Earth and Planetary Science Letters 362 (2013) 272–282

Contents lists available at SciVerse ScienceDirect

Earth and Planetary Science Letters

0012-82

http://d

n Corr

Tel.: þ3

E-m1 N

Den Bu

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

Impact of the Messinian Salinity Crisis on Black Sea hydrology—Insightsfrom hydrogen isotopes analysis on biomarkers

Iuliana Vasiliev a,b,n, Gert-Jan Reichart b,c,1, Wout Krijgsman b

a Organic Geochemistry, Department of Earth Sciences, Utrecht University, Budapestlaan 4, 3584 CD, Utrecht, The Netherlandsb Paleomagnetic Laboratory ‘Fort Hoofddijk’, Department of Earth Sciences, Utrecht University, Budapestlaan 17, 3584 CD, Utrecht, The Netherlandsc Alfred Wegener Institute for Polar and Marine Research, Biogeosciences, Am Handelshafen 12 (E), D-27570 Bremerhaven, Germany

a r t i c l e i n f o

Article history:

Received 24 May 2012

Received in revised form

15 November 2012

Accepted 20 November 2012

Editor: G. Hendersonformation. Here, we present compound specific hydrogen isotope (dD) data from Messinian Black Sea

Available online 3 January 2013

Keywords:

hydrogen isotopes

alkenones

n-alkanes

Black Sea

Messinian

1X/$ - see front matter & 2012 Elsevier B.V.

x.doi.org/10.1016/j.epsl.2012.11.038

esponding author.

1 302531361; fax: þ31 302531677.

ail addresses: i.vasiliev@uu.nl, iuli.iuliana@ya

ow at:Royal Netherlands Institute of Sea Re

rg, Texel, The Netherlands.

a b s t r a c t

The Messinian Salinity Crisis (5.96–5.33 Ma ago) was a dramatic oceanographic event, when evaporites

kilometers thick precipitated in a desiccating Mediterranean basin, trapping more than 5% of the

world’s oceanic salt. Hydrological changes in the adjacent Black Sea and water exchange with the

Mediterranean region are crucial, but poorly understood factors, influencing Messinian evaporite

sedimentary rocks that show a rapid change to heavy waters at 5.8 Ma, when major glaciations

occurred. At the same time, highly depleted dD values of long chain n-alkanes derived from plant waxes

indicate that fresh, river transported water originated from colder northern latitudes. The dD values of

alkenones, biosynthesized by haptophyte algae, show an unprecedented increase of 60% within

�100 kyr. The corresponding rapid change to þ110% for dD of the Black Sea waters seem unrealistic,

being heavier than anywhere in the present day oceans. Regardless of the applied relation between the

dD values of the alkenones and dD of the waters where they were produced, the 60% enrichment in the

dD values of alkenones indicates strongly enhanced evaporitic conditions. Still, the relative distribution

of the alkenones implies in-situ growth and reproduction of haptophyte algae, requiring sustained

marine conditions in the Black Sea up to 5.6 Ma. This indicates that Mediterranean–Black Sea

connectivity persisted during the first MSC phase when gypsum precipitated in the Mediterranean

basin. When the Black Sea became isolated, at the peak of the MSC (�5.6 Ma), it had a strongly negative

hydrological budget and rapidly desiccated due to excess evaporation.

& 2012 Elsevier B.V. All rights reserved.

1. Introduction

Intracontinental marine basins like the Mediterranean andBlack Seas react rapidly to changing environments. Variations inrivers runoff, evaporation and precipitation are reflected quickly inseawater salinity and circulation patterns. For the Mediterranean,changes in the basin hydrology during the latest Miocene weredramatic. The region was affected by the so-called MessinianSalinity Crisis (MSC; 5.96–5.33 Ma), when evaporites kilometersthick accumulated in the desiccating Mediterranean basin (Hsuet al., 1973; Krijgsman et al., 1999). During the latest phase of theMSC, the Mediterranean basin seemed to have been affected byinflow of (fresh?) water from the Paratethys, area currentlyoccupied by the circum Black and Caspian Seas regions (Citaet al., 1990, 1978). This resulted in a brackish to fresh water

All rights reserved.

hoo.com (I. Vasiliev).

search, P.O. Box 59, 1790AB

lake-sea (Lago Mare) system, extending over Mediterraneantowards the northeast, in the regions of Black Sea, Caspian Seaand Aral Lake (Fig. 1). The MSC ended with Pliocene re-floodingwhen marine sedimentation resumed (Hsu et al., 1973).

An intermittent Black Sea–Mediterranean connection can betraced back during the late Miocene to early Pliocene (�11–3 Ma).During those times the Black Sea was the central part of the EasternParatethys (Fig. 1), an epicontinental sea whose connections to theopen ocean became progressively restricted, resulting in the forma-tion of several subbasins with environments marked by salinitiesvarying from marine to brackish and fresh water conditions (e.g.Popov et al., 2006; Rogl, 1998). An impact of the Mediterranean’sMSC on its neighboring Paratethyan sub-basins has been speculated,although the precise mechanism involved remains unknown (Hsuand Giovanoli, 1979). Frequent sea level changes in the Black Seahave been described from the Messinian (7.24–5.33 Ma) (Gilletet al., 2007), when the Black Sea area was still the central part ofthe eastern Paratethys (Fig. 1) (Rogl, 1996). Cores from the Deep SeaDrilling Project Leg 42 B revealed the presence of supratidal andintertidal Messinian deposits at more than 1700 m below currentsea level (Hsu and Giovanoli, 1979). Also seismic data show deep

Fig. 1. Palaeogeographic map of the late Miocene, showing the Paratethys area on the presented day land configuration. Major rivers draining into the Paratethys are

indicated. The values of the present day precipitation dD are reported according to IAEA (IAEA, 2001). Long-term means were calculated by selecting yearly means in which

isotope content have been measured at least in 75% of the precipitation for that year and at least over eight months (IAEA, 2001). The star locates Taman peninsula (TM)

with Zheleznyi Rog section. Deep sea drilling project 42B sites 379, 380 and 381 are located. Eastern and Western Paratethys (largely overlapping to the Pannonian basin)

were subbasins exiting during Late Miocene.

Fig. 2. Total-ion current chromatogram (TIC) of the apolar fraction with increasing predominance of the n-alkanes towards the younger sedimentary rocks. The examples

are shown from old (a) to young (c), i.e. from upper Meotian to Kimmerian. Stratigraphic levels are in meters. C17–C34 refers to n-alkanes with odd (green) over even (blue)

predominance in chain length distribution; C37–C39 (purple) are alkenones; other important compounds also indicated. The star represents co-injected standard. (a) TM 07

sample with the predominance of the alkenones; (b) TM 09 sample with comparable amplitudes of the n-alkanes and alkenones; (c) TM 13 sample where the n-alkanes are

abundant in the a-polar fraction while the alkenones could not be detected. The stratigraphic level is indicated in all three panels. (For interpretation of the references to

color in this figure legend, the reader is referred to the web version of this article.)

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282 273

Messinian erosional surfaces (Gillet et al., 2007), in agreement witha desiccating Black Sea as a consequence of a negative hydrologicalbudget (Bartol and Govers, 2009). Still, evidence such as Late Miocene

salt deposits are lacking from the Black Sea area. In contrast,overspilling of Black Sea water as a consequence of a positive ratherthan a negative hydrological budget has been suggested as well, since

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282274

this would explain the widespread expansion of brackish environ-ments in the Mediterranean during the latest MSC (Lago Mare) phase(Cita et al., 1978). This controversy was impossible to resolve becauseof poor time control in the Black Sea area and also because ofdifficulties in interpreting local proxy data. The limited number ofcarbonate test building foraminifera species, which are moreoverhighly endemic, has prevented the construction of carbonate basedstable isotope and trace metal records.

Recently, we established a robust time frame for the MessinianBlack Sea successions of the Taman peninsula (Russia) (Krijgsmanet al., 2010; Vasiliev et al., 2011). Using this time frame we herepresent compounds specific hydrogen isotope data from excel-lently preserved n-alkanes and alkenones, derived from plantwaxes and haptophyte algae respectively (Fig. 2), to reconstructpast changes in the hydrology of the Paratethys. The long-chainn-alkanes are part of the leaf waxes, protective layers on top ofhigher plants leaves (Eglinton and Hamilton, 1967). These bimo-lecules retain the hydrogen isotopic composition (dD) of thewater used at the time of biosynthesis in the leaf. Several studieshave shown that dD ratios of n-alkanes reflect the dD compositionof precipitation (e.g. Sachse et al., 2004a,b, 2006) and are affectedby the evapotranspiration operating in the plant (e.g. Sachse et al.,2006). In this research the compound-specific dD data of long-chain n-alkanes reflect the dD values of the original environmen-tal water used by the higher plants and hence, serve as paleoprecipitation/aridity proxies for the circum-Black Sea region.Alkenones is the colloquial name of long strait chain ketonesand are some of the most widely biomarkers used. These arebiosynthesized by unicellular eukaryotic haptophyte marine algaeand are widespread in the photic zone of the modern ocean. Inthis study, the alkenones are used as recorders of the dDcomposition of the water they live in and use (i.e. Black Sea,Paratethys basin waters). Therefore, whereas the plant waxesprovide information on terrestrial precipitation and evaporation,the alkenones are used to assess basin hydrology. This data issubsequently used to investigate transport of water from theMediterranean to the Paratethys or vice versa. This providesinsight into impact and underlying mechanisms of the MessinianSalinity Crises in the larger circum Mediterranean area.

2. Material and methods

2.1. The Zheleznyi Rog section

The Zheleznyi Rog section is located on the Taman Peninsula(Russia; Fig. 1) and covers the upper Miocene to Pliocene,represented by Meotian, Pontian and Kimmerian regional EasternParatethys stages (Krijgsman et al., 2010; Popov et al., 2006;Vasiliev et al., 2004). The development of regional stratigraphicscheme was required because the progressive restriction of theEastern Paratethys from the global ocean and, as consequence, theexpansion of endemic Paratethyan type of biotas. The new agemodel for this section was earlier developed on the basis ofextensive paleomagnetic and radio-isotopic dating (Krijgsmanet al., 2010; Vasiliev et al., 2011).

The sampled part of the section starts in the upper Miocene(upper Meotian). The upper Meotian is represented by alterna-tions of grey-clays and whitish marls and diatomites and ends atthe levels with the last occurrence of Congeria novorossica.The Meotian–Pontian transition, dated at 6.04 Ma, is marked bya marine transgression identified by an incursion of planktonicforaminifera (Krijgsman et al., 2010). Like the Upper Meotian, thePontian consists of an alternation of grey-marls and whitishdiatomite layers. The Pontian–Kimmerian boundary is markedby a highly characteristic reddish layer, with silty concretions,

gypsum crystals and jarosite, followed by two meters of oolithicsandstone with iron oxides indicating sudden lowering of BlackSea water levels at �5.6 Ma (Krijgsman et al., 2010). Above thisreddish interval, a sharp transition to Pliocene marls is observed.The upper 90 m of section consists of dark-grey marls with inter-bedded limonite layers and alternations of grey-clays with meterthick coarser rocks (silts and sands) at the very top of the exposedsuccession.

In an attempt to reconstruct the large scale past changes in thehydrology of central part of the Eastern Paratethys (i.e. Black Searegion–Taman section) and assess the impact of the MSC weanalyzed 12 samples from stratigraphic levels that span the MSCperiod (from �6.2 to 5.5 Ma) for organic geochemistry analysisand hydrogen isotope measurements. Marls and clayey intervalswere sampled because these lithologies preserve organic matterbetter and hence have the highest potential for organic lipidbiomarkers study. To avoid weathered material, samples werecollected having removed the weathered part (�0.3 m) of theexposure.

2.2. Lipid extraction, separation and analyses

Between 10 and 65 g of sedimentary rock of each sample wasdried and then powdered. The total organic lipids of the largersamples, weighing between 20 and 60 g, were extracted withSoxleth extraction using a dichloromethane–methanol (7.5:1)mixture. From the smaller samples (weighing up to 20 g) theorganic compounds were extracted using Accelerated SolventExtraction (Dionex 200) equipment and a dichloromethane–methanol (9:1) mixture. The extracts were rotary-evaporated tonear dryness and subsequently further dried under a nitrogenflow. The total lipid extracts (TLE) were rinsed over anhydrousNa2SO4 column to remove any remaining molecular water. Ele-mental sulfur was removed by stirring over night with activatedcopper in dichloromethane. Copper flakes were activated with2 M HCl and afterwards rinsed with MilliQ ultra pure water,methanol and dichloromethane. The total lipid extract wasseparated using column chromatography with pre-combustedAl2O3 as stationary phase and a mixture of hexane and dichlor-omethane (9:1, v-v) to elute the a-polar fraction, a mixture ofhexane and dichloromethane (1:1, v-v) to elute the ketonefraction and, finally, a mixture of dichloromethane and methanol(1:1, v-v) was used to elute the polar fraction. Straight chainn-alkanes were isolated from the a-polar fraction using urea-adduction. The dry a-polar fraction was dissolved in 200 mlmethanol/urea (�10%, H2NCONH2, Merck) solution. Subse-quently, 200 ml acetone and 200 ml hexane were added to thesolution, frozen (�20 1C) and dried under N2 flow. The n-alkanecompounds were captured during the formation of the ureacrystals. These were subsequently washed with hexane to removethe non-adductable branched and cyclic compounds. Urea crys-tals, containing the adductable n-alkanes, were dissolved in500 ml methanol and 500 ml MilliQ ultra pure water mixture.The n-alkanes were subsequently extracted from the solutionusing hexane. The urea-adduction procedure was repeated two–three times to fully eliminate the non-adductable alkanes fromthe a-polar fraction. The n-alkanes and alkenones were identifiedusing mass spectra, molecular ion mass and retention timeusing a Gas Chromatography (GC) Mass Spectrometer (Thermo-Finnigan Trace DSQ). The fractions (dissolved in hexane) wereinjected on-column at 70 1C (CP-Sil 5CB fused silica column)(30 m�0.31 mm i.d, film thickness 0.1 mm). The oven programmewas set at constant pressure (100 kPa) and then programmed toincrease to 130 1C at 20 1C min�1, and then at 5 1C min�1 to320 1C at which it was held isothermal for 10 min. The individual

Fig. 3. Representative gas chromatograms of typical samples of Taman section. The presented example is TM 08 located at the uppermost Meotian sequence. (a) The

a-polar adductable fraction after second urea adduction step; note the n-alkanes with a distinctly odd (green) over even (blue) predominance in chain length distribution;

(b) the alkenones fraction from the same rock sample as the one shown in panel a; (c) detailed view of the alkenones in panel b. (For interpretation of the references to

color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4. The calculated carbon preference index (CPI) compared to the average

chain length (ACL). On the left side the regional Paratethys time scales is

presented. The age model is according to Krijgsman et al. (2010) and the ages

are in bold italic. The foraminifer symbol indicates the level of marine transgres-

sion into Paratethys. The different shadings in the lithological columns represent

schematically the alteration of marls (less protruding) and siltstones (more

indurate). The two yellowish layers at ��170 and �180 represent diatomites.

Note the significant change in the CPI occurring at the Meotian/Pontian transition

and less significant shift in the ACL. The change in vegetation composition is

marked. (For interpretation of the references to color in this figure legend, the

reader is referred to the web version of this article.)

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282 275

n-alkanes and alkenones were quantified using a GC equippedwith a flame ionization detector (FID).

2.3. Compound specific hydrogen isotope analyses

Compound-specific hydrogen isotopes of n-alkanes and alke-nones were determined by gas chromatography-isotope ratiomass spectrometry (GS-IRMS). The dD composition of individualn-alkanes and alkenones was measured on a-polar adductableand alkenones fractions on a HP 6890N Gas Chromatograph (GC)coupled to a Thermo-Finningan Delta Plus XP Isotope Ratio MassSpectrometer (IRMS) at the Organic Geochemistry laboratory ofthe Earth Sciences department, Utrecht University. The fractions(dissolved in hexane) were injected on-column at 70 1C, the ovenbeing programmed to increase to 130 1C at 20 1C min�1, and thenat 5 1C min�1 to 320 1C at which it was held isothermal for10 min.

GC conditions were similar to conditions for GC analysisexcept that the film thickness of the CP-Sil 5 column was0.4 mm and that a constant flow of He was used at 1.5 ml min�1.Compounds were pyrolyzed in an empty ceramic tube heated at1450 1C which was pre-activated by a 5 min methane flow of0.5 ml min�1. H3þ factors were determined daily on the isotopemass spectrometer and were at any time less than 5. Each extractwas measured between two and 14 times. The large number ofmultiple analyses was related to the unusual results, whichneeded verification. H2 gas with known isotopic compositionwas used as reference and a mixture of C16–C32 n-alkanes withknown isotopic composition (ranging from �42% to �256% vs.Vienna Standard Mean Ocean Water (V-SMOW)) was used tomonitor the performance of the system (Schimmelman Mixture Aand B, Biogeochemical Laboratories, Indiana University). A squa-lane standard was co-injected with every sample and its averagevalue was �17173%, which compared favorably with its offlinedetermined value of �168.9%. For plotting the results data, foreach n-alkane were averaged to obtain the mean value for thatn-alkane. Error bars plotted are based on the standard deviationof the full set of multiple analyses.

2.4. ACL and CPI

Each sample contains long chain (C25–C35) n-alkanes with adistinct odd over even predominance in chain length. Averagechain length (ACL) and the degree of oddity (CPI) are expressed asfollows:

ACL¼(25A25þ27A27þ29A29þ31A31þ33A33)/(A25þA27þA29þA31þA33)

CPI¼(((A25þA27þA29þA31þA33)/(A24þA26þA28þA30þA32))þ((A25þA27þA29þA31þA33)/(A26þA28þA30þA32þA34)))�0.5 whereA represents the area under the chromatogram peak for individualn-alkanes.

3. Results

3.1. Molecular biomarkers

Initially total lipid extracts were separated into polar anda-polar fractions only. The a-polar fractions show a series ofn-alkanes and ketones and a minor contribution from hopenes(Fig. 2). Since the shorter ketones partly co-elute with then-alkanes a separate ketones fraction was made. The ketonesfraction shows in addition to a series of shorter ketones high

Table 1

dD isotopes measured on alkenones from Taman. The dD of the source waters (dDwater) were calculated using the relations of Englebrecht and Sachs (2005) for C37 and C38

separately. Average, standard deviation (STDEV) and standard error of the means (SEM) are listed.

Sample Level Alkenone Amplitude dD Average dD STDEV SEM dDw Englebrecht and Sachs (2005)

(m) (mV) (%) (%) C37 C38

TM 12 �93.8 C37 1042 �170.6 �160.2 9.6 4.81 88.5

C38 777 �157.3 97.7

801 �164.6

2746 �148.3

TK 21 �104.6 C38 1076 �148.2 �149.1 1.3 - 112.6

1586 �150.0

TK 28 �109.7 C38 578 �158.7 �158.4 1.8 0.89 100.1

1656 �159.6

1656 �159.6

2194 �155.8

TM 11 �116.2 C37 2794 �146.4 �153.9 6.8 3.95 97.2

C38 865 �155.4 106.2

871 �159.8

TM 10 �133.6 C37 725 �191.0 �188.3 3.3 1.23 50.1

1162 �185.6

1534 �194.3

C38 496 �185.3 60.0

1046 �188.0

1681 �186.9

2528 �187.1

TR 121 �143.0 C37 1498 �205.3 �207.7 3.6 1.46 23.7

1526 �205.2

1958 �213.5

C38 1625 �205.9 34.0

1640 �205.4

2210 �210.7

TM 09 �184.9 C37 571 �227.0 �223.5 6.0 2.99 2.0

615 �230.1

C38 702 �218.8 12.7

757 �218.1

TM 08 �197.1 C37 1206 �228.2 �223.8 8.2 2.36 1.7

1301 �227.4

1533 �216.0

1562 �216.1

3072 �231.7

C38 1643 �225.9 12.4

1775 �222.2

1778 �226.6

1848 �213.3

1876 �210.6

2890 �229.8

4557 �237.3

TM 07 �226.6 C37 731 �222.5 �216.8 4.0 1.06 11.3

737 �220.8

1815 �218.4

707 �213.4

725 �212.6

477 �212.0

472 �209.3

C38 529 �213.7 21.8

2012 �216.4

739 �216.5

512 �218.7

741 �218.9

744 �220.2

733 �221.0

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282276

concentrations of long chain ketones (Fig. 2). These long chainketones (C37–C39) are unsaturated ethyl and methyl ketones(alkenones) (Fig. 3). The relative distribution of C37:2 and C37:3

alkenones, expressed in the UK037 index (Prahl and Wakeham,

1987), indicates SSTs between 19.5 and 23.8 1C.The a-polar fractions show a series of n-alkanes ranging from

n-C21 to n-C35, with the long-chain (C27–C29) n-alkanes having thehighest concentration (Figs. 2 and 3a). The long-chain n-alkanes

also show a strong odd over even predominance (Figs. 2a and 3c).A suite of C27 to C35 hopanes (Fig. 2), dominated by 22R-17a(H),21b(H)-Homohopane, is also present, with Hop-17(21)-ene being identified in the samples most frequently.

There is a correlation in the trends of the calculated ACL andCPI for the Taman samples (Fig. 4). The ACL has the highest valueof 29.5 at �197.1 m level coinciding with the maximum valuesfor the CPI of 7. Above, towards the younger part of the section

Fig. 5. dD isotopes measured on n-alkane and long chain alkenones from Taman and the time equivalent climate records. On the left side Mediterranean and Paratethys

time scales are presented next to oxygen isotope record. TG indicates glacial–interglacial marine stages. Events of the Messinian Salinity Crisis and regional Paratethys

stage names are listed. See also captions to Fig. 4. In the right side are simplified stratigraphic column and geomagnetic polarity reversals (Krijgsman et al., 2010). Ages in

million years are in green italics. dD values for n-C27, n-C29 and n-C31 n-alkanes are presented. The dD precipitation values are calculated for two scenarios (1) in the case of

the European ‘humid’ temperate climate using the relation of Sachse et al. (2006) and (2) in the case of ‘arid’ climate using the relation of Feakins and Sessions (2010).

Different scales are indicated for dry and humid climate effect on evapotranspiration. The level where the switch in hydrological balance takes place is marked on the

figure and represents the level where different evapotranspiration effect should be used. dD values for C37, C38 and C39 alkenones and the calculated dD of the source

waters using Englebrecht and Sachs (2005) calibration are plotted. The error bars indicate standard errors of the mean. (For interpretation of the references to color in this

figure legend, the reader is referred to the web version of this article.)

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282 277

the calculated ACL for the samples ranges between 28.5 and 29,while the CPI is between 5 and 3.5 (Fig. 4).

3.2. Compound specific hydrogen isotope analysis

3.2.1. Alkenones hydrogen isotope ratios

The stable hydrogen isotopic composition of the C37 and C38

alkenones (dDalkenone) is between �220% and �150% (%VSMOW), showing a sharp increase in values between �140and �120 m stratigraphic level (Table 1 and Fig. 5). Values of theC37 and C38 alkenones closely correspond, showing the sametrend through time. In this record, up to �5.85 Ma, values remainrather constant around �220%. Subsequently, within 100 kyrthere is a 60% shift to heavier values of �150%. Although onlyone point falls within this shift (�188%), the values measured at�133.6 m stratigraphic level indicates a steady, gradual transi-tion towards heavier values. Values remain high towards the topuntil in the Kimmerian alkenones are no longer present.The lightest values recorded (�220%) coincide, albeit with aconsiderable stratigraphic range, with the Meotian–Pontian tran-sition (at 6.04 Ma).

3.2.2. n-Alkane hydrogen isotope ratios

The dD of the C27, C29 and C31 n-alkanes co-vary, rangingbetween �213.5% and �167% (Table 2 and Fig. 5). At the loweststratigraphic levels n-alkane concentrations were too low to allow

dD analyses. The first sample with concentrations high enough islocated before the Meotian–Pontian transition, showing values ofabout �197%. At the Meotian–Pontian transition (at 6.04 Ma andat �184.9 m in our section) the dDn-alkanes increase, showing theheaviest values of about �167%. Subsequently values decreaseagain, showing a steady trend towards values of �220% at the topof the record. In general, dD of C27 n-alkane shows somewhatheavier stable isotopic compared to the C29 and C31 n-alkanes.Overall the spread in stable isotopic values observed in then-alkanes (C27, C29 and C31) seems slightly higher than thatbetween the different alkenones (C37 and C38). This is expectedsince the (long chain) n-alkanes are produced by a variety ofhigher plants (e.g. grasses, trees) and the signal trapped in thesedimentary succession record the normal vegetation variation.The lower variation in the stable isotopic composition of C37 andC38 alkenones is also expected because they are produced by aspecific group of organisms (haptophyte algae).

4. Discussion

The long chain (C37, C38 and C39) unsaturated ethyl and methylketones (alkenones) are some of the best known and widely usedbiomarkers in paleoceanography and they are produced byprymnesiophyte algae (Marlowe et al., 1984; Volkman et al.,1980). They are frequently used for sea surface temperaturereconstructions (Prahl and Wakeham, 1987) and for estimating

Table 2

dD isotopes measured on n-alkane from Taman. The dDprecipitation is calculated for a ‘wet’ climate using Sachse et al. (2006) and for a ‘dry’ climate using Feakins

and Sessions (2010). See also the caption of Table 1.

Sample Level n-alkane Amplitude dD AveragedD STDEV SEM dDprecipitation

(m) (mV) (%) (%) Sachse et al. (2006) Feakins and Sessions (2010)

TM 13 �88.8 C27 518 �211.5 �213.5 3.4 1.7 �95.9 �131.9

C29 638 �217.4

C31 544 �211.5

TM 12 �93.8 C27 541 �206.4 �208.1 4.2 3.0 �89.8 �126.0

C27 2253 �208.0

C27 1966 �212.2

C29 602 �200.1

C29 2435 �214.2

C29 2123 �211.7

C31 552 �205.9

C31 2615 �207.5

C31 2246 �207.2

TK 21 �104.6 C27 1228 �208.4 �202.8 4.7 2.4 �83.7 �120.1

C27 1245 �199.8

C29 1454 �206.0

C29 1483 �206.4

C31 1520 �198.2

C31 1458 �197.9

TK 28 �109.7 C27 2325 �212.6 �210.1 5.5 2.6 �92.1 �128.2

C27 1378 �201.3

C29 2963 �213.9

C29 1537 �211.5

C29 426 �214.0

C31 3160 �214.3

C31 1391 �203.4

TM 11 �116.2 C27 675 �194.6 �200.6 9.6 3.5 �81.2 �117.7

C27 2131 �203.2

C27 2149 �189.4

C27 521 �187.3

C29 771 �201.2

C29 2905 �219.4

C29 2916 �193.9

C29 687 �203.7

C31 701 �196.5

C31 3616 �216.0

C31 3682 �199.5

C31 602 �202.8

TM 10 �133.6 C27 805 �189.9 �199.5 10.3 2.4 �79.9 �116.5

C27 951 �189.8

C29 1115 �213.7

C29 1349 �210.5

C31 1082 �195.0

C31 1310 �198.2

TR 121 �143.0 C27 745 �168.9 �183.5 11.4 2.4 �61.5 �98.8

C27 651 �171.0

C29 920 �193.9

C29 850 �195.8

C31 735 �183.3

C31 696 �188.1

TM 09 �184.9 C27 846 �157.8 �166.9 7.9 2.6 �42.4 �80.4

C27 414 �155.8

C29 1366 �170.6

C29 646 �171.1

C31 1376 �176.2

C31 629 �173.3

C31 596 �163.2

TM 08 �197.1 C29 671 �201.3 �197.0 7.5 2.0 �77.0 �113.7

C29 533 �185.9

C31 812 �201.0

C31 622 �200.0

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282278

paleo-pCO2 via their d13C isotopic composition (Pagani et al.,1999). In the Taman section relative abundances of the C37, C38

and C39 alkenones (Fig. 3b and c) closely resemble those in recentBlack Sea sediments (de Leeuw et al., 1980; Freeman and

Wakeham, 1992). Whereas in most open marine settings C39 ispresent only in relative low concentrations, both the recent BlackSea and the Taman section shows appreciable concentrations ofC39 alkenones. The C39:2 in the Taman section (Fig. 3b and c) is

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282 279

somewhat higher than observed in marine settings, and moresimilar to what has been found in high alkalinity lakes (Thielet al., 1997). However, whereas high alkalinity lakes show adominant C37:4 alkenone, this compound is absent from theTaman section. The absence of C37:4 alkenones corroborated withthe dominance of the C37:2, C37:3, C38:2 and C38:3 (Fig. 3b and c) isin line with an open marine depositional setting while theappreciable amount of C39:2 in Taman samples (may) indicate amodern Black Sea-like connection to the marine realm.

4.1. Alkenones and their hydrogen isotope ratios

dDalkenones has been shown to increase linearly with thedeuterium stable isotopic ratio of the water (dDwater) in whichthe algae live (Englebrecht and Sachs, 2005; Paul, 2002). Paul(2002) found that the fractionation between the dD of varyingculture medium water and C37 alkenones produced by theEmiliania huxleyi was relatively constant at �232%. Englebrechtand Sachs (2005) proposed a similar fractionation of �225%.Recently, Schwab and Sachs (2011), observed no change in D/Hfractionation (aalkenone–water) along a salinity gradient in theChesapeake Bay, suggesting that individual alkenone D/H ratiosare primarily determined by source water dD values and that theymay therefore be used as proxy to reconstruct dDwater. However,the values of the alkenones–water D/H fractionation reported bySchwab and Sachs (2011) are up to 60% lower than thosereported in the studies of Englebrecht and Sachs (2005) suggest-ing that coastal species may fractionate less than oceanic hapto-phyte algae (Schwab and Sachs, 2011). Comparing values fromthe sedimentary record and the water column Schwab and Sachs(2011) suggested a fractionation between 170% and 190%, forthese coastal settings. Still, since the deuterium isotopic values ofestuarine waters are relatively depleted, values measured on thealkenones are depleted as well (�220% to �170%).

Several physiological factors have been shown to affect hap-tophyte deuterium isotopic fractionation. Variations in dDalkenone

values can also be related to changes in growth rates (malk). A 60%shift in dDalkenone would, however, require unrealistic changes inmalk. The maximum difference observed related to changes ingrowth rate is 13% only, between the early and late phases ofexponential growth (Wolhowe et al., 2009). This implies thatchanges in growth rate alone cannot explain the observeddifference. Moreover, since the record is generated from a timespan of 6.25 to 5 Myr, with an average accumulation rate of22 cm/kyr, each sample represents �4 kyr. The dDalkenone valuesare thus averaging these 4 kyr. Changes in dDalkenone due tovariations in the specific growth rates during different phases ofhaptophyte blooms would, therefore, average out. The impact ofchanges in growth rate will, therefore, probably be considerablysmaller even than the maximum 13% observed in an experi-mental setting. Glacial–interglacial variation in the dD due tobuildup of continental ice sheets could account at most 8% fromthe observed shift in the dDwater of the Black Sea since the oceanwater in an ice-free world is 8% lighter (Schmidt et al., 1999).

Recent work on two marine haptophyte, Emiliania huxleyi

and Gephyrocapsa oceanica, cultured at different salinities andtemperatures suggests that hydrogen isotope fractionation bythese algae depends on the dD of the water, salinity (Swater) andgrowth rate (malk) (Schouten et al., 2005) and possibly irradiance(Pagani, 2002) but not on temperature. Thus, the dD value of C37

alkenones of E. huxleyi mainly depends on Swater and the dD ofwater, which itself is again strongly correlated to Swater (Schoutenet al., 2005). Based on these findings van der Meer et al. (2007)concluded that a 25% depletion in dD measured on the C37

alkenones corresponded to a 6% reduction in the salinity ofMediterranean water during the last interglacial’s sapropel

formation. In the adjacent Black Sea basin a similar fresheningof the sea water over the past 3000 yr has been inferred from�20% dD C37 alkenones depletion (van der Meer et al., 2008). Ifwe apply one of these relationships: dDalkenone¼4.8� Swater�347(for G. oceanica); or dDalkenone¼4.2� Swater�354 (for E. huxleyi) tothe values measured on the Taman alkenones the calculated Swater

increases from 31% to 48.8% (G. oceanica ) or from 25.7% to41.2% (E. huxleyi). Calculated salinities are within a reasonablerange for E. huxleyi and G. oceanica to survive and reproducesuggesting sustained marine conditions. However, as in the caseof the highly used SST calculations using UK370 index (Prahl andWakeham, 1987), there are limitations when applying salinitycalculations based on dDalkenone because the calibration wasperformed on species that did not exist in the MessinianBlack Sea.

In the present day ocean measured dDalkenone range from�181% in the warm Sargasso Sea (at 311N) to �200% intemperate Gulf of Maine and Emerald Basin (at 431N)(Englebrecht and Sachs, 2005). For the present day Black Sea thevalues of dDalkenone are approximately �230% (van der Meeret al., 2008), values much lighter than the rest of the ocean at thesame latitude because of the influence of large fresh water input.The high values of �150% for the dDalkenone observed in theMiocene Taman deposits of the Black Sea have not been reportedpreviously from any marine natural environment. Thesevalues indicate that extraordinary conditions affected the Messi-nian Black Sea. Applying the proposed relations (Englebrechtand Sachs, 2005) between the dDalkenone and dDwater

(dDalkenone¼0.732� dDwater�225 for C37 alkenones anddDalkenone¼0.745�dDwater�233 for C38 alkenones) we calculatethat dDwater in the Black Sea increased from þ20% to þ110%,within approximately 100 kyr, between �143 and �116.2 m(Figs. 5 and 6b). Such enriched waters are nowadays typical forextreme environments such as high salinity desert lakes. Part ofthe offset observed could, however, also be caused by changes infractionation. When using the latest proposed relations betweenthe dDalkenone and dDwater (Schwab and Sachs, 2011) the change indDwater would be 110%, from �70% to þ40% (applying thecalibration for C37:3) or 145%, from �115% to þ30% (using thecalibration for C37:2). When using the relation between the C38:3

dDalkenone the dDwater of the basin would vary with 115% (from�60% to þ55%), while using the relation between the C38:2

dDalkenone would be 131%, from �95% to þ35%. All existingcalibrations (Englebrecht and Sachs, 2005; Paul, 2002; Schoutenet al., 2005; Schwab and Sachs, 2011) clearly indicate a largeincrease in dDwater, reaching extreme positive values. Regardlessof the applied relation, our dDalkenone results clearly indicate a fastswitch of the Black Sea hydrological budget from positive tonegative, with extreme evaporation prevailing.

4.2. n-Alkane hydrogen isotope ratios

In the same samples equally high concentration of long chainn-alkanes with a clear odd over even predominance were found,indicating a higher plant wax origin (Eglinton and Hamilton,1967). The high-molecular weight, straight-chain hydrocarbons(C27–C33 n-alkanes) are frequently used biomarkers for dD assess-ment of the terrestrial environment. Recent studies observed that,despite the complex factors determining the fractionation of thehydrogen isotopes in plants, dDprecipitation values are the funda-mental control on the plant-wax dD composition (Polissar andFreeman, 2010; Sachse et al., 2004b; Sachse et al., 2006; Smithand Freeman, 2006). A change in plant community could poten-tially affect the biosynthetic fractionation and would be reflectedin the average chain length of the n-alkanes (Feakins and Sessions,2010; Pedentchouk et al., 2007; Polissar and Freeman, 2010;

Fig. 6. Schematic representations of important changes in connectivity and

impact on the dD of the precipitation and basins waters. (a) The flooding event

in Paratethys at 6.04 Ma (Krijgsman et al., 2010). (b) The upper evaporites time

when the Mediterranean and Paratethys were (recurrently) connected to the

ocean. (c) The time of total disconnection of Paratethys and Mediterranean from

the open ocean.

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282280

Sachse et al., 2004b; Sachse et al., 2006; Smith and Freeman,2006). Despite plant-physiology-induced limitations to the quan-titative interpretation of the dD lipid based records, the dDn-alkanes

have been successfully used in reconstruction paleo precipitation(Andersen et al., 2001; Pagani et al., 2006; Sachse et al., 2004a;Schefuss et al., 2005; Speelman et al., 2010; Tipple and Pagani,2010).

The dD values from the Messinian in Taman of the C27, C29 andC31 n-alkanes (dDn-alkane) ranged from �214% to �167% andindicate lighter precipitation towards the younger part of thesection (Fig. 5 and Table 2). The heaviest value however coincideswith the Meotian–Pontian transition, at 6.04 Ma, when an influxof marine waters was identified (Krijgsman et al., 2010).To calculate dDprecipitation we applied a constant biosyntheticfractionation between source water and n-alkane of 157%(Sachse et al., 2006; Sessions et al., 1999) and an additionalevapotranspiration effect. We are probing two extreme valuesfor the evapotranspiration enrichment effect: (1) of �30%

(Sachse et al., 2006) found in the modern Western Europe humidconditions and (2) of �60% as described for arid ecosystem(Feakins and Sessions, 2010). The latter is compatible with theinterpretation of the stable deuterium isotope data from thealkenones which indicates strong evaporation. Assuming a Wes-tern Europe like humid environment dDprecipitation values wouldhave varied between �96% and �42% (Fig. 5 and Table 2).Currently, at stations north of the Black Sea values around �62%are reported (IAEA, 2001), suggesting that at least for the middlepart of the record, a comparable, mainly Atlantic, source andisotopic pathway for local precipitation in Messinian times ispossible (Fig. 1). In the case of prevailing arid conditions thecalculated dDprecipitation would have varied between �132% and�80%, typical values exclusively observed in cold northern highlatitudes in Eurasia (IAEA, 2001). Since the dDalkenones indicate ashift from relatively humid to dry conditions a different evapo-transpiration effect could be applied to the lower and upper partof the record. This would imply a change from �42% in thehumid (older) period to �132% in the driest (younger) period.The much lighter precipitation towards the younger part of therecord can be explained by three different mechanisms, or acombination thereof. First, a change in the dominant water vaporsource could offset overall values. Second, distance to the vaporsource might have increased. Third, continental temperaturesmight have decreased. Which of these factors has the most impactdepends on the timing of the observed changes. All threeproposed mechanisms would strongly influence the compositionof the higher-plant vegetation which is the major producer of thelong-chain n-alkanes and indirectly, alter the recorded dDn-alkanes

values. The calculated ACL (28.5–29.5) and CPI (3.5–7) values aretypical for n-alkanes producers by terrestrial higher plants(Fig. 4). However, the switch towards lower ACL and CPI (Fig. 4)appears at Meotian–Pontian transition (6.04 Ma) which wouldindicate a change in the n-alkanes producers’ plant assemblages,still of terrestrial origin but different so that it could affect thebiosynthetic fractionation

4.3. Timing of events

The heaviest inferred dDprecipitation values and the lightestBlack Sea dDwater coincide with a marine incursion from theMediterranean into the Eastern Paratethys, at the Meotian–Pontian transition (6.04 Ma) (Figs. 5 and 6a). The heavydDprecipitation most probably indicate reduced distance to thevapor source for the rain, i.e. this site was closest to the ocean(Fig. 5), interpretation supported also by the lightest Black SeadDwater. At this specific time slice the Black Sea had a positivewater balance (Fig. 6a) and a connection to the Mediterranean assupported by the foraminifers’ influx (Krijgsman et al., 2010).Earlier modeled output suggests that the rivers supplying the BlackSea and Caspian today fed more run-off into Paratethys in the LateMiocene (Gladstone et al., 2007). The timing (�5.83–5.7 Ma) of theswitch from a positive to a negative hydrological budget in the BlackSea area coincides conspicuously with glaciations TG 20 and TG 22(Fig. 5), suggesting a causal link to the expansion of glacial ice caps.Because the change in dDwater due to the formation of continentalice sheets (78%) is clearly much smaller than the observed shift,this implies that a causal link must be indirect, being one of thecontributing factors in the observed dDwater shift. For the next�100 kyr (between �5.7 and 5.6 Ma), Black Sea dDwater valuesremained stable, albeit clearly enriched (Figs. 5 and 6b). Intensifiedevaporation at relative low humidity, possibly enhanced by a dryglacial atmosphere, would progressively enrich the remaining BlackSea seawater in deuterium (Fig. 6c). Still, the magnitude of theenrichment suggests that the water sources for the Black Sea musthave been isotopically enriched as well.

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282 281

The Mediterranean at that time experienced its first MSCphase (Krijgsman et al., 1999) with precipitation of gypsum inshallow water environments under evaporative conditions(Roveri et al., 2008) and associated elevated dDwater values asfound in the crystal water of laminar gypsum (Bellanca et al.,1986). dD analyses on individual biomarkers from MSC succes-sions showed dDwater of the Mediterranean increased up to þ66%(Andersen et al., 2001) during deposition of the Lower Gypsum(5.96–5.6 Ma) (Fig. 6b). The moisture evaporating from theMediterranean Sea surface transported eastward on westerliesmay have rained out over the Black Sea region supplying heavy,fresh water. The alkenone pattern from our samples stronglysuggest that the Black Sea remained marine type of basin even attimes of persistent negative water budget (Fig. 6b) implying that aconnection must have persisted to the salty Mediterranean or tothe open ocean. The inflow of deuterium enriched seawater and/or deuterium enriched atmospheric moisture originating from theMSC Mediterranean are both likely contributors to the increasingdDwater of the Black Sea. The relative contribution of freshwaterfrom rivers must have been limited as the isotopic signature ofthis water was, au contraire, very deuterium depleted (Fig. 6b).The source for the light river water at that time was the lightprecipitation, as indicated by and calculated from our dDn-alkanes.

The connection between the Mediterranean and the Atlanticclosed at �5.6 Ma (Krijgsman et al., 1999), causing a dramatic sealevel lowering in the Mediterranean and thus a disruption of theBlack Sea–Mediterranean connection as well (Fig. 6a). Indeed, theTaman section shows an unconformity with a sharp transition ofmarine Messinian marls to a coastal unit of reddish iron rich sandsindicating sudden lowering of Black Sea water levels at �5.6 Ma(Krijgsman et al., 2010). Above this reddish interval, a sharp transitionto Pliocene marls is observed with strontium isotope ratios similar tothe values recorded for the Black Sea in glacial times, when it was alake, indicating that no connection to the Mediterranean existed(Major et al., 2006). We did not find alkenones in any of thesedimentary rock higher than �90 m (Fig. 5), which suggests thatthe Pliocene transgression in the Black Sea was caused by a change toa positive hydrological budget resulting from increased river inflow.This is in agreement with a marked climate change in the latestMessinian–early Pliocene, where stable isotope records indicate muchhigher temperatures (Hodell et al., 2001).

5. Summary and implications

The combined hydrogen isotopic composition of the excel-lently preserved alkenones and n-alkanes from our section reflectimportant changes in both dDwater and dDprecipitation of the BlackSea area during the Messinian Salinity Crisis. Regardless of theapplied relation between the dDalkenones and dDwater they wereproduced in, the 60% enrichment in the dDalkenones indicatestrongly enhanced evaporitic conditions. The calculated 110%increase for the dD of the Black Sea waters corresponds to achange that seems unrealistic, being heavier than anywhere inthe present day marine ecosystem. The dDn-alkanes data fromTaman peninsula reflect changes towards much lighter precipita-tion for the younger parts of the section. This trend can beexplained by three main mechanisms: (1) a precipitation sourcediverted to the cooler, northern latitudes; (2) increased distanceto the vapor source occurred and (3) temperature over Eurasia hasdropped significantly. For the Meotian–Pontian transition at6.04 Ma, the dDalkenones and dDn-alkanes data point more to proxi-mity of the sampled site to the marine realm with precipitationfalling soon after evaporation from the sea surface (Figs. 5 and6a). Between 5.8 and 5.6 Ma, the isotopic contrast between thedDwater and dDprecipitation is the highest from the entire studied

interval. The calculated dDprecipitation values for 5.8–5.6 Ma timeinterval are typical for cold climate or deep into continent zoneswhile the inferred Black Sea dDwater suggest extreme evaporationresulting in negative hydrological budget.

During the MSC the Mediterranean experiences its majordesiccation phase and a physical link to Black Sea is likely.However, the lack of equivalent hydrogen isotope data from theMediterranean realm hampers the understanding of hydrologicalexchange between Paratethys subbasins and Mediterraneanduring the MSC. Future work will focus on the integration andadjustments of the used methods to the specific Mediterranean–Paratethys system with the final goal of understanding the largeenvironmental changes of the Messinian Salinity Crisis. The datawill be used for modeling experiments that need the boundaryconditions to be set by paleo proxy acquisitions (Lunt et al., 2008).

Our results from Taman peninsula indicate a rapid switch of theBlack Sea hydrological budget from positive to negative (Fig. 5).Overall, our isotopic data show that the Black Sea had a negativehydrological budget when it became isolated from the Mediterranean(Fig. 6c) during the peak MSC event (�5.6 Ma). We thus concludethat the water level in this Messinian ‘‘Black Lake’’ could indeed havedropped substantially as suggested by seismic profiles (Gillet et al.,2007). However, the amplitude of changes in the Black Sea waterlevel is still a matter of discussion. The seismic profiles (Gillet et al.,2007) plead for hundreds of meters of sea level drop while thepaleontological content of the land-based Taman section indicates atmaximum tens of meters of level drop. Lowering of the Black Seawater level would cause isostatic uplift in the gateway areas (Bartoland Govers, 2009), and additional tectonic rearrangement in theCarpathians and Caucasus could close important sea ways, e.g.leading to the isolation of the Caspian Sea. The Caspian Basin indeedshowed a marked transition at �5.6 Ma, where the North CaspianBasin desiccated and the Volga Delta progressed southward into theSouth Caspian Basin forming the hydrocarbon reservoirs of theProductive Series. The Messinian water level lowering in the CaspianSea was estimated to have been �1000 m (Dumont, 1998), indicatingthat paleo-Volga’s debit was reduced, switching the Caspian Seawater budget to a negative one.

The reconfiguration of the Paratethys into separate basins playedan important role in the Pliocene and Pleistocene hydrologicalchanges of the Black Sea region. The Black Sea has been, and still is,positioned at the latitude where the overall hydrological budgetchanges from evaporation controlled to precipitation and river run-offcontrolled, and such a reconfiguration might have had a dispropor-tionally large impact on the regional hydrological balance. Futurechanges in the hydrological cycle related to anthropogenic inducedclimate change may strongly and rapidly impact the Black Seaenvironment and its unique ecosystem.

Acknowledgments

We are thankful to Marius Stoica, Cor Langereis, Viktor Popov,Alexandr Iosifidi, and Ekaterina Grundan for their help in the fieldand fruitful discussions. I.V. thanks Sietske Batenburg for help inthe geochemical laboratory. Rachel Flecker and Bill Ryan arethanked for their constructive comments that improved themanuscript. This work was financially supported by the Nether-lands Life and Earth Sciences Foundation (ALW) with supportfrom the Netherlands Organization for Scientific Research (NWO).

References

Andersen, N., et al., 2001. Large and rapid climate variability during the Messiniansalinity crisis: evidence from deuterium concentrations of individual biomar-kers. Geology 29, 799–802.

I. Vasiliev et al. / Earth and Planetary Science Letters 362 (2013) 272–282282

Bartol, J., Govers, R., 2009. Flexure due to the Messinian–Pontian sea level drop inthe Black Sea. Geochem. Geophys. Geosyst. 10 (10).

Bellanca, A., Calderone, S., Neri, R., 1986. Isotope geochemistry, petrology anddepositional environment of the diatomite-dominated Tripoli Formation(lower Messinian), Sicily. Sedimentology 33, 729–743.

Cita, M.B., Santambrogio, S., Melillo, B., Rogate, F., 1990. Messinian paleoenviron-ments: new evidence from the Tyrrhenian Sea (ODP Leg 107). Proc. ODP Sci.Results 107, 211–227.

Cita, M.B., Wright, R.C., Ryan, W.B.F., Longinelli, A., 1978. Messinian Paleoenviron-ments. Initial Reports of the Deep Sea Drilling Project, pp. 1003–1035.

de Leeuw, J.W., v.d. Meer, F.W., Rijpstra, W.I.C. Schenck, P.A., 1980. On theOccurrence and Structural Identification of Long Chain Unsaturated Ketonesand Hydrocarbons in Sediments. Physic. Chem. Earth, 12,211-12,217.

Dumont, H.J., 1998. The Caspian Lake: history, biota, structure, and function.Limnol. Oceanogr. 43 (1), 44–52.

Eglinton, G., Hamilton, R.J., 1967. Leaf epicular waxes. Science 156, 1322–1335.Englebrecht, A.C., Sachs, J.P., 2005. Determination of sediment provenance at drift

sites using hydrogen isotopes and unsaturation ratios in alkenones. Geochim.Cosmochim. Acta 69 (17), 4253–4265.

Feakins, S.J., Sessions, A.L., 2010. Controls on the D/H ratios of plant leaf waxes inan arid ecosystem. Geochim. Cosmochim. Acta 74, 2128–2141.

Freeman, K.H., Wakeham, S.G., 1992. Variation in the distribution and isotopiccomposition of alkenones in Black Sea particles and sediments. Org. Geochem.19, 277–285.

Gillet, H., Lericolais, G., Rehault, J.-P., 2007. Messinian event in the black sea:evidence of a Messinian erosional surface. Mar. Geol. 244, 142–165.

Gladstone, R., Flecker, R., Valdes, P.J., Lunt, D., Markwick, P., 2007. The Mediterra-nean hydrologic budget from a Late Miocene global climate simulation.Palaeogeogr. Palaeoclimatol. Palaeoecol. 251, 254–267.

Hodell, D.A., Curtis, J.H., Sierro, F.J., Raymo, M.E., 2001. Correlation of late Mioceneto early Pliocene sequences between the Mediterranean and North Atlantic.Paleoceanography 16, 164.

Hsu, K.J., Giovanoli, F., 1979. Messinian event in the Black Sea. Palaeogeogr.Palaeoclimatol. Palaeoecol. 29, 75–93.

Hsu, K.J., Ryan, W.B.F., Cita, M.B., 1973. Late Miocene desiccation of the Mediter-ranean. Nature 242, 240–244.

IAEA, 2001. GNIP Maps and Animations. International Atomic Energy Agency,Vienna.

Krijgsman, W., Hilgen, F.J., Raffi, I., Sierro, F.J., Wilson, D.S., 1999. Chronology,causes and progression of the Messinian salinity crisis. Nature 400, 652–655.

Krijgsman, W., Stoica, M., Vasiliev, I., Popov, V.V., 2010. Rise and fall of theParatethys Sea during the Messinian Salinity Crisis. Earth Planet. Sci. Lett. 290,183–191.

Lunt, D.J., Flecker, R., Valdes, P.J., Salzmann, U., Gladstone, R., Haywood, A.M., 2008.A methodology for targeting palaeo proxy data acquisition: A case study forthe terrestrial late Miocene. Earth Planet. Sci. Lett. 271, 53–62.

Major, C.O., et al., 2006. The co-evolution of Black Sea level and compositionthrough the last deglaciation and its paleoclimatic significance. Quat. Sci. Rev.25, 2031–2047.

Marlowe, I.T., et al., 1984. Long-chain (n-C37–39) alkenones in the Prymnesiophy-ceae. Distribution of alkenones and other lipids and their taxonomic signifi-cance. Br. Phycol. J. 19, 203–216.

Pagani, M., 2002. The alkenone–CO2 proxy and ancient atmospheric carbondioxide. Philos. Trans. R. Soc. London A 360, 609–632.

Pagani, M., Freeman, K.H., Arthur, M.A., 1999. Late Miocene atmospheric CO2

concentrations and the expansion of C4 grasses. Science 285, 876–879.Pagani, M., et al., 2006. Arctic hydrology during global warming at the Palaeocene/

Eocene thermal maximum. Nature 442, 671–675.Paul, H.A., 2002. Application of Novel Stable Isotope Methods to Reconstruct

Peleoenvironments: Compound-Specific Hydrogen Isotopes and Pore-WaterOxygen Isotopes. Ph.D. Thesis. Swiss Federal Institute of Technology, Zurich,141 pp.

Pedentchouk, N., et al., 2007. Comparison of d13C and dD values of n-alkanes fromangiosperms and gymnosperms in Western Europe. Geochim. Cosmochim.Acta 71, A770.

Polissar, P.J., Freeman, K.H., 2010. Effects of aridity and vegetation on plant-waxdD in modern lake sediments. Geochim. Cosmochim. Acta 74, 5785–5797.

Popov, S.V., et al., 2006. Late Miocene to Pliocene palaeogeography of theParatethys and its relation to the Mediterranean. Palaeogeogr. Palaeoclimatol.Palaeoecol. 238, 91–106.

Prahl, F.G., Wakeham, S.G., 1987. Calibration of unsaturation patterns in long-chain ketone compositions for paleotemperature assessment. Nature 330,367–369.

Rogl, F., 1996. Stratigraphic correlation of the Paratethys Oligocene and Miocene.Mitt. Ges. Geol. Bergbaustud. 41, 65–73.

Rogl, F., 1998. Paleogeographic Consideration for the Mediterranean andParatethys Seaways (Oligocene to Miocene). Ann. Naturhist. Mus. Wien 99A,279–310.

Roveri, M., Lugli, S., Manzi, V., Schreiber, B.C., 2008. The Messinian Sicilianstratigraphy revisited: new insights for the Messinian salinity crisis. TerraNova 20, 483–488.

Sachse, D., et al., 2004a. Reconstruction of palaeohydrological conditions in alagoon during the 2nd Zechstein cycle through simultaneous use of dD valuesof individual n-alkanes and d18O and d13C values of carbonates. Int. J. EarthSci. (Geol. Rundsch.) 93, 554–564.

Sachse, D., Radke, J., Gleixner, G., 2004b. Hydrogen isotope ratio of recentlacustrine sedimenatry n-alkanes record modern climate variability. Geochim.Cosmochim. Acta 68 (23), 4877–4889.

Sachse, D., Radke, J., Gleixner, G., 2006. dD values of individual n-alkanes fromterrestrial plants along a climatic gradient—implications for the sedimentarybiomarker record. Org. Geochem. 37, 469–483.

Schefuss, E., Schouten, S., Schneider, R.R., 2005. Climatic controls on central Africanhydrology during the past 20,000 years. Nature 437, 1003–1006.

Schmidt, G.A., Bigg, G.R., Rohling, E.J.,1999. Global sea water oxygen18 data base./http://www.data.giss.nasa.gov/o18data/1999S.

Schouten, S., et al., 2005. The effect of temperature and salinity on the stablehydrogen isotopic composition of long chain alkenones produced byEmiliania huxleyi and Gephyrocapsa oceanica. Biogeosci. Discuss. 2, 1681–1695.

Schwab, V., Sachs, J.P., 2011. Hydrogen isotopes in individual alkenones from theChesapeake Bay estuary. Geochim. Cosmochim. Acta 75, 7552–7565.

Sessions, A.L., Burgoyne, T.W., Schimmelmann, A., Hayes, J.M., 1999. Fractionationof hydrogen isotopes in lipid biosynthesis. Org. Geochem. 30, 1193–1200.

Smith, F.A., Freeman, K.H., 2006. Influence of physiology and climate on dD of leafwax n-alkanes from C3 and C4 grasses. Geochim. Cosmochim. Acta 70,1172–1187.

Speelman, E.N., et al., 2010. Modeling the influence of a reduced equator-to-polesea surface temperature gradient on the distribution of water isotopes in theEarly/Middle Eocene. Earth Planet. Sci. Lett. 298, 57–65.

Thiel, V., Jenisch, A., Landmann, G., Reimer, A., Michaelis, W., 1997. Unusualdistributions of long-chain alkenones and tetrahymenol from the highlyalkaline Lake Van, Turkey. Geochim. Cosmochim. Acta 61, 2053–2064.

Tipple, B.J., Pagani, M., 2010. A 35 Myr North American leaf-wax compound-specific carbon and hydrogen isotope record: implications for C4 grasslandsand hydrologic cycle dynamics. Earth Planet. Sci. Lett. 299, 250–262.

van der Meer, M.T.J., et al., 2007. Hydrogen isotopic composition of long-chainalkenones record freshwater flooding of the Eastern Mediterranean at theonset of sapropel deposition. Earth Planet. Sci. Lett. 262, 594–600.

van der Meer, M.T.J., et al., 2008. Molecular isotopic and dinoflagellate evidence forLate Holocene freshening of the Black Sea. Earth Planet. Sci. Lett. 267,426–434.

Vasiliev, I., et al., 2011. Magnetostratigraphy and radio-isotope dating of upperMiocene–lower Pliocene sedimentary successions of the Black Sea Basin(Taman Peninsula, Russia). Palaeogeogr. Palaeoclimatol. Palaeoecol. 310(3–4), 163–175.

Vasiliev, I., et al., 2004. Towards an astrochronological framework for theeastern Paratethys Mio-Pliocene sedimentary sequences of the Focsani basin(Romania). Earth Planet. Sci. Lett. 227, 231–247.

Volkman, J.K., Eglinton, G., Corner, E.D.S., Forsberg, T.E.V., 1980. Long chain alkenesand alkenones in the marine coccolithophorid Emiliania huxleyi. Phytochem-istry 19, 2619–2622.

Wolhowe, M.D., Prahl, F.G., Probert, I., Maldonado, M., 2009. Growth phasedependent hydrogen isotopic fractionation in alkenone-producing hapto-phytes. Biogeosci 6, 1681–1694.