What mistakes are we making while interpreting salt? Could FWI help?

Model studies indicate that our conventional salt-interpretation workflow, consisting of cascaded sequences of flooding with either salt or sediment velocities, migration, and picking, produces two distinct types of velocity errors: 1) small but ubiquitous positioning errors of the margins of the salt, and 2) large “chunky” errors where salt boundary reflections were grossly misinterpreted. Full-Waveform Inversion might be a solution, but to achieve success may require new kinds of data, improved algorithms, or most likely both.