Dependence of Possible Characteristic Earthquakes on Spatial Sampling: Illustration for the Wasatch Seismic Zone, Utah

Abstract

An important issue for regional tectonics and earthquake hazard estimation is whether large earthquakes are “characteristic”, more frequent than would be inferred from the rates of smaller events. A challenge in resolving this question is that the rates of small earthquakes are typically determined from the seismologically recorded earthquake history, whereas the rates of large earthquakes are inferred from paleoseismic observations. As a consequence, different results from comparing the two can arise depending on the specific assumptions made and time and space sampling used. In general, earthquake recurrences approximately follow a log-linear, b -value, or Gutenberg-Richter relation, log N = a - bM , with b ∼1, such that the logarithm of the annual number ( N ) of earthquakes above a given magnitude ( M ) decreases linearly with magnitude (Ishimoto and Iida, 1939; Gutenberg and Richter, 1944). Studies of specific areas, however, which commonly address the short history of seismological observations by combining seismological data for smaller earthquakes with paleoseismic data or geologic inferences for larger earthquakes, sometimes infer that large “characteristic” earthquakes occur more frequently than expected from the log-linear frequency-magnitude relation observed for smaller earthquakes (Schwartz and Coppersmith, 1984). Whether characteristic earthquakes are real or apparent in any given region is an interesting question (Kagan, 1996; Wesnousky, 1996). A number of effects can give rise to apparent characteristic earthquakes or “uncharacteristic” earthquakes, ones that appear to occur less frequently than expected from the rates of smaller earthquakes (Stein and Newman, 2004). One bias can result from a short recorded earthquake history, in particular if its length is comparable to the mean recurrence time of large earthquakes predicted by a Gutenberg-Richter distribution. A second bias can result from errors in estimating the size or frequency of the largest earthquakes from the paleoseismic record (Stein and Newman, 2004; …