AN UPDATED APPROACH FOR ESTIMATING PALEO-EARTHQUAKE MAGNITUDES FROM MULTIPLE TRENCH OBSERVATIONS

One of the major contributions of paleoseismology to seismic hazard assessment are paleoearthquake magnitudes. Especially in regions with low or moderate seismicity, paleomagnitudes are normally much larger than those of historical earthquakes. Biasi & Weldon (2006) proposed a probabilistic magnitude estimate given one observed displacement, taking into account the natural variability of surface displacement along a fault. However, in case of more observation points (i.e. trench sites) per fault, single-event displacements at different locations can be correlated and result in a larger set of observed displacements for each earthquake. This can be used to narrow down the possible magnitude range associated with the generating earthquake. Extrapolating the approach of Biasi and Weldon (2006), the single-observation probability density functions (PDF) are assumed to be independent of each other, since observations of each trench do not depend on the findings of other trenches. Following this line, the common PDF for all observed displacements generated by one earthquake is the product of all single-displacement PDFs. In order to test this idea to practicability, we used the results of a paleoseismological investigation within the Vienna Pull-Apart Basin (Austria), where 3 trenches were opened along the normal Markgrafneusiedl Fault (MGNSF). Even if the Vienna Basin is characterized by low to medium seismicity (Mmax = 5.3/Imax = 8), there is no historical seismicity recorded along the MGNSF. However, our studies provide evidence for at least 5 major earthquakes with M > 6.0 along the MGNSF during the last ~ 100 ka. Using events that are observed within all 3 trench sites, we derived common magnitude PDFs for each set of observed displacements related to a single event. As expected, the final magnitude PDFs are narrower than those of the single-event magnitude PDFs. In addition, they are dominated by the largest observed displacement, especially with respect to the lowest magnitude that could have generated all observed displacements. In total, this approach seems to be a suitable method to combine observations from different locations to one magnitude value accounting for the natural variances of observed along-strike surface displacement. Reference: Biasi & Weldon (2006), BSSA 96, pp. 1612-1623.