4.2 Synthetic Modeling Results
Since we were only modeling the features at slab depth and only considering the upgoing Ps conversion, we calculated correlation coefficients of the predicted and the observed waveforms from 2 s after P arrival to 10 seconds after P arrival. Model 1 (Fig. 3a) produced the worst fitting synthetics of all three models (average correlation coefficient of -0.06). Model 2 (Fig. 3b) which consisted of a featureless upper crust and an 8 km-thick oceanic crust is better fit compared to the first model (average correlation coefficient of 0.45). Model 3 (Fig. 3c), the simple two-layer model with an increase in velocity at the slab Moho depth is the best fitting model with an average correlation coefficient of 0.53. The results suggest that the Vp, Vs and density above the slab Moho must be similar to obtain optimal fit to the observed data. In other words, introducing additional features in the model above the Moho, even an oceanic crust, creates synthetics that poorly match the observational data.