Reappraising elastic thickness variation at oceanic trenches

We reassess the variation of elastic thickness as a function of lithospheric plate age using a global database of bathymetric and free- air gravity profiles which are perpendicular to oceanic trenches. As in many previous studies, our starting point is the well- known floating elastic plate model. In order to remove the influence of short- wavelength features not associated with lithospheric bending, adjacent profiles from 10- Myr bins have been stacked together to construct average profiles with standard deviations. Each average profile was then inverted in a two- stage procedure. First, singular value decomposition was used to determine two unknown flexural parameters, together with a regional slope and offset, for any given elastic thickness. This procedure was repeated for a range of elastic thicknesses. Second, residual misfit was plotted as a function of elastic thickness, and the global minimum was identified. This two- stage procedure makes no prior assumptions about magnitude of the load, size of the bending moment, or whether the elastic plate is broken/ continuous. We obtained excellent fits between theory and observation for both bathymetric and gravity profiles from lithosphere with an age range of 0 - 150 Ma. The shape of the residual misfit function indicates the degree of confidence we have in our elastic thickness estimates. The lower limit of elastic thickness is usually well determined but upper limits are often poorly constrained. Inverse modeling was carried out using a range of profile lengths ( 250 - 300, 500, and 700 km). In general, our estimates show no consistent increase of elastic thickness as a function of plate age. This surprising result is consistent with recent reassessments of elastic thickness beneath seamounts and implies either that elastic thickness is independent of plate age or that elastic thickness cannot be measured with sufficient accuracy to reveal such a relationship. Modeling of short free- air gravity profiles ( 250 - 300 km) does tentatively suggest that elastic thickness increases linearly from 5 to 10 km between 0 and 20 Ma and from 10 to 15 km between 20 and 150 Ma. This variation roughly matches the depth to the 200 degrees C isotherm which corresponds to an homologous temperature of 0.4 for wet peridotite. Unfortunately, for longer profile lengths, there is no temporal dependence, and elastic thicknesses vary considerably for all plate ages. Bathymetric profile modeling yields similar results although uncertainties are larger.