Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

Seismic Zones in Texas - Some USGS Comments

[Subject Prev][Subject Next][Thread Prev][Thread Next]
	Frank McClure informed me a few days ago about the recent round of
correspondence regarding ground motion maps on this site, starting with the
subject "Seismic Zones in Texas".  After an absence of about 3 weeks I have
once again subscribed to the server.   I have been reading the
correspondence Frank referred to as well as subsequent correspondence
regarding ground motion maps - those published by USGS, those included with
the 1997 NEHRP Seismic Design Provisions (hereafter referred to as the
Provisions), and those proposed for the 2000 IBC.  The proposed IBC maps
are the same as the Provisions maps except in number and in scale.

	Since I work with the group that prepared the USGS maps I can say what
they represent, how they were developed, and how they can be used.  I also
worked on the BSSC committee that proposed the rules for developing the
Provisions maps and I (and one of my colleagues) applied committee rules to
prepare those maps.  We also prepared the originals of the maps that are
proposed for the IBC, following the recommendations of the BSSC and the
IBC.  Individuals may or may not like the results of the processes that
were used but this posting is an attempt to have all of us on the same page
when critiquing the results of the processes.

TWO SETS OF MAPS

It is important to understand that there are two sets of maps.  USGS
prepared one set of maps that are probabilistic (more on this later).  Two
of the USGS maps (0.2 sec spectral acceleration for 2% in 50 years and 1.0
sec spectral acceleration for 2% in 50 years)  were selected by a BSSC
technical committee to serve as the basis of the so-called Maximum
Considered Earthquake (MCE) Ground Motion Maps that serve as the basis for
the Provisions and the 2000 IBC.  These two maps were modified by the BSSC
committee to form a second set of maps for use in design.  This second set
of "design" maps were subsequently balloted on and approved by the
membership of BSSC.

It is important to know that the maps and data on the USGS Web site are NOT
the same as the Provisions maps since the latter maps were modified by
committee.

SOURCES OF INFORMATION

Discussion has to be abbreviated so I will refer to places where additional
information can be obtained.  The USGS Geologic Hazards Team in Golden,
Colorado has a Web site (http://geohazards.cr.usgs.gov) on this subject as
well as other topics.  The Web site has documentation on how the first set
of maps, the probabilistic ones, were developed, page size maps, and data
used to prepare the maps.  The user can do searches on the data by zip code
or by latitude-longitude.  There is also a FAQ page which provides answers
to frequently asked questions.  There is an experimental interactive
mapping area and there is the capability to make your own map by entering
bounding latitudes and longitudes.

Documentation of the second set of maps (the Provisions maps) is included
in appendices A and B of FEMA 303, the commentary to the Provisions.
Currently a special edition of Spectra is in preparation which will add
more information.

The map package accompanying the Provisions is considerable.  It includes
the MCE maps for the U.S. as well as regional maps for areas where
additional detail is needed beyond that visible on the national maps.  This
map package also includes a set of maps for 10% probability of exceedance
in 50 years.  The 10% in 50 year maps, which are the same as USGS maps, are
not used by either the Provisions or the IBC, but they are used by FEMA 273
as are the MCE maps.  One map package was prepared so the same package
could be distributed with both reports.

Two additional sources of information that will be completed over the next
few months are on CD-ROMs.  One CD is a source for USGS maps and one is for
the Provisions maps.  The USGS CD contains both zip code and lat-long
lookup for spectra at different probability levels and hazard curves plus
maps.  The Provisions CD, which I am preparing,  will be released by
USGS-BSSC-FEMA and will contain both zip code and lat-long lookup for the
design parameters used in the Provisions (and proposed for IBC) and will
include the Provisions maps and the IBC maps.  These are in pdf format and
include some zoom capability.  The software required for reading pdf files
(Adobe Acrobat reader) is included for those that do not have it.  Both CDs
have been prepared using Visual Basic so that they are interactive and use
the point and click approach that many users are familiar with.  Files
developed during work with the program can be saved in an ascii format and
can be imported into many programs such as Excel.

Review of the USGS CD has been completed and review of the Provisions CD is
in progress.

USGS MAPS

The USGS maps were prepared following a series of six regional workshops.
Participants included earth scientists, geologists, geotechnical engineers,
and structural engineers.  The maps were prepared and revised in parallel
with the workshops. Review maps were posted on the WEB site for comment.
They were revised and published following a review period.   Some of the
differences from earlier USGS maps are briefly described below.  (Not all
of the maps between 1982 and 1996 are discussed.  There is a report on the
WEB site "USGS Spectral Response Maps and their Relationship with Seismic
Design Forces in Building Codes", by Leyendecker, et al which does this.)

1969 - The Algermissen map was a zone map based on intensities. 

1976 - The Algermissen and Perkins map was based on historical seismicity
for rates and sources.  Only peak ground acceleration was mapped and only
for one probability level.

1982 - The Algermissen and others' map of 1982 was based on historical
seismicity rates, with geology guiding the assumed seismic sources.  Map
parameters included peak ground acceleration and velocity for three
different probability levels.

1996 - The Frankel et al maps are probabilistic maps of ground motion.
They use historical seismicity to determine earthquake rates and sources
and fault slip rates for fault sources where available.  These data serve
as INPUT into a probabilistic model.  A limited "logic tree" analysis was
done incorporating different models and different attenuations.  Maps were
prepared by calculating hazard on a grid of about 150,000 sites in the 48
states.  Gridded calculations were also done for Alaska and Hawaii.

Mapped parameters include peak ground acceleration and spectral
accelerations for 0.2, 0.3, and 1.0 sec. These are available in large
format hard copy. Spectral accelerations were also calculated for 0.1, 0.5,
and 2.0 sec but these were not mapped.  However, all spectral values are
being included on the CD ROM.  Maps were prepared for each of the mapped
parameters for 2%, 5%, and 10% probability of exceedance in 50 years.

The trend as described above is to include additional information as it
becomes available.  This is as it should be.  Earth science has certainly
not remained static between 1969 and 1996.  The 1996 maps differ
significantly from an almost 30-year old 1969 map.  The capability to map
additional parameters has also improved dramatically in this time period.
I would expect this trend to continue although I do not see major changes
in the next release of the USGS maps.

SITE CONDITIONS

The USGS maps are for ground motion using a site condition with a shear
wave of 760 m/sec in the first 30 meters, corresponding to the boundary of
site conditions B and C as defined in the Provisions.  The ground motions
are NOT for bedrock.   For different site conditions site factors can be
used to adjust the mapped values.  Certainly for critical structures, site
studies are still important and should be done.

The Provisions maps equate the USGS maps with site condition B.  This adds
a small amount of conservatism to the maps but simplifies application of
the site factors.

MCE GROUND MOTION MAPS

The BSSC committee that arrived at the MCE maps modified the USGS maps by
considering engineerings needs.  This is discussed in the Commentary as
mentioned above.  Briefly, the committee selected one probability level, 2%
in 50 years, and two parameters, 0.2 sec and 1.0 sec, spectral acceleration
as the key parameters for design.  These maps were selected in part to
obtain greater protection against collapse for large rare events in the
central and eastern U. S. than would be provided by using 10% in 50 year
maps.  At the same, the map values were not considered drastically
different from 1997 UBC values in California when the MCE values were
multiplied by 2/3.  The implications of multiplying map values by 2/3 is
discussed in the following paragraph.

A key decision was made to multiply mapped values by 2/3 (this is done in
the design procedure).  The rationale for the 2/3, which is related to the
margin of safety against collapse, is discussed in Commentary.  The net
effect of this multiplication means that areas such as California are
designing for something close to a 10% in 50 year level, and the central
and eastern U.S. are designing for something around a 5% in 50 year level.
The reader should refer to appendix B of the Commentary for a discussion of
why this is so.

A second key decision was made to use the probabilistic contours up to a
plateau level corresponding to 1994 UBC Zone 4 design.  Faults located
inside this plateau use deterministic ground motion attenuating away from
the fault.  So the MCE maps are a combination of probabilistic and
deterministic values.  In concept this approach is similar to the use of
near-fault factors in the 1997 UBC.  The end result is that values may
exceed the zone 4 value as one moves closer to a fault.

USE OF SPECTRAL ACCELERATIONS

Buildings and other structures are designed to resist earthquake ground
motions, not earthquake magnitude.  The ground motion from a particular
earthquake depends on the magnitude, the distance of the site in question
from the fault generating the earthquake, site conditions, and other
factors.  In addition, there is a natural variability or randomness in
earthquake ground motions.  All these factors are taken into account in
calculating the probabilistic USGS maps for PGA and spectral accelerations.

UBC currently uses a zone factor, constant within a zone, that is related
to what is often referred to as effective peak acceleration.  The zone
factor is converted to a spectral response value by multiplying by a
factor, such as 2.5 for the short period value.  The spectral shape is
fixed by one ground motion parameter.  However, we know that the spectral
shape varies by region and takes at least two values of ground motion to
determine the spectral shape.  We also know that a single multiplier does
not apply for the entire U. S.  The short period value of 2.5, while not
bad for California, is not appropriate for large portions of the U. S.
Thus the new design procedures use two spectral accelerations (at 0.2 sec
and 1.0 sec) rather than peak ground accelerations or zone values to
determine the shape.  Furthermore, since we start with spectral values
there are NO MULTIPLIERS used.  Values are read directly from the map and
then used in the design procedure.   Interpolation is used between contours
to obtain intermediate values instead of using a constant value between
contours.  This avoids the former  problem of large changes in design
values for sites near each other but across zone boundaries.  Any
difficulty of obtaining design values for a site will be simplified when
the CD-ROMs are available.  Small location errors of a site won't lead to
large differences in design values.

Using the two spectral values, a design spectrum is approximated by using
the 0.2 sec spectral acceleration for the "flat" portion of the spectrum
and passing a curve varying as 1/T through the 1.0 sec spectral
acceleration.  But again, there are no multipliers (except for the 2/3
factor) unless soil factors are required to modify the site condition.  The
Provisions and Commentary provide additional details on calculating a
spectrum.

The USGS probabilistic data can be used to determine a uniform hazard
spectrum simply by plotting the spectral accelerations at each available
period (pga, 0.1, 0.2, 0.3, 0.5, 1.0, and 2.0 sec) from the USGS data set.
These are not code design values, and the resulting spectrum for a specific
site may be different from the spectrum determined using the MCE maps and
design procedure.

IBC MAPS

The IBC made an effort to provide the user fewer maps at a smaller size
than those included with the Provisions.  Otherwise the maps are the same.
The maps in review copies I have seen aren't the best quality, but we have
original digital files that we are sharing with IBC that should improve the
quality in the final printed document.  The maps that are in the proposals
for change are already improved over what was out for review last summer
although the county boundaries are faint.  These are reproduction matters
that are easy to remedy.

It is my understanding that there is strong consideration being given to
including the Design CD-ROM that was described earlier with the printed
IBC.  The CD can be used to obtain actual design values.  The maps will be
a good first approximation of the design values as well as a sanity check
against careless errors, such as entering the wrong lat-long in the CD.

USGS MAP CONTOURS

Finally, we frequently see comments about the values on the probabilistic
maps.  I have selected a range of our maps and listed the maximum contour
that we show (I won't bore you with values from each and every map).

Probability             PGA    0.2 sec spectral acceleration   1.0 sec
spectral acceleration 
10% in 50 years        1.0 g	    2.0 g                               1.0 g
2% in 50 years         2.0 g          4.0 g
2.0 g

If I search through the 150,000 point data grid, I can certainly find
instances of spot values near some faults that are larger than the contours
we have shown.  While we could have used different maximum contours, it was
our judgement that these were reasonable choices.

For comparision, unless I missed something while doing a quick review for
this note, the MCE maps have a maximum 0.2 sec spectral acceleration of
about 2.6 g and a maximum 1.0 sec spectral acceleration of about 1.6.
Remember that these are multiplied by 2/3 in the design procedure.

I have not attempted to address each and every comment that might have been
made, that was not my purpose, I have tried to provide factual information
about the USGS, PROVISIONS, and IBC maps.


E. V. Leyendecker
USGS National Seismic Hazard Mapping Project
leyendecker(--nospam--at)usgs.gov