Effects of temperature on Mg/Ca in neogloboquadrinid shells determined by live culturing

Authors


Abstract

[1] New culturing protocols allowed the development of a Mg/Ca temperature calibration equation for Neogloboquadrina pachyderma (d.). When cultured between 9°C and 19.2°C, Mg/Ca in N. pachyderma (d.) shell calcite exhibits a clear dependence on temperature. Shell calcite Mg/Ca increased exponentially with temperature: Mg/Ca = 0.51−0.15+0.23 * exp(0.10±0.02*T). Limited data suggest N. dutertrei exhibit similar elemental temperature relationships to N. pachyderma (d.). We applied the temperature relationship established by culturing to Mg/Ca in neogloboquadrinid shells collected using plankton tows. Compared to measured SST, the calculated temperatures were within ±2°C for Mg/Ca (95% CI). Multiple samples from single tows generally had calculated temperatures that ranged from near-surface values to temperatures consistent with calcification depths down to ∼70 m. Our work indicates that Mg/Ca temperature calibration equations for temperate and polar foraminiferal species have a similar exponential constant but a higher pre-exponential constant than those for surface-dwelling tropical species. Therefore accurate paleotemperature reconstructions from temperate and polar species require the use of an equation distinct from those used for tropical species.

1. Introduction

[2] Stable isotopic and elemental signals recorded in shells of the non-spinose planktonic foraminifera Neogloboquadrina dutertrei and Neogloboquadrina pachyderma are widely used to reconstruct past ocean conditions in temperate and high latitude environments. N. pachyderma (d.) and N. dutertrei are abundant in upwelling environments, where their depth habitat ranges from the surface to approximately 100 m in association with the thermocline and chlorophyll maximum [Fairbanks et al., 1982; Kroon and Ganssen, 1989; Sautter and Thunell, 1989, 1991]. Similarly, N. pachyderma (s.) range from deeper pycnocline depths (∼100–200 m) in high latitude environments, to surface waters in the highest latitudes [Bauch et al., 1997], allowing for the reconstruction of ocean conditions at these depths of calcification [Kanfoush et al., 2000]. In addition, thickly calcified neogloboquadrinid shells are more resistant to dissolution than thinner shells of other species in sediments exposed to corrosive deep waters. This often results in the abundance of well preserved neogloboquadrinid shells from sediments in which there are few intact shells of other species [Keany and Kennett, 1972; Bé et al., 1985]. Finally, N. pachyderma are the only planktonic foraminifera living in polar waters, where the sinistral morphotype dominates, along with a small percentage of genetically identical dextral N. pachyderma [Bé et al., 1977; Hemleben et al., 1989; Bauch et al., 2003].

[3] Although δ18O records based on foraminiferal shells are extremely valuable, variations in the δ18O of seawater complicates any interpretation of the temperature signal. Recent work has focused on N. pachyderma Mg/Ca as a temperature proxy for temperate and high latitude sites [Nürnberg, 1995; Nürnberg et al., 1996; Lea et al., 1999; Mashiotta et al., 1999]. Mg incorporation into foraminiferal calcite is primarily thermodynamic, leading to a Mg increase of approximately 10% per °C [Nürnberg et al., 1996; Mashiotta et al., 1999; Elderfield and Ganssen, 2000; Lea et al., 2000]. However, the pre-exponential constant appears to vary among species, requiring species- or genus-specific Mg/Ca temperature calibration equations [Lea et al., 1999; Anand et al., 2003; Lea, 2003]. Previous work on mono-specific core-top N. pachyderma (s.) yielded Mg/Ca temperature calibration equations from the Norwegian Sea and South Atlantic of (Mg/Ca = 0.55*exp(0.099T)) and (Mg/Ca = 0.41*exp(0.083T)), respectively. Additional calibration equations for temperate/subpolar species have been described using mixed planktonic species from core tops [Elderfield and Ganssen, 2000].

[4] It is preferable to use live culturing whenever possible to develop relationships that directly link shell chemistry with environmental conditions. Spatial and temporal integration of material may bias calibrations based on core-tops, sediment traps and plankton tows. Furthermore, cryptic speciation near frontal boundaries can complicate field-based calibrations because of the possibility of distinct species-specific relationships [Bauch et al., 2003; Darling et al., 2004]. Culturing spinose planktonic foraminifera, on the other hand, has established geochemical relationships between seawater and shell chemistry [Spero and DeNiro, 1987; Lea and Spero, 1994; Nürnberg et al., 1996; Spero et al., 1997; Bemis et al., 1998]. In several cases, culturing has elucidated previously unknown influences on foraminiferal shell geochemistry that could not be recognized by other approaches [Erez and Luz, 1983; Spero and Williams, 1988; Nürnberg et al., 1996; Spero and Lea, 1996; Mashiotta et al., 1997].

[5] Despite the importance of neogloboquadrinids for understanding past climate changes, controls on their shell isotopic and elemental signatures are largely unknown because they have not been calibrated by geochemical experiments. This is largely due to difficulties in culturing non-spinose planktonic foraminifera. Techniques to culture neogloboquadrinids and other non-spinose planktonic foraminifera in controlled geochemical experiments are described elsewhere [von Langen, 2001]. These newly devised protocols make it possible to calibrate the effect of temperature on Mg/Ca values incorporated in N. pachyderma (d.) and N. dutertrei shell calcite. This paper describes these culturing experiments and the application of these relationships to neogloboquadrinid shells collected by plankton tows from the Santa Barbara Channel (SBC) and San Pedro Channel (SPC) in the Southern California Bight.

2. Experimental Methods

2.1. Collection of Neogloboquadrinids

[6] Geochemical experiments based on live culturing of Neogloboquadrina pachyderma (d.) and N. dutertrei have lagged behind those for spinose foraminifera due to difficulties in collection and handling. N. pachyderma's small size and adaptation to deeper colder waters make it difficult to collect with SCUBA, requiring the use of plankton nets, which are more stressful to the foraminifera and lead to greater mortality during collection. In addition, the small compact morphology and absence of spines make it difficult to observe, handle and feed living N. pachyderma specimens and to remove and analyze shell chambers grown in culture. As a result, existing culturing protocols [Hemleben and Spindler, 1983; Spero and DeNiro, 1987] required significant adaptation for non-spinose foraminifera.

[7] For this experiment, N. pachyderma (d.) and N. dutertrei were collected from small boats by plankton tows in the SBC (spring 1994-early spring 2000) and SPC (summer 1996 and spring 2000) (Figure 1). Generally smaller (∼100–200 μm) specimens were collected in shallower (<50 m) compared to deeper (∼200 m) tows that regularly contained larger (∼300 μm), more encrusted specimens, which appeared to be more advanced in their lifecycle. Two surface tows were made during each trip to maximize the collection of N. pachyderma (d.) (one near the base or below the thermocline (∼30–50 m) and one in near surface waters (5–10 m)). Because the nets were open during ascent and descent through the water column, deeper tows also collected specimens from the surface and there was no distinction made between specimens collected in the near-surface or sub-surface tow after collection. Specimens were collected by a double plankton net with an inside mesh diameter ∼500 μm and outside mesh diameter of 63 μm. Before collections, thermocline depths were located using a custom made thermometer connected to 200 m of cable. Surface temperatures were calibrated with a mercury thermometer (accuracy of ∼0.1°). Simultaneous CTD casts allowed us to verify temperature profiles on several occasions. Plankton samples were immediately transferred to an insulated cooler after collection and kept at approximate collection temperatures with seawater and/or blue ice. A digital thermometer recorded minimum and maximum temperatures within the cooler during the transit to the laboratory.

Figure 1.

Map of the Southern California Bight showing locations of the Santa Barbara and San Pedro Channels. The approximate site of surface tows (shown by the black symbols) and their proximity to the University of California at Santa Barbara (UCSB), Point Conception, Catalina Island, and the Mid Santa Barbara Buoy (46053) and San Pedro Buoy (46025) used for SST data are indicated.

2.2. Culturing of Neogloboquadrinids

[8] Juvenile neogloboquadrinids were distinguished from adult specimens by their small size (<150 um) and thin transparent shells. As adults generally add calcite by thickening rather than adding chambers, juveniles were picked from the tow material, transferred to polystyrene culture containers (Falcon™), and allowed to recover overnight from the stress of collection. Each healthy specimen was transferred to an individual culture container over the next several days. The health of specimens was estimated by observing length, abundance and activity of rhizopodia, color and texture of cytoplasm and number of chambers with cytoplasm. The individual culture containers were placed in insulated PVC tanks connected by Tygon© tubing to heater/refrigerator circulating temperature baths that controlled the temperature (±0.2°C) of the culturing environment over the range of 9.0°C to 19.2°C. Salinity and pH of seawater were monitored during the beginning and end of experiments.

[9] N. pachyderma (d.) were grown at 9.0°C, 12.9°C, 16.2°C and 19.2°C. Within each temperature group, we cultured approximately 100 individuals for trace metal analysis. All specimens were cultured under low light levels (20–30 μEinst m−2 s−1). As non-spinose planktonic foraminifera cannot hold onto healthy Artemia in culture, specimens were fed every other day with a freeze-killed Artemianauplius (San Francisco brine shrimp) until observations revealed deteriorated health or a pre-reproductive condition. When observations of dead or empty shells revealed the end of ontogeny, they were removed from culture containers with artist brushes and transferred through an ultrapure water rinse to micropaleoslides, where they were stored for later preparation and analysis.

2.3. Preparation and Analysis for Mg/Ca

2.3.1. Sample Preparation and Cleaning

[10] Initial shell size of each N. pachyderma specimen was measured upon collection using a binocular microscope (Wild M3Z) fitted with a reticle. After culturing, the final shell size was measured, and the difference between final and initial shell size was used to determine the number of chambers grown in culture. Chambers grown in culture were amputated with a No. 11 surgical scalpel blade. Cultured N. pachyderma (d.) chambers were extremely small, averaging <0.5 μg each. Because each Mg/Ca or isotope analysis required approximately 10 μg of calcite, it was necessary to pool approximately 10–30 chambers from 10–20 cultured specimens for a single analysis. However, in one case (PVLAJ16) we were able to pool three whole shells when calculations based on 48Ca/44Ca spike and initial and final shell measurements indicated that 87% of the shell mass was precipitated in the laboratory [von Langen, 2001]. The much larger specimens of N. dutertrei only required 3–5 chambers for each analysis. For each tow sample, we ran individual N. dutertrei and combined 3–7 N. pachyderma (d.) shells. Tow specimens were cut open with a scalpel to increase the exposure of shell interiors to cleaning reagents.

[11] Culturing observations suggest N. pachyderma (d.) grown at higher temperatures sometimes exhibit morphology intermediate between N. pachyderma (d.) and N. dutertrei [de Vargas et al., 1997; Sautter, 1998]. Specimens with more than four chambers and a compact morphology were designated as “P-D intergrade” and were included with N. pachyderma (d.) in culture and tow samples. Shell weights helped to distinguish between N. dutertrei and P-D intergrade when morphology and size of specimens were ambiguous. N. dutertrei shells weighed at least 15 μg, while P-D intergrade weighed less than 8 μg, and N. pachyderma (d.) weighed approximately 2–4 μg. These results are consistent with neogloboquadrinid morphotypes from the Sulu Sea [Sautter, 1998].

[12] Foraminifera chambers precipitated in the laboratory were separated and cleaned to remove remnant organic material. Initial runs were cleaned using previously established bleach cleaning protocols [Mashiotta et al., 1997, 1999]. During the study, we observed that replicate samples collected by plankton tows and cleaned with bleach sometimes exhibited anomalously high Mg/Ca values and high variability. We interpreted high Mg/Ca to result from remnant organic material, likely composed of foraminiferal cytoplasm that was not completely oxidized during cleaning. Recent work suggests that hydrogen peroxide may be a more efficient cleaning method for live collected foraminifera [Pak et al., 2004]. Therefore, midway through the study we adopted the following peroxide cleaning protocol: shells were soaked for 30 minutes in a buffered H2O2 solution (50% H2O2 (30%) + 50% 0.1 N NaOH) in a ∼60–70°C water bath. Note that the peroxide cleaning uses a much higher concentration than regimens designed for sediment samples [Boyle and Keigwin, 1986]. In order to maximize the number of points available for analysis, both bleach- and peroxide-cleaned samples were included. Of the cultured samples, 20% were bleach-cleaned, whereas 36% of the tow specimens were bleach-cleaned (Tables 1 and 2).

Table 1. Elements in Neogloboquadrinids Cultured Under Low Light
SampleSpeciesCulture T, °CCleaning MethodMg/Ca, mmol/molNa/Ca, mmol/molCalcite,a μgSample Type
  • a

    μg calcite calculated by ICP-MS.

  • b

    Note: Sample not rejected based on 48Ca value of 0.170.

PVLAL07N. pachyderma (d.)9peroxide1.285.39chambers
PVLAI08N. pachyderma (d.)12.9peroxide1.985.15chambers
PVLAE24bN. pachyderma (d.)12.9bleach1.8711.92chambers
PVLAH06N. pachyderma (d.)16.2peroxide2.925.86chambers
PVLAJ08N. pachyderma (d.)16.2peroxide2.326.37chambers
PVLAE23N. pachyderma (d.)19.2bleach3.367.44chambers
PVLAH09N. pachyderma (d.)19.2peroxide3.836.06chambers
PVLAH19N. pachyderma (d.)19.2peroxide3.105.911chambers
PVLAI06N. pachyderma (d.)19.2peroxide3.727.04chambers
PVLAJ16N. pachyderma (d.)19.2peroxide4.196.93whole shells
PVLAL23N. dutertrei16.2peroxide2.495.017chambers
PVLAH21N. dutertrei19.2peroxide3.806.829chambers
Table 2. Elemental Ratios in Neogloboquadrinid Shells Collected by Shallow (0–50 m) Plankton Tows From the SBC and SPCa
SampleCollection DateCleaning MethodSpeciesMg/Ca, mmol/molNa/Ca, mmol/molCalcite, μgBuoy T, °CBuoy T, °C, SDMeas. T, °CMg/Ca T, °C
  • a

    Also included are temperatures measured during collection and hourly averaged buoy temperatures with standard deviation for a period of two weeks prior to collection.

Peter131-Apr-94bleachN. pachyderma2.015.2614.10.6 13.4
Peter141-Apr-94bleachN. pachyderma2.415.6814.10.6 15.2
PVLAJ2627-Apr-94peroxideN. pachyderma1.846.3512.40.7 12.6
PVLAJ2727-Apr-94peroxideN. pachyderma1.336.2612.40.7 9.4
PVLAC0827-Apr-94bleachN. pachyderma1.796.6512.40.7 12.3
PVLAC1027-Apr-94bleachN. pachyderma1.757.2512.40.7 12.1
PVLAC1127-Apr-94bleachN. pachyderma1.736.5612.40.7 12
PVLAC0927-Apr-94bleachN. pachyderma2.519412.40.7 15.6
Peter124-Sep-97bleachN. dutertrei3.9911.710  2020.2
Peter525-Sep-97bleachN. pachyderma3.825.99  20.419.7
Peter425-Sep-97bleachN. dutertrei2.446.313  20.415.3
Peter225-Sep-97bleachN. dutertrei4.2677  20.420.8
Peter731-Oct-97bleachN. dutertrei2.745.514  21.516.5
Peter631-Oct-97bleachN. dutertrei4.845.811  21.522.1
Peter914-Jan-98bleachN. pachyderma1.925.65  12.113
Peter814-Jan-98bleachN. dutertrei1.655.514  12.111.5
PVLAH0417-Jun-98peroxideN. pachyderma2.136.21014.90.416.514
PVLAH018-Jul-98peroxideN. pachyderma2.817.4216.71.517.816.7
PVLAH0822-Oct-98peroxideN. pachyderma2.66.1417.30.617.415.8
PVLAJ1413-Jan-99peroxideN. pachyderma1.746.3614.40.414.212
PVLAE1424-Jul-99bleachN. pachyderma1.756.4618.40.417.712.1
PVLAE1924-Jul-99bleachN. pachyderma1.777.2918.40.417.712.2
PVLAE2024-Jul-99bleachN. pachyderma2.249.91018.40.417.714.5
PVLAE2524-Jul-99bleachN. pachyderma2.438.2318.40.417.715.3
PVLAE1824-Jul-99bleachN. pachyderma2.527.4518.40.417.715.6
PVLAG3024-Jul-99peroxideN. pachyderma2.566.9818.40.417.715.8
PVLAG2924-Jul-99peroxideN. pachyderma2.855.8918.40.417.716.9
PVLAE1724-Jul-99bleachN. pachyderma3.188.2318.40.417.718
PVLAH238-Oct-99peroxideN. dutertrei2.56.5516.40.51415.6
PVLAJ2311-Feb-00peroxideN. pachyderma2.186.9613.90.41214.2
PVLAJ042-Apr-00peroxideN. pachyderma2.297.6614.60.415.714.7
PVLAJ036-Apr-00peroxideN. pachyderma1.816.5414.80.515.512.4
PVLAJ026-Apr-00peroxideN. pachyderma2.456.6814.80.515.515.4
PVLAK0113-Apr-00peroxideN. pachyderma1.786.7614.80.515.512.2
PVLAK1328-Apr-00peroxideP-D intergrade2.685.91815.21.11716.3
PVLAK1528-Apr-00peroxideN. pachyderma2.327315.21.11714.8
PVLAK1428-Apr-00peroxideN. dutertrei2.6461815.21.11716.1
PVLAK0810-May-00peroxideN. pachyderma2.196916.50.617.514.3
PVLAK0710-May-00peroxideN. pachyderma3.188.3116.50.617.517.9
PVLAK1010-May-00peroxideN. pachyderma3.259116.50.617.518.2
PVLAK0910-May-00peroxideN. dutertrei2.956.3916.50.617.517.2
PVLAK2817-May-00peroxideN. pachyderma1.9210.41015.8116.513
PVLAK2717-May-00peroxideN. dutertrei3.352515.8116.518.3
PVLAK2925-May-00peroxideN. pachyderma2.86.61715.5117.516.7
PVLAK2925-May-00peroxideN. pachyderma2.86.61715.5117.516.7
PVLAK2625-May-00peroxideN. dutertrei2.36.3915.5117.514.8
PVLAL1830-May-00peroxideN. pachyderma2.136.7517.11.719.414
PVLAL1330-May-00peroxideN. pachyderma2.225.71117.11.719.414.4
PVLAL1230-May-00peroxideN. pachyderma2.956.11517.11.719.417.2
PVLAL1130-May-00peroxideN. dutertrei2.88 117.11.719.417
PVLAL2515-Feb-01peroxideN. pachyderma1.466.4912.40.811.910.3
PVLAL2915-Feb-01peroxideN. pachyderma1.469.2412.40.811.910.3
PVLAL2715-Feb-01peroxideN. pachyderma1.666.61012.40.811.911.6
PVLAL3015-Feb-01peroxideN. pachyderma1.736.61012.40.811.912
PVLAL0610+17+19 May-00peroxideN. pachyderma2.145.71115.8116.514

2.3.2. Mg/Ca Analysis

[13] Samples were simultaneously dissolved and spiked with 500 μl of spike containing known amounts of 25Mg, 45Sc and 89Y in 0.1 N HNO3 (Fisher Optima). Ratios of 25Mg/24Mg, 45Sc/44Ca and 89Y/88Sr were collected and converted to elemental ratios by means of standard isotope dilution and internal standardization equations [Lea and Martin, 1996; Lea et al., 1999]. Initial sample runs (PAKBR + PETERS, PVLAC, PVLAE, PVLAG, PVLAH and PVLAI; Tables 1 and 2) were analyzed with a Fisions/VG PlasmaQuad 2+ Turbo inductively coupled plasma-mass spectrometer (ICP-MS) coupled to a Cetac-ultrasonic nebulizer. Reproducibility of small sample foraminiferal Mg/Ca using isotope dilution is estimated to be ∼2.5% (1 sigma) on the basis of repeated analysis of spiked gravimetric standards [Lea et al., 1999]. Later sample runs (PVLAJ, PVLAK and PVLAL; Tables 1 and 2) were analyzed by a ThermoQuest Finnigan Element 2 double focusing magnetic sector Inductively Coupled Plasma Mass Spectrometer (ICP-MS) with a self-aspirating low flow nebulizer. Long-term precision of the Element 2 is about 1% (1 sigma) for Mg/Ca.

[14] Because it was necessary to combine amputated chambers from multiple specimens in order to pool enough calcite for a single analysis, the approximately 400 individual neogloboquadrinids cultured for this study yielded 19 data points. Sample size and shell Na/Ca were monitored as a measure of data quality. Due to greater difficulty in cleaning and handling, the smallest samples were generally associated with higher Na/Ca values resulting from incomplete removal of cleaning reagents. Approximately 20% of the neogloboquadrinid Mg/Ca data were rejected on the basis of a combination of a small sample size (<6 μg) and high Na/Ca (>7.5 mmol/mol) values. Another 15% of the Mg/Ca data were rejected due to cleaning with expired reagent (Tables 1 and 2). Unfortunately, the majority of the rejected data (5 of 7 data points) were from the colder temperature groups, resulting in only one and two data points at 9 and 12.9°C, respectively.

3. Results and Discussion

3.1. Mg/Ca in Cultured Neogloboquadrinid Shells

[15] Mg/Ca shows a strong increase with temperature in cultured N. pachyderma (d.), increasing from 1.3 to 4.2 mmol/mol over the 10.2°C temperature range investigated (Figure 2 and Table 1). Fitting the 10 individual data points with an exponential relationship yields:

equation image

The standard error of equation (1) is equivalent to ±1.0°C. The lower and upper bounds (95% CI) for the pre-exponential constant of 0.51 are 0.36 and 0.74, respectively. The 95% CI for the exponential constant, 0.10 is ±0.02. Although the colder culture temperatures have admittedly fewer data points, the greater variability at the 19.2°C culture temperature may be in part due to the exponential Mg-temperature response, with greater variance expected at higher temperatures. The two data points for N. dutertrei grown at 16.2°C and 19.2°C under identical conditions are consistent with equation (1).

Figure 2.

Experimental Mg/Ca versus temperature for cultured N. pachyderma (d.) cleaned with peroxide (solid red circles) or bleach (open red circles) with exponential relationship fit through 10 data points (Mg/Ca = 0.51−0.15+0.23 * exp(0.10±0.02*T)R2 = 0.90 P < 0.001) composed of either amputated chambers or whole shells grown under low light conditions (20–30 μEinst m−2 s−1). Also included are two data points for N. dutertrei (large blue circles) grown under identical conditions at 16.2°C and 19.2°C. Uptake of Mg/Ca in neogloboquadrinid calcite roughly doubles every 7°C and is consistent with relationships previously described for cultured Globigerina bulloides [Mashiotta et al., 1999]. The standard error of the N. pachyderma culturing equation is ±0.10 units of ln(Mg/Ca), which is equivalent to ±1.0°C. The lower and upper 95% confidence intervals for the pre-exponential constant (0.51) are 0.36 and 0.74, respectively. The 95% confidence for the exponential constant (0.10) is ±0.02.

3.2. Mg/Ca in Plankton Tow Samples

[16] We evaluated the utility of the culture-based Mg/Ca (equation (1)) paleotemperature relationship by analyzing Mg/Ca in N. pachyderma (d.) and N. dutertrei shells collected by plankton tows. We compared Mg/Ca measured in neogloboquadrinid shells collected from the SBC and SPC directly against sea surface temperature (SST) (Figure 3 and Table 2).

Figure 3.

Mg/Ca measured in N. pachyderma (d.) (red squares) and N. dutertrei (blue circles) shells collected by shallow (0–50 m) plankton tows from the Santa Barbara Channel. Hourly averaged sea surface temperatures (SST) continuously measured by the Mid Santa Barbara Channel buoy (46053) and Catalina Buoy (46025) are plotted on the x axis. The culture-based N. pachyderma (d.) Mg/Ca calibration equation (this study) is shown by the black line. Points below the line likely result from subsurface calcification.

[17] Because of the complex, rapidly changing circulation of the Southern California Bight, it is unlikely that discrete SST measurements made during collection always represented the actual calcification temperatures. Neogloboquadrinids often survived longer than 4 weeks in culture and those that added secondary calcite in culture could survive for approximately a week before death or the end of ontogeny [von Langen, 2001]. Therefore thick-shelled neogloboquadrinids collected by plankton tows could have undergone significant secondary calcification up to a week before collection. As temperatures may have varied considerably during this period, it is difficult to directly relate the neogloboquadrinid calcification temperatures to discrete temperatures measured during collections that were often weeks apart. This is especially true for regions such as the Southern California Bight that experience rapid SST changes [Harms and Winant, 1998].

[18] We hypothesize that the integration of continuous SST measurements from buoys close to the collection site more accurately represents calcification temperatures than discrete temperature measurements made during collection [von Langen, 2001]. Because culturing experiments revealed that juvenile chambers contribute very little to total shell calcite, we assume neogloboquadrinids collected by plankton tows formed during the two prior weeks by adding larger and thicker chambers and/or a calcite crust. We therefore compare shell geochemistry and reconstructed temperatures to SST measurements averaged hourly over two weeks before collection by buoys belonging to the National Data Buoy Center (NDBC) (http://www.ndbc.noaa.gov/). We used Catalina Buoy 46025 (33°44′42″N 119°05′02″W) in the middle of the SPC to reconstruct calcification temperatures for specimens collected from sites approximately 10 km offshore of Catalina Island (Figure 1). In the SBC, we estimate calcification temperatures from measurements by the Mid SBC Buoy 46053 (34°14′10″N 119°50′48″W) (Figure 1).

[19] Although measured temperatures were generally within ±1.0°C of measurements from buoys, there were several instances when these temperatures differed by up to ∼2.5°C (Table 2). Similarly, surface temperatures from concurrent CTD casts in the Santa Barbara Basin also differ from measured and/or buoy temperatures by up to 2.5°C. It is likely these differences reflect SST changes of several degrees over time spans of hours to a few days caused by the complex circulation of fronts, eddies, or upwelling [Auad et al., 1999]. Removal of Buoy 46053 from the SBC during 1997 by the Minerals Management Service required the comparison of foraminiferal geochemistry to discrete measurements of SST made during collections (Table 2). However, discrete temperature measurements during this period should adequately reflect calcification temperatures because fall 1997 was characterized by a strong El Niño, when the thermocline was depressed and other buoys offshore California exhibited relatively stable SSTs.

[20] Across a SST range of 12.1 to 21.5°C, Mg/Ca values ranged between 1.3 to 3.8 mmol/mol and 1.7 to 4.8 mmol/mol for tow-collected N. pachyderma (d.) and N. dutertrei, respectively (Table 2). When plotted against SST, these Mg/Ca values were generally consistent with the exponential shape of the temperature relationship established by culturing (Figure 4). On the basis of the fit from culturing, most samples exhibited Mg/Ca values similar to or lower than would be expected if they calcified at the SST measured during collection (Figure 3). In a few cases, Mg/Ca values were only 55–60% of the ratios expected if specimens calcified at the measured SST (Figure 3). With the exception of a N. dutertrei sample with a Mg/Ca value of 3.3 mmol/mol, very few samples exhibited Mg/Ca values higher than expected by the culturing relationship (Figure 3).

Figure 4.

Experimental Mg/Ca versus temperature relationship for cultured N. pachyderma (d.) (red dashed line) compared with published calibration equations for temperate and subpolar planktonic foraminifera. Equations shown include Mashiotta et al. [1999]G. bulloides culture-based equation (light blue line), Elderfield and Ganssen [2000] core-top equation based on eight planktonic species including both G. bulloides and N. pachyderma (black dashed line), Nürnberg et al. [1995] N. pachyderma (s.) equation based on Norwegian Sea core-tops (orange line), and Nürnberg et al. [1995] N. pachyderma (s.) equation based on South Atlantic core-tops (dark blue line). Culturing data for N. pachyderma and N. dutertrei are shown in the red and blue circles, respectively.

[21] Lower Mg/Ca-based temperatures observed in tow neogloboquadrinids likely reflect subsurface calcification of these individuals. Previous work based on sediment trap and plankton studies in the California Bight indicates that N. pachyderma (d.) has a depth habitat ranging from the surface to well below the thermocline [Sautter and Thunell, 1991; Pak and Kennett, 2002; Field, 2004; Pak et al., 2004], with larger, more heavily encrusted individuals found at greater depths. Plankton tow samples for this study were collected from either near-surface (5–10 m) or slightly subsurface (30–50 m) depths. Although vertical temperature profiles are not available for every collection date, periodic CTD casts confirm that Mg-based temperatures are generally consistent with measured temperatures of the upper 50 m [Shipe and Brzezinski, 2001]. Applying the culture-based N. pachyderma (d.) temperature equation to sediment trap N. pachyderma (d.) from Santa Barbara Basin also yields below-thermocline temperatures, indicating a deep subsurface habitat for mature specimens with significant secondary calcite [Pak et al., 2004], and supporting subsurface calcification for plankton tow N. pachyderma (d.) from this study.

3.3. Comparison With Previous Temperate/Polar Foraminiferal Mg/Ca Temperature Relationships

[22] Our culturing work indicates that neogloboquadrinid Mg/Ca increases 10% per °C (equation (1)). The derived Mg/Ca temperature response is virtually indistinguishable from a culture-based calibration equation for the temperate planktonic species Globigerina bulloides [Mashiotta et al., 1999] (Figure 4 and Table 2), as well as from a previous relationship based on core top N. pachyderma (s.) shells from the Norwegian Sea (Mg/Ca = 0.55−0.90+0.10 * exp(0.099±0.03*T) [Nürnberg, 1995]). Similarly, the culture-derived pre-exponential constant (0.51) agrees well with a pre-exponential constant derived from mixed temperate planktonic species from North Atlantic core-tops and an assumed constant of 0.1 [Elderfield and Ganssen, 2000] (Figure 4 and Table 2). An equation developed from N. pachyderma (s.) Mg/Ca from South Atlantic core-tops (Mg/Ca = 0.41−0.11+0.13 * exp(0.083±0.02*T)) [Nürnberg, 1995] indicates lower Mg/Ca values for equivalent temperatures than found for other temperate and polar foraminiferal species (Table 3, Figure 4). We hypothesize that this offset reflects partial shell dissolution in the undersaturated waters of the South Atlantic, biasing the Mg/Ca data toward lower values. Alternatively, N. pachyderma (s.) may be a distinct genotype from N. pachyderma (d.) [Darling et al., 2004] whereby the observed relationship reflects a species-specific response to Mg/Ca and temperature. Sea surface temperatures derived from a downcore N. pachyderma (s.) Mg/Ca record from the deep subantarctic Pacific using the Nürnberg [1995] South Atlantic calibration equation are consistent with independent evidence for SSTs in this region [Mashiotta et al., 1999], further supporting the inference that this equation is genotype-specific or that it applies to specimens that have experienced some dissolution.

Table 3. Comparison of Mg-Temperature Relationships for Temperate/Subpolar Planktonic Foraminifera
ReferenceMg/Ca = b*e(mT)b
m
Nürnberg [1995], core-top  
N. pachyderma (s.) Norwegian Sea0.0990.55
N. pachyderma (s.) South Atlantic0.0830.41
Mashiotta et al. [1999], culturing  
G. bulloides0.1020.53
Elderfield and Ganssen [2000], core-top  
Eight planktonic species0.10a0.52
This study, culturing  
N. pachyderma (d.)0.1010.51

[23] With the exception of the South Atlantic N. pachyderma equation, Mg/Ca temperature calibration equations based on temperate and polar species have similar pre-exponential constants of 0.5 ± 0.15 (Table 3). In contrast, calibration equations based on tropical species from culturing [Nürnberg et al., 1996], core-top [Dekens et al., 2002] and sediment trap [Anand et al., 2003] studies are all characterized by lower pre-exponential constants of (0.38 ± 0.06). Both tropical and high latitude calibration equations indicate an exponential temperature increase of ∼ 10%/°C. Although the errors of the pre-exponential constants overlap, tropical and high-latitude equations suggest that accurate temperature reconstructions based on foraminiferal Mg/Ca require the use of latitudinally appropriate calibration equations in at least this case.

4. Conclusions

[24] Neogloboquadrina pachyderma (d.) cultured over a temperature range of 9 to 19.2°C reveal that shell Mg/Ca values follow an exponential relationship of Mg/Ca = 0.51−0.15+0.23 * exp(0.10±0.02*T).

[25] Applying the culture-based Mg/Ca temperature relationship to neogloboquadrinids collected by shallow plankton tows results in reconstructed temperatures consistent with measured water temperatures of the upper 50 m in Santa Barbara and San Pedro Channels. With the exception of a South Atlantic calibration equation that likely incorporates a dissolution signal, the N. pachyderma culture-based Mg/Ca: temperature relationship is indistinguishable from previous G. bulloides and N. pachyderma (s.) culturing and core-top results. However, it has a higher pre-exponential constant than temperature calibration equations based on tropical species. Our work suggests that latitudinally appropriate calibration equations are necessary for accurate paleoceanographic reconstructions using foraminiferal Mg/Ca.

Acknowledgments

[26] Laboratory and field assistance was provided by Jonathon Blythe, Dayle Gates, Gerick Bergsma, Bryan Bemis, Daniel Schuller, Laurie Juranek, Krista Ehrenchlou, Annalise Schilla, Hawkeye Sheene, Brian Krstich, Lauren Deeb, Robin Cathcart, and Zemeda Ainekulu. Shane Anderson and Chris Gotschalk assisted with collections and culturing setup at UCSB. George Paradis, Howard Berg, Dave Winter, and Qianli Xie provided technical expertise. We thank two anonymous reviewers for comments that improved the manuscript. This research was funded by NSF grants OCE-9415991 (to D.W.L.) and OCE-9416595 and OCE-9729203 (to H.J.S.), and a UCSB Regents Fellowship and NSF Graduate Student Traineeship in Coastal Ocean Processes to P.V.L.

Ancillary