Modes of Faulting at Mid-Ocean Ridges

by W. Roger Buck, Luc L. Lavier and Alexei N. B. Poliakov

3-4pm
Friday October 11, 2002
Refreshments served at 2:45pm
Munk Conference Room
Cecil and Ida M. Green Institute of Geophysics and Planetary Physics
Scripps Institution of Oceanography
University of California, San Diego
http://mahi.ucsd.edu/seminar/

Abstract

Many differences between normal faults generated at fast and slow
spreading ridges are recognized.  Fast spreading faults show about a
tenth the vertical displacement seen for typical faults formed at slow
spreading ridges.  Essentially all faults mapped at slow spreading
ridges dip toward the spreading axis while about 50% of normal faults
seen at fast spreading ridges dip away from the axis.  The length -
frequency distribution and displacement - length ratios are markedly
different for fault populations mapped on fast and slow spreading ridges.
Teleseismic earthquakes are generated on slow spreading ridges and not
on fast spreading ridges.  Given all these differences it is somewhat
surprising that the standard model for faults at ridges holds that all
ridge faults are produced by tectonic stretching.  The standard model
explains differences in faults in terms of the thickness of lithosphere
cut by faults and the time interval in which no magma is supplied to the
ridge, during which time plate separation is accomplished by stretching.

We report on a set of numerical models for faulting at ridges that
does not follow these assumptions.  We consider model magmatic dikes
that cut the entire thickness of brittle lithosphere at a spreading
axis and that widen at a constant rate described by the fraction, M,
of the total rate of plate separation.  The model produces areas of
concentrated brittle strain, analogous to faults, because we allow
for strain dependent weakening of the brittle lithosphere.  For an
axial thermal structure inspired by data from slow-spreading ridges
we see stretching faults as long as M<1.  The amount of horizontal
offset on the average stretching fault depends on M, but the vertical
relief across the resultant axial valley does not depend strongly on M.
For M close to 0.5 the model produces faults on one side of the ridge
with tens of kilometers offset analogous to the "core complex faults"
seen at some slow spreading segment ends.

For our model of a fast spreading ridge a low-density, presumably a
partially molten zone, buoys up the axis in these model cases and we
get an axial high that is wider than the region of low-density material.
This is consistent with the observed topography and gravity of typical
axial highs.  M is set to 1 so that no stretching faults are produced.
The bending of model lithosphere as it moves away from the axis produces
extensional bending faults at the surface that dip half toward and half
away from the axis.  The models only fit observations for a range of
rates and amounts of fault strain weakening.