2017 May Evening Public Lecture

Video Transcript
Download Video
Right-click and save to download

Detailed Description

Title: Underwater Secrets of the Hayward Fault Zone: Integrated 3D imaging to understand earthquake hazards 

  • Underwater imaging provides a unique opportunity to study urban fault hazards.
  • How do we link surface structures to depths where earthquakes occur?
  • How does "acoustic trenching" help us understand earthquake history?

Details

Date Taken:

Length: 01:14:11

Location Taken: Menlo Park, CA, US

Transcript

[background conversations]

- No.
- No.

[inaudible conversations]

- Can you hear me now?
- Good evening.

Good evening. And welcome to
the U.S. Geological Survey.

I’m Helen Gibbons from the Pacific
Coastal and Marine Science Center.

And I’m glad to see you all here
for the May public lecture.

As a quick reminder, I want to let
you know that the June lecture will be

on June 22nd, and it is on climate
change and ocean acidification.

You can pick up a flier in the back –
at the back table there.

Tonight’s lecture is titled Underwater
Secrets of the Hayward Fault Zone –

Integrated 3D Imaging to
Understand Earthquake Hazards,

presented by Janet Watt.

Janet received her M.S. in marine
science with a emphasis on geological

oceanography from Moss Landing
Marine Laboratories in 2004.

While she was earning
her M.S., she worked at the

U.S. Geological Survey
here in Menlo Park.

She was part of the Geophysical Unit
of Menlo Park, also known as GUMP.

[laughter]
That’s a great name.

Working with that group, she combined
3D geological mapping with

potential field methods, that is,
measurements of gravity and magnetism,

to understand earthquake and
volcano hazards in the western U.S.

She also applied these methods
to mineral and water resources.

In 2010, Janet returned to her
marine geology roots and joined

the Pacific Coastal and Marine
Science Center in Santa Cruz

where she currently works
as a research geophysicist.

Janet’s research focuses on
connecting onshore and offshore

geologic structures, deformation,
and active tectonics to address problems

focused on geologic
hazards and processes.

Her recent work involves integrating
various geophysical techniques

to characterize faults and fault
interactions in three dimensions.

This provides a good way to
detect recent fault movements

and assess earthquake hazards.
The USGS is pleased to bring you

this program on using underwater
imaging to study urban faults.

As always, please hold any
questions until the end of the lecture.

And now please
welcome Janet Watt.

[ Applause ]

- Thank you, Helen.

[ Applause ]

Can you all hear me?
- Yes.

- Yes.
- Okay. Okay.

- [inaudible]
- Well, it’s a pleasure to

be here tonight,
and thank you all for

coming out to hear me talk
about my work in San Pablo Bay.

Tonight I’m going to tell you
about using a suite of geophysical tools

to image the Hayward Fault
beneath San Pablo Bay

and better understand the
earthquake hazard of this fault zone.

Now, often when I’m giving
talks to folks, I talk about

studying a particular fault to
understand the earthquake hazards.

But I want to start by explaining,
what does that actually mean?

How do we take a
site-specific fault study

and understand
earthquake hazards?

And then, what does that mean for you,
the residents in the local area?

So here on this slide
I have four products

that the USGS puts out
related to earthquake hazards.

And so when we go out and
we look for active faults,

we try to get this
information integrated into

the USGS Quaternary Fault Database,
which is the map on the left.

And this map identifies faults that cut
young deposits that are active faults.

And these are considered the
most likely to cause earthquakes.

And so this fault map
on the left is used as the basis

for all of the following
hazard maps on the right.

The next map in line shows
the probability of a

major earthquake on Bay Area
faults within the next 30 years.

And these probabilities take into
consideration decades of earthquake

research on the size and frequency of
past events on the Bay Area faults.

Now, given the estimated
earthquake size and frequency,

plus knowledge about local geology
and soil conditions, seismologists

can estimate the amount of shaking that
is likely to occur in any given area, and

that’s portrayed on the U.S. National
Seismic Hazard Maps – the next one.

And this information is used
in designing building codes

and insurance rates
used in the United States.

Now, the map on the far right
shows a ShakeMap – a scenario –

of a scenario earthquake, magnitude 7,
on the Hayward Fault.

And the ShakeMap shows the
shaking intensity expected in the area.

And the highest shaking
is shown in the red colors.

And these maps are used
in planning and coordinating

emergency response
in local areas.

So here’s a roadmap of what I’m
going to talk to you about tonight.

First I’ll explain the motivation
of the work we’re doing in

San Pablo and explain what the scientific
questions are we’re trying to answer.

And then I’ll explain
how we use sound

to image faults in the
marine environment.

How we connect surface faults
to depths where earthquakes occur

in 3D geologic mapping.
And how we use acoustic trenching

to unravel the history of
earthquakes along the fault.

So all this is based on
the idea that understanding

past earthquakes will help us
prepare for future ones.

So as a researcher, I have an affinity –
some might say an obsession –

for fault discontinuities,
and those are fault intersections,

step-overs, bends –
little kinks in faults.

And one of the reasons I find
these areas so fascinating

is that because they likely hold the
key to understanding why some

earthquakes cascade into multi-fault
devastating events and others do not.

And I’ll explain
what I mean.

So until recently, the prevailing
theory was that earthquakes are

confined to fault segments.
And here I portray that.

The gray line is a fault line, and the
red part is an earthquake rupture.

It’s defined – and its endpoints
are defined by what we call

segment boundaries
in the blue boxes.

However, recent large events
have shown us that

sometimes earthquakes are not
confined to single fault segments.

And on the far right, I show an example
of a rupture that breaks through

multiple segment boundaries and
becomes a very large earthquake.

And a recent example of this is the
Kaikoura earthquake in New Zealand

that broke as many as
10 surface faults in one event

and likely the megathrust
subduction fault also.

So a lot of large events are
showing us that, while sometimes

earthquakes only break
single segments on a fault,

we know that they can and do break
through those segment boundaries.

And this is how – the need to take a
closer look at these segment boundaries

and re-assess their role in
earthquake hazard assessment.

So these fault discontinuities – places
where faults bend and there are kinks

in these faults – are inherently unstable.
The fault wants to be straight.

So it’s constantly sort of reorganizing
itself to try to straighten itself out.

And what this means is, often at the
surface, you have a very complex

network of faults at the surface,
as I’ve portrayed here on the left.

And what we need to do to
understand the hazard in an area

is understand which of
these faults are currently active

and how they connect to
one another and link up.

And then, once we figure out which
of the faults at the surface are active,

we need to understand how those
faults connect down to where

earthquakes occur, which is
much deeper in the Earth.

And so here on the left, I’ve shown a
cross-section, basically, of the surface

trace at the top and trying to connect
down to where these earthquakes occur.

And so we need to know
what the dip is – down-dip geometry,

or shape, of the fault
from the surface down to

where the earthquakes
are occurring.

And then we need to know how often
these earthquakes rupture this fault.

So here, this is a
screen grab from the

USGS Quaternary Fault Database
that you can get to online.

And it shows the map faults
in the Bay Area and their ages.

And here is the Hayward/Rodgers
Creek Fault highlighted in red.

Note that there are no faults mapped
in San Pablo Bay between the

Hayward and Rodgers Creek
Fault currently on this map.

This has always bothered me
ever since I came to the Survey.

I wondered, why? [laughter]
Why can’t we find this fault?

Is it just that it’s underwater? I mean,
we’ve got to be able to figure this out.

So it was an honor to get to
try to go out and figure this out.

So why is it important to understand
the geometry beneath San Pablo Bay?

Well, this area has been
defined as a segment boundary

in a number of earthquake
models in California.

But yet, we don’t understand
exactly where the active fault goes.

And also, the Hayward/Rodgers
Creek Fault cuts through the heart of

the San Francisco Bay Area,
which boasts the fourth-largest

economy in the U.S. and ranks
among the top 20 in the world.

So there’s a lot at stake
along this fault zone.

The most recent large
earthquake on the Hayward Fault

was in 1868, and that’s
about 150 years ago.

The five earthquakes
before that one occurred

approximately at 140-year intervals.
And that’s why the Hayward

and Rodgers Creek Faults are
considered the most likely Bay Area

fault to experience a major
earthquake within the next 30 years.

So how big an earthquake could that
be if these faults are ready to go?

That depends on the fault length.
And according to recent mapping,

the Hayward and Rodgers Creek Faults
together are 190 kilometers long and

extend from north of Healdsburg in
the north to Alum Rock in the south.

And based on that length, it can create
up to a magnitude 7.4 earthquake.

So here we’ve zoomed in a bit
on that Quaternary Fault map

on the left into where you
can see San Pablo Bay.

And the ability of an earthquake
on the Hayward Fault to continue

onto the Rodgers Creek Fault,
or vice versa, depends on the

geometrical relationship between
these faults beneath San Pablo Bay.

And until now, that relationship
has remained uncertain.

So here are the goals of the project.
Definitively locate and characterize

the most recent active
Hayward/Rodgers Creek Fault

in 3D beneath
San Pablo Bay.

And that involves what I call
integrated geophysical approach.

And the second part is to constrain the
earthquake history along this portion

of the fault, and that involves geological
sampling – coring beneath the bay.

So here we are
looking at San Pablo Bay.

We have the Hayward Fault going
offshore in the south at Pinole Point.

And up to the north, we have the
Rodgers Creek Fault coming down

and going offshore near Sears Point
by the raceway up there to the north.

So where the faults enter the bay,
they’re separated by about 5 kilometers.

Now, San Pablo Bay is very shallow.
I don’t know if many of you know this,

but most of the bay is less than
5 meters deep, except for the

shipping channel in the south,
which is that dark blue color,

is about 12 meters deep. And that’s
where the big barges go through.

And this shallow water is
one of the reasons – it makes it

very difficult to get ships in there
to collect data and image the fault.

And this is likely part of the
reason why the relationship

between these faults has
remained a mystery for so long.

The other reason is that there
is widespread gas in the bay

sediments from decaying organic matter. 
It’s a natural process.

There’s shallow gas in
many estuaries across the U.S.

But this shallow gas
scatters the acoustic signal

and makes imaging the geologic
structure in this area very difficult.

But we chose to target
the very shallow layers

above that gas layer to
look for signs of activity.

So the geometry and connectivity
of these fault zones has been

discussed and debated for years
with workers describing the relationship

between the two faults as either what’s
called a step-over or as a fault bend.

And resolving this uncertainty
is particularly important

because these alternative fault
models have different implications

for earthquake dynamics
and local ground motion.

And you can imagine, if two faults are
separated by a distance, it’s difficult –

more difficult for an earthquake
to jump from one to the other.

It’s possible.
And we know it’s possible.

However, but if there’s a fault bend,
and there’s a direct connection,

it’s like the faults are holding hands,
and it’s easier for an earthquake

to travel from
one to the other.

So the problem is this previous work –
none of these interpretations were based

or supported by direct evidence of
near-surface active faulting.

So Phase I is our integrated geophysical
approach to try to solve this problem.

We want to document the
active faulting in San Pablo Bay.

And to do this, we used an ultra-high-
resolution chirp seismic reflection.

I’ll explain that in
more detail in a minute.

But again, we’re focusing –
we’re trying to image

small deformation – or, small centimeter-
scale offsets in the very near surface.

And then we want to
link that surface structure

to depths where earthquakes occur –
figure out its 3D shape.

And to do that, we integrate our
near-surface observations

with marine magnetics as well as
existing geologic and geophysical

information as well as
deformation modeling.

So I talk – I’m using subsurface
imaging techniques – seismic reflection.

And this is done on land as well.
But on land, imaging with

seismic reflection can be very tricky.
And as an example, this is the

East Bay Seismic Investigation that
was carried out by the USGS recently.

And they wanted to image the
structure of the Hayward Fault

and the Chabot Fault here.
And this is one seismic line, and it took

an amazing amount of work to collect
seismic data along this one transect.

And they had to drill holes.
It was an amazing amount of

manpower to just get one cross-section.
We have – we’re lucky in the marine

environment because here’s a map of
all the cross-sections that we get

throughout San Pablo Bay from towing
our seismic source behind our vessel.

We got 13 fault cross-sections
spaced about a kilometer apart

shown by the yellow lines.
And when we have this sort of

spatial grouping of cross-sections,
we can characterize changes

in the faults along strike.
And we can image at higher resolutions

than you can on land – up to – again,
we can measure centimeter offsets.

The active marine sedimentation
within the bay is a natural

tape recorder of past events.
So you can’t record earthquake events

unless you have an active
sedimentary record to look at,

and we have that in
the marine environment.

So here is a primer on seismic
reflection just to explain how it works.

So basically,
you have a survey vessel,

and we tow what’s called
an acoustic source.

And basically, we send acoustic waves,
and they propagate through the water

column and into the subsurface.
They’re reflected from layers

that show a contrast in what’s
called acoustic impedance.

Basically, changes in
physical properties –

a sand layer, a clay layer –
reflect sound differently.

Those reflected signals are recorded
by hydrophones on the receiver.

And then data is processed to
produce 2D cross-sections, kind of like

a road cut of the subsurface –
the sediment layers below the sea floor.

So I mentioned that what
we use is called a CHIRP.

And that stands for compressed
high-intensity radar pulse.

And this uses
very high frequencies.

The difference between this instrument
and the image I showed before

is that here the source and the
receiver are on the same unit.

And this is ideal for imaging
shallow structures – faults.

We can image centimeter-scale vertical
offsets, and it’s capable of – the system

is capable of imaging tens of meters
below the seafloor in ideal conditions.

But I mentioned, San Pablo Bay
is not an ideal condition because of

the shallow gas in the area.
But luckily, the 2 to 5 meters below

the sub-seafloor, above the gas,
was quite revealing in terms of showing

us what the Hayward Fault is doing.
So we got lucky.

Because the bay is very shallow,
we towed our instrument at the surface.

We had to put it on these pontoons
to keep is from sinking and dragging

in the mud. So this method worked
very well to keep the fish at the surface.

Normally, we tow the
fish down behind the boat.

It sinks below the
sea surface a few meters.

But in San Pablo Bay, that wouldn’t
work. It would drag it in the mud.

So here is an image.
It’s called a seismic profile –

basically a road cut showing
the sediments below the seafloor.

And this is an image from
the northern part of the bay –

the yellow line on the
map in the upper right.

And you can see that very strong black-
red-black line is the seafloor reflector.

So that’s the seafloor
at the very top there.

You can see the layers. There’s layers –
flat-lying layers below that.

But then, as you get
further down in the section,

those layers start to
be deformed, or warped.

And down the middle of this
cross-section, the layers actually –

it’s like they’re cut. They don’t match
on either side. They’re offset.

You can’t – and that’s the Hayward
Fault that we’re imaging right there.

And the – at the bottom,
where the signal goes from

seeing layers to sort of fuzzy,
that’s the gas.

So how do we map this fault
across the bay using these data?

We basically look at all these
cross-sections that I have shown

on the right, and we identify the fault.
Where is the sediment section broken?

And we put a
mark on the map.

And then we essentially
connect the dots.

And so we see here, as the fault traverses
the bay from the south to the north,

it’s bending slightly,
about 10 degrees,

to the right towards
the Rodgers Creek Fault.

So with these chirp data,
we have the first direct evidence

of active faulting along the
Hayward Fault in San Pablo Bay.

Now, what happens in this area,
is the next question.

So here we see the southern part of the
Rodgers Creek Fault at Sears Point.

And previously it was –
so the southern Rodgers Creek Fault

branches into two faults.

The eastern strand follows the
main trend further to the north.

And the western strand
diverges by about 22 degrees.

So the western strand of the
Rodgers Creek Fault points

directly towards our offshore
fault that we had mapped along –

and it extends along – there is
distinct linear gradients in the

topography along the side of
Sears Point there that it follows.

Previously it was thought that the
Rodgers Creek Fault continued along its

main eastern trend into San Pablo Bay.
However, our seismic data show

no definitive evidence for offset in the
near surface along this projected trace.

And in the next slide,
I’ll show you a couple profiles

of this lack of
offset on that trace.

So here we have two seismic
profiles crossing the bay –

lines A and B are shown
in the map in the upper right.

And the projected trace of the
Rodgers Creek Fault cuts through here.

You’ll see the layers beneath the
seafloor are not obviously cut

where that fault projects offshore.
And that leads us to believe that

either that fault was never there,
or it’s presently not active.

So the integrated onshore/offshore
evidence described above strongly

suggests the Hayward and Rodgers
Creek Faults are directly connected,

as I’ve portrayed here
in the red dashed line.

And what the – the other
faint lines you see behind there

are previous interpretations of
the step-over, which is shown

in the white dashed line, and the
fault bend by the black dashed line.

It’s not that much different,
but it – our new result shows

that they are
directly connected.

So since our onshore evidence
for faulting is not as strong as

the offshore evidence,
we wanted to see if our

proposed connection
made physical sense.

And so to do this, I enlisted the help
of a colleague named Tom Parsons

to help me with what’s called
deformation modeling.

We basically take the fault –
our interpreted fault here and then

model what happens when you
slip two blocks past each other

with that – with that geometry.
And that model shows areas where –

that are deformed where
the land goes up and

where the land goes
down in this model.

And so we want to compare the
model deformation to what we

observe and see if they’re the same –
see if it makes physical sense.

So what we see –
here are the results.

The zone of greatest subsidence,
which is shown in the blue, is consistent

with the diffuse seismicity in an area
of active extensional deformation.

Seismicity is shown by those
black dots you see in the map.

And also there’s a subtle
elongate depression, or sag,

observed in the seismic data
offshore in the area of

predicted greatest subsidence
where the land goes down.

So in the next slide,
I’m going to show you a couple

profiles that cross this area
of subsidence, or the sag.

So here I have two profiles at
the northern end of the bay.

And you can see the sediments are
flat-lying at the top, but as you go down

in the sections, they bow towards each
other and look like a basin, basically.

And this basin is elongated
along the Hayward Fault zone.

We do not see this depression
in the southern half of the bay.

It’s just isolated in this one
corner of the bay where the

model predicted that we’d
have the greatest subsidence.

We believe this depression
is indicative of the formation

of what’s called a lazy Z
fault bend basin.

These are observed all over the world,
where strike-slip faults make a

right bend. Often there is extension.
A basin forms.

There’s subsidence, and the
land goes down in that bend.

So this basin has likely been trapping
sediment here and is likely a good area –

has potential for preserving
an earthquake record.

It might be a good place to go coring,
and so that’s what we did eventually.

So the onshore/offshore evidence,
combined with our deformation

modeling, shows that the
Hayward and Rodgers Creek Faults

are directly connected
through a fault bend.

And this direct connection
makes it easier for an earthquake

to rupture from the Hayward
to the Rodgers Creek or vice versa.

So now that we know where the
active trace is, we want to understand

how that trace extends into the
subsurface to where earthquakes occur.

So to do this, there are a number of
different geophysical tools that we

can and did use in San Pablo Bay.
And I’ve listed them here on the right.

However, in the interest of time,
I’m just going to talk about two

that I’m most familiar with,
and that’s gravity and magnetics.

It’s important to note that the
interpretation of all these data types

involve assumptions and uncertainties.
They’re not direct observations.

Hence the need for
an integrated approach.

If you use different –
a number of different techniques,

and we get the same answer,
we have more confidence in our results.

So how do we use gravity
and magnetics to map faults?

Well, we take advantage of the fact
that faults often juxtapose rock units

with different physical properties –
different types of rocks, especially on

strike-slip faults where faults
move laterally from one another.

So you have a dense object here, and a
fault moves, and it offsets that from the

other, and you end up getting a contrast
across the fault in the physical properties.

And that shows up in a
gravity or magnetic anomaly.

So we can map edges of magnetic
or dense rock units using the horizontal

gradient maxima of that anomaly.
So the slope.

If those boundaries
correspond to fault traces,

then we can use those
anomalies to estimate fault dip.

So here is an aeromagnetic
anomaly map of San Pablo Bay.

And this was the magnetic data
that existed before we went out and

collected new
marine magnetic data.

And the previous interpretation of the
fault was that the Hayward Fault just

extended straight across the bay, sort of
coincident with this magnetic anomaly.

So here these anomalies reflect
the amount of magnetite, in a sense,

within the crustal rocks.
And warm colors are

magnetic highs. And cool colors
are dipole, or magnetic lows.

So this is what existed
before we collected data.

Sort of a fuzzy
view of things.

But a colleague of mine, Dave Ponce,
went out and collected marine magnetic

data along this dense grid of track
lines spaced about 200 meters apart.

And that’s the resulting image.
Much more detail.

And you can see that the previous
interpretation doesn’t quite work

anymore, and this is the trace of
the Hayward Fault that we mapped

with the chirp data, and it aligns
very nicely and follows the northeast

boundary of two prominent
magnetic anomalies here – A and B.

The sources of which are likely
serpentinite, coast range ophiolite,

or Tertiary volcanic rocks, all with a
high degree of magnetite in them.

But what you – I also want to draw
your attention to here are these subtle

what we call short-wavelength anomalies
and gradients along the fault trace.

And these likely reflect
folding and/or vertical offset

of tertiary volcanic rocks
within the fault zone itself.

And we can only identify these very
small features with these new data –

very high resolution, close spacing,
new magnetic data.

So here is an isostatic
gravity map of San Pablo Bay.

And isostatic gravity
anomalies reflect the

lateral density variations
in the upper crust.

And so the strong gradient you see
here in the bay, the changing color,

reflects the juxtaposition of dense
Franciscan basement rocks on the left

in the warm colors with
lower-density Tertiary

sedimentary rocks on
the right in the blue colors.

And the maximum horizontal gradient,
shown by the white dashed line,

corresponds to the fault
mapped with our seismic

and marine magnetic data
shown in the black X’s.

So now we have a correspondence
between our surface trace and

this maximum horizontal
gradient in the gravity data.

So we can now make
an estimate of fault dip.

So if a fault juxtaposes rocks of different
densities, and the surface trace is known,

which it is, one can determine the dip of
the fault at depth from the position of

the maximum horizontal gradient relative
to the slope of that anomaly.

And so here I’ve shown that if that
surface trace is sort of in the middle of

that slope, then it indicates that
that’s likely a vertical boundary.

And if it’s at either the top or the bottom
of that curve, the fault is dipping.

And so, in San Pablo Bay, the shape of
the gradient is shown over on the right,

and it’s sort of a hybrid between a
vertical and a northeast-dipping fault.

So the shape of the gradient
here suggests a vertical to

steeply northeast-dipping fault.

Okay, so we’re
done with Phase I.

We’ve identified the faulted surface and
estimated its dip from the gravity data.

Now onto Phase II, which is in progress.
We haven’t completed this phase yet.

But it involves geological sampling
and coring in San Pablo Bay.

So the path an earthquake rupture takes
depends not only on the fault geometry

and connectivity, but on other factors,
including the earthquake history.

So there’s a large body of
existing work on earthquake history

along this fault system.

And it all indicates that the Hayward
and Rodgers Creek Faults have

each accumulated enough stress
to produce a major earthquake.

Now, now that we know these two
faults are essentially holding hands,

we need to ask if, or how often,
they have ruptured together in the past.

And so estimates of the timing of the
most recent events along the Rodgers

Creek Fault, shown in the purple star
furthest to the north, and prehistoric

events on the northern and
southern Hayward Fault allow

the possibility of a combined
rupture in the mid-1700s.

Now, there are a number of
uncertainties associated with

the dates of these past events
on any one location along the fault.

But the uncertainties on these
three events overlap, so it allows

the possibility that they all ruptured
at one time, or they ruptured within

decades of each other. We can’t quite
tell because of the uncertainties.

So the more observations we have
of fault history along a fault zone,

the more confidence we build in
sort of what events are the same.

So perhaps San Pablo Bay has a record
of past earthquakes that could help.

So much of the previous work on
fault history on land involved the

use of paleoseismology, or the
study of ancient earthquakes.

And this is based
on the concept –

first, you know where
your active fault trace is.

And you dig a trench across it, which
is shown in the picture on the right.

And this is actually a fault trench
across a fault in Lake Tahoe.

But the concept is, before an earthquake,
you have a flat-lying surface.

An earthquake occurs,
you create a scarp for movement

on the fault – vertical
movement on the fault.

And then that scarp erodes and
deposits what’s called a

colluvial wedge – a pile of sediment
from the erosion of that scarp.

And then, over time, sediment
fills in that different elevation.

And you get a flat-lying layer over the
top, and the sequence repeats itself.

And what we try to do – what’s done
in paleoseismology is you date the

pre- and post-earthquake sedimentary
deposits exposed in the trench walls.

But of course you can’t
just trench anywhere.

You need to target areas
with active sedimentation

that can record a
history of fault movement.

So you have to pick sites where sediment
is actively accumulating to record events.

We use the same technique
in the marine environment,

but we do it a
little bit differently.

So the northern end of San Pablo Bay,
where we modeled the subsidence,

and our chirp data show that we
have a basin where you’d have

active sedimentation, this is an area
that’s likely to record past events.

So what we look for in
our chirp data is what’s called

evidence of growth faulting.
And here on the right, I have an

animation to show what I mean by
growth faulting, or event stratigraphy.

So you have a fault,
shown on the red dashed line,

and sediment layers
are those black lines.

An earthquake occurs.

You have a scarp forming.

Sediment fills in the area,
or accommodation space, we call it.

And then new sediment is deposited
on top in between earthquakes.

Then there’s another event.

Sediment fills in
that depression.

Flat-lying sediments
are laid on top.

So what we do – we can identify
those sedimentary sequences

in the seismic data, and then
we can go core those sequences

and try to get dates above
and below these event horizons.

So that’s what we did in the
fall of 2016. That’s a typo.

So just this past fall,
on a four-day coring cruise

on a barge called the
Retriever in San Pablo Bay.

So here on the left is a picture of that –
the back of the barge and what’s

called the A-frame –
that black A-shaped frame.

And then our coring device is on the
back of the boat lying down there.

And this A-frame lifts and guides
our about-4-meter-long core

to the seafloor where it’s
pushed into the bay mud

with a vibrating head that sits
on top of that core barrel there.

And it basically sits there
and vibrates into the ground

until it doesn’t go any further.
And then we pull it back out,

pull it onto the deck, retrieve the core,
cut it up, and take it back to the lab.

Oh, I forgot – the map
on the right shows all the

locations that we
cored in the blue circles.

And so to understand the
earthquake history of the fault zone,

we collected pairs of cores on either
side of the fault, particularly in the

northern part of the bay where there
is the most evidence for subsidence.

Then we take the cores from the
barge back to the lab and run analyses.

The first thing we do is
run the cores through what’s called

a multi-sensor core logger.
And we measure things like

P wave velocity,
magnetic susceptibility,

electro resistivity, gamma density,
and we photograph the core.

Basically measuring
its physical properties

with a bunch of
different methods.

Then we split the core in half –
take half the core, keep it as an archive.

Use the other half
for sampling.

We do a detailed
description of the core –

describe its sedimentology,
color, grain size.

And then we identify areas
to subsample – get dates.

So what we’re trying to figure out
is when did this fault last move.

Here is a seismic section from
the northern part of the bay – the area

that has evidence for subsidence,
or where the land is going down.

We have two cores that we drilled there,
marked 11 and 12, on either side of the

Hayward Fault, which cuts the
center of that – of that profile.

You can see the layers at the top of that
profile are horizontal. They’re not broken.

But then where that red star is on the
left, below that, those layers are broken.

And so in order to understand
when the fault last moved,

we need to date the sediments
above that star and below that star.

So we can bracket in
the age of that movement.

And we use a combination
of dating methods.

Just like use a combination of
geophysical techniques

because they all have uncertainties,
we want to use many different

techniques to do the dating so
we’re confident in our results.

And right now, we’re waiting for
those dates to come back from the lab.

[laughter]

So stay tuned.
Maybe come back next year.

And I’ll have more to say about the
earthquake history of these faults.

It’s a time-consuming job,
but we’ll get there.

So in summary, Phase I –
we integrated onshore and

offshore analysis to show that the
most recent trace of the Hayward Fault

connects to the Rodgers Creek
Fault through a fault bend.

And the gravity data suggests the fault
is vertical to steeply northeast-dipping.

We’ve collected and began
analyzing cores from the bay

to understand how the fault zone
has evolved through time

and when the last earthquake
ruptured this part of the fault.

So what does this mean
for earthquake hazard?

Well, the discovery of a direct
link enables simultaneous rupture

of the Hayward and Rodgers
Creek Faults, and that’s a scenario

that could result in a
magnitude 7.4 earthquake,

which would cause extensive damage
with global economic impact.

And for a more local perspective –
I don’t know how many of you

in the room remember the
Loma Prieta earthquake.

But a magnitude 7.4 would release five
times the energy as the Loma Prieta

earthquake, which caused 63 deaths
and 6 to 10 billion in property loss.

The other thing to remember
about the Loma Prieta earthquake

was that it was
not in the Bay Area.

It was in the Santa Cruz Mountains
quite a distance from here.

And the Hayward and
Rodgers Creek Fault and

many other faults
run through the Bay Area.

To provide some more context,
this is a slide that shows a comparison

of insured and economic losses
from recent natural disasters,

past quakes and hurricanes, and
earthquake scenarios on the right.

And I want to draw your attention
to a magnitude 7 Hayward scenario

and its total economic loss
compared to Hurricane Katrina.

More significant loss
than Hurricane Katrina.

But what’s really –
what’s really disturbing is the

insured loss for Hurricane Katrina.
About half of that loss was insured, but

no one has earthquake insurance, or very
few people have earthquake insurance.

And so that fact means it’ll be more
difficult to recover from an earthquake.

So you all might be wondering,
what would shaking from a

strong earthquake on the Hayward/
Rodgers Creek Fault look like?

Well, here’s estimated shaking
intensity from a magnitude 7.2

scenario on the Hayward and Rodgers
Creek earthquake with an epicenter –

where the earthquake began –
in San Pablo Bay.

And the red colors show the
strongest shaking, and you can see the

strong shaking is distributed throughout
the entire Bay Area, but it’s focused in

shallow basins between cities like Santa
Rosa, Livermore, Fremont, and San Jose.

Because the basin sediments
actually amplify ground motion.

So it’s interesting that –
you need to note that this scenario

was based on what was then
considered to be an unlinked step-over

between the Hayward
and Rodgers Creek.

And given the documented sensitivity of
ground motions to the shape of faults,

ground motions in a linked scenario
may be somewhat different

than what’s
pictured here.

The good news in
all this is that cities

like San Francisco are, in fact,
taking steps to prepare for

big earthquakes by passing
landmark earthquake retrofit laws.

And in terms of hazard planning,
thanks to the foresight of folks that

work at the USGS and other
geological agencies, they’ve prepared

for a linked Hayward/Rodgers Creek
Fault event, even though we didn’t have

direct evidence that the two
faults connected. So that’s good.

The Working Group on California
Earthquake Probabilities first included

a combined Hayward/Rodgers
Creek scenario in 1999.

And shaking estimates from such a
scenario are part of the current

National Seismic Hazard Map,
upon which building codes are made.

So they planned for the shaking
for an event of this size, even though

there wasn’t direct evidence
that the faults were connected.

And the Uniform California Earthquake
Rupture Forecast, UCERF3 –

not sure if you’ve heard of that –
considers the two faults as

separate segments capable of
rupturing together, which is good.

The new detailed information
on the connectivity of the Hayward

and Rodgers Creek Faults
will be incorporated into

future scientific studies that
model the behavior and

ground shaking of combined
Hayward/Rodgers Creek events.

And those results can be incorporated
into the next generation of USGS

hazard products and perhaps
go into improving some of

the hazard models
that we use today.

So, in closing, I want to remind
you all that – I like to tell people

that there’s no sense worrying about
things we can’t control – earthquakes.

But we can prepare for those things.
So we all need to be prepared

for the next big earthquake.
And even fault-obsessed geoscientists

need to be reminded to prepare
and take care of the earthquake kit,

replace your water, make sure
your family plan is in order.

And I just wanted to remind you
that there’s a lot of information

for what you can do to be
prepared online on websites.

Here I’ve taken a screen grab from the
Association of Bay Area Governments.

The USGS has information on their
websites that’s easily available.

If you want to go, how do I make
an earthquake kit, how do I –

how do I know if I need to
retrofit parts of my home,

how do I make a
family emergency plan.

So there are things
you can do to prepare.

And thank you for coming.

[ Applause ]

- Thank you, Janet.
That was a great talk.

We’ll now take questions
from the audience.

And I’d like you to use the microphones
that are in the aisles there for the benefit

of our online listeners. Or if you’d like,
I can bring you this microphone.

Thanks.

- It does work. You showed a slide
a few slides back with vertical bars

of the insured losses for – what
counties was that 24 billion for?

- That’s a good question. I’m not sure I …
- Oop. That’s the one.

[laughter]

So the 29.4 billion.

I’m not sure the answer to that question.
- Oh, okay.

- It’s a good question.
- Because I’ve considered the

earthquake insurance possibilities,
but apparently they’re only

underwritten to
about 11 or 12 billion.

And no one explains where the
extra money comes from.

[laughter]

- I’m just stating the facts.
I don’t – yeah. [laughs]

- Next question would be a question
of the fact that you said that the

energy of a fault on a magnitude
depends on the length of the fault.

- Mm-hmm.
- But that – does that also mean, though,

that since the longer fault
has greater energy,

it’s also spread out over a greater area,
so the energy per unit area

may not be that much worse
if it has a larger magnitude?

- Well, it depends how
close you are to the epicenter.

The shaking – it would be, in a longer
earthquake, in a larger earthquake,

that energy would be
split over a larger area.

But the closer you are to the epicenter,
the stronger that energy would be.

- Thank you.

- Yeah, I’m wondering about
the banding that occurred in the –

on the sediments.
So I’m assuming that’s a high –

or a small-particle, large-particle,
you know, sand versus mud.

And would that be due to
different flooding that would

come from – transport out of the
Sierras to the – to the delta?

- It could. Could be from
many different things.

- Okay. So could you name a
couple other things that it could be?

- Well, you can get different
thicknesses of sediment would

create a different acoustic
signal if the – thick versus thin.

Different sediment sizes –
grain sizes.

Yeah. You can get big
sand impulses seasonally.

You can get shifts in – hydraulic mining
occurred in the – in the bay during a

certain period, and there was a lot of
sedimentation associated with that.

And we see a marker in the bay,
in the cores, where we see basically

a change from where you’d
have that high sedimentation

to deeper in the core where you
see mostly sand layers and less mud.

So it can be environmental changes
that occur that cause sort of

discontinuities in the sedimentary
layers. Things like that.

- Okay. Thanks.
- Mm-hmm.

- I have a general question.
It goes beyond what you talked about.

And that is, when one looks at
the Hayward Fault and the

San Andreas system, they’re parallel.
They don’t seem too far apart.

Do they interact? If one moves,
does it relieve stress on the other?

If one moves, can it trigger the other?
- That’s a great question.

[laughter]

Possibly.
[laughter]

Yes. All these fault zones interact on
some level – the answer to your question.

Depends how close they are and
where they are relative to one another.

But you’re right. In an event
on the San Andreas Fault,

it imparts stress on certain
parts of other faults.

And it releases stress on
certain parts of other faults.

And yes, that has an effect on
this overall stress budget.

And there is evidence that
earthquakes on one fault do,

in fact, dynamically trigger
earthquakes on other nearby faults.

That’s not my expertise, but that
is known to happen, yeah.

- Great talk. Thanks.
One of your early maps had the

shake profile, but the color
of the Bay Area was blue.

I mean, does that mean that the
bay doesn’t shake?

[laughter]

- Maybe I’ll try to find that map.
I’m trying to – was it, like,

the second slide?
- Yeah.

- Okay.

Here?
- Yeah, over on – over on the right.

There’s the …
- It should come back up. Okay.

Oh. That’s because the underwater areas
have just been blued out. [laughter]

Yeah. They need to have a better
appreciation for the underwater areas.

In reality, that red should
extend across the bay.

- You have a later map that
projected that showed it …

- Yeah, I think the – on the last –
second-to-last slide, that water body

blue color was removed, and you could
actually see the extent of the shaking.

So that’s a good observation.
- When you’re looking for data at a

place like that, how accurately can you
measure the location of small faults?

I mean, could you just surround
the place with a whole bunch of

seismometers and wait
a couple years and look at

all the data from
the little ones?

- Well, how we found –
find active faults is not usually

with earthquake seismometers.
It’s with actually geology – going on

the ground and looking at morphology –
scarps that form on the surface.

And then we actually
dig into the earth and look at

what happens below the surface.
Sometimes active faults that we see

on the surface are associated
with seismicity, or earthquakes,

at depth, but sometimes
they’re not.

So …

- On the shaking – on the
shaking map that you did show us,

where was the epicenter?
- In San Pablo Bay.

- Oh, in the bay.
- In the bay.

Because it’s a segment boundary,
and segment boundaries are thought to

be places where stress builds up and
earthquakes can begin. Or end.

- Thank you. So on your cores –
so at some point, the bay was dry.

It was about 10,000 years ago,
during the Ice Age, or so.

So do you know if your 4-meter cores –
was that all underwater sediments?

Or were there any sediments that
were laid while it was dry?

- That’s a great question. Based on 
preliminary dates that we have

from those sediments, they were all –
they’re all too young.

They were underwater.
So our cores don’t go back

that far in time.
- Okay, thank you.

- Yeah.
Mm-hmm.

- That’s roughly what I was going to ask.
If, instead of measuring in meters,

you measure in years, how deep are
those profiles you were showing?

- Well, the two cores that we
have some preliminary dates on,

which are in the
northern part of the bay,

our cores extend into the
subsurface about 3 meters.

And the deepest date we have
is around 1400 A.D.

- Okay. So what …
- Recent.

- A seismologist says “recently active.”
What’s … [laughter]

Sorry. Wasn’t recently.
It was “currently active.” Yeah.

What’s current on your –
on your scale?

- Recently active, in terms of
earthquake hazard and the

Quaternary Fault Database –
Holocene active – the last 10,000 years.

- Thank you.

- Research …
- Is somebody asking a question? Oh.

- A research question.
It seems like you’ve got a really good,

dense set of data from the
acoustic chirp methods.

There’s a lot of software now
that’s been developed in the

petroleum industry for
three-dimensional interpretation.

I notice your plots are
all two-dimensional.

Do you have access, or might you be
able to put that into a three-dimensional

imager for a more detailed study?
- Yeah. We do have the ability

to put it into 3D. So that could
be a direction that we go in.

And that’s probably what
I’m going to end up doing next

to look at certain unconformities
or distinct layers I see in the bay

that I think indicate a time horizon.
And I’d like to trace those through

the seismic data and look at what
that looks like. So that’s on the list.

- I thought that your first image –
seismic section was outstanding.

You might want to bring that up.
On the left side, the sediments

are extremely deformed.
The right side, they’re not.

Did all that deformation
take place in one event?

Or is that just accumulation
of deformation on the left side

of the fault as imaged?

It’s your first seismic
section that came up.

- It’s – yeah.
- It’s really outstanding.

All that deformation
is to the left.

When did it happen?
At once or over time?

- We’re not quite sure yet.
It could have happened in one event,

or it could be a cumulative event.
The fact that the layers are all

bent the same amount from the
bottom through the top tells me

it’s probably close together in time
because they’re all the same shape.

There’s not a progression going on.
But we still don’t know.

The one thing you have to keep in mind
with looking at these sections

that I didn’t explain is, because
we’re looking at a very shallow section,

these sections
have been squeezed.

And so they’re highly
vertically exaggerated.

And that can sort of
mess you up – your eyes.

But that’s a good question.
That’s partly what we want to do

with the coring is get sort of
indications of time and sort of back out

how these events occurred and if there –
you know, what amount of time

would happen between, say,
that folding event and then

the layers on top of it
are relatively flat-lying.

- And the second question is, you, you
know, talked about risk and insurance.

Do you recommend buying
earthquake insurance?

[laughter]

- I will make no recommendations
on that topic. [laughter]

And USGS does not study risk.
[laughter]

- I’m curious about tsunami
hazard in – from a – up here …

- Sorry.

- … from a, you know, earthquake
along the San Andreas or this fault,

you know, from water inside the
bay hitting shores – you know, other side

of the bay or whatever. Just any interest
on – or any information on that?

- Well, anytime you have a
fault that could accommodate

vertical displacement – move the
seafloor up and down in an event,

could displace water
and create a tsunami.

But tsunamis can come on
many orders of magnitude.

So you can have a 20-centimeter
tsunami created from an earthquake.

Or you can have a 10-meter tsunami
created from an earthquake.

The largest tsunamis are usually created
in subduction zones where you have

vertical displacement at deep depths
further offshore can create a larger wave.

You can get tsunamis from
offsets of strike-slip faults –

the vertical offsets – and also
the horizontal offsets, they’re finding,

can create – can displace the
seafloor and create a wave.

And there was actually – there was an
earthquake – I’m blanking on the year.

It’s the Mare Island earthquake.
It happened in the 1800s,

and they don’t know exactly what fault
it happened on, but they recorded a

very small tsunami in the bay from that
on the order of, I think, a meter or less.

So, yes is the answer to –
they can happen. Yeah.

- I agree that that’s
a wonderful picture.

To my eye, there’s another fault there
that you didn’t mention anything about.

[laughter]

Just to the right of
the one you’ve labeled.

- Just to the right of
the one labeled?

- Yeah. There – well, a tenth of
the screen – one meter to the right,

but the vertical …
- This way?

- No. Over a little farther.
- Over here?

- Little farther to the right.
That one.

- Right there.
- Oh. Are you talking

about this vertical …
- Yeah.

- … apparently vertical – that’s the
tricky gas. You have to be careful.

Gas can jump from layer to layer without
a fault and have it appear that there is

a break there. When in fact, it’s just
the gas jumping around in the section.

So – and the reason I – the reason
I say it’s – I think it’s the gas is because

I can actually see concordant features
below that that are not offset.

So if you have a fault, all the
features below that, where you think

the fault is, need to be offset.
Does that make sense? Okay.

- In one of your slides, you described
the fault as a strike-slip fault,

which I thought meant lateral.
And you also describe it as have

fault dip – vertical. Which is it,
and how do those two work?

- Strike-slip faults predominantly
move laterally from one to the other.

But they also can accommodate
significant vertical motion.

- Both.
- Both. Yeah.

So technically, it’s an oblique slip fault.
[laughter]

Thank you.
That’s a great question.

[laughter]

- You had a slide showing a quake
in the Rodgers Creek/Hayward Fault

is three times as much
uninsured damage as Katrina.

And was that for a magnitude 7 quake
or a magnitude 7.4?

- Magnitude 7.
- Magnitude 7. So it could be

as much as a 7.4, and which really …
- Would likely be worse, yeah, damage.

Yeah.

- I’d like to explore a
worst-case scenario with you.

You mentioned earlier, you can have an
earthquake that triggers another quake

on another fault, perhaps – possibility
have one in the Bay Area now.

I think if we had one in the Hayward
Fault at 7 or more, we could feel it down

here on the peninsula too – this far away.
It wouldn’t be that big a deal.

Possibility, we have an earthquake
that triggers [inaudible] on a number of

different faults, you want
something that’s about 9 or 9.6.

Is that – you know, one reaction
to another one on different faults

at the same time, I get the [inaudible]
this place would look like Hiroshima

after we dropped the bomb on.
Wouldn’t be anything left.

What’s the possibility of that?

- It’s possible that all
these faults can link up.

But the likelihood of
that is extremely small.

Because I’ll remind you that large
earthquakes – very large earthquakes

happen very, very infrequently.
- Yeah.

- Moderate earthquakes happen
much more often in our lifetimes.

- Yeah.
- So – yeah.

- You have a number of 5s
in your lifetime.

Occasional 6 or 7 once in a while,
but we had – last time we had 8, that –

the ’06 quake in San Francisco.

- Yeah.
A large one in – and not

very many people were
living here at that time.

- No. No. I think those people
all passed away by now too.

- Yeah. I’m not – I’m not sure.
[laughter]

- All right, thanks.

- Hi. So I have
a quick question,

which is actually a
follow-up on this slide.

And it’s a two-part question.
So during a potential earthquake,

looking at the gaseous layer at the
bottom, how much of that could

potentially be released, in terms of –
given the fact that it’s a slip

that’s sliding and not
sort of an upward movement?

- Well, I don’t really know
the answer to your question.

But given the fact that it’s
not under a lot of pressure,

because it’s not very deep gas,
I don’t think it would

be released
catastrophically.

I don’t think it would
be released catastrophically.

But, yeah, likely some would be released
in some form, yeah, during shaking.

- So given that’s the case, how have you
been able to capture all this data so that,

if such a disruption takes place,
we still have all these beautiful

pictures that you’ve taken
that still show where the fault exists,

given the gas would
probably mess up this image?

- Well, I’m not sure
what the gas would do.

It might – if the gas were released,
it might clear up some of the image

for a short period of time [laughter]
before it has a change to re-accumulate.

Yeah. That’s a good question, but yeah,
don’t worry. We’ll hold onto these.

[laughter]

In the event that happens.

- You described the Hayward Fault’s
primary motion as being up and down.

How does – what does
the San Andreas Fault do?

And is its movement consistent along the
entire length of the San Andreas Fault?

- So if I – if you heard me describe
that the predominant movement

of the Hayward Fault is up
and down, then I misspoke.

What this technique measures,
we are only capable of detecting

vertical fault movement
with our 2D cross-sections.

Okay? So I can only measure
vertical fault motion.

Because I know this is a strike-slip
fault that has accommodated

a lot of motion through time,
I know that the horizontal motion

is likely much greater than the
vertical motion we measure here.

But what happens along a
fault zone is that, as the fault bends

and changes trajectory, it’s – the amount
of vertical motion to horizontal motion

changes quite dramatically within
tens of meters, hundreds of meters.

You can get – in one any on a fault,
depending if it’s a right bend or

a left bend, you can get the
land going down on one side,

and then all the sudden, that land
is going up on the other side.

So the motion along the fault
can change quite dramatically

from one place to the other –
the vertical motion.

So that’s another reason
why you want to – if you’re

characterizing a fault zone, you want
to take more than a look in one spot.

You need to look in many spots
along that fault to make sure you’re

properly characterizing its motion.
Because it does change.

- I have an
interactive question.

Okay, so you pointed
this out as the fault.

- Mm-hmm.

- But I was also curious –
can you explain these guys?

- Yeah. It’s another
form of deformation.

But because there’s no break,
it’s what I would call folding,

where the layers fold,
but they don’t actually break.

They could be folding
from a fault that’s actually

breaking layers further below
where we can’t image.

Or it could just be folding from …
- From pressure.

- From pressure, motions, yeah.
- Okay. Thank you.

- Well, I can make
a personal observation about

earthquake insurance,
if you don’t mind.

I think that in the next 50 years,
there’s probably about –

some reasonable probability
that my house in Palo Alto

could have a million dollars’
worth of damage.

Well, 50 years is a long time from now.
But it costs me about $2,000 a year

in earthquake insurance.
And so in 50 years, I’m only

spending $100,000 to save a
million dollars’ worth of damage.

And even if it took 100 years, I’d still be
cheaper off paying earthquake insurance.

If you don’t believe those probabilities,
of course, that’s doesn’t count.

If the earthquake is really not going
to take place for 200 years,

I’m wasting my money.
Who knows? [laughter]

- The research that you have done so far,
so we have found that the two creeks –

or, the two faults are connected.
So are they going to be considered as

one fault going forward? Or are they
still considered as two separate ones?

- That’s a good question.
I think it’s worth discussing

how they are considered.
Yeah.

- We could rename it the Watt Fault.

- No. We shouldn’t rename the fault.
[laughter]

But, yeah, I think what this work
and other work that are – that’s looking

at these fault intersections and proposed
segment boundaries is understanding,

okay, how do these
faults really connect?

And when we’re calling them a segment
boundary, and what that means when

you’re doing a hazard assessment is what
there needs to be a conversation about.

Recently – the last earthquake
rupture model for California, UCERF3,

has these segment boundaries in it,
but it allows those faults to

break together if those – if the two faults
are spaced a certain distance apart.

So based on just a simple, okay,
are they 5 kilometers or less

from each other, then they can –
then they can break.

But in the future, perhaps we could
use a more sophisticated way

of choosing whether or not these
faults go together based on earthquake

histories or folks that can do dynamic
models of how an earthquake rupture

behaves in any
given fault geometry.

There’s room
for improvement.

- Well, is it called a segment boundary
because when they first mapped the

Hayward Fault, they got to the northern
end of it, and it sort of stopped at

at the bay. And then when they’re doing
Rodgers Creek, at the southern end,

it sort of stopped at the bay.
So you’d say that’s a segment boundary?

- Yeah.
- Can you repeat his question?

- It’s like taking the train halfway
[inaudible] point, you know.

And they haven’t finished the last 100
yards, and you can see the tracks ahead.

- Well, I think it was –
the question is, why –

how is this defined as a
segment boundary to begin with?

Was it because they mapped it
to the Pinole Point and

couldn’t see it anymore?
And then could map it from

the north to the south and
couldn’t see it off Sears Point?

And they project the faults straight
from that point, and they’re separated.

And so, yeah, they call it
a segment boundary.

But there are other reasons that it’s
called – it’s a segment boundary.

And it’s because the
behavior of the Hayward Fault

is quite different from the behavior
of the Rodgers Creek Fault.

Okay? So the Hayward
Fault is a creeping fault.

Means it’s – at the surface,
it’s moving aseismically without –

you know, creeping –
offsetting curbs and the like.

And some – and that fault creep
takes up some of the stress

in between large earthquakes.
It’s still – the Hayward Fault

still has large earthquakes because the
full depth of the fault does not creep.

We know that only the surface
portion of the fault creeps.

The lower portion of the Hayward
Fault in a large area is locked.

Okay? So from what we
know of the Rodgers Creek Fault,

it is fully locked
north of the bay.

It may exhibit some creep further
to the north near Santa Rosa.

But there are other reasons
to create segment boundaries

besides just the
geometrical relationship.

Yeah.

- Question about the top gas that you
have marked here. It’s methane, right?

- Yes.

- Could you …
- Biogenic.

- … take a – or could you
speculate on the amount of

methane that might be there and
whether it’s commercially viable?

And would it be worthwhile …
[laughter]

… to frack it?
[laughter]

- I’m afraid not.
[laughter]

But if anyone would like to look
into that, they may use these data.

[laughter]

- Your – from my point of view,
what you’ve been talking about is –

from my point of view, what you’ve been
talking about is rather local, in a sense.

And the question is, how does
that relate to the – to the motion

of the platelets over
each other over time?

Is the source of all this the motion of the
platelets, or is it a local effect? [laughs]

- Both.
- Both.

- Yeah. So the San Andreas Fault
system is made up of a number

of different faults.
And especially in the Bay Area,

there are a number of faults that
accommodate this motion between

the Pacific Plate and
the North American Plate.

That’s where all this is coming from.
Okay, and so that motion between

the Pacific and North American Plate
is separated – divided up onto

these individual fault strands.
And we’ve got a pretty good idea

of how much is
proportioned onto each one.

But, yeah, and then there
are definitely local effects.

There are fault intersections,
and one fault will lump into

another one and impart
some motion on it.

So it’s gets more
complicated at a local level.

But basically, all this
deformation is because

these two plates are
sliding past one another.

- Thank you.
- Mm-hmm.

- Two quick questions. Or one quick
question and maybe a longer one.

So the vertical scale is 1 meter.
What’s the horizontal

scale on this graph?
- That’s a great question.

- Well, the vertical exaggeration
is 200 to 1, so …

- Okay.
- I can’t do math this late in the day, but …

- Okay.
[laughter]

- Yeah. It’s a few – I think it’s
probably a kilometer. Ish. [laughter]

- Okay. And the second thing
related to fracking out the methane.

I was at a presentation by another
scientist here about a year ago maybe,

and apparently there is quite a significant
signature of some heavy metals,

like arsenic and other stuff, contained in
the muds from the mining that went on.

And so they – basically,
don’t mess with the mud.

Just leave it in place.
It’ll stay there.

So if you start disturbing stuff, then
that might come out and be a problem.

- I agree.
We’re doing it on a very local scale.

And actually, those trace elements might
provide good markers of dates for offset

when we see the arrival of mercury
or something in the – in the core.

So they might
actually be helping –

helpful to understand
the earthquake history.

- I have a question about the more
southerly part of the county where it –

there probably are – is fracking
going on now, and there certainly

is an injection well where they dispose
of the huge amounts of water,

now contaminated, that’s been
used in the fracking process.

And there are multiple
faults in that area.

One goes right
through that area.

The San Andreas is about 2 miles away
from where the wells are now.

How close or – does the injection
of water in the ground have to be

to the fault to loosen it up
and have it move differently?

- I don’t know the answer
to your question. [laughs]

But it’s a good question.
- Huh? It was – yeah.

- But one wouldn’t …
- It would have to be probably

injected very deep to affect the fault.
I don’t know.

- All right. Looks like it’s
time to wrap up.

So thank you all for coming,
and thank you, Janet.

[ Applause ]

[background conversations]

[ Silence ]