[CEUS-earthquake-hazards] reply to Joe Tomasello; buildings codes and earthquake hazard

Fri Feb 15 11:12:52 MST 2008

Dear All,

Wonderful things can be done with statistics, when there is plenty of data.
Demographic numbers are available as data sets of millions of people and they
can tell insurers exactly how many will die at any age. However, those
statistics cannot tell you at what age any individual will die. Those
statistics, like earthquake statistics, are not site specific. Worse, for
earthquakes there is hardly any data.  Every new earthquake changes the local
data base enormously and that changes the Gutenberg-Richter probability
projection. Additionally, the projected b-line is represented through time by
no data at all. Dozens of studies show that earthquake recurrences are not
regular and that the linear-projected b-line is not valid. So, what does a
2-percent exceedance in x-years really mean?

You need to go back to the basics and ask some hard questions.

Ellis

-----Original Message-----
From: ceus-earthquake-hazards-bounces at geohazards.usgs.gov
[mailto:ceus-earthquake-hazards-bounces at geohazards.usgs.gov] On Behalf Of
Wang, Zhenming
Sent: Friday, February 15, 2008 9:35 AM
To: Arthur D Frankel
Cc: ceus-earthquake-hazards at geohazards.usgs.gov; James Cobb; Keifer, John D
Subject: Re: [CEUS-earthquake-hazards] reply to Joe Tomasello; buildings
codes and earthquake hazard

Art,

This is the first time we see these comparisons:

"When I talk to code committees and other groups, I compare the relative
level of protection that designing to different probability levels of ground
shaking will provide to buildings.  This can be assessed by comparing the
ground-motion values for the probability levels in the building codes to the
median ground motions expected when the next 1811-12 type New Madrid
earthquake occurs and by comparing code values to intensities observed in the
Memphis area from the 1811-12 earthquakes."

The selected design ground motion should be consistent with the scientific
facts. However, these comparisons seem to be contradictory to your early
statements:

1.	"In fact, we release seismic hazard curves (a range of ground motion,
from 0.0 to 10g or larger) for a grid of sites across the nation, so that
users can calculate the ground motions at any probability level they choose."
"It should also be reiterated that the national seismic hazard maps are based
on the average hazard curves from a variety of input models and attenuation
relations; they are not worst-case maps."  
2.	"It is not correct to compare the intensity observations from
1811-1812 with the probabilistic hazard maps that also include the hazard
from earthquakes closer to St. Louis." 

Thanks.

Zhenming

________________________________

From: ceus-earthquake-hazards-bounces at geohazards.usgs.gov
[mailto:ceus-earthquake-hazards-bounces at geohazards.usgs.gov] On Behalf Of
Arthur D Frankel
Sent: Thursday, February 14, 2008 3:35 PM
To: ceus-earthquake-hazards at geohazards.usgs.gov
Subject: [CEUS-earthquake-hazards] reply to Joe Tomasello; buildings codes
and earthquake hazard

Joe, 

The USGS policy is to support the process of the Building Seismic Safety
Council (BSSC) establishing probability levels and design procedures for the
national model building codes, such as the International Building Code.  The
BSSC membership consists of a large group of engineers and stakeholders. The
BSSC is a council of the National Institute of Building Sciences. The code
development process of the BSSC is funded by FEMA. The design procedures are
published in the NEHRP Recommended Provisions for the Development of Seismic
Regulations for New Buildings, which is written by the BSSC/NIBS and
published by FEMA. 

I think a key responsibility of the USGS is to provide the best scientific
information to decision makers.  Part of this scientific information is
assessment of the ground motions from the 1811-12 earthquakes and estimation
of the ground motions for the next 1811-12 type earthquake. 

When I talk to code committees and other groups, I compare the relative level
of protection that designing to different probability levels of ground
shaking will provide to buildings.  This can be assessed by comparing the
ground-motion values for the probability levels in the building codes to the
median ground motions expected when the next 1811-12 type New Madrid
earthquake occurs and by comparing code values to intensities observed in the
Memphis area from the 1811-12 earthquakes.   

For example, the value of ground motions around 1 Hz with a 10% probability
of exceedance in 50 years (10%/50) is substantially lower than the median 1
Hz ground motion expected for the next 1811-12 type earthquake.   The new
Memphis code adopted in 2006 uses the 10%/50 year ground motions from the
1996 vintage of the national maps (the 2002 maps are higher).  Here I am
considering 1 Hz spectral accelerations (S.A.), which are used for the design
of buildings with about 10 stories.  For a site in Memphis (35.15 N; 90.05
W), the 10%/50 value of  1 Hz spectral acceleration is 0.16g (from the 1996
maps and using an amplification factor of 2.4 for class D stiff-soil site
relative to firm-rock site from the NEHRP amplification factors).  This is
much lower than the median 1 Hz S.A. of 0.36g expected in Memphis from a
scenario earthquake with moment magnitude 7.7 located on the portion of the
current New Madrid seismicity trend northwest of Memphis (using the
stiff-soil amplification factor from the NEHRP factors). This calculation of
the expected spectral acceleration is based on the average of the five
attenuation relations used in the 2002 national maps.  If the next large New
Madrid earthquake was a moment magnitude 7.4, the calculated median 1 Hz S.A.
at Memphis would be 0.29g for a stiff-soil site, still much higher than the
10%/50 value (0.16g) from the 1996 maps.   

The 10%/50 values for 1 Hz S.A. from the 2002 hazard maps would still be
significantly lower than the scenario ground motions. For 5 Hz S.A., the
expected values of the median ground motions for a M7.7 earthquake are more
sensitive to assumptions on the nonlinearity and attenuation of sediments in
the Mississippi Embayment. 

The International Building Code (IBC) uses spectral accelerations that are
2/3 times the values with a 2% probability of exceedance in 50 years for most
of the nation (there are some areas where IBC uses the median deterministic
ground motions, depending on the level of the probabilistic motions). For 1
Hz S.A., the 2006 IBC specifies a value of 0.42g for Memphis (stiff-soil
site). This is similar to the median value of 0.36g expected for a M7.7
earthquake (see above). 

In my presentations, I also compare the code values to the ground motions
estimated from intensity reports in the Memphis area from past earthquakes.
Here I use peak ground accelerations (PGA) rather than spectral
accelerations, because intensities are generally correlated in the literature
with PGA's or peak ground velocities. 

Intensities in the Memphis area during the 1811-12 earthquakes have been
assigned as intensity VIII by Hough et al. (2000) and as intensity X by
Johnston (1996), depending on their interpretation of earthquake effects.
Intensity VIII corresponds to a peak ground acceleration between about 0.34
and 0.65g, based on the work David Wald did for Shakemap.  This range is
consistent with the calculated median PGA at Memphis of 0.39g for a M7.7
earthquake determined from the average of 5 attenuation relations used in the
2002 hazard maps (using the NEHRP amplification factors). The calculated PGA
for a M7.4 earthquake is 0.32g, close to the range of the PGA's estimated for
intensity VIII. 

 The new Memphis code procedure of using the 10%/50 values from the 1996 maps
results in a PGA of 0.23g (for a stiff soil site), which is substantially
lower than the range of ground motions estimated from the intensities
reported in Memphis during the 1811-12 earthquake sequence (0.34-0.65g for
intensity VIII).  A similar value of PGA (0.22g) is found by taking the 5 Hz
S.A. with 10%/50 and dividing by 2.0, which is the factor relating PGA to 5
Hz S.A. derived for M7.4-7.7 earthquakes from the average of the five
attenuation relations. 

 Using a PGA that is 2/3 times the PGA with 2% probability of exceedance in
50 years (2%/50), which corresponds to the procedure used for spectral
accelerations in the 2006 International Building Code for the Memphis area,
gives a PGA value of 0.50g for Memphis (stiff soil site), which is in the
range of the values estimated from the 1811-12 intensities.  This is similar
to the PGA value of 0.47g derived from the 5 Hz S.A.in the IBC divided by a
factor of  2.0 to convert to PGA. So, there is evidence from intensity data
that the ground motions specified in the IBC have been experienced in Memphis
during the 1811-12 earthquakes 

In summary, the 1 Hz spectral accelerations with a 10% probability of
exceedance in 50 years, as used in the current Memphis code, are
substantially lower than the median 1 Hz spectral accelerations expected for
the next 1811-12 type earthquake. The 10%/50 value of PGA is  probably lower
than the ground shaking experienced in Memphis during the 1811-12
earthquakes, based on intensity data. 

The 1 Hz spectral accelerations specified in the International Building Code
(2/3 times the motions with a 2% probability of exceedance in 50 years) are
similar to the  median  1 Hz spectral accelerations expected for the next
1811-12 type earthquake. The PGA with 2/3 times the value with 2%/50 is
probably comparable to the PGA experienced in Memphis from the 1811-12
earthquakes, based on intensity data. 

-Art 

Art Frankel
U.S. Geological Survey
MS 966, Box 25046
DFC
Denver, CO 80225
phone: 303-273-8556
fax: 303-273-8600
email: afrankel at usgs.gov