Figure 1 Map of tectonic environment of the Sichuan-Yunnan and surrounding regions with seismic stations, events and faults superimposed. Red circles and green triangles show respectively event and station locations selected for the P-wave velocity model inversion in this study. The black dot and triangles show the locations of the event and stations whose records are displayed in Figure 2. Major active fault zones are shown by thick black lines with abbreviated names in black, including LMSFZ (Longmenshan Fault Zone), XSHFZ (Xianshuihe Fault Zone), ANH-ZMHFZ (Anninghe-Zemuhe Fault Zone), XJFZ (Xiaojiang Fault Zone), LJ-XJHFZ (Lijiang-Xiaojinhe Fault Zone), JSJFZ (Jinshajiang Fault Zone), RRFZ (Red River Fault Zone), LZJFZ (Luzhijiang Fault Zone), and LCJFZ (Lancangjiang Fault Zone). Major tectonic blocks are indicated by abbreviated names in blue, including SYRB (Sichuan-Yunnan Rhombic Block), SB (Sichuan Basin), SGB (Songpan-Ganzi Block), QB (Qiangtang Block), and YB (Yangtze Block). The yellow dashed lines with arrows show the general directions of the velocity field of crustal deformation (Wang and Shen, 2020). The plot in the lower-right corner shows the P-wave ray paths (43,540) coverage used in this study.
The strong crustal deformation and relative motions between various blocks make the Sichuan-Yunnan region one of the most tectonically active zones with a high level of earthquake activities. Frequently occurring small and moderate earthquakes, coupled with the deployment of increasing number of broadband seismic stations in recent years provide valuable data for us to improve the quality of regional structural images. In this study, we develop a high-resolution P-wave velocity model for the lower crust and uppermost mantle for the Sichuan-Yunnan region by applying the fast marching Method (FMM) of Rawlinson and Urvoy (2006) to a large dataset of P waves obtained by an automatic picking algorithm based on deep learning (Wang et al., 2019). The true resolution of the model has been estimated by the statistical resolution matrix approach (An, 2012). Our results provide further support for the existence of the lower crust flow and reveal more details on the complex structure of the uppermost mantle of the Sichuan-Yunnan region.