The production of cold, deep waters in the Southern Ocean is an important factor in the Earth's heat budget. The supply of deep water to the Pacific Ocean is presently dominated by a single source, the deep western boundary current east of New Zealand. Here we use sediment records deposited under the influence of this deep western boundary current to reconstruct deep-water properties and speed changes during the Pleistocene epoch. In physical and isotope proxies we find evidence for intensified deep Pacific Ocean inflow and ventilation during the glacial periods of the past 1.2 million years. The changes in throughflow may be directly related to an increased production of Antarctic Bottom Water during glacial times. Possible causes for such an increased bottom-water production include increasing wind strengths in the Southern Ocean or an increase in annual sea-ice formation, leaving dense water after brine rejection and thereby enhancing deep convection. We infer also that the global thermohaline circulation was perturbed significantly during the mid-Pleistocene climate transition between 0.86 and 0.45 million years ago.

Abstract. Diffuse reflectance records from Feni Drift in the North Atlantic faithfully record sediment percent carbonate. A high-resolution, reflectance-based age model for these sediments derived from an orbitally tuned age model for western equatorial Atlantic, Ceara Rise sediments was generated by spectral frequency mapping. Power spectra of the Feni Drift record indicate statistically significant subMilankovitch cyclicity at 7.6-8.4 and 4.8-6.1 kyr. We infer that these -8 and-5 kyr cycles document a linkage between North and equatorial Atlantic climate given our ability to correlate these records. These climate cycles influence Atlantic basin carbonate prior to the intensification of Northern Hemisphere glaciation and thus must arise from some portion of the climate system other than the dynamics of large ice sheets. The presence of these peaks, which could be related to equatorial clipped precession, implies a possible non-linear response to Milankovitch forcing.

Ocean Drilling Program (ODP) Leg 138 was designed to study the late Neogene paleoceanography of the equatorial Pacific Ocean at time scales of thousands to millions of years. Crucial to this objective was the acquisition of continuous, high-resolution sedimentary records. It is well known that between successive advanced piston corer (APC) cores, portions of the sedimentary sequence often are absent, despite the fact that core recovery is often recorded as 100%. To confirm that a continuous sedimentary sequence was sampled, each of the 11 drill sites was multiple-APC-cored. At each site, continuously measured records of magnetic susceptibility, gamma-ray attenuation porosity evaluator (GRAPE), wet-bulk density, and digital color reflectance were used to monitor section recovery. These data were used to construct a composite depth section while at the site. This strategy often verified 100% recovery of the complete sedimentary sequence with two or three offset piston-cored holes.Here, these initial efforts have been extended to document the recovery of a complete sediment section and to investigate sources of error associated with sediment density measurements and changes in local sedimentation rates. At Sites 846 through 852, fine-scale correlation (on the order of centimeters) of the GRAPE records was accomplished using the inverse correlation techniques of Martinson et al. (1982). Having a common depth scale for all holes at each site facilitated comparison of highresolution data from different holes. After refining the interhole correlation, GRAPE records from adjacent holes were "stacked" to produce a less noisy estimate of sediment wet-bulk density for Sites 846 through 852. The continuity of the stacked GRAPE record is confirmed with reflectance and susceptibility records. The resulting stacked GRAPE records have a temporal resolution of less than 1000 yr for the past 5 m.y. Moreover, the stacking procedure allows for development of error estimates for measurements present in more than one hole. An important advantage provided by this framework is that one can determine the range of sedimentation variability between adjacent holes at a given site. This variability is caused by local sedimentation variability and by artifacts of the coring process. We demonstrate that the depth domain changes in sedimentation variability required to correlate among adjacent holes are larger than the changes induced by time-scale tuning procedures.

Terrigenous sediments from Ceara Rise in the western tropical Atlantic Ocean record Pleistocene Amazon Basin climate variability. Iron oxides and oxyhydroxides in this region originate mainly from chemically leached Amazon lowland soils. Concentrations of goethite and hematite in the terrigenous fraction consistently peak during transitions from glacial to interglacial periods, suggesting an increased proportion of erosive products derived from the Amazon lowlands compared to the physically weathered highlands. Lowland Amazon Basin precipitation changes, monitored by the percentage of goethite relative to total iron oxides, lead ice age extremes with maximum aridity during ice growth and maximum precipitation during ice melt. Rapid climate changes over the Amazon Basin may reflect shifts in the position of the Intertropical Convergence Zone forced by northern hemisphere insolation at precessional (1/23,000 yr−1) and obliquity (1/41,000 yr−1) frequencies. Variance in the orbital eccentricity bands (1/100,000 and 1/413,000 yr−1) may be explained by nonlinear amplification of insolation forcing at precessional frequencies. The early response of Amazon precipitation to insolation, ahead of high-latitude ice volume (δ18O) at all orbital frequencies, suggests that tropical aridity is part of the chain of events leading to ice ages, rather than a response to glacier oscillations.

No abstract