Methodology Description

In order to link the seismic interval velocity with the proposed rock-physics models, we need to know either the porosity of the sediments or the actual brine, gas and hydrate saturations. Solving for both porosity and saturation simultaneously would pose a non-unique problem: a known interval velocity could be reproduced from any combination of porosity and saturation, however unphysical. Since there is no direct well-control on the seismic data from the Blake Outer Ridge, both porosity and saturation are unknown. Therefore, an initial assumption has to be made about either one of them. Since there is no real physical relationship that would govern the behavior of the porosities throughout the sediments, I base my first analytical step on fixing the saturation: I assume the entire sediment section to be 100% brine-saturated. Physically, this assumption will be true everywhere but for the regions of hydrate and gas saturation, where it will produce porosity anomalies. I then calculate what I call the baseline porosities by posing an ``inverse'' problem: using a saturation of 100% brine, I change the porosity until the rockphysics model can reproduce the seismic interval velocities. This results in anomalies where the sediment is not 100% brine-saturated. The porosity is underestimated in areas of hydrate saturation, while it is overestimated when free gas is present.

In order to relate these anomalies to the actual amount of hydrate and gas saturation, a measure of the anomaly size is needed. The anomalous baseline porosity has to be related to a porosity without the effects of both hydrate and free gas. I offer two techniques of determining such a reference ``normal'' porosity:

• calculating the average porosity trend from the region without BSR
• calculating the porosity trend on a trace-by-trace basis by fitting polynomials to the porosities above the hydrate and below the gas (thus assuming that those parts of the sediment is fully brine-saturated)

Both of those methods have pitfalls. In the first, I will neglect any possible lateral variations of porosity based on actual lithology changes. Furthermore, any error in interval velocity which was directly mapped into the porosities will be propagated. Also, by calculating the average trend from the region without BSR, I assume that those sediments do not contain gas hydrates. This assumption, however, may or may not be true for the sediments at the Blake Outer Ridge, as the drilling has shown Matsumoto et al. (1996). The trace-by-trace approach does include lateral lithology variations and does not propagate velocity errors. However, it suffers from the lack of reflectors below the water bottom towards the end of the seismic line (see Figure , close to 50 km lateral distance). Combining both methods, however, should enable me to minimize the errors in the hydrate and gas saturation estimates and to obtain reasonable upper and lower bounds of possible saturations.

After having obtained estimates of the size of the porosity anomalies caused by the presence of hydrate and gas, I can relate them to the actual hydrate and free gas saturation. This is done by using the original baseline porosities, obtained by assuming 100% brine saturation everywhere, as a starting porosity model. I then pose again an ``inverse'' problem: I decrease the assumed brine saturations and increase hydrate or gas saturation until I can minimize the porosity anomalies, calculated by either the averaging or the trace-by-trace technique, and until I can fit the seismic interval velocities. Negative anomalies are considered to be caused by the hydrate, thus the hydrate saturation is increased in these regions. In areas of positive anomalies, on the other hand, the gas saturation is increased.