Knowledge of homogeneous and heterogeneous fluid-distribution conditions in unconsolidated sediments is important for the selection of remediation techniques for groundwater contamination. However, for unconsolidated sediments, fluid-distribution conditions from laboratory tests on core samples may not be representative of in situ conditions. We have developed a seismic inversion method to determine in situ fluid-distribution conditions that involves inverting experimental seismic P- and S-wave velocities using Hertz-Mindlin and Biot-Gassmann models with different averaging methods (Wood and Hill averages) and different fluid-distribution condition assumptions. This method can determine whether seismic velocity-versus-depth profiles are better explained assuming heterogeneous or homogeneous saturation conditions in shallow ( depth) unconsolidated sands. During the imbibition and drainage of shallow unconsolidated sands, we have observed nonmonotonic relationships between P-wave velocity and water levels (WLs) as well as an S-wave velocity and WLs that were consistent with other field and laboratory observations. This relationship can be explained by transitions between the lower Wood bound and the higher Hill bound. The transition is possibly caused by the alternation in the size of fluid patches between small and large during the imbibition and drainage. Inverted results can be verified by a good correlation (difference ) between the inverted and measured water saturation using moisture sensors.