Rockfall, Cliff Retreat in Yosemite Valley since Last Glacial Maximum

Video Transcript
Download Video
Right-click and save to download

Detailed Description

The granitic cliffs of Yosemite Valley produce frequent rockfalls, modifying the landscape but also posing risk to park visitors. Analyses of terrestrial lidar and historical structure-from-motion photogrammetry data provide relatively precise short-term (approximately 40 years) rates of rockfall and cliff retreat. Those same data can also be used to estimate long-term (postglacial) rates through analysis of talus accumulation. Comparison of short- and long-term rates reveals complex spatial and temporal patterns of rockfall and provides a broader context for evaluating modern hazard conditions.

Stock G (2021) Pace of rockfalls and cliff retreat in Yosemite Valley since the Last Glacial Maximum. USGS Landslide Hazards Program Seminar Series, 27 October 2021.

Details

Date Taken:

Length: 00:40:28

Location Taken: US

Video Credits

 

Video thumbnail courtesy of Greg Stock, National Park Service.

Transcript

[silence]

Okay, folks. Well, I can see we’re just
about to hit one minute after the hour,

so we’re going to
go ahead and get started.

My name is Matt Thomas,
and thank you for tuning in

to the USGS Landslide Hazards
Program seminar series.

For those of you that are new to this
meeting, you have the ability to submit

questions via the chat window or to
use the raise-your-hand feature

in combination with your
microphone and video camera.

We’re going to wait until the end of
the presentation to take questions.

So, in the meantime, please just do
your best to make sure your

microphone is muting – is muted
when you’re not intending to speak.

Brian, thank you for introducing
today’s speaker. I’ll turn it over to you.

- Okay, great. Hi, everyone.
It is definitely my distinct pleasure

to introduce Greg Stock.
He is the national park geologist

in Yosemite in California.
He is a lifelong Californian.

And he might deny this,
but from my perspective,

he has a comprehensive understanding
of nearly all things Sierra Nevada.

[voice in background]
I’m sure he’d say there’s some topics

he doesn’t have full focus on, but,
you know, flora, fauna, geology,

geomorphology, glaciology. He
answers all of my questions and more.

Greg was educated at Humboldt State
and UC-Santa Cruz working with

Bob Anderson on the landscape
evolution of the Sierra Nevada

and did a lot of cosmogenic
nuclide studies of cave sediments

related to the tilting
of the mountain range.

He then took a Turner postdoctoral
fellowship at the University of

Michigan, where he did more
work with cosmogenic nuclides.

And then, since 2006, he moved back
to California and has been in Yosemite

working on just a really wide range
and incredible number of cutting-edge

geomorphological topics, including
rockfall triggering, hazard and risk

assessments – and that’s what he’ll
be sharing information with today,

rock avalanche dating and
reconstruction, Pleistocene glacial

mapping that included finishing a
mapping product that was left by

Clyde Wahrhaftig at the USGS,
so that was really a strong and

really important contribution
and showcases some of the ties

that Greg has between the National
Park Service and the USGS.

Greg studied glacial retreat and
climate change, effects of

post-wildfire debris flows.
He’s actually in King’s Canyon

Sequoia National Park now
dealing with some of those issues.

He studies fluvial sediment delivery.
And the list goes on.

These topics have led to him being
collaborators with a host of high-profile

partners all over the world.
And he’s been sought out to provide

scientific advice for projects in other
national parks, both in the United States

and elsewhere – in Switzerland.
He’s worked in Indonesia as well.

Given all these contributions, I hope
that you’ll agree that he is a very –

we’re lucky to have him talk today.
I consider him – I think a lot of us

consider him sort of an
adopted member of the USGS.

And I’ve been lucky to have him as
a colleague and friend for the past

15 years. And so, with that, I’ll turn
it over to Greg. Thank you.

- Great. Thanks, Brian.
That was really nice.

And thanks, everyone,
for joining today.

I’m going to be talking about rockfalls
in Yosemite Valley – not surprising.

And, in particular,
I’m going to be taking kind of

a geomorphic angle on this.
As Brian mentioned, a lot of the

work that I do in Yosemite is focused
specifically on rockfall triggering,

especially as it relates to
hazard and risk.

That’s clearly the information that
the park is most interested in.

But the fact that rockfalls are so
common in Yosemite, and the fact

that it’s been a national park for
a really long time – it’s arguably the

first national park, certainly in
the U.S., arguably around the world,

with a really long record
of information about rockfalls.

So today I’m going to be kind of taking
the longer view – a short view and

a longer view on rockfalls and
landscape evolution in Yosemite.

So this dramatic photo is –
I think illustrates the issues

of rockfalls in Yosemite.
I’m going to be talking about

this particular rockfall
at length today.

This was actually one of a series of
rockfalls that came off of El Capitan

in September of 2017.
And this is the kind of rockfall

that creates a lot of
headaches for the park service.

This one in particular
was really challenging.

In the sequence of rockfalls,
there was one fatality, two injuries,

and the closure of the main road
leading out of Yosemite Valley.

So quite a bit of hazard and
risk associated with this event.

But it also moved something like
10,000 cubic meters of rock

in a matter of seconds.
So, from a geomorphic standpoint,

there was quite a bit of work that was
accomplished by this single rockfall.

And that brings to mind, you know,
some questions about the role that

rockfalls play in evolving this iconic
landscape of Yosemite Valley.

So briefly, just to touch on this,
rockfalls are really common.

I mean, I think it’s safe to say
that they really are the dominant

geologic process operating
in Yosemite Valley since

glaciers retreated about
16,000 years ago.

We have a database that I’ll tell
you more about in a moment.

But we have something like 1,500
rockfalls and other mass-wasting

events, like rockslides,
debris flows, etc., recorded since 1857.

The vast majority of those events
are rockfalls in Yosemite Valley.

So I’ll pretty much just be
focusing on rockfalls today.

Since I started in 2006, and we’ve
been keeping pretty careful track

of rockfalls, there’s one event
somewhere in Yosemite

about every five days or so.
Again, usually in Yosemite Valley.

The most recent one
was yesterday at 1:00 p.m.

So, again, it makes it a nice laboratory
for studying rockfalls because we have

a large – a large data set to work with
and new data coming in all the time.

But, again, there is hazard and risk
associated with these rockfalls.

There have been 16 fatalities
and more than 100 injuries

from rockfalls in Yosemite
going back to 1857.

So, for the purposes of this talk, I’ve
got three motivating questions, right?

Again, these are sort of
geomorphic in nature.

And they are related to hazard
in the sense that, if we – if our

historical record, and even our more
recent observations of rockfall activity,

are fairly indicative of long-term
geologic rates, then we have

confidence sort of
extrapolating out what we think

the hazard is
in certain areas.

If not – if, say, the modern rockfall
rates are anomalously low or

anomalously high, you know, that
can affect how we view the hazard.

So the first thing we want to do is just
see how those short- and long-term

rates of rockfall expressed as
cliff retreat, or cliff erosion rates –

how do they compare?

Are these modern rates representative
of longer-term rates going all the

way back to just after
the Last Glacial Maximum?

And also, how important are the largest
rockfalls to cliff erosion rates?

We know that there’s a lot of hazards
associated with those events,

although they are infrequent,
but what role do they play in the

long-term landscape evolution?
So we’re going to be looking at two

cliffs – El Capitan and Middle Brother.
These are sort of our two study areas.

And these cliffs are basically side
by side on the north wall –

south-facing wall
of Yosemite Valley.

And basically, they’re made up
of just about identical rock types.

These are all some variant
of granite or granodiorite.

They have very similar relief –
about 900 meters tall.

The cliff areas are a little bit different.
El Capitan is a smaller cliff area.

It’s actually quite a bit steeper.
It’s even overhanging in some places.

Lower Brother is a little bit lower angle
and has more of a stair-step shape,

which gives it a
slightly larger cliff area.

But, by and large, I would argue
that these two study cliffs are

actually quite similar and
therefore a really good basis

for comparing rates
against one another.

As far as the data sources that we’re
going to use for this analysis,

we’re using a couple
of different things.

First of all, we have
a rockfall inventory database.

Again, that goes from 1857,
and we’ve updated it to 2020.

We’ll be using aerial
structure-from-motion photogrammetry,

and I’ll tell you more about that,
but these are photographs that

were taken from a helicopter for
search-and-rescue purposes in 1976.

And – as well as some more
recent structure-from-motion –

terrestrial structure-
from-motion data.

And then terrestrial Lidar
data starting in 2010 up to 2016.

As well as geologic maps
and field observations

that we’ve made
as part of this study.

So let me first tell you about
the rockfall inventory database.

Again, it records something like 1,500
rockfalls and other events in Yosemite.

Each event contains information
about the location, timing, volume,

and potential trigger for the rockfall
as well as a narrative description.

And it’s not quite out yet, but within
a year, we’ll actually be probably

re-publishing the database. It was last
published in 2013 with data up to 2011.

Next year, we’ll be publishing it
up to 2020, and it will be in

a different format. Rather than
kind of a tedious spreadsheet,

we’re going to have a web-based
interactive GIS-type map that people

can use to investigate
each rockfall in detail.

There’s a lot of great things that I could
say about the inventory database.

I actually think it’s probably
one of the very best in the world.

But I actually kind of want to focus
more on the downsides of the database

for this talk, and the main one is
that it’s really prone to

human observation bias.
The vast majority of events that

are recorded in the database,
because we demand such specific

information about location and timing,
are based on human observations –

eyewitness-type accounts.
And that’s pretty good for

Yosemite Valley, especially
in the summer months.

But it’s – you know, it’s not so good
at 3:00 a.m. in February, right?

There’s just fewer people around.
There are a lot of rockfalls that

don’t get reported for that reason.
They tend to be smaller ones.

The larger-volume rockfalls,
people tend to notice,

and so the reporting
is better for those.

One of the things that’s kind of
lacking in the database is

a precise estimate of volume, right?
Most of the volumes were just kind of

spitballed, especially back in the day.
And so the very first iteration of this

database that was published by Gerry
Wieczorek of the USGS in 1992 took

sort of an order of magnitude approach
so that rockfalls were recorded as,

say, 20 cubic meters,
200 cubic meters, etc.

And what that tends to do is
over-represent those specific volumes

compared to the other volumes that
are determined more accurately.

And, when it comes to looking at
the volume frequency distributions,

you see these frequency
jumps at those volumes

because there’s a lot of
data clustered in those bins.

So these will all – these are all things
that sort of come up when

we’re analyzing the database.
And it’s actually why, when I first

started here, it became clear that
it would be nice to have a way of

remotely detecting rockfalls so that
we don’t demand eyewitness accounts

or human observations all the time.
So actually starting in about 2006

and 2007, we did our first terrestrial
Lidar scans of the valley using

kind of primitive scanners that could
just barely get data for the cliffs

a kilometer and a half away.
That’s no problem now.

And, beginning in 2010, we’ve done
quite a bit of terrestrial scanning.

At this point, we have just about
every square meter of Yosemite Valley

scanned at least once,
and many of the cliffs,

like El Capitan, have been
scanned six or eight times.

We’ve also adopted structure-from-
motion photogrammetry more recently.

And sometimes this is helicopter-based,
as I’ll be showing you later. Sometimes

it’s just on the ground, but we do have
quite a bit of SfM data now as well.

Okay. So let me tell you about sort of
the signature rockfalls for this talk.

These were rockfalls that occurred
on the 27th and 28th of September

in 2017, and they came off the
southeast face of El Capitan.

They were large and dramatic events,
and in late September, there were

a lot of people in Yosemite Valley.
And, in fact, late September is sort of

peak climbing season
for the big walls in Yosemite.

I’m guessing there were probably
at least 100 climbers on El Capitan

on the days that these
rockfalls happened.

And they had –
they were consequential.

The very first rockfall in this sequence
occurred in the middle of the day

on the 27th of September.
There happened to be two climbers

walking along the base of the cliff
directly underneath this rockfall.

And one of those climbers was killed as
he was actually protecting his partner –

shielding her from falling rocks.
And she was also seriously injured.

And, after that initial event,
reports starting coming into

our park dispatch about it.
There were another six rockfalls

that occurred from that same location
over the next five or six hours.

So our search-and-rescue team and
our protection rangers and myself

were involved in some pretty stressful
decisions about how to go in with

helicopter and with ground crews
and try to extract the injured climber

and do a body recovery in the
midst of these ongoing rockfalls.

And fortunately,
that all worked out okay.

Everything was successful in that
regard, but it was still a really sad day.

And, at the end of the day,
it wasn’t exactly clear to me

what would continue to happen.
So, the following morning,

started taking a good look at it.
And also starting to look at the initial

data that came in when there was
another rockfall from that same

location that was quite a bit larger than
anything that had happened before.

As I recall, it was something like –
at least 10 to 12 times larger.

So it was the largest rockfall
of the whole sequence.

It was about 10,000
cubic meters in size.

And this rockfall was large
enough that – actually, in that

upper-right photo, you can see
a large dust cloud and sort of

sandy cloud of rock debris racing
across the floor of Yosemite Valley.

It covered Northside Drive,
which is the one-way road

leading out of Yosemite Valley.
And there was one rock fragment

about the size of a football that had
just the right trajectory to come through

the sunroof of a moving vehicle
and hit the driver in the head.

And he was injured.
He is okay now, but it was –

it was a significant injury,
and he had to be helicoptered out.

And we closed the road.
So, at the end of this two-day period,

we had one injury – or, sorry – two
injuries, one fatality, and a closed road.

And I, at that point, called Brian
[chuckles] and officially asked for

USGS assistance,
and Brian came right up to assist.

So, fortunately, we were able to
respond to these rockfalls rather

quickly compared to in previous years,
primarily because we had

a lot of baseline data already.
So, for example, we had taken

very high-resolution photographs
of the cliffs. We were able to

re-occupy those positions.
It’s pretty obvious, in this case,

where the rockfalls
were happening.

Note not just the size of that rockfall
scar, but the big runout path,

and then the new area of fresh talus
at the base of the cliff, right?

That’s kind of a
landscape-changing event.

I mean, El Capitan looks
different now after that rockfall,

and it will forevermore.

We also were able to sort of
coincidentally pull off some repeat

photographs from a helicopter
before and after the biggest of the

event on the 28th of September.
So the photograph on the left there

shows the source area for the smaller
rockfalls on the 27th of September.

And I’ve also shown some rock
climbers on the cliff for scale there.

And then, the following day,
you can see the very large failure

that occurred sort of starting
from the bottom of those rockfall scars

the previous day
and propagating upwards.

I was up in a helicopter on
the afternoon of each day.

And, in addition to making observations
of the source areas and assisting with

the search-and-rescue operations,
I was also just taking a lot of

photographs out the window of the –
of the helicopter as we were

flying around, knowing that
these could be incorporated

into a structure-from-motion model.

So, because Brian and I have some
really great collaborators at the

University of Lausanne in Switzerland,
as soon as I ended the day, I was able

to ship them all of these photographs,
and, overnight, they cranked out this

SfM model and did some differencing
against Lidar data that we had before.

So one of the things that I’m kind of
proud about with this rockfall series

is that we were able to do a lot
of the remote sensing analysis

in rapid fashion. And, in less than
24 hours, we had pretty much the

full story in terms of the source, the
volume, the geometry of the failures.

We knew the history really well.
And that set us up for doing the kind of

analysis that the park needed,
which was, are we going to be able to

open the road again, which was
a decision that Brian and I had to

make within about 24 hours.
But we felt confident doing that

based on what we
were seeing in the data.

So, when we compare the SfM point
clouds and the Lidar data,

for the whole the sequence of rockfalls,
there’s a cumulative volume of

about 10,000 cubic meters.
Most of these rockfalls were

sort of classic
Yosemite exfoliation.

They were relatively thin failures –
1 to 2 meters thick.

But eventually, as those
rockfalls propagated up the cliff,

they sort of tapped into more
structural control in the cliff.

They were interacting with
more regional joints,

as opposed to the exfoliation joints.
And it thickened up to greater than

7 meters, which is why that last
rockfall had such a large volume.

And then, yeah, Brian and I basically
looked at the volume of what a

potential future failure could look like
if it continued to propagate upward

and where – you know,
where that rock mass would be

bounded by existing fractures.
And we basically determined that it

was unlikely that there’d be a rockfall
bigger than what had already happened.

And there have been additional
smaller rockfalls from there,

but nothing on the scale of
what happened in 2017.

We also found a couple of
other interesting things.

For the most part, the cliff is retreating
back as these rockfalls occur,

and that’s expressed
by the color bar there.

But the pinks and purples actually
represent places where the cliff

has moved forward.
And what we’re looking at there

is a roughly 23-by-14-meter
exfoliation sheet that moved

20 centimeters outward.
It actually kind of rotated out as if it

was like a giant barn door swinging
open slightly by about 20 centimeters.

Part of that slab has since failed.
Part of it remains.

But that was obviously something
that we were interested in in terms of

additional future failures.
But volumetrically, this sheet

was relatively small, so we
weren’t too worried about it.

One of the interesting things we can
pull out of the longer-term Lidar data

set, as well as photographs of the cliff
taken prior to 2010, is that this area

really was not active from about 1976,
as I’ll show you in a minute, to 2010.

Right, there was really
no activity there.

But, beginning in 2010, there were
rockfalls that started low on the

cliff and, over multiple years,
slowly propagated upwards.

And then there was a whole bunch of
failures on those two days in September

of 2017. And, again, there have been
smaller subsequent rockfalls since then.

So this is actually a pattern that we
see repeating fairly often in Yosemite

with these exfoliation-type rockfalls.
So they begin at a point.

Why that happens initially, I don’t know.
Or, I don’t always know.

But then it’s common for
rockfalls to sort of radiate

outward from that initial point –
upward and outward.

And this is a series of rockfalls that
occurred from the Rhombus Wall

over a two-year period starting
in 2009. And it just shows

a similar sort of pattern.
These rockfalls are not strongly tied

to environmental triggers,
like rainfall, for example.

They seem to be more related to
changing stress conditions on the cliff,

often associated with a prior rockfall.
So, for this sequence, we actually

observed a number of occasions
where a rockfall would happen,

and hours or days later, there’d be
audible cracking sounds from the cliff.

There would be another rockfall,
more cracking sounds,

and another rockfall, right?
And eventually, the spacing

between those events would extend
out and then just taper off entirely.

To my knowledge, there haven’t
been any rockfalls at the

Rhombus Wall since 2010.
So you see these clusters of activity,

and then things
quiet down a bit.

Okay. So that’s – that was a sort of
short-term example of a large

volume sequence of rockfalls.
We wanted to probe a little deeper into

the past, and so we were able to use
a series of black-and-white

photographs taken from a helicopter
in 1976 by a member of our

search-and-rescue team
who still works in Yosemite.

And we scanned these.
We brought them into Agisoft,

and we made SfM models
from these historical photos.

In this case, there were only
21 photographs for the southeast face

of Middle Brother, but we were
able to generate a point cloud

with about 6 million points,
which really is, I think, not bad.

It’s kind of interesting to have a
structure-from-motion model

[background noises] [chuckles] –
somebody needs to watch their mic –

structure-from-motion model,
basically, that’s 40 years old,

that represents the cliff prior to
a number of big rockfalls.

And we were able to do
the same thing for El Capitan.

Here, we only used 16 photographs.
But, again, there is sufficient overlap

and change in perspective
to generate a pretty decent

point cloud model of
about 5 million points.

So, by combining the historical –
we’re calling it historical SfM models

against the more recent Lidar data,
we can start to compare the two

and come up with rates of
rockfall over a 40-year period.

I will note that this
doesn’t work everywhere.

On El Capitan, for example, you know,
I was really interested in trying to get

a model for a whole cliff,
but that whole left side –

the west-faced Salathé Wall,
if you know it, we just did not have

sufficient overlap in the photographs.
So we basically weren’t able to build

a model for that, which is why our
analysis – my whole study here is

kind of limited to the southeast faces
of Middle Brother and El Capitan.

When we combine the structure-from-
motion data and the Lidar data,

the first thing we see is that there
are a fair number of artifacts.

Right, these are just places
that mostly are not resolved well

in the historical
SfM model.

And so they appear as differences,
but they’re not actually places

where rockfalls occurred.
We can confirm that by just doing

a simple photographic analysis.
When we compare the actual photos

that make up these models, we can
see that there’s really no change there.

It’s more of a change in shadow or just
not sufficient overlap in the photos.

But we can also see, in the case of
Middle Brother, eight large-volume

rockfalls that were detected. And I
want to just mention a couple of those.

So, for example, there was a rockfall
in March of 2010 – sorry, in March

of 1987 – the 10th of March, 1987.
That was a rockfall that, in our

database, was reported as being
600,000 cubic meters, right?

It was by far the largest-volume
rockfall in our database.

And I was always kind of suspicious
of that because, on the ground, it just

did not look like a rockfall of more
than half a million cubic meters.

And, when we actually did
this analysis, we found that it’s

more on the order of
20,000 cubic meters.

So it’s way smaller than
the original volume estimate.

And I talked with the guy who made
that volume estimate, and it turned out

that, when the rockfall occurred,
he was almost directly below it,

running for his life as rock fragments
were sort of flying past him.

So that may have influenced his ability
to estimate that particular volume.

Subsequent estimates that
he made were really good.

So this was kind of a one-off,
but a major correction

to our rockfall database.

And just to show you how good
some of the other estimates are,

from the same cliff in February
of 2000, the previous estimate

was about 14,000 cubic –
sorry, 17,000 cubic meters.

This analysis showed it was about
14,000 cubic meters. So it’s different,

but, considering that these are just
estimates from the ground using

binoculars or a spotting
scope, not too bad.

So, by and large, we found that
most of the volumes

were estimated decently well
with one glaring exception.

Okay. So, when we look at the
short-term rockfall rates by comparing

the terrestrial Lidar against the older
structure-from-motion model –

so it’s a 40-year observation period,
we detect 120 rockfalls from Middle

Brother over that 40-year period.
That contrasts with 62 rockfalls

reported in the database.
Now, I want to say that pretty much

all of the large-volume rockfalls
in the database are – sorry, all the

large-volume rockfalls that we
found in this analysis were

also recorded in the database.
The under-reporting is mostly

for the small rockfalls. So the
cumulative volumes are pretty close.

But the number of rockfalls is
much greater using this analysis.

When we take that volume of material,
and we normalize by the contributing

cliff area, we come up with a cliff
retreat rate of 1.6 millimeters per year,

which is pretty fast for a geomorphic
rate in the Sierra Nevada.

And one interesting thing that
comes out of this – we’re looking at

two different areas on Middle Brother.
On the left, we’re looking at two events

that occurred in February of 2000.
The event on the left – so the initial

failure is in red, and then subsequent
failures in the other colors.

And you can see that, within that scar,
after the initial failure, which was

in the year 2000, you know,
there’s been another 16 years

of smaller failures
from within the scar.

So ongoing activity
within that rockfall scar.

And, by contrast, the rockfall scar
to the right, which happened,

more or less, at the same time,
has had no additional activity.

I don’t have a good explanation for that,
but it does illustrate in one frame the

sometimes complexity that’s involved
in these rockfalls and that they are

highly progressive, and these scars can
be active over a long period of time.

The panel to the right just shows
another example of one of these

progressive rockfalls that’s
occurred mostly over a single year.

But you can see that there were some
smaller failures in the years leading up

to that, and those may, in retrospect,
actually have been precursor events

to these larger failures.
But that’s a difficult thing to work out.

Okay. So that analysis that I just walked
you through from Middle Brother,

we did the same thing for El Capitan.
We added one additional year of

observation to account
for these rockfalls in 2017.

Over that 41-year period,
we detected 115 rockfalls,

compared to only 38
reported in the database.

The total rockfall volume there
was about 16,000 cubic meters.

And, when we normalize by that
contributing cliff area, the cliff retreat

rate is about 0.8 millimeters per year.
So that’s half of what we measured at

the adjacent cliff – Middle Brother –
over that same time period.

From both the rockfall database and
from the SfM and Lidar monitoring,

we can develop rockfall
volume frequency distributions.

I think last week’s talk –
I wasn’t able to attend,

but I think it’s about this topic.
So you’ve seen this sort of thing before.

These distributions exhibit a power law
relationship, and the b-value exponents

in this case are much less than 1, right?
They’re sort of in the – in the realm

of 0.32 to 0.41. And so that’s indicating
to us that the large-volume rockfalls

are dominantly responsible for
moving rock from the cliff down

to the floor of the valley, right?
So there’s not sort of an equal

contribution of rockfalls
of all sizes to cliff retreat.

Cliff retreat is dominated by these
infrequent large-volume rockfalls,

just as understood by the slope
of the line of this relationship.

We can also, from this, extract
frequency of occurrence in terms of

number of rockfalls per year per cliff.
And, for these two cliffs in particular,

the recurrence period for something
10,000 to 20,000 cubic meters is

anywhere from to 50 to 100 years.
When you include all of Yosemite

Valley, that comes
down quite a bit –

like, it’s less than 10 years for
something of that volume.

So this is all useful information for
understanding frequency,

recurrence, and sort of
the geomorphic effectiveness

of rockfalls of
different volumes.

So that’s our short-term
understanding of rockfalls, right?

That’s our
40-year snapshot.

What we want to do now is compare
those rates that we derived from

remote sensing to longer-term rates.
And the way we’re going to do this

is by looking at talus accumulations
at the base of the cliff.

And I would argue that Yosemite
Valley is nearly a perfect place

to do this. Because it was glaciated
during the LGM, we assume that

that basically swept out any
pre-existing talus and sort of reset

the system at about 60,000 years ago.
The floor of Yosemite Valley

is remarkably flat. We know the
amount of post-glacial aggradation.

It’s on the order of 3 to 5 meters.
And, once the talus falls on the

valley floor, there is virtually no
modification of that talus, right?

It’s not – it’s not eroding
at any appreciable rate.

Given our time scales,
the river is doing nothing to it.

It just bends around the talus.
So it’s a really nice closed system

for talus accumulation.
And so, if we can determine

the volume of talus underneath these
cliffs, again, normalized by area,

and divide by 16,000 years,
we can come up with a cliff retreat

rate – post-glacial cliff retreat rate.
And the way we do this is

relatively straightforward.
In three dimensions, we’re just

working to do our best to estimate
what the cliff is like underneath

that talus pile, right?
It’s clearly hidden from us,

but we can use the adjacent surfaces
of the cliff and the flat valley floor

to come up with our best sense of
what that cliff surface looks like.

For Middle Brother, it was pretty
straightforward – sorry, that’s the

situation here on the left,
where we assume just a continuation

of the broad U-shaped valley,
which is partially buried in Yosemite.

But, based on some information
we know from old seismic data

of the shape of the valley floor
underneath that sediment, have a pretty

good estimate of what the cliff was
like underneath Middle Brother.

So that’s about 4-1/2 million
cubic meters, once we account for

talus porosity. When we divide
by the contributing cliff area,

that gives us a cliff retreat rate
over a 16,000-year period

of about 0.3 millimeters per year.
So recall that the modern rate –

the 40-year rate for the same cliff
was 1.6 millimeters per year.

So, in this case, a huge difference
between what I’ll call the modern rate

and the longer-term
post-glacial rate.

At El Capitan, the analysis was
a little more complicated because

there’s actually some bedrock
poking up in various places in

the middle of this talus slope. So it has
a different subsurface bedrock shape.

And we basically did our best using
these outcrops to develop a more

sophisticated bedrock shape there.
But still, we can do the same sort of

analysis – calculate the total talus
volume here, which is about

7.8 million cubic meters,
divide by the contributing cliff area.

Over a 16,000-year period, our cliff
retreat rate is 0.8 millimeters per year.

All right. So finally, we can compare
these long-term versus short-term

cliff erosion rates and get back to
these kind of motivating questions

that I mentioned early on.
The first, how do long-term

and short-term cliff
retreat rates compare?

So, if we look at El Capitan,
they’re remarkably similar.

The long-term post-glacial rate at El Cap
is just under 0.8 millimeters per year.

The short-term rate measured
over that 40-year period

is just a little
more than that.

So, if we were only looking at
El Capitan, we would probably say

that the modern rates are, indeed,
representative of the longer-term rates.

But just next door at Middle Brother,
it’s an entirely different story, right?

For starters, we see that the long-term
rates are lower than at El Capitan –

less than half. And the modern
rates are much, much greater –

like, 5 times greater
than the long-term rates.

So the first thing that this suggests to
me is that Middle Brother has either

recently experienced or is still in the
midst of a pulse of activity that is not

representative of longer-term rates.
It would be easy to use that explanation

if all of these rockfalls were occurring
from the same progressive scar,

but they’re not. They are actually
a number of discrete source areas

on the cliff. Many of those
are relatively large volume.

And so, for whatever reason,
over a 40-year period, we’ve just

seen a lot of rocks come
off of Middle Brother.

And my sense is that that rate is
anomalously high and,

in the coming decades,
we would probably start to see

that rate come down to something
close to the post-glacial rate.

And, in that way, it may be a little bit
akin to clustering of earthquakes.

But, as far as why that’s happening
from a number of discrete source

areas on the cliff, I don’t have
a good explanation for that.

Nor do I have a great explanation
for why Middle Brother actually has

a lower post-glacial
rate than El Capitan.

As I said at the outset, these cliffs
are really pretty similar in most of

their key metrics. If anything,
El Capitan has a steeper slope,

and maybe that’s the driver for it, but
beyond that, I’m a little hard-pressed

to understand exactly why those
long-term rates are so different.

Okay, so the second one –
are modern rockfall and cliff retreat

rates representative of rockfall
rates in the geologic past?

Sort of answered that.
At El Capitan, the answer appears

to be yes because those long- and
short-term rates match pretty well.

And Middle Brother, the answer
is clearly no, and it remains kind of

an open question
as to why that is.

And then the final question,
how important are the largest

rockfalls to erosion rates?
And I think the answer is that

they are quite important.
So our volume-frequency relationships

indicate the contribution of
large-volume rockfalls is

most important to
drive cliff erosion.

And furthermore, we have kind of
an interesting little example in that

the analysis for Middle Brother
went from 1976 to 2016.

We extended that for a year at
El Capitan because of those

big rockfalls in 2017
that I’ve talked about.

But, if we were to subtract those
out and only go with the analysis

up to 2016, the rate would be
something like 30% lower.

In other words, one year’s worth of
rockfalls at El Capitan caused the cliff

retreat rate to jump up quite a bit, right?
They added a lot of volume to the total

rockfalls that have come off that cliff.
So the large-volume rockfalls are really

important, and it does kind of get at
this issue of, what is the relevant time

period of observation that you need if
you’re gauging something like rockfall

frequency for hazard purposes?
You know, basically, how long is

a long enough period of observation
in order to try to capture meaningful

longer-term rates? And that’s
kind of [chuckles] –

that’s kind of
an open question.

But, in this case, there’s a pretty
big difference between

a 40-year observation period and
a 41-year observation period.

So, to conclude, we see that cliff
erosion rates for El Capitan

are similar across both short-
and long-term time scales.

But they are significantly different
at Middle Brother with the recent rates

being much higher than
the post-glacial average.

And these, again, are adjacent
cliffs with similar characteristics.

And yet they’re displaying markedly
different rockfall behavior over these

time scales. And so that – again,
I don’t have a great answer for that.

I think there are some things
we could explore and ways

we could try to quantify that.
And that’s one of our next steps here.

I can say for sure that these infrequent
large-volume rockfalls are the dominant

contributor to cliff erosion and
consequently, the observation period

is important because, if you have one
year of really large-volume rockfalls,

that’s going to have an outsized
influence on your short-term average.

But I still don’t have a clear answer
on how long is long enough

for those observations.
So with that, I will thank you

for your time, and I’m happy
to take any questions.

[silence]