Fig. 2

Comparison of slab top and slab Moho temperatures for a NE Japan, b Alaska, and c Cascadia as predicted by three different approaches: (1) Updated D80 Sepran models from Syracuse et al. (2010) as discussed in text. (2) Sepran models following D80new description as discussed in Sect. 2.2. (3) Fully independent models from Wada and Wang (2009). The models agree moderately well—main differences are due to the shallower decoupling depth \(d_c\) used in Wada and Wang (2009) and the difference in mantle temperature between the original D80 models and the new model set presented here. In addition the use of a younger age of the incoming lithosphere in Wada and Wang (2009) for Cascadia (8 Myr vs. 10 Myr) leads to a pronounced warming of the slab thermal structure—even minor differences in slab age have a strong influence in young subduction zones due the the change in thermal gradients in the shallow lithosphere. The Cascadia model from Wada and Wang (2009) is also warmer particularly at shallow depth because of the different treatment of modeling the effects of the thick sediment section which leads to a warmer initial thermal structure of the oceanic crust compared to the D80 model. Combined these different model assumptions cause a relatively significant difference in the predicted forearc thermal structure for Cascadia