ˡ͡ʪõز
143(2020ǯٽ)ؽѹֱ



ֱ׻(ʸ)

ֱ׻(ʸ)
Japan Trench is a dynamic environment with characteristic tectonic activity due to the converging plate margins at the subduction zone. The basal erosion of the upper plate causes normal faults in the forearc slope, which leads to the subsidence of the structures above them. Seismic velocity model building and depth imaging is a challenging task in the forearc slope, where severely faulted and deformed sedimentary strata overly Cretaceous continental framework with an erosional unconformity in between. In order to develop a reliable velocity model which could explain the complex subsurface geology in this area, we applied acoustic full waveform inversion (FWI) on 2D seismic data acquired using a towed streamer. FWI is a robust method which iteratively updates a smooth initial model by fitting observed and calculated data. We carefully designed a workflow for data preconditioning, initial model building, and waveform inversion, which could achieve improved depth images. A layer stripping method was used to build an accurate initial velocity model to avoid cycle skipping issues. A matching filter was implemented in the frequency domain for correct data fitting before misfit evaluation. Reverse time migration (RTM) and Kirchhoff prestack depth migration (KPSDM) images were produced to confirm the accuracy of the resulting velocity model. We noticed that some of the faults in the study area offset the seafloor with evident sub-seafloor sediment subsidence, implying that the faults are likely active.