Strontium isotope stratigraphy has the potential to provide important age information, especially in sections where deep burial, poor core recovery, and/or lack of age-diagnostic microfossils have confounded stratigraphic interpretation. However, before Sr isotope stratigraphy and the ages predicted from it can be accepted, the extent of diagenetic alteration must be assessed.
Diagenetic alteration is known to affect original Sr isotope values of carbonates as a result of exchange with pore fluids during recrystallization. Minor alteration in sediments with a significant clay component that often possess a highly radiogenic Sr isotope signature can have a relatively large effect on the final 87Sr/86Sr ratio (e.g., Bralower et al., 1997; Mearon et al., 2003). For chalks and limestones that are nearly pure biogenic carbonate, major recrystallization usually has a relatively small effect on Sr isotope ratios, as the system is well buffered to contemporary seawater values. However, these small effects can translate into significant age errors (i.e., 1-2 m.y.) in parts of the section such as the Campanian where 87Sr/86Sr values change slowly through time. Moreover, even minor amounts of clay can impart a radiogenic signature during recrystallization. When analyses of these sediments are made on components such as foraminifers, the effect of recrystallization can be gauged by scanning electron microscopy (Bralower et al., 1997); with bulk carbonate analyses such discrimination is impossible, and diagenetic concerns can only be addressed indirectly.
One indicator of potential diagenetic alteration of Sr isotope ratios is the occurrence of visible residue, often composed of clay, after carbonate samples are dissolved. An alternate method of gauging the effect of diagenesis on Sr isotopic values is by comparison of predicted ages with those obtained from biostratigraphy. Consistent offsets between biostratigraphy and Sr isotope stratigraphy can arise from differences in the way absolute ages are assigned to biostratigraphic datums and the seawater Sr isotopic curve; alternatively, such offsets may be an indication of diagenetic alteration. Because more indurated samples are more likely to have experienced minor alteration of Sr ratios during recrystallization, analysis of lithology can explain deviation of results from those expected from stratigraphy. In Hole 1183A, where samples span a considerable depth range (Table T1) and lithology gradually changes downward from chalk to limestone over the course of this interval, the scatter increases downsection as does the deviation from Sr isotope values expected from biostratigraphy (Figs. F1, F3).
Ages of nannofossil datums used here were calculated using the Gradstein et al. (1994) timescale, assuming constant sedimentation rates in key sections in the pre-Campanian and combining this assumption with ages of magnetic chrons in the Campanian and Maastrichtian (e.g., Bralower, Premoli Silva, Malone, et al., 2002). Ages of Sr isotope values used to derive the LOWESS curve (McArthur et al., 2001) are taken directly from publications and calibrated to the Gradstein et al. (1994) timescale. The biostratigraphy on which the Cretaceous part of the LOWESS curve is based includes nannofossils (Bralower et al., 1997; McArthur et al., 1993) and a range of macrofossils (ammonites, belemnites, and bivalves [e.g., McArthur et al., 1993, 1994; Jones et al., 1994]). Part of the LOWESS curve is directly calibrated with bentonite radiometric ages (e.g., McArthur et al., 1994), but in other cases ages are derived from stage boundary identifications in the sections studied (McArthur et al., 1993) and from zonal boundary ages determined assuming constant sedimentation rates (e.g., Bralower et al., 1997) and constant ammonite zonal durations (e.g., Jones et al., 1994).
Finally, a limitation exists in the mid-Cretaceous where an Sr isotope value on its own might indicate two or three different ages. This uncertainty exists because the seawater Sr isotopic curve changes direction several times (Fig. F3). In this interval, paleontological or other stratigraphic data are necessary to constrain the possible ages to a small portion of the seawater curve and allow a unique age to be predicted from Sr isotope stratigraphy. In this case, Sr isotope data can then be used to provide a more precise age estimate and test for minor errors in biostratigraphic ages, but circularity becomes a problem in addressing stage-level questions.
In Hole 1183A, predicted ages from Sr isotope values agree moderately well with ages from nannofossil biostratigraphy (Fig. F4). In the Campanian-Maastrichtian section, predicted ages are similar to or slightly (~2 m.y.) younger than biostratigraphic ages. Part of the offset could result from minor variations in 87Sr/86Sr that are a result of recrystallization. These minor offsets are also possibly a result of uncertainty in age estimates derived from nannofossil biostratigraphy, combined with differences in the way absolute ages are calculated from nannofossil biostratigraphy and Sr isotope stratigraphy. Similar offsets between predicted and biostratigraphic ages also occur at Ocean Drilling Program Sites 738, 1049, 1050, and 1052, suggesting that the ages of datums used to calibrate LOWESS are too low or those estimated for the nannofossil datums are too high. A possible explanation for this is that the LOWESS curve in this interval is hinged on boreal and western interior macrofossil (e.g., McArthur et al., 1993, 1994) and microfossil (e.g., Burnett [1990], used in McArthur et al. [1993]) zones, which may be systematically offset from stage boundary definitions in the deep-sea sites based on microfossil zones. For example, the one common nannofossil datum used by Burnett (1990) that is found routinely in the deep-sea sites, the first occurrence (FO) of Aspidolithus parcus, lies in the middle lower Campanian in the former study but close to the Santonian/Campanian boundary in Bralower, Premoli Silva, Malone, et al. (2002) and other publications. This difference would lead to the observed offset whereby predicted ages from LOWESS are younger than those from nannofossil biostratigraphy.
More significant disparities exist in the section between 1090 and 1130 mbsf, where nannofossil biostratigraphy suggests at least three unconformities (Sikora and Bergen, submitted [N1]). The first occurs at ~1090 mbsf and spans ~7 m.y. from the early Campanian (82.5 Ma) to the latest Turonian (89.3 Ma); the second, at ~1108 mbsf, spans ~11 m.y. from the late Albian to the latest Aptian (101.7-112.6 Ma); and the third, near the sediment/basement contact at ~1130.37 mbsf, spans ~2 m.y. from the early Aptian (119.0 Ma) to the latest Barremian (121.1 Ma). Sr isotope values suggest breaks in sedimentation at two of these levels (likely hiatuses of ~10 m.y. between ~1079 [74.38 Ma] and 1090 mbsf [84.49 Ma] and ~25 m.y. at ~1112 mbsf [85.2-109.9 Ma]). On the other hand, Sr isotope data suggest a relatively thick Santonian section at 1091-1111 mbsf and an apparently continuous lower Albian to upper lower Aptian section at 1112-1130 mbsf and do not confirm the lowest of the three proposed hiatuses.
The lowermost hiatus indicated by nannofossil biostratigraphy in Hole 1183A may be an artifact of the truncation of ranges near the sediment/basement contact. The FOs of Hayesites irregularis and Eprolithus floralis lie immediately above basement and may not reflect true datum levels. The FO of Rhagodiscus achlyostaurion, just over 1 m above basement, may also be unreliable because of disparities between the taxonomies of different workers who have applied this datum (e.g., Erba, 1991; Bralower et al., 1993; Bergen, 1994; Tremolada, 2002). Thus, Sr isotope stratigraphy may provide a more precise age interpretation of the lowermost 18 m of section in Hole 1183A.
The hiatus at ~1108 mbsf indicated by nannofossil biostratigraphy is based on several well-established and reliable datums including the FOs of Prediscosphaera columnata and Eiffellithus turriseiffelii. These datums are inconsistent with Sr isotope values of three samples between 1108 and 1111 mbsf (Fig. F4). Two of these samples contained dark residues after dissolution, possibly reflecting a contribution of radiogenic Sr from clays that increased their Sr isotope values above original seawater levels. The 1090-mbsf unconformity corresponds to the FOs of Marthasterites furcatus, Micula staurophora, Aspidolithus parcus parcus, and Aspidolithus parcus constrictus. The FOs of the latter two datums fit well with Sr isotope stratigraphy, but the FOs of M. furcatus and M. staurophora are inconsistent with Sr isotope values of the three samples between 1098 and 1101 mbsf. These samples contain higher Sr isotope values than predicted by nannofossil biostratigraphy, and two of them had a dark residue after dissolution. One, however, did not, and the other sample with similar residues matches expected values; thus, it is difficult to dismiss these values as diagenetic artifacts. An alternative explanation is that the disparity arises from differences in the stratigraphic position of M. staurophora in different timescales and the paucity of specimens of M. furcatus toward the base of its range in Hole 1183A.
In the Campanian-Maastrichtian interval of Hole 1186A, predicted ages from Sr isotope values are as much as ~2.5 m.y. younger than ages from nannofossil biostratigraphy, about the same magnitude as disparities in this part of the section in Hole 1183A (Figs. F4, F5). These disparities are likely a result of minor recrystallization combined with differences in the way that absolute ages are calibrated to nannofossil datums and the LOWESS curve. The disparity between ages predicted by Sr isotope stratigraphy and biostratigraphy at ~813 mbsf likely results from the truncation of the range of Lithraphidites quadratus at the Cretaceous/Paleocene unconformity (Mahoney, Fitton, Wallace, et al., 2001).