Part I. Slip behavior of the San Andreas Fault through several earthquake cycles Part II. A structural interpretation of the aftershock 'Cloud' of the 1992 Mw 7.3 Landers earthquake

摘要

Part 1 of this thesis addresses the question of how variable fault slip per event is through time. This question is important, because progress in understanding and forecasting large earthquakes depends critically on precise reconstruction of the variation of rupture magnitude with time. Well-documented examples of slip measurements of successive past earthquakes rarely span more than three earthquake cycles. The dearth of accurate measurements of serial slip is partly due to the obliteration of piercing lines that we can use to separate the offsets associated with individual earthquakes.ududThe special configuration of a series of channels offset across the San Andreas fault, near Wallace Creek has enabled me to determine the styles and the magnitudes of slip of the most recent 6 events at this locality. At the site, a feeder channel cuts a Pleistocene alluvial fan on the upstream side of the fault. On the downstream side, several small channels were offset dextrally from the source and sequentially abandoned. We opened a latticework of trenches across the offset channels on both sides of the fault. The trenches across the mouth of the upstream feeder channel exposed a set of nested channels; downstream trenches revealed several singular channels. The elevations, shapes, stratigraphy and ages of channels provide reliable information for correlating channels across the fault. 3-D excavations have allowed me to locate accurately the offset channel pairs and to determine the amounts of motion with small uncertainties. The dextral slips associated with the latest 6 events are, from the youngest to oldest, ~ 8 m, ~ 7.5 m, ~5.5 m, ~ 1.5 m, ~ 8.0 m, and ~ 5.5 m. The high occurrence of events with slip of about 7.5 m suggests that the magnitude of slip at a point along the fault does not result from a random process. But the slip is not as regular as predicted by characteristic models either. Thus, data at the site do not support perfectly characteristic behavior, but do show a significant degree of regularity.ududPart II of the thesis demonstrates that Landers aftershocks constituted primarily a several-km-wide damage zone centered on the mainshock rupture plane. Most aftershocks probably did not occur on the same surfaces that had moved during the mainshock. Rather, the aftershock populations revealed the nature of the fractured medium around the principal faults, and the general structure of a fault zone.