Regarding the idea of detecting gravitational variations prior
to earthquakes:

The gravitational force is an effect between two masses m1 and m2
separated by some distance r. For a large object, the center of mass
of the objects is the reference distance. The idea that gravity comes
"from the core" of the earth and "cuts across" some part of the surface
is rather far fetched. This implies that the gravitational force is
some sort of "beam", rather than the omnidirectional force that it is.
The formula  for the force (F = G*m1*m2/r^2, G = 6.67 x 10^-11 N*m^2/kg^2)
has no directive components; it is simply between the two objects.

A good reference for studies of the core-mantle-boundary (CMB) is:

	"Earth's Deep Interior", edited by D.J.Crossley; Gordon and Breach 
	Science Publishers; ISBN 9056990322.

(Dr. Crossley is the Chair of the Department of Earth and Atmospheric
Sciences at St. Louis University.  He is a specialist in superconducting
gravimeter systems and in interpreting the data).

The book contains 11 research papers on the CMB conditions, and how
they are determined from geomagnetism, gravimetry, seismology, and
geochemistry. It contains hundreds of references.

Several papers estimate the velocity at the CMB, which is the rotation
of the outer core with respect to the lower mantle across the transition
zone called the D" layer. Velocities of 4 x 10^-4 m/sec or 0.15
degree/year result from the most recent models;  converting this to 
more familiar rates results in a velocity of 0.009mph. RMS values range
from 12.7 to 14.3 km/year (4.03 to 4.53 x10^-4 m/sec), depending on 
the theoretical model chosen. All the models and the math are very complex.
The flow of the core surface is also spatially irregular in both rate
and direction, with the favored model called a "tangentially geostrophic"
flow ( in the paper by Whaler and Davis in the book).

It is not difficult to expect that the CMB is a rather irregular surface.
And if there are irregularities in the mass distribution, they will cause
variations in the gravity vectors that involve them. But detecting them
has not yet been possible, even with superconducting gravimeters with a
resolution of 1 nanogal ( 10^-11 of g).

Gravimeter data is recorded in micro-gals, (10^-8 of g) for historical
reasons. I have been operating a quartz-spring recording gravimeter at
SLU since 1969. The dominant output is the 12 to 24 hour lunar-solar
tides at 50 to 300 microgals (ugals). Instrument drift is around 100
ugal/year. To detect smaller effects, these have to be removed, as well
as barometric effects, ocean tide loading, etc. This can reduce the
noise level to less han 1 ugal, where oscillations of the core that are
excited by large earthquakes can be detected. The primary nutation of the 
outer core is 433 days.

Roughness or texture of the CMB HAS been detected by tomographic studies
using modern broadband data. But the resolution is still poor until the
density of VBB stations is significantly increased. Modern VBB seismometers
reliably record earth tides as well, with resolution at the ugal level.

But since the gravitational effect of any texture at the CMB is not
currently detectable at the 0.001ugal level, the effect is orders of
magnitude less than the lunar-solar gravitational changes of 50 to
300 ugals that occurr daily. Some have proposed these forces as trigger
mechanisms for earthquakes, but the numbers don't agree any better than
random chance.

Regards,
Sean-Thomas

_____________________________________________________________________

Public Seismic Network Mailing List (PSN-L)