RESULTS AND DISCUSSION

The method described above yields a set of three synthetic CMP gathers that can be compared with data near Site 1276 (Fig. F9). All three sets of synthetic data reproduce most of the first-order features seen in observed data, including the series of bright horizontal reflections observed around ~7 s and the U reflection and underlying bright reflections toward the base of the stratigraphic section (~7.62–7.75 s). The reflections corresponding to Units 1–4 are similar between the three scenarios, despite the modifications made to the layer thicknesses associated with high velocities in Scenario 3. However, there are differences among the three scenarios for Unit 5. In the synthetic for Scenario 1 (Fig. F9A), a series of bright reflections can be observed in Unit 5 that do not appear to correspond to reflections in the MCS reflection data (e.g., ~7.3, 7.35, 7.39, 7.57, and 7.62 s). These reflections are caused by high-velocity measurements, some of which are from very thin lithologic intervals. In Scenario 2, many of these reflections are no longer present because all high-velocity (>3 km/s) measurements from thin (<1 m) intervals in this unit were removed (Fig. F9B). The complete elimination of some measurements from Unit 5 makes this synthetic the least realistic, but it provides a useful way to consider the effects of high-velocity measurements. Scenario 3 represents a compromise between these two end members, where the thickness associated with high-velocity measurements has been adjusted to better approximate the thickness of corresponding lithologies in the core. In this synthetic, reflections are found at TWTs similar to those in Scenario 1, but they are typically lower in amplitude (Fig. F9C). Although all three sets of synthetics reproduce the first order features, Scenario 3 appears to best represent subtle waveform characteristics observed in coincident reflection data in Unit 5 above the sills (Fig. F9C). Thus, we use this synthetic for the detailed discussion of links between seismic and core data that follows.

Below, we consider three alternative correlations between synthetic-seismogram Scenario 3 and reflection data, as illustrated in Figure F10. These correlations are identical at the base of the hole but differ in the precise linkages made between synthetic and observed reflections for Units 1–4. The largest uncertainty in correlations occurs in this portion of the hole because of the absence of a definitive depth-time link for the top of the hole, as discussed earlier. After considering the Unit 1–4 correlations, we discuss correlations between the synthetic seismogram and reflection data for Unit 5, with particular focus on the U reflection and the two postrift sills at the base of Site 1276.

Lithologic Units 1–4

Lithologic Unit 1 is dominantly represented by mudrock, and the associated velocity and density measurements are correspondingly homogeneous (Fig. F3). As a result, primarily low-amplitude reflections appear in the synthetic seismogram for this interval (Fig. F9). However, one bright reflection (7.02 s in Fig. F10) is generated in synthetic seismograms near the base of lithologic Unit 1; this results from carbonate-cemented claystones and sandstones that are associated with velocities as high as 3.659 km/s between 841.3 and 852.5 mbsf. In contrast to Unit 1, laboratory measurements of density and velocity taken from Units 2–4, particularly Unit 2, show dramatic changes in velocity (1.961–5.545 km/s) and density (1.95–2.647 g/cm³) (Fig. F3) that produce a series of high-amplitude reflections in the synthetic seismogram (7.04–7.23 s) (Fig. F9). These variations in physical properties and the resulting bright reflections in the synthetic seismogram are generated by the interlayering of high-velocity carbonate-cemented claystones and sandstones that are found in turbidites (20–140 cm thick) with low-velocity marlstone and claystone. Strong velocity contrasts are also found, to a lesser extent, in Units 3 and 4. The deepest reflection in this bright series is located at ~7.23 s in the synthetic seismogram and appears to arise from a negative impedance contrast at the base of Unit 4.

Because of the nature of the velocity and density measurements used to generate the synthetic seismograms and the lack of a definitive depth-time tie (normally supplied by a checkshot survey), we can envisage three different possible correlations between the reflections in the synthetic seismogram produced by Units 1–4 and the coincident seismic reflection data:

1. Interpretation 1: The Unit 1/2 boundary correlates with Reflection A^U1 (~7.02 s), and the Unit 4/5 boundary correlates to the apparent seismic unconformity at ~7.19 s in reflection data; this correlation of the Unit 4/5 boundary was originally proposed in the Leg 210 Initial Reports volume (Shipboard Scientific Party, 2004) (Fig. F10A).
2. Interpretation 2: The Unit 1/2 boundary corresponds to Reflection A^U2 (~6.98 s), and the Unit 4/5 boundary corresponds to the reflection at ~7.16 s (Fig. F10B). This correlation of the Unit 1/2 boundary was originally proposed in the Leg 210 Initial Reports volume (Shipboard Scientific Party, 2004).
3. Interpretation 3: The Unit 1/2 boundary corresponds to the reflection beneath Reflection A^U1 (~7.04 s), and the Unit 4/5 boundary corresponds to the reflection at ~7.22 s (Fig. F10C).

To determine which of these three interpretations is most likely, we compare reflection characteristics between the synthetic seismogram and MCS data and consider the implications for velocity structure implied by each interpretation. For Interpretation 1, the match between reflection characteristics in the synthetic seismogram and MCS data (Fig. F10A) is good overall. Reflection A^U1 is lower in amplitude in the MCS data than in the synthetic seismogram. The brightest reflections in the synthetic seismogram, which are associated with Unit 2, correlate with the brightest reflections in the seismic reflection data. Below this, a series of reflections with slightly lower amplitudes are observed in both the synthetic seismogram and the MCS data, and the Unit 4/5 boundary is matched to the deepest reflection in the bright package between 7.04 and 7.19 in the reflection data.

Matches between reflections in the synthetic seismogram and MCS reflection data for Interpretation 2 are generally poor for Units 1, 2, and the upper portion of Unit 3 but are very good for Units 3 and 4 (Fig. F10B). Although the amplitudes of the Unit 1/2 boundary reflection in the synthetic seismogram and Reflection A^U2 in the MCS data are more similar than the correlation in Interpretation 1 (Fig. F10A), the underlying reflections associated with Unit 2 in the synthetic seismogram are much stronger than the corresponding reflections in the seismic data.

Finally, correlations between reflections in the synthetic seismogram and reflection data for Interpretation 3 are relatively good for Units 1, 2, and 3 but are poor for Unit 4 (Fig. F10C). The strong reflections arising in the synthetic seismogram from the Unit 1/2 boundary and from Unit 2 correspond to strong reflections in the seismic reflection data. However, the reflections associated with the base of Unit 4 in the synthetic seismogram have higher amplitudes than the corresponding reflections in the MCS data.

Each of these interpretations also carries implications for the velocity structure between the seafloor and the lithologic Unit 1/2 boundary (~865 mbsf) and between this boundary and the underlying upper sill (the U reflection at 1613 mbsf, see "Lithologic Unit 5, Including Sills and Undercompacted Sediments"). We can compare the velocity structure implied by each of these interpretations with the velocities estimated by modeling MCS reflection data described above in "Estimating Reflection-Time Depth of the Top of Cored Section" (Fig. F8). We chose to compare average velocities above and below 865 mbsf, the Unit 1/2 boundary (Fig. F10). Modeling of MCS reflection data yields an average velocity of 1839 m/s between the seafloor (6.04 s) and 865 mbsf and of 2508 m/s between this level and the U reflection. The following velocities are implied for each of these two intervals by each of the interpretations presented in Figure F10: Interpretation 1: 1765 m/s and 2473 m/s; Interpretation 2: 1840 m/s and 2319 m/s; Interpretation 3: 1648 m/s and 2579 m/s (Table T1; Fig. F11). If the other synthetic seismograms in Figure F9 (Scenarios 1 and 2) were used for these interpretations, the implied velocities would only vary by 15 m/s. Interpretation 1 predicts velocities above and below 865 mbsf that are close to those estimated by modeling MCS reflection data, with differences of ~74 m/s above and ~35 m/s below this level. Interpretation 2 predicts average velocities above 865 mbsf that are very close to those from MCS reflection modeling (within ~1 m/s) but much lower (~189 m/s) below. Finally, Interpretation 3 predicts much lower velocities (~191 m/s) above 865 mbsf than those estimated by modeling of MCS reflection data but velocities that are somewhat higher (~71 m/s) than modeled velocities below this level (Table T1; Fig. F11).

Based on both reflection characteristics and implied velocity structure, we favor Interpretation 1. This correlation provides an acceptable fit between reflection characteristics for lithologic Units 1–4 and implies a realistic velocity structure throughout the sedimentary section. Conversely, for Interpretation 2, the mismatch between reflection characteristics for Unit 2 and the unrealistically low velocities implied below 865 mbsf suggest that this interpretation is not viable. Finally, the mismatch between reflections at the base of Unit 4 and the low velocities implied for the interval above 865 mbsf argue against Interpretation 3. The results of adjusting the synthetic seismogram up to match the observed data, consistent with Interpretation 1, are shown in Figure F12.

Tucholke and Sibuet (this volume) have identified two possible reflections (A^U1 and A^U2) that might correspond to Horizon A^U observed farther south in the main North Atlantic basin (Fig. F2). The A^U reflection there correlates with a hiatus near Eocene/Oligocene boundary and is commonly expressed as a seismic unconformity (Miller and Tucholke, 1983; Tucholke and Mountain, 1979). Thus the reflection has been correlated to the onset of strong abyssal circulation in the North Atlantic Ocean (Miller and Tucholke, 1983). The association of the A^U1 reflection with the Unit 1/2 boundary described above for Interpretation 1 implies that this reflection does not correspond to Horizon A^U. Wood et al. (submitted [N1]) conducted a detailed study of the nannofossil biostratigraphy of the interval around the Unit 1/2 boundary and found that it corresponded to a hiatus of between 1.2 and 6.9 m.y. in the middle Eocene (47.3–40.4 Ma). This timing is some ~6 m.y. older than predicted for the onset of circulation in the North Atlantic Ocean. Furthermore, seismic sequence characteristics associated with the flow of abyssal currents, particularly sediment waves, are developed above the A^U2 reflection. Thus Tucholke and Sibuet (this volume) suggest that the A^U2 reflection at ~6.96 s is more likely to correlate with Horizon A^U farther to the south in the western North Atlantic. In Interpretation 1, the A^U2 reflection lies close to the top of the cored interval beginning at 800 mbsf, and according to shipboard biostratigraphy, this level dates to the latest Eocene–earliest Oligocene (Shipboard Scientific Party, 2004).

The correlation between the reflection arising from the Unit 4/5 boundary and the reflection at ~7.19 s in the seismic data agrees well with shipboard biostratigraphy. This reflection truncates underlying reflections over much of the transition zone and thus appears to represent an unconformity (Fig. F2). Biostratigraphic data suggest a period of either very slow sedimentation or a hiatus at about this level (Campanian–Turonian), although this is not well constrained because some cores at this boundary were barren of microfossils (Shipboard Scientific Party, 2004).

Lithologic Unit 5, Including Sills and Undercompacted Sediments

In contrast to the bright reflections predicted for lithologies in Units 2–4, much lower amplitude reflections are observed in the synthetic seismogram associated with Unit 5 (7.24–7.68 s, Fig. F10), which is dominated by mudrock. Velocity and density in this interval show much less variation than in the overlying lithologic units (Shipboard Scientific Party, 2004) (Fig. F3), and weak reflections found in this interval are caused by velocity and density variations associated with intermittent turbidites. According to our interpretation, this section of low reflectivity likely corresponds to the low-amplitude interval in the coincident seismic reflection data spanning 7.19–7.62 s (Fig. F10A).

The velocity and density structure at the base of Site 1276 is considerably more complicated than in the remainder of the hole, with variations of more than 4.0 km/s in velocity and 2 g/cm³ in density over tens of meters in depth (Fig. F3). These changes in core properties are associated with the presence of both diabase sills (tops at 1612.7 and 1719.2 mbsf) and a thin, ~17-m-thick interval of undercompacted sediments (1693–1710 mbsf). The large changes in velocity and density associated with these features result in high-amplitude reflections in the synthetic seismogram. A strong reflection beginning at 7.68 s in the synthetic is caused by a combination of the positive and negative impedance contrasts from the top and base of the upper sill, respectively (Fig. F10A). At 7.76 s, an even stronger reflection is predicted in the synthetic seismogram, caused by a positive impedance contrast between the base of the undercompacted sediments and the top of the lower sill. Overall, the pattern of synthetic reflections closely resembles the U reflection and the underlying reflections in the MCS reflection data (Fig. F10), implying that the U reflection at Site 1276 is created by the upper 10-m-thick diabase sill.

For all of the synthetic-seismic correlations discussed above, the difference in TWT for reflections in the synthetic seismogram and the MCS data increases progressively with depth. For example, in Scenario 3, Interpretation 1, the difference increases from 21 ms at the level of the A^U1 reflection to 62 ms at the U reflection, yielding a total change of 41 ms (Fig. F10A). This indicates that the average velocity used to create the synthetic seismogram is too low, and this low velocity is likely is an artifact of the velocity function derived from laboratory measurements. As mentioned earlier, samples from cores brought to the surface from depth undergo decompaction and microfracturing, causing laboratory measurements of velocity and density taken at atmospheric pressure to have lower values than they would have in situ. Additionally, we used vertical velocities to create our synthetic seismograms. Sediments at Site 1276 exhibit increasing transverse anisotropy with depth, so that horizontal velocities were greater than vertical velocities (Shipboard Scientific Party, 2004). Thus, the large offsets recorded in this MCS reflection data set would have a contribution from comparatively high horizontal velocities. All of these effects suggest that laboratory measurements of vertical velocity provide a slow-velocity end-member for depth-time conversion. To resolve the difference in TWT for the interval noted above (assuming the correlation in Interpretation 1, Fig. F10A), an ~10%–11% increase in velocity would be required for synthetic seismogram Scenarios 2 and 3 (where high velocities were removed or the thickness associated with them were modified), and a 3% increase would be needed for Scenario 1. Such shifts are consistent with the effects described above.