High detail fault segmentation: deep insight into the anatomy of the 1983 Borah Peak earthquake rupture zone (Mw 6.9, Idaho, USA)

Following the fieldwork along the Lost River Valley (Idaho, USA), we integrate different subsets of both bibliographic and original data to obtain an extremely detailed segmentation of the fault portion that released the 1983 Borah Peak earthquake (Mw 6.9). The earthquake ruptured the topographic surface with a normaloblique faulting mechanism, activating two SW-dipping segments (i.e., Thousand Springs and Warm Springs) and a branching SSW-dipping fault (i.e., Arentson Gulch Fault), and producing coseismic surface ruptures up to 3 m high. We augment the 1983 earthquake knowledge by investigating and interpreting high-resolution topography and scarps mapping obtained through structure-from-motion. A large dataset of quality selected vertical separation (VS) data, combined with rupture zone width measurements, new fault/slip data, and an analysis of major and minor structural-geometric complexities, highlights a partition of the deformation and a segmentation of the fault up to the detail of four orders (i.e., segments, sections, subsections, sectors), providing new useful details of the Borah Peak earthquake and new constraints for paleoseismic and seismotectonic studies. The fault/slip data show variations along the fault strike that we interpret from the perspective of a kinematic partitioning, which supports segmentation. In 1983, the two main activated segments had completely different rupture behaviors, with important Rupture Zone Widths in the southern portions and with the deformation concentrated along the main fault trace in the northern portions. We show that the distributed ruptures, in addition to being a large percentage of all deformation in terms of a total length of the ruptures (~19.5 km vs 31 km in total for the main ruptures) also accommodate most of the surface deformation (~66%). We also show that 83% of the deformation in terms of length of the surface faulting is located at the hanging wall of the main rupture, while at the footwall it is located the 17%, which becomes 80% and 20% respectively if in terms of VS. We observe a significant correlation between VS, rupture zone width, the position of the rupture (footwall or hanging wall), and fault geometry. From these correlations, we highlight, for example, the control of obliquity and kinematic partitioning in the surface expression of the earthquake propagation. We interpret the coseismic (i.e., 1983) and long-term (i.e., Quaternary) behavior, showing that the two segments activated had similar cumulated behaviors in distributing the deformation among synthetic and antithetic ruptures over time, even if with different geometries. We calculate ratios applying probability density functions, that are consistent with most of the literature but not spatially uniform. These results must be considered in studies on active faulting hazard and suggest caution in establishing rules for land use planning close to active faults capable of rupturing the surface.