Sea surface temperature changes in the southern California borderlands during the last glacial-Interglacial cycle

A variety of evidence suggests that average sea surface temperatures (SSTs) during the last glacial maximum in the California Borderlands region were significantly colder than during the Holocene. Planktonic foraminiferal δ18O evidence and average SST estimates derived by the modern analog technique indicate that temperatures were 6°–10°C cooler during the last glacial relative to the present. The glacial plankton assemblage is dominated by the planktonic foraminifer Neogloboquadrina pachyderma (sinistral coiling) and the coccolith Coccolithus pelagicus, both of which are currently restricted to subpolar regions of the North Pacific. The glacial-interglacial average SST change determined in this study is considerably larger than the 2°C change estimated by Climate: Long-Range Investigation, Mapping, and Prediction (CLIMAP) [1981]. We propose that a strengthened California Current flow was associated with the advance of subpolar surface waters into the Borderlands region during the last glacial.

weakens and a surface countercurrent (the Davidson Current) flows northward, landward of the California Current. These warm surface waters from the south result in reduced surface productivity [Tont, 1976].

Oceanographic Setting
The Southern California Borderlands are a series of basins that underlie the highly productive waters of the California Current and its associated eddies (Figure 1). The strength of the California Current changes in response to seasonal variations in the winds. Upwelling along the west coast of North America is forced by atmospheric circulation around the North Pacific highpressure cell that seasonally varies in both strength and position [Huyer, 1983]. The strength of the pressure gradient between the North Pacific high and the continental thermal low is the driving force for the surface winds in this area. The gradient is strongest during the summer when the low deepens (due to heating) and the North Pacific high increases in intensity [Reid et al., 1958]. Strong northerly winds during the spring and summer strengthen the current flow, with offshore Ekman transport causing coastal upwelling and high productivity [Huyer, 1983;Eppley, 1986]. Offshore Ekman transport is lower in the Borderlands than along the regions immediately to the north and south [Huyer, 1983] due to the offshore position of the California Current and the unfavorable orientation of the coastline relative to the prevailing winds. However, significant upwelling does occur in the Borderlands due to local wind stress curl [Thunell and Sautter, 1992]. In the winter, the northerly component of the wind Basins were used in this study (Figure 2). Table 1 provides a summary of core information including latitude, longitude, water depth, and core length. Each core was sampled at 5-to10-cm intervals for this study. Samples were disaggregated in a 100øC Calgon solution and wet sieved through a 63-gm screen. Following drying, the sediment samples were sieved at 150 gm and split into two separate aliquots for faunal and stable-isotope analyses.

Stratigraphy and Chronology
Oxygen-isotope analyses were performed on the planktonic foraminiferal species, Globigerina bullaides (150-to 250-gm size fraction; Table 2). This species was chosen for two reasons.
First, it was found to be sufficiently abundant at most sample depths and thus provides a continuous record. Second, this species primarily lives above the thermacline in this region [Sautter and Thunell, 1991 The depths of the oxygen-isotope stage 1-2 and stage 2-3 boundaries are indicated for each core (Figure 3). These boundaries were estimated by taking the midpoint between light and heavy extremes in isotopic values. The stage 1-2 and 2-3 boundaries have been dated at approximately 12 and 24 ka, respectively [Imbrie et al., 1984;Martinson et al., 1987]. The •5180 profiles were used to establish a basic glacial-interglacial stratigraphy for each core.
Carbon 14 (14C) dating of foraminiferal shells by accelerator mass spectrometry (AMS) was used to establish a chronology for each core. Mixed planktonic foraminiferal species were picked from several depths in each core and analyzed at the AMS •4C dating facility at the University of Arizona (Table 3) Figure 3). Using the age model described above, sedimentation rates were calculated for each core by assuming that the top of each core has an age of 0 years and that sedimentation rates are linear between 14C datums. Temporal changes in sedimentation rate for each core are presented in Table 4. Age estimates were linearly interpolated for each core sample, and the •5180 records were converted to a timescale ( Figure 4). Two of the cores (from Tanner and San Nicolas Basins) had poor recovery of the youngest material, and data are therefore not presented for the uppermost sections.

Micropaleontological Analyses
A total planktonic foraminiferal faunal analysis was carried out on Tanner Basin core AHF-11343 (>150-gm size fraction). The key species from this core are discussed as relative abundance data (Table 5; Figure 5). For all four cores, the abundances of sinistral-and dextral-coiling specimens of N. pachyderma were determined. All faunal counts were based on approximately 300 specimens per sample. Two types of N. pachyderma data were recorded: the number of sinistral-coiling specimens per gram of sediment and the percentage of total N. pachyderma specimens that are dextral coiling. These two sets of data were plotted against one another in order to illustrate their inverse trends ( Figure 6).

Temperature Estimation
In the present study, we evaluate average SST changes in the Southern California Borderlands for the last glacial-interglacial cycle using several different approaches. First, planktonic foraminiferal relative abundance patterns for the past 30,000 years are compared with modern seasonal distribution patterns for this region. This comparison allows us to document the response of plankton to changing climate-oceanographic conditions from the LGM to the Holocene. Second, we document temporal changes in the relative abundances of both sinistral-and dextralcoiling varieties of N. pachyderma. Bandy [ 1960] was the first to propose that there is temperature control on the coiling direction of this species. In the Borderlands region, Bandy [1960] found that dextral forms have dominated for the last 11,000 years, whereas sinistral populations dominated during the LGM. Sediment trap results from Station Papa (50øN, 145øW) in the subpolar northeast Pacific [Reynolds and Thunell, 1986] show that the dextral variety of this species dominates when SSTs are 12ø-14øC, whereas sinistral varieties are only present when SSTs at Station Papa are 6ø-8øC. A similar study in the North Atlantic [Wolfteich, 1994] also found that sinistral N. pachyderma were restricted to SSTs of less than 8øC. Although stable-isotopic evidence suggests that dextral-coiling N. pachyderma dwells beneath the surface mixed layer in the Borderlands region [Thunell and Sautter, 1992], we compare the two forms of this species with SST changes in the higher latitudes of the modern ocean since both coiling varieties are only found in these areas. Sediment trap results from the Borderlands are not useful for this distinction since only dextral-coiling N. pachyderma are observed in the modern population from this region [Thunell and Sautter, 1992]. Third, single-species planktonic foraminiferal •5180 changes are used to estimate maximum average sea surface temperature changes from the LGM to the present, once the ice volume effect is accounted for [Fairbanks and Matthews, 1978;Fairbanks, 1989] and ignoring unknown surface salinity changes which have not been independently estimated. Fourth, planktonic foraminiferal assemblages are used to derive quantitative estimates of average SST using the modern analog technique (MAT) [Hutson, 1979;Overpeck et al., 1985;Prell, 1985]. Finally, changes in the spatial occurrence of indicator coccolithophore species relative to their current biogeography are used to infer SST changes since the LGM.

Tanner Basin Planktonic Foraminiferal Species Census Data
The planktonic foraminiferal data for the Tanner Basin core

Abundance and Coiling Direction Changes of Neogloboquadrina Pac hyderma
All four cores show a peak in the number of sinistral-coiling N. pachyderma at approximately 15 ka (Figure 6). This is followed by a sharp increase in the percentage of dextral-coiling N. pachyderma at the oxygen-isotope stage 1-2 boundary (termination I). This change in coiling direction provides a reliable indicator of the last glacial termination and was previously noted by Gorsline and Prensky [1975].
The abrupt glacial-interglacial changes observed in sinistraland dextral-coiling varieties of N. pachyderma can be compared with the seasonal changes observed in this species at Station Papa (50øN, 145øW) in the North Pacific. Reynolds and Thunell [1986] found that these two taxa prefer very distinct SST and thermocline structure conditions. The sinistral-coiling variety prefers SSTs in the range of 6ø-8øC, whereas the dextral-coiling variety prefers SSTs in the range of 12 ø-14øC at this location. On the basis of sediment trap results from several locations in the North Atlantic, Wolfteich [1994] also found that sinistral N. pachyderma preferred temperatures less than 8øC, whereas the dextral-coiling form was more common at warmer temperatures. These changes in N. pachyderma coiling associated with seasonal SST changes may serve as a modern analog for the glacialinterglacial faunal changes observed in our cores.
While the abrupt change in coiling direction at termination I is common to all four cores, there are intercore differences in the abundance of sinistral-coiling N. pachyderma. The Tanner and San Nicolas Basin cores contain multiple peaks in sinistral N. pachyderma abundance during glacial stage 2 ( Figure 6). In contrast, the cores from Velero and No Name Basins are marked by only a single peak in the abundance of this species at the end of stage 2 ( Figure 6).

Planktonic Foraminiferal •5180 Changes
The planktonic foraminiferal oxygen-isotope records for each core are plotted against age in Figure 4. In each of the four

Borderlands cores, a 2.5-3.5%0 •5 ]80 change is observed from the
LGM to the Holocene. Subtracting the 1.3%o ice volume effect [Fairbanks and Matthews, 1978;Fairbanks, 1989] from the total glacial-interglacial •5180 change leaves a residual of 1.2-2.2%o that must be attributed to salinity and/or temperature changes at the depths where G. bulloides lives. If the residual is due entirely to salinity changes, this would mean a freshening of approximately 3-5%0 (according to a relationship plotted by Berger and Gardner [1975]) during the last deglaciation, which seems unlikely. We suspect that some combination of both salinity and temperature change leads to the observed residual, but the relative importance of each is unknown. If salinity changes are ignored and the entire 1.2-2.2%o •5180 residual is attributed to temperature, this would indicate a maximum average temperature difference of 5 ø-10øC between the LGM and present (Table 6). This estimation is based on the paleotemperature equation of Epstein et al. [1953], which predicts a 4.5øC change for each 1.0 %0 change in •5 ]80.

Other Evidence for Glacial-Interglacial SST Changes
In order to evaluate glacial-interglacial SST changes in the Borderlands further, we used the modern analog technique (MAT) for paleotemperature estimation [Hutson, 1979;Overpeck et al., 1985;Prell, 1985  with a core-top calibration database [Prell, 1985], and the 10 most similar analogs are selected using the squared-chord dissimilarity measurement (a 0.4 cutoff value was used) [Overpecket al., 1985]. Winter and summer SSTs associated with each of the analogs are averaged to produce SST estimates for the downcore samples. Glacial samples (~18 ka) were selected from each core, and winter and summer average SSTs were estimated using this procedure (Table 7). These glacial SST estimates are compared with present-day temperatures in the region and indicate that the average glacial-interglacial temperature differences for summer and winter were approximately 7øC and 9øC, respectively (Table 7). In terms of the modern biogeography of coccolithophores, it has been well established that Coccolithus pelagicus is primarily restricted to subpolar regions where water temperatures range between 6 ø and 14øC [Okada and Mclntyre, 1977  individuals and an increase in the abundance of G. quinqueloba. The Holocene assemblage is dominated by dextral-coiling N. pachyderrna, with minor occurrences of G. quinqueloba and G. bulloides. The decrease in abundance of G. bulloides during deglaciation may be an indicator of reduced upwelling throughout this period. Similarly, the decrease in G. quinqueloba throughout most of the Holocene may indicate warmer SSTs and less upwelling after approximately 7 ka. During deglaciation, the number of sinistral-coiling N. pachyderrna specimens per gram of sediment decreases by at least a factor of 3 in each core, coincident with an increase in the percentage of dextral-coiling specimens (Figure 6). This is clearly a biotic response to the rapid changes in climateoceanographic conditions in this region at the termination of the last glaciation. Several workers [Kheradpir, 1968[Kheradpir, , 1970Gorsline and Prensky, 1975] have used the coiling ratio of N. pachyderrna in the Borderlands to study the last glacialinterglacial transition. Gorsline and Prensky [1975] suggested that the change in coiling direction represents the passage through this region of average annual surface water temperatures of 8 ø-10øC. Our observation of an abrupt coiling change during deglaciation agrees well with these findings. The MAT-derived SST estimates based on planktonic foraminifera also suggest that average glacial SSTs were approximately 7ø-9øC lower than at present in the Borderlands region (Table 7). In addition, the occurrence of the subarctic coccolith C. pelagicus in the Borderlands during the LGM suggests that SSTs in this region were significantly cooler during the LGM. In summary, a combination of faunal, floral, and stable-isotopic evidence indicates that a substantial cooling (6 ø-10øC) of average SST occurred since the LGM in the Southern California Borderlands. This observation is supported by recent findings from Santa Barbara Basin, where planktonic foraminiferal assemblage changes suggest a 7ø-8øC warming throughout the last deglaciation [Kennett and Venz, 1995], and planktonic foraminiferal •5180 contrasts [Kennett and Ingram, 1995] are similar to those observed in the Borderlands. Our findings are further supported by the work of Prahl et al. [ 1995], in which they used the uk37 technique to demonstrate that SSTs at 42øN in the California Current were approximately 6øC during the last glacial and warmed to 12øC during the Holocene.
The dominance of sinistral-coiling N. pachyderma and the presence of the coccolith C. pelagicus in our cores during the LGM suggests that subpolar surface waters extended southward along the eastern margin of the North Pacific to at least the Borderlands region at this time. It has been suggested that the Laurentide ice sheet caused the North Pacific high-pressure cell to move farther south and closer to the continent during the LGM relative to its present position [COHMAP, 1988]. Because the position and strength of this atmospheric high controls the flow of the Califomia Current [Hickey, 1979;Huyer, 1983]

Climatic Instability During the Last Glacial
The four cores used in this study are marked by two different patterns in the abundance of sinistral N. pachyderma ( Figure 6). All four cores show a peak in sinistral-coiling N. pachyderrna/gram at approximately 15 ka (end of the last glacial).

For the cores from the two southern basins (Velero and No
Name), the early part of glacial stage 2 is not marked by any intervals of high abundance of this taxa. In contrast, the cores from the two northern basins (Tanner and San Nicolas) contain three peaks in sinistral-coiling N. pachyderma abundance during glacial stage 2 ( Figure 6). If high abundances of this species reflect the presence of subpolar waters, then the observed differences between the basins may provide information on the frequency and spatial extent of these subpolar incursions. We propose that the northern part of the Borderlands experienced repeated advances and retreats of subpolar waters during the last glacial. In contrast, subpolar waters only extended to the southern sector of the Borderlands at the end of the last glacial.
The high-frequency changes in abundance of sinistral-coiling N. pachyderrna during the last glacial in the northern Borderlands region may have more than just regional significance. Micropaleontological and stable-isotopic data from four piston cores in the Southern California Borderlands region have been used to estimate average SST changes during the last glacialinterglacial cycle. On the basis of our observations, the following conclusions can be drawn. 1. Changes in the relative abundances of planktonic foraminifera indicate that major climate-oceanographic changes occurred in the Borderlands region during the last 25,000 years. Specifically, an abrupt change in the coiling direction of N. pachyderma occurred during deglaciation, with sinistral-coiling forms being replaced by dextral-coiling individuals at 13-14 ka. The presence of sinistral-coiling N. pachyderma in this region during the last glacial is interpreted as indicating that subpolar surface waters extended into the Borderlands region during that time.

A planktonic foraminiferal •5•80 change of 1.2-2.2%o in
excess of the ice volume effect in each of our cores suggests that average glacial SSTs may have been as much as 5ø-10øC cooler than present, ignoring any changes in salinity.

MAT-derived
SST estimates for the LGM are 7ø-9øC cooler than at present and are thus consistent with the stable-isotope results. In addition, a subarctic coccolith, C. pelagicus, is abundant in glacial Borderlands samples. This suggests that a southward advancement of the subpolar waters along the eastern margin of the North Pacific may have occurred during the last glacial, bringing with it considerably cooler average SSTs. 4. Our results collectively suggest a 6ø-10øC cooling of average SSTs in the Borderlands during the last glacial. This is considerably larger than that predicted by CLIMAP [1981], probably due to their lack of core coverage in the eastern North Pacific, an inaccurate modern database for their temperature calibration, and use of radiolarians (versus foraminifera) in their transfer function equations. 5. Repeated advances and retreats of subpolar surface waters occurred in the Borderlands region during the last glacial. On the basis of abundance patterns of sinistral-coiling N. pachyderma in our cores, the northern basins contain evidence for multiple advances/retreats of the subpolar waters during the last glacial, whereas the southern basins record only a single incursion of subpolar waters. These observations are a further indication of the general instability of climate conditions during the last glacial period.