WEBVTT
Kind: captions
Language: en-US

00:00:01.773 --> 00:00:05.360
[silence]

00:00:05.360 --> 00:00:08.160
Hello, greetings.
My name is Sean Ahdi.

00:00:08.160 --> 00:00:12.320
It is snowing here in Denver,
Colorado, but it’s sunny in

00:00:12.320 --> 00:00:16.480
San Francisco, as my background
shows. I would like to introduce

00:00:16.480 --> 00:00:20.000
myself and thank you for
inviting me to this workshop.

00:00:20.000 --> 00:00:24.880
I am a research geophysicist newly
hired at the Geological Hazards

00:00:24.880 --> 00:00:30.480
Science Center in Golden, Colorado,
and also a lecturer at the UCLA

00:00:30.480 --> 00:00:34.560
Department of Earth, Planetary, and
Space Sciences. And today I’m going to

00:00:34.560 --> 00:00:40.153
talk about why we are still using
Vs30 for seismic site characterization.

00:00:42.800 --> 00:00:47.120
Like to acknowledge my USGS
colleagues Alan Yong, Grace Parker,

00:00:47.120 --> 00:00:51.680
my supervisor Morgan Moschetti,
the National Seismic Hazard Model

00:00:51.680 --> 00:00:56.296
Group, and the Near-Real-Time
Products Group, and other folks

00:00:56.320 --> 00:00:59.760
in the Earthquake Science Center
with whom I’ve collaborated over

00:00:59.760 --> 00:01:03.760
the years and discussed this work.
And a special shout-out to

00:01:03.760 --> 00:01:08.240
Jack Boatwright, who very graciously
invited me to give a presentation

00:01:08.240 --> 00:01:12.320
at this workshop a few years ago
when I was still just a student.

00:01:12.320 --> 00:01:15.120
I’d like to acknowledge my
UCLA Shear Wave Velocity

00:01:15.120 --> 00:01:19.920
Profile Database project team.
Professor, PI, and my adviser

00:01:19.920 --> 00:01:23.256
Jonathan Stewart and
Professor Scott Brandenberg.

00:01:23.280 --> 00:01:27.430
Dong Youp Kwak, Paolo Zimmaro,
Pengfei Wang, and Professor Yousef

00:01:27.430 --> 00:01:31.736
Bozorgnia, and many others
who helped with this effort.

00:01:31.760 --> 00:01:35.040
Numerous individuals and agencies
provided their data and their time to

00:01:35.040 --> 00:01:39.600
input this data into our databases that
I will talk about briefly at the end.

00:01:39.600 --> 00:01:43.976
And support from the USGS
California State Geologic Survey,

00:01:44.000 --> 00:01:48.400
COSMOS, PEER, UCLA,
PG&E, SoCal Edison, and BC Hydro

00:01:48.400 --> 00:01:51.016
at various times
throughout the projects.

00:01:51.040 --> 00:01:54.400
A quick outline of my talk.
I will talk about the history of use of

00:01:54.400 --> 00:01:59.360
Vs30 for seismic site characterization.
Arguments for and against Vs30

00:01:59.360 --> 00:02:02.856
and other site parameters
used for site characterization.

00:02:02.880 --> 00:02:07.840
Ongoing efforts supported by Vs30
data collection in NorCal and beyond

00:02:07.840 --> 00:02:13.096
and the products that exist that
show shear wave velocity data.

00:02:13.120 --> 00:02:16.536
How about some odes
to Vs30, if you will?

00:02:16.560 --> 00:02:18.720
Some quotes I’ve
heard over the years.

00:02:18.720 --> 00:02:21.656
It is what everyone
loves to hate.

00:02:21.680 --> 00:02:25.280
It’s a curious dilemma.
We don’t really like to use Vs30,

00:02:25.280 --> 00:02:29.256
but using anything better
introduces other challenges.

00:02:29.280 --> 00:02:34.936
Are there any other things
we can look into using?

00:02:34.960 --> 00:02:40.640
With a laugh, watching people who try
to correlate Vs30 with Z2.5, which is

00:02:40.640 --> 00:02:46.000
the depth to the 2-1/2 kilometer per
second shear wave velocity surface,

00:02:46.000 --> 00:02:52.696
which correlates with basin depth.
And the use and abuse of Vs30.

00:02:52.720 --> 00:02:56.400
So what exactly is Vs30 we
keep waving our arms about?

00:02:56.400 --> 00:02:59.680
It’s the time-averaged shear wave
velocity in the upper 30 meters

00:02:59.680 --> 00:03:04.376
of a site. I emphasized “time”
so that we don’t average over depth.

00:03:04.400 --> 00:03:08.960
In our one-dimensional
Vs profile shown here, the depth –

00:03:08.960 --> 00:03:14.000
maximum depth is Z-sub-p.
And we can take that and divide it

00:03:14.000 --> 00:03:16.320
by the total travel time to
go through all these layers,

00:03:16.320 --> 00:03:19.600
which is the time-averaged
velocity in the whole profile.

00:03:19.600 --> 00:03:24.880
That travel time is the sum of
these velocities divided by –

00:03:24.880 --> 00:03:27.840
layer thicknesses divided
by velocities – excuse me.

00:03:27.840 --> 00:03:32.880
Plug that in. Get that travel time.
And replace the profile depth

00:03:32.880 --> 00:03:37.142
with 30 meters, and you
have Vs30. It’s very simple.

00:03:37.920 --> 00:03:41.760
Over time, in the past, we had –
originally had qualitative and

00:03:41.760 --> 00:03:46.160
subjective site descriptions, such as
soil and rock, stiff and soft, etc.

00:03:46.160 --> 00:03:50.640
But Roger Borcherdt of the USGS
in 1970 identified the relation of

00:03:50.640 --> 00:03:55.040
local geology, including its shear
wave velocity of the materials to

00:03:55.040 --> 00:03:59.360
the amplification of ground motion.
And the USGS commenced a

00:03:59.360 --> 00:04:02.160
Vs measurement campaign,
starting with downhole profiles,

00:04:02.160 --> 00:04:09.280
in the early to mid-1970s and released
many Open-File reports to date,

00:04:09.280 --> 00:04:13.840
including the starting views of
non-invasive methods as well –

00:04:13.840 --> 00:04:16.456
besides invasive
ones in boreholes.

00:04:16.480 --> 00:04:20.400
Dave Boore and others first adopted
Vs30 for their ground motion

00:04:20.400 --> 00:04:24.720
attenuation relations in 1993,
and there’s a screenshot of their

00:04:24.720 --> 00:04:30.000
model here, and the Vs30 values
for different qualitative site classes

00:04:30.000 --> 00:04:32.720
that they used
in their effort.

00:04:32.720 --> 00:04:36.720
The National Earthquake Hazards
Reduction Program, or NEHRP, started

00:04:36.720 --> 00:04:42.320
adopted Vs30 for building codes after
Roger Borcherdt’s paper in 1994.

00:04:42.320 --> 00:04:46.400
And preliminary – it’s – I’m sorry –
it’s the primary term that is used

00:04:46.400 --> 00:04:49.920
in ergodic site response models
implemented in ground motion

00:04:49.920 --> 00:04:55.256
models – GMMs –
or called GMPEs – to date.

00:04:55.280 --> 00:04:58.800
So why are we using Vs30?
It’s a single continuous variable

00:04:58.800 --> 00:05:02.696
instead of those discrete classes
we talked about earlier.

00:05:02.720 --> 00:05:06.160
As Dave Boore showed in his
2004 study, there is a reduction

00:05:06.160 --> 00:05:10.480
in uncertainty when you use
Vs30 compared to discrete site

00:05:10.480 --> 00:05:14.536
classifications or using nothing
to classify your site at all.

00:05:14.560 --> 00:05:18.720
It effectively captured first-order
linear and nonlinear site effects.

00:05:18.720 --> 00:05:21.040
The nonlinear is shown
at the top of the page,

00:05:21.040 --> 00:05:26.136
and the linear correlates –
linear ground – linear scaling

00:05:26.160 --> 00:05:30.776
of Vs30 with ground motion
in the bottom set of plots.

00:05:30.800 --> 00:05:34.160
Vs30 correlates with shallower
and deeper sedimentary structure,

00:05:34.160 --> 00:05:37.920
as Dave Boore and his colleagues
showed in 2011 in this global

00:05:37.920 --> 00:05:42.720
data set of Vs30, which correlates
to deeper time-average values –

00:05:42.720 --> 00:05:47.656
say, Vs100 or even
up to Vs600 in Japan.

00:05:47.680 --> 00:05:51.360
And Vs30 has been shown to correlate
with basin sediment thickness terms

00:05:51.360 --> 00:05:56.960
like that Z2.5 parameter or
Z1.0 that we discussed earlier.

00:05:56.960 --> 00:06:02.640
Ned Field of the USGS showed that –
in 2000 that ground motion residuals

00:06:02.640 --> 00:06:05.680
correlate at long periods
with these basin depth terms.

00:06:05.680 --> 00:06:10.720
And Ebuka Nweke and others at
UCLA in recent years have updated

00:06:10.720 --> 00:06:16.640
correlation terms for Z1.0, which is
the 1-kilometer isosurface depth –

00:06:16.640 --> 00:06:22.456
1 kilometer per second shear
wave velocity correlated to Vs30.

00:06:22.480 --> 00:06:25.920
Continuing on, other parameters
may not be obtainable at every site,

00:06:25.920 --> 00:06:28.320
such as fundamental
or dominant site frequency,

00:06:28.320 --> 00:06:30.696
or those basin
depth terms – Z-sub-X.

00:06:30.720 --> 00:06:36.160
This plots shows horizontal-to-vertical
spectral ratio – HVSR – curves that

00:06:36.160 --> 00:06:41.256
are collected at a granite outcrop
site in southern – near Ridgecrest.

00:06:41.280 --> 00:06:48.456
And we cannot identify a predominant
site frequency at this site.

00:06:48.480 --> 00:06:50.640
Vs30, as I mentioned,
as been started –

00:06:50.640 --> 00:06:54.616
has been adopted by building codes,
and we continue to use them.

00:06:54.640 --> 00:06:58.240
It correlates to high-frequency
attenuation parameter, kappa,

00:06:58.240 --> 00:07:01.176
and Fourier amplitude
spectral space.

00:07:01.200 --> 00:07:04.080
And it’s much more readily
available than other site parameters.

00:07:04.080 --> 00:07:08.400
This is a screenshot of the USGS
Vs30 data compilation website,

00:07:08.400 --> 00:07:13.600
which is active and available, has over
4,000 Vs30 sites and counting.

00:07:13.600 --> 00:07:17.576
We are ongoingly
updating this database.

00:07:17.600 --> 00:07:22.400
So why would we not use Vs30?
Arguments against its use.

00:07:22.400 --> 00:07:28.560
Well, in our classic schematic of
interpreting how ground motions

00:07:28.560 --> 00:07:32.776
occur at a site, we know that
it’s a combination of source,

00:07:32.800 --> 00:07:39.760
wave propagation path, and site effects.
And one Vs30 value does not indicate

00:07:39.760 --> 00:07:45.576
our complexity here.
And our one Vs profile for our site

00:07:45.600 --> 00:07:50.080
can actually give us a Vs30 value
which can be represented by

00:07:50.080 --> 00:07:54.480
many Vs profiles. It’s not
really a unique problem.

00:07:54.480 --> 00:07:58.640
Site amplification is directed also
by the directivity of the ground motion

00:07:58.640 --> 00:08:02.240
and topographic effects.
And the complex geology we see

00:08:02.240 --> 00:08:06.880
in this plot in active tectonic regions,
especially with basin edges,

00:08:06.880 --> 00:08:09.840
cannot be captured by
one parameter as well.

00:08:09.840 --> 00:08:14.320
H-over-V peak frequency
better predicts the site amplification

00:08:14.320 --> 00:08:17.520
at frequencies near the
fundamental period of the site,

00:08:17.520 --> 00:08:20.136
or frequency
of the site.

00:08:20.160 --> 00:08:23.920
And, just in summary, it’s not
a fundamental site parameter.

00:08:23.920 --> 00:08:28.160
It’s a proxy for site response,
as has been stated in the literature

00:08:28.160 --> 00:08:30.535
and discussed
over the years.

00:08:31.360 --> 00:08:36.640
Just a quick argument against the use
of fundamental frequency in ground

00:08:36.640 --> 00:08:41.040
motion models instead of Vs30.
As Grace Parker, our colleague

00:08:41.040 --> 00:08:44.960
in the ESC, has shown, a consensus
has not yet emerged on the

00:08:44.960 --> 00:08:48.800
appropriateness of estimating site
parameters from attributes of

00:08:48.800 --> 00:08:52.720
ground motion recordings when the
performance of the resulting ground

00:08:52.720 --> 00:08:55.600
motion models is then judged
against those same recordings.

00:08:55.600 --> 00:09:01.816
So this is a circular argument that
we are doing ongoing research

00:09:01.840 --> 00:09:05.600
to ascertain whether or not
this is going to be figured out.

00:09:05.600 --> 00:09:08.880
I’m going to put a plug in for my
co-authors for two abstracts that

00:09:08.880 --> 00:09:15.920
will be presented at SSA to understand
the consistency of using H-over-V

00:09:15.920 --> 00:09:19.120
from earthquakes and from
other measurements to then

00:09:19.120 --> 00:09:25.200
assign them to station
characteristics for recording stations

00:09:25.200 --> 00:09:28.456
in ground motion
model development.

00:09:28.480 --> 00:09:31.360
I’m going to make a quick argument
for slowness, which is the inverse

00:09:31.360 --> 00:09:36.160
of velocity. You see two figures here.
One is shear wave velocity and

00:09:36.160 --> 00:09:39.360
one is shear wave slowness
at a site in southern California.

00:09:39.360 --> 00:09:44.800
Travel time is linearly proportional to
S-S, instead of inversely proportional

00:09:44.800 --> 00:09:49.120
to Vs, which makes it less complicated.
And thus, it can be linearly averaged

00:09:49.120 --> 00:09:52.960
to get the depth average values,
which would alleviate the concern

00:09:52.960 --> 00:09:56.000
I mentioned earlier when we
defined Vs30, that you have to

00:09:56.000 --> 00:09:58.880
do a time-averaged velocity.
And some people have been

00:09:58.880 --> 00:10:02.616
making this mistake
in industry applications.

00:10:02.640 --> 00:10:07.416
Statistical properties fitting travel time
data apply to S-S rather than Vs.

00:10:07.440 --> 00:10:12.000
And S-S emphasizes the material
properties that amplify ground motions

00:10:12.000 --> 00:10:15.760
rather than diminishing them in
the velocity profile or emphasizing

00:10:15.760 --> 00:10:19.656
stiffer materials that are
less important for site response.

00:10:19.680 --> 00:10:24.000
Another plug for slowness is recent
work we’ve done with Utkarsh Mital

00:10:24.000 --> 00:10:31.256
and others at the USGS and Japanese
colleagues that we’re looking at

00:10:31.280 --> 00:10:36.056
S-S30 distributions that can
be fit with a normal distribution.

00:10:36.080 --> 00:10:39.440
But then the inverse of that,
like the inverse of S-S is Vs,

00:10:39.440 --> 00:10:42.160
would be the reciprocal
normal distribution.

00:10:42.160 --> 00:10:46.720
And we’re trying to investigate the
efficacy of this in describing Vs30

00:10:46.720 --> 00:10:51.360
data sets as compared to the
longstanding belief that Vs30 data sets

00:10:51.360 --> 00:10:54.000
have been fit with
log normal distributions.

00:10:54.000 --> 00:10:57.749
That might just be
statistical and not physical.

00:10:58.880 --> 00:11:03.280
Quick plug for V-R40, which is the
40-meter wavelength Rayleigh wave

00:11:03.280 --> 00:11:07.440
phase velocity that can be
easily obtained from surface wave

00:11:07.440 --> 00:11:11.840
measurements, such as the
SASW surface – spectral analysis

00:11:11.840 --> 00:11:15.280
of surface waves. And Alan Yong
and others have shown that

00:11:15.280 --> 00:11:19.120
there’s almost a 1-to-1 correlation
between Vs30 and V-R40.

00:11:19.120 --> 00:11:22.640
So, while we’re still – would be
using Vs30 if we implemented

00:11:22.640 --> 00:11:29.176
V-R40 correlation, it gives us another
way to quickly obtain the parameter.

00:11:29.200 --> 00:11:33.256
So concluding remarks for
why we’re using these terms.

00:11:33.280 --> 00:11:37.680
Most major ground motion models
and building codes use Vs30 because

00:11:37.680 --> 00:11:40.800
of widespread data accessibility
that continues to grow, and the

00:11:40.800 --> 00:11:43.760
momentum that is built, it just –
we’re just kind of taking advantage

00:11:43.760 --> 00:11:45.040
of that and moving
along with it.

00:11:45.040 --> 00:11:49.096
Admittedly, it has been
used for a long time.

00:11:49.120 --> 00:11:53.440
It’s a stable measurement across
multiple measurement methods

00:11:53.440 --> 00:11:58.880
so that we can now start to ascertain
epistemic uncertainties at our sites

00:11:58.880 --> 00:12:04.470
across methods and across
soil types and geologic regimes.

00:12:05.520 --> 00:12:09.200
The arguments against Vs30
are still value – it is just a proxy.

00:12:09.200 --> 00:12:12.376
It’s not a
physical parameter.

00:12:12.400 --> 00:12:15.360
And it’s an index parameter.
It cannot capture the effects of

00:12:15.360 --> 00:12:20.696
basins and other complicated
three-dimensional geologic

00:12:20.720 --> 00:12:23.760
constraints at our site.
And so that means there’s

00:12:23.760 --> 00:12:28.456
definitely room for more work,
and this is an active area of research.

00:12:28.480 --> 00:12:33.440
I will – I will conclude this part
of the talk with John Anderson from

00:12:33.440 --> 00:12:38.480
University of Nevada-Reno reminding
us that the upper 30 meters only

00:12:38.480 --> 00:12:43.120
represents less than 1% of the path
from the source to the site, but this

00:12:43.120 --> 00:12:47.920
Vs30, the observed value of kappa,
attenuation, and resonant frequencies

00:12:47.920 --> 00:12:50.880
that we mentioned might be
sufficient to characterize a site

00:12:50.880 --> 00:12:55.176
where 1D models are appropriate.
So there’s still use in it.

00:12:55.200 --> 00:12:59.680
Shotgun efforts showing you what
Vs data is being used and the

00:12:59.680 --> 00:13:05.760
data sources that we have so far.
Ongoing work to update the National

00:13:05.760 --> 00:13:11.040
Seismic Hazard Model in Hawaii
where we have compared Vs30 and

00:13:11.040 --> 00:13:14.640
site amplification models from
existing ground motion models

00:13:14.640 --> 00:13:20.696
with Hawaiian data. This is
unpublished work, and it’s ongoing.

00:13:20.720 --> 00:13:25.496
CEUS – central and eastern U.S.,
Atlantic and Gulf Coastal plain

00:13:25.520 --> 00:13:28.800
amplification models that have been
published, and we are collecting

00:13:28.800 --> 00:13:36.720
a new database of Vs30 and Vs profiles
to determine the efficacy of those

00:13:36.720 --> 00:13:40.960
amplification models and ongoing
updates to the National Seismic Hazard

00:13:40.960 --> 00:13:46.136
Model that will be published,
hopefully, in 2023.

00:13:46.160 --> 00:13:51.840
The USGS hosts a Vs30 Map Viewer,
which has an underpin of the

00:13:51.840 --> 00:13:55.840
topographic slope-based Vs30 that
was initially proposed by Dave Wald

00:13:55.840 --> 00:14:00.480
and Trevor Allen and has since been
updated by Eric Thompson and others

00:14:00.480 --> 00:14:05.600
and put into a mosaic that the USGS –
mosaic map that the USGS hosts that

00:14:05.600 --> 00:14:11.200
includes effects of surface geology
and topography and those Vs30

00:14:11.200 --> 00:14:14.160
measurements from the data
compilation that I mentioned earlier.

00:14:14.160 --> 00:14:18.480
We are hoping to expand these
mosaic inlays for specific

00:14:18.480 --> 00:14:22.545
geographic regions in
other parts of the U.S.

00:14:23.200 --> 00:14:27.520
The data compilation that I mentioned
earlier is online, and we encourage

00:14:27.520 --> 00:14:31.976
people to access it and
collect the data that they need.

00:14:32.000 --> 00:14:39.040
And the U.S. – the UCLA
organized project – the Vs Community

00:14:39.040 --> 00:14:44.936
Profile Database, which now has
H-over-V, or HVSR, measurements

00:14:44.960 --> 00:14:49.280
implemented in it – these are –
this is the map of sites to date.

00:14:49.280 --> 00:14:52.960
And a smattering of Lakers
colors across the Bay Area

00:14:52.960 --> 00:14:56.616
for you Warriors
fans out there.

00:14:56.640 --> 00:15:02.240
And this is the online web interface
for the UCLA Vs Profile Database

00:15:02.240 --> 00:15:06.320
and now has – that now has the
HVSR measurements as well.

00:15:06.320 --> 00:15:11.920
We are – we have thousands of sites,
and we continue to add to them.

00:15:11.920 --> 00:15:15.360
And I thank you for your time.
If, for no other reason that

00:15:15.360 --> 00:15:18.880
I mentioned herein, we should
continue to collect Vs30 and

00:15:18.880 --> 00:15:23.816
other site data so that we can get out
in the field and enjoy field work.

00:15:23.840 --> 00:15:25.840
Thank you.

00:15:27.550 --> 00:15:32.375
[silence]