Distribution of intact and core tetraether lipids in water column profiles of suspended particulate...

Organic Geochemistry 72 (2014) 1–13

Contents lists available at ScienceDirect

Organic Geochemistry

journal homepage: www.elsevier .com/locate /orggeochem

Distribution of intact and core tetraether lipids in water column profilesof suspended particulate matter off Cape Blanc, NW Africa

http://dx.doi.org/10.1016/j.orggeochem.2014.04.0070146-6380/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: MARUM Center for Marine Environmental Sciences,Department of Geosciences, University of Bremen, D-28359 Bremen, Germany. Tel.:+49 1787803644.

E-mail address: abasse@marum.de (A. Basse).

Andreas Basse a,b,⇑, Chun Zhu b, Gerard J.M. Versteegh b, Gerhard Fischer b, Kai-Uwe Hinrichs b,Gesine Mollenhauer a,b

a Alfred-Wegener-Institut Helmholtz-Zentrum für Polar- und Meeresforschung (AWI), D-27570 Bremerhaven, Germanyb MARUM Center for Marine Environmental Sciences, Department of Geosciences, University of Bremen, D-28359 Bremen, Germany

a r t i c l e i n f o

Article history:Received 9 April 2014Accepted 9 April 2014Available online 24 April 2014

Keywords:GDGTArchaeaTEX86

Suspendet particulate matterNW-AfricaUpwellingIntact polar lipidsOxygen minimum zoneNepheloid layersIn situ production

a b s t r a c t

In the reconstruction of sea surface temperature (SST) from sedimentary archives, secondary sources, lat-eral transport and selective preservation are considered to be mainly negligible in terms of influencingthe primary signal. This is also true for the archaeal glycerol dialkyl glycerol tetraethers (GDGTs) thatform the basis for the TEX86 SST proxy. Our samples represent four years variability on a transect off CapeBlanc (NW Africa). We studied the subsurface production, vertical and lateral transport of intact polar lip-ids and core GDGTs in the water column at high vertical resolution on the basis of suspended particulatematter (SPM) samples from the photic zone, the subsurface oxygen minimum zone (OMZ), nepheloid lay-ers (NL) and the water column between these. Furthermore we compared the water column SPM GDGTcomposition with that in underlying surface sediments. This is the first study that reports TEX86 valuesfrom the precursor intact polar lipids (IPLs) associated with specific head groups (IPL-specific TEX86).We show a clear deviation from the sea surface GDGT composition in the OMZ between 300 and600 m. Since neither lateral transport nor selective degradation provides a satisfactory explanation forthe observed TEX-derived temperature profiles, with a bias towards higher temperatures for both core-and IPL-specific TEX86 values, we suggest that subsurface in situ production of archaea with a distinctrelationship between lipid biosynthesis and temperature is a responsible mechanism. However, in theNW-African upwelling system the GDGT contribution of the OMZ to the surface sediments did not seemto affect the sedimentary TEX86 as it showed no bias and still reflected the signal of the surface watersbetween 0 and 60 m.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Marine planktic Thaumarchaeota produce glycerol dibipytanylglycerol tetraethers (GDGTs) as their major membrane buildingblocks. These lipids occur in the natural environment as intactpolar lipids (IPLs) with glycosidic and/or phosphate head groupsattached to the GDGT core, or as core lipids (CLs) without thesehead groups. Core lipid GDGTs (CL-GDGTs) have the potential tobe preserved in sediments up to hundreds of millions of years(Kuypers et al., 2001; Jenkyns et al., 2012). They may contain oneto four cyclopentyl moieties and up to one cyclohexyl moiety inthe isoprenoid chains. The relative abundance of the cyclopentylmoieties is expressed in the TEX86 proxy (Schouten et al., 2002).

Since core-TEX86 for surface sediments mainly correlates well withthe annual mean sea surface temperature or integrated tempera-ture of the upper water column (Kim et al., 2008, 2010) it has beenwidely used to reconstruct sea surface temperature (SST) valuesfrom ancient sediments. The procedure assumes an accurate trans-fer of core-TEX86 from the overlying surface water through thewater column to the sediment and negligible influence from selec-tive preservation or lateral transport (Mollenhauer et al., 2008;Shah et al., 2008; Kim et al., 2009a) and from other sources ofGDGTs from e.g. deep cold water (e.g., Wuchter et al., 2005;Huguet et al., 2006a). The validity of these assumptions has beenprimarily empirical. It is based on correlations between sedimentsurface core-TEX86 values with satellite-derived SST data buthardly on the water column analysis required to assess the abovemodifying processes. There is only a relatively small number ofstudies providing comprehensive information on GDGTs insuspended particulate matter (SPM) in the water column, espe-cially > 1000 m water depth, available. The SPM data indicate that,

2 A. Basse et al. / Organic Geochemistry 72 (2014) 1–13

in the water column > 100 m, neither core-TEX86 (Wuchter et al.,2005; Ingalls, 2006; Turich et al., 2007; Schouten et al., 2012,2013; Taylor et al., 2013) nor IPL-derived TEX86, considered toreflect bulk IPL-TEX86 (Schouten et al., 2012), correlates within situ water temperature. This has been explained by the hypoth-esis that the CL-GDGTs from surface sediments are derived onlyfrom non-living matter transported from the upper water column(Wuchter et al., 2005). However, this is debable, and a ‘‘subsurface’’GDGT contribution has also been hypothesized (Huguet et al.,2006a; Lee et al., 2008; Schouten et al., 2013; Taylor et al., 2013).Lateral transport, e.g. in nepheloid layers, may introduce an allo-chthonous component to the sediment. As this material may orig-inate up to several hundred km from the site of sedimentation(Ohkouchi et al., 2002; Nowald et al., 2006), this component mightdiffer completely from the autochthonous signal and thus poten-tially cause strong bias. However, some evidence exists that trans-port processes do not affect GDGTs to the same extent as otherlipids like, e.g. alkenones (Mollenhauer et al., 2008; Shah et al.,2008).

Studies have shown that marine picoplankton communitieschange with depth: While the photic zone is primarily dominatedby bacteria, archaea contribute up to 50% or more to the totalmicrobial community in deep water (Karner et al., 2001; Teiraet al., 2004). Changes in archaeal ecotypes in deep vs. shallowwater have been observed (Francis et al., 2005; Hallam et al.,2006; Mincer et al., 2007). Variation in O2 concentration may alsohave an influence on the archaeal community (Herndl et al., 2005;Labrenz et al., 2010; Loescher et al., 2012). This is consistent withobservations of anoxic or suboxic environments where the TEX86

signal in the water column correlates neither with in situ nor withsurface temperature (Schouten et al., 2012).

In general, intact polar GDGTs with head groups (IPL-GDGTs)were thought to be degraded rapidly after cell death (Harveyet al., 1989) and thus could be used as biomarkers for livingmicrobes (Sturt et al., 2004; Biddle, 2006; Biddle et al., 2006; Lippand Hinrichs, 2009; Schubotz et al., 2009). However, more recentlythe predominant glycosidic IPL-GDGTs have been suggested to beaffected by a fossil contribution, especially in anoxic environmentssuch as sediments, and may even be preserved on geologicaltimescales (Schouten et al., 2010). It is still under discussion whetherdifferences in the lability of IPL-GDGTs with different head groupscan significantly influence core-GDGT composition in the water col-umn and sediments (Lengger et al., 2012a,b; Schouten et al., 2012).

To obtain a better insight into the potential modification of theGDGT composition and TEX86 by lateral transport, in situ produc-tion and differential degradation, we collected SPM from the watercolumn at different depths at multiple stations along an onshore–offshore transect off the upwelling area near Cape Blanc (Fig. 1a) in

Fig. 1. Map of study area off Cape Blanc with sampling stations (dark gray circles) on tintermediate nepheloid layer (INL) and bottom near layer (BL) in the water column alonscales are not identical in (a and b).

four consecutive years. For the first time not only core-TEX86 val-ues but also TEX86 values from IPL groups associated with specifichead groups (IPL-specific TEX86) were systematically examined inorganic matter (OM) from different water depths by applying anew high performance liquid chromatography–mass spectrometry(HPLC–MS) protocol (Zhu et al., 2013). This allowed us to constrainthe influences of (i) productivity along the onshore–offshore gradi-ent, (ii) lateral transport by comparison with samples from neph-eloid layers, (iii) degradation by comparing samples fromdifferent water depths, (iv) in situ production through differencesin IPL composition and (v) low O2 concentrations within the localoxygen minimum zone (OMZ).

2. Material and methods

2.1. Study area

The northwest African upwelling system is one of the four larg-est eastern boundary upwelling systems in the world and is char-acterized by high productivity (Carr, 2001). A considerable part ofthe biomass produced in the photic zone sinks to the deep sea inthe form of aggregates, fecal pellets and marine snow (Nowaldet al., 2006; Fischer and Karakas, 2008, 2009). On its way, the sink-ing OM undergoes degradation and remineralization, wherebymicroorganisms degrade up to 13% of the organic carbon (OC)per day (Iversen et al., 2010). As a result, the concentration of lipidsgenerally decreases exponentially with water depth (Martin et al.,1987; Sarmiento, 1993). Nevertheless, a relevant amount of OMstill reaches the seafloor and is buried in the deep-sea sediments,which contain 0.4–2% OC (Seiter et al., 2004). Long term observa-tions (Ratmeyer et al., 1999; Nowald et al., 2006) with sedimenttraps and particle cameras (ParCa) as well turbidity sensors showthat transport is not only vertical. Particles are also transported lat-erally (Karakas et al., 2009), mainly in nepheloid layers (NLs) con-taining a significantly higher concentration of particles than thewater column in between. In particular, an intermediate nepheloidlayer (INL) at ca. 250–600 m water depth and a bottom layer (BL) at50–100 m above the seafloor could be detected during severalyears of observation (Müller and Fischer, 2001; Karakas et al.,2006; Fischer and Karakas, 2009). The INL seems to originate atthe shelf break and seems to be derived at least partly from re-suspended shelf material. The BL generates particle ‘‘clouds’’detaching from the slope at 1900–2800 m.

2.2. Sampling

Samples were taken from an east–west transect off Cape Blancat three main stations (CB, CBi and Slope) during four cruises: MSM

he offshore transect (a) and general distribution of chlorophyll maximum (Chlmax),g the offshore transect derived from CTD–turbidity-sensor data (b). Note: horizontal

A. Basse et al. / Organic Geochemistry 72 (2014) 1–13 3

11-2 from 2009/03/14–2009/04/09, POS 396 from 2010/02/24 to2010/03/08, MSM 18-1 from 2011/04/17 to 2011/05/05 and POS425 from 2012/01/16 to 2010/01/30. In 2009, 2011 and 2012, thetransect also included additional sampling sites located betweenthe main stations (Fig. 1a, Table 1). During all the cruises, depthprofiles of SPM were collected via in situ filtration using battery-powered pumps (WTS 6-1-142LV; McLane Research Laboratories,Falmouth, MA) for 3–6 h at each depth, corresponding to 350–2500 l (Table 1). Sampling volumes were determined by the pumpcontrol software and via a mechanical flowmeter fixed to the pumpoutlet. All samples were collected on pre-combusted 142 mm/0.7 lm Whatman GF/F glass fiber filters. In addition, surface waterSPM samples were taken from the ships seawater inlets (90–200l/filter, determined via a mechanical flow meter) during the MariaS. Merian cruises MSM 11-2 and MSM 18-1 as well as Poseidoncruise POS 396. During POS 425, water samples (ca. 40 ml each)for nutrient measurements were taken from the same waterdepths as the in situ pump samples using Niskin bottles on arosette water sampler. All surface sediment samples (0–2 cm) usedwere taken during MSM 11-2 with a multi-coring system (Table 1).

2.3. Water temperature, salinity, oxygen, turbidity, chlorophyllfluorescence

Water temperature, salinity, O2, turbidity and chlorophyll fluo-rescence were measured with a self-contained SBE 19 CTD profilerequipped with a conductivity-temperature-depth probe plus O2

sensor, a CHELSEA-fluorometer and a WETLABS turbidity sensor(no turbidity sensor was available during MSM 11-2). SatelliteSST values and 40-yr-mean SST were retrieved from the IGOSS-database (Reynolds et al., 2002).

2.4. Nitrate

Water samples for NO3� measurement were only taken during

POS 425. Analyses was performed using ion chromatography witha Metrohm 882 Compact IC plus at the University of Bremen. Themethod is described in DIN EN ISO 10304-1 (2008).

2.5. Lipid extraction

Filters were stored at �20 �C immediately after recovery andair-dried (at 36 �C, 24 h) immediately before extraction. Fourpieces were cut out of the dried GF/F filters with a broach (diame-ter = 12 mm) for total OC content (TOC) and 14C analysis, assumingthat the composition of the cut out pieces was representative of thewhole filter. Lipids were extracted from the remaining part of thefilters and from freeze-dried and homogenized sediments using amodified method after Müller et al. (1998) with ultra sonicationusing an UP 200H sonic disruptor probe (200 W, amplitude 0.5,pulse 0.5) with successively less polar solvents: MeOH, MeOH/dichloromethane (DCM) (1:1 v/v) and DCM, each for 5 min. Aninternal C46 GDGT standard was added before extraction. Onesub-sample of a internal laboratory sediment standard wasextracted every 11 samples using the same extraction method.Each combined total lipid extract (TLE) was desalted and an aliquotof 25% was stored for IPL-GDGT analysis. The rest of the TLE wassaponified with 300 ll of 0.1 M KOH in MeOH with 10% H2O at80 �C for 2 h. The solution was evaporated using dried N2 and theneutral lipids were extracted into hexane five times. The hexanefraction was separated into three fractions of increasing polarityusing self-packed silica gel columns (1% H2O deactivated SiO2,0.063–0.2 mm mesh size, column: £ 6 mm � 4 cm), elutinghydrocarbons with 4 ml hexane, ketones with 4 ml hexane:DCM(1:2 v/v) and alcohols with 4 ml MeOH.

2.6. TOC

TOC content of SPM samples was assessed using an ELEMENTARvario EL element analyzer in CN-mode after removal of CO3

� with6 N HCL as described by Müller et al. (1994). Carbon concentrationwas normalized to the sample volume equivalent of each filterpiece.

2.7. Core lipid GDGT analysis

For core lipid GDGT (CL-GDGT) analysis we used a modifiedliquid chromatography method derived from Hopmans et al.(2000). The MeOH fraction was dried under N2 and dissolved inhexane:isopropanol (99:1; v/v) at a concentration of 2 mg/ml andthen filtered through a 4 lm diameter PTFE filter (0.45 lm poresize) to prevent clogging of the column. An aliquot (20 ll) of eachsample was injected onto a high performance liquid chromatogra-phy (HPLC) system (Agilent 1200 series HPLC system) coupled toan Agilent 6120 MSD mass spectrometry (MS) instrument, operat-ing with atmospheric pressure chemical ionization (APCI). ForGDGT separation we used a Prevail Cyano 3 lm column (GRACE150 mm � 2.1 mm) at 30 �C. Samples were eluted with a mixtureof solvent A (100%) and solvent B (5% isopropanol 95% hexane).Flow was 0.2 ml/min, with 20% of solvent B from 0–5 min, then lin-early increasing to 36% of solvent B from 5–45 min. The columnwas back flushed with 100% solvent B for 8 min after each analysis.APCI-MS was run with following spray chamber conditions: Dryinggas flow 5 l/min and 350 �C, nebulizer 50 psi, vaporizer gas 350 �C,capillary voltage �3 kV and corona current +5 lA. GDGTs weredetected using single ion monitoring (SIM) of the [M+H]+ ions(dwell time 67 ms) for m/z 744.0 (internal C46 standard), 1022.0,1036.0, 1050.0, 1292.3, 1296.3, 1298.3, 1300.3 and 1302.3.

The TEX86 (core and IPL-specific) values were calculated on thebasis of the relative abundances of GDGTs (Schouten et al., 2002)using the respective peak areas [Eq. (1)]. SST values werecalculated according to the core top calibration equation of Kimet al. (2010) [Eq. (2)].

TEX86 ¼ ½GDGT2� þ ½GDGT3�þ ½crenarcheol isomer�=½GDGT1� þ ½GDGT2�þ ½GDGT3� þ ½crenarcheol isomer� ð1Þ

A standard deviation of 0.01 units for TEX86 determination wascalculated from 47 measurements of GDGTs extracted from a labinternal standard sediment between December 2009 andMarch 2012.

SST ¼ 68:4� log TEX86 þ 38:6 ð2Þ

Concentration of core lipids was calculated relative to the C46-GDGT internal standard, assuming equal response factors. Due tothe unavailability of purified C86-GDGTs this assumption was nec-essary. As a result, the concentration values calculated here cannotbe regarded as absolute values (cf. Huguet et al., 2006b).

2.8. IPL specific-TEX86 analysis

A newly developed reverse phase (RP) liquid chromatography-electrospray ionization-MS (RP-ESI-MS) protocol (Zhu et al.,2013) was used to measure CL-GDGTs and IPL-GDGTs and TEX86

values from CL-GDGTs (core-TEX86, RP) and from specific IPL classesclustered by different head groups (IPL-specific TEX86). In brief,samples from POS 425 and 396 were analyzed with an Agilent1200 series HPLC instrument coupled to an Agilent 6130 MSDMS instrument via an ESI interface. Samples from MSM 18-1 and11-2 were analyzed using a Dionex Ultimate 3000 UHPLC coupledto a BrukermaXis Ultra High Resolution orthogonal accelerated

Table 1Sample data (N.A., not available).

Cruise Sample number Station Depth (m) Lat (�N) Long (�W) Volume (l)

POS 396 POS 396 OW4 CB 5 21.28 �21 77POS 396 POS 396 OW5 CB 5 21.26 �21 93POS 396 POS396 ISP1 CB 45 21.29 �21 811POS 396 POS396 ISP2 CB 200 21.29 �21 870POS 396 POS396 ISP3 CB 500 21.29 �21 N.A.POS 396 POS396 ISP4 CB 700 21.29 �21 1730POS 396 POS396 ISP5 CB 1500 21.29 �21 1757POS 396 POS396 ISP6 CB 2500 21.29 �21 1456POS 396 POS396 ISP7 CB 3500 21.29 �21 2028POS 396 POS396 ISP8 CB 4120 21.29 �21 1416POS 396 POS 396 OW6 Cbi 5 20.74 �19 74POS 396 POS396 ISP9 CBi 45 20.72 �19 411POS 396 POS396 ISP10 CBi 150 20.72 �19 591POS 396 POS396 ISP11 CBi 300 20.72 �19 1272POS 396 POS396 ISP12 CBi 1000 20.72 �19 1096POS 396 POS396 ISP13 CBi 1900 20.72 �19 1183POS 396 POS396 ISP14 CBi 2000 20.72 �19 1040POS 396 POS396 ISP15 CBi 2200 20.72 �19 1138POS 396 POS396 ISP16 CBi 2620 20.72 �19 1103POS 396 POS 396 OW7 Slope 5 20.59 �18 51POS 396 POS396 ISP17 Slope 45 20.58 �18 636POS 396 POS396 ISP18 Slope 100 20.58 �18 451POS 396 POS396 ISP19 Slope 170 20.58 �18 984POS 396 POS396 ISP20 Slope 300 20.58 �18 946POS 396 POS396 ISP21 Slope 600 20.58 �18 840POS 396 POS396 ISP22 Slope 720 20.58 �18 1220MSM 11-2 MSM11-2 OW 20 Slope 5 20.58 �18 151MSM 11-2 MSM11-2 ISP1 Slope 55 20.58 �18 575MSM 11-2 MSM11-2 ISP2 Slope 280 20.58 �18 1116MSM 11-2 MSM11-2 ISP3 Slope 370 20.58 �18 690MSM 11-2 MSM11-2 ISP4 Slope 720 20.58 �18 1057MSM 11-2 GeoB 13613-3 0-1cm Slope 739 20.58 �18 SedimentMSM 11-2 MSM11-2 OW 19 CBi 5 20.77 �19 115MSM 11-2 MSM11-2 ISP5 CBi 55 20.77 �19 1274MSM 11-2 MSM11-2 ISP6 CBi 400 20.77 �19 1344MSM 11-2 MSM11-2 ISP7 CBi 1095 20.77 �19 1370MSM 11-2 MSM11-2 ISP8 CBi 2640 20.77 �19 1196MSM 11-2 GeoB 13612-3 0-1cm CBi 2690 20.77 �19 SedimentMSM 11-2 MSM11-2 ISP9 Between CB&CBi 45 21 �20 1236MSM 11-2 MSM11-2 ISP10 Between CB&CBi 400 21 �20 1333MSM 11-2 MSM11-2 ISP11 Between CB&CBi 1800 21 �20 1324MSM 11-2 MSM11-2 ISP12 Between CB&CBi 3683 21 �20 1147MSM 11-2 MSM11-2 OW 22 CB 5 21.26 �21 156MSM 11-2 MSM11-2 OW 23 CB 5 21.63 �20 300MSM 11-2 MSM11-2 ISP13 CB 45 21.26 �21 1224MSM 11-2 MSM11-2 ISP14 CB 450 21.26 �21 1140MSM 11-2 MSM11-2 ISP15 CB 625 21.26 �21 1348MSM 11-2 MSM11-2 ISP16 CB 4130 21.26 �21 1183MSM 11-2 GeoB 13616-6 0-1cm CB 4260 21.29 �21 SedimentPOS425 POS 425 ISP1 CB 55 21.27 �21 484POS425 POS 425 ISP2 CB 250 21.27 �21 1567POS425 POS 425 ISP3 CB 350 21.27 �21 870POS425 POS 425 ISP4 CB 1000 21.27 �21 821POS425 POS 425 ISP19 CBi 50 20.78 �19 348POS425 POS 425 ISP20 CBi 150 20.78 �19 1072POS425 POS 425 ISP21 CBi 280 20.78 �19 388POS425 POS 425 ISP22 CBi 350 20.78 �19 1559POS425 POS 425 ISP23 CBi 400 20.78 �19 980POS425 POS 425 ISP24 CBi 500 20.78 �19 869POS425 POS 425 ISP13 CBi 700 20.78 �19 1730POS425 POS 425 ISP16 CBi 2000 20.78 �19 1597POS425 POS 425 ISP17 CBi 2350 20.78 �19 1530POS425 POS 425 ISP18 CBi 2685 20.78 �19 654POS425 POS 425 ISP7 Slope 50 20.62 �18 377POS425 POS 425 ISP8 Slope 200 20.62 �18 785POS425 POS 425 ISP9 Slope 300 20.62 �18 1819POS425 POS 425 ISP10 Slope 450 20.62 �18 1338POS425 POS 425 ISP11 Slope 600 20.62 �18 2042POS425 POS 425 ISP12 Slope 720 20.62 �18 839POS425 POS 425 ISP25 Between CBi&Slope 50 20.67 �18 412POS425 POS 425 ISP26 Between CBi&Slope 150 20.67 �18 797POS425 POS 425 ISP27 Between CBi&Slope 280 20.67 �18 2POS425 POS 425 ISP28 Between CBi&Slope 600 20.67 �18 1753POS425 POS 425 ISP29 Between CBi&Slope 1150 20.67 �18 1480

4 A. Basse et al. / Organic Geochemistry 72 (2014) 1–13

Table 1 (continued)

Cruise Sample number Station Depth (m) Lat (�N) Long (�W) Volume (l)

POS425 POS 425 ISP30 Between CBi&Slope 1300 20.67 �18 12POS425 POS 425 ISP31 Between CB&CBi 50 20.92 �19 358POS425 POS 425 ISP32 Between CB&CBi 150 20.92 �19 903POS425 POS 425 ISP33 Between CB&CBi 400 20.92 �19 1128POS425 POS 425 ISP34 Between CB&CBi 1000 20.92 �19 220POS425 POS 425 ISP35 Between CB&CBi 2200 20.92 �19 1177POS425 POS 425 ISP36 Between CB&CBi 3400 20.92 �19 53MSM 18�1 MSM18-1 ISP1 CB 60 21.25 �21 944MSM 18-1 MSM18-1 ISP2 CB 130 21.25 �21 2471MSM 18-1 MSM18-1 ISP3 CB 390 21.25 �21 2380MSM 18-1 MSM18-1 ISP4 CB 880 21.25 �21 2392MSM 18-1 MSM18-1 ISP5 CB 2150 21.25 �21 754MSM 18-1 MSM18-1 ISP6 CB 3300 21.25 �21 741MSM 18-1 MSM18-1 ISP7 CBi 50 20.77 �19 217MSM 18-1 MSM18-1 ISP8 CBi 350 20.77 �19 1561MSM 18-1 MSM18-1 ISP9 CBi 960 20.77 �19 N.A.MSM 18-1 MSM18-1 ISP10 CBi 1250 20.77 �19 2198MSM 18-1 MSM18-1 ISP11 CBi 1900 20.77 �19 2273MSM 18-1 MSM18-1 ISP12 CBi 2150 20.77 �19 922MSM 18-1 MSM18-1 ISP13 Slope 20 20.59 �18 78MSM 18-1 MSM18-1 ISP14 Slope 150 20.59 �18 1925MSM 18-1 MSM18-1 ISP15 Slope 350 20.59 �18 1406MSM 18-1 MSM18-1 ISP16 Slope 450 20.59 �18 1716MSM 18-1 MSM18-1 ISP17 Slope 550 20.59 �18 526MSM 18-1 MSM18-1 ISP20 Between CB&CBi 120 21.01 �19 2077MSM 18-1 MSM18-1 ISP21 Between CB&CBi 440 21.01 �19 2147MSM 18-1 MSM18-1 ISP22 Between CB&CBi 1400 21.01 �19 2264MSM 18-1 MSM18-1 ISP23 Between CB&CBi 2600 21.01 �19 790

A. Basse et al. / Organic Geochemistry 72 (2014) 1–13 5

quadrupole-time-of-flight (qTOF) tandem MS/MS instrument,equipped with an ESI source (Bruker Daltonik, Bremen, Germany).For both instruments, ether lipid mixtures were dissolved in MeOHand separated with an ACE3 C18 column (3 lm, 2.1 � 150 mm;Advanced Chromatography Technologies Ltd., Aberdeen, Scotland)equipped with a guard cartridge and maintained at 45 �C. Ether lip-ids were eluted isocratically with 100% A for 10 min, followed by arapid gradient to 24% B in 5 min, and then a slow gradient to 65% Bin 55 min at 0.2 ml/min, where solvent A was 100:0.04:0.10MeOH/HCO2H/14.8 M NH3aq and B was 100:0.04:0.10 of 2-propa-nol/HCO2H/14.8 M NH3(aq). The column was washed with 90% Bfor 10 min and re-equilibrated with 100% A for another 10 min.

ESI-MS conditions on the qTOF instrument were optimized asfollows: capillary voltage 4500 V, nebulizing gas 0.8 bar and dryinggas 4 ml/min at 200 �C, in source collision-induced energy (ISCID)0 eV. Intact and core ether lipids were scanned in positive modefrom m/z 100–2000 without further fragmentation (for improvingpeak quality) at a scan rate of 1 Hz. However, in order to confirmthe ether lipid structures, typical samples were re-analyzed andscanned over the same range but with automated data-dependentfragmentation of the three most abundant ions. For the ESI-MSconditions on the MSD instrument, the optimum conditions were:capillary voltage 4000 V, nebulizing gas 4.14 bar (60 psi), dry gas5 ml/min at 200 �C, vaporizer 150 �C, fragmentor voltage 180 V.

The MS data obtained from the MSD or qTOF instrument wereprocessed through Data Analysis 4.0 software (BrukerDaltonik, Bre-men, Germany). Protonated [M+H]+, ammoniated [M+NH4]+ and[M+Na]+ ions are included for lipid peak integration. If detected,doubly charged ions such as [M+NH4+NH4]2+, [M+H+H]2+, and[M+NH4+Na]2+ were also included in the integration. Lipids wereidentified from retention times, m/z values and diagnostic frag-ments (Sturt et al., 2004). IPL-specific TEX86 values obtained fromthe RP-ESI-MS protocol were validated by Zhu et al. (2013). Thisnovel method did not allow determination of absolute concentra-tion values; so only relative concentration values are reported.

We studied the following IPL-GDGTs: monoglycosidic-GDGT (1G-GDGT); diglycosidic-GDGT (2G-GDGT); monoglycosidic-phospho-GDGT-monoglycosidic (1GP-GDGT-1G; or hexose-phosphohexose

(HPH)-GDGTs in Pitcher et al., 2011a). Fig. 2 illustrates structuresof head groups (Fig. 2a) and core-GDGTs (Fig. 2b). Typical reversephase chromatograms are shown in Fig. 3.

2.9. Four year weighted mean core TEX86 for water column SPM

The 4 yr weighted mean core-TEX86 for each station(wmTEX86

Station) was calculated from the mean of all core TEX86 val-ues obtained for all depths at one station from all four cruises,weighted for the water depth interval (Di) defined for each pumpas spanning from half way between the next higher and lowerpumps and assuming uniform core GDGT abundance within thatdepth interval. Initially, the hypothetical contribution from an indi-vidual water depth interval to the total TEX86 signal at one stationis calculated as

TEXDi86 ¼ ðDi � Cpump=Dtot � CtotÞ

X1�n

Di � Cpump

!, ,ðDtot � CtotÞ

ð3Þ

Subsequently, the weighted mean core-TEX86 at one station isdetermined as

wTEXStation86 ¼

X1�n

TEXDi86 ð4Þ

Finally, the mean core-TEX86 of station for all years is

wmTEXStation86 ¼

X2009�2012

wTEXStation86 : 4 ð5Þ

For the above equations, the following nomenclature was used: Dm isthe mid-point between the deployment depth of one pump and thatof the next deeper pump. Di is the depth interval for which a specificpump is taken as representative and is calculated as the differencebetween Dm for this pump and Dm for the next shallower pump [forthe shallowest and deepest pumps, the upper and lower limits of Di

are the surface (0 m) and the water depth at the station (Dtot),respectively]. Cpump is the summed concentration of the GDGTsused in the calculation of TEX86 (i.e. [GDGT1] + [GDGT2] +

Fig. 2. Structures of head groups (a) and core-GDGTs (b).

Fig. 3. Partial chromatograms of 1G-GDGTs (a), 2G-GDGTs (b) and 1G-GDGTs-PI (c) determined with RP(C18)-ESI-MS(qTOF) (sample: POS 396 ISP11, 300 m). Lipids werescanned using [M+H]+, [M+NH4]+ and [M+Na]+, with peaks of signal to noise (S/N) > 5 integrated and filled in gray.

6 A. Basse et al. / Organic Geochemistry 72 (2014) 1–13

[GDGT3] + [crenarchaeol isomer]) measured from the sample fromone pump. Ctot is the sum of all Cpump values at one stationði:e:

P1�nCpumpÞ. Numbers (1–n) refer to the shallowest to deepest

pumps deployed at one station.

For the discussion, we used only the weighted mean value forall 4 yr, as this value was compared with sediment values thatwere integrated over many years. Besides, interannual differenceswere negligible (see data in Supplementary Table).

Fig. 4. Summary of CTD-O2 concentration profiles from all cruises and stations.Note that O2 content is expressed as relative values due to lack of CTD calibrationbetween cruises. Light gray box shows depth interval of the INL. See main text forfurther description.

A. Basse et al. / Organic Geochemistry 72 (2014) 1–13 7

3. Results

3.1. Turbidity, O2 and NO3�

Turbidity sensor profiles from all cruises showed a distinct INLwith elevated particle concentration that seemed to originate atthe shelf break and then spread out westward between 250 and750 m water depth (Fig. 1b). A second high turbidity zone was

Fig. 5. Concentration of NO3�, O2 and core TEX86 in the water column during POS425

transect; sediment trap stations CB and CBi as well as the slope station are indicated. R

identified at the CBi station between 1900 and 2800 m waterdepth. During POS 425 we found nearly the same distribution ofNLs at additional stations between CB and CBi, and between CBiand the slope, supporting the inference that the INL formed a con-tinuous layer between all stations.

During all four cruises, a OMZ was observed between 150 and800 m water depth (Fig. 4). Minimum O2 concentration (0.8 ml/land 2.2 ml/l) was measured in the INL, between 300 and 600 m.The highest values (4.5–5.5 ml/l) occurred in surface waters. Sincethe O2 sensor was not calibrated continuously during the cruises,the measurements represent only relative concentration values.Our measured values were generally about 1 ml/l lower than theabsolute values stored in long term observation databases (fromGLODAP Gridded and Bottle Data Files; Key et al., 2004), so wecould assume safely that anoxic conditions were not present atany water depth.

NO3� concentration varied between 0.2 and 1.6 mg/l (Fig. 5). The

lowest values were observed in surface waters, except for the CBistation, where the signal was strongly scattered. At 2350 m NO3

�

was below detection limit. An absolute maximum value of4.8 mg/l was obtained for the deepest sample at 2690 m, fromthe BL. The deepest sample from the station between CBi and slopewas from the upper boundary of the BL and had a slightly increasedNO3� concentration (Fig. 5).

3.2. TOC, IPL- and core-GDGT concentrations

In the following, trends common to all stations in all 4 yr aredescribed (Fig. 6). TOC concentration decreased exponentially fromup 240 lg/l in the surface to values between 2 and 9 lg/l at below1000 m (Fig. 6c). At some stations, elevated values of up to 23 lg/lwere observed near the seafloor. We observed a slight increase inTOC up to 25 lg/l for the INL/OMZ. Here, the highest concentra-tions generally co-occurred with the core- and IPL-specific-TEX86

maxima (compare with Fig. 8) and the O2 minimum (compare withFig. 4). Slightly increased TOC values (up to 9 lg/l) were alsoobserved between 1500 and 2300 m, which corresponded to thedepth interval of the BL particle ‘‘clouds’’ (Fig. 1). The trend ofdecreasing concentration with depth, with some small excursionsto increased values in the INL/OMZ and BL particle ‘‘clouds’’, was

in January 2012. The five panels represent the stations sampled on the east–westelative O2 concentration is plotted on a reversed scale.

Fig. 6. Depth profile of concentration of IPL-GDGTs (a), core-GDGTs (b) and TOC (c); all cruises and stations.

Fig. 7. Differences in temperature between calculated core-TEX86 SST after Kim et al. (2010) and in situ water temperature from CTD (a), ship thermometer SST (b) and 40 yrmean SST (c) plotted vs. water depth. Different symbols are used for each cruise and for sediment samples. Gray shaded vertical bars and error bars show residual standarderror in the temperature calibration (�C, ±1). Horizontal gray shaded bar indicates depth interval of the INL.

8 A. Basse et al. / Organic Geochemistry 72 (2014) 1–13

also observed for all the individual lipids. The sums of concentra-tions of TEX86

H relevant core-GDGT lipids (Fig. 6b) ranged between0.04 and 2.30 ng/L. The highest concentrations values were in sub-surface waters between 100 and 200 m. Generally, the IPLs showedsimilar trends with depth (Fig. 6a); 2G-GDGTs and 1GP-GDGT-1Gshad their highest concentration in the subsurface between 100 and200 m, while the 1G-GDGTs showed maxima at the surfacebetween 20 and 60 m.

3.3. Core lipid GDGT TEX86 and BIT-index

In surface waters (0–60 m), core-TEX86-based temperatures val-ues after Kim et al. (2010) generally agreed with the in situ surfacewater temperature within the standard error of the calibration(Fig. 7a). Below the Chlmax (20–60 m) and above ca. 1000 m, a dis-tinct maximum in core-TEX86-values occurred, particularly in theINL and OMZ (Fig. 8). Here, core-TEX86 based temperature wasup to 12 �C higher than SST values obtained with the ships’

thermometers (Fig. 7b), up to 8 �C higher than the 40 yr meanSST derived from the IGOSS database (Reynolds et al., 2002;Fig. 7c) and up to 21 �C higher than CTD in situ values (Fig. 7a).Below 1500 m, the core-TEX86 based values remained around20 �C higher than the in situ values. Simultaneously, for thesewater depths the core-TEX86 based temperatures fit well withthe 40-yr-mean SST within the residual standard error of the tem-perature calibration (Kim et al., 2010).

Results from reverse phase measurements [core-TEX86(RP)]were, with a few exceptions, within error bars of measurementsobtained with the conventional normal phase method [core-TEX86(NP)] (Fig. 8). BIT values could only be calculated for a fewSPM samples and the sediment. In most SPM samples, at leastone of the branched GDGTs used for the calculation of BIT wasbelow detection limit. For all of the SPM samples with sufficientbranched GDGTs for the calculation of BIT, values were < 0.01(for sediment < 0.03), indicating negligible soil organic mattercontribution to sedimentary GDGTs.

Fig. 8. Depth profiles of core-TEX86 calculated using normal phase (NP) and reverse phase (RP) HPLC methods, in addition to IPL-specific TEX86 values (different symbols foreach head group) and O2 concentration. Data are presented separately for each station and cruise. Relative O2 concentration is plotted on a reversed scale (values decreasingto the right). Gray shaded areas show the position of the INL.

A. Basse et al. / Organic Geochemistry 72 (2014) 1–13 9

Table 2Weighted mean core-TEX86 values for water column SPM and core-TEX86 for surfacesediment.

CB CBi Slope

SPM 0–200 m 0.59 0.58 0.56SPM without INL/OMZ 0.63 0.6 0.58SPM entire water column 0.65 0.61 0.61Sediment 0.6 0.61 0.58

Fig. 9. Fractional abundance of TEX86-relevant core-GDGTs.

10 A. Basse et al. / Organic Geochemistry 72 (2014) 1–13

3.4. Sediment and weighted mean core-TEX86 of water column SPM

Core-TEX86-SST values from the surface sediments fit well withthe 40 yr mean SST (Fig. 7). While weighted mean SPM core-TEX86

for the upper water column (0–200 m) was lower than that of thesediment, the weighted mean of the entire water column SPM wasslightly higher (Table 2). The best fit of sediment and weightedmean SPM core-TEX86 was observed when the values from theINL (250–600 m) were excluded (Table 2).

3.5. IPL-specific TEX86

The IPLs with different head groups, e.g. 1G-GDGTs, 2G-GDGTsand 1GP-GDGT-1G, (Fig. 2) each showed different but relativelyconstant offsets between their IPL-specific TEX86 and thecore-TEX86 above 1500 m (Fig. 8). The 1GP-GDGT-1G-TEX86 wasthe lowest, and, although the values scattered more than those ofthe core-TEX86, an increase in the INL can be observed. Below1500 m water depth the 1GP-GDGT1-1G 1GP-GDGT3-1G and1GP-crenarchaeol-1G regioisomer were below detection limit and,consequently, the corresponding TEX86 could not be calculated.1G-GDGT-TEX86 had intermediate but still lower values than core-TEX86 and showed more scatter, but the general trend to increasedvalues in the INL was still obvious. The 2G-GDGT-TEX86 showed aconstant offset to higher values compared with core-TEX86 above1500 m. Below 1500 m values decreased with depth and below2500–3000 m reached levels below core-TEX86 values (Fig. 8).

Systematic differences in the core and IPL-specific-TEX86 depth-distribution between the 4 yr and the different stations ofsampling could not be observed.

3.6. Fractional abundance of core-GDGTs

The core-GDGT distribution (Fig. 9) showed that the elevatedTEX86 values in the INL/OMZ co-occurred with a relative increasein fractional abundance of the crenarchaeol isomer (up to 23%)and GDGT2 (up to 48%) and a decrease in that of GDGT1 (downto 30%). In the surface, crenarchaeol isomer and GDGT2 were attheir lowest abundance (3% and 24%) and GDGT1 had its maximum(up to 57%). Below the INL/OMZ fractional abundances stayed rel-atively constant (crenarchaeol isomer. between 12% and 19%;GDGT2 around 39% and GDGT1 around 41%). GDGT3 was mostabundant (up to 18%) in the surface water and exponentiallydecreased with depth down to ca. 500 m. Below that it stayed con-stant at around 2–3%. Sediment values for the crenarchaeol isomer,GDGT1 and GDGT2 were similar to those of SPM at similar waterdepth (Fig. 9). GDGT3 fractional abundances in the slope and CBistation sediments were slightly higher than in the deep watercolumn.

4. Discussion

In samples from > 60 m water depth, we observed a consistentdiscrepancy between the temperature from SPM-core-TEX86 andtemperature from both in situ measurements and the IGOSS

database (Reynolds et al., 2002). Similar observations of SPM-TEX86 profiles were made in the Bermuda region and North Sea(Wuchter et al., 2005), the Mediterranean Sea and North Atlantic(Turich et al., 2007), the Arabian Sea (Schouten et al., 2012) andthe eastern tropical north Pacific. Fractional abundances of theGDGT2 and crenarchaeol isomer increased in concert with TEX86,in particular between 250 and 600 m, implying that the changesin TEX86 were governed by processes controlling the relative abun-dance of these two lipids. In contrast, GDGT3 seemed to play aminor role, as its total abundance was rather low and decreasedexponentially with depth down to ca. 500 m and stayed rather con-stant below. These observations suggest that SPM-core-TEX86 isinfluenced not only by temperature but by other processes suchas in situ production and/or lateral transport and/or selectivedegradation.

Since planktonic archaea inhabit the entire water column(Karner et al., 2001), in situ production could involve differentarchaeal community members and/or a changed metabolism inresponse to water depth-specific environmental conditions suchas concentration of O2 and nutrients (e.g. Francis et al., 2005;Herndl et al., 2005; Hallam et al., 2006; Mincer et al., 2007;Labrenz et al., 2010; Loescher et al., 2012). Since the small-sizedplanktonic archaeal cells cannot be quantitatively captured by0.7 lm filters (Pitcher et al., 2011a; Ingalls and Huguet, 2012;Schouten et al., 2012; Wakeham et al., 2012) this may bias the coreand IPL-specific TEX86 signal from SPM. Especially, the in situ sig-nal from living archaea might be underrepresented. However, theeffective pore size decreases during the filtration, and previousinvestigations indicated generally good agreement between lipidprofiles obtained from 0.7 lm filters and marker gene distributionsfrom 0.2 lm filters, suggesting that a representative fraction ofarchaea was still retained on the 0.7 lm filters (Pitcher et al.,2011b,c; Schouten et al., 2012). Moreover, as we used the same fil-ter type for all samples, any potential under sampling would affectall samples, so they remained comparable between depths, sitesand years, and with other studies.

A. Basse et al. / Organic Geochemistry 72 (2014) 1–13 11

4.1. Influence of laterally transported re-suspended organic matter

The observed distribution of the nepheloid layers fit remarkablywell with distributions modeled on the basis of ParCa-data, particleflux rate obtained during earlier cruises and with sediment trapdata (Karakas et al., 2006). Increased TEX86 values co-occurredwith increased particle abundance in the water column, suggestingthat the TEX86 signal may be allochthonous, imported with re-sus-pended organic-rich material and thus reflecting the TEX86 of thesource area of this material. The material is likely derived fromre-suspended sediment from the shelf and the shelf break. It istransported westward in an INL at 300–600 m water depth and anear bottom layer (BL) downslope, generating particle ‘‘clouds’’ at1900–2800 m water depth. The larger amount of particles in theNLs also coincides with high TOC and lipid concentration, implyingthat the NLs have a large potential to maintain their ‘‘own’’ lipidsignal derived at least partly from resuspended shelf sediment.

However, the strong coastal upwelling on the shelf results inlower water temperature, such that lower, not higher, TEX86 valueswould be expected in the INLs if the GDGTs were laterally trans-ported from the shelf. We found TEX86 temperature between 25and 29 �C in the INL, while satellite SST on the shelf is 17–21 �Cduring upwelling. The annual mean SST values on the shelf arebetween 19.5 and 22 �C and the most likely temperatures reflectedby the re-suspended material, do not fit with the signal in the INL.Only in late summer does satellite SST on the shelf reach 25–27 �C(Reynolds et al., 2002). In addition, we did not observe increasedTEX86 values in the SPM samples from the BL and the particle‘‘clouds’’ at station CBi, which seem to derive from the same sourceas the INL. It is thus rather unlikely that a warm signal istransported from the shelf to the offshore stations within there-suspended organic matter.

4.2. Selective degradation of IPL-GDGTs?

Selective degradation between different CL-GDGTs, if present,does not seem to influence TEX86 (Huguet et al., 2009; Kim et al.,2009b; Bogus et al., 2012) and it is therefore unlikely to accountfor the TEX86 differences in the water column. Alternatively, pref-erential degradation of IPL-GDGTs with phosphate head groupsover those with glycosidic head groups during the sinking processcould lead to selective enrichment of certain core lipids andincrease TEX86 at some depths (Lengger et al., 2012a,b; Schoutenet al., 2012). Although the IPL-specific TEX86 values differ system-atically from the core-TEX86, the patterns of increase in the INL arenearly identical for all IPLs. Furthermore, the 1GP-GDGT-1G,assumed to be the most labile compound (Lengger et al.,2012a,b), exhibits much lower TEX86 values than the core lipids.Therefore, its degradation products should decrease core-TEX86

whereas we observe the opposite in the respective water depthinterval [assuming that all 1GP-GDGT-1G compounds degrade atthe same rate (cf. Schouten et al., 2004)]. This pattern stronglyargues against selective degradation of polar lipid precursors beingresponsible for anomalously warm TEX86 patterns in the watercolumn off Cape Blanc. This implies that the changes in TEX86

patterns in the deeper water column are likely strongly influencedby in situ production of the GDGT lipids below the photic zone,while processes such as differential degradation of selectedcompounds or sediment dynamics appear less important.

4.3. In situ production

After excluding lateral transport and selective degradation asmechanisms responsible for the discrepancy between SST andTEX86 derived temperature values at depth, in situ productionremains the most plausible explanation. The anti-correlation

between TEX86 values and O2 concentrations supports this. Else-where, O2 depletion also coincides with TEX86 values deviatingfrom SST values and, in these cases, a similar explanation has beensuggested (e.g. Blaga et al., 2008; Weijers et al., 2011; Liu et al.,2012). The correlation between the OMZ and the INL may beexplained by the additional consumption of O2 by degradingOrganic matter imported laterally through the INL. Compared withthe surface waters, the different environmental parameters such asreduced O2 concentration at depth may induce a change in thearchaeal community and/or its metabolism. It has been shown thatO2-deficient settings may result in a changed archaeal community(eg. Francis et al., 2005; Herndl et al., 2005; Labrenz et al., 2010;Loescher et al., 2012). This seems to be also implied by the NO3

�

maxima coinciding with the highest TEX86 values (core and IPL-specific) and the OMZ and INL (Fig. 4). However, NO3

� and TEX86

are decoupled in the particle ‘‘clouds’’ and the BL, leading to theconclusion that it might be a (unique) feature of the INL. It appearsthat low O2 conditions induce the elevated TEX86 signals, andco-occurring high NO3

� might be a hint for the influence of NH3

oxidizing archaea (AOA; Schouten et al., 2007; Turich et al.,2007). The relative increase in the crenarchaeol isomer in theOMZ/INL (Fig. 9) may be an additional indication of AOA. Tosubstantiate these hypotheses, further analyses (like DNA-basedcommunity analysis) are required. Taking into account that thein situ signal from living archaea might be underrepresented insamples taken with 0.7 lm filters (Pitcher et al., 2011a; Ingallsand Huguet, 2012; Schouten et al., 2012; Wakeham et al., 2012),it is likely that the true IPL-specific TEX86 signal in the INL/OMZis even higher than values observed in this study.

4.4. Effect on core-TEX86 in underlying sediments

The good fit of surface-sediment derived core-TEX86 based tem-perature values with long term mean SST values suggests that theelevated core-TEX86 signal in the INL/OMZ does not influence thecore-TEX86 of the material that reaches the seafloor. The weightedmean core-TEX86 of the water column SPM (Table 2) shows thatfor the upper 200 m the core-TEX86 values tend to be slightly colderthan sediment data. This is to be expected since all SPM sampleswere taken during late winter and early spring when SSTs were attheir minimum, while the sediments reflect mean SST over severalyears. The weighted mean SPM-core-TEX86 (including INL/OMZ) isslightly higher in water column samples than sediment samples.An overall slightly better agreement of the values is obtained whenthe INL/OMZ is excluded (Table 2). This implies that, although theelevated TEX86 values from the INL/OMZ could theoretically affectthe sediment signal, in practice they do not. Further support for thiscomes from the good agreement between fractional abundances ofthe individual core GDGTs in SPM and sediments. It is also in accordwith previous studies of the modern permanently anoxic Black Seaand Cariaco Basin, where core-TEX86 values in sinking particles andsurface sediments were found to reflect conditions in the near sur-face oxic layer (Wakeham et al., 2003,2004; Turich et al., 2013).Moreover, in the anoxic Santa Barbara Basin, a potential influenceof GDGTs produced in the deeper part of the water column led tolower rather than higher temperature estimates than SST (Huguetet al., 2007). These observations support previous inferences thata lack of particles (aggregates or fecal pellets) capable of sinkingand incorporating GDGTs formed at depth accounts for the underrepresentation of subsurface signals in the sediments (Wuchteret al., 2005; Kim et al., 2008; Huguet et al., 2006a).

Our study indicates that TEX86 values of sediments from areaswith highly productive surface waters can reliably be used toreconstruct long term mean SST. This is supported by the findingthat TEX86 values below the OMZ are within the statistical errorof those of sediments and surface waters. This would suggest that

12 A. Basse et al. / Organic Geochemistry 72 (2014) 1–13

in situ production of archaeal lipids has declined substantially anddoes not contribute significantly to SPM below this depth. Thismeans that, as hypothesized previously, SPM below 1000 m waterdepth is dominated by sinking ‘‘dead‘‘ material derived from thesurface waters.

5. Conclusions

On the basis of the analysis of intact and core-GDGT distribu-tions in water column profiles (from four cruises on a transect with3–5 stations) and surface sediments off Cape Blanc, NW Africa, wehave shown that, in the setting of a stable system of nepheloid lay-ers and an OMZ between 200 and 750 m water depth:

- The core-TEX86 derived temperature values for SPM below 60 mwater depth do not reflect the in situ water temperatures, sug-gesting that the signal either has not been produced at the sam-pling depth [and in case of the INL/OMZ also could not (only)derive from the surface water above it], or that the TEX-temper-ature transfer function is not valid for this depth, becausearchaea living here may produce GDGTs with different relativeabundances of cyclopentyl moieties (e.g. due to a different com-munity/metabolism).

- In the water column, the IPL-specific TEX86 values systemati-cally differ from, while being largely parallel to, the core-TEX86. This excludes selective degradation as a mechanismresponsible for the core-TEX86 trends through the water col-umn. Rather, the patterns of core- and IPL-TEX86 values are con-sistent with both production and degradation of IP-GDGTsin situ, i.e. intermediate core-TEX86 values likely derive fromhigh-TEX 2G-GDGT and low-TEX 1G-GDGT and 1GP-GDGT-1G.

- Elevated core- and IPL-specific-TEX86 occur specifically in theINL and OMZ. By analogy with other low O2 environments, wepropose that metabolically and possibly also genetically adaptedplanktonic archaeal communities (e.g. AOA crenarchaeota)hosted in these zones give rise to these elevated TEX86 values.

- Core-TEX86 values found for the sediments seem to originatefrom above 60 m water depth. The good correlation betweenthe core-TEX86 values from the deep waters and surface sedi-ment with the core-TEX86 values from the photic zone indicatethat the NL/OMZ GDGTs are not exported to the sediment anddeep water GDGT pools, a possible reason being that aggregatesand/or fecal pellets capable of efficient export of GDGTs to thesediment are not formed in the INL/OMZ.

Acknowledgements

We thank the captains and crews of RV Poseidon and RV MariaS. Merian. Laboratory assistance from R. Kreutz, R. Himmelsbach,and S. Pape is gratefully acknowledged as well as support fromJulius Lipp. Constructive reviews from K. Taylor and J. Sepúlveda,helped improve the manuscript. The research was funded by theGerman Research Foundation (DFG) – Center for Marine Environ-mental Sciences (MARUM) and the Alfred-Wegener-Institut Helm-holtz-Zentrum für Polar- und Meeresforschung (AWI).

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.orggeochem.2014.04.007.

Associate Editor—S. Schouten

References

Biddle, J.F., 2006. Microbial Populations and Processes in Subseafloor MarineEnvironments. PhD Thesis of the Pennsylvania State University.

Biddle, J.F., Lipp, J.S., Lever, M.A., Lloyd, K.G., Sørensen, K.B., Anderson, R., Fredricks,H.F., Elvert, M., Kelly, T.J., Schrag, D.P., Sogin, M.L., Brenchley, J.E., Teske, A.,House, C.H., Hinrichs, K.-U., 2006. Heterotrophic archaea dominate sedimentarysubsurface ecosystems off Peru. Proceedings of the National Academy ofSciences of the United States of America 103, 3846–3851.

Blaga, C.I., Reichart, G.-J., Heiri, O., Damste, J.S.S., 2008. Tetraether membrane lipiddistributions in water-column particulate matter and sediments: a study of 47European lakes along a north–south transect. Journal of Paleolimnology 41,523–540.

Bogus, K.A., Zonneveld, K.A.F., Fischer, D., Kasten, S., Bohrmann, G., Versteegh, G.J.M.,2012. The effect of meter-scale lateral oxygen gradients at the sediment–waterinterface on selected organic matter based alteration, productivity andtemperature proxies. Biogeosciences 9, 1553–1570.

Carr, M.-E., 2001. Estimation of potential productivity in Eastern Boundary Currentsusing remote sensing. Deep-Sea Research Part II 49, 59–80.

Fischer, G., Karakas, G., 2008. Sinking rates of particles in biogenic silica-andcarbonate-dominated production systems of the Atlantic Ocean: implicationsfor the organic carbon fluxes to the deep ocean. Biogeosciences Discussions 5,2541–2581.

Fischer, G., Karakas, G., 2009. Sinking rates and ballast composition of particles inthe Atlantic Ocean: implications for the organic carbon fluxes to the deep ocean.Biogeosciences 6, 85–102.

Francis, C.A., Roberts, K.J., Beman, J.M., Santoro, A.E., Oakley, B.B., 2005. Ubiquity anddiversity of ammonia-oxidizing archaea in water columns and sediments of theocean. Proceedings of the National Academy of Sciences of the United States ofAmerica 102, 14683–14688.

Hallam, S.J., Mincer, T.J., Schleper, C., Preston, C.M., Roberts, K., Richardson, P.M.,DeLong, E.F., 2006. Pathways of carbon assimilation and ammonia oxidationsuggested by environmental genomic analyses of marine Crenarchaeota. PLoSBiology 4, 520–536.

Harvey, H., Ohara, S., Eglinton, G., Corner, E., 1989. The comparative fate ofdinosterol and cholesterol in copepod feeding: implications for a conservativemolecular biomarker in the marine water column. Organic Geochemistry 14,635–641.

Herndl, G.J., Reinthaler, T., Teira, E., van Aken, H., Veth, C., Pernthaler, A., Pernthaler,J., 2005. Contribution of archaea to total prokaryotic production in the deepAtlantic Ocean. Applied and Environmental Microbiology 71, 2303–2309.

Hopmans, E.C., Schouten, S., Pancost, R.D., van der Meer, M.T.J., Sinninghe Damsté,J.S., 2000. Analysis of intact tetraether lipids in archaeal cell material andsediments by high performance liquid chromatography/atmospheric pressurechemical ionization mass spectrometry. Rapid Communications in MassSpectrometry 14, 585–589.

Huguet, C., Cartes, J., Damsté, J.S.S., Schouten, S., 2006a. Marine crenarchaeotalmembrane lipids in decapods: implications for the TEX86 paleothermometer.Geochemistry, Geophysics, Geosystems 7, Q11010.

Huguet, C., Hopmans, E.C., Febo-Ayala, W., Thompson, D.H., Damsté, J.S.S., Schouten,S., 2006b. An improved method to determine the absolute abundance ofglycerol dibiphytanyl glycerol tetraether lipids. Organic Geochemistry 37,1036–1041.

Huguet, C., Schimmelmann, A., Thunell, R., Lourens, L.J., Damste, J.S.S., Schouten, S.,2007. A study of the TEX86 paleothermometer in the water column andsediments of the Santa Barbara Basin, California. Paleoceanography 22.

Huguet, C., Kim, J.-H., de Lange, G.J., Damsté, J.S.S., Schouten, S., 2009. Effects of longterm oxic degradation on the UK0

37, TEX86 and BIT organic proxies. OrganicGeochemistry 40, 1188–1194.

Ingalls, A.E., 2006. From the cover: quantifying archaeal community autotrophy inthe mesopelagic ocean using natural radiocarbon. Proceedings of the NationalAcademy of Sciences of the United States of America 103, 6442–6447.

Ingalls, A.E., Huguet, C., 2012. Distribution of intact and core membrane lipids ofarchaea (GDGTs) among size fractionated particulate organic matter in theHood Canal. Applied and Environmental Microbiology 78, 1480–1490.

Iversen, M.H., Nowald, N., Ploug, H., Jackson, G.A., Fischer, G., 2010. High resolutionprofiles of vertical particulate organic matter export off Cape Blanc, Mauritania:degradation processes and ballasting effects. Deep-Sea Research Part I –Oceanographic Research Papers 57, 771–784.

Jenkyns, H.C., Schouten-Huibers, L., Schouten, S., Damsté, J.S.S., 2012. Warm MiddleJurassic–Early Cretaceous high-latitude sea-surface temperatures from theSouthern Ocean. Climate of the Past 8, 215–226.

Karakas, G., Nowald, N., Blaas, M., Marchesiello, P., Frickenhaus, S., Schlitzer, R.,2006. High-resolution modeling of sediment erosion and particle transportacross the northwest African shelf. Journal of Geophysical Research – Oceans111, C06025.

Karakas, G., Nowald, N., Schäfer-Neth, C., Iversen, M.H., Barkmann, W., Fischer, G.,Marchesiello, P., Schlitzer, R., 2009. Impact of particle aggregation on verticalfluxes of organic matter. Progress in Oceanography 83, 331–341.

Karner, M.B., DeLong, E.F., Karl, D.M., 2001. Archaeal dominance in the mesopelagiczone of the Pacific Ocean. Nature 409, 507–510.

Key, R.M., Kozyr, A., Sabine, C.L., Lee, K., Wanninkhof, R., Bullister, J., Feely, R.A.,Millero, F., Mordy, C., Peng, T.-H., 2004. A global ocean carbon climatology:results from GLODAP. Global Biogeochemical Cycles 18, GB4031.

A. Basse et al. / Organic Geochemistry 72 (2014) 1–13 13

Kim, J.-H., Schouten, S., Hopmans, E.C., Donner, B., et al., 2008. Global sediment core-top calibration of the TEX86 paleothermometer in the ocean. Geochimica etCosmochimica Acta 72, 1154–1173.

Kim, J.-H., Crosta, X., Michel, E., Schouten, S., Duprat, J., Damsté, J.S.S., 2009a. Impactof lateral transport on organic proxies in the Southern Ocean. QuaternaryResearch 71, 246–250.

Kim, J.-H., Huguet, C., Zonneveld, K.A.F., Versteegh, G.J.M., Roeder, W., Damsté, J.S.S.,Schouten, S., 2009b. An experimental field study to test the stability of lipidsused for the TEX86 and UK0

37 palaeothermometers. Geochimica et CosmochimicaActa 73, 2888–2898.

Kim, J.-H., van der Meer, J., Schouten, S., Helmke, P., Willmott, V., Sangiorgi, F., Koc,N., Hopmans, E.C., Damsté, J.S.S., 2010. New indices and calibrations derivedfrom the distribution of crenarchaeal isoprenoid tetraether lipids: implicationsfor past sea surface temperature reconstructions. Geochimica et CosmochimicaActa 74, 4639–4654.

Kuypers, M.M.M., Blokker, P.P., Erbacher, J.J., Kinkel, H.H., Pancost, R.D.R., Schouten,S.S., Damsté, J.S.S., 2001. Massive expansion of marine archaea during a mid-Cretaceous oceanic anoxic event. Science 293, 92–95.

Labrenz, M., Sintes, E., Toetzke, F., Zumsteg, A., Herndl, G.J., Seidler, M., Juergens, K.,2010. Relevance of a crenarchaeotal subcluster related to CandidatusNitrosopumilus maritimus to ammonia oxidation in the suboxic zone of thecentral Baltic Sea. ISME Journal 4, 1496–1508.

Lee, K.E., Kim, J.-H., Wilke, I., Helmke, P., Schouten, S., 2008. A study of the alkenone,TEX86, and planktonic foraminifera in the Benguela Upwelling System:implications for past sea surface temperature estimates. Geochemistry,Geophysics, Geosystems 9, Q10019.

Lengger, S.K., Hopmans, E.C., Damsté, J.S.S., Schouten, S., 2012a. Comparison ofextraction and work up techniques for analysis of core and intact polartetraether lipids from sedimentary environments. Organic Geochemistry 47,34–40.

Lengger, S.K., Hopmans, E.C., Reichart, G.-J., Nierop, K.G.J., Damsté, J.S.S., Schouten,S., 2012b. Intact polar and core glycerol dibiphytanyl glycerol tetraether lipidsin the Arabian Sea oxygen minimum zone. Part II: Selective preservation anddegradation in sediments and consequences for the TEX86. Geochimica etCosmochimica Acta, 1–15.

Lipp, J.S., Hinrichs, K.-U., 2009. Structural diversity and fate of intact polar lipids inmarine sediments. Geochimica et Cosmochimica Acta 73, 6816–6833.

Liu, M., Shuai, Y., Liang, H., Huang, L., Song, N., 2012. Comparative study onextraction methods of microbial lipids. Journal of Earth Science 21, 300–303.

Loescher, C.R., Kock, A., Koenneke, M., LaRoche, J., Bange, H.W., Schmitz, R.A., 2012.Production of oceanic nitrous oxide by ammonia-oxidizing archaea.Biogeosciences 9, 2419–2429.

Martin, J., Knauer, G., Karl, D., 1987. VERTEX: carbon cycling in the northeast Pacific.Deep Sea Research Part A 34, 267–285. http://dx.doi.org/10.1016/0198-0149(87)90086-0.

Mincer, T.J., Church, M.J., Taylor, L.T., Preston, C., Karl, D.M., DeLong, E.F., 2007.Quantitative distribution of presumptive archaeal and bacterial nitrifiers inMonterey Bay and the North Pacific Subtropical Gyre. EnvironmentalMicrobiology 9, 1162–1175.

Mollenhauer, G., Eglinton, T.I., Hopmans, E.C., Damsté, J.S.S., 2008. A radiocarbon-based assessment of the preservation characteristics of crenarchaeol andalkenones from continental margin sediments. Organic Geochemistry 39,1039–1045.

Müller, P., Fischer, G., 2001. A4-year sediment trap record of alkenones from thefilamentous upwelling region off Cape Blanc, NW Africa and a comparison withdistributions in underlying sediments. Deep Sea Research Part I: OceanographicResearch Papers 48, 1877–1903.

Müller, P., Schneider, R., Ruhland, G., 1994. Late quaternary PCO2 variations in theAngola current: evidence from organic carbon d13C and alkenone temperature.In: Zahn, R., Pedersen, T.F., Kaminski, M.A., Labeyrie, L. (Eds.), Carbon Cycling inthe Glacial Ocean: Constraints on the Ocean’s Role in Global Change, NATO ASISeries I. Springer-Verlag, Berlin, pp. 343–366.

Nowald, N., Karakas, G., Ratmeyer, V., Fischer, G., Schlitzer, R., Davenport, R.A.,Wefer, G., 2006. Distribution and transport processes of marine particulatematter off Cape Blanc (NW-Africa): results from vertical camera profiles. OceanScience Discussions 3, 903–938.

Ohkouchi, N., Eglinton, T.I., Keigwin, L.D., Hayes, J.M., 2002. Spatial and temporaloffsets between proxy records in a sediment drift. Science 298, 1224–1227.

Pitcher, A., Hopmans, E.C., Mosier, A.C., Park, S.J., Rhee, S.K., Francis, C.A., Schouten,S., Damsté, J.S.S., 2011a. Core and intact polar glycerol dibiphytanyl glyceroltetraether lipids of ammonia-oxidizing archaea enriched from marine andestuarine sediments. Applied and Environmental Microbiology 77, 3468–3477.

Pitcher, A., Villanueva, L., Hopmans, E.C., Schouten, S., Reichart, G.J., Damsté, J.S.S.,2011b. Niche segregation of ammonia-oxidizing archaea and anammox bacteriain the Arabian Sea oxygen minimum zone. ISME Journal 5, 1896–1904.

Pitcher, A., Wuchter, C., Siedenberg, K., Schouten, S., Damsté, J.S.S., 2011c.Crenarchaeol tracks winter blooms of ammonia-oxidizing Thaumarchaeota inthe coastal North Sea. Limnology and Oceanography 56, 2308–2318.

Ratmeyer, V., Fischer, G., Wefer, G., 1999. Lithogenic particle fluxes and grain sizedistributions in the deep ocean off northwest Africa: implications for seasonalchanges of aeolian dust input and downward transport. Deep Sea Research PartI: Oceanographic Research Papers 46, 1289–1337.

Reynolds, R.W., Rayner, N.A., Smith, T.M., Stokes, D.C., Wang, W., 2002. An improvedin situ and satellite SST analysis for climate. Journal of Climate 15, 1609–1625.

Sarmiento, J., 1993. Atmospheric carbon dioxide and the ocean. Nature 365, 119–125.Schouten, S., Hopmans, E.C., Schefuß, E., Damsté, J.S.S., 2002. Distributional

variations in marine crenarchaeotal membrane lipids: a new tool forreconstructing ancient sea water temperatures? Earth and Planetary ScienceLetters 204, 265–274.

Schouten, S., Hopmans, E.C., Damsté, J.S.S., 2004. The effect of maturity anddepositional redox conditions on archaeal tetraether lipid palaeothermometry.Organic Geochemistry 35, 567–571.

Schouten, S., Forster, A., Panoto, F.E., Damsté, J.S.S., 2007. Towards calibration of theTEX86 palaeothermometer for tropical sea surface temperatures in ancientgreenhouse worlds. Organic Geochemistry 38, 1537–1546.

Schouten, S., Middelburg, J.J., Hopmans, E.C., Damste, J.S.S., 2010. Fossilization anddegradation of intact polar lipids in deep subsurface sediments: A theoreticalapproach. Geochimica et Cosmochimica Acta 74, 3806–3814.

Schouten, S., Pitcher, A., Hopmans, E.C., Villanueva, L., van Bleijswijk, J., Damsté,J.S.S., 2012. Intact polar and core glycerol dibiphytanyl glycerol tetraether lipidsin the Arabian Sea oxygen minimum zone: I. Selective preservation anddegradation in the water column and consequences for the TEX86. Geochimicaet Cosmochimica Acta, 1–16.

Schouten, S., Hopmans, E.C., Damsté, J.S.S., 2013. The organic geochemistry ofglycerol dialkyl glycerol tetraether lipids: a review. Organic Geochemistry 54,19–61.

Schubotz, F., Wakeham, S.G., Lipp, J.S., 2009. Detection of microbial biomass byintact polar membrane lipid analysis in the water column and surfacesediments of the Black Sea. Environmental Microbiology 11, 2720–2734.

Seiter, K., Hensen, C., Schröter, J., Zabel, M., 2004. Organic carbon content in surfacesediments—defining regional provinces. Deep Sea Research Part I:Oceanographic Research Papers 51, 2001–2026.

Shah, S.R., Mollenhauer, G., Ohkouchi, N., Eglinton, T.I., Pearson, A., 2008. Origins ofarchaeal tetraether lipids in sediments: insights from radiocarbon analysis.Geochimica et Cosmochimica Acta 72, 4577–4594.

Sturt, H., Summons, R., Smith, K., Elvert, M., Hinrichs, K.-U., 2004. Intact polarmembrane lipids in prokaryotes and sediments deciphered by high-performance liquid chromatography/electrospray ionization multistage massspectrometry – new biomarkers for biogeochemistry and microbial ecology.Rapid Communications in Mass Spectrometry 18, 617–628.

Taylor, K.W.R., Huber, M., Hollis, C.J., Hernandez-Sanchez, M.T., Pancost, R.D., 2013.Re-evaluating modern and Palaeogene GDGT distributions: implications for SSTreconstructions. Global and Planetary Change 108, 158–174.

Teira, E., Reinthaler, T., Pernthaler, A., Pernthaler, J., Herndl, G.J., 2004. Combiningcatalyzed reporter deposition-fluorescence in situ hybridization andmicroautoradiography to detect substrate utilization by bacteria and archaeain the deep ocean. Applied and Environmental Microbiology 70, 4411–4414.

Turich, C., Freeman, K.H., Bruns, M.A., Conte, M., Jones, A.D., Wakeham, S.G., 2007.Lipids of marine archaea: patterns and provenance in the water-column andsediments. Geochimica et Cosmochimica Acta 71, 3272–3291.

Turich, C., Schouten, S., Thunell, R.C., Varela, R., Astor, Y., Wakeham, S.G., 2013.Comparison of TEX86 and UK0

37 temperature proxies in sinking particles in theCariaco Basin. Deep-Sea Research Part I 78, 115–133.

Wakeham, S.G., Lewis, C.M., Hopmans, E.C., Schouten, S., Damsté, J.S.S., 2003.Archaea mediate anaerobic oxidation of methane in deep euxinic waters of theBlack Sea. Geochimica et Cosmochimica Acta 67, 1359–1374.

Wakeham, S.G., Hopmans, E.C., Schouten, S., Damsté, J.S.S., 2004. Archaeal lipids andanaerobic oxidation of methane in euxinic water columns: a comparative studyof the Black Sea and Cariaco Basin. Chemical Geology 205, 427–442.

Wakeham, S.G., Turich, C., Schubotz, F., Podlaska, A., Li, X.N., Varela, R., Astor, Y.,Sáenz, J.P., Rush, D., Damsté, J.S.S., Summons, R.E., Scranton, M.I., Taylor, G.T.,Hinrichs, K.-U., 2012. Biomarkers, chemistry and microbiology showchemoautotrophy in a multilayer chemocline in the Cariaco Basin. Deep SeaResearch Part I: Oceanographic Research Papers 63, 133–156.

Weijers, J.W.H., Lim, K.L.H., Aquilina, A., Damsté, J.S.S., Pancost, R.D., 2011.Biogeochemical controls on glycerol dialkyl glycerol tetraether lipiddistributions in sediments characterized by diffusive methane flux.Geochemistry, Geophysics, Geosystems 12, Q10010.

Wuchter, C., Schouten, S., Wakeham, S.G., Damsté, J.S.S., 2005. Temporal and spatialvariation in tetraether membrane lipids of marine Crenarchaeota in particulateorganic matter: Implications for TEX86 paleothermometry. Paleoceanography 20,1–11.

Zhu, C., Lipp, J.S., Wörmer, L., Becker, K.W., Schröder, J., Hinrichs, K.-U., 2013.Comprehensive glycerol ether lipid fingerprints through a novel reverse-phaseliquid chromatography–mass spectrometry protocol. Organic Geochemistry 65,53–62.