We have shown how we can successfully infer hydrate saturation estimates from surface seismic using three different micromechanical models of hydrate deposition in the pore space. The three models result in different amounts of hydrate present in the pore space. Using P-wave velocity information alone, we are not able to distinguish between those different models. Therefore, we use AVO analysis in an attempt to determine the internal structure of hydrated sediment present at the Blake Outer Ridge.
Figure 14 shows a representative CMP gather after normal moveout correction which displays the amplitude variation of the BSR reflection. This is the reflection caused at the transition from hydrate bearing sediments to gas saturated sediments. The right panel shows the amplitude picks taken along the BSR. It displays an increasingly negative amplitude with increasing offset. Since the different hydrate deposition models will have different effects on the elastic sediment moduli. Therefore, forward modeling and subsequent AVO analysis might be a tool to differentiate between them and thus determine the actual scenario present.
 |
First, we construct a simple 1-D model representing the three different hydrate deposition schemes. The model assumes a thick water layer overlaying brine sediments, hydrated sediments, gas sediments and, again, brine saturated sediments. The model parameters are shown in Appendix C. Using those parameters, we forward model the elastic sediment properties using the same techniques and sediment mineralogy (see Appendix B) as described before. The elastic moduli are then used as input into a Kirchhoff modeling program Lumley and Beydoun (1993), resulting in three CDP gathers. Those three gathers are shown in Figure 15.
 |
Models 1 and 2 result in a BSR reflection with amplitudes which seem to be slightly increasing with increasing offset, while model 3 shows a clear decrease of amplitudes with increasing offset. It is obvious that hydrate cementation causes a significantly different amplitude effect than deposition of hydrate in the pore space. Comparing the synthetic gathers with the seismic gather in Figure 14 suggests that the in-situ conditions are more likely to be represented by either hydrate deposition model 1 or 2. This is consistent with the actual physical sediment properties in this region. Due to their significant unconsolidation, a cementation model is not likely. The synthetic seismograms show, however, how clearly distinguishable model 3 is from the other two models. In geological setting where all three hydrate models pose a physical possibility, seismic modeling will clearly distinguish between model 3 and model 1 or 2.
 |
In order to emphasize the effect of the different models on the amplitudes at the BSR, we evaluate the AVO response directly by picking the amplitudes along this reflection. The result is shown in Figure 16. The left panel shows the amplitudes characterizing model 1 and 2, the right panel those of model 3. The amplitude plot indicates that we can not differentiate between hydrate model 1 and hydrate model 2 in this setting. Both result in increasing negative amplitude with increasing offset. Both model 1 and 2 are ,however, clearly separable from model 3. The amplitude picks furthermore show that there is a relatively good agreement between the synthetic zero-offset reflection coefficient and the in-situ one. This indicates that our model could reproduce the acoustic impedances of the in-situ conditions considerably well.