Methane hydrates are increasingly recognized as a potential future energy resource. In order to evaluate the actual reservoir potential of hydrated sediments, it is important to know the actual amount of hydrate present, as well as its mobility and recoverability. Recently, several attempts have been made to estimate the amount of hydrate directly from seismic velocities and/or amplitudes. Most studies determine the hydrate saturation using Wyllie's time average equation Wyllie et al. (1958), which directly relates acoustic velocity, porosity and saturation Korenaga et al. (1997); Sholl and Hart (1993); Wood et al. (1994). In order to obtain a more reliable estimate in high-porosity sediments, Lee et al. 1992 used a weighted mean of the Wyllie equation together with Wood's equation Wood (1941). Constructing a velocity-porosity relationship based on core and drilling results, Yuan et al. 1996 translated the observed velocity increase associated with the appearance of hydrate directly into a porosity reduction caused by the presence of hydrate and thus hydrate saturation. All of these investigations estimated 1-D hydrate saturations and assumed known porosity, but their analyses did not consider the effect of the position of hydrate in the pore space. This, however, might have an considerable effect on the saturation estimates.
In this paper, our goal is to provide a theoretical tool for estimating the amount of hydrate without prior well information and for inferring the internal structure of the hydrated sediment at the Blake Outer Ridge. We develop new theoretical rock-physics models that link the elastic wave velocities in high-porosity sediments to density, porosity, effective pressure, mineralogy, and water/gas and hydrate saturation. The first model assumes that hydrate is part of the pore fluid, and does not affect the solid frame at all. In the second model, the hydrate becomes part of the solid frame, thus reducing porosity and slightly affecting the sediment properties. Since it is often assumed that hydrate is cementing the sediments, we also employed a third model where hydrate is cementing the grain contact and therefore strongly reinforcing the sediments. This third model is probably physically unlikely at the Blake Outer Ridge because of the highly unconsolidated sediments in this regions. However, in order to show the significant variations in hydrate saturations that result when different models are applied, we have included it into our investigation.
We use the three rock-physics models together with the interval velocities obtained from seismic stacking velocity analysis to calculate a porosity section with the assumption that all of the sediment is fully brine saturated. This results in porosity anomalies where gas hydrate and free gas is present. These porosity anomalies between brine saturated sediments and the hydrate/gas saturated sediments can be related to hydrate and gas saturations. We calculate the needed background brine porosity in two ways. First, assuming uniform lithology, we estimate an average porosity trend from the part of the seismic line that presumably does not contain either gas or hydrate. Second, we determine the background porosity on a trace-by-trace basis by fitting polynomials to the near-surface and deep sediment porosities thus excluding the anomalous porosities in between. Inverting the resulting residuals for hydrate saturation, we obtain maps of lateral hydrate and gas distribution and saturation. Applying our analysis to well-logs -from the region using known velocity and porosity trends, we find that the results are consistent. In order to obtain more structural saturation maps, we repeat the analysis by adding the high frequency impedance information.
Based on the different rock-physics models and obtained hydrate saturations, we generate synthetic seismograms using Kirchhoff modeling. Comparing the seismic AVO responses of the different models, we conclude that it is is not possible to differentiate between the model (1) and (2). However, only those two models can reproduce the actually observed AVO responses of the seismic data. Thus, we conclude that hydrate is either part of the fluid or part of the solid and thus reduces porosity and slightly changes the sediment stiffness.