The QuakeFinder Earthquake Forecasting Technology – A Signal Processing Problem
Speaker: Tom Bleier, QuakeFinder
Presentation
Abstract
Earthquakes are the last major natural disaster lacking a near-term forecasting system. Large, complex statistical models are constantly being refined that attempt to predict the probability of major seismic events decades in advance. And the technology exists to provide seconds of warning using an array of seismic sensors near fault lines monitored real-time. But no system has been successfully demonstrated that provides weeks or days of advanced notice—enabling more precise and actionable forecast. Predicting an earthquake’s approximate location and magnitude weeks in advance could save thousands of lives and billions of dollars in damage. QuakeFinder, started in 2000, is a research and technology development effort at Stellar Solutions, Inc—a Palo Alto-based company that provides engineering services to the aerospace industry. For decades, it has been acknowledged that strange electrical phenomena are associated with Earthquakes—animal behavior changes, earthquake "lights," etc. Based on research findings by Stanford professor Tony Frazier-Smith prior to the 1989 Loma Prieta earthquake, QuakeFinder has attempted to capture and characterize changes in the Earth’s magnetic signals near the epicenter of earthquakes—weeks ahead of the seismic event. The team has deployed 125 sensors suites along the San Andreas and adjacent fault lines in California and have collected data at 50 Hz sample rate for over a decade. QuakeFinder also has 40 sensors in 5 other earthquake prone countries (Chile, Peru, Indonesia, Greece, and Taiwan). The intent is to capture the electrical activity of earthquakes, characterize the signals, and enable an earthquake warning system. In parallel, researchers are attempting to understand the underlying science behind the electrical pulses. Dr. Friedemann Freund of NASA Ames Research Center and the SETI Institute has developed a theory and is undertaking lab and field experiments to demonstrate the source of the electromagnetic signals. The QuakeFinder team has over 65 TB of data and has captured numerous earthquakes. However, the magnetometer sensors are extremely sensitive and also capture other natural and man-made noise. The problem at hand is to develop a signal processing algorithm capable of recognizing the earthquake pulse ahead of the event—and save lives.
The QuakeFinder team will be joined by Dr. Freund and provide an overview of the project, the data, and the problem statement.
Biography
For thirty-seven years Tom Bleier has developed, built, and tested complex defense and commercial satellites and ground control systems. Early in his career Bleier became interested in understanding the physics of the pre-earthquake process. Based on the theories developed by Stanford University scientists after the Loma Prieta earthquake and the USGS observations of earthquake lights, he developed ultra-low frequency (ULF) magnetometers to detect electromagnetic signals at the onset of large earthquakes that may provide the key to understanding the warning signs of deadly seismic hazards. What started as an educational outreach program, donating time and materials for the ULF magnetometer kits to a handful of high-schools, has expanded over the last decade into a web-based network of over 165 sensors that detect and record ULF background signals—looking for earthquakes. Bleier serves as the "epicenter" of a growing community of researchers worldwide studying Electromagnetic phenomenon associated with earthquakes.