HR: 08:15h
AN: T21C-02    [Abstracts]
TI: Toward a Next Generation of Isostatic Analysis
AU: * Lowry, A R
EM: arlowry@cc.usu.edu
AF: Department of Geology, Utah State University, 4505 Old Main Hill, Logan, UT 84322-4505, United States
AB: Much of the effort to improve upon coherence analysis of isostatic response since Forsyth's&p [1985] introduction has focused on optimizing resolution/bias using more advanced power spectral estimators or wavelets. Less attention has been paid to Forsyth's innovative recognitions that (1) internal loads significantly impact T_e estimates, (2) any arbitrary choice of T_e exactly models observed gravity and topography given the right combination of surface and internal loading, and (3) an incorrect choice of T_e or other parameter to deconvolve surface from internal loads artificially correlates load estimates, making the parameterization that minimizes correlation the likely optimal choice. However, a drawback of parameterizing isostasy by minimizing load correlation is that accuracy of the result hinges on the assumption that surface and internal load processes truly are uncorrelated. Simulations demonstrate that error in recovery of T_e arises because, even when synthetic load fields are truly uncorrelated, sub-sampling the data in windows, tapers or wavelets incurs random, spurious correlation at certain locations and wavelengths. As the inexorable march of EarthScope's transportable array of seismic stations proceeds across the country, we are approaching a time when assumptions about load statistics may be downweighted or even unnecessary in isostatic analyses (in the conterminous US at least). Seismic velocities (e.g. from P_n, body and surface wave tomography) and layer thicknesses (from P-S conversions) may be used via regression and/or correlation analyses with gravity/topography data to arrive at independent estimates of internal loading that can then be used to better constrain T_e. Seismic data will also aid in overcoming another significant limitation of Forsyth's deconvolution approach. Load deconvolution assumes the internal load mass occurs at a single depth (dictated by the fact that we can solve for at most two unknown amplitudes from the two observations of gravity and topography), whereas true internal loading occurs at multiple layer interfaces and as density variations within layers. Incorporation of seismic data into the isostatic analyses holds the promise of overcoming these limitations altogether and substantially reducing errors in isostatic analyses introduced by erroneous load estimates. I will present a new approach to incorporating seismic data in T_e estimation procedures. Preliminary tests on data profiles from the Himalayas and European Alps suggest that adding even a skeletal seismic constraint may provide a more robust estimate of T_e than minimization of the load coherence alone.
DE: 7218 Lithosphere (1236)
DE: 8120 Dynamics of lithosphere and mantle: general (1213)
DE: 8138 Lithospheric flexure
DE: 8159 Rheology: crust and lithosphere (8031)
DE: 8194 Instruments and techniques
SC: Tectonophysics [T]
MN: 2007 Fall Meeting