Preliminary Assessment of Seismic CTBT/NPT Monitoring Capability

• Main Authors: Fisk, Mark D., Gray, Henry L., McCartor, Gary D.
• Other Authors: MISSION RESEARCH CORP SANTA BARBARA CA
• Format: Text
• Language: English
• Published: 1994
• Subjects: Seismic Detection and Detectors; Mining Engineering; Nuclear Weapons; *UNDERGROUND EXPLOSIONS; *MINING ENGINEERING; *NUCLEAR EXPLOSION DETECTION; *SEISMIC DISCRIMINATION; MONITORING; THRESHOLD EFFECTS; FALSE ALARMS; GEOLOGY; SEISMIC DATA; FREQUENCY BANDS; SPECTRA; TRANSPORTABLE; STANDARDIZATION; CALIBRATION; EARTHQUAKES; RANGE(DISTANCE); CORRECTIONS; RANK ORDER STATISTICS; MINES(ORDNANCE); TREMORS; CTBT(COMPREHENSIVE TEST BAND TREATY); PE62301E; WUPLNM93GMAL; Novaya Zemlya
• Online Access: http://www.dtic.mil/docs/citations/ADA293188; http://oai.dtic.mil/oai/oai?&verb=getRecord&metadataPrefix=html&identifier=ADA293188

Our recent work focuses on assessing seismic event identification performance using a method we developed for outlier detection. This is motivated by the lack of seismic calibration data for underground nuclear explosions in most regions and the difficulty associated with transporting regional discriminants. The procedure may be fully automated to flag events warranting special attention and to test all appropriate assumptions to ensure validity. The method also allows control of false alarm rates and a natural way to rank events. The procedure is applied to nuclear explosions, earthquakes, mining blasts and mine tremors in diverse geological regions, recorded by the ARCESS and GERESS arrays, CDSN station WMQ, and LNN stations KNB and MNV. Discriminants used include Pn/Lg and Pn/Sn in several frequency bands from 3 to 8 Hz, as well as Lg spectral ratios. Identification and false alarm ratios were estimated for each region using a standard set of discriminants. At 0.01 significance level, between 92-100% of the explosions and quarry blasts were detected as outliers of the earthquake groups in the various regions, except at KNB where the outlier detection rate was 80%. There were 3 false alarms out of 158 earthquakes, slightly higher than the target rate of 1%. Effects of contaminated training data were also studied by intentionally including quarry blasts or rock bursts in earthquake training sets to determine if the outlier test can detect them and to assess potential impacts on monitoring performance. Pn/Lg transportability was also examined to assess how precisely a discrimination threshold must be transported in order to be effective. Last, identification analysis of the 31 December 1992 Novaya Zemlya event was repeated after first applying distance corrections. Prepared in cooperation with Souther Methodist Univ., Dallas, TX.