A Knowledge-Based System to Support Nuclear Test Ban Treaty Verification

All Papers

The major technical obstacle to the signing of nuclear test ban treaties is the issue of verifying compliance. Since the banning of atmospheric and oceanic testing (Limited Test Ban Treaty of 1963) pushed testing underground, seismic monitoring has been one of the most important technologies available for monitoring compliance with test ban treaties. The goal of these treaties is to progressively reduce the allowed yield of nuclear tests (Threshold Test Ban Treaty) and, possibly, eliminate testing completely (Comprehensive Test Ban Treaty). Increasing the sensitivity of the seismic monitoring requires processing rapidly increasing volumes of naturally occurring background seismicity (earthquakes, mine blasts, and so on) that must be discriminated from nuclear tests. Scientists in the nuclear monitoring community realized that the key to improving the implied search was to bring to bear the rich body of knowledge used by human seismic analysts. This chapter describes a knowledge-based system that was built to apply this approach to the problem. To address the problem of verification, the Defense Advanced Research Projects Agency (DARPA) has been engaged in an ongoing effort to extend U.S. capabilities to monitor the testing of nuclear weapons throughout the world by analyzing remotely sensed seismic signals. As part of this effort, in 1989, DARPA funded the Nuclear Monitoring Research and Development (NMRD) project for the development of a major system for collecting, organizing, analyzing, and archiving seismological data from many geographically dispersed seismic stations. Science Applications International Corporation (SAIC) is the prime contractor.