Contrasting variability in foraminiferal and organic paleotemperature proxies in sedimenting particles of the Mozambique Channel (SW Indian Ocean)

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 75 (2011) 5834–5848 www.elsevier.com/locate/gca

Contrasting variability in foraminiferal and organic paleotemperature proxies in sedimenting particles of the Mozambique Channel (SW Indian Ocean) Ulrike Fallet a,⇑, Jenny E. Ullgren b, Isla S. Castan˜eda c, Hendrik M. van Aken b, Stefan Schouten c, Herman Ridderinkhof b, Geert-Jan A. Brummer a a

Department of Marine Geology, NIOZ Royal Netherlands Institute for Sea Research, P.O. Box 59, NL-1790 AB Den Burg, The Netherlands b Department of Physical Oceanography, NIOZ Royal Netherlands Institute for Sea Research, P.O. Box 59, NL-1790 AB Den Burg, The Netherlands c Department of Marine Organic Biogeochemistry, NIOZ Royal Netherlands Institute for Sea Research, P.O. Box 59, NL-1790 AB Den Burg, The Netherlands Received 5 November 2010; accepted in revised form 8 February 2011; available online 11 August 2011

Abstract Accurate sea surface temperature (SST) proxies are important for understanding past ocean and climate systems. Here, we examine material collected from a deep-moored sediment trap in the Mozambique Channel (SW Indian Ocean) to constrain 0 and compare both inorganic (d18O, Mg/Ca) and organic (U k37 , TEX86) temperature proxies in a highly dynamic oceanographic setting for application in paleoceanography. High-resolution time-series current velocity data from long-term moorings (2003 – present) deployed across the Mozambique Channel reveal the periodic migration of four to six meso-scale eddies through the channel per year. These meso-scale eddies strongly influence water mass properties including temperature and salinity. Despite the dynamic oceanographic setting, fluxes of the surface-dwelling planktonic foraminifera Globigerinoides ruber and Globigerinoides trilobus follow a seasonal pattern. Temperatures reconstructed from G. ruber and G. trilobus d18O and Mg/Ca closely mirror seasonal SST variability and their flux-weighted annual mean SSTs of 28.1 °C and 27.3 °C are in close agreement with annual mean satellite SST (27.6 °C). The sub-surface dwelling foraminifera Neogloboquadrina dutertrei and Globigerinoides scitula recorded high-frequency temperature variations that, on average, reflect conditions at water depths of 50–70 m and 200–250 m, respectively. Concentrations and fluxes of organic compounds (alkenones and crenarchaeol) dis0 play no or only moderate seasonality but flux weighted means of the associated temperature signatures, U k37 , and TEX H86 of 28.3 °C and 28.1 °C, respectively, also closely reflect mean annual SST. We analyzed all time-series data using multiple statistical approaches including cross-correlation and spectral analysis. Eddy variability was clearly expressed in the statistical analysis of physical oceanographic parameters (current velocity and sub-surface temperature) and revealed a frequency of four to six cycles per year. In contrast, statistical analysis of proxy data from the sediment trap did not reveal a significant coupling between eddy migration and organic compound fluxes or reconstructed temperatures. This is likely a result of the relatively low resolution (21 days) and short (2.5 years) duration of the time series, which is close to the detection limit of the eddy frequency. Ó 2011 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +0031 35 737 0271.

E-mail address: [email protected] (U. Fallet). 0016-7037/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2011.08.009

Foraminiferal and organic paleotemperature proxies in sedimenting particles

1. INTRODUCTION In order to understand past ocean and climate systems, proxies are needed that accurately record environmental conditions (Elderfield, 2002). Several organic and foraminiferal temperature proxies have been established (e.g. Erez and Luz, 1983; Nu¨rnberg et al., 1996; Schouten et al., 2002) and subsequently fine-tuned to ambient surface ocean temperatures (e.g. Mu¨ller et al., 1997; Spero et al., 1997; Sadekov et al., 2008). The temperature proxies generally used can be broadly divided into two groups; those obtained from calcareous microfossils and those from organic compounds. Together they allow for a multi-proxy approach for the reconstruction of past environmental conditions. Foraminifera record temperatures via the changing 18O/16O (Emiliani, 1955; Shackleton and Vincent, 1978; Bemis et al., 1998) and Mg/ Ca ratios (Nu¨rnberg et al., 1996) of their calcite shells. The 0 U k37 index (Prahl et al., 1988) is based on the ratio of long-chain di- and tri-unsaturated ketones produced by haptophyte algae (Brassell et al., 1986), which increases with increasing growth temperature. Another organic geochemical SST proxy is the TEX86-index (Schouten et al., 2002), which is derived from the distribution of several different glycerol dialkyl glycerol tetraethers (GDGTs) produced by Marine Group 1 Crenarchaeota (Thaumarchaeota), organisms that are ubiquitous in marine environments. The underlying prerequisite for all these proxies is that temporal coupling between the different signal carriers is maintained upon deposition and burial in the sediment. For foraminiferal shells this coupling between the water column and ocean floor is strong since foraminifera are little affected by ocean current transport due to their relatively fast sinking speeds (Takahashi and Be´, 1984; Gyldenfeldt et al., 2000), especially in the frequently used larger grain sizes (>250 lm). In contrast, organic proxies can be strongly influenced by current advection and subsequent lateral transport over long distances (Thomsen et al., 1998; Benthien and Mu¨ller, 2000; Ohkouchi et al., 2002; Ru¨hlemann and Butzin, 2006). This process could cause a mismatch between reconstructed temperatures from sedimentary foraminifera and organic proxies (Sicre et al., 2005; Lee et al., 2008; Saher et al., 2009). The effect of currents has also been documented by differing radiocarbon ages of foraminifera, alkenones and GDGTs in surface sediments (Mollenhauer et al., 2007). Additionally, varying 0 current effects have been noted on the same proxy: U k37 values of sinking particles were found to differ from those of suspended particulate matter (Hamanaka et al., 2000). In this study, we compare different temperature proxies from a time-series sediment trap with satellite SST to establish how each proxy records temperature changes in the Mozambique Channel and how they are affected by strong current transport. In the Mozambique Channel, two major driving forces periodically occur that might influence the different temperature proxies: annual SST variation and eddy-induced transport. In order to determine the differential effects of these variables on particle fluxes and paleotemperature proxies, the (sub-) tropical Mozambique Channel offers a unique setting because it has a relatively large annual temperature gradient of 5.2 °C and four to

5835

six fast-rotating eddies migrate through the channel every year (Schouten et al., 2003; Harlander et al., 2009). As this study is embedded in the Long-term Ocean Climate Observation (LOCO) program (de Ruijter et al., 2004; Palastanga et al., 2006), we can assess highly resolved long-term measurements of temperature, salinity and current transports for the Mozambique Channel, allowing us to link them with the flux patterns of inorganic and organic particles 0 and the temperature proxies d18O, Mg/Ca, U k37 and TEX86. 1.1. Oceanographic setting The Mozambique Channel hydrography is strongly dominated by the southward migration of fast-rotating, mesoscale eddies. These eddies form at 10°S off the East-African coast after the bifurcation of the South Equatorial Current (Fig. 1). Satellite altimetry has shown that Mozambique Channel eddies can exceed 300 km in diameter (Schouten et al., 2003), have high rotational velocities often above 1.5 m/s and pass through the channel at a mean frequency of about four to six per year before joining the Agulhas Current. Backeberg and Reason (2010) suggested that the formation of meso-scale eddies in the Mozambique Channel is connected to variability in South Equatorial Current transport whereas Harlander et al. (2009) argue that the formation of eddies is due to Rossby waves migrating at the eigenperiod of the Mozambique Channel. Ongoing current observations beginning in November 2003 with a large array of eight moorings across the narrowest part of the channel (Figs. 1 and 2) indicate a mean southward transport of 17 Sv, with large fluctuations ranging between 45 Sv northward and 65 Sv southward with a deep undercurrent flowing northwards along the continental slope (de Ruijter et al., 2002; Harlander et al., 2009; Ridderinkhof et al., 2010). These long-term observations also provide in situ evidence that eddies occasionally reach the 2.5 km deep channel bottom thereby likely affecting lateral and vertical particle transport.

2. MATERIALS AND METHODS 2.1. Instrumental data Within the LOCO program we monitored current velocities along a transect across the Mozambique Channel with 9 upwards facing Acoustic Doppler Current Profilers (ADCPs) and 16 Recording Current Meters (RCMs) moored at specific depths from November 2003 to December 2009 sampling at 30 min resolution (Fig. 2). Current velocity data were low pass filtered with a 3.5 day Butterworth filter and sub-sampled daily. We here describe current velocities in North–South direction (velocity v) since it is the dominant flow direction in the channel (Schouten et al., 2003). For this study, we used current velocity v and temperature data at 110 m water depth recorded by an ADCP and a CTD deployed on mooring lmc5A (16.8°S, 41.1°E), which is closest to the trap site (see Section 2.2; Figs. 1 and 2). Additionally, time-series temperature data are available for 200 m water depth from the same mooring (Table 1, supplementary data). From these

5836

U. Fallet et al. / Geochimica et Cosmochimica Acta 75 (2011) 5834–5848

Fig. 1. The study area in the Mozambique Channel with the mooring locations indicated by the black dots and numbers (4, 5, 5A, 6, 7, 8 and 9), and channel depth is shown by the 500 m isobars. The black box in enlarged map roughly indicates the area from which satellite data were used.

time-series we calculated daily averages to eliminate periodicities of less than a day such as tides. For the SST time-series, we obtained weekly averaged re-analyzed Reynolds SST from the Giovanni database for the area 16–17°S and 40–41°E (trapsite) for the period from January 1997 to December 2009 (http://www. nomad1.ncep.noaa.gov/cgi-bin/pdisp_sst.sh). For surface mixed layer depth (SML-depth) and chlorophyll a, we used daily averages for the trapsite (see Section 2.2) from January 1997 to December 2006 that were obtained from the NASA satellite remote sensing database (http://www. gdata1.sci.gsfc.nasa.gov/daac-bin/G3/gui.cgi?instance_id = ocean_model_day). 2.2. Sediment trap moorings At 16.8°S and 40.8°E in the central Mozambique Channel nearby the aforementioned mooring lmc5A, we deployed four consecutive time-series sediment trap moorings (Fig. 1). Each time the Technicap PPS 5 sediment trap, with a baffled collecting area of 1 m2 and a 24-cup automated sampling carousel, was positioned at 250 m above the channel floor at 2250 m water depth (Fig. 2).

Prior to deployments, sample cups were filled with a HgCl2-poisoned and borax-buffered solution of seawater collected from the deployment depth (Loncaric et al., 2007). The sediment traps were programmed to a 21 day sampling interval. Here, we examined the foraminiferal proxies for a 2.5 year period and the organic proxies for a 3.5 year period (Table 1, supplementary data). Organic carbon concentrations were additionally analyzed on the most recently recovered time-series, which was programmed to a higher resolution of 17 days (also see Sections 3 and 4). 2.3. Foraminiferal analysis Time-series sediment trap samples were wet-split and foraminiferal shells cleaned as described by Fallet et al. (2009, 2010). Shells from Neogloboquadrina dutertrei (d’Orbigny, 1839) and Globigerinoides scitula. (Brady, 1882) were handpicked from the 250–315 lm fraction. Both species have been shown to consistently live in sub-surface waters (Fairbanks et al., 1982; Sautter and Thunell, 1991). Each species was analyzed for shell flux, d18O and Mg/Ca (Table 1, supplementary data); for details on d18O and Mg/Ca sample preparation and analysis see Fallet et al.

Foraminiferal and organic paleotemperature proxies in sedimenting particles

5837

Fig. 2. Cross-section of the Mozambique Channel showing bathymetry with mooring arrays indicating the instrument positions and depths with 9 upwards facing Acoustic Doppler Current Profilers (ADCP), 16 recording current meters (RCM), 23 conductivity temperature depth sensors (CTD) and a consecutively moored sediment trap (also Supplementary Table 1).

(2010). Flux-weighted averages were calculated by multiplying each proxy value with its corresponding flux value, dividing by the total flux, and summing the resultant values. 2.4. Organic analysis A second and third sediment aliquot of the sediment trap material were filtered over a 0.45 lm polycarbonate filter, frozen, freeze-dried, weighed and subsequently ground into a homogeneous powder. One of the two aliquots was analyzed for carbon and nitrogen separately before and after removal of the carbonate-carbon with 2 M ultrapure hydrochloric acid applied directly onto the sample (Bonnin et al., 2002). Analysis with a Carlo/Erba-Flash EA-1112 yielded total carbon (Ctot), organic carbon (Corg) and nitrogen, which were converted into weight percent organic matter (OM) by OM = Corg  2 (Romero et al., 2002). The other aliquot was ultrasonically extracted 4 using a solvent mixture of 2:1 dichloromethane (DCM) to methanol (MeOH). After extraction, excess solvent was removed from the total lipid extract (TLE) via rotary evaporation and sample extracts were run through a Na2SO4 pipette column. Known amounts of three internal standards were added to the TLE; squalane, 10-nonadecanone (C19 ketone), and a C46 GDGT. The TLE was next separated into apolar, ketone and polar fractions via alumina pipette column chromatography using solvent mixtures of 9:1 (vol:vol) hexane/DCM, 1:1 (vol:vol) hexane/DCM, and 1:1 (vol:vol) DCM/MeOH, respectively. The squalane internal standard eluted in the apolar fraction, the C19 ketone eluted in the ketone fraction and the C46 GDGT eluted in the polar fraction.

For quantification of alkenones, the ketone fractions were analyzed on an HP 6890 GC using a 50 m CP Sil-5 column (0.32 mm diameter, film thickness of 0.12 lm) and helium as the carrier gas. The oven program was initiated at 70 °C, increased at a rate of 20 °C/min to 200 °C then to 3 °C/min until 320 °C. The final temperature of 320 °C was held for 25 min. Selected fractions were analyzed by a Thermo Finnigan Trace Gas Chromatograph (GC) Ultra coupled to Thermofinnigan DSQ mass spectrometer (MS) to confirm the identification of the C37:2 and C37:3 alkenones. Compound concentrations were determined by relating chromatogram peak areas to the concentration of the internal standard. 0 The U k37 Index (Prahl et al., 1988) was used to estimate SST and values were converted to temperature using the calibrations of Mu¨ller et al. (1998) and Conte et al. (2006). The polar fractions, containing the GDGTs, were ultrasonically dissolved in a mixture of 99:1 (vol:vol) hexane:propanol and filtered through 0.45 lm PTFE filters. GDGTs were analyzed on an Aglient 1100 series LC/ MSD SL with an auto-injector and Chemstation software by High Pressure Liquid Chromatography – Mass Spectrometry (HPLC/MS) following the methods described by Hopmans et al. (2004), with minor modifications (Schouten et al., 2007). TEX86 ratios were calculated following Schouten et al. (2002) converted to SST using the TEX H 86 global core top calibration of Kim et al. (2010). 2.5. Statistical analysis Time-series spectral analysis was used to examine whether meso-scale eddies influence sub-surface temperatures, salinities, particle fluxes, or inorganic or organic

5838

U. Fallet et al. / Geochimica et Cosmochimica Acta 75 (2011) 5834–5848

temperature proxies in the Mozambique Channel. To identify different signals in our time-series data we applied spectral analysis as it explores the cyclical pattern of data. The purpose of the analysis is to decompose a complex time-series with cyclical components into a few underlying sinusoidal functions of particular wavelengths. This allows estimating the power spectrum of a random signal from a sequence of time samples and detecting any periodicities in the data. We choose spectral analysis over e.g. auto-correlation analysis (which yielded similar results), as it directly allocates a signal’s power to a certain frequency. For spectral analysis, we used the Welch method (Welch, 1967) as it significantly reduces the loss of information at the edge of each window as the individual data sets are overlapped in time. The resulting periodogram is time-averaged, which reduces the variance of the individual power measurements. Welch’s method is an improvement on

the standard periodogram spectrum estimating method in that it reduces noise in the estimated power spectra in exchange for reducing the frequency resolution. Due to the noise caused by imperfect and finite data, the noise reduction from the Welch-method is desired for our short, low-resolution time-series data. For spectral analysis, we linearly interpolated all organic concentrations, foraminiferal fluxes and temperature proxies to account for missing values and breaks between consecutive deployment periods. All time-series were de-trended prior to spectral analysis since the spectrum would otherwise be overwhelmed by a large value for the first cosine coefficient (for frequency 0.0). Apart from the time-series with 21-day resolution, we applied spectral analysis on the time-series of organic carbon with 17-day resolution to analyze the influence of resolution artifacts.

Fig. 3. Time-series of weekly resolved satellite SST, daily resolved satellite chlorophyll a and daily resolved satellite surface mixed layer depth (SML-depth) for the area 16–17°S and 40–41°E (highlighted in Fig. 1). Daily resolved current velocities in N–S direction (v) at 110 m and 200 m water depth from mooring 5A (Figs. 1 and 2) closest to the sediment trap mooring. We plotted the maximum time-series length for each parameter starting 01/01/1997 also used for spectrum analyses.

Foraminiferal and organic paleotemperature proxies in sedimenting particles

We calculated a one-sided power spectrum density (PSD) of each time-series, which is windowed into eight sections with a Hanning window with 50% overlap and specified window length. Window length was chosen subjectively so as to balance resolution against variance. pffiffiffiffi As a rough guide we used a window length of about 2 N , with N being the length of the time-series. All specific window lengths used for each time-series are given in the figure captions. The number (N) of Fast Fourier Transform (FFT) points used to calculate the PSD estimate is 256 or the next power of two greater than the length of each section of the timeseries. Note that if NFFT is greater than the segment, the data is zero-padded. If NFFT is less than the segment, the segment is wrapped to make the length equal to NFFT. The vector of normalized frequencies at which the PSD is estimated spans the interval 0 to Pi. Significance of results is evaluated against the 95% confidence intervals. The 95% confidence limit is estimated from the variance of the averaged periodograms using the Matlab statistics software.

5839

3. RESULTS The Mozambique Channel displays a dominant interannual SST contrast between summer and winter of 5.2 °C, with minima of 25.0 °C in late austral winter and maxima of 30.2 °C in summer (Fig. 3). This annual forcing is also revealed in the SML depth and chlorophyll a concentrations (Fig. 3). Chlorophyll a in the Mozambique Channel is relatively low with background values around 0.3 mg m3 and maxima of up to 0.8 mg m3. SML-depth varies annually between about 10 m in austral winter and 50 m in summer. Both parameters peak between June and August and are anti-correlated with SST. Current velocities from the Mozambique Channel at 110 m water depth vary (Fig. 3) strongly between 50 (directed northwards) and 150 cm s1 (directed southwards). The mean transport is directed southwards. Similar to velocity v, temperature at 110 m water depth is also highly variable fluctuating rapidly within days between 14 °C and 28 °C with an average of 21.6 °C. This trend is continued in

Fig. 4. Times series data from the sediment trap mooring collected from November 2003 until Nov. 2007 at a 21 day sampling resolution. The lower three panels show time-series concentrations of organic carbon (wt.%), the GDGT crenarchaeol (mg g1 sediment) and alkenones (lg g1 sediment) while the upper two panels show foraminiferal fluxes (specimens m2 day1) of sub-surface N. dutertrei and deep-dweller G. scitula.

5840

U. Fallet et al. / Geochimica et Cosmochimica Acta 75 (2011) 5834–5848

deeper waters as time-series temperature data from 200 m water depth also show a high sub-annual variability of 10.4 °C (10.8–21.2 °C), with an average of 16.1 °C (Fig. 3). From the particulate matter, we determined organic carbon, alkenone and GDGT concentrations (Fig. 4, Table 1, supplementary data). Organic carbon varies between minima of 4.7 wt.% and maxima of 10.6 wt.%, with an average of 6.9 wt.% over the entire deployment period from November 2003 to February 2009. The concentration of crenarchaeol, a specific biomarker for Thaumarchaeota (Sinninghe Damste´ et al., 2002), varies from 4.7 mg g1 to 15.1 mg g1 sediment, with an average of 10.4 mg g1 sediment from November 2003 to September 2007. In contrast, the concentration of alkenones is much lower ranging from 0.7 lg g1 to 11.6 lg g1 sediment with an average of 3.9 lg g1 sediment. Similarly, the shell flux of G. scitula to the sediment is highly variable ranging between 0.2 and 13.2 specimens m2 day1 from November 2003 to March 2006 (Fig. 4). A substantially higher shell flux was observed for N. dutertrei with minima of 0.4 and maxima of 25.7 specimens m2 day1. The average shell flux of N. dutertrei is 8.0 specimens m2 day1, which is roughly 2.5 times higher than that of G. scitula with 3.1 specimens m2 day1. From the organic compounds and the different foraminifera species, we also analyzed the associated temperature

0

proxies U k37 , TEX86, d18O and Mg/Ca (Fig. 5, Table 1, supplementary data). The results showed that the TEX86 ranges from 0.67 to 0.71 with an average value of 0.69, 0 while the U k37 -index varies between 0.95 and 1. It should 0 be noted that the U k37 index was at unity in some samples and thus, at its uppermost temperature limit. High variance was observed in the d18O and Mg/Ca of N. dutertrei and G. scitula (Fig. 5, Table 1, supplementary data). For N. dutertrei d18O values between 2.5& and 1.4& are recorded and Mg/Ca values range from 2.7 mmol/mol to 4.0 mmol/mol (Fig. 5, Table 1, supplementary data). The d18O values of G. scitula range between 0.49& and 4.31& with an average of 1.95&. Paired Mg/Ca values for the time-series vary less strongly between 1.6 mmol/mol and 1.9 mmol/mol with an average of 1.8 mmol/mol. 4. DISCUSSION 4.1. Organic fluxes and proxies 0

Flux-weighted annual averages of U k37 yielded SSTs of 28.3 °C and 28.1 °C using calibrations developed by Mu¨ller et al. (1998) and Conte et al. (2006), respectively. Both are close to the annual mean satellite SST of 27.6 °C (Fig. 3) and in situ measured SST (Fig. 6). However, the seasonal

0

Fig. 5. SST proxies reconstructed from sinking particles at the sediment trap mooring in the Mozambique Channel. TEX86 and U k37 values are plotted together with d18O and Mg/Ca values for the sub-surface species N. dutertrei and the deep-dwelling species G. scitula.

Foraminiferal and organic paleotemperature proxies in sedimenting particles

5841

Fig. 6. Continuous temperature data and standard deviation from shipboard CTD-casts taken close to the trap-site mooring (mooring 5A) in 2000, 2001, 2003, 2005, 2006, 2008 and 2009 and from moorings deployed at 110 m, 200 m and 400 m water depth (2003–2009) with error bars 0 indicating the total temperature range in these six years. Additionally, we plotted d18O-, U k37 -(Fig. 6a) and Mg/Ca-, TEX86 derived temperatures (Fig. 6b) and their total temperature range (error bars) to indicate the probable depth habitat of each organism. 0

U k37 minima of 0.95 is equal to a SST of 27.5 °C using the calibration of Conte et al. (2006), which is well above the winter minimum of 25.0 °C (Fig. 7). The fact that 0 the U k37 -index does not reflect the seasonal SST maximum of 30.2 °C is not surprising as this is above the uppermost 0 limit of the U k37 proxy, i.e. a value of 1, but we can only 0 speculate as to why the U k37 -proxy fails to record the seasonal SST minimum. A possible explanation for this phenomenon could be the constant mixing of freshly produced alkenones in surface waters and re-suspended (older) organics from the seafloor by current advection and transport (Thomsen et al., 1998; Benthien and Mu¨ller, 2000; Ru¨hlemann and Butzin, 2006). Given that the flux 0 weighted U k37 -temperature is in close agreement with mean annual SST in the Mozambique Channel, the alkenones are likely transported only relatively short distances through the channel. Total alkenone concentrations do not show a clear seasonal pattern although peak values are noted in December to January of 2004 and 2005 (Fig. 4). Similarly, total alkenone fluxes (not shown) also do not display any clear seasonal patterns. This contrasts with other studies showing

that alkenone concentrations or fluxes display pronounced seasonality, which has been observed in a wide variety of oceanographic settings. For example, seasonality is noted in sediment traps from NW Africa (Mu¨ller and Fischer, 2001) and the Mediterranean Sea (Sicre et al., 1999) as well as in particulate matter filtered from the water column of Chesapeake Bay (Mercer et al., 2005). A potential difference explaining the lack of seasonality in the alkenone signal of the Mozambique Channel is the presence of strong oceanographic currents, which could lead to enhanced lateral transport of alkenones from more distal sites, in contrast to the relatively more tranquil waters of (semi-enclosed) Chesapeake Bay and the Mediterranean Sea. 0 Like the alkenone concentrations and the U k37 -index, no pronounced annual cycle is noted in the crenarchaeol concentrations or the TEX86-values, although these records show considerable variability (Figs. 4 and 5). The average flux-weighted TEX H86 -SST of 27.9 °C (Fig. 7) closely reflects annual mean SST in the Mozambique Channel (Fig. 3) and in situ measured SST (Fig. 6). A lack of a pronounced seasonality in TEX86-derived SST has been previously observed in a deep-moored sediment trap from the Arabian

5842

U. Fallet et al. / Geochimica et Cosmochimica Acta 75 (2011) 5834–5848

Fig. 7. Organic and inorganic SST proxies calibrated to temperature for sinking particles at the sediment trap mooring including d18O, U 037 (left panel), Mg/Ca and TEX86 (right panel). The grey circles representing the flux-weighted annual mean temperature for each proxy, are indicated on the axes. (Data for G. ruber and G. trilobus are taken from Fallet et al., 2010). Note that d18O and Mg/Ca of G. ruber and G. 0 trilobus together with organic U k37 and TEX86 yield flux-weighted annual mean SSTs between 27.3 °C and 28.1 °C, which is very close to the annual mean satellite SST of 26.7 °C. Flux-weighted average means were calculated by multiplying each proxy value with its corresponding flux value, averaging the results and then dividing by the average flux.

Sea, with its mean temperature signal corresponding well with annual mean SST (Wuchter et al., 2006). Given that TEX H86 -temperatures are in close agreement with mean annual SST in the Mozambique Channel, the GDGTs, like the alkenones, are likely transported from relatively short 0 distances. Similar to U k37 -SST, TEX H86 -SST also does not

record the entire seasonal SST range in the Mozambique Channel, which might be caused by lateral advection and admixing of re-suspended material. Alternatively, it is also possible that the TEX86 reflects a subsurface signal as has been observed in a few studies (Huguet et al., 2007; Lee et al., 2008; Lopes dos Santos et al., 2010). However,

Foraminiferal and organic paleotemperature proxies in sedimenting particles

TEX H86 reconstructed SST agrees well with annual mean SST at this site. Furthermore, there is no strong correspondence between TEX H86 -temperatures and the d18O or Mg/Ca of subsurface dwelling foraminifera (see Section 4.2) suggesting that TEX H86 -temperatures reflect an upper water column signal. 4.2. Fluxes and proxies of foraminifera On the same sediment trap series studied here, we previously analyzed the surface-dwelling foraminifera Globigerinoides ruber and Globigerinoides trilobus for shell fluxes, d18O and Mg/Ca (Fallet et al., 2010). This study revealed that shell fluxes of surface-dwelling foraminifera are strongly seasonal with G. ruber fluxes peaking in late austral summer (March) and those of G. trilobus in early austral winter (July). Additionally, it has been shown that d18O and Mg/Ca of both G. ruber and G. trilobus is strongly coherent with SST on seasonal timescales, capturing the entire SST range of 25.0–30.2 °C (Fallet et al., 2010). Despite contrasting seasonal preferences, both d18O and Mg/Ca yield average flux-weighted SSTs of 28.1 °C for G. ruber and 27.2 °C for G. trilobus (Fig. 7) thereby closely reflecting annual mean SST of 27.6 °C (Fallet et al., 2010). Thus, surface-dwelling planktonic foraminifera of the Mozambique Channel carry a strong seasonal imprint both in fluxes and in reconstructed SSTs. For the subsurface-dweller N. dutertrei we observe a strong and frequent sub-seasonal variation in shell fluxes (Fig. 4). As several studies have shown that N. dutertrei mostly lives below the surface mixed layer (Dekens et al., 2002), the large variation in shell fluxes is probably caused by subsurface processes occurring on frequencies that are too high to be well-resolved with the current 3-weekly resolution of our sediment trap sampling. We found a strong sub-annual variability of 1.1& for d18O and 1.3 mmol/ mol for Mg/Ca (Fig. 5), which equals temperature differences of 10.0 °C and 4.3 °C using temperature equations developed by Epstein et al. (1953) and Anand et al. (2003), respectively (Fig. 7, Table 1, supplementary data). The d18O of N. dutertrei corresponds to the large temperature variability between 50 m and 70 m water depth recorded with shipboard CTDs (Fig. 6). Flux-weighted annual mean temperature derived from d18O is 24.3 °C and from Mg/Ca is 23.1 °C, both slightly higher than the annual average of in situ measured temperature at 110 m of 21.6 °C (Figs. 3 and 6). As the calibration error for d18O and Mg/Ca analyses range around 1.2 °C (e.g. Dekens et al., 2002) in situ measured temperature and proxy temperature are in relatively good agreement. These results suggest that in the Mozambique Channel N. dutertrei predominantly lives below the surface mixed layer at depths between 40 m and 70 m as observed by other studies in different parts of the world’s oceans (e.g. Dekens et al., 2002; Be´ and Hutson, 1977; Peeters, 2002). The smaller temperature variability in Mg/Ca in contrast to d18O might be caused to a large extent by the more homogeneous sample aliquot (solution of 20 specimens) for Mg/Ca measurements. For accurate d18O-measurements only 10 lg of material, equal to the weight of about one

5843

shell, is needed for optimal results (Barker et al., 2003). Yet, even though aliquots from a minimum of three crushed shells were analyzed for d18O, this does not equal a homogeneous solution of twenty specimens used for Mg/Ca analyses. Similar to N. dutertrei, the deep-dwelling species G. scitula shows a large, sub-seasonal variability in shell fluxes suggesting a pronounced influence of quick, high-frequency changes in sub-surface water mass characteristics. G. scitula has been found to occur at water depths of 200–300 m (Be´ and Hutson, 1977; Kahn and Williams, 1981; Hemleben et al., 1989) where in situ measured temperature in the Mozambique Channel varies rapidly between minima of 10.8 °C and maxima of 21.2 °C (Fig. 3, Table 1, supplementary data). This is captured in the strong variability in d18O of G. scitula, which fluctuates between 0.49& and 4.31& (Fig. 5). The d18O of in situ seawater (d18Ow) at 10 m depth is 0.56&, and is relatively stable across the Mozambique Channel (Fallet et al., 2010). When the d18O of G. scitula is corrected for d18Ow, this corresponds to a wide temperature range of 3.2–17.3 °C using the calibration developed by Horibe and Oba (1972) as suggested by Kahn and Williams (1981), (Figs. 6 and 7). As observed for N. dutertrei, G. scitula Mg/Ca is less variable in comparison with d18O. Mg/ Ca varies only about 0.35 mmol/mol, which equals a temperature range of 2.2 °C using the calibrations developed for deep-dwelling G. hirsuta (Anand et al., 2003). As G. hirsuta has a very similar Mg/Ca range of 1–2 mmol/mol this calibration seems appropriate as so far no species-specific calibration for G. scitula is available. The flux-weighted annual mean temperature of G. scitula is11.8 °C for d18O when corrected for d18Ow, and 15.9 °C for Mg/Ca (Fig. 7). Although low temperatures of 3.2 °C have been observed at depths below 1500 m in the Mozambique Channel (Donohue and Toole, 2003; Harlander et al., 2009), this was not observed in the upper 500 m of the water column where G. scitula is expected to live. It is therefore unlikely that temperature changes alone caused the wide temperature range recorded in d18O of G. scitula. Another possible explanation for the large differences noted in d18O- and Mg/Ca-derived temperatures is that rapid changes in salinity may have influenced d18Ow values, and thus d18O values of G. scitula. The salinity variability induced by migrating eddies at 200 m from 2003 to the present can be up to 0.6 psu (min: 35.0; max: 35.6). This translates into a d18Ow value of 0.3& using the relation between d18Ow and salinity (Loncaric et al., 2005). However, this can only explain about 1.0 °C to 1.5 °C of the total temperature variation of G. scitula. It may be possible that the extremely low temperatures of G. scitula could be related to a relatively large “temperature error” associated with this species. An earlier study on d18O of G. scitula (Kahn and Williams, 1981) showed that this species does not calcify in equilibrium with the surrounding seawater (“vital effect”). This vital effect was estimated to be approximately 3.9 °C (Kahn and Williams, 1981). Taking this vital effect into consideration, the d18O-based flux-weighted temperature of G. scitula is 15.7 °C in close agreement with the Mg/Ca-based temperature of this species and in situ measured temperatures at

5844

U. Fallet et al. / Geochimica et Cosmochimica Acta 75 (2011) 5834–5848

200 m water depth that are on average 16.1 °C (Fig. 6, Table 1, supplementary data). This observation supports earlier suggestions that G. scitula has a large vital effect and implies that this species preferentially excretes shells at about 200 m water depth, in agreement with observations from other open ocean settings (Kahn and Williams, 1981; Hemleben et al., 1989). 4.3. Effect of lateral and vertical transport on sinking organic particles and foraminiferal shells The underlying prerequisite for all SST proxies is that temporal coupling between the different signal carriers is maintained upon deposition and burial in the sediment. In foraminiferal shells this coupling between the water column and ocean floor is strong since foraminifera are only little affected by ocean current transport due to their relatively fast sinking speeds of 500–1000 m per day (Takahashi and Be´, 1984), especially in the frequently used larger grain sizes of >250 lm (Gyldenfeldt et al., 2000). It has been

shown that organic carbon can sink slowly through the water column at speeds of less than 2 m day1 when uncompacted (Fowler and Knauer, 1986) but more often it accumulates as marine snow causing the organic carbon to sink much faster at rates of >50 m day1 (de la Rocha and Passow, 2007). Depending on size, sinking rates of marine snow aggregates can equal the sinking rates of foraminifera (Bienfang, 1981; Takahashi and Be´, 1984; Berelson, 2002). If, however, organics are only slightly aggregated and not associated with mineral ballast, they can sink very slowly through the water column. In such a case, the associated organic proxies may be strongly affected by current advection and subsequent lateral transport over long distances (Thomsen et al., 1998; Benthien and Mu¨ller, 2000; Ru¨hlemann and Butzin, 2006). The lack of seasonality in the organic temperature proxies and the relatively weak seasonality in their fluxes are in contrast to the foraminifera, which reveal a strong seasonality. This likely indicates that at our study location the organic proxies are more affected by lateral advection in comparison to

Fig. 8. Welch spectrum analysis of a number of observed parameters. Stippled lines indicate the 95% confidence level. All spectra were computed using a Hanning window of size approximately 2  sqrt(N) where N is the record length, resulting in a window size w = 120 for velocity, temperature, chl-a and SML depth; w = 66 for satellite SST; w = 12 for crenarchaeol and alkenone fluxes and their respective temperature proxies; and w = 10 for the shorter time series of foraminiferal fluxes and associated temperature proxies d18O and Mg/Ca.

Foraminiferal and organic paleotemperature proxies in sedimenting particles

the foraminifera, and that at least a portion of the organic matter is likely derived from lateral and vertical transport. This hypothesis could be tested by radiocarbon dating the foraminifera, alkenones and GDGTs. 4.4. Potential influence of mesoscale eddies on particle fluxes and temperature proxies An alternative cause for the contrasting variability of the inorganic versus organic proxies could be an influence from eddies, which migrate through the Mozambique Channel at a frequency of about five cycles per year (Harlander et al., 2009; Ridderinkhof et al., 2010). Indeed, time-series analysis of temperatures at 110 m and 200 m water depth shows the same dominant peak at five cycles per year, strongly suggesting that deep-reaching eddies are a major driving force for changes in water mass characteristics below the surface-mixed layer (Figs. 3 and 8). Interestingly, spectral analysis (and cross-correlation) of the fluxes of sub-surface and deep-dwelling N. dutertrei and G. scitula also reveals peaks at four to six cycles per year, which are significant at the 95% confidence level (Fig. 8). The foraminiferal temperature proxies d18O and Mg/Ca, alkenone and cren0 archaeol concentrations, as well as U k37 -derived and TEX H86 –derived SSTs, all display peaks at four to six cycles per year although these are less prominent and close to their statistical significance (Fig. 8). As periodicities of four to six cycles per year are in agreement with those of the meso-scale eddies that strongly influence water mass properties in the Mozambique Channel, we hypothesize that particle fluxes and temperature proxies, in particular for the sub- and deep-dwelling foraminifera, are to some degree influenced by eddy transport. We note that the statistical analysis of our time-series data is relatively difficult because of the few available data-

5845

points (N = 39, 63) paired to a low resolution (21 days) sampling interval. These features restrict time-series analysis as the statistical significance of a signal increases with length and resolution of a time-series. The resolution of 21 days can result in a three-point problem especially for the higher frequency such as the one caused by eddy migration through the Mozambique Channel. To address this issue, a follow-up time-series was collected from the same site in 2009 using a 17 day sampling interval, thereby increasing the number of data points to a minimum of four per migrating eddy (Table 1, supplementary data). We analyzed this time series for organic carbon content (Fig. 9) and found a pronounced, statistically significant peak at five cycles per year, which is also observed in the lower-resolution time-series. It thus seems that the time-series fluxes are indeed affected by eddy migration through the channel that is, however, not well-resolved in the lower-resolution sediment trap study discussed here. Ideally, future studies specifically aimed at examining the influence of eddy migration on particle fluxes and temperature proxies should use a sampling resolution of 8 days. This would be possible with sediment traps that are equipped with a double carrousel of 24 sampling cups. With such a setting, both resolution and sampling length could be expanded to capture the eddy migration through the channel as well as allowing sampling through an annual cycle. 5. CONCLUSION Analysis of time-series satellite data reveals that the Mozambique Channel is influenced by annual variability (5.2 °C SST range) and the periodically recurring migration of meso-scale eddies. The annual SST variability is transposed to fluxes and temperature proxies (d18O and Mg/ Ca) of the surface-dwelling foraminifera G. ruber and G.

Fig. 9. Time-series % organic carbon content from sediment trap mooring with a 17 day resolution for the period from November 2007 to January 2009 (9a, left panel). P-Welch spectrum analysis using a Hanning window with window size 7. The stippled line indicates the 95% confidence level (9b, right panel). A pronounced, statistically significant peak at five cycles per year is noted.

5846

U. Fallet et al. / Geochimica et Cosmochimica Acta 75 (2011) 5834–5848

trilobus (Fallet et al., 2010). We find that the d18O of the thermocline dwelling species N. dutertrei closely records temperatures just below the surface-mixed layer whereas the d18O of the deep-dwelling species G. scitula reflects a deeper temperature at around 200 m. While the Mg/Ca data of both species is more homogeneous, their annual flux-weighted mean temperatures also closely mirror in situ annual mean temperatures below the seasonal thermocline (40–70 m) and at 200 m, respectively. We find that concentrations and fluxes of organic compounds (alkenones and crenarchaeol) and their associated temperature proxies 0 U k37 and TEX86) did not display pronounced seasonality, which may be related to lateral transport and mixing of the signal over a larger area. Yet, the organic proxies yielded flux-weighted annual mean SSTs that are in good agreement with the annual mean satellite SST, suggesting that alkenones and GDGTs are likely transported from relatively short distances. Spectral analysis hints that both the inorganic and organic proxy records may be influenced to some degree by strong currents such as eddies. This is worth investigating in future high-resolution sediment trap studies. ACKNOWLEDGEMENTS This research was supported by the Netherlands Organization for Scientific Research (NWO), through the Paleosalt (ESFEuroClimate), Gateways, Netherlands-Bremen Oceanography II (NEBROC) and LOCO programs. We are grateful to Santiago Gonzalez for sample preparation and to Sharyn Crayford for C/N and biogenic silica analyses. We thank Denise Dorhout and Wim Boer for d18O and Mg/Ca analyses as well as Michiel Kienhuis and Evaline van Weerlee for support with mass spectrometry. Jort Ossebaar, Marianne Baas and Ellen Hopmans are thanked for their assistance with the organic geochemical analyses. We thank editor Josef Werne and three anonymous reviewers for their constructive criticism that substantially improved the paper. We also acknowledge all technical support during research cruises onboard the RRV Charles Darwin, RRV Discovery, FS Meteor and RV Pelagia.

APPENDIX A. SUPPLEMENTARY DATA Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.gca.2011.08.009. REFERENCES Anand P., Elderfield H. and Conte M. H. (2003) Calibration of Mg/Ca thermometry in planktonic foraminifera from a sediment trap time series. Paleoceanography 18, 20–22. Backeberg B. C. and Reason C. J. C. (2010) A connection between the South Equatorial Current north of Madagascar and Mozambique Channel Eddies. Geophys. Res. Lett. 37, L04604. Barker S., Greaves M. and Elderfield H. (2003) A study of cleaning procedures used for foraminiferal Mg/Ca paleothermometry. Geochem. Geophys. Geosyst. 4, 20–22. Be´ A. W. H. and Hutson W. H. (1977) Ecology of planktonic foraminifera and biogeographic patterns of life and fossil assemblages in the Indian Ocean. Micropaleontology 23, 369– 414.

Bemis B. E., Spero H. J., Bijma J. and Lea D. W. (1998) Reevaluation of the oxygen isotopic composition of planktonic foraminifera: experimental results and revised paleotemperature equations. Paleoceanography 13, 150–160. Benthien A. and Mu¨ller P. J. (2000) Anomalously low alkenone temperatures caused by lateral particle and sediment transport in the Malvinas Current region, western Argentine Basin. Deep Sea Res. Part I Oceanogr. Res. Pap. 47, 2369–2393. Berelson W. M. (2002) Particle settling rates increase with depth in the ocean. Deep Sea Res. Part II Top. Stud. Oceanogr. 49, 237– 251. Bienfang P. K. (1981) Sinking rates of heterogeneous, temperate phytoplankton populations. J. Plankton Res. 3, 235–253. Bonnin J., van Raaphorst W., Brummer G. J., van Haren H. and Malschaert H. (2002) Intense mid-slope resuspension of particulate matter in the Faeroe-Shetland Channel: short-term deployment of near-bottom sediment traps. Deep Sea Res. Part I Oceanogr. Res. Pap. 49, 1485–1505. Brady H. B. (1882) Report on the Foraminifers. Exploration of the faro¨e Channel during the summer of 1880, in Her Majesty’s ship Knight Errant, with subsidiary reports. Proc. R. Soc. 11, 708–717. Brassell S. C., Eglinton G., Marlowe I. T., Pflaumann U. and Sarnthein M. (1986) Molecular stratigraphy – a new tool for climatic assessment. Nature 320, 129–133. Conte M. H., Sicre M. A., Ruhlemann C., Weber J. C., Schulte S., Schulz-Bull D. E. and Blanz T. (2006) Global temperature calibration of the alkenone unsaturation index (U-37(K’)) in surface waters and comparison with surface sediments. Geochem. Geophys. Geosyst. 7. de la Rocha C. L. and Passow U. (2007) Factors influencing the sinking of POC and the efficiency of the biological carbon pump. Deep Sea Res. Part II Top. Stud. Oceanogr. 54, 639–658. de Ruijter W. P. M., Ridderinkhof H., Lutjeharms J. R. E., Schouten M. W. and Veth C. (2002) Observations of the flow in the Mozambique Channel. Geophys. Res. Lett. 29. de Ruijter W. P. M., van Aken H. M., Beier E. J., Lutjeharms J. R. E., Matano R. P. and Schouten M. W. (2004) Eddies and dipoles around South Madagascar: formation, pathways and large-scale impact. Deep Sea Res. Part I Oceanogr. Res. Pap. 51, 383–400. Dekens P. S., Lea D. W., Pak D. K. and Spero H. J. (2002) Core-top calibration of Mg/Ca in tropical foraminifera: Refining paleotemperature estimation. Geochem. Geophys. Geosyst. 3. Donohue K. A. and Toole J. M. (2003) A near-synoptic survey of the Southwest Indian Ocean. Deep Sea Res. Part II Top. Stud. Oceanogr. 50, 1893–1931. d‘Orbigny, A.d. (1839). Foraminife`res; In Histoire physique, politique et naturelle de l‘ Ile de Cuba 8, (ed. R. de la Sagra). Paris, pp. 1-224. Elderfield H. (2002) Climate change: carbonate mysteries. Science 296, 1618. Emiliani C. (1955) Pleistocene temperatures. J Geol. 63, 538–578. Epstein S., Buchsbaum R., Lowenstam H. A. and Urey H. C. (1953) Revised carbonate-water isotopic temperature scale. Geol. Soc. Am. Bull. 64, 1315–1326. Erez J. and Luz B. (1983) Experimental paleotemperature equation for planktonic foraminifera. Geochim. Cosmochim. Acta 47, 1025–1031. Fairbanks R. G., Sverdlove M., Free R., Wiebe P. H. and Be´ A. W. H. (1982) Vertical distribution and isotopic fractionation of living planktonic foraminifera from the Panama Basin. Nature 298, 841–844. Fallet U., Boer W., van Assen C., Greaves M. and Brummer G. J. A. (2009) A novel application of wet oxidation to retrieve

Foraminiferal and organic paleotemperature proxies in sedimenting particles carbonates from large organic-rich samples for ocean-climate research. Geochem. Geophys. Geosyst. 10, 20–22. Fallet U., Brummer G. J., Zinke J., Vogels S. and Ridderinkhof H. (2010) Contrasting seasonal fluxes of planktonic foraminifera and impacts on paleothermometry in the Mozambique Channel upstream of the Agulhas Current. Paleoceanography 25, A4223. Fowler S. W. and Knauer G. A. (1986) Role of large particles in the transport of elements and organic compounds through the oceanic water column. Prog. Oceanogr. 16, 147–194. Gyldenfeldt A. B., Carstens J. R. and Meincke J. (2000) Estimation of the catchment area of a sediment trap by means of current meters and foraminiferal tests. Deep Sea Res. Part II: Top. Stud. Oceanogr. 47, 1701–1717. Hamanaka J., Sawada K. and Tanoue E. (2000) Production rates of C-37 alkenones determined by C-13-labeling technique in the euphotic zone of Sagami Bay, Japan. Org. Geochem. 31, 1095– 1102. Harlander U., Ridderinkhof H., Schouten M. W. and de Ruijter W. P. M. (2009) Long-term observations of transport, eddies, and Rossby waves in the Mozambique Channel. J. Geophys. Res. -Oceans 114, 20–22. Hemleben C., Spindler M. and Anderson O. R. (1989) Modern Planktonic Foraminifera. Springer Verlag, New York. Hopmans E. C., Weijers J. W. H., Schefuss E., Herfort L., Sinninghe Damste´ J. S. S. and Schouten S. (2004) A novel proxy for terrestrial organic matter in sediments based on branched and isoprenoid tetraether lipids. Earth Planet. Sci. Lett. 224, 107–116. Horibe Y. and Oba T. (1972) Temperature scales of aragonitewater and calcite–water systems (in Japanese). pp. 69–79. Huguet C., Schimmelmann A., Thunell R., Lourens L. J., Sinninghe Damsth J. S. and Schouten S. (2007) A study of the TEX86 paleothermometer in the water column and sediments of the Santa Barbara Basin, California. Paleoceanography 22, A3203. Kahn M. I. and Williams D. F. (1981) Oxygen and carbon isotopic composition of living planktonic foraminifera from the Northeast Pacific Ocean. Palaeogeogr. Palaeoclimatol. Palaeoecol. 33, 47–69. Kim J. H., van der Meer J., Schouten S., Helmke P., Willmott V., Sangiorgi F., Kocß N., Hopmans E. C. and Damste´ J. S. S. (2010) New indices and calibrations derived from the distribution of crenarchaeal isoprenoid tetraether lipids: implications for past sea surface temperature reconstructions. Geochim. Cosmochim. Acta 74, 4639–4654. Lee K. E., Kim J. H., Wilke I., Helmke P. and Schouten S. (2008) A study of the alkenone, TEX86, and planktonic foraminifera in the Benguela Upwelling System: implications for past sea surface temperature estimates. Geochem. Geophys. Geosyst. 9, Q10019. Loncaric N., Brummer G. J. A. and Kroon D. (2005) Lunar cycles and seasonal variations in deposition fluxes of planktic foraminiferal shell carbonate to the deep South Atlantic (central Walvis Ridge). Deep Sea Res. Part I Oceanogr. Res. Pap. 52, 1178–1188. Loncaric N., van Iperen J., Kroon D. and Brummer G. J. A. (2007) Seasonal export and sediment preservation of diatomaceous, foraminiferal and organic matter mass fluxes in a trophic gradient across the SE Atlantic. Prog. Oceanogr. 73, 27–59. Lopes dos Santos R. A., Prange M., Castaneda I. S., Schefuss E., Mulitza S., Schulz M., Niedermeyer E. M., Sinninghe Damste´ J. S. S. and Schouten S. (2010) Glacial-interglacial variability in Atlantic meridional overturning circulation and thermocline adjustments in the tropical North Atlantic. Earth Planet. Sci. Lett. 300, 407–414.

5847

Mercer J. L., Zhao M. X. and Colman S. M. (2005) Seasonal variations of alkenones and Uk’37 in the Chesapeake Bay water column. Estuar Coast Shelf Sci. 63, 675–682. Mollenhauer G., Inthorn M., Vogt T., Zabel M., Sinninghe Damste´ J. S. S. and Eglinton T. I. (2007) Aging of marine organic matter during cross-shelf lateral transport in the Benguela upwelling system revealed by compound-specific radiocarbon dating. Geochem. Geophys. Geosyst. 8, Q09004. Mu¨ller P. J., Cepek M., Ruhland G. and Schneider R. R. (1997) Alkenone and coccolithophorid species changes in late Quaternary sediments from the Walvis Ridge: Implications for the alkenone paleotemperature method. Palaeogeogr. Palaeoclimatol. Palaeoecol. 135, 71–96. Mu¨ller P. J. and Fischer G. (2001) A 4-year sediment trap record of alkenones from the filamentous upwelling region off Cape Blanc, NW Africa and a comparison with distributions in underlying sediments. Deep Sea Res. Part I Oceanogr. Res. Pap. 48, 1877–1903. Mu¨ller P. J., Kirst G., Ruhland G., von Storch I. and Rosell-Mele A. (1998) Calibration of the alkenone paleotemperature index U-37(K’) based on core-tops from the eastern South Atlantic and the global ocean (60 degrees N-60 degrees S). Geochim. Cosmochim. Acta 62, 1757–1772. Nu¨rnberg D., Bijma J. and Hemleben Ch. (1996) Assessing the reliability of magnesium in foraminiferal calcite as a proxy for water mass temperatures. Geochim. Cosmochim. Acta 60, 803– 814. Ohkouchi N., Eglinton T. I., Keigwin L. D. and Hayes J. M. (2002) Spatial and temporal offsets between proxy records in a sediment drift. Science 298, 1224–1227. Palastanga V., van Leeuwen P. J. and de Ruijter W. P. M. (2006) A link between low-frequency mesoscale eddy variability around Madagascar and the large-scale Indian Ocean variability. J. Geophys. Res. -Oceans 111, 20–22. Peeters, F.J.C., and Brummer, G.J.A. (2002) The seasonal distribution of living planktic foraminfera in the NW Arabian Sea. In The tectonic and climatic evolution of the Arabian Sea region, The Geological Society London. vol. 195, pp. 463-497. Prahl F. G., Mu¨hlhausen L. A. and Zahnle D. L. (1988) Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions. Geochim. Cosmochim. Acta 52, 2303– 2310. Ridderinkhof H., van der Werf P. M., Ullgren J. E., van Aken H. M., van Leeuwen P. J. and de Ruijter W. P. M. (2010) Seasonal and interannual variability in the Mozambique Channel from moored current observations. J. Geophys. Res. 115, C06010. Romero O. E., Boeckel B., Donner B., Lavik G., Fischer G. and Wefer G. (2002) Seasonal productivity dynamics in the pelagic central Benguela System inferred from the flux of carbonate and silicate organisms. J. Mar. Syst. 37, 259–278. Ru¨hlemann C. and Butzin M. (2006) Alkenone temperature anomalies in the Brazil-Malvinas confluence area caused by lateral advection of suspended particulate material. Geochem. Geophys. Geosyst. 7, 20–22. Sadekov A., Eggins S. M., de Deckker P. and Kroon D. (2008) Uncertainties in seawater thermometry deriving from intratest and intertest Mg/Ca variability in Globigerinoides ruber. Paleoceanography 23. Saher M. H., Rostek F., Jung S. J. A., Bard E., Schneider R. R., Greaves M., Ganssen G. M., Elderfield H. and Kroon D. (2009) Western Arabian Sea SST during the penultimate interglacial: a comparison of U37’K; and Mg/Ca paleothermometry. Paleoceanography 24, A2212. Sautter L. R. and Thunell R. C. (1991) Planktonic foraminiferal response to upwelling and seasonal hydrographic conditions –

5848

U. Fallet et al. / Geochimica et Cosmochimica Acta 75 (2011) 5834–5848

sediment trap results from San-Pedro Basin, Southern California Bight. J. Foraminiferal Res. 21, 347–363. Schouten M. W., de Ruijter W. P. M., van Leeuwen P. J. and Ridderinkhof H. (2003) Eddies and variability in the Mozambique Channel. Deep Sea Res. Part II Top. Stud. Oceanogr. 50, 1987–2003. Schouten S., Hopmans E. C., Schefuss E. and Sinninghe Damste´ J. S. S. (2002) Distributional variations in marine crenarchaeotal membrane lipids: a new tool for reconstructing ancient sea water temperatures? Earth Planet. Sci. Lett. 204, 265–274. Schouten S., Huguet C., Hopmans E. C., Kienhuis M. V. M. and Sinninghe Damste´ J. S. S. (2007) Analytical methodology for TEX86 paleothermometry by high-performance liquid chromatography/atmospheric pressure chemical ionization-mass spectrometry. Anal. Chem. 79, 2940–2944. Shackleton N. J. and Vincent E. (1978) Oxygen and carbon isotope studies in recent foraminifera from the Southwest Indian Ocean. Mar. Micropaleontol. 3, 1–13. Sicre M. A., Labeyrie L. D., Ezat U., Duprat J., Turon J. L., Schmidt S., Michel E. and Mazaud A. (2005) Mid-latitude Southern Indian Ocean response to Northern Hemisphere Heinrich events. Earth Planet. Sci. Lett. 240, 724–731. Sicre M. A., Ternois Y., Miquel J. C. and Marty J. C. (1999) Alkenones in the Northwestern Mediterranean Sea: interannual variability and vertical transfer. Geophy. Res. Lett. 26, 1735– 1738.

Sinninghe Damste´ J. S. S., Schouten S., Hopmans E. C., van Duin A. C. T. and Geenevasen J. A. J. (2002) Crenarchaeol: the characteristic core glycerol dibiphytanyl glycerol tetraether membrane lipid of cosmopolitan pelagic crenarchaeota. J. Lipid Res. 43, 1641–1651. Spero H. J., Bijma J., Lea D. W. and Bemis B. E. (1997) Effect of seawater carbonate concentration on foraminiferal carbon and oxygen isotopes. Nature 390, 497–500. Takahashi K. and Be´ A. W. H. (1984) Planktonic foraminifera – factors controlling sinking speeds. Deep Sea Res. A 31, 1477– 1500. Thomsen C., Schulz-Bull D. E., Petrick G. and Duinker J. C. (1998) Seasonal variability of the long-chain alkenone flux and the effect on the U-37(k0 )-index in the Norwegian Sea. Org. Geochem. 28, 311–323. Welch P. (1967) The use of fast Fourier transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms. IEEE Trans. Audio Electroacoust. 15, 70–73. Wuchter C., Schouten S., Wakeham S. G. and Sinninghe Damste´ J. S. S. (2006) Archaeal tetraether membrane lipid fluxes in the northeastern Pacific and the Arabian Sea: Implications for TEX86 paleothermometry. Paleoceanography 21. Associate editor: Josef Werne