A composite section for Site 1123 yielded overlapping records for the upper ~167 mcd. Data overlaps from Holes 1123A, 1123B, and 1123C document complete recovery of the sedimentary sequence to this depth. Two high-resolution data sets proved most useful for correlations at this site: magnetic susceptibility (MS), measured on whole cores on the MST, and spectral reflectance at 550 nm (the center wavelength of the range measured), measured on split cores. Reflectance records provided unambiguous ties for matching nearly all APC cores for the composite section, whereas MS provided reliable ancillary support for the majority of these ties. These two parameters are inversely correlated over the depth range of the splice, where the lithology alternates between white and green nannofossil ooze. The final composite section is illustrated in Figure F24.
Above ~40 mbsf, offsets between mbsf and mcd do not follow a clear trend (Fig. F25). We suggest these offsets are influenced by coring disturbance, including flow-in at the ends and in the middle sections of cores. Flow-in occurred in several APC cores from Holes 1123A and 1123B. The flow-in most likely resulted from slower-than-average APC strokes, which, when combined with ship heave, can introduce spurious material at any interval in a core. For example, in Core 181-1123A-4H, flow-in recovered in Section 4H-5 caused distortion of color reflectance and MS records of >2 m relative to the same stratigraphic intervals in Holes 1123B and 1123C (Fig. F24). Matches between different holes with and without flow-in can result in both negative and positive offsets relative to the mbsf scale. Data were edited to remove disturbed intervals before composite section development, and edited data are shown in Figure F24. Below ~40 mbsf, downhole core offsets in the composite section parallel a model of 10% stretch between the mbsf and mcd depth scales (Fig. F25). Table T12 (also in ASCII format) contains the offsets between the mbsf and mcd scales that result from composite section construction.
The continuous spliced record, based on MS and reflectance data (Fig. F26), extends to 167 mcd. The splice is composed entirely of records from Holes 1123B and 1123C. Coring difficulties, including several intervals of flow-in, complicated the upper section of Hole 1123A, making portions of the hole unsuitable for inclusion in the splice. Additionally, shipboard sampling was concentrated in Hole 1123A; therefore, post-cruise high-resolution sampling following the splice will benefit from splice concentration in Holes 1123B and 1123C. Wherever possible, splice tie points (Table T13, also in ASCII format) were picked at well-defined maxima or minima where the overlaps in data from Holes 1123B and 1123C are correlated. Typically, parameter values differed by less than 10% at tie levels. In all cases, ties were selected so that the spliced record was as free from noise (high-frequency variability) as possible. (Note: Because of a proof-stage correction to the splice, Tables T12 and T13 are correct, but Figures F25-F29 have not been altered.)
Several tephras were recovered in only one of three holes, even in intervals where the major shapes of sedimentary cycles correlate between holes. Core 181-1123C-11H contains a particularly prominent example, where a 20-cm-thick tephra layer occurred near the base of the core (Fig. F26). Core 181-1123B-12H clearly spans other features recovered in Cores 181-1123C-11H and 181-1123C-12H (Fig. F24), though this thick tephra is missing. We did not include this tephra in the splice because (1) it was recovered in only one hole, and (2) it represents an "instantaneous" event that does not reflect the dominant pattern of sedimentation through this interval. If desired, one may add this (and other) features into the splice using the information in Tables T12 and T13.
Overlap between Holes 1123B and 1123C was deliberately attempted deeper in the section. Hole 1123C was APC cored to refusal, washed down to approximately the base of Hole 1123B (~484 mbsf), then XCB cored to a depth of ~625 mbsf. The purpose of this drilling plan was to continue drilling deeper than the base of Hole 1123B, while recovering an overlapping section between the two holes. Magnetic susceptibility data document ~6 m of overlap between Cores 181-1123B-51X and 52X and Cores 181-1123C-18X and 19X (Fig. F27). The overlap illustrated here agrees with independent paleomagnetic results, which show a matching reversal at the boundary between paleomagnetic Chrons C5Br and C5Cn.1n in both holes (Fig. F27).
Recovery (meters of core relative to meters advanced by the drill string) in XCB Cores 181-1123B-23X, 26X, 41X through 50X, and Core 181-1123C-20X was typically poor. However, through this interval, a complete sequence of paleomagnetic reversals was documented, a surprising result for poorly recovered intervals (see "Paleomagnetism"). A comparison between core and log data supports the view that a more complete section is recovered than indicated by "core recovery" percentages. Rather than losing material at either end of an XCB core or from a continuous interval in the middle, shipboard data are consistent with the hypothesis that loss of material between XCB "biscuits" is distributed throughout the core, resulting in a compressed recovered section. To test whether recovery of the sedimentary signal at decimeter-meter scale is higher than indicated by meters of core retrieved, we "stretched" the core data for several XCB cores by linearly interpolating between fixed end-points for each core. The end-points in this scheme are (1) the top of the core fixed at its assigned mbsf depth, and (2) the bottom of the core fixed at 9.5 m below the top. The resulting "stretched" section compares well to downhole log magnetic susceptibility records from Hole 1123B (a continuous record of MS), with few cycles missing from the stretched core data (see Fig. F22 for a core-log MS comparison including stretched paleomagnetic records). Based on this comparison, we suggest that the stretched core data represent a reasonable approximation of the real compositional variations (via MS and reflectance) through these intervals. To translate between mbsf and stretched mbsf, or mbsf and stretched mcd depths, refer to Table T14.
The continuity of the splice, plus good recovery in single-cored XCB sections, provides an excellent opportunity for development of a high-resolution, orbitally tuned time scale at this site. Figure F28 illustrates a preliminary "map" between reflectance percentage (550 nm) and the benthic oxygen isotope record from eastern equatorial Pacific Site 846 for 0-3 Ma (Mix et al., 1995b). In many intervals, the shape and duration of features in the reflectance percentage curve mirror the oxygen isotope curve. Isotope stages are reflected in Site 1123 core data back to the end of the spliced record at ~4.6 Ma. This age model will be refined postcruise, incorporating revised biostratigraphic and paleomagnetic datums, and oxygen isotope stratigraphy.