BSR AVO Response

In order to emphasize the effect of the different models on the amplitudes at the BSR, I evaluate the AVO response directly at the BSR. This is done by picking the maximum (negative)BSR amplitudes.

 

syntavo
Figure 9 Synthetic BSR AVO picks for the case of hydrate overlying gas. The upper panel shows the picks for hydrate models A and B. The lower panel shows the picks for hydrate model C.

The resulting AVO curves for the case of gas underlying hydrate sediments are shown in Figure . The AVO behavior on the upper panel represents those obtained from hydrate models A and B. The AVO behavior in the lower model represents the one obtained from hydrate model C. Neither model A (hydrate is part of the fluid) and model B (hydrate becomes part of the solid frame) can be distinguished based on their amplitudes. Both of them result in an increasingly negative amplitude with increasing offset. Model B does affect the shear wave velocity in the hydrated sediment differently. However, since this model requires less hydrate saturation to reproduce the acoustic interval velocities, the effect on the shear properties become less pronounced as when both model A and model B would require the same amount of hydrate saturation. The cementation model, on the other hand, displays a clearly opposite amplitude behavior. The amplitudes are decreasing with increasing offset.

The picked AVO curves for the case of hydrate underlain by brine-saturated sediment can be seen in Figure . As before, the upper panel represents the AVO behavior of hydrate models A and B, while the lower panels shows the AVO behavior of model C. Once again, models A and B cannot be differentiated from seismic. They show a slightly less increasing amplitude trend with increasing offset. The most pronounced difference from the gas-saturated case is the considerable decreased zero-offset reflection coefficient, which can be clearly attributed to the presence of brine underneath the hydrate.

 

syntavo-ng
Figure 10 Synthetic BSR AVO picks for the case of hydrate overlying brine sediments. The upper panel shows the picks for hydrate models A and B. The lower panel shows the picks for hydrate model C.

Comparison of these AVO trends with that of the real data (Figure ) shows first of all that hydrate underlain by gas-saturated sediment could reproduce the zero-offset reflection coefficient reasonably well. Furthermore, the hydrate cementation model results in an AVO behavior which is clearly opposite to the one shown in the real data. Even though the magnitude of the amplitude increase displayed by the data cannot be quantitatively matched by either model A or B, the overall trend is in good agreement.

Based on this modeling approach, I can conclude that seismic cannot differentiate between hydrate as part of the pore fluid and hydrate becoming part of the solid. However, one of those two models or a combination of both appears to match the seismic data qualitatively. Thus, the effect of hydrate on the in-situ structure must be fairly small. The sediment frame is not significantly stiffened, as suggested by a cementation model, but appears to be only weakly, if at all, affected by the presence of hydrate.

It furthermore appears that cement positioned in the sediment may strongly affect permeability: at the same high porosity, rocks with contact cement may have higher permeability than those with pore-filling cement. This results from the numerical modeling of Cade et al. Cade et al. (1994). The inferred sediment structure of the hydrated sediment at the Blake Outer Ridge requires hydrate to be either part of the fluid or the solid frame. Such hydrate deposition schemes can lower permeability considerably compared to the cementation model. Hydrate clogging large pore space conduits could therefore seal the BSR. This could explain why free gas is trapped beneath the BSR.