Geologic Deposit Strength Inversion from Landslide Inventories

Mapped landslide databases and estimates of the seasonal groundwater conditions provide an opportunity to back-calculate statistical distributions of macroscopic geologic deposit strengths. Based on the general assumption that landslides are most often active, with their strength fully mobilized, within existing landslide deposits during wet conditions, we develop a maximum a posteriori inversion method to back-calculate strengths for various geologic deposits. In this method, multiple landslide modes are evaluated for dry and wet conditions, and the factor of safety is computed for a wide range of deposit strengths and landslide geometries. The optimal strength parameters (1) plausibly represent the available laboratory and field index test data while also producing a distribution of landslides that (2) best matches the landslide inventory maps during wet conditions and (3) minimizes overprediction of landslides during dry conditions. The resulting strength estimates are more appropriate to evaluate a broad set of landslide failure modes than estimates from previous studies, which often rely on sparse laboratory strength data or inversion from a single mode of failure. We demonstrate the applicability of the back-calculated strengths in a test area of the northern Santa Cruz Mountains of California, USA, for M6.9 Loma Prieta and M7.9 San Francisco earthquake scenarios.