We compared total P as calculated from the sum of all of the P extraction steps to shipboard total P values (Fig. F1A; Tables T1, T3). The shipboard P concentrations are lower than the sequential extraction P concentrations by ~25% (slope = 0.74 ± 0.06) (Fig. F1B), especially where total P is high (e.g., at ~154.00 mbsf and in the interval 154.2–154.3 mbsf) (Fig. F1A). At low to moderate values, the shipboard total Ba measurements are higher than the barite Ba concentrations (Fig. F2A). However, for samples with high barite concentrations at ~154.1 mbsf and between 154.2 and 154.3 mbsf, the shipboard-measured total Ba values are lower than the barite Ba concentrations (Fig. F2A). Shipboard Ba values are systematically offset from barite Ba concentrations by ~25% (slope = 0.75) (Fig. F2B).
Consistent with many studies of deep-sea sedimentary P (Delaney and Anderson, 1997, 2000; Filippelli and Delaney, 1995, 1996; Faul and Delaney, 2000; Faul et al., 2003), most reactive P in these sediments (96% ± 14%) is in the form of authigenic P (Table T1). The concentrations of P in the other reactive sedimentary components (water soluble, oxide-associated, and organic P) are extremely low and compose <2% of total P (Table T1). Authigenic P determined according to the operationally defined sequential extraction procedure (Ruttenberg 1992) may include fish teeth (biogenic carbonate fluorapatite). Indeed, several fish teeth were found in samples from Section 199-1221C-11X-3 (samples from 24 to 44 cm [~153.65–153.85 mbsf], R. Norris, pers. comm., 2004). However, both Froelich et al. (1982) and Anderson et al. (2001) have demonstrated that fish debris constitutes an insignificant percentage of both the global P budget and typical oceanic authigenic P concentrations. A typical fish tooth contributes ~0.12 µmol P/g sediment to authigenic P (Faul et al., 2003). Authigenic P concentrations for this site are typically two to three orders of magnitude higher (Table T1). Given these P concentrations, fish teeth are an insignificant portion of authigenic P concentrations in this study.
We separate reactive P from detrital P (Fig. F3) because detrital P is not involved in biological cycling of P. In these sediments, it is particularly important because detrital P concentrations are high (as high as 76 µmol P/g or 0.24 wt%) and compose a significant portion of total P from 154.5 to 154.18 mbsf (19% of total P) but drop to almost zero (1.4 µmol P/g or 0.004 wt%) at 154.16 mbsf (e.g., 1% of total P) (Table T1, Fig. F3). Because this detrital P is included in total P but is not biologically reactive, total P would overestimate reactive P in some intervals. Moreover, a change in biologically related P burial but not in detrital P (154.04 mbsf) could be masked when looking only at total P. Reactive P concentrations peak at 300 µmol P/g immediately after the BEE (Lyle, Wilson, Janecek, et al., 2002) then decrease to as low as 7.5 µmol P/g, coincident with the shipboard-measured Mn peak, and then increase 20 cm upcore, showing a secondary peak of ~250 µmol P/g at 154.04 mbsf (Table T1; Fig. F3).
Barite concentrations in the sediments range from 0.05 wt% to as high as 5.4 wt%, peaking simultaneously with reactive P burial (Table T1; Fig. F3). Like P, barite concentrations exhibit a double peak; however, in the case of barite, the relative size of the peaks is reversed (Table T1; Fig. F3). Peak values of barite concentrations (~5.4 wt% or ~31,000 ppm) are higher than any others observed for P/E boundary sediments to date. For example, maximum biogenic Ba concentrations for the North Atlantic (Hole 1051B) are ~2000 ppm and are ~1000 ppm (Bains et al., 2000) for the Southern Ocean (Hole 690B). Examination of all separated barite samples from this core by SEM indicates that our samples resemble modern unaltered biogenic barite and thus are not of diagenetic origin. The barite crystals are the size (2–5 µm) and morphology (euhedral) of modern unaltered biogenic barite and are considerably better preserved than barite from other P/E boundary cores (e.g., Hole 1051B) (Bains et al., 2000).
CaCO3 values as measured by Murphy et al. (this volume) decrease from 74 to 1 wt% in the interval from 154.70 to 153.26 mbsf, remain at <5 wt% until 154.02 mbsf, and then recover to ~80 wt% at 153.85 mbsf (Table T4; Fig. F1). To remove the effect of changes in CaCO3 content on the calculated barite and P concentrations, we determined these concentrations also on a CaCO3-free basis. When calculated on a CaCO3-free basis, most of the variation in reactive P is removed (Table T4, Fig. F4A). CaCO3-free reactive P concentrations hover around 1 wt% (323 µmol P/g) except in the interval from 154.2 to 154.0 mbsf, where CaCO3-free reactive P values drop to as low as 0.02 wt% (Table T4, Fig. F4A). Changes in CaCO3 significantly affect the shape of the record of reactive P, suggesting that P concentration changes are mainly a function of the dilution rate by carbonate, although they do not completely drive the record (Figs. F3, F4). For example, the reactive P concentrations increase (154.06 mbsf) before the CaCO3 concentrations increase (154.0 mbsf). Detrital P, barite, and shipboard Mn concentrations yield similar distribution trends (although they change in magnitude) whether calculated on a CaCO3-free basis or not (Figs. F3, F4; Tables T1, T4).
Organic C to reactive P ratios yield values (0.26 ± 0.34) (Table T4) consistently lower than the Redfield Ratio (117:1) (Anderson and Sarmiento, 1994), indicating preferential loss of C in the water column during deposition or in the sediment at some point after burial. A calculated barite/reactive P ratio for a modern equatorial Pacific core top (JGOFS TT013-69MC, 0.11°N, 139.72°W; 4307 meters below sea level) is 1.4 (molar ratio), based on measured values of Ba associated with barite (Eagle et al., 2003) and reactive P (Faul et al., in press) in this core top. In comparison, the barite/reactive P ratio for the PETM range at ODP Site 1221 ranges from much lower than modern value (as low as 0.05 at 154.14 mbsf) to higher than modern value (2.9 at 154.08 mbsf) (Table T4; Fig. F4). These changing ratios may indicate changes in export productivity relative to nutrient burial during the PETM interval.
Shipboard results show an Mn peak (~12 wt%) (Lyle, Wilson, Janecek, et al., 2002) when CaCO3 values are low (Fig. F3; Tables T3, T4). Visual inspection revealed the presence of small manganese oxide nodules in several samples from the section (R. Norris, pers. comm., 2004).