Complexity of Lithosphere and Earthquake Prediction

Monday, September 24, 2001 - 11:00am - 12:00pm
Keller 3-180
Vladimir Keilis-Borok (University of California, Los Angeles)
1. Earthquakes occur in the lithosphere - an upper shell of the solid Earth. Its thickness ranges from few km near oceanic ridges, to few hundred km in some continental regions. The lithosphere is set in motion by the large-scale currents in the underlying Earth's mantle and some internal processes like gravitational and chemical differentiation. In seismically active regions large part of this motion is realized through the earthquakes in a stick-slip fashion.

2. Two major factors cause complexity of the lithosphere: (i) Hierarchical structure, extending from about 10 tectonic plates to the about 10^25 grains of rocks. (ii) Instability, caused by a multitude of non-linear mechanisms, controlling the {strength - stress} field. In the time scale relevant to earthquake prediction, these factors turn the lithosphere into a hierarchical dissipative complex system. Strong earthquakes are regarded as the critical phenomena; an earthquake may be a critical phenomenon in certain volume of lithosphere, and a part of the background seismicity in a larger volume.

3. Development of the earthquake prediction algorithms brought together three methodologies: (i) phenomenological analysis of observations - I. Gelfand's type of pattern recognition and J. Tukey's kind of exploratory data analysis; (ii) universal lattice models of complex systems such as considered in statistical mechanics and non-linear dynamics; and (iii) Earth-specific models of tectonic faults' networks. In addition, (iv) theory of optimal control is used to link earthquake prediction with the earthquake preparedness.

4. Ongoing global test of the intermediate term prediction algorithms is discussed.