RUPTURE LENGTH AND EARTHQUAKE MAGNITUDE ESTIMATES FROM PALEOSEISMIC MEASUREMENTS OF DISPLACEMENT AT A POINT

Paleoseismic investigations are a primary means for discovering pre-instrumental earthquake behavior of a fault. Individual paleoseismic excavations necessarily recover only data at a point on the fault. However, fault length and earthquake magnitude are necessary in order to apply paleoseismic findings to earthquake hazard estimation. Thus a challenge for paleoseismology is to extend results from a point to some estimate of rupture length and magnitude. A leading complication is that observed surface slips tend to vary greatly within a single rupture, so one cannot say whether measured offsets are representative of the paleorupture, or whether they are larger or smaller than the average. In addition detailed studies suggest that sites such as minor transtensive step-overs best suited for the preservation of dating information can hide or distribute slip, leading to underestimates of slip and earthquake size.

We have developed a means of estimating the magnitude and rupture length of a paleoearthquake in a probabilistic way given a point estimate of slip. Hemphill-Haley and Weldon (BSSA, 1999) show that by scaling slip distributions by rupture length and average displacement, well-mapped ruptures have some common features. Their normalized variability plots can be inverted for a probability distribution of possible earthquake magnitudes given point displacement after they are scaled using magnitude-length and magnitude-surface displacement regressions from the literature. Beside paleomagnitude estimates, the inverse probabilities provide quantitative help in assessing whether events at adjoining paleoseismic sites are likely to correlate, and thus compliment radiocarbon evidence that might not be compelling on its own. We are presently applying these relationship to assess event correlations among paleoseismic sites on the southern San Andreas fault.