Cooling history of the Paci c lithosphere

Data coverage, expressed as path density of Rayleigh wave group velocities, is shown in Figure 1. Coverage is best at intermediate periods, and reduces particularly at very long periods. It is also considerably lower for Love waves than for Rayleigh waves, especially at periods above about 100 sec. Resolution estimates are shown in Figure 2. Average resolution across the Paci c is about 600 km at 20 sec, 720 km at 50 sec, 850 km at 100 sec, and 980 km at 150 sec period. The joint inversion of phase and group velocities improves the vertical resolution of the estimated shear velocity model because the sensitivity kernels of short period group velocities compress into the upper lithosphere while the long period phase velocities provide better deep sensitivity than the group velocities. Figure 3 shows an example of an inversion at a point in the Central Paci c, demonstrating this e ect. The ensemble of acceptable models that t both the phase and group velocities is substantially narrower than the ensembles that t either wave type alone. Principal trends in seismic velocities and temperatures are revealed in both the seismic and thermal parameterizations. In the main text, we show results using the temperature parameterization based on a thermal model only. Figures 4 and 5 here display aspects of the 3-D seismic and temperature models estimated with the seismic parameterization. These gures are to be contrasted with Figures 3 and 4 in the main text. Although there are subtle di erences between the seismic and temperature models that are estimated with these two di erent parameterizations, the general features and age trends are very similar, particularly the trend of apparent thermal age shown in Figure 4 in the main text. We estimate apparent thermal age from the seismic parameterization by converting Vs to temperature and tting the thermal model shown in Figure 2b in the main text to the temperature pro le. This results in a range of apparent thermal age ( ) and potential temperature (Tp) estimates that de ne the ensemble of acceptable temperature models. To construct Figure 5 we use the middle of the ensemble. Reheating inferred in the Central Paci c is absent in the lithosphere beneath other oceans. Figure 6 shows the average temperature structure beneath the other oceans world-wide where