Assessing the influence of geometrically non-linear effects on
earthquake fault plane inversions.
G.J. van Zwieten, M.A. Gutierrez and R.F. Hanssen
To deduce from coseismic surface displacement the exact source
parameters of an earthquake through inversion, one needs a forward
model that predicts displacements for given fault geometry and
slip. Often used for this purpose are the analytical solutions by
Okada for constant dislocations in an elastic half space, thereby
assuming that the earth behaves approximately elastically over short
time scales.
Although the Okada solutions solve the governing differential
equations exactly, they do have a potential flaw. That is: the
assumption of geometric linearity that lies at the basis of these
equations is valid everywhere in the solid except at the discontinuous
displacement jump that is formed by the fault. A large jump will lead
to a non-physical discontinuity in the computed stress field.
This contribution presents a study of this potential problem. Using
finite element code written specially to enforce continuity of stress
everywhere in the domain is possible to compare the results of linear
and non-linear computations. The main question that will be answered
is what can be gained by using a forward model in fault plane
inversions that correctly deals with non-linearity.