“…As mentioned above, examining the deeper part of those models is more difficult because correcting for the effect of attenuation is model-dependent. Richards et al (2020), for example, used the anelastic parameterization of Yamauchi and Takei (2016) to compare half-space cooling and plate models against SL2013, SEMum2, and one more tomographic model (CAM2016; Ho et al, 2016); their results seem to indicate that SEMum2 and CAM2016 are consistent with half-space cooling and plate models, respectively, and SL2013 is somewhere in-between. Richards et al (2020), however, calibrated the anelastic parameters for each of these different tomographic models, and some of the calibrated parameters vary substantially among models (see their Table F3).…”