The mcd scale is expanded relative to the mbsf scale in all holes at Site 1151 by varying amounts (Fig. F8). Hole 1151A showed the greatest expansion, increasing by 7.0% from the original mbsf length over the upper 93.3 m cored. The mcd scale was 5.9% longer than the mbsf scale for Hole 1151C, and 3.7% longer in Hole 1151D. Differences in drill types between Hole 1151A and Holes 1151C and 1151D, as well as difficulties in the measurements of mbsf depths as described below, are likely causes of the differences in scales. These results are consistent with the mcd scales developed at other sites, which reported lengthening of the mcd scale relative to the mbsf scale by about 10% (e.g., Alexandrovich and Hays, 1989; Farrell and Janecek, 1991; Hagelberg et al., 1995; MacKillop et al., 1995; Lyle, Koizumi, Richter, et al., 1997; Acton et al., 2001).
Drilling and recovery procedures are potential sources of the differences in depths to lithologic boundaries observed among different holes (Ruddiman et al., 1987; Ruddiman, Kidd, Thomas, et al., 1987; Ruddiman, Sarnthein, Baldauf, et al., 1988; Alexandrovich and Hays, 1989; Murray and Prell, 1991; Farrell and Janecek, 1991; Harris et al., 1995; Lyle, Koizumi, Richter, et al., 1997). For example, the release of overburden pressure upon recovery of a core causes expansion of the sediments. Heave and drift of the ship may result in inaccurate estimation of the drill string length, as well as result in repeat coring of some sections of the hole, causing further error in mbsf depth measurements. The cohesiveness of the sediments, affected by such factors as the water content, may also contribute to the expansion of a core. Small-scale faulting or slumping within a hole during coring may result in further core expansion and recovery that exceeds 100%.
It should be kept in mind that the composite depth scale is a first-order correlation in which the length scales of individual cores are not expanded or contracted. Because relative expansion occurs between coeval intervals, correlative features may be misaligned by a few tens of centimeters or less. For the log magnetic intensity data in Figure F9, for example, the distinct peaks at ~69 mcd in Holes 1151C and 1151D are tied, but this causes an observable offset in the data immediately above and below this correlation. Figure F10 shows another example in the magnetic susceptibility data, with data in Hole 1151D between 8.5 and 12.8 mcd expanded relative to the coeval interval in Hole 1151C. Although it is possible this mismatch reflects lateral variation in sedimentation rates, it is more likely a result of relative core expansion given the short distance between holes. This trade-off in matching some features at the expense of the misalignment of other features within the coeval interval is repeated throughout the data sets. By relaxing the constraint that a core's length scale be fixed, higher-order correlations can be made that would permit the alignment of features on the order of a centimeter or better. Studies that require correlation of features at this scale can be achieved using Analyseries or methods such as those outlined by Martinson et al. (1982). The first-order correlations presented here yield an ~30% improvement in the correlation coefficients between holes and establish an important depth template for building higher-order scales.