Figure F2 shows the synthetic seismograms and MCS data along line EW95-1366 for Sites 1118, 1109, and 1115. A 200-ms AGC window has been applied to both data sets. The units defined herein for Sites 1118, 1109, and 1115 (Figs. F3, F4, F5; organized south-north on the northern margin) are consistent with regional seismic stratigraphic units that exist beyond line EW95-1366 (Fig. F2) (Taylor et al., unpubl. data). Tables T3, T4, and T5 present the depth, age, and two-way traveltime of the boundaries between these units, key reflectors described herein, and bio- and magnetostratigraphic markers (Takahashi et al., this volume). The corresponding density, check shot-corrected velocity, and reflection coefficient profiles are presented in Figs. F3, F4, and F5. To aid in cross-site correlations, the lithostratigraphic units, paleodepth, and age-depth curves (with symbols projected into two-way traveltime) are also shown (Shipboard Scientific Party, 1999b, 1999c, 1999d; Takahashi et al., this volume).
Most of the major features of the MCS data in the vicinity of each site are well reproduced by the synthetic seismograms, confirming the accuracy of the density and check shot-corrected velocity profiles. Check-shot results show that the maximum error in the initial velocity profiles was ~3.14% (2.41% if the top of Site 1118, where data was extrapolated from Site 1109, is ignored), with a mean of 0.55% (not including any estimate of error beneath the deepest check shots). Some of this error may be due to shipboard measurements of velocity underestimating in situ velocities resulting from loss of overburden (Hamilton, 1979; Fulthorpe et al., 1989; Urmos et al, 1993). The results also show that the source signal has been well characterized, a key step in obtaining good synthetic seismograms. The source signal dominates for approximately the first 70 ms of the synthetic seismogram (the approximate length of the source signal at peak amplitude). In this interval, it is difficult to extract any geological information from the MCS data. This represents, on average, the top 55 m of each site, or a little over 1 m.y. of depositional history.
The discussion below details the correlation of prominent reflectors between Sites 1118, 1109, and 1115 (Figs. F3, F4, F5). It should be noted when interpreting these results that the synthetic seismograms presented herein are minimum phase (i.e., the first break of the waveform is coincident with the causative spike in the reflection coefficient). For example, the main positive peak of the seafloor reflection in the synthetic seismogram for Site 1118 is centered on ~3100 ms. However, the seafloor is actually at 3083 ms. At 1500 m/s, this is a difference of 12.75 m. Seismic correlations typically pick the main positive peak of a reflector, but this time would only represent the causative event if the waveform were zero phase. This is commonly not the case with seismic reflection data, which are typically minimum phase; therefore, the physical feature causing the event is at the first break. To aid in assigning ages to particular seismic reflectors, the biostratigraphic, magnetostratigraphic, and radioisotope ages of samples in depth have been presented in two-way traveltime between the synthetic and MCS data in Figures F3, F4, and F5. A shift of 18 ms at Sites 1118 and 1115 and 22 ms at Site 1109 (the lag between the first break and the peak amplitude of the first positive spike in the source signal) has been added to the two-way traveltime of the stratigraphic markers, facilitating direct correlation.