Comparison of Regional 3-D Subduction Models in the Western Pacific to Regional Models from Slab 1.0
To create accurate models of subduction zone geometries we use Slab1.0, a three-dimensional model of global subduction zone geometries created using primarily teleseismic data. We assess the accuracy of Slab1.0 3-D models by comparing a selection of these models to models derived from previous regional studies. By digitizing models of subduction zones collected from readily available geoscience literature, we quantitatively compare these models to Slab1.0, identifying discrepancies in slab interface depth. Plausible reasons for such discrepancies include differences in data coverage and bias in the location of the earthquake epicenters in regional and teleseismic catalogs. Here, we show comparisons between subduction zone models in Japan, Sumatra and the Tonga trench, all well defined in Slab1.0. In Japan, where we have collected several regional models, we find that the Slab1.0 models are consistently shallower than their regional counter-parts in the shallow sections of the subduction zone. This may indicate that discrepancies are due to well-known bias in teleseismic earthquake locations in subduction zones, or that perhaps that the regional data are biased because of poor azimuthal coverage in recordings of offshore events.