Saturation Uncertainties

Using the above described interval velocities obtained after introducing a reasonable picking error into the RMS velocities, I repeat the method of estimating hydrate and gas saturations. This will yield the robustness of the hydrate and gas saturations to errors in velocity.

The previous analysis (section 4.3) showed that in the case of hydrate cementing the grains, only very little ($\le$ 1%) hydrate saturation is required to fit the seismic data. I do not anticipate that errors in velocity will considerably change these estimates based on the drastic effects only small amounts of hydrate have on the sediment stiffness. Therefore, I only investigate the effect of velocity errors on the saturations for hydrate models A and B.

As described in section 4.3, I assume the sediment to be 100% brine-saturated and use the interval velocities to calculate the baseline porosities. These baseline porosities are then compared with the normal reference porosity in this region, i.e., a porosity which is not affected by the presence of hydrate and gas in the pore space. I chose to compare the baseline porosities with the calculated average porosity in the region (calculated in section 4.3). I also could have compared the baseline porosities with the normal porosity calculated by the trace-by-trace method. However, the sensitivity of both the average and the trace-by-trace method to the velocity errors is similar. Therefore, I only show the results using the averaging method.

The three baseline porosities obtained from the three different interval velocities (see Figure ) are shown in Figure . They are overlain with the average porosity trend that was calculated in section 4.3. The left panel shows the baseline porosity calculated from the original interval velocity. The middle panel represents the porosity in the case of enhanced hydrate and gas velocity zones. It shows an increased hydrate and gas anomaly. The right panel represents the porosities obtained from the velocity, which increased steadily. The anomaly caused by the gas disappears completely, while there is still a hydrate-related anomaly in the very near-surface sediments. However, additional velocity errors in this region could make this anomaly disappear as well.

 

base-error
Figure 14 Baseline porosities for the three velocities shown in Figure 4.13, overlain by the average average porosity calculated in section 4.3. The left panel represents the porosity trend obtained from the original velocity. The middle panel represents the porosity determined from the velocity which enhances both the hydrate and gas velocity zones.; and the right panel shows the porosity calculated from the velocity which suppresses the velocity increase due to hydrate or decrease due to gas.

From these new hydrate and gas porosity anomalies, I calculated again the hydrate and gas saturation. The results, using hydrate model A and hydrate model B, can be seen in Figure . The left panel shows the saturations obtained for hydrate model A, in which hydrate is solely part of the fluid. The left panel represents the saturations obtained for hydrate model B, in which hydrate is assumed to become part of the solid sediment frame. The solid curve represents the saturations obtained from the original velocity; the dashed line the saturations obtained from the velocity trend that increases the hydrate and gas velocity anomalies; and the double dashed line the saturations for the velocity trend that suppresses both gas and hydrate anomalies. The saturations show that there is an approximately a 14% discrepancy between the original saturation and the one using enhanced velocities. The saturations resulting from the suppressed velocity trend (double dashed line) still display hydrate saturation in the near-seafloor sediments. Additional errors in velocity in that region could, however, cause this saturation to drop all the way to zero. Therefore, it appears to be reasonable to assume that a $\pm$ 14% uncertainty can theoretically be introduced into the hydrate saturations by errors in the velocity. The gas saturation displays an error of about $\pm$ 2% (note that % does not refer to percentage of saturation but to saturation itself.)

 

sat-error
Figure 15 Saturation estimates for hydrate model A (left panel) and hydrate model B (right panel) resulting from errors in the interval velocity. The solid line represents the saturations based on the original velocity. The dashed line represents the saturations calculated from the velocity enhancing both hydrate and gas velocity anomalies; and the double dashed line represents the saturations for the velocity trend suppressing hydrate and gas anomalies.

This uncertainty assessment has shown that the hydrate saturation is sensitive to errors in interval velocity. The saturation values can theoretically vary up to $\pm$ 14%. A Comparison with VSP velocities (see Chapter 2, section 2.3.2) and with other seismic velocity measurements in the region Katzman et al. (1994); Korenaga et al. (1997); Wood et al. (1994) suggests, however, that my initially obtained interval velocities represent the overall velocity trends in the region of the Blake Outer Ridge fairly well. Thus, an error of $\pm$ 14% in saturation represents an upper uncertainty bound, but the actual errors might be much smaller.