A 136.1-mcd thick (116.8 mbsf) late Pleistocene (~170 ka) hemipelagic sediment sequence was recovered at Site 1233. Age control points for the last 7 k.y. were derived from a nearby sediment record (core GeoB 3313-1; Lamy et al., 2002) that was dated by AMS 14C. For the interval >7 ka, variations in paleomagnetic directions and intensities at Site 1233 could be correlated to dated records elsewhere (Table T19). Linear sedimentation rates (LSRs), MARs, and carbonate MARs were calculated at 10-k.y. intervals (see "Age Models and Mass Accumulation Rates" in the "Explanatory Notes" chapter).
The late Pleistocene interval recovered at Site 1233 provided no biostratigraphic datum to define an age model. The continued presence of E. huxleyi to the bottom of Hole 1233B indicates that the entire sequence is younger than 260 ka. The E. huxleyi acme zone (base at 80 ka) was not observed either because of the generally low abundance of the species or, more likely, because the zone cannot be applied to this part of the southeast Pacific (see "Biostratigraphy").
The magnetic susceptibility record of the uppermost 9 mcd recovered at Site 1233 could be correlated directly to the magnetic susceptibility of the 14C-dated sediment record of core GeoB 3313-1 (Fig. F28), which justified the transfer of six AMS 14C ages ranging from 0.260 to 7.090 ka to Site 1233 (Table T19). The 14C age of 5.590 ka was not considered as an age control point at Site 1233 because of an uncertain match between the magnetic susceptibility records in this interval.
Site 1233 provided an unprecedented high-resolution Southern Hemisphere record of centennial- to millennial-scale variability in paleomagnetic intensities (Fig. F29), which offered the opportunity for developing a preliminary age model. We used the paleointensity record from Site 1089 in the sub-Antarctic South Atlantic (Stoner et al., in press) that was dated by oxygen isotope stratigraphy as a target record for correlation with Site 1233. The Sint-200 paleointensity stack (Guyodo and Valet, 1996) is also shown for reference in Figure F29. To avoid overinterpretation regarding the small-scale variability, we correlated only the large-scale pattern of changes in paleointensity between Sites 1233 and 1089 (Fig. F29). Among them, the most prominent feature in late Pleistocene paleointensity fluctuations of the geomagnetic field is the intensity low associated with the Laschamp Event (41 ka). At Site 1233, the directional record of this event is observed to occur over a 2-m interval between 65 and 70 mcd in Holes 1233B, 1233C, and 1233D (see "Paleomagnetism"). The correlation of magnetic intensity variations between Sites 1233 and 1089 provided additional age control points for the interval from 41 to 121 ka (Table T19). We chose to add age uncertainties of ±3 k.y. and depth uncertainties of ±2 m to these tie points, which resulted in a conservative straight-line model for depths >75 mcd.
Average MARs and LSRs are generally very high at Site 1233, ranging from 40 to 150 g/cm2/k.y. and from 40 to 160 cm/k.y., respectively (Fig. F30; Table T20). The lowest average MAR and LSR mark the interval from ~150 to ~50 ka (MIS 5 to 3). The rates increased by a factor of three on average from 50 ka to ~10 ka (marine isotope stage [MIS] 3) and the maximum MAR and LSR values at that time represent a rough estimate for the last glacial maximum (LGM). During the Holocene, the average MAR dropped to one third the LGM level, to ~50 g/cm2/k.y.
Siliciclastic material dominates the late Quaternary sediment at Site 1233 (generally >90 wt%). The total MAR therefore reflects the input of terrigenous material largely derived from the Andes and the Coastal Range provinces (see "Lithostratigraphy"). The current source is dominated by fluvial input. The late Pleistocene maximum in total MARs between ~50 and 10 ka indicates that this source has been enhanced during the LGM, possibly associated with temporal changes in continental rainfall patterns and/or refocusing of fluvial sediment discharge during glacial sea level lowstand. We cannot exclude the possibility that the tripling of MAR during MISs 2 and 3 is a random event, perhaps driven by a newly formed channel that favored a higher sediment flow to Site 1233.
Carbonate MARs account for only a small portion of the total MAR at Site 1233 but they are the highest carbonate MARs compared to any of the Leg 202 sites. Similarly, TOC MARs are very high (up to ~2 g/m2/yr after 50 ka) even if compared to other high-productivity regions like the upwelling zone offshore northwest Africa and California (1-4 g/m2/yr) (Stein et al., 1989; Lyle et al., 2000). The carbonate MAR increased from the lowest values of 1-2 g/cm2/k.y. at ages >50 ka to a maximum of ~10 g/cm2/k.y. at 20-10 ka, and then decreased to half that value, suggesting a significantly reduced productivity during the late Holocene. TOC accumulation parallels this trend. This is consistent with the findings of Hebbeln et al. (2002), who reconstructed the paleoproductivity off central Chile based on records of TOC, carbonate, and biogenic opal accumulation rates for the last 33 k.y. and found maximum values for the LGM and lower values for the late Holocene. Studies of productivity indicators in surface sediments from the Chile margin (Hebbeln et al., 2002) and in the Holocene sedimentary record of gravity core GeoB 3313-1 located at Site 1233 (Lamy et al., 2002) suggest that productivity changes off southern Chile are primarily controlled by the inflow of high-nutrient and low-chlorophyll waters of the Antarctic Circumpolar Current supplemented by micronutrients (e.g., iron) derived from the continental detritus. Future studies will address whether or not additional variations in the presently very restricted seasonal upwelling at Site 1233 contributed to the biogenic MAR changes.