Poor reproducibility or discrepancies in phases dissolved by different procedures are unlikely to account for the systematic offset observed in shipboard to shore-based P and barite values. Neither the sequential extraction reproducibility (~4%) nor the shipboard reproducibility (10%–15%) (Quintin et al., 2002) can account for a systematic difference. It is unlikely that the difference could be accounted for in terms of detrital P (or some other phase) not being extracted in the shipboard procedure, causing the shipboard total P values to be systematically lower than shore-based total P values. Detrital P values from the sequential extraction are high in the interval from 154.1 to 154.5 mbsf (Figs. F3, F4), but the largest difference between the sequential extraction total P and the shipboard total P exists in the interval between 154.2 and 154.3 mbsf (Fig. F1). In contrast, some of the offset between shipboard Ba and shore-based barite could be explained by the fact that the shipboard procedure includes detrital Ba in addition to barite. Additionally, the fluxing may not have quantitatively dissolved all of the barite where the barite content was substantial. Indeed, the cross-plot between the shipboard total Ba and the barite extraction–based Ba indicates a residual amount of detrital Ba of ~3000 ppm, (Fig. F2B). However, the fact that the slope between shipboard and shore-based values is 25% lower than 1:1 cannot be accounted for by the addition of detrital Ba. We believe that the discrepancy between shipboard and shore-based measurements can only be accounted for by problems with calibration of the shipboard standards for high P and Ba concentrations. Indeed, Quintin et al. (2002) point out that it was a challenge to find standard reference materials to cover the range of elemental concentrations encountered during Leg 199.
As a preliminary substitute for mass accumulation rates and to account for dilution effects, we determine concentrations on a CaCO3-free basis. The interval of low CaCO3 concentrations (<5 wt%; 154.3–154.0 mbsf) (Fig. F3) indicates either dissolution (most likely) or nondeposition of CaCO3, the predominant biogenic sedimentary component. The sharp and simultaneous drop-offs of reactive P, detrital P, and barite at ~154.2 mbsf (Figs. F3, F4) may indicate that an interval of extremely low deposition of all components may be present in this section, although it is impossible to determine definitively if indeed the deposition rates in this interval were low without a determination of precise sedimentation rates (on the timescale of the observed sedimentary changes). If, however, the sedimentation rates throughout the section were constant, the accumulation rate records would show the same trends in reactive P and barite abundances as those presented here. In any case, interpretations made on a CaCO3-free basis should be a good first-order substitute for mass accumulation rate trends. Moreover, the relative magnitudes of the different proxies at the same depth horizons will remain the same regardless of absolute values. That is, the organic C to reactive P ratios as well as the barite to reactive P ratios and any other comparisons of the different proxies at the same depth horizons will be unaffected by sedimentation rate or dilution effects by carbonate. Thus, we believe the lack of accumulation rate data does not preclude preliminary paleoceanographic interpretation of our results.
The presence of Mn oxide nodules, the high barite concentrations, and the similarity of the barite to modern unaltered barite, along with low organic C in the sediments, imply that there were oxygenated conditions in the sediments during the PETM at this site in the eastern equatorial Pacific and that at some point, either during the PETM or postdepositionally, organic C was lost from the section. The presence of solid MnO2 (nodules) forming at the sediment/water interface indicates oxygenated bottom water conditions (Calvert and Pedersen, 1993). The Mn peak may be attributed to diagenetic mobilization of Mn in reducing conditions followed by precipitation upon oxidation. This, however, is not consistent with the occurrence of the nodules. Sulfate concentrations in the pore waters at present throughout the core are similar to seawater value (28 mM) (Lyle, Wilson, Janecek, et al., 2002). Both the morphology and the sulfur isotopes in the barite indicate that the barite is not diagenetically formed in reducing pore fluids. Thus, there is no indication that sulfate was deficient enough in these sediments to dissolve barite at the time of the PETM or at any time since, supporting the use of barite at this site as an export production proxy.
Because sink-switching retains the reactive P initially delivered to the sediment predominantly with organic C, reactive P may be used as an indicator of P burial and for initial organic C deposition even at sites where postdepositional oxidation removed the organic matter (Anderson et al., 2001; Anderson and Delaney, in press, Faul et al., in press). Reactive P concentrations are high (~1 wt%) and relatively constant on a CaCO3-free basis (except from 154.2 to 154.0 mbsf) (Fig. F4). In contrast, the organic C values are very low (average = 0.06 wt% calculated on a CaCO3-free basis). Organic C to reactive P ratios (0.02–2.2) (Table T4) are much lower than the Redfield Ratio (117) and thus indicate preferential loss of organic C relative to reactive P at this site. These combined observations along with high barite burial suggest that much of the organic C may have been oxidized and/or captured in forms that are difficult to measure. The biogenic silica records also show no changes in biogenic opal deposition across the boundary when calculated on a CaCO3-free basis (Murphy et al., this volume), and if diatoms are responsible for a large fraction of organic matter flux or if their abundance is proportional to organic matter flux, this also indicates that no significant change in organic C delivery to the sediments across the PETM has occurred, assuming no changes in preservation over the sampled time. Reactive P trends are inversely correlated with CaCO3 content throughout much of the interval. Presumably, as indicated by our carbonate-free calculations, reactive P deposition remained constant (except for the interval from 154.2 to 154.0 mbsf) and the CaCO3 diluted or enhanced the reactive P signal. This scenario is consistent with little change in P and presumably organic C burial across the section.
In the modern ocean, productivity tracers in the sediment do not always co-vary, even under areas of high production. In oxygenated sediments under the high export productivity region of the equatorial Pacific, organic C contents of the sediments are low whereas barite contents are high (Paytan et al., 1996). This is the case for these PETM sediments. High barite concentrations may be interpreted as indicating high export productivity at this site. This high export production scenario may be consistent with low organic C burial if organic C is lost postdepositionally and/or if organic C is efficiently regenerated in the deep ocean before burial. The barite/reactive P ratio can act as a measure of the efficiency of internal nutrient cycling (Nilsen et al., 2003). A high barite Ba/reactive P ratio may indicate a time of higher export production associated with a low net nutrient burial, whereas a low barite Ba/reactive P ratio may indicate a time of lower export production and high net nutrient burial. The records of barite and reactive phosphorus during the PETM show variation in barite/reactive P ratios both above and below the modern measured core top equatorial Pacific value of 1.4 (Table T4; Fig. F4). Lower than modern barite Ba to reactive P ratios before the BEE and immediately afterward indicate efficient nutrient burial compared to export production relative to today's equatorial Pacific sites. In contrast, a high barite Ba to reactive P ratio (2.9) at 154.08 mbsf and subsequent values similar to those today (but higher than the pre-BEE values) may indicate higher export production, as evidenced by elevated barite, coupled with more efficient nutrient recycling, as evidenced by unchanging nutrient (P) burial. Efficient nutrient regeneration (e.g., reactive P regeneration preferentially to organic C) could drive and sustain high primary productivity while increasing net organic C burial, but the low organic C to reactive P ratios found in this core do not support preferential reactive P regeneration, unless C was lost postdepositionally.
Changes in export production and nutrient-recycling efficiency that are decoupled from each other could reconcile differing signals observed in reactive P, organic C, and barite for this site in the eastern equatorial Pacific. Constant and relatively low organic C and biogenic opal sedimentation, along with relatively constant reactive P, indicate little change in the organic C burial and possibly organic C delivery to the site. However, both export productivity, as evidenced by high barite concentrations, and nutrient burial efficiency, as evidenced by high reactive P concentrations, were higher before the BEE than today. If reactive P and organic C burial are linked, this could signify a situation where organic C could have been sequestered in the sediments, therefore drawing down CO2. The low organic C to reactive P ratios in the sediment at present most likely result from postdepositional loss of organic C.
After the interval of presumably low sedimentation (beginning at ~154.18 mbsf), both nutrient burial (P) and export productivity (barite) increase but export productivity increases more than nutrient burial, as evidenced by higher average barite Ba/reactive P ratios, indicating that nutrient recycling efficiency has increased approximately to the level it is at today. Two specific scenarios may be envisioned for this time interval. The first requires high export of organic matter below the euphotic zone, which results in barite formation and burial while the reactive P is recycled within the euphotic zone efficiently to sustain productivity. This would require decoupling of organic C from reactive P in the water column. Alternatively, reactive P is recycled in the deep ocean and/or at the sediment/water interface along with organic C, resulting in lower burial of reactive P and organic C. The latter option will result in C sequestration only on timescales of ocean water turnover (thousands of years) and not on longer timescales. The former option results in more effective long-term C sequestration. Thus, high export flux (as indicated by a peak in barite burial) and efficient P recycling in the euphotic zone (as indicated by constant reactive P burial), presumably decoupled from C recycling, would result in C being delivered and buried in the sediment (as suggested by the barite peak). Organic C would then be lost to oxidation at some time after deposition (consistent with the very low C to P ratios and Mn peak).
As indicated above, changes in the relative rates of export production, nutrient and organic C recycling efficiency, and organic C and nutrient burial along with postdepositional loss of organic C may explain the observed fluctuations in the sedimentary record at this site across the P/E boundary. Because our data are limited to one location, we cannot globally extrapolate our conclusions for the nature of either export productivity or nutrient burial. However, efficient nutrient and organic C recycling may explain widespread indications of increases of productivity in the surface waters and high export production (e.g., Sr/Ca ratios in coccolith carbonate at Site 690 [Stoll and Bains, 2003]). When postdepositional oxidation of organic C is invoked, this is also consistent with low organic C burial for the PETM. If, indeed, C was originally buried in the sediment and was subsequently lost, as may be suggested by the high (albeit constant) reactive P, the Mn peak, and the high barite, then longer-term C sequestration may have taken place. However, until further direct constraints regarding the timing of C regeneration are available it is not possible to distinguish between the two scenarios postulated above. Regardless, these mechanisms provide an explanation for high barite concentrations despite low organic C deposition without necessarily invoking Ba release from methane clathrates (Dickens et al., 2003).