IODP Proceedings    Volume contents     Search


Discussion: high-resolution CaCO3 profile from Site U1338

Calibrating the XRF measurements with discrete CaCO3 data makes it possible to examine high-resolution variability in the carbon cycle since the late early Miocene. Although the crustal age at Site U1338 is ~18 Ma, variable sedimentation at the base of the core, including additions of hydrothermal sediment as well as unresolved issues about the core splice, make us hesitant to interpret the record below 16 Ma. Above that level, sedimentation responds to a large-scale pattern common to most drill sites in the equatorial Pacific. Sedimentation increases as the site is carried to the Equator by Pacific plate movement and then decreases as the site moves away from the Equator. The crust at Site U1338 formed at ~1.5°S and crossed the Equator at ~11.5 Ma, where sedimentation rates were ~30 m/m.y. (see the “Site U1338” chapter [Expedition 320/321 Scientists, 2010b]). Sedimentation rates slowed dramatically to ~12 m/m.y. at ~4.65 Ma, as Site U1338 passed 1.5°N. These changes in sedimentation make the resolution of CaCO3 sampling above 4.65 Ma ~2 k.y. between samples, whereas the deeper record has a time resolution of <1 k.y. Responses to orbital forcing should be easily resolved, provided that the high-resolution variability in the profile reflects changes in sediment chemistry and not measurement variability.

Is the high-resolution XRF CaCO3 estimate signal or noise?

High-resolution discrete CaCO3 measurements were taken every 10 cm between 180 and 230 cmcd versus a spacing of roughly 75 cm in other intervals. Even in the highly sampled interval, XRF measurements are a factor of 4 higher in resolution than the discrete sampling. In Figure F7 the discrete and XRF depth series are compared to each other to determine whether the higher resolution XRF estimate actually represents CaCO3 variations. The discrete CaCO3 and XRF CaCO3 profiles strongly resemble each other, despite the coarser discrete sample spacing. All of the major features and most of the minor features in the XRF CaCO3 profile are reproduced by the discrete CaCO3 data. A few points show significant mismatches between discrete and XRF CaCO3. In each case, the difference appears to be either a sample with the wrong depth or a sample where there is an actual difference in CaCO3 caused by the different sample positions of each measurement.

Figure F7 also shows the gamma ray attenuation (GRA) bulk density profile from Site U1338 (see Fig. F29 in the “Site U1338” chapter [Expedition 320/321 Scientists, 2010b]). The GRA profile is another way to estimate high-resolution CaCO3 trends based on the well-known relationship between CaCO3 content and bulk density in equatorial Pacific sediment (Lyle and Dymond, 1976; Mayer, 1991; Lyle et al., 1995). In order to make CaCO3 estimates from bulk density, it is first necessary to “decompact” the sediment (i.e., remove the porosity decrease caused by increasing the sedimentary overburden) (Mayer, 1991). In short intervals such as the one shown in Figure F7, decompaction can be ignored when comparing trends.

Small features not clearly resolved in the 10 cm discrete sampling appear in both the GRA bulk density profile and the XRF CaCO3 estimate. For example, in the interval between 206 and 207.5 cmcd, both bulk density and XRF clearly identify an intermediate peak between the initial and final peaks. The discrete data are ambiguous about the middle peak, as it was sampled by only 1 data point. The higher resolution XRF estimate of CaCO3 thus contains an actual signal from the sediment and not sample-to-sample noise.

Large-scale features of the Site U1338 CaCO3 profile

Large-scale chronostratigraphic CaCO3 events have been previously identified in the equatorial Pacific and are the basis for the Neogene equatorial Pacific seismic stratigraphy (Mayer et al., 1986). One of the objectives of Expedition 320/321 was to better identify these events and put them into a modern chronostratigraphic framework.

Figure F8 is an illustration of the Site U1338 CaCO3 curve and identifies important low CaCO3 intervals by blue shading. There are three potential processes that cause low CaCO3—additional dissolution by more corrosive bottom water, low carbonate productivity, or extreme relative increases in productivity by diatoms (a bio-SiO2 producer; e.g., Kemp and Baldauf, 1993). Further work is being done to determine how each process contributes to each low-CaCO3 interval. Nevertheless, we note that low-CaCO3 intervals have been identified as causes of the distinct Neogene seismic stratigraphy of the equatorial Pacific (Mayer et al., 1986; Bloomer et al., 1995) and can be correlated for 3000 km across the equatorial Pacific (Tominaga et al., 2011). In Figure F8 we use a revision of the Site U1338 age model (J. Backman and J. Baldauf, unpubl. data) to correlate the Mayer et al. (1985) central equatorial Pacific seismic stratigraphy to the Site U1338 carbonate curve.

Inspection of Figure F8 shows that CaCO3 content varies at multiple age scales. The high-resolution “chatter” in the record represents the ±20% variation in CaCO3 associated with orbitally forced climate change. This scale of variation will be investigated more fully in the future as the age models are orbitally tuned. The larger scale events shown in Figure F8 have amplitudes of 40% or greater and last for varying amounts of time, as little as 300 k.y. to ~2 m.y. These low-CaCO3 events mark significant reorganizations of the equatorial Pacific that are important to understand, as they probably do not represent linear responses to orbital forcing. There are now new tools to investigate their history with the newly collected near-continuous sediment records from Expedition 320/321 and new high-resolution analyses using calibrated XRF scans.