Sedimentation and burial of organic and inorganic temperature proxies in the Mozambique Channel, SW Indian Ocean

Deep-Sea Research I 59 (2012) 37–53

Contents lists available at SciVerse ScienceDirect

Deep-Sea Research I journal homepage: www.elsevier.com/locate/dsri

Sedimentation and burial of organic and inorganic temperature proxies in the Mozambique Channel, SW Indian Ocean ˜ eda b,1, Aneurin Henry-Edwards a, Thomas O. Richter a, Wim Boer a, Ulrike Fallet a,n, Isla S. Castan b Stefan Schouten , Geert-Jan Brummer a,c a

NIOZ, Royal Netherlands Institute for Sea Research, Department of Marine Geology, P.O. Box 59, NL-1790 AB Den Burg, The Netherlands NIOZ, Royal Netherlands Institute for Sea Research, Department of Marine Organic Biogeochemistry, P.O. Box 59, NL-1790 AB Den Burg, The Netherlands c Faculty of Earth- and Life Sciences, VU University Amsterdam, de Boelelaan 1085, 1081 HV Amsterdam, The Netherlands b

a r t i c l e i n f o

abstract

Article history: Received 15 December 2010 Received in revised form 12 October 2011 Accepted 13 October 2011 Available online 6 November 2011

Paleoceanographic studies strongly rely on proxies to reconstruct past environmental conditions. However, several factors influence the reliability of these proxies, particularly during sedimentation and burial. In this study, we measured both inorganic (d18O and Mg/Ca on three species of planktonic foraminifera 0 (Globigerinoides ruber, Globigerinoides trilobus and Neogloboquadrina dutertrei) and organic (U k37 and TEX86) sea surface temperature (SST) proxies on core top sediment taken from an east-west transect across the Mozambique Channel. We contrast our findings with previously published modern time-series temperature proxy data from a sediment trap moored in the mid-channel. The coretop sediment was analyzed for 14C, 210 Pb and excess (xs) 234Th. While 234Thxs data indicate a flux of fresh particulate matter to the bottom sediment, radiocarbon dating shows that the core top sediments are composed of material that is on average about 1000 years old. The fine organic carbon is consistently (even though only slightly) younger than the coarser foraminiferal calcite, which is likely caused by preferential downcore mixing of the fine fraction. Besides vertical mixing by bioturbation, stable lead isotope ratios from the time-series particle fluxes indicate episodic lateral transport of old particles from the shelf to the deep Mozambique Channel as an additional source of pre-aged material in core tops. Core top temperature proxies show warmer values close to the channel flanks while colder values are found in the mid-channel. These could be associated with higher maximum summer temperatures in modern coastal waters in contrast to the mid-channel. Additionally, we find an offset in all foraminiferal proxies between core top samples and time-series data that corresponds to 1–3 1C, which probably reflects climate variability over the past  2000 years. However, this temperature 0 difference is not observed in the organic proxies U k37 and TEX86, which may result from current transport of unconsolidated organic matter that can homogenize the organic proxy signal prior to burial. & 2011 Elsevier Ltd. All rights reserved.

Keywords: Mozambique Channel SW Indian Ocean Temperature proxies Biomarkers 0 U k37 TEX86 Foraminifera d18O Mg/Ca Radionuclides (234Th, 210Pb, Sediment Bioturbative mixing Sediment trap

14

C)

1. Introduction In order to understand past ocean and climate systems, proxies are needed that accurately record environmental conditions (Elderfield, 2002; Lea, 2003). However, many factors can influence proxy signals during sedimentation and burial. One important aspect is the contribution of pre-aged material to the sedimentary record (Gardner and Sullivan, 1981; Laine et al., 1994) and the timeaveraging effect of bioturbation (Keigwin and Guilderson, 2009). Proxy parameters from core top sediments are frequently used for calibration purposes to modern oceanographic parameters in the absence of sediment trap series (Dekens et al., 2002; Kim et al., 2008;

n

Corresponding author. E-mail address: [email protected] (U. Fallet). 1 Present address: Department of Geosciences, University of Massachusetts Amherst, 611 North Pleasant Street, 233 Morrill Science Center, Amherst, MA 01003, USA. 0967-0637/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.dsr.2011.10.002

¨ Muller et al., 1998). Yet, calibration of pre-aged allochthonous and bioturbated sediment proxies to modern SST can potentially introduce uncertainties or biases. Furthermore, sedimentary records containing variable contributions of pre-aged material can result in pooled proxies that do not reflect the environmental conditions at ¨ their original place of deposition (Benthien and Muller, 2000). Upward bioturbation of older sediment from the deeper core and downward bioturbation of fresh material from the top play an important role in time-averaging proxy-temperature in sediments. Local variability in the intensity of bioturbation in deepsea sediment may be as important as regional differences in sediment properties (de Master and Cochran, 1982). Bioturbation of sediment usually reaches as deep as 10 cm (Boudreau, 1998), which implies a potentially substantial age bias in proxy signals particularly in deep sea sediments where sedimentation rates are usually low. Other studies concluded that mixing depth is not constant on a global scale, but varies as a predictable function of particulate organic carbon flux (Smith and Rabouille, 2002;

38

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

Trauth et al., 1997). Finally, Bard (2001) argues that differential mixing between fine and coarse fractions can lead to age offsets for a given sediment horizon, potentially decoupling foraminiferal and organic proxy records. The relative importance of present-day fluxes of fresh and preaged (transported) material together with downcore mixing of sediment can be assessed using a suite of radionuclides with different half-lives (234Th, 210Pb, 14C) and stable lead isotope ratios (206Pb/207Pb). 234Th has the shortest half-life of 24.3 day which makes it the ideal tracer for short-time processes on time scales of about three months (Pope et al., 1996). In contrast, the half-lives of 210Pb (22.3 years) and 14C (5730 years) are longer and can be used for establishing sedimentation processes and age models for the past hundred years for 210Pb (Masque et al., 2003) and for the past 50,000 years for 14C (Fairbanks et al., 2005; Hughen et al., 2004). Stable lead isotope ratios (206Pb/207Pb) trace natural and anthropogenic sources of Pb, with anthropogenic (pollutant) lead generally displaying lower 206Pb/207Pb (Reuer and Weiss, 2002; Weiss et al., 1999). Hence, higher 206Pb/207Pb reflects input of older pre-industrial material. With radionuclides and stable lead isotope ratios modern and past sedimentation processes can thus be reconstructed. Previously, we reported on the behavior of planktonic forami0 niferal d18O and Mg/Ca proxies and organic U k37 and TEX86 (Fallet et al., 2010; Fallet et al., 2011). We found that in the Mozambique Channel maximum fluxes of Globigerinoides. ruber occur in late austral summer (February–March) when SST ranges between 28.7 and 30.2 1C. By contrast, Globigerinoides trilobus maxima appear in early winter (June–July) at a lower SST between 25.3 and 27.0 1C. Despite this seasonality in shell fluxes, flux-weighted means for both species closely reflect annual mean SST within a margin of þ0.5 1C for G. ruber and  0.3 1C for G. trilobus. In contrast, the 0 organic SST proxies U k37 and TEX86 from the same sediment trap display no or only moderate seasonality. Yet, the flux weighted means of the associated temperature signatures also closely reflect 0 mean annual SST with 28.370.3 and 28.170.2 1C for U k37 and H TEX 86 , respectively. In the study presented here, we analyzed core top sediment taken from an east–west transect across the Mozambique Channel for the same foraminiferal and organic temperature proxies. Together with the radionuclide data, we examine the burial history of these proxies and assess the importance of sediment transport and vertical mixing by bioturbation.

2. Study location The Mozambique Channel forms a suitable area to study the effect of sedimentation and burial on temperature proxies as particle fluxes have been intensively monitored over the last years (Fallet et al., 2010; Fallet et al., 2011). The Mozambique Channel is characterized by a relatively narrow shelf on the African side of only 15 km width and a broader shelf area on the Madagascar side that is on average 100 km wide. Both continental slopes are steep with an average gradient of about 1.61 (Fig. 1). Currently, research focuses on the Mozambique Channel as it contributes more than 30% of the volume transport to the Agulhas Current (van der Werf et al., 2009) and therefore significantly influences the Indian–Atlantic exchange (Beal et al., 2011; de Ruijter et al., 1999). The variability of modern circulation in the Mozambique Channel is well known because it is monitored in the Long-term Ocean Climate Observation (LOCO) program since 2003 (Palastanga et al., 2006). Studies within this program have shown that the strong bottom water currents (Harlander et al., 2009) and presence of bottom water oxygen make it a good locality to study the potential effects of sediment transport and bioturbation.

3. Material and methods 3.1. Sampling and preparation of core top sediment Across the narrowest transect of the Mozambique Channel, we took an ensemble of multi-cores during scientific cruises in 2003 (CD153), 2006 (D301), 2008 (M75-1b) and 2009 (64PE304) (Fig. 1a). The shallowest multi-core analyzed for foraminiferal temperature proxies was taken at 1500 m water depth, which is only 20 km from the shoreline because of the narrow African shelf. For foraminiferal shell weights, d18O and Mg/Ca analysis we used core tops obtained during cruises CD153 (Fig. 1a and b, crosses) and D301 (Fig. 1a and b, squares). Analysis of organic 0 U k37 and TEX86 was performed on core top material from cruises M75-1b and 64PE304 (Fig. 1a and b, rhombi). The multi-cores from CD153, D301 and 64PE304 were sliced into 0–1 cm slices and multi-cores from M75-1b into 0–0.25 cm and 0.25–0.5 cm slices. For proxy comparison across the channel, we averaged all measured proxy values from the same coring site. Core top sediments were wet-split using a binary Folsom splitter with a precision of 495% (Loncaric et al., 2007; Sell and Evans, 1982). Reconstructed core top proxy temperatures were subsequently compared with modern satellite SST and to proxy temperatures from a sediment trap deployed between November 2003 and March 2006 in the mid-channel (Fallet et al., 2010; Fallet et al., 2011). Flux-weighted proxy temperatures were calculated by multiplying each proxy value with its corresponding flux value, averaging the results and then dividing it by the total flux.

3.2. Satellite SST, in situ salinity and d18Ow For the SST data, we obtained monthly averaged re-analyzed Reynolds SST data from the Giovanni website (http://gdata1.sci. gsfc.nasa.gov/daac-bin/G3/gui.cgi?instance_id=ocean_month, Acker and Leptoukh, 2007). We used the longest available time-series data ranging from January 1999 to October 2010 for coastal waters (16.3–16.41S, 39.8–39.91E) and the mid-channel (16.3–16.41S, 42.0–42.11E, Fig. 2a). The temperature data are from the 11 mm daytime MODIS Aqua satellite, where 11 mm refers to the wavelength of the infrared emission used for SST estimation. We also produced maps for monthly averaged SST data for austral summer (February 2004), fall (May 2004), winter (August 2004) and spring (November 2004) from the Giovanni database (Fig. 2b). For the SST areal maps we also used the 11 mm daytime MODIS Aqua data with 9 km resolution for the area 351E–501E and 101S–201S. Salinity was measured with conductivity–temperature–depth sensors (CTDs) in situ at the trapsite (40.81E, 16.41S) during scientific cruises to the Mozambique Channel in November 2003 and March 2006 (Fig. 3a). We also obtained monthly averaged satellite sea surface salinity maps from the IRI/LDEO climate library (http://iridl.ldeo.columbia.edu/SOURCES/.CARTON-GIESE/ .SODA/.v2p0p2-4/figviewer.html) for the area 351E–501E and 101S–201S (Fig. 3b). We used representative figures for the four seasons of austral summer (February 2004), fall (May 2004), winter (August 2004) and spring (November 2004). For details regarding sample preparation and analysis of surface seawater d18O from the Mozambique Channel, see Fallet et al. (2010). From the global seawater databases (http://cchdo.ucsd. edu/ and http://data.giss.nasa.gov/o18data/), we generated a d18Ow–salinity calibration for the SW Indian Ocean (Supplementary Fig. S1) using d18Ow data from 5–100 m water depth in the area 10–401S and 30–801E. This depth range matches (sub-) surface waters in the Indian Ocean with salinities of between 34.8 and 35.8 psu.

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

Juan de Nova

Mozambique

39

Madagascar

depth [m]

0

0

500

500

1000

1000

1500

1500 sediment trap

2000

2000

CD153 (forams) D301 (forams) M75-1b (organics) 64PE304 (organics)

2500

Davie Ridge

2500

3000

3000 39.5

40

40.5

41

41.5 42 longitude [°E]

42.5

43

43.5

Fig. 1. (a) Study area with schematic current pattern illustrating the fast rotating eddies that migrate through the channel every about 70 days before joining the Agulhas Current. The inset shows the location and depth of the multicores across the Mozambique Channel in greater detail. (b) Bathymetry of the Mozambique Channel with core locations and trap mooring (November 2003–March 2006, inverted triangle), multi-cores taken during cruise CD153 (crosses) and D301 (squares) were used for 0 foraminiferal d18O- and Mg/Ca-analyses and multi-cores taken during cruises M75 (rhombi) were used for organic matter U k37 and TEX86 analyses. Location of the sediment trap is indicated by the triangle.

40

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

32 31 30

SST [°C]

29 28 27 26 25 24

coastal SST mid-channel SST 18/11/10

22/01/10

28/03/09

06/08/07

01/06/08

10/10/06

14/12/05

17/02/05

23/04/04

28/06/03

01/09/02

05/11/01

15/03/00

09/01/01

20/05/99

23

date

Fig. 2. (a) Time-series SST data with monthly and 0.11 grid resolution across the Mozambique Channel for coastal waters (16.3–16.41S, 39.8–39.91E) and the mid-channel (16.3–16.41S, 42.0–42.11E). Note that on the shelves there is often a larger SST gradient than in the open ocean. (b) Maps of monthly averaged SST data for summer (upper left), fall (upper right), winter (lower left) and spring (lower right) for the area 351E–501E and 101S–201S. The color bar is adjusted for SSTs from 23–30 1C. Note that the wider Madagascar shelf is often warmer than the smaller Mozambique shelf resolved even in monthly averaged SST data.

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

41

δ O = 0.4 ‰

δ O = 0.15 ‰

salinity [psu] 34.6

34.8

35

35.2

35.4

0

0

100

100

200

200

300

300

400

400

500

depth [m]

depth [m]

δ O = 0.45 ‰

500 salinity (end rain season) salinity (end dry season)

600

600 34.6

34.8

35 salinity [psu]

35.2

35.4

Fig. 3. In situ salinity profiles from the end of the rain (March 2006) and dry season (November 2003) measured close to the trapsite in the middle of the Mozambique Channel. A 0.4 psu increase in salinity is expressed as a 0.3% increase in d18Ow using the salinity–d18Ow equation generated from surface water data (5–100 m) in the SW Indian Ocean (Fig. S1). (b) Maps of monthly averaged SSS data for summer (upper left), fall (upper right), winter (lower left) and spring (lower right) for the area 351E–501E and 101S–201S. The color bar is adjusted for SSSs ranging from 33–36 g kg  1.

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

3.3. Age of core top sediment and age model of sediment core

3.4. Analysis of time-series lithogenic matter, from sediment traps

210

Pb and

206/207

8000

7000

6000

5000

4000

2000

1000

0

10

20

outlier

0

tot

500

0

210 Pb

[Bq/kg] 4000

30

forams TOC

3900

14C-age

3800

14C-age

1000

cumulative dry mass [g cm-2]

outlier

trap A (250 mab) trap B (2 mab)

10 core depth [cm]

The age of surface sediments in the Mozambique Channel was assessed using three different techniques: 234Thxs, 210Pb and 14C. We applied 14C dating to planktonic foraminiferal calcite (mixed G. trilobus and Globigerinoides sacculifer from the 4315 mm fraction) and total organic carbon (TOC) from multicore D301–MC2 at 2311 m water depth taken close to the trap site (40.841E, 16.701S), (Fig. 4a, Table 1). The 14C-dating was performed at the Poznan Radiocarbon Laboratory using Accelerator Mass Spectrometry (AMS). For conversion of the 14C dates to calendar ages the program CALIB 6.0 (http://intcal.qub.ac.uk/calib/, Stuiver and Reimer, 1993) and the Marine09 calibration dataset (Reimer et al., 2009) were used. A local reservoir age of 433757 years was applied, as given for the region around the island Mayotte (Southon et al., 2002), the closest location to our study area for which this information is available. Sedimentation rates were calculated by linear regression excluding outliers. Core top samples CD153-MC3 and D301-MC2 were taken close to the sediment trap site in 2003 and 2006 for analysis of 234 Thxs- and 210Pb-activity, respectively. Both core top samples were filtered over a 0.45 mm polycarbonate filter, subsequently frozen, freeze-dried, weighed and ground into a homogeneous powder. To assess deposition of fresh ( o3 months) material, the 234 Th activity was measured by counting the 63 and 92 keV gamma emissions with a Canberra low energy germanium detector. Calibration was achieved by means of the external standard method using a uranium ore diluted with silica powder in which uranium was at secular equilibrium with all its daughters. Samples were analyzed for 210Pb by means of a-spectrometry indirectly via its granddaughter 210Po, using 10 mg of material for sediment trap samples, and 200mg for core top sediments (Fig. 4b). Sediment was spiked with 209Po and leached with concentrated HCl. For further details about sample preparation and analyses of 210Pb and 234Th see Boer et al. (2006). Using the 14 C sedimentation rate (Fig. 4a), the 210Pb data were fitted with a mixing model (Soetaert et al., 1996) to estimate a 210Pb mixing rate and mixing depth (Fig. 4b). We also examined stable lead isotope ratios (206Pb/207Pb), which can be used to trace natural and anthropogenic sources of Pb. For analysis of the bulk sediment 206Pb/207Pb ratio, we used the method described by Richter et al. (2009).

0

calibrated 14C-age [years] 3000

42

20

Pb 30

For comparison, we contrast the core top sediment results with time-series data from two consecutively moored sediment traps deployed between November 2003 and March 2006 (Fig. 5, Table 2), which we have previously studied (Fallet et al., 2010, Fallet et al., 2011). These Technicap PPS 5 sediment traps were positioned at 16.421S and 40.851E in the central Mozambique Channel at 2000 and 2248 m water depth, 250 m (‘‘upper trap’’) and 2 m (‘‘lower trap’’) above the channel floor, respectively (Fallet et al., 2010). However, the lower trap did not function during sampling period 1 (November 2003–March 2005), which restricts our dataset to sampling period 2 (March 2005–March 2006). Time-series samples were taken from a 0.45 mm polycarbonate filter and subsequently frozen, freeze-dried, weighed and ground the sample into a homogeneous powder. Samples were analyzed for carbon, nitrogen and opaline silica separately before and after removal of the carbonate-carbon with 2 M ultrapure hydrochloric acid applied directly on to the sample as described by 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)

40 Fig. 4. (a) Age reconstruction of core D301-MC2 from radiocarbon analysis of foraminiferal calcite (G. trilobus/G. sacculifer in the 4315 mm fraction—‘‘forams’’, rhomboids) and total organic carbon (‘‘TOC’’, circles) with error bars. Sedimentation rates of derived from dating of foraminifera is 4.2 cm/1000 years and 5.0 cm/1000 years for TOC. Calculation of cumulative dry mass, which accounts for compaction of the sediment, indicates that sedimentation was relatively stable over the past  8000 years. (b) Profile of 210Pb for the core from the trapsite at 40.841E, 16.81S and 2250 m water depth. For comparison, average 210Pb from the upper (250 m above bottom/ mab) and bottom (2 mab) sediment trap are shown. Note that 210Pb values of the bottom trap are slightly higher than those from the upper trap, which is probably caused by the increased lithogenic matter input in comparison with the upper trap (also see Fig. 5).

by OM¼Corgn2 (Romero et al., 2002). Opaline silica was obtained from about 14 mg of bulk material using the continuous alkaline leaching technique (Koning et al., 2002). Lithogenic matter content

Table 1 Time-series 210Pb, total mass flux, opaline silica (BSi), organic matter (OM), calcium carbonate (CaCO3), lithogenic matter (lith) and the ratio between central Mozambique Channel. Sample

Start date sampling

Upper trap 210

23-11-03 14-12-03 04-01-04 25-01-04 15-02-04 07-03-04 28-03-04 18-04-04 09-05-04 30-05-04 20-06-04 11-07-04 01-08-04 22-08-04 12-09-04 03-10-04 24-10-04 14-11-04 05-12-04 26-12-04 16-01-05 06-02-05 27-02-05 09-03-05 01-04-05 24-04-05 17-05-05 09-06-05 02-07-05 25-07-05 17-08-05 09-09-05 02-10-05 25-10-05 17-11-05 10-12-05 02-01-06 25-01-06 17-02-06 12-03-06

Pb/207Pb for the upper (250 mab) and lower trap (2 mab) moored in the

Lower trap

Pb (Bq/kg)

Error (Bq/kg)

Total mass flux (mg m  2 day  1)

BSiO (wt%)

OM (wt%)

CaCO3 (wt%)

Lith (wt%)

206

Ratio Pb/207Pb

210 Pb (Bq/kg)

Error (Bq/kg)

Total mass flux (mg m  2 day-1)

BSiO (wt%)

OM (wt%)

CaCO3 (wt%)

Lith (wt%)

3786 3704 3393 3925 3756 1975 3528 4661 – 3099 3494 4003 4292 3413 3226 3550 3098 3184 – 2984 2973 – – 3474 3821 3464 5413 4895 4809 4198 4082 4743 4162 3235 4207 4261 4372 4750 3936 –

196 149 150 171 164 97 160 189 – 136 151 166 171 97 93 101 95 96 – 52 54 – – 103 114 101 145 139 135 119 63 72 65 55 66 69 69 73 62 –

89.16 47.29 83.88 120.10 78.79 566.45 97.11 58.83 27.25 158.06 592.07 82.88 103.33 166.27 167.24 104.44 82.53 39.27 57.02 78.46 32.43 24.78 70.54 39.95 102.93 100.70 122.15 112.21 75.70 63.57 96.28 121.87 81.72 85.42 156.96 112.20 137.10 102.16 89.46 140.06

12 11 11 11 10 6 7 8 5 7 6 9 11 12 12 11 12 10 11 10 9 10 10 9 10 7 8 9 9 9 9 10 11 9 7 11 12 10 9 10

8 7 8 7 10 6 7 8 6 6 5 7 7 6 6 7 7 8 10 7 6 7 7 8 7 6 7 8 8 8 8 7 9 8 6 10 9 8 7 7

37 46 37 35 36 15 27 37 58 22 14 40 34 37 40 44 38 48 41 40 57 61 46 51 50 48 33 33 38 39 38 40 43 39 56 44 38 41 43 41

44 36 45 46 44 73 59 48 31 65 75 44 49 44 42 38 43 34 38 44 28 23 37 32 32 39 52 50 45 44 45 43 38 44 31 35 41 42 40 42

1.15 1.17 1.18 1.16 1.18 1.22 1.20 1.17 – 1.21 1.22 1.17 1.18 1.18 1.18 1.17 1.18 1.17 – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – 3282 3991 3523 4246 4349 3938 4215 4296 3545 3604 3914 4317 3658 4431 4573 3281 4234

– – – – – – – – – – – – – – – – – – – – – – – 54 66 58 71 73 68 72 70 60 61 67 75 51 58 63 48 60

– – – – – – – – – – – – – – – – – – – – – – – 244 205 346 367 273 110 305 279 281 236 212 322 260 181 251 322 96

– – – – – – – – – – – – – – – – – – – – – – – 7 8 7 6 6 6 7 7 5 6 6 7 10 8 7 9 6

– – – – – – – – – – – – – – – – – – – – – – – 5 6 5 5 6 6 5 5 4 5 6 6 7 6 5 6 6

– – – – – – – – – – – – – – – – – – – – – – – 33 41 27 26 27 38 26 34 29 29 39 33 35 35 30 19 35

– – – – – – – – – – – – – – – – – – – – – – – 54 45 62 63 61 49 62 53 62 60 49 54 48 52 58 65 53

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

MOZ1-01 MOZ1-02 MOZ1-03 MOZ1-04 MOZ1-05 MOZ1-06 MOZ1-07 MOZ1-08 MOZ1-09 MOZ1-10 MOZ1-11 MOZ1-12 MOZ1-13 MOZ1-14 MOZ1-15 MOZ1-16 MOZ1-17 MOZ1-18 MOZ1-19 MOZ1-20 MOZ1-21 MOZ1-22 MOZ1-23 MOZ2-01 MOZ2-02 MOZ2-03 MOZ2-04 MOZ2-05 MOZ2-06 MOZ2-07 MOZ2-08 MOZ2-09 MOZ2-10 MOZ2-11 MOZ2-12 MOZ2-13 MOZ2-14 MOZ2-15 MOZ2-16 MOZ2-17

206

43

44

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

was taken as the residual mass after subtracting the CaCO3, organic matter and opaline silica (Fallet et al., 2011).

Pb/

Pb

1.24 1.22

natural (unpolluted) material = old

1.20

3.5. Foraminiferal shell weights, d18O and Mg/Ca analysis

1.18 1.16 1.14

Pb [Bq kg ]

6000 5000 4000 3000 2000 1000

lithogenic matter [wt %]

80.0

high lithogenic matter

60.0 40.0 20.0

total mass flux [mg m day ]

600

high-flux events

400 200

24/03/2006

14/12/2005

05/09/2005

28/05/2005

17/02/2005

09/11/2004

01/08/2004

23/04/2004

14/01/2004

06/10/2003

0

date Fig. 5. Time-series total mass flux, lithogenic matter, 210Pb and the 206Pb/207Pb ratio for the upper (open circles) and bottom (closed squares) sediment trap at 250 and 2 m above the bottom (mab), respectively. The 206Pb/207Pb ratio is a measure for anthropogenic lead pollution in the southern hemisphere since pre-industrial times (prior to 1880 AD). Samples containing old material (ratio  1.22) where lead pollution is absent are indicated with circles.

Core top samples were wet-split and foraminiferal shells precleaned as described by Fallet et al. (2009; 2010). Shells from G. ruber, white sensu stricto (d’Orbigny, 1839), G. trilobus (Reuss, 1850) and Neogloboquadrina dutertrei (d’Orbigny, 1839) were handpicked from the 250–315 mm fraction. The species G. ruber and G. trilobus have been shown to consistently live in the surface mixed layer whereas N. dutertrei lives at the bottom of or below the surface mixed layer (Fairbanks et al., 1982; Sautter and Thunell, 1991). Each species was analyzed for shell weight (Fig. 6), d18O and Mg/Ca (Table 3). To minimize possible seasonal or size-dependant biases (e.g. Barker et al., 2003), we weighed 20 shells in the 250–315 mm fraction to the nearest microgram on a high-precision micro-balance (Section 4.5) for all paired foraminiferal d18O and Mg/Ca measurements. Before weighing, foraminifera were gently cleaned by ultra-sonication in demineralized water (3  ) and ethanol (2  ) for both core top and sediment trap specimens. After cleaning no broken shells were observed. Paired d18O and Mg/Ca measurements were subsequently carried out on the same specimens, for details on sample preparation and analysis see Fallet et al. (2010). For Mg/Ca measurements, further cleaning followed the protocol of Barker et al. (2003), which has also been applied by Anand et al. (2003). Hence temperature calibrations by the latter authors apply to our study. For G. ruber and G. trilobus we used species-specific calibrations for the SW Indian Ocean to calculate d18O- and Mg/Ca-based temperatures (Fallet et al., 2010) and contrasted these with satellite SST and SSS measured across the channel (Figs. 7–9): 18

18

SSTrub ¼ 5:11nðd Orub 2d Ow Þ þ 12:56 18

18

SSTtril ¼ 6:26nðd Otril 2d Ow Þ þ 10:33 SSTrub ¼ 2:32nMg=Carub þ 15:71 SSTtril ¼ 5:46nMg=Catril þ 7:00, where d18Orub, d18Otril, Mg/Carub and Mg/Catril are the oxygen isotope and the Mg/Ca composition of G. ruber and G. trilobus shells, respectively and d18Ow is the oxygen isotope composition of the surrounding seawater.

Table 2 Shown are the results for 234Thtotal from core CD153-MC3, which were corrected for decay between seafloor sampling and time of measurement as well as for background values (234Thxs). Avg depth 234Thtotal (Bq/kg) (cm)

234

Thxs (Bq/kg)

234 Thxs 1s error (%)

210

PB (Bq/kg)

210

Pb 1s error (Bq/kg)

14 C forams (yr BP)

Error (cal yr)

Corr. 14C forams (cal. yr)

14

C TOC (yr BP)

Error (cal yr)

Corr. 14C TOC (cal yr)

0.3 0.5 0.8 1.5 2.5 3.5 4.5 5.0 5.5 6.5 10.5 15.0 15.5 20.5 25.5 29.5

14.15 – 458.20 Constant Constant – – Constant – – – Constant – – – –

26.5 – 4.5 22.0 20.0 – – 14.0 – – – 9.5 – – – –

– 743 – 400 271 281 140 – 88 69 61 – 45 73 46 29

– 17 – 10 8 8 5 – 3 3 2 – 2 3 2 1

– 1545 – – – 2240 – – – 4840 3480 – 4160 5170 7180 8980

– 35 – – – 40 – – – 50 35 – 40 50 60 60

– 1066 – – – 1805 – – – 5092 3324 – 4190 5495 7621 9630

– 1465 – – – – – – – 2340 – – 3880 – 3430 –

– 30 – – – – – – – 30 – – 35 – 35 –

– 983 – – – – – – – 1924 – – 3803 – 3267 –

67.01 – 511.06 47.80 53.10 – – 52.35 – – – 58.20 – – – –

Additionally, the 210Pb and radiocarbon 14C of foraminifera and organic matter (TOC) from core D301-MC2 are given. We calculated calendar ages using CALIB 6.0 with an assumed reservoir age of 433 7 57 years. Age reversals are shown in italics. Sediment density ranges between 0.53 g cm  3 in the coretop (0–1cm core depth) and 0.97 g cm  3 in the lower core (29–30 cm core depth).

Aliquots of the bulk samples were freeze-dried and analyzed for 0 alkenones (to determine U k37 temperatures) and glycerol dialkyl glycerol tetraethers (GDGTs) to determine TEX H 86 temperatures (Fig. 10, Table 4). These aliquots were 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

0.07 0.03 0.06 0.08 0.05 0.04 0.06 0.01 0.00 0.05 0.56 0.56 0.59 0.58 0.49 0.56 0.66 0.57 0.60 0.57 16.60 16.65 16.77 16.82 16.88 16.88 17.00 17.10 17.17 17.22 40.38 40.61 40.88 41.07 41.30 41.48 41.93 42.48 42.76 43.03 0.01 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 0.01 2.18 2.29 1.99 2.19 2.10 2.16 2.87 2.08 2.49 3.03 0.03 0.00 0.01 0.01 0.003 0.00 0.01 0.02 0.02 0.01 3.50 3.39 3.28 3.37 3.28 3.43 3.42 3.42 3.40 3.74 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 4.77 4.57 4.11 4.56 4.35 4.30 4.76 4.27 4.55 5.29 0.02 0.02 0.01 0.02 0.02 0.02 0.02 0.03 0.01 0.01  1.10  0.62  0.27  0.66  0.95  0.77  1.21  0.64  0.88  1.34 0.12 0.12 0.08 0.06 0.05 0.09 0.03 0.06 0.02 0.01  1.96  0.96  1.51  1.72  1.86  1.34  1.82  1.69  1.59  2.15 0.08 0.01 0.02 0.08 0.01 0.06 0.16 0.17 0.04 0.02

STD

d18Ow Lat (1S) lon (1E) St dev Mg/Cadut St dev Mg/Catril St dev Mg/Carub

Also shown are seawater d18O values (d18Ow) measured in surface waters (upper 10 m) across the Mozambique Channel. Again, values measured close to the trap site are indicated in bold.

0

3.6. U k37 and TEX86 analysis

 2.24  2.16  2.20  2.04  2.37  1.95  2.68  2.21  2.42  2.46

For direct comparison of the d18O with Mg/Ca proxy SST, we corrected all d18O data for d18O of ambient seawater that was measured in surface waters across the channel (Fallet et al., 2010). Additionally, we contrasted core top data with in situ SST and SSS measured across the channel (391E–441E and 161S–171S, 0.51 grid) and flux-weighted annual averages of modern d18O and Mg/Ca of the same species obtained from time-series sediment traps moored at 2000 m water depth (Figs. 8 and 9).

0–1 0–1 0–1 0–1 0–1 0–1 0–1 0–1 0–1 –

SSTdut ¼ LNðMg=Ca=0:342Þ=0:09

1540 1820 2080 2247 2311 2490 2650 2110 1570 2000

18

16.57 16.62 16.68 16.72 16.70 16.80 16.94 17.05 17.25 16.42

18

40.22 40.45 40.67 40.84 40.84 41.13 41.71 42.23 42.99 40.85

18

CD153, MC6 CD153, MC5 CD153, MC4 CD153, MC3 (trap) D301, MC2 (trap) CD153, MC2 D301, MC3 D301, MC5 D301, MC6 Sediment trap

18

SSTdut ¼ 16:424:2nðd Odut 2d Ow Þ þ0:13nðd Odut 2d Ow Þ2

St dev

For N. dutertrei, we used the d18O calibration to calculate temperatures (Fig. 8) developed by Epstein et al. (1953) as it is also applied by Thunell and Sautter (1992) for temperature estimates based on N. dutertrei. Other possible d18O temperature calibrations (e.g. Kim and O’Neill, 1997) yield similar results. For Mg/Ca, we chose the species-specific calibration developed by Anand et al. (2003) to calculate Mg/Ca-based temperatures (Fig. 9) as it yields temperatures found at 50–70 m water depth in accordance with the preferential depth habitat of N. dutertrei (Dekens et al., 2002):

d18Odut

Fig. 6. Size-normalized weight (SNW) for three species of planktonic foraminifera from 39 sediment trap (traps) and 10 core top samples (sed). The caps at the end of each box indicate the extreme values (minimum and maximum), the box is defined by the lower and upper quartiles and the line in the center is the median.

St dev

traps sed N. dutertrei

d18Otril

sed traps G. trilobus

St dev

traps sed G. ruber

d18Orub

9

Core depth (cm)

12

Depth (m)

15

Lat (1S)

18

lon (1E)

SNW [μg]

21

45

Core

24

Table 3 Foraminiferal d18Oc and Mg/Ca results measured on core top sediment taken during cruises CD153 (November 2003) and D301 (April 2006). Annual, flux-weighted averages from sediment traps and the results from the core top samples close to the trap-site are indicated in bold. Values in italics are probably outliers and not included in calculations.

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

46

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

trap Juan Mozambique location de Nova 31 max summer temperatures

trap location

Mozambique

Madagascar

Juan de Nova

Madagascar

31 Jan 2003 (summer)

30

29 trap values 27 δ O-temperature [°C]

staellite SST [°C]

29

28

27

26

25

trap value

23

21 July 2003 (winter)

25

G. ruber G. trilobus N. dutertrei satellite SST 2003

19 min winter temperatures

24

17 38

39

Mozambique 35.4

40

trap location

41 42 43 longitude [°E]

Juan de Nova

44

45

46

max winter (dry season) salinities

satellite sea surface salinity

35.2

35

34.8

July 2003 (winter) Jan 2003 (summer)

34.4 min summer (wet season) salinities 34.2 38

39

40

41 42 43 longitude [°C]

44

45

39

40

41 42 longitude [°E]

43

44

45

Fig. 8. Variation of foraminiferal d18O-temperatures in core top sediment taken across the Mozambique Channel. For comparison flux-weighted annual mean d18O derived from sediment trap samples are also plotted (grey symbols, trap values). For plotting purposes, we averaged values from cores CD153-MC3 and D301-MC2 that are located close to the trap site. Temperatures were calculated with calibrations developed by Fallet et al. (2010) for G. ruber and G. trilobus and Epstein et al. (1953) for N. dutertrei. For comparison, summer and winter satellite SSTs (grey lines) are plotted (also see Fig. 7). For seawater d18Ow, we used a mean value of 0.56% measured in surface waters across the Mozambique Channel. As symbol size is larger than analytical precision, no error bars are shown. For all original d18O values and standard deviations, see Table 3.

Madagascar

34.6

38

46

column (0.32 mm diameter, film thickness of 0.12 mm) and helium as the carrier gas. The oven program was initiated at 70 1C, increased at a rate of 20 1C/min to 200 1C then to 3 1C/min until it reached 320 1C. The final temperature of 320 1C 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 0 areas to the concentration of the internal standard. The U k37 index (Prahl et al., 1988) was used to estimate SST and values were converted to temperature using the calibration from Conte et al. (2006): 0

Fig. 7. (a) Austral summer and winter SST together with minimum and maximum values (stippled lines) measured with the MODIS Aqua satellite across the Mozambique Channel (Giovanni database: 39.0–44.01E, 16–171S, 0.51 grid). As G. ruber and G. trilobus fluxes peak in January and July (Fallet et al., 2010), monthly averages from January (summer) and July (winter) 2003 are plotted for comparison with proxy temperatures (also see Figs. 8–10). (b) Sea surface salinity (SSS) data together with minimum and maximum values (stippled lines) measured with satellites across the Mozambique Channel (IRI/LDEO climate library: 39.0–44.01E, 16–171S, 0.51 grid) as monthly averages for January and July 2003.

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 (Fig. 10, Table 4), the alkenone fractions were analyzed on an HP 6890 GC using a 50 m CP Sil-5

0

0

SST ¼ 20:957 þ 54:293nðU k37 Þ252:894nðU k37 Þ2 þ28:321nðU k37 Þ3: The analytical error ranged from 0.002 to 0.04 which corresponds to standard deviations expressed in temperature of 0.03 and 0.55 1C, respectively. The polar fractions, containing the GDGTs, were ultrasonically dissolved in a mixture of 99:1 (vol:vol) hexane:proponol and filtered through 0.45 mm 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 according to Schouten et al. (2002) and converted to SST using the TEX H 86 global core top calibration of Kim et al. (2010): SST ¼ ð68:4nTEX H 86 Þ þ 38:6

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

trap location

Mozambique

Juan de Nova

Madagascar

31

Table 4 Organic matter proxy results measured on core top sediment taken during cruise M75-1 (Jan 2008) and 64PE304 (Mar 2009). Core

lon (1E)

Lat (1S)

0 Depth Core U k37 (m) depth (cm)

St dev TEX86

St dev

64PE304-56 64PE304-63 64PE304-68 M75-1b, MC1 M75-1b, MC2 M75-1b, MC3 M75-1b, MC4 M75-1b, MC5 M75-1b, MC6 Sediment trap

41.32 40.36 40.03 43.10 43.16 41.60 40.13 40.87 42.48 40.85

16.85 16.59 16.50 17.25 17.27 16.92 16.56 16.73 17.10 16.42

2652 1695 756 1027 455 2768 1345 2266 2198 2000

0.004 – 0.014 0.036 0.015 0.004 0.006 0.002 0.003 0.004

0.005 0.001 0.003 0.004 0.003 0.002 0.010 0.002 0.003 0.004

29

Mg/Ca-temperature [°C]

trap values 27

25

trap value

23

0–1 0–1 0–1 0–0.5 0–0.5 0–0.5 0–0.5 0–0.5 0–0.5 –

0.973 0.986 0.951 0.946 0.943 0.977 0.973 0.973 0.976 0.977

0.686 0.689 0.692 0.707 0.705 0.687 0.700 0.687 0.698 0.697

Annual, flux-weighted averages from sediment traps and core top sample from trap site are indicated in bold.

21 G. ruber G. trilobus N. dutertrei satellite SST 2003

19

17 38

39

40

41 42 longitude [°E]

43

44

45

Fig. 9. Variation of foraminiferal Mg/Ca in core top sediment taken across the Mozambique Channel. For comparison flux-weighted annual mean Mg/Ca derived from sediment trap samples are also plotted (grey symbols, trap). For plotting purposes, we averaged values from cores CD153-MC3 and D301-MC2 that are located close to the trap site. Temperatures were calculated with calibrations developed by Fallet et al. (2010) for G. ruber and G. trilobus and Anand et al. (2003) for N. dutertrei. For comparison, summer and winter satellite SSTs (grey lines) are plotted (also see Fig. 7). As symbol size is larger than analytical precision, no error bars are shown. For original Mg/Ca values and standard deviations, see Table 3.

trap location

Mozambique

Juan de Nova

Madagascar

31.0

30.0 trap values 29.0

28.0

27.0

U

and TEX SST [°C]

47

26.0

where TEX H 86 is derived from a global coretop dataset that was calibrated against SST but excludes data from the (sub-) polar oceans (Fig. 10, Table 4). The uncertainty based on duplicate measurements for TEX86 ranged from 0.0009 and 0.0097 (Table 4) which converts to standard deviations expressed in temperatures of 0.02 and 0.27 1C, respectively.

4. Results 4.1. Satellite SST, in situ salinity and d18Ow data Annual SST variation measured with satellites from November 2003 to March 2006 in coastal waters ranges between 24.8 and 31.8 1C (Fig. 2a, grey line) and for the mid-channel between 25.0 and 30.2 1C (Fig. 2a, black stippled line). The typical annual SST range in the open Mozambique Channel is  5 1C but can be 6.0 1C in coastal waters. Annual mean SST derived from these monthly SST data are 27.6 1C in the mid-channel and 28.3 1C on the shelf. Even monthly averaged SST show an across-channel gradient, particularly in austral summer when SST can vary by  2 1C between the shelf and the open channel (Fig. 7a). In situ measured salinities for the surface waters in the Mozambique Channel at the trap site vary between 35.2 in austral winter (dry season) and 34.8 in austral summer (wet season; Figs. 3 and 7b). Similarly to the area-averaged monthly satellite SST, sea surface salinities show an across-channel gradient (Figs. 3b and 7b). It appears that especially the Madagascar shelf is influenced by increased river-input during both summer and winter as it consistently shows lower salinity values than elsewhere in the channel. Available d18Ow and salinity data from the SW Indian Ocean show a linear correlation with an r2 of 0.5 (Fig. S1):

d18 Ow ¼ 0:49nS17:16 U

25.0

TEX satellite SST 2003 24.0 38

39

40

41 42 longitude [°E]

43

44

45

Both, d18Ow and salinity data are from 3 m water depth. The data were taken from the databases http://cchdo.ucsd.edu/ and http://data.giss.nasa.gov/o18data/ and from scientific cruises in spring, summer and autumn to the Mozambique Channel (2003– 2009). This d18Ow–salinity equation is similar to one obtained for the subtropical South Atlantic (Loncaric et al., 2006).

0

Fig. 10. Variation of organic SST proxies U k37 and TEX H 86 in core top sediment taken 0 across the Mozambique Channel. For comparison flux-weighted annual mean U k37 and TEX86 derived from sediment trap samples are also plotted (grey symbols, trap). The sea surface temperature was calculated with the Conte et al., (2006) and 0 Kim et al. (2010) calibrations for U k37 and TEX H 86 , respectively. For comparison, summer and winter satellite SSTs (grey lines) are plotted (also see Fig. 7). For 0 original U k37 and TEX86 values see Table 4.

4.2. Radionuclides of a mid-channel sediment core The 210Pb(tot) profile from the core taken at the trap-site decreases from a core top (0–1 cm) activity of 743717 Bq kg  1 smoothly to background values of 50 Bq kg  1 at about 8 cm core depth (Fig. 4b,

48

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

Table 2). The 210Pb profile shows a diffusive mixing pattern with little non-local mixing as described by Soetaert et al. (1996). The 14C dates of foraminiferal calcite and TOC yield absolute core top ages of 1066735 yr BP and 983730 yr BP, respectively and downcore sedimentation rates of about 5 cm/1000 years (Fig. 4a, Table 2). We corrected for compaction of the sediment by calculating the cumulative dry mass (Fig. 4b) which shows that linear sedimentation and accumulation rates have remained constant over the past 8000 years. We observed two outliers in the 14C of the TOC and the foraminiferal calcite profile which might have been caused by locally intensive bioturbation by, e.g. feeding tubes (Aller, 1984; Hughes et al., 2005). Using the 14C sedimentation rate in the 210Pb model, we calculated a mixing coefficient of 0.1 cm2 year  1 and a mixed layer depth of about 10 cm consistent with observations from a global dataset (Boudreau, 1998). The excess (xs) 234Th-data indicate a high activity of 458 Bq kg  1 in the uppermost sediments (0.25–0.5 cm) but supported background levels are reached at intervals deeper than 1 cm (Table 2). This indicates deposition and incipient burial of fresh material at the core sites within the last three months before sampling, without indications of significant non-local mixing by bioturbation during the same period. 4.3. Time-series mass flux, lithogenic matter,

210

Pb and

206

Pb/207Pb

Time-series lithogenic matter collected in the upper sediment trap varies between 21 and almost 80wt% of total fluxes over the 2.5 years deployment period from November 2003 to March 2006 (Fig. 5). As the lower trap started collecting material only in March 2005, we only directly compare values of the upper and lower trap of the second deployment period. These show that lithogenic matter in the lower trap is overall by about 10% higher than the corresponding samples in the upper trap with the general pattern being similar (lower panel, Fig. 5, Table 1). Total mass flux varies strongly between 30 and 600 mg m  2 day  1 with an annual average of 110 mg m  2 day  1. The high total flux events are found in samples MOZ1-A06 and A11 (Fig. 5, circles), which also correspond to maxima in lithogenic matter. The total mass flux for the lower trap is about 2.5 times higher than for the upper trap with an annual average of 250 mg m  2 day  1. Average 210Pb activities recorded in the upper and lower trap are about 4000 Bq kg  1 with minima around 2000 Bq kg  1 (middle panel, Fig. 5, Table 1) that indicate the episodic input of older sediments. The ratio 206Pb/207Pb, measured in the deployment period November 2003 to March 2005, varies between 1.15 and 1.22 (upper panel, Fig. 5, Table 1). The upper values of 1.22 correspond to samples MOZ1-A06 and MOZ1-A11, which also show highest lithogenic matter and total mass fluxes as well as lowest 210Pb values.

time series are 2.5% for d18O and 5.3 mmol/mol for Mg/Ca (Fallet et al., 2010), significantly more positive than coretop values. The surface-dwelling species G. trilobus yields core top d18O values across the channel ranging between –1.3% and  2.0% with no particular trend (Table 3). Mg/Ca core top values of G. trilobus are also quite homogeneously distributed across the channel ranging from a minimum of 3.3 mmol/mol in the midchannel to a maximum of 3.5 mmol/mol at the African continental slope (Table 3). Similarly to d18O and Mg/Ca of G. ruber, modern flux-weighted averages of G. trilobus ( 2.2% and 3.7 mmol/mol, Fallet et al., 2010) are also significantly more positive than the respective coretop values. At the channel flanks the d18O core top values of the subsurface-dweller N. dutertrei range between 1.2% and 0.9% and mid-channel values between  0.6% and  0.3% (Table 3). The Mg/Ca core top values for N. dutertrei range from 2.3 to 2.9 mmol/ mol close to the channel flanks and the island Juan de Nova and from 2.0 to 2.2 mmol/mol in the open Mozambique Channel (Table 3). For N. dutertrei modern flux-weighted averaged values for d18O and Mg/Ca are also significantly more positive than the core top specimens with values of 1.3% and 3.0 mmol/mol, respectively (Fallet et al., 2011). For d18O of all three analyzed species, the uncertainty based on duplicate measurements ranges between 0.01% and 0.17% and for Mg/Ca between 0.005 and 0.69 mmol/mol (Table 3). 0

4.5. U k37 and TEX86 of core top sediments 0

The U k37 core top values are homogeneously distributed mid channel (Table 4) but show scattered values closer to the channel 0 flanks. The highest U k37 -values in the core top sediments are found in the mid-channel and range from 0.99 to 0.98. The lowest values of 0.94 and 0.95 are found close to the African and Madagascan 0 continental slope. Modern flux-weighted averages of U k37 are 0.98, identical to the core top values from the same location. Also TEX86 core top values are relatively homogeneously distributed across the Mozambique Channel (Table 4) but display scattered values near the channel flanks. We found highest values of 0.70 and 0.71 close to the African and Madagascan continental slope and the 0 lowest value of 0.69 in the mid-channel. (Table 4). Similar to U k37 , modern flux-weighted averages of TEX86 are 0.70 akin to core top values from the same location.

5. Discussion 5.1. Age and proxy offsets between sediment trap and core top sediment

4.4. Shell weights, d18O and Mg/Ca of core top foraminifera Core top and sediment trap shell weights of G. ruber specimens average at 12.1 and 11.8 mg, for G. trilobus at 17.2 and 16.8 mg and for N. dutertrei at 16.8 and 16.2 mg, respectively (Fig. 6). The shell weights for each individual foraminifera species are thus similar for core top and time-series specimens in the 250–315 mm size fraction. For the surface-dwelling G. ruber in the core top sediments we find d18O values of  2.7% and –2.4% from cores taken closest to the Madagascan shelf and the island of Juan de Nova (Table 3). The lowest value of  2.0% is observed in the mid-channel. For Mg/Ca of G. ruber, highest values occur near the Madagascan shelf and the island of Juan de Nova (4.6–4.8 mmol/mol) but also in the core top sediment taken closest to the African shelf (Table 3). The lowest Mg/Ca value of 4.1 mmol/mol is found in the mid-channel. In comparison, the flux-weighted values of the sediment trap

Radiocarbon analysis of the core top sediment in the Mozambique Channel revealed an average age of about 1000 years (Fig. 4a) and particles are thus partly not of modern 20th century origin. The possible causes for this are (1) low accumulation rates combined with (2) continuous vertical mixing of sediment and (3) lateral transport of old particles. Additionally, we observed that the inorganic proxies from the core top sediment at the site of the sediment trap show a striking difference with the flux-weighted average of the time series and modern SST. For the surface- and the subsurface-dwelling foraminifera species, all d18O- and Mg/Ca-based inferred temperatures from modern time-series (Fallet et al., 2010) yield warmer temperatures than those from the core top sediments (Figs. 8 and 9). For G. ruber, both coretop proxy values are around 1–2 1C colder than the time-series average. Similar or larger offsets occur for surfacedwelling G. trilobus and the subsurface-dwelling 0 species N. dutertrei. However, we do not observe any offset in U k37 and TEX86 values

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

5.2. Bioturbation Our 234Thxs data imply that there is a flux of fresh sediment to the bottom of the Mozambique Channel while the 210Pb data suggest that both old and young sediments are continuously mixed downcore (Table 2). This is supported by the radiocarbon results of both TOC and foraminiferal calcite, which are respectively 983 730 yr BP and 1066735 yr BP old (Table 2). The mixture of old TOC and foraminiferal calcite could be caused by advection and transport of pre-aged matter. However, if this was the dominant mechanism in the Mozambique Channel, one would expect an older 14C age for the TOC than for the foraminifera as TOC is associated with the finer fraction and thus much more prone to advection (Mollenhauer et al., 2007; Ohkouchi et al., 2002). As this is not the case, it seems more likely that the 14C ages are mainly controlled by bioturbation, which mixes deeper older sediment layers with surface layers for both TOC and the coarser foraminiferal calcite. If bioturbation and downcore mixing is taken into account with an average mixing depth of about 10 cm (Fig. 4b) at a mean sedimentation rate of 5 cm/1000 years (Fig. 4a), changes in SST over the past  2000 years might cause the offset in proxy temperature between modern and coretop foraminifera. Part of the offset between modern and core top foraminifera could reflect the rapid recent warming that started with the onset of industrialization at the end of the 19th century. For example, instrumental temperature data for the southern Equatorial Indian Ocean showed a significant warming trend of 0.7 1C for the period 1901–2002 (Kothawale et al., 2008). However, natural climate variability over the past 2000 years such as the ‘‘Little Ice Age’’ and the ‘‘Medieval Climate Anomaly’’ also has to be considered (Jones and Mann, 2004; Mann et al., 2009). Studies of lacustrine and terrestrial temperature proxies from (sub-) tropical Africa spanning the past ca. 2000 years have shown that such a temperature offset would amount to about 2 1C (Holmgren et al., 1999; Powers et al., 2011; Tierney et al., 2010) and fall directly into the observed temperature offset of 1–3 1C between modern and core top foraminifera. Both the anthropogenic warming trend and the natural climate variability may perhaps be less visible in the organic proxies in core top sediments because these can be more affected by downcore mixing than their foraminiferal counterparts. The observation that TOC-ages are systematically (even though only slightly) younger than foraminiferal calcite (Fig. 4) indeed suggests preferential downcore mixing of the fine fraction (Bard, 2001). 5.3. Sediment transport

distal areas (Ohkouchi et al., 2002) or continuous cycles of resuspension and redeposition during cross-shelf transport (Mollenhauer et al., 2003; Mollenhauer et al., 2007). For the Mozambique Channel, the lack of a seasonal signal in the organic proxies from the sediment traps time-series (Fallet et al., 2011) and the relatively homogeneous distribution of values in coretop sediments suggest that organic proxies are substantially affected by current transport. Foraminifera are usually heavier than organic matter and rapidly sink to the ocean floor within days (Gyldenfeldt et al., 2000; Takahashi and Be´, 1984), although organic matter in ballasted ‘marine snow’ aggregates may attain sinking speeds approaching those of foraminifera (Wakeham et al., 2009). Strong current transport of (unconsolidated) pre-aged organic matter that causes a homogenization of the organic proxy signal would account for the lacking temperature offset as organic proxies in both sediment trap and core tops are probably time- and area-averaged. Further clues can be obtained from the stable lead isotope analysis of sediment trap material. Most of the 206Pb/207Pb values fall between 1.15 and 1.18 but during several periods with elevated fluxes the 206Pb/207Pb ratio increased to  1.22, suggesting input of pre-industrial material (Fig. 5). Maximum 206Pb/207Pb in our study area is comparable to pre-industrial (prior to 1880 AD) ratios from Antarctic ice (Vallelonga et al., 2002; Van de Velde et al., 2005) and to the average lead isotopic composition of upper continental crust (Hemming and McLennan, 2001; Millot et al., 2004). Most other values are intermediate between this tentative natural endmember and late 20th century urban aerosols from South Africa with 206 ¨ Pb/207Pb values around 1.07 (Bollhofer and Rosman, 2000), suggesting a mix of natural and anthropogenic sources of Pb. As the time-series 206Pb/207Pb isotope composition in the Mozambique Channel shows periodic events with values around 1.22 in combination with maxima in lithogenic fluxes (Fig. 11), we can assume that pre-industrial sediment is intermittently resuspended from the upper margin and transported downslope to the deep channel. Comparable transport events have been described by Richter et al. (2009) for submarine canyons in the NE Atlantic Ocean. Two such resuspension events in the Mozambique Channel were found in the 1.24

1.22

Pb/ Pb

between trap and the core top sediment (Fig. 10). Below we discuss potential causes for this mismatch between inorganic and organic proxies.

49

older sediment

resuspension events (MOZ1-A06, A11)

1.2

1.18

In addition to bioturbation and downcore mixing, the discrepancy between organic and inorganic proxies can also be caused by differential transport processes affecting both fresh and pre-aged sediments. Time-series mooring data show that in the Mozambique Channel strong rotational current velocities often exceed 1.5 m/s (Harlander et al., 2009). Such high current velocity could (re-) suspend and transport particles over long distances as shown for study areas with similarly high current speeds (McCave et al., 1980; Mollenhauer et al., 2003; Ohkouchi et al., 2002). Especially, organic proxies associated with the fine fraction of sediments can be transported over long distances in certain settings (Benthien and ¨ Muller, 2000; Kim et al., 2009; Sicre et al., 2005). Lateral transport can induce substantial age offsets between fine and coarse fractions through resuspension of pre-aged particles by bottom currents in

1.16

younger sediment 1.14 0

100 200 300 400 lithogenic matter flux [mg m day ]

500

Fig. 11. Lithogenic matter flux from the upper sediment trap plotted against the 206 Pb/207Pb ratio showing that younger sediment has lower lithogenic matter fluxes and that resupension events have by far the highest lithogenic matter fluxes.

50

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

upper trap that was moored 250 m above the bottom of the ocean (Fig. 5) indicating the force of these events. This suggests intense sediment resuspension by strong (bottom) currents and/or frequent input of sediment originating from off-slope intermediate nepheloid layers. Similar transport events along continental slopes have been observed in many previous studies (e.g. Mollenhauer et al., 2003; Ransom et al., 1998; Thomsen and Gust, 2000). 5.4. Across-channel variability in temperature and salinity So far, we discussed the observed proxy variance in the light of downcore mixing and transport of sediment. Yet, vertical fluxes of foraminifera from areas with a different temperature and salinity regime could also cause differences between inorganic and organic proxies. Our data show that there is a large variance in the d18O- and Mg/Ca-temperature signal between foraminifera from the mid-channel and the channel flanks (Figs. 8 and 9). Indeed, satellite SST can be almost 2 1C higher in coastal areas as opposed to the mid-channel during summer whereas for the rest of the year it is more homogeneous across the channel (Figs. 2a and 7a). In addition to the temperature effect in d18O and Mg/Ca, salinity is known to affect foraminiferal d18O (Shackleton and Vincent, 1978) and to some extent Mg/Ca (Arbuszewski et al., 2010; Ferguson et al., 2008; Mathien-Blard and Bassinot, 2009) 0 but, as far as we know, does not impact U k37 and TEX86 values for the range of salinities encountered in open ocean settings (Herbert, 2001 and references therein; Wuchter et al., 2004). Long-term satellite observations for the Mozambique Channel show that salinity in coastal waters is influenced by seasonal variation in freshwater input by rivers. On the broad Madagascar shelf, we find lowest salinity values varying seasonally between 34.4 and 35.0 (Fig. 7b). As (large-scale) salinity variation can potentially exert an influence on the analyzed foraminiferal temperature proxies, we subsequently discuss this effect for the Mozambique Channel. However, the quantitative estimation of the salinity effect on the proxies is uncertain as the relationship between excess d18O and Mg/Ca and salinity is variable between studies (culture, core tops) and in different oceanic regions. In order to account for the observed d18O difference between core top and sediment trap foraminifera, seawater d18Ow has to be raised by 0.2–0.5%, corresponding to an increase in salinity of 0.6 g kg  1 using the salinity–d18Ow calibration for the SW Indian Ocean (Fig. S1). For Mg/Ca, salinity changes would also have to be around 1.0 g kg  1 in order to yield the observed offset between time-series and core top specimens. It has been shown that an increase of 1.0 g kg  1 in salinity can cause a 20% increase in the ¨ Mg/Ca content of a G. ruber shell (Ferguson et al., 2008; Kisakurek et al., 2008; Mathien-Blard and Bassinot, 2009). The lower salinities in coastal regions would lead to more positive d18Oand lower Mg/Ca-values compared to open water sites causing colder d18O- and Mg/Ca-based proxy temperatures. As the opposite is seen in the Mozambique Channel, we argue that salinity does not cause the observed variation in foraminiferal d18O- and Mg/Ca-based proxy temperatures. As both organic proxies are probably independent of (smallscale) salinity changes, less saline coastal waters would not affect their signature, which leaves SST variation as the dominant controlling factor. Since SST variation is larger close to the shores than in the mid-channel, organic proxies that are produced in coastal waters under more variable temperature conditions (warmer in summer, rarely colder in winter) could record more variable proxy temperatures. In addition to more variable organic 0 proxy-values, we also observed a slight opposite trend in U k37 and TEX86 in coastal samples (Fig. 10). This observation is intriguing but it should be noted that the temperature differences are well 0 within the calibration errors of around 1.5 1C for U k37 (Conte et al.,

2006; Prahl et al., 1988) and 2.5 1C for TEX86 (Kim et0 al., 2010). Another possible explanation for the differences in U k37 and TEX86 is that the surface sediments across the Mozambique Channel are not of the same age. 5.5. Effects of diagenesis on (in-)organic proxies Another potential cause for the discrepancy between organic and inorganic temperature proxies could be diagenetic effects on the 0 foraminiferal shells, which would not affect the U k37 and TEX86. It has been repeatedly shown that calcite dissolution preferentially removes 16O and Mg2 þ (Brown and Elderfield, 1996; Dekens et al., 2002) while secondary calcite precipitation adds those components (Hoogakker et al., 2009) thereby influencing proxy-based temperature reconstruction. To ascertain whether diagenesis has affected the core top Mozambique Channel foraminifera, we determined their shell weight (Fig. 6) and examined their wall-structure by SEM-microphotography (Fig. S2, supplementary data). As shell weights of time-series and core top specimens for all three foraminifera species are statistically identical and SEM images do not show calcite dissolution nor precipitation, we assume that foraminifera shells have not been altered. Additionally, Tachikawa et al. (2008) showed that minor to moderate dissolution does not affect Mg/Ca of G. ruber in the western Indian Ocean. Moreover, the deep Mozambique Channel (2700 m maximum water depth) is too shallow to sustain a pressure and temperature regime where calcite is easily dissolved (above the calcite lysocline), which is also testified by the excellent preservation of fragile foraminiferal species such as Hastigerina pelagica. Finally, we observed no proxy–depth relationship for any of the six measured parameters (Table 3) that could hint at shallow depth dissolution (Brown and Elderfield, 1996; Regenberg et al., 2006) or the inorganic precipitation of bottom/ deep-water carbonate (aged 14C, cold d18O–Mg/Ca). Thus, it is unlikely that the lower proxy temperatures of the core top foraminifera compared to those of the sediment trap are caused by diagenetic effects. 5.6. Seasonal fluxes, interspecies effects, pH and carbonate ion It has been suggested that seasonal variation in foraminiferal shell fluxes might skew the sedimentary temperature signal and that it varies in different species (Berger and Wefer, 1990; Curry et al., 1992; Deuser et al., 1981). For the Mozambique Channel, the two surface-dwelling species G. ruber and G. trilobus have been shown to capture seasonal variation in satellite SST well with G. ruber predominantly calcifying during austral summer and G. trilobus during winter (Fallet et al., 2010). Therefore, the summer species G. ruber yields a slightly warmer and the winter species G. trilobus a colder, flux-weighted SST than the annual average satellite SST of 27.6 1C (Figs. 8 and 9, grey symbols). Yet, this temperature shift caused by seasonal variation in shell fluxes is well within the calibration error of about 1.3 1C for d18O (e.g. Spero et al., 1997) and 1.2 1C for Mg/Ca (Dekens et al., 2002). In 0 contrast, the organic fluxes and proxies U k37 and TEX86 from the same sediment trap series show no seasonal signal but both return flux-weighted SST of 28.1 and 28.3 1C, respectively (Fallet et al., 2011) that also correspond well with annual mean satellite SST (Fig. 10, grey symbols). As modern flux-weighted SST should reflect the sedimentary signal, we would expect that fluxweighted SST from the sediment trap series and core top proxy temperatures return similar results for both foraminiferal and organic proxies. As this is only observed for the organic proxies, seasonal variation in fluxes cannot fully explain the observed discrepancy in temperature proxies between modern and core top foraminifera.

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

Finally, intra-species and ontogenetic-size variations that may affect the foraminiferal d18O and Mg/Ca ratios (Elderfield et al., 2002) are minimized in the small size range of 250–315 mm used in this study. We can also rule out ambient pH (Lea et al., 1999) and carbonate ion concentration (Russell et al., 2004) as factors affecting foraminiferal d18O and Mg/Ca ratios as the surfacemixed-layer waters of the Mozambique Channel can be considered constant with respect to total alkalinity variation (Lee et al., 2006). As size variation, pH, carbonate-ion concentration and seasonal variation in shell fluxes do not seem to significantly affect the measured foraminiferal proxies, we conclude that transport of fresh and pre-aged material together with downcore mixing of sediment probably causes the observed offset between modern and coretop foraminiferal temperature proxies.

6. Conclusions 0

We measured a number of organic (U k37 and TEX86) and inorganic (d18O and Mg/Ca in G. ruber, G. trilobus and N. dutertrei) temperature proxies in core top sediments taken across the Mozambique Channel and contrasted these results with the same temperature proxies from a sediment trap moored in the deep channel. Radionuclide analysis shows that core top sediments are a mixture of fresh and pre-aged particles contributed by offshore advection and down-slope transport combined with bioturbative vertical mixing of the accumulating sediment. The organic and inorganic temperature proxies from the core top sediment also give evidence that higher (and more variable) temperatures are recorded on the channel flanks and lower temperatures in the more open waters, consistent with higher modern maximum temperatures and salinities in coastal areas as compared to the central Mozambique Channel. Additionally, we observed a significant positive temperature offset of 1–3 1C between time-series and core top foraminifera, which is not seen in the organic temperature proxies. We suggest that part of this temperature offset in foraminifera is due to warming of the Indian Ocean over the past  2000 years. Such a warming trend is not observed in the organic proxies, which is likely due to transport of fine sediments by strong currents together with bioturbative mixing of sediment downcore that might homogenize proxy signatures.

Acknowledgments This research was supported by the Netherlands Organization for Scientific Research (NWO) through the Netherlands-Bremen Oceanography II (NEBROC2) and LOCO programs, as well as Paleosalt (ESF-EuroClimate) and GATEWAYS (EU-FP7, MC-ITN). We are grateful to Santiago Gonzalez for assisting with sample preparation and to Sharyn Crayford for C/N and biogenic silica analyses. We thank Denise Dorhout for Mg/Ca analyses, Michiel Kienhuis and Evaline van Weerlee for support with the oxygen isotope analysis, and Jort Ossebaar, Marianne Baas and Ellen Hopmans for support with the organic proxy analysis. We are also grateful to Eric Epping for the intensive discussion on the 210Pb age model. We also acknowledge all technical support during research cruises onboard the RRV Charles Darwin, RRV Discovery, FS Meteor and RV Pelagia. Analyses and visualizations of satellite SST used in this paper were produced with the Giovanni online data system, developed and maintained by the NASA GES DISC. Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at doi:10.1016/j.dsr.2011.10.002.

51

References Acker J.G. and Leptoukh G. (2007) Online analysis enhances use of NASA Earth science data, pp. 14–17. Aller, R.C., 1984. The importance of relict burrow structures and burrow irrigation in controlling sedimentary solute distributions. Geochim. Cosmochim. Acta 48, 1929–1934. Anand, P., Elderfield, H., Conte, M.H., 2003. Calibration of Mg/Ca thermometry in planktonic foraminifera from a sediment trap time series. Paleoceanography 18. Arbuszewski, J., de Menocal, P., Kaplan, A., Farmer, E.C., 2010. On the fidelity of shell-derived d18O-seawater estimates. Earth Planet. Sci. Lett. 300, 185–196. Bard, E., 2001. Paleoceanographic implications of the difference in deep-sea sediment mixing between large and fine particles. Paleoceanography 16, 235–239. Barker, S., Greaves, M., Elderfield, H., 2003. A study of cleaning procedures used for foraminiferal Mg/Ca paleothermometry. Geochem. Geophys. Geosyst. 4. Beal, L.M., De Ruijter, W.P.M., Biastoch, A., Zahn, R., 2011. On the role of the Agulhas system in ocean circulation and climate. Nature 472 (429–436). ¨ Benthien, A., Muller, 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. Paper 47, 2369–2393. Berger, W.H., Wefer, G., 1990. Export production—seasonality and intermittency, and paleoceanographic implications. Global Planet Change 89, 245–254. Boer, W., van den Bergh, G.D., de Haas, H., de Stigter, H.C., Gieles, R., van Weering, T.C.E., 2006. Validation of accumulation rates in Teluk Banten (Indonesia) from commonly applied Pb-210 models, using the 1883 Krakatau tephra as time marker. Mar. Geol. 227, 263–277. ¨ Bollhofer, A., Rosman, K.J.R., 2000. Isotopic source signatures for atmospheric lead: the Southern Hemisphere. Geochim. Cosmochim. Acta 64, 3251–3262. Bonnin, J., van Raaphorst, W., Brummer, G.J., van Haren, H., 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. I Oceanogr. Res. Paper 49, 1485–1505. Boudreau, B.P., 1998. Mean mixed depth of sediments: the wherefore and the why. Limnology and Oceanography 43, 524–526. Brown, S.J., Elderfield, H., 1996. Variations in Mg/Ca and Sr/Ca ratios of planktonic foraminifera caused by postdepositional dissolution: evidence of shallow Mg-dependent dissolution. Paleoceanography 11, 543–551. Conte, M.H., Sicre, M.A., Ruhlemann, C., Weber, J.C., Schulte, S., Schulz-Bull, D.E., Blanz, T., 2006. Global temperature calibration of the alkenone unsaturation index (U-37(K0 )) in surface waters and comparison with surface sediments. Geochem. Geophys. Geosyst., 7. Curry, W.B., Ostermann, M.V., Gupta, S., Ittekkot, V., 1992. Foraminiferal production and monsoonal upwelling in the Arabian Sea. J. Geol. Soc. London 64, 93–106. d’Orbigny (1839) Foraminife res. In: Histoire Physique, Politique et Naturelle de l’ile de Cuba2 (ed. R. de la Sagra), pp. 1–224. de Master, D.J., Cochran, J.K., 1982. Particle mixing rates in deep-sea sediments determined from excess Pb-210 and Si-32 profiles. Earth Planet. Sci. Lett. 61, 257–271. de Ruijter, W.P.M., Biastoch, A., Drijfhout, S.S., Lutjeharms, J.R.E., Matano, R.P., Pichevin, T., van Leeuwen, P.J., Weijer, W., 1999. Indian–Atlantic interocean exchange: dynamics, estimation and impact. J. Geophys. Res.- Oceans 104, 20885–20910. Dekens, P.S., Lea, D.W., Pak, D.K., Spero, H.J., 2002. Core-top calibration of Mg/Ca in tropical foraminifera: refining paleotemperature estimation. Geochem. Geophys. Geosyst. 3. Deuser, W.G., Ross, E.H., Hemleben, Ch., Spindler, M., 1981. Seasonal changes in species composition, numbers, mass, size, and isotopic composition of planktonic foraminifera settling into the deep Sargasso Sea. Palaeogeogr. Palaeoclimatol. Palaeoecol. 33, 103–127. Elderfield, H., 2002. Climate change: carbonate mysteries. Science 296, 1618–1621. Elderfield, H., Vautravers, M., Cooper, M., 2002. The relationship between shell size and Mg/Ca, Sr/Ca, delta O-18, and delta C-13 of species of planktonic foraminifera. Geochem. Geophys. Geosyst. 3. Epstein, S., Buchsbaum, R., Lowenstam, H.A., Urey, H.C., 1953. Revised carbonatewater isotopic temperature scale. Geol. Soc. Am. Bull. 64, 1315–1326. Fairbanks, R.G., Mortlock, R.A., Chiu, T.C., Cao, L., Kaplan, A., Guilderson, T.P., Fairbanks, T.W., Bloom, A.L., Grootes, P.M., Nadeau, M.J., 2005. Radiocarbon calibration curve spanning 0 to 50,000 years BP based on paired Th-230/ U-234/U-238 and C-14 dates on pristine corals. Quat. Sci. Rev. 24, 1781–1796. Fairbanks, R.G., Sverdlove, M., Free, R., Wiebe, P.H., 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., Brummer, G.J.A., 2009. A novel application of wet oxidation to retrieve carbonates from large organic-rich samples for ocean-climate research. Geochem. Geophys. Geosyst. 10. Fallet, U., Brummer, G.J., Zinke, J., Vogels, S., 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. Fallet, U., Ullgren, J.E., Castaneda, I.S., van Aken, H.M., Schouten, S., Ridderinkhof, H., Brummer, G.J., 2011. Contrasting variability in foraminiferal and organic paleotemperature proxies in sedimenting particles of the Mozambique Channel (SW Indian Ocean). Geochim. Cosmochim. Acta 75, 5834–5848.

52

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

Ferguson, J.E., Henderson, G.M., Kucera, M., Rickaby, R.E.M., 2008. Systematic change of foraminiferal Mg/Ca ratios across a strong salinity gradient. Earth Planet. Sci. Lett. 265, 153–166. Gardner, W.D., Sullivan, L.G., 1981. Benthic storms: temporal variability in a deepocean nepheloid layer. Science 213, 329–331. Gyldenfeldt, A.B., Carstens, J.r., Meincke, J., 2000. Estimation of the catchment area of a sediment trap by means of current meters and foraminiferal tests. DeepSea Res. II: Top. Stud. Oceanogr. 47, 1701–1717. Harlander, U., Ridderinkhof, H., Schouten, M.W., de Ruijter, W.P.M., 2009. Longterm observations of transport, eddies, and Rossby waves in the Mozambique Channel. J. Geophys. Res. -Oceans 114. Hemming, S.R., McLennan, S.M., 2001. Pb isotope compositions of modern deep sea turbidites. Earth Planet. Sci. Lett. 184, 489–503. Herbert, T.D., 2001. Review of alkenone calibrations (culture, water column, and sediments). Geochem. Geophys. Geosyst. 2. Holmgren, K., Karlen, W., Lauritzen, S.E., Lee-Thorp, J.A., Partridge, T.C., Piketh, S., Repinski, P., Stevenson, C., Svanered, O., Tyson, P.D., 1999. A 3000-year highresolution stalagmite-based record of palaeoclimate for northeastern South Africa. Holocene 9, 295–309. Hoogakker, B.A.A., Klinkhammer, G.P., Elderfield, H., Rohling, E.J., Hayward, C., 2009. Mg/Ca paleothermometry in high salinity environments. Earth Planet. Sci. Lett. 284, 583–589. Hopmans, E.C., Weijers, J.W.H., Schefuss, E., Herfort, L., Sinninghe Damste´, J.S.S., 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. Hughen, K., Lehman, S., Southon, J., Overpeck, J., Marchal, O., Herring, C., Turnbull, J., 2004. C-14 activity and global carbon cycle changes over the past 50,000 years. Science 303, 202–207. Hughes, D.J., Brown, L., Cook, G.T., Cowie, G., Gage, J.D., Good, E., Kennedy, H., MacKenzie, A.B., Papadimitriou, S., Shimmield, G.B., Thomson, J., Williams, M., 2005. The effects of megafaunal burrows on radiotracer profiles and organic composition in deep-sea sediments: preliminary results from two sites in the bathyal north-east Atlantic. Deep-Sea Res. I: Oceanogr. Res. Papers 52, 1–13. Jones, P.D., Mann, M.E., 2004. Climate over past millennia. Rev. Geophys. 42. Keigwin, L.D., Guilderson, T.P., 2009. Bioturbation artifacts in zero-age sediments. Paleoceanography 24, A4212. Kim, J.H., Crosta, X., Michel, E., Schouten, S., Duprat, J., Sinninghe Damste´, J.S.S., 2009. Impact of lateral transport on organic proxies in the Southern Ocean. Quat. Res. 71, 246–250. Kim, J.H., Schouten, S., Hopmans, E.C., Donner, B., Sinninghe Damste´, J.S.S., 2008. Global sediment core-top calibration of the TEX86 paleothermometer in the ocean. Geochim. Cosmochim. Acta 72, 1154–1173. Kim, J.H., van der Meer, J., Schouten, S., Helmke, P., Willmott, V., Sangiorgi, F., Koc- , N., Hopmans, E.C., 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. Kim, S.T., O’Neill, 1997. Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates. Geochim. Cosmochim. Acta 61, 3461–3475. ¨ ¨ ¨ Kisakurek, B., Eisenhauer, A., Bohm, F., Garbe-Schonberg, D., Erez, J., 2008. Controls on shell Mg/Ca and Sr/Ca in cultured planktonic foraminiferan, Globigerinoides ruber (white). Earth Planet. Sci. Lett. 273, 260–269. Koning, E., Epping, E., van Raaphorst, W., 2002. Determining biogenic silica in marine samples by tracking silicate and aluminium concentrations in alkaline leaching solutions. Aquat. Geochem. 8, 37–67. Kothawale, D.R., Munot, A.A., Borgaonkar, H.P., 2008. Temperature variability over the Indian Ocean and its relationship with Indian summer monsoon rainfall. Theor. Appl. Climatol. 92, 31–45. Laine, E.P., Gardner, W.D., Jo Richardson, M., Kominz, M., 1994. Abyssal currents and advection of resuspended sediment along the northeastern Bermuda Rise. Mar. Geol. 119, 159–171. Lea, D.W., 2003. Elemental and isotopic proxies of past ocean temperatures. In: Heinrich, K.T., Karl, K.T. (Eds.), Treatise on Geochemistry, Pergamon, Oxford, pp. 1–26. Lea, D.W., Mashiotta, T.A., Spero, H.J., 1999. Controls on magnesium and strontium uptake in planktonic foraminifera determined by live culturing. Geochim. Cosmochim. Acta 63, 2369–2379. Lee, K., Tong, L.T., Millero, F.J., Sabine, C.L., Dickson, A.G., Goyet, C., Park, G.H., Wanninkhof, R., Feely, R.A., Key, R.M., 2006. Global relationships of total alkalinity with salinity and temperature in surface waters of the world’s oceans. Geophys. Res. Lett. 33. Loncaric, N., Peeters, F.J.C., Kroon, D., Brummer, G.J.A., 2006. Oxygen isotope ecology of recent planktic foraminifera at the central Walvis Ridge (SE Atlantic). Paleoceanography 21. Loncaric, N., van Iperen, J., Kroon, D., 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. Mann, M.E., Zhang, Z., Rutherford, S., Bradley, R.S., Hughes, M.K., Shindell, D., Ammann, C., Faluvegi, G., Ni, F., 2009. Global signatures and dynamical origins of the Little Ice Age and Medieval Climate Anomaly. Science 326, 1256–1260. Masque, P., Fabres, J., Canals, M., Sanchez-Cabeza, J.A., Sanchez-Vidal, A., Cacho, I., Calafat, A.M., Bruach, J.M., 2003. Accumulation rates of major constituents of hemipelagic sediments in the deep Alboran Sea: a centennial perspective of sedimentary dynamics. Mar. Geol. 193, 207–233.

Mathien-Blard, E., Bassinot, F., 2009. Salinity bias on the foraminifera Mg/Ca thermometry: correction procedure and implications for past ocean hydrographic reconstructions. Geochem. Geophys. Geosyst. 10, Q12011. McCave, I.N., Lonsdale, P.F., Hollister, C.D., Gardner, W.D., 1980. Sediment transport over the Hatton and Gardar contourite drifts. J. Sediment. Petrol. 50, 1049–1062. Millot, R., Allegre, C.J., Gaillardet, J., Roy, S., 2004. Lead isotopic systematics of major river sediments: a new estimate of the Pb isotopic composition of the Upper Continental Crust. Chem. Geol. 203, 75–90. ¨ Mollenhauer, G., Eglinton, T.I., Ohkouchi, N., Schneider, R.R., Muller, P.J., Grootes, P.M., Rullkotter, J., 2003. Asynchronous alkenone and foraminfera records from the Benguela Upwelling System. Geochim. Cosmochim. Acta 67, 2157–2171. Mollenhauer, G., Inthorn, M., Vogt, T., Zabel, M., Sinninghe Damste´, J.S.S., 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. ¨ Muller, P.J., Kirst, G., Ruhland, G., von Storch, I., Rosell-Mele, A., 1998. Calibration of the alkenone paleotemperature index U-37(K0 ) based on core-tops from the eastern South Atlantic and the global ocean (601N–601S). Geochim. Cosmochim. Acta 62, 1757–1772. Ohkouchi, N., Eglinton, T.I., Keigwin, L.D., 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., de Ruijter, W.P.M., 2006. A link between lowfrequency mesoscale eddy variability around Madagascar and the large-scale Indian Ocean variability. J. Geophys. Res.-Oceans 111. Pope, R.H., Demaster, D.J., Smith, C.R., Seltmann, H., 1996. Rapid bioturbation in equatorial Pacific sediments: Evidence from excess Th-234 measurements. Deep-Sea Res. II Top. Stud. Oceanogr. 43, 1339–1364. Powers, L.A., Johnson, T.C., Werne, J.P., Castaneda, I.S., Hopmans, E.C., Sinninghe Damste´, J.S.S., Schouten, S., 2011. Organic geochemical records of environmental variability in Lake Malawi during the last 700 years, Part I: The TEX86 temperature record. Palaeogeogr. Palaeoclimatol. Palaeoecol. 303, 133–139. ¨ Prahl, F.G., Muhlhausen, L.A., Zahnle, D.L., 1988. Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions. Geochim. Cosmochim. Acta 52, 2303–2310. Ransom, B., Shea, K.F., Burkett, P.J., Bennett, R.H., Baerwald, R., 1998. Comparison of pelagic and nepheloid layer marine snow: implications for carbon cycling. Mar. Geol. 150, 39–50. ¨ ¨ Regenberg, M., Nurnberg, D., Steph, S., Groeneveld, J., Garbe-Schonberg, D., Tiedemann, R., Dullo, W.C., 2006. Assessing the effect of dissolution on planktonic foraminiferal Mg/Ca ratios: evidence from Caribbean core tops. Geochem. Geophys. Geosyst., 7. Reimer, P., Baillie, M., Bard, E., Bayliss, A., Beck, J., Blackwell, P., Ramsey, C., Buck, C., Burr, G., Edwards, R., Friedrich, M., Grootes, P., Guilderson, T., Hajdas, I., Heaton, T., Hogg, A., Hughen, K., Kaiser, K., Kromer, B., McCormac, F., Manning, S., Reimer, R., Richards, D., Southon, J., Talamo, S., Turney, C., van der Plicht, J., Weyhenmeye, C., 2009. Intcal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 51, 1111–1150. Reuer, M.K., Weiss, D.J., 2002. Anthropogenic lead dynamics in the terrestrial and marine environment. Philos. Trans. R Soc. London Ser. A—Math. Phys. Eng. Sci. 360, 2889–2904. ¨ sterreichischen Reuss, A.E., 1850. Neue Foraminiferen aus den Schichten des O ¨ Tertiarbeckens, 365–390. Richter, T.O., de Stigter, H.C., Boer, W., Jesus, C.C., van Weering, T.C.E., 2009. Dispersal of natural and anthropogenic lead through submarine canyons at the Portuguese margin. Deep-Sea Res. I: Oceanogr. Res. Papers 56, 267–282. Romero, O.E., Boeckel, B., Donner, B., Lavik, G., Fischer, G., Wefer, G., 2002. Seasonal productivity dynamics in the pelagic central Benguela System inferred from the flux of carbonate and silicate organisms. J. Marine Syst. 37, 259–278. Russell, A.D., Honisch, B., Spero, H.J., Lea, D.W., 2004. Effects of seawater carbonate ion concentration and temperature on shell U, Mg, and Sr in cultured planktonic foraminifera. Geochim. Cosmochim. Acta 68, 4347–4361. Sautter, L.R., Thunell, R.C., 1991. Planktonic foraminiferal response to upwelling and seasonal hydrographic conditions—sediment trap results from San-Pedro Basin, Southern California Bight. J. Foraminiferal Res. 21, 347–363. Schouten, S., Hopmans, E.C., Schefuss, E., 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., Sinninghe Damste´, J.S.S., 2007. Analytical methodology for TEX86 paleothermometry by high-performance liquid chromatography/atmospheric pressure chemical ionizationmass spectrometry. Anal. Chem. 79, 2940–2944. Sell, D.W., Evans, M.S., 1982. A statistical analysis of subsampling and an evaluation of the Folsom plankton splitter. Hydrobiologia 94, 223–230. Shackleton, N.J., 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., Mazaud, A., 2005. Mid-latitude Southern Indian Ocean response to Northern Hemisphere Heinrich events. Earth Planet. Sci. Lett. 240, 724–731. Smith, C.R., Rabouille, C., 2002. What controls the mixed-layer depth in deep-sea sediments? The importance of POC flux. Limnol. Oceanogr. 47, 418–426. Soetaert, K., Herman, P.M.J., Middelburg, J.J., 1996. A model of early diagenetic processes from the shelf to abyssal depths. Geochim. Cosmochim. Acta 60, 1019–1040.

U. Fallet et al. / Deep-Sea Research I 59 (2012) 37–53

Southon, J., Kashgarian, M., Fontugne, M., Metivier, B., Yim, W.W.S., 2002. Marine reservoir corrections for the Indian Ocean and southeast Asia. Radiocarbon 44, 167–180. Spero, H.J., Bijma, J., Lea, D.W., Bemis, B.E., 1997. Effect of seawater carbonate concentration on foraminiferal carbon and oxygen isotopes. Nature 390, 497–500. Stuiver M. and Reimer P.J. (1993) Extended 14C Data base and Revised CALIB 3.0 14C Age Calibration Program, pp. 215–230. Tachikawa, K., Sepulcre, S., Toyofuku, T., Bard, E., 2008. Assessing influence of diagenetic carbonate dissolution on planktonic foraminiferal Mg/Ca in the southeastern Arabian Sea over the past 450 ka: comparison between Globigerinoides ruber and Globigerinoides sacculifer. Geochem. Geophys. Geosyst. 9, Q04037. Takahashi, K., Be´, A.W.H., 1984. Planktonic foraminifera—factors controlling sinking speeds. Deep-Sea Res. A 31, 1477–1500. Thomsen, L., Gust, G., 2000. Sediment erosion thresholds and characteristics of resuspended aggregates on the western European continental margin. DeepSea Res. I Oceanogr. Res. Papers 47, 1881–1897. Thunell, R., Sautter, L.R., 1992. Planktonic Foraminiferal Faunal and Stable Isotopic Indices of Upwelling: A Sediment Trap Study in the San Pedro Basin, Southern California Bight. Geological Society, London, Special Publications 64, 77–91. Tierney, J.E., Mayes, M.T., Meyer, N., Ch., Johnson, Swarzenski, P.W., Cohen, A.S., Russell, J.M., 2010. Late-twentieth-century warming in Lake Tanganyika unprecedented since AD 500. Nat. Geosci. 3, 422–425.

53

Trauth, M.H., Sarnthein, M., Arnold, M., 1997. Bioturbational mixing depth and carbon flux at the seafloor. Paleoceanography 12, 517–526. Vallelonga, P., Van de Velde, K., Candelone, J.P., Morgan, V.I., Boutron, C.F., Rosman, K.J.R., 2002. The lead pollution history of Law Dome, Antarctica, from isotopic measurements on ice cores: 1500 AD to 1989 AD. Earth Planet. Sci. Lett. 204, 291–306. Van de Velde, K., Vallelonga, P., Candelone, J.P., Rosman, K.J.R., Gaspari, V., Cozzi, G., Barbante, C., Udisti, R., Cescon, P., Boutron, C.F., 2005. Pb isotope record over one century in snow from Victoria Land, Antarctica. Earth Planet. Sci. Lett. 232, 95–108. van der Werf, P.M., Schouten, M.W., van Leeuwen, P.J., Ridderinkhof, H., de Ruijter, W.P.M., 2009. Observation and origin of an interannual salinity anomaly in the Mozambique Channel. J. Geophys. Res.-Oceans 114. Wakeham, S.G., Lee, C., Peterson, M.L., Liu, Z.F., Szlosek, J., Putnam, I.F., Xue, J.H., 2009. Organic biomarkers in the twilight zone–time series and settling velocity sediment traps during MedFlux. Deep-Sea Res. II Top. Stud. Oceanogr. 56, 1437–1453. Weiss, D., Shotyk, W., Kempf, O., 1999. Archives of atmospheric lead pollution. Naturwissenschaften 86, 262–275. Wuchter, C., Schouten, S., Coolen, M.J.L., Sinninghe Damste´, J.S.S., 2004. Temperaturedependent variation in the distribution of tetraether membrane lipids of marine Crenarchaeota: implications for TEX86 paleothermometry. Paleoceanography, 19.