WEBVTT
Kind: captions
Language: en-US

00:00:00.960 --> 00:00:04.080
Hi, everyone. I’m glad that
I can be with you all virtually.

00:00:04.080 --> 00:00:06.800
Thank you to the session organizers
for inviting me to give this talk

00:00:06.800 --> 00:00:10.000
on some of my Mendenhall
postdoc work using tree rings

00:00:10.000 --> 00:00:12.000
as a proxy and deriving
different types of records

00:00:12.000 --> 00:00:17.256
from trees to get precise dates
of past Cascadia seismicity.

00:00:17.280 --> 00:00:21.040
So paleoseismology studies really
do characterize the seismic hazards

00:00:21.040 --> 00:00:25.360
in Cascadia because we have not had
a Cascadia subduction zone event

00:00:25.360 --> 00:00:27.680
during the instrumental record.
So, when we’ve had seismometers

00:00:27.680 --> 00:00:29.600
out there, we actually
have not had an event.

00:00:29.600 --> 00:00:34.640
And, because of this, we are really
dependent on paleoseismic studies

00:00:34.640 --> 00:00:37.600
for our understanding of earthquake
recurrence as well as their associated

00:00:37.600 --> 00:00:42.000
hazards. And the next logical step
in Cascadia subduction zone

00:00:42.000 --> 00:00:44.640
paleoseismology is to
improve the geochronology.

00:00:44.640 --> 00:00:48.640
And this is really a key step because
we are starting to ask questions about

00:00:48.640 --> 00:00:51.760
how we are tying our terrestrial
paleoseismology records with our

00:00:51.760 --> 00:00:56.560
offshore paleoseismology records
to better understand any potential

00:00:56.560 --> 00:01:01.920
clustering we’re seeing in the event
beds as well as the real continuity along

00:01:01.920 --> 00:01:04.800
the Cascadia subduction zone
and along the Cascadia coast.

00:01:04.800 --> 00:01:11.520
So do we see a full extent in 1700?
Is 1700 a full rip event, if you will?

00:01:11.520 --> 00:01:15.840
Or what might we expect sequential
large-magnitude 8 earthquakes

00:01:15.840 --> 00:01:20.456
instead of one large magnitude 9?
So, by improving the geochronology,

00:01:20.480 --> 00:01:23.040
we can start to ask some of
these really important questions

00:01:23.040 --> 00:01:25.840
about Cascadia subduction
zone seismicity.

00:01:25.840 --> 00:01:31.200
So utilizing trees as a proxy for time
in the Cascadia subduction zone

00:01:31.200 --> 00:01:34.720
environment is not novel.
Back in the 1980s and 1990s,

00:01:34.720 --> 00:01:39.920
Gordon Jacoby, David Yamaguchi,
Brian Atwater, they used tree ring

00:01:39.920 --> 00:01:45.120
width records from trees that were
killed in the 1700 earthquake event

00:01:45.120 --> 00:01:49.440
to obtain a very precise
age of the 1700 earthquake.

00:01:49.440 --> 00:01:55.200
They utilized the tree ring width record,
and they compared that tree ring width

00:01:55.200 --> 00:01:58.160
record, which is like a barcode,
to the master chronology for the area.

00:01:58.160 --> 00:02:02.320
So they created a master chronology
from living trees, and they determined

00:02:02.320 --> 00:02:06.640
that the final year of the tree’s life
was 1699. And then, combining that

00:02:06.640 --> 00:02:11.760
piece of evidence with the stratigraphy
from the marshes as well as the written

00:02:11.760 --> 00:02:15.280
records from Japan across the Pacific
ocean and the oral – and the oral

00:02:15.280 --> 00:02:18.320
histories of the people who were
living in the region, they came up

00:02:18.320 --> 00:02:22.000
with the January 26, 1700, date.
And what you’re seeing in this figure

00:02:22.000 --> 00:02:27.040
here, and why we continue to return
to trees in Cascadia are some of the

00:02:27.040 --> 00:02:30.800
age probability functions you might
get from a radiocarbon age from

00:02:30.800 --> 00:02:34.800
the other types of paleoseismic
proxies in the Cascadia region.

00:02:34.800 --> 00:02:38.160
And so the length of the gray bar,
the paleo – or, excuse me,

00:02:38.160 --> 00:02:41.040
the probability age distribution.
You can see that, in turbidites,

00:02:41.040 --> 00:02:45.520
we have quite a large age probability.
We can get slightly improved ages

00:02:45.520 --> 00:02:48.720
in our terrestrial cores.
Even better ages if we have

00:02:48.720 --> 00:02:52.480
bracketing – radiocarbon
bracketed the event bed –

00:02:52.480 --> 00:02:55.840
the tsunami bed, for example.
But really the gold standard here is to be

00:02:55.840 --> 00:02:58.960
able to get to the year and to the season.
And one of the only ways you can do

00:02:58.960 --> 00:03:03.016
that in this environment is
to utilize tree rings themselves.

00:03:03.040 --> 00:03:07.040
So why are trees such an incredible
geochronometer for Cascadia?

00:03:07.040 --> 00:03:10.480
Well, it’s an abundant resource.
Trees are found along most of

00:03:10.480 --> 00:03:14.895
the entire Cascadia margin –
the coastline offshore.

00:03:14.895 --> 00:03:17.680
Or, excuse me, the coastline or the
Cascadia subduction zone is offshore.

00:03:17.680 --> 00:03:21.920
And they stay in one place their
entire lives and record the climate

00:03:21.920 --> 00:03:24.400
and ecology that they feel in
their annual growth ring.

00:03:24.400 --> 00:03:27.840
So, each and every year, a tree puts
on an annual growth ring that is that

00:03:27.840 --> 00:03:30.880
transition from the early wood, late –
that light-colored wood to the dark

00:03:30.880 --> 00:03:34.400
wood, that dark-colored wood.
And within the physical and chemical

00:03:34.400 --> 00:03:37.760
characteristics of these trees,
we can derive records.

00:03:37.760 --> 00:03:41.520
And the patterns of those records can
be matched with master records that

00:03:41.520 --> 00:03:46.160
will then precisely date that tree in time.
And so what’s been used in Cascadia

00:03:46.160 --> 00:03:49.760
primarily is tree ring width records.
So the actual width of one year to

00:03:49.760 --> 00:03:53.600
another, this is dictated by climate.
So it makes for good paleoclimate

00:03:53.600 --> 00:03:57.440
records. But, by utilizing the width
record, you can match a tree you

00:03:57.440 --> 00:04:00.560
might find in a paleoseismic site
with the local master chronology.

00:04:00.560 --> 00:04:03.896
So it is dependent on having
that master chronology.

00:04:03.920 --> 00:04:07.600
Now, we can also look at stable isotope
records that we can derive from trees.

00:04:07.600 --> 00:04:10.776
I’m not going to talk about this in this
talk, but I think it’s worth mentioning.

00:04:10.800 --> 00:04:14.400
Because it is a real frontier for
utilizing trees. They found that there

00:04:14.400 --> 00:04:19.920
was a large jump in carbon isotopes
after the tsunami in Japan.

00:04:19.920 --> 00:04:23.360
The trees actually have water stress, and
that is reflected in their carbon isotopes.

00:04:23.360 --> 00:04:28.560
So this is a potential frontier to see the
extent of tsunamis as well as thinking

00:04:28.560 --> 00:04:32.376
about the ecological impacts
of large flooding events.

00:04:32.400 --> 00:04:35.120
And another type of record we
can derive from the annual growth

00:04:35.120 --> 00:04:37.680
rings are radiogenic isotopes.
And I am going to talk about that in

00:04:37.680 --> 00:04:42.400
this talk, where we can actually pull
radiogenic isotopes like carbon-14

00:04:42.400 --> 00:04:46.560
from the rings themselves and get this
annually resolved radiocarbon record.

00:04:46.560 --> 00:04:49.840
This greatly reduces the uncertainty
in the age if you don’t have a master

00:04:49.840 --> 00:04:53.920
chronology, but we can also utilize
that annual radiocarbon record to

00:04:53.920 --> 00:04:57.680
get even uncertainties of zero.
And I’ll talk about that in a few slides.

00:04:57.680 --> 00:05:01.600
So the over-arching message here is,
we can extract multiple types of

00:05:01.600 --> 00:05:05.040
records from trees that all
ultimately assist in their dating.

00:05:05.040 --> 00:05:09.360
and the reason why we want to think
about multiple ways to use trees here

00:05:09.360 --> 00:05:12.080
is because, as I mentioned,
dendrochronology and using trees

00:05:12.080 --> 00:05:14.880
really does become the
gold standard in Cascadia.

00:05:14.880 --> 00:05:22.240
So we have radiocarbon as a potential
to date back to a few – 50,000 years

00:05:22.240 --> 00:05:25.360
or so if you really want to push it.
And dendrochronology has the ability

00:05:25.360 --> 00:05:28.960
to date events for thousands of years.
So the potential is there.

00:05:28.960 --> 00:05:33.280
Unfortunately, the length of the local
master chronologies in Cascadia really

00:05:33.280 --> 00:05:36.056
don’t extend much further
back than 500 years or so.

00:05:36.080 --> 00:05:40.480
But the precision on
dendrochronology is zero uncertainty,

00:05:40.480 --> 00:05:42.856
and so that’s really
what we want to go for.

00:05:42.880 --> 00:05:47.280
Now, the way that I’ve been utilizing
radiocarbon and trees to get a precise

00:05:47.280 --> 00:05:51.760
age is I’ve been utilizing these global
radiocarbon excursion events.

00:05:51.760 --> 00:05:54.960
And so this allows us to exactly
date a tree ring record without

00:05:54.960 --> 00:05:57.760
a master chronology.
And what they are are large jumps

00:05:57.760 --> 00:06:00.240
in radiocarbon in the atmosphere
in both the northern and southern

00:06:00.240 --> 00:06:05.200
hemisphere that occur at a very
specific time. And so this an example.

00:06:05.200 --> 00:06:09.760
This is the large jump
between the year 774 and 775.

00:06:09.760 --> 00:06:13.200
This jump is about 20 times more
than we might expect to attribute to

00:06:13.200 --> 00:06:17.760
any normal solar modulation.
And so this large jump in delta-14-C,

00:06:17.760 --> 00:06:21.360
if you capture it in your annual tree
rings records, you have then precisely

00:06:21.360 --> 00:06:23.840
dated that sample in time without
the use of a master chronology.

00:06:23.840 --> 00:06:28.400
This is very important if we want to
go back further in time than just 1700.

00:06:28.400 --> 00:06:31.040
And so I’m going to now show
an example of how we utilize this

00:06:31.040 --> 00:06:34.480
multi-proxy approach at paleoseismic
sites in the Puget Lowlands.

00:06:34.480 --> 00:06:37.360
So the Puget Lowlands are up
in the northern part of the

00:06:37.360 --> 00:06:39.920
Cascadia subduction zone.
I’m going to zoom in here.

00:06:39.920 --> 00:06:43.600
And all of the different sites that
we are working with for this study

00:06:43.600 --> 00:06:47.200
are these small black dots that
are highlighted in yellow.

00:06:47.200 --> 00:06:50.240
So there are multiple sites –
paleoseismic sites across the

00:06:50.240 --> 00:06:54.480
Puget Lowlands that all dated within
radiocarbon error of each other.

00:06:54.480 --> 00:06:57.200
So they all dated to
about 1,000 years ago or so.

00:06:57.200 --> 00:06:59.840
And the real question was,
were all these events triggered

00:06:59.840 --> 00:07:02.720
by the same earthquake?
Were they potentially a clustering

00:07:02.720 --> 00:07:05.336
of earthquakes that happened
within a few years of each other?

00:07:05.360 --> 00:07:08.400
And what are the really underlying
fault mechanics of this if it was

00:07:08.400 --> 00:07:11.816
one large event or even
just a series of events?

00:07:11.840 --> 00:07:14.880
And so the first step was to take
the trees that were killed at these

00:07:14.880 --> 00:07:18.240
paleoseismic sites and to look at
the tree ring width records to

00:07:18.240 --> 00:07:21.920
see if they dated with each other’s.
And what we found – so this is

00:07:21.920 --> 00:07:24.560
an example from three of the sites
that had some of the longest tree rings

00:07:24.560 --> 00:07:28.960
records in the Puget Lowlands –
is that the trees themselves were coeval.

00:07:28.960 --> 00:07:32.000
And not only were they coeval,
but they also all died at the same time.

00:07:32.000 --> 00:07:35.040
So they all died in the same year.
Now, once again, our master

00:07:35.040 --> 00:07:38.240
chronologies didn’t go back further
enough to actually say exactly which

00:07:38.240 --> 00:07:42.160
year, but we knew that they
all dated within the same year.

00:07:42.160 --> 00:07:44.960
And so this is where we turn to that
annual radiocarbon measurement now.

00:07:44.960 --> 00:07:48.080
So, sure enough, we found and
captured, within our annual

00:07:48.080 --> 00:07:52.160
radiocarbon measurements,
this large jump from 774 to 775.

00:07:52.160 --> 00:07:54.800
That exactly placed these
tree ring records in time.

00:07:54.800 --> 00:07:58.560
And we can now say that
all of these trees died between

00:07:58.560 --> 00:08:02.160
the years 923 and 924. So they
had a full complete growing season.

00:08:02.160 --> 00:08:06.320
They went dormant in the fall of 923.
They did not start growing again in the

00:08:06.320 --> 00:08:11.520
year 924, so they were killed by the
landslide, the tsunami, or actually the

00:08:11.520 --> 00:08:15.840
drowning from a fault that dammed
up a creek in the same season.

00:08:15.840 --> 00:08:20.080
So, if we now go back to that map,
all of these sites that I have circled

00:08:20.080 --> 00:08:24.720
in red, they are all sites that have trees
in them that died in that season

00:08:24.720 --> 00:08:28.720
in that year at paleoseismic sites,
so sites that have clear other indications

00:08:28.720 --> 00:08:31.920
that this was an earthquake event.
So this is pretty dramatic.

00:08:31.920 --> 00:08:35.336
These are events that span across
the entire Puget Lowlands area

00:08:35.360 --> 00:08:38.320
across multiple different faults.
And, importantly, we also found that

00:08:38.320 --> 00:08:42.400
there are sites, including the one that’s
circled in blue here, that do not date

00:08:42.400 --> 00:08:47.440
to that same – that same 923 to 924.
So we are able – we have to now think

00:08:47.440 --> 00:08:52.000
of other ways to – other underlying
mechanisms to invoke the subsidence

00:08:52.000 --> 00:08:53.680
that we saw at Little
Skookum Inlet down here.

00:08:53.680 --> 00:08:56.720
So thinking about other types
of fault dynamics for the region.

00:08:56.720 --> 00:09:00.800
So that’s one example of the power of
dendrochronology – power of tree rings

00:09:00.800 --> 00:09:04.240
and radiocarbon and how we can utilize
them together to get very, very precise

00:09:04.240 --> 00:09:08.160
ages, down to the season, and then
thinking more broadly about what this

00:09:08.160 --> 00:09:11.600
means for interactions of faults and
the potential hazards they might pose.

00:09:11.600 --> 00:09:13.840
So where can these
techniques be applied?

00:09:13.840 --> 00:09:17.040
Well, obviously the low-hanging fruit
is to continue work along the Cascadia

00:09:17.040 --> 00:09:21.200
subduction zone along the coast,
looking again at 1700 with a closer look

00:09:21.200 --> 00:09:24.480
at the more southern sites outside of
western Washington, but also older

00:09:24.480 --> 00:09:28.400
Cascadia subduction zone events.
And I’ll go down to northern California

00:09:28.400 --> 00:09:31.920
in honor of this particular workshop
to talk about some of the projects

00:09:31.920 --> 00:09:35.440
we’re doing in northern California
to get at this question of, can we use

00:09:35.440 --> 00:09:37.680
trees to date earthquakes
in northern California?

00:09:37.680 --> 00:09:40.800
And specifically, can we get
at the penultimate Cascadia subduction

00:09:40.800 --> 00:09:45.360
zone earthquake event?
So this is a figure from Erin’s paper

00:09:45.360 --> 00:09:48.960
that shows that the shaking that
we might expect from a magnitude 9

00:09:48.960 --> 00:09:52.560
Cascadia subduction zone earthquake
in the area that I will be focusing in on

00:09:52.560 --> 00:09:55.680
in Humboldt County in northern
California is still expected to be

00:09:55.680 --> 00:09:58.160
very strong and severe.
And so the implications of very

00:09:58.160 --> 00:10:02.320
strong and severe shaking in this
landscape of these – on these

00:10:02.320 --> 00:10:05.280
mountainous landscapes and these
coastal mountains could potentially

00:10:05.280 --> 00:10:08.000
trigger massive landslide
and debris avalanches.

00:10:08.000 --> 00:10:13.600
And so this is actually a Lidar image
of two debris avalanches that are very

00:10:13.600 --> 00:10:17.360
close to each other that we visited this
past summer and took samples from.

00:10:17.360 --> 00:10:20.240
There’s the Red Lassic debris
avalanche that we are attributing

00:10:20.240 --> 00:10:23.416
to the penultimate Cascadia
subduction zone event.

00:10:23.440 --> 00:10:28.720
The radiocarbon ages so far are
tantalizingly close to other dates that

00:10:28.720 --> 00:10:32.400
we have for the penultimate event.
And as well as the Mule Slide debris

00:10:32.400 --> 00:10:35.200
avalanche, which is a much younger
event that we are thinking might

00:10:35.200 --> 00:10:37.680
actually be attributed to the
1906 earthquake.

00:10:37.680 --> 00:10:41.120
So not Cascadia subduction zone, but
still potentially seismically triggered.

00:10:41.120 --> 00:10:44.936
So what do these two landslides,
or debris avalanches look like?

00:10:44.960 --> 00:10:49.120
Mule Slide – this is a photo taken
from far away. It’s not re-vegetated.

00:10:49.120 --> 00:10:52.720
This is another reason why
we think it might be a younger event

00:10:52.720 --> 00:10:55.336
because there isn’t
vegetation across the slide.

00:10:55.360 --> 00:11:01.200
It almost looks like entire blocks
from the land slid down the slope.

00:11:01.200 --> 00:11:05.680
And you can actually see trees –
relatively old living trees –

00:11:05.680 --> 00:11:09.600
and I can tell you this because
I sampled them, that are intact on these

00:11:09.600 --> 00:11:12.720
blocks that have slid down the slope.
And so these are the little trees

00:11:12.720 --> 00:11:14.560
you see right here.
It looks like a block landslide.

00:11:14.560 --> 00:11:18.560
And so the idea was, can we utilize
the stress indicators of the trees to get

00:11:18.560 --> 00:11:21.816
a precise age on potentially when
this block landslide occurred?

00:11:21.840 --> 00:11:23.520
And so there’s a couple
ways we went at this.

00:11:23.520 --> 00:11:26.800
We took some samples from
dead trees with our chainsaw.

00:11:26.800 --> 00:11:31.920
And what you’re seeing here is a sample
of a tree that clearly grew on a slope.

00:11:31.920 --> 00:11:33.920
This will illustrate to you the
point that I’m trying to get at.

00:11:33.920 --> 00:11:37.280
And you can see the tree is
putting more wood on the

00:11:37.280 --> 00:11:41.096
downslope side to right itself up.
This is something that conifers do.

00:11:41.120 --> 00:11:43.920
Hardwood trees actually pull
themselves up – put tension wood

00:11:43.920 --> 00:11:47.840
on the uphill slope side.
But soft wood, conifers, they’ll put

00:11:47.840 --> 00:11:50.480
more wood on the downslope side.
So the idea is you can potentially

00:11:50.480 --> 00:11:54.240
start dating when this slide occurred
by dating when you start to see

00:11:54.240 --> 00:11:56.872
that reaction
wood in the tree.

00:11:57.680 --> 00:12:02.240
you can also look at the trees that are
surviving in the landslide environment

00:12:02.240 --> 00:12:05.280
and look at the indicators to see
if there’s any other types of clues

00:12:05.280 --> 00:12:07.760
we can derive potentially
when this slide occurred.

00:12:07.760 --> 00:12:12.000
And so, for example, this is a tree
we named The Ship, where the

00:12:12.000 --> 00:12:17.680
tree itself clearly is not righted.
But it has two large trunks, stems,

00:12:17.680 --> 00:12:20.400
growing out of it that
are in proper position.

00:12:20.400 --> 00:12:24.376
So they are growing towards
the sun, righted themselves.

00:12:24.400 --> 00:12:27.760
And we can date those stems.
And those stems date to the 1930s.

00:12:27.760 --> 00:12:32.000
So we know that the slide
has to have pre-dated 1930s.

00:12:32.000 --> 00:12:36.080
We can actually take cores of the tree
itself and also start to zoom in on the

00:12:36.080 --> 00:12:40.160
stress and the reaction wood of the
trees to see if we can get a clearer date.

00:12:40.160 --> 00:12:45.120
And not all the trees are agreeing at
this point, but that particular tree that

00:12:45.120 --> 00:12:48.560
I showed you, The Ship, we will see
that these numbers are pretty small,

00:12:48.560 --> 00:12:51.200
that this is the year 1900.
This is the year 1920.

00:12:51.200 --> 00:12:55.040
We see quite a bit of suppression
around the year 1906 and 1907.

00:12:55.040 --> 00:12:58.400
And so this does agree with the idea
that this landslide could have been

00:12:58.400 --> 00:13:02.160
triggered by the
1906 earthquake.

00:13:02.160 --> 00:13:05.520
The tree, then, is severely stressed
for the next five or so years.

00:13:05.520 --> 00:13:10.616
So we’re going at this question
with multiple types of evidence.

00:13:10.640 --> 00:13:14.560
Now, the other landslide we’re
looking at is the Red Lassic slide.

00:13:14.560 --> 00:13:18.480
And this is a slide that we are thinking
might be attributed to the penultimate

00:13:18.480 --> 00:13:22.640
quake because we have radiocarbon
dates from one of the trees that was

00:13:22.640 --> 00:13:27.600
buried in this avalanche debris
that dates to around the years

00:13:27.600 --> 00:13:29.200
of the penultimate.
And so we went back,

00:13:29.200 --> 00:13:33.416
and we found some of these bark-
bearing trees in this avalanche debris.

00:13:33.440 --> 00:13:37.040
This is a picture of Harvey and Steve
taking a cross-cutting saw to it.

00:13:37.040 --> 00:13:41.520
And we – nice samples from these trees.
So there’s quite a few rings in

00:13:41.520 --> 00:13:44.720
the samples we were able to take.
And so the next step now is to start to

00:13:44.720 --> 00:13:47.760
do the carbon-14 wiggle-matching
process to see if we can potentially

00:13:47.760 --> 00:13:51.760
capture one of the jumps and exactly
date this tree that might give us a target

00:13:51.760 --> 00:13:55.040
for the penultimate earthquake event.
And so we’re still waiting on

00:13:55.040 --> 00:13:57.200
these radiocarbon dates,
but we’re really excited to see

00:13:57.200 --> 00:13:59.360
what comes out of
this particular study.

00:13:59.360 --> 00:14:03.520
So, in closing, dendrochronology
can address longstanding

00:14:03.520 --> 00:14:05.680
paleoseismic questions
in Cascadia.

00:14:05.680 --> 00:14:08.400
There are multiple proxies we can
derive from tree rings records, and they

00:14:08.400 --> 00:14:11.600
can answer a wide variety of questions –
everything from when did the event

00:14:11.600 --> 00:14:15.360
occur, but also potential tsunami
inundation and ecological resilience

00:14:15.360 --> 00:14:18.056
in the face of some of these
large landscape changes.

00:14:18.080 --> 00:14:20.640
High-precision radiocarbon,
combined with dendrochronology,

00:14:20.640 --> 00:14:25.120
will definitely improve ages on other
earthquakes happening in the Cascadia

00:14:25.120 --> 00:14:28.000
region including the penultimate
Cascadia subduction zone earthquake.

00:14:28.000 --> 00:14:30.800
And the preliminary dates we’re getting
from northern California avalanche

00:14:30.800 --> 00:14:35.255
debris indicate that earthquakes might
be responsible for the largest slides

00:14:35.280 --> 00:14:38.240
we see in the area and some of the
smaller slides we might be able to

00:14:38.240 --> 00:14:41.040
attribute to large weather events.
But these very, very large,

00:14:41.040 --> 00:14:46.880
deep-seated avalanches – or, excuse me,
debris avalanches seem to be tied

00:14:46.880 --> 00:14:50.240
to large earthquake events based
upon the dendrochronology so far.

00:14:50.240 --> 00:14:55.141
So, with that, I would like to thank
you all very much, and I [audio cuts out]