Our observations made on dry-sandstone ultrasonic velocity data relate
to the variation in velocity (or modulus) with effective stress, and
the ability to predict a velocity for a rock under one effective press
ure when it is known only under a different effective pressure. We fin
d that the sensitivity of elastic moduli, and velocities, to effective
hydrostatic stress increases with decreasing porosity. Specifically,
we calculate the difference be tween an elastic modulus, M(P-1, phi),
of a sample of porosity phi at effective pressure P-1 and the same mod
ulus. M(P-2, phi), at effective pressure P-2. If this difference, Delt
a M = M(P-1, phi) - M(P-2, phi), is plotted versus porosity for a suit
e of samples, then the scatter of Delta M is close to zero as porosity
approaches the critical porosity value, and reaches its maximum as po
rosity approaches zero. The dependence of this scatter on porosity is
close to linear. Critical porosity here is the porosity above which ro
ck can exist only as a suspension-between 36% and 40% for sandstones.
This stress-sensitivity pattern of grain-supported sandstones (clay co
ntent below 0.35) practically does not depend on clay content. In prac
tical terms, the uncertainty of determining elastic moduli al a higher
effective stress from the measurements at a lower effective stress is
small at high porosity and increases with decreasing porosity. We exp
lain this effect by using a combination of two heuristic models-the cr
itical porosity model and the modified solid model. The former is base
d on the observation that the elastic-modulus-versus-porosity relation
can be approximated by a straight line that connects two points in th
e modulus-porosity plane: the modulus of the solid phase at zero poros
ity and zero at critical porosity, The second one reflects the fact th
at at constant effective stress, low-porosity sandstones (even with sm
all amounts of clay) exhibit large variability in elastic moduli. We a
ttribute this variability to compliant cracks that hardly affect poros
ity but strongly affect the stiffness. The above qualitative observati
on helps to quantitatively constrain P- and S-wave velocities at varyi
ng stresses from a single measurement at a fixed stress. We also show
that there are distinctive linear relations between Poisson's ratios (
nu) of sandstone measured at two different stresses. For example, in c
onsolidated medium-porosity sandstones nu(40) = 0.018 + 0.913 nu(20),
where the subscripts indicate hydrostatic stress in MPa. Linear functi
ons can also be used to relate the changes (with hydrostatic stress) i
n shear moduli to those in compressional moduli. For example, G(40) -
G(20) = 0.084 + 0.344 (M(40) - M(20)), where G = rho V-S(2) is shear m
odulus and M = rho V-P(2) is compressional modulus. both in GPa. and t
he subscripts indicate stress in MPa.