Spatial variation of slip behavior beneath the Alaska Peninsula along Alaska-Aleutian Subduction Zone

The Alaska Peninsula, including the Shumagin and Semidi segments in the Alaska-Aleutian subduction zone, is one of the best places in the world to study along-strike variations in the seismogenic zone. Understanding the cause of along-strike variations on the plate interface and seismic potential is significant for better understanding of the dynamic mechanical properties of faults and the rheology of the lower crust and lithospheric mantle in subduction zones. GPS measurements can be used to study these properties and estimate the slip deficit distribution on the plate interface. We re-surveyed pre-existing (1992-2001) campaign GPS sites in 2016 and estimated a new dense and highly precise GPS velocity field for the Alaska Peninsula. We find evidence for only minimal time variations in the slip distribution in the region. We used the TDEFNODE software package to invert for the slip deficit distribution from the new velocities. There are long-wavelength systematic misfits to the vertical velocities from the optimal model that fits the horizontal velocities well, which cannot be explained by altering the slip distribution on the subduction plate interface. Possible explanations for the systematic misfit are still under investigation since the plate geometry, GIA effect and reference frame errors do not explain the misfits. In this study, we use only the horizontal velocities. We divided the overall Alaska Peninsula area into three sub-areas, which have strong differences in the pattern of the observed deformation, and explored optimal models for each sub-area. The width of the locked region decreases step-wise from NE to SW along strike. Then we compared each of these models to all of the data to identify the locations of the along-strike boundaries that mark the transition from strongly to weakly coupled segments of the margin. We identified three sharp boundaries separating segments with different fault slip deficit rate distributions. Significant change in fault coupling from strong to weak are spatially correlated with the change in pre-existing plate fabric caused by cessation of the Kula-Pacific spreading and reorientation of the northern section of Farallon-Pacific spreading, which also correlate with changes in the degree of outer rise normal faulting and hydration of the downgoing plate.