Apparent discrepancies between velocity measurements made with differe
nt frequencies in a formation at the Gypsy test site are explained in
terms of elastic scattering and intrinsic attenuation. The elastic sca
ttering component of the dispersion (38%) in a marine interval above t
he Gypsy sandstone is estimated via simple models constructed from wel
l log information. Any dispersion above the predicted value for elasti
c scattering in this interval is assigned to intrinsic attenuation (62
%). Using the vertical measurements in the well, the marine interval d
irectly above the Gypsy sandstone has an estimated intrinsic Q(I) = 51
and an effective Q because of the scattering of Q(SC) = 85. The total
Q of the combined mechanisms is 32. The dispersion of the vertical me
asurements through the heterogeneous sands and shales of the Gypsy for
mation can be explained using an intrinsic Q(I) = 30 and neglecting th
e effects of scattering. The horizontal observations require a more de
tailed modeling effort to unravel the relative roles of path and volum
e effects, elastic scattering, attenuation, and intrinsic anisotropy.
Thin layers barely resolvable on the sonic logs play a significant rol
e in modifying the crosswell response. Potentially, the dispersion can
be a key to mapping reservoir properties using crosswell and surface
seismic data.