Previous Work in Mongolia
In 1995, several geologists from the USGS participated in an international field trip led by Peter Molnar (MIT) and A. Bayasgalan (Mongolian Centre for Informatics and Remote Sensing) to view the surface rupture from the great 1957 Gobi-Altay earthquake. While most of us were aware of the 1957 event and the spectacular surface faulting associated with it from the classic post-earthquake study of Florensov and Solonenko (1965), and the more recent work of Baljinnyam et al (1993), we were far less knowledgeable about other aspects of Mongolian earthquake activity and faulting. The 1995 trip not only exposed us to the complexities of Mongolian earthquakes, but also to the physical beauty, history, traditions, culture, and isolation of a country struggling to find its political and economic way after emerging from the Soviet sphere. Our 1995 trip raised fundamental questions about earthquake and fault behavior. The complexity of the 1957 Gobi Altay rupture with stress release on both strike-slip and thrust fault segments raised questions about fault segmentation theory, which is an underpinning of current seismic hazard analysis. Another important question was whether the 20th century clustering of earthquakes was random, or whether this clustering cycle repeats itself. The only way to answer these questions is to obtain information on the dates of past earthquakes. This is done through paleoseismology, which the study of fossil earthquakes that are typically exposed in trenches excavated across faults and are dated using carbon-14 or other techniques.
In 1996, armed with these questions, we returned to the Gobi-Altay rupture with 12 geologists and a goal of trenching each of the major rupture components to expose the stratigraphic evidence of the pre-1957 (penultimate) earthquake and to date it with radiocarbon and thermoluminescence analysis (this is a technique that directly dates when grains of sand and silt were last deposited). Through examination of hand-excavated trenches (backhoes and other mechanized equipment are difficult to find in the Gobi desert), and by mapping of offset gullies and fault scarps, we exposed the penultimate event on five of the 1957 rupture components, including two on the main Bogd strike-slip fault. In the summer of 1997, a smaller group returned to the southern (Gurvan Bulag) thrust fault to undertake work not completed in 1996. (A summary of the 1997 study.)
To date, our work suggests that the pattern of surface rupture along the Bogd strike-slip fault and associated thrust faults prior to 1957 may have been different from what occurred in 1957. The penultimate event on the eastern part of the Bogd strike-slip fault occurred about 2400 years ago, whereas on the western segment only two large earthquakes occurred during the 12,000 years prior to 1957. The Dalan Turuu thrust appears to have slipped every 8000 years on average, the Gurvan Bulag thrust appears to produce earthquakes about every 4000 years, and the Toromhon thrust last moved several tens of thousands of years ago. (see Figure 2 for the locations of these fault segments)
