RMS and Interval Velocity Errors

I evaluate the accuracy of my velocity models on one CMP gather, located above both the hydrate-saturated sediments and the gas-saturated sediments. Figure displays its velocity scans overlain by the obtained RMS velocity trend (left and middle panel) as well as the resulting NMO corrected gather (right panel). The main reflections appear to be moved out with an accurate velocity and the velocity scans display a good agreement between the RMS picks and the maximum semblance. The interval velocity resulting from the displayed RMS velocity trend can be see in Figure . It shows, as described in the previous section, an increase to about 1.9 km/s in the hydrated zone, and a subsequent decrease in velocity to approximately 1.7 km/s in the region where free gas might be present.

The accuracy of the interval velocity depends on the accuracy of the RMS velocity picks. RMS velocity picking errors mainly depend on the width of the maximum semblance at a reflector. Based on the width of the semblance at the BSR and the reflector underneath, I introduce a possible RMS velocity picking error of $\pm$ 10 m/s. The resulting RMS velocity trends are displayed in the left panel of Figure . The solid line represents the original RMS velocity trend, while the two dashed lines represent the velocities resulting from $\pm$ 10 m/s picking errors.

Using the Dix equation these different RMS trends are converted into interval velocities which are shown in the right panel of Figure . The solid line represents the interval velocity obtained from the original RMS velocity. The interval velocities are characterized by discrepancies which are as high as $\pm$ 200 m/s (more than 10%). The relatively small errors in RMS velocity introduced by picking can therefore yield considerable uncertainties in the interval velocities. As a consequence, the anomalous velocity zones in the hydrate and the layer underneath might be enhanced (dashed line) or suppressed (double dashed line). In the latter case, where the velocity appears to be steadily increasing with increasing time, the large negative reflection coefficient at the BSR needs to be caused by a strong negative density contrast with simultaneous increase in velocity. Considering the lithologically fairly homogeneous sediment structure at the Blake Outer Ridge, this does not seem feasible.