TURKEY FLAT, USA SITE EFFECTS
This is the official website for the Turkey Flat “blind” strong-motion prediction test, based on recordings obtained by the CSMIP instruments in the Turkey Flat array during the September 28, 2004 M6.0 Parkfield earthquake. The purpose of the test is to validate current methods of estimating the effects of soil/ground conditions on earthquake shaking that are used in designing earthquake resistant structures. Nearly all transactions with participants in the test will be conducted through this website, including online registration and distribution of prediction criteria, input records, format standards, etc. References and reports regarding the Turkey Flat test site will also be made available through this site.
Overview of Turkey Flat Project
Systematically compare and evaluate the reliability of
contemporary methods used to estimate the effect of surface
geology on, and to test the linearity of, shallow stiff-soil
Approach: Collect high quality ground motion data from local earthquakes, quantify the geotechnical properties of the site geology, and distribute the information to experts in California and around the world. Conduct “blind” prediction tests using predicted calculations made by ground motion experts for sensor sites where the response is known, but not made available until all predictions have been received. Results of each prediction will be compared with one another and with actual observed ground motion. As with a similar test using weak-motion recordings in 1990’s, submitted calculations will be kept anonymous.
Products: A database of weak and strong-motion records from the array, and predictions based on a variety of ground motion models; a database of geotechnical properties of the test area; a series of reports describing each principal phase of the project (see references); an analysis of all prediction results, including statistical comparisons of the predictions and the recorded data; and technical papers from participants.
Background: At the 1985 General Assembly of the International Association of Seismology and Physics of the Earth’s Interior (IASPEI), held jointly with the International Association of Earthquake Engineering (IAEE) in Tokyo, Japan, a resolution was passed forming a Joint Working Group on the Effects of Surface Geology on Seismic Motion (IASPEI/IAEE JWG-ESG). The purpose of the group is to coordinate the establishment of an international series of test areas designed to provide a database for comparing and testing contemporary methods, and develop new methods, to predict the effects of local geology on earthquake ground motion. Although methods for assessing site effects are being used in the design of critical facilities around the world, the reliability of these methods has not been rigorously tested. It is the goal of this program to fulfill this need. An international program provides a forum for experts around the world to exchange ideas, and significantly increases the prospects for acquiring the necessary data much sooner than would otherwise be possible. Recognizing the importance of the effort to earthquake safety in California, in 1986 the California Geological Survey established a test site at Turkey Flat, California. At the XIX General Assembly of the International Union of Geodesy and Geophysics in Vancouver, British Columbia, 1987, a resolution was passed incorporating the experiment at Turkey Flat into the international program.
Summary of Strong-Motion Test
Eighteen years ago, in anticipation of the Parkfield earthquake,
the California Geological Survey (CGS) established a test area
in a sedimentary valley at Turkey Flat, near Parkfield,
California, where the California Strong Motion Instrumentation
Program (CSMIP) installed a strong motion array. CGS partnered
with the IASPEI/IAEE Joint Working Group on Effects of Surface
Geology on Seismic Motion as well as members of the geotechnical
community to thoroughly characterize the geophysical properties
of the site. The strong motion array consists of surface and
downhole accelerometers, with surface instruments at the two
valley edges, at one quarter of the valley width, and at the
center of the small, shallow stiff-soil (25 m) sedimentary
valley. The instruments at the valley center also include a
downhole array, with instruments just beneath the rock interface
and at mid-height in the sediments.
The September 28, 2004 M6.0 Parkfield earthquake was well recorded throughout the test site array, and this provides the ground motion records necessary to conduct the long awaited blind prediction test. In this prediction experiment, acceleration time histories recorded on bedrock near one valley edge will be provided along with a “standard” model of the geotechnical properties at all recording sites to interested participants. Participants will be asked to make predictions of the ground motions at the five other recording locations (3 surface and 2 downhole), where, as part of a long-term plan, recordings are being withheld by CSMIP until the predictions have been received and officially logged. Details of the test area and the test procedure will be distributed in an open invitation for all interested parties to participate in the blind test. A workshop at which predictions can be presented and comparisons made with the recorded motions is planned for Fall 2005.