wedge.

00:05:56.640 --> 00:06:00.776
In the next series of slides, I’ll focus
just on the portable OSL age map

00:06:00.800 --> 00:06:04.696
and highlight fine to coarse gradients
in age across the exposure.

00:06:04.720 --> 00:06:08.640
For example, as expected, the alluvial
fan age of both the footwall and

00:06:08.640 --> 00:06:13.256
hanging wall decreases vertically,
but not horizontally.

00:06:13.280 --> 00:06:17.976
We also observe a strong age gradient
from the paleosol to the colluvial wedge

00:06:18.000 --> 00:06:23.157
and also across the buried free face
from the footwall to the hanging wall.

00:06:23.920 --> 00:06:27.976
However, we find more nuanced
age complexity in the paleosol.

00:06:28.000 --> 00:06:31.600
For example, on the right side,
adjacent to the fault and buried free face,

00:06:31.600 --> 00:06:35.736
the portable OSL map suggests
rapid burial of the paleosol.

00:06:35.760 --> 00:06:38.320
In comparison, at a more
fault-distal position,

00:06:38.320 --> 00:06:42.640
the map suggests continued soil
formation at the top of the paleosol

00:06:42.640 --> 00:06:46.296
and possibly after the
rupture may have occurred.

00:06:46.320 --> 00:06:49.360
Within the wedge, we note a prominent
area of poorly bleached sediment

00:06:49.360 --> 00:06:53.600
close to the fault and buried free face.
Sediment in this part of the wedge

00:06:53.600 --> 00:06:58.000
was likely not fully exposed to
sunlight during transport, either as

00:06:58.000 --> 00:07:01.920
the result of block formation
or a short transport path.

00:07:01.920 --> 00:07:04.800
Elsewhere within the wedge,
the age decreases laterally

00:07:04.800 --> 00:07:09.016
from the fault and vertically
toward the surface, as expected.

00:07:09.040 --> 00:07:11.920
I just wanted to quickly show that,
with these data, we can quantify

00:07:11.920 --> 00:07:15.920
horizontal and vertical trends
in the portable OSL data.

00:07:15.920 --> 00:07:18.640
In all cases shown here,
we’re plotting the model

00:07:18.640 --> 00:07:23.200
portable OSL ages or, that is,
the OSL photon counts converted to

00:07:23.200 --> 00:07:27.256
age using our linear
regression versus distance.

00:07:27.280 --> 00:07:31.280
The left column highlights horizontal
trends as portable OSL age is plotted

00:07:31.280 --> 00:07:35.520
against distance from the fault.
So 1 meter on the X axis indicates

00:07:35.520 --> 00:07:40.000
a meter outboard of the fault.
The data are shown in the upper figure,

00:07:40.000 --> 00:07:44.856
and in the lower figure, grouped
into 0.5-meter distance bins.

00:07:44.880 --> 00:07:48.136
Horizontal trends in age are especially
apparent in the lower figure.

00:07:48.160 --> 00:07:52.400
For example, the blue and orange arrows
show that the colluvium age decreases

00:07:52.400 --> 00:07:56.560
laterally with distance from the fault,
whereas, red and black arrows mark

00:07:56.560 --> 00:08:00.696
more laterally constant paleosol
and alluvial fan ages.

00:08:00.720 --> 00:08:03.920
In the right column, distance above
or below the paleosol is plotted

00:08:03.920 --> 00:08:08.880
against age to highlight vertical trends in
the data. In these figures, the paleosol is

00:08:08.880 --> 00:08:13.896
marked by the shaded area between
the solid and dashed and dotted lines.

00:08:13.920 --> 00:08:17.120
In these figures, 1 meter on
the Y axis is a meter above

00:08:17.120 --> 00:08:21.176
the top of the paleosol and
negative 1 meter is a meter below it.

00:08:21.200 --> 00:08:23.280
In the lower panel,
the portable OSL results

00:08:23.280 --> 00:08:27.896
are grouped into 0.1- to
0.3-meter distance bins.

00:08:27.920 --> 00:08:31.200
In this case, vertical trends
in portable OSL are apparent.

00:08:31.200 --> 00:08:34.960
For example, the red and black arrows
show that the alluvial fan and paleosol

00:08:34.960 --> 00:08:38.456
ages decrease consistently
toward the ground surface.

00:08:38.480 --> 00:08:41.520
In contrast, the blue arrow
shows that the colluvium age

00:08:41.520 --> 00:08:44.296
is mostly
constant vertically.

00:08:44.320 --> 00:08:47.920
Finally, we can use these results
to explore how the spatial age

00:08:47.920 --> 00:08:50.456
variability relates to
earthquake timing.

00:08:50.480 --> 00:08:53.496
Here we compare
two earthquake horizons.

00:08:53.520 --> 00:08:57.440
The first is the stratigraphic event
horizon, or stratigraphic position

00:08:57.440 --> 00:09:00.720
at which sediments are interpreted
as pre-dating or post-dating the

00:09:00.720 --> 00:09:03.576
earthquake, shown here
by the white arrows.

00:09:03.600 --> 00:09:07.760
We also define a chronologic event
horizon, or position at which sediments

00:09:07.760 --> 00:09:10.216
are older or younger
than the earthquake.

00:09:10.240 --> 00:09:14.000
Here we show three chronologic event
horizons based on Bayesian model

00:09:14.000 --> 00:09:18.160
variations with the mean earthquake
times converted to portable OSL age

00:09:18.160 --> 00:09:22.960
contours and shown on the age map.
We find that the portable OSL contours

00:09:22.960 --> 00:09:27.120
at around 500 to 700 years,
which are shaded gray and labeled

00:09:27.120 --> 00:09:30.240
as the apparent chronologic
event horizon, have the greatest

00:09:30.240 --> 00:09:33.736
spatial overlap with the
stratigraphic event horizon.

00:09:33.760 --> 00:09:36.880
However, notably, the stratigraphic
and chronologic event horizons

00:09:36.880 --> 00:09:40.400
diverge close to the fault.
This suggests that there are sedimentary

00:09:40.400 --> 00:09:44.160
facies within the fault-proximal
colluvial wedge that were not mappable

00:09:44.160 --> 00:09:47.280
from sedimentary texture alone.
Although these sediments are

00:09:47.280 --> 00:09:49.440
interpreted as younger than
the earthquake based on the

00:09:49.440 --> 00:09:52.720
stratigraphic event horizon,
the chronologic event horizon

00:09:52.720 --> 00:09:55.736
suggests a pre-earthquake
sediment age.

00:09:55.760 --> 00:09:58.880
This shows that portable OSL can
provide important chronological

00:09:58.880 --> 00:10:03.360
context for interpreting radiocarbon
and traditional OSL ages as well as

00:10:03.360 --> 00:10:07.520
Bayesian models of earthquake timing
and that these timing interpretations,

00:10:07.520 --> 00:10:12.056
based on stratigraphy alone,
can be misleading.

00:10:12.080 --> 00:10:15.280
Okay, at this point, I’d like to
shift gears and talk about practical

00:10:15.280 --> 00:10:18.800
applications of this method.
If you recall, in portable OSL,

00:10:18.800 --> 00:10:21.600
we were measuring luminescence
from bulk sediment samples.

00:10:21.600 --> 00:10:25.816
Because we don’t correct for organics,
grain size, or class lithology,

00:10:25.840 --> 00:10:29.416
it’s ideal to try to keep variations
in these factors to a minimum.

00:10:29.440 --> 00:10:33.656
So ideal site conditions would include
low variability in sediment texture,

00:10:33.680 --> 00:10:37.760
minimal post-depositional modification,
consistent dose rate across the

00:10:37.760 --> 00:10:41.736
exposure or information on how
dose rate changes spatially,

00:10:41.760 --> 00:10:45.576
and sediments that are generally
suitable for luminescence dating.

00:10:45.600 --> 00:10:48.400
Data density is also
an important aspect of this.

00:10:48.400 --> 00:10:51.040
For the Deep Creek data set,
we generated a 10-centimeter

00:10:51.040 --> 00:10:53.760
cell raster, however
lower-resolution rasters

00:10:53.760 --> 00:10:57.576
are also suitable for resolving
the spatial age relations.

00:10:57.600 --> 00:11:01.040
The figure here shows variations
in raster density from 10 centimeters

00:11:01.040 --> 00:11:04.480
in the upper left to 100 centimeters
in the lower right.

00:11:04.480 --> 00:11:07.680
We find that key details in the
portable OSL data, such as the strong

00:11:07.680 --> 00:11:11.680
age gradient from the alluvial fan,
the colluvial wedge sediments, and the

00:11:11.680 --> 00:11:14.560
area of poorly bleached sediment in
the fault-proximal part of the colluvial

00:11:14.560 --> 00:11:19.576
wedge are still resolvable with the 20-
to 47-centimeter cell rasters shown here.

00:11:19.600 --> 00:11:21.840
Thus, even in a lower
density field application,

00:11:21.840 --> 00:11:26.376
this method could be useful for
interpreting paleoseismic exposures.

00:11:26.400 --> 00:11:29.496
The last aspect to
consider is sampling strategy.

00:11:29.520 --> 00:11:32.480
Although our sampling density is
quite high for the Deep Creek site,

00:11:32.480 --> 00:11:36.320
we found the lower-density sampling
strategies also reproduce the primary

00:11:36.320 --> 00:11:39.976
spatial relations observed
in the full density data set.

00:11:40.000 --> 00:11:42.960
This figure shows portable
OSL rasters generated using

00:11:42.960 --> 00:11:46.400
subsets of the original data.
Points in black were included

00:11:46.400 --> 00:11:49.976
in the raster calculation,
and those in white were excluded.

00:11:50.000 --> 00:11:52.640
So, for example, in the upper
left-hand corner, the raster

00:11:52.640 --> 00:11:57.176
was generated using every
other point, or 50% of the data.

00:11:57.200 --> 00:12:01.040
For the Deep Creek site, rasters
generated using about 40 to 50%

00:12:01.040 --> 00:12:04.560
of the data reasonably
reproduce the spatial age relations

00:12:04.560 --> 00:12:07.465
apparent in the
original results.

00:12:08.320 --> 00:12:10.640
In general, sample transects
oriented normal to the

00:12:10.640 --> 00:12:13.416
stratigraphic contexts
are preferable.

00:12:13.440 --> 00:12:17.120
Again, this highlights that portable OSL,
even with a low-density sampling

00:12:17.120 --> 00:12:23.096
strategy of, say, 100 or so samples can
be used in a rapid field application.

00:12:23.120 --> 00:12:26.640
In conclusion, portable OSL
is an additional tool to consider

00:12:26.640 --> 00:12:30.480
in paleoseismic investigations.
At the Deep Creek site, the portable

00:12:30.480 --> 00:12:35.016
OSL age map yields insight into
how spatiotemporal age variability

00:12:35.040 --> 00:12:38.160
relates to sedimentary
and pedogenic processes.

00:12:38.160 --> 00:12:42.536
Our work suggests that stratigraphic and
chronologic event horizons may diverge,

00:12:42.560 --> 00:12:45.920
which has implications for
paleoseismic sampling strategies

00:12:45.920 --> 00:12:48.696
and Bayesian models
of earthquake timing.

00:12:48.720 --> 00:12:52.960
In summary, portable OSL, even in
a moderate-density field application,

00:12:52.960 --> 00:12:56.160
has the potential to provide insight
into pre- and post-earthquake

00:12:56.160 --> 00:12:59.200
depositional processes and
improve the accuracy

00:12:59.200 --> 00:13:02.800
of paleoseismic rupture histories.
And so finally, a huge thanks to

00:13:02.800 --> 00:13:06.456
Harrison Gray and Sylvia Nicovich
who spent many hours sampling.

00:13:06.480 --> 00:13:10.160
And I’m excited to announce that this
research was just released in Geology.

00:13:10.160 --> 00:13:11.040
Thank you.