Well Log Comparison

In order to validate the technique of estimating hydrate and gas saturations using only interval velocities, I apply it to the known well-log velocities from wells 994 and 995. Besides known velocities, the wells also provide known porosity functions from cores which represent the normal porosities without the effect of hydrate or gas. Therefore, the application of the technique without known porosities can be compared to the equivalent analysis using the known core porosities.

The smoothed well velocities are shown in Figure . They are used together with the rock-physics models to invert for a baseline porosity trend by assuming 100% brine-saturation throughout the sediments. Figure shows the baseline trend overlain by the fitted porosity trend and the true (core) porosity trend. The comparison shows that for well 994, the true normal porosity (dotted line) and the inferred normal porosity (polynomial fitting) are in reasonable agreement. At well 995, there is a discrepancy between the true normal porosities and the one obtained by fitting a polynomial to the baseline porosity.

 

wellvel
Figure 16 Smoothed well-log velocities from well 994 (solid line) and well 995 (dashed line).

Using the porosities displayed in Figure , I calculate an estimate of the hydrate anomaly by (1) subtracting the true porosity from the baseline porosity and (2) subtracting the fitted porosity from the baseline porosity. The resulting anomalies are then used to calculate hydrate saturation.

 

wells
Figure 17 Porosities for well 994 (left panel) and well 995 (right panel). The real core porosities are given by the dotted lines, the baseline porosities by the solid lines, and the fitted normal porosities by the dashed lines.

The resulting saturations for hydrate models A and B can be seen in Figure . The saturations obtained using the true porosity are represented by the solid lines, while the ones determined using purely the velocities are given by the dashed lines. Since the cementation model (model C) would again result in both cases in having less than 1% hydrate saturation, I did not investigate it here. The comparison of the results indicates that there is a good fit between the saturations in well 994, both for hydrate model A and hydrate model B. The maximum hydrate saturation is at about 30 % for hydrate model A and 21% for hydrate model B. The range of these estimates is consistent with the results I obtained from the seismic data. As expected from the porosity misfits at well 995, the resulting saturations agree less than those in well 994. The discrepancy is less than 10% above a depth of 3.3 km. The more significant discrepancy below 3.3 km can be attributed to the lack of velocity data in the lower portion of the well which affects the accuracy of the polynomial fitting. It has to be noted, however, that the overall magnitude of the hydrate saturation is again consistent with the results obtained from seismic. Therefore, the methodology offered to estimate hydrate saturation from only velocity appears to be quantitatively accurate.

 

wellsat
Figure 18 Hydrate Saturations for well 994 (upper figures) and well 995 (lower figures). The solid line represents the saturations obtained using the real porosities. The dashed line represents the saturation obtained using the technique of inverting for porosity.