“…In the past few decades, FWI has been successfully applied to active‐source ocean‐bottom and streamer datasets to characterize the oceanic crustal structures (e.g., Górszczyk et al., 2017, 2021; Gray et al., 2019; Kamei et al., 2012; Morgan et al., 2016; Operto et al., 2006; Ravaut et al., 2004), and the land datasets to characterize the continental structures (e.g., Bleibinhaus et al., 2007; Morgan et al., 2011). In particular, FWI has also been applied to deep reflection seismic profiles in the study of the upper‐crustal structures (Adamczyk et al., 2015; Li, Gao, et al., 2021; Smithyman & Clowes, 2013; Zhang et al., 2021). Still, the application of FWI to the field data faces considerable difficulties, as FWI is an inverse problem with strong nonlinearity (e.g., Virieux & Operto, 2009) and prone to suffer from ill‐posedness originating from the crosstalk between different parameters.…”