Abstract In 1984, the National Earthquake Prediction Evaluation Council (NEPEC) recognized that the 40 km long Parkfield section of the San Andreas fault is the best location to trap a significant earthquake and recommended that the U.S. Geological Survey (USGS) establish a focused earthquake prediction experiment at Parkfield. The Parkfield earthquake prediction experiment established in 1985 was designed to determine if short-term earthquake prediction was feasible and to attempt to issue a public warning before the next Parkfield mainshock. Networks of instruments to record all credible earthquake precursors were installed, and Parkfield was soon the best-instrumented earthquake region in the world. A prediction plan was reviewed and approved by NEPEC and by the California Earthquake Prediction Evaluation Council (CEPEC). The Director of the USGS and the Governor of California delegated their authority to issue a prediction if the plan was followed. The Parkfield earthquake prediction remains the only earthquake prediction ever officially recognized by the U.S. government or by the State of California. Data recorded before, during, and after the 28 September 2004 Parkfield earthquake have resulted in fundamental advances in our understanding of earthquake physics (e.g., Harris and Arrowsmith, 2006). Only subtle precursory changes in strain were recorded before the 2004 earthquake, too small to provide a basis for reliable public warnings. Reliable short-term earthquake prediction remains unattainable. Ground motion at 8 sites within 1 km of the 2004 rupture and at 40 sites within 1 and 10 km were recorded, nearly doubling the global data set of strong-motion records at these distances. Large spatial variations in strong-motion amplitudes were observed. Parkfield remains the best place to trap a significant earthquake, and the next Parkfield earthquake will provide another opportunity to advance our understanding of earthquake physics.
William H. Bakun (Mon,) studied this question.