The five parts of the stack section analyzed are illustrated in Figure F4. We have picked seven to nine reflectors with spacing of about 100 ms or less (Fig. F5). Purposely, we did not pick reflectors too close to the seafloor reflector in order to avoid the interference of bubble-induced ringing or streamer ghosts. The picking of reflectors was particularly easy in the topsets of Unit S1. In the foresets of Unit S2 and in deeper dipping reflectors of Units S3 and S4, the coherency of reflections is rather weak. The coherency of the reflectors was found to be highest, particularly in common-offset gathers, in traces from 10 to 29 (offsets from 329.5 to 567 m), and these traces have been selected for picking on common-trace and common-offset gathers.
The tomographic inversion has been applied to one of every four shots. As expected, the coherency of the inversion of the traveltimes is weaker in foreset reflectors (e.g., see Fig. F3), but it has always been possible to obtain a satisfactory final solution for all picked reflectors. The uncertainty in the velocity estimation with this method is typically a few percent. Although we have no estimates of the uncertainty in this application, a ±2% error was estimated in a recent application of this method by us in an area with generally poorer reflector continuity on the nearby South Shetland margin (Tinivella et al., 1998, fig. 8).
The two-dimensional (2-D) velocity structure across the location of Site 1102 is illustrated in Figure F6A. All reflectors picked are within Unit S1. The shelf break occurs at shotpoint (SP) 538, about 130 m seaward of the site location. The upper part is divided into two topset layers of moderately increasing velocity with depth. Weak lateral velocity changes that do not suggest significant trends are present in these upper layers. Within the third interval, the transition from topset reflectors in the southeast to foreset reflectors in the northwest occurs. A sharp lateral velocity change within this interval, from 1.8 to 2.7 km/s, with increasing velocity toward the northwest, is coincident with such a transition. All underlying intervals are composed of foreset reflectors. Weak lateral velocity changes that do not suggest significant trends are also present in these lower layers. A sharp velocity increase from 1.8 to 3 km/s occurs to the southeast of the site location, between the topset reflectors of the third interval and the foreset reflectors. This coincides with the presence of foreset reflector terminations against the topset reflectors.
A comparison between the tomographic velocity profile at the location of the site and the nearest available stacking velocity profile (SP 575, 875 m southeast of the site location) is illustrated in Figure F7A.
The 2-D velocity structure across the location of proposed Site APSHE-02A is illustrated in Figure F6B. The topsets of Unit S1 are divided into four layers of increasing velocity with depth. A sharp velocity increase occurs between the first and the second layer of the topset unit. Weak lateral velocity changes that do not suggest significant trends are present in this unit. The velocity increases from about 1.6 to 2.7 km/s in the upper 350 m of sediment. In foresets of Unit S2, the velocity increases from 2.8 to 3.3 km/s. A sharp velocity increase occurs between the lower boundary of the topsets and the upper boundary of the foresets at the southeast end of the analyzed section. This coincides with foreset reflector terminations against the topset reflectors. Northwest of the site, the velocity across the topset/foreset boundary does not increase sharply. A sharp lateral velocity gradient, with increasing velocity to the southeast, is present within the second interval of the foresets of Unit S2.
A comparison between the tomographic velocity profile at the proposed site location and the nearest available stacking velocity profile (SP 895, 175 m northwest of the site location) is illustrated in Figure F7B.
The 2-D velocity structure across the location of proposed Site APSHE-03A is illustrated in Figure F6C. Unit S1 is divided into four layers of increasing velocity with depth. Weak lateral velocity changes are present. The velocity increases from about 1.7 to 2.7 km/s in the upper 400 m of sediment. In the foresets of Unit S2, the velocity increases from 2.95 to 3.3 km/s. Qualitatively, the differences in the velocity distribution between topsets and foresets found at this site location are similar to those described above for the location of proposed Site APSHE-02A location, including lateral gradients within foresets.
Figure F7C illustrates the comparison between tomographic and stacking velocity 700 m southeast of the site location (SP 975).
The 2-D velocity structure across the location of proposed Site APSHE-04 is illustrated in Figure F6D. The topsets of Unit S1 are divided into three layers of increasing velocity with depth. Lateral velocity changes are present in this unit: in the first layer an increase occurs from the southeast end of the analyzed section toward the northwest; in the third layer, an even sharper increase occurs from the northwest end of the analyzed section toward the site, and hence, gradually decreases. Within this unit, the velocity increases from ~1.7 to ~2.7 km/s in the upper 300 m of sediment. Unit S2 is divided into two layers of increasing velocity with depth. Also, at this location, velocity increases sharply at the boundary between topsets and foresets. An exception is found directly northwest of the site location, where the velocity across the topset/foreset boundary does not increase sharply because of the relatively high velocity in the topset layer. A velocity decrease (velocity inversion) occurs at the boundary between glacial reflectors of Unit S2 to "early glacial or non-glacial" reflectors of Unit S3. Within S3, the velocity increases from 2.8 to about 3.5 km/s. A lateral gradient is present in the lowest layer, with a decrease toward the outer shelf (northwest).
See Figure F7D for a comparison between the tomographic velocity at the site location and the nearest available stacking velocity profile (SP 1375, 125 m southeast of the site location).
The 2-D velocity structure across the Site 1103 location is illustrated in Figure F6E. The topsets of Unit S1 are divided into three layers of increasing velocity with depth. The velocity increases from about 1.6 to 2.5 km/s in the upper 250 m of sediment. Unit S2 is missing at this location. Within Unit S3, the velocity increases from 2.7 to 3.0 km/s. The sharp velocity increase below the lower boundary of Unit S1 found at all other outer shelf sites is not present here. The velocity increases more gradually throughout and lateral gradients are present only in the second layer of Unit S1, with an increase toward the inner shelf (southeast). The only sharp increase downward is between Units S3 and S4 (from 3 to 3.4-3.5 km/s).
See Figure F7E for a comparison between the tomographic velocity profile at the site location and the nearest available stacking velocity profile (SP 1695, 300 m northeast of the site location).
According to the velocity fields presented, we have identified the time-depth relationships for the seafloor, the bottoms of the drilled holes at the three drilled sites, and the S1/S3 boundary at Site 1103 (Table T2).