Science of the Total Environment 672 (2019) 657–668

Contents lists available at ScienceDirect

Science of the Total Environment

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

Discriminating sources and preservation of organic matter in surface

sediments from five Antarctic lakes in the Fildes Peninsula (King George
Island) by lipid biomarkers and compound-specific isotopic analysis
Daniel Carrizo a,⁎, Laura Sánchez-García a, Rodrigo Javier Menes b, Felipe García-Rodríguez c,d
a
Centro de Astrobiología (CSIC-INTA), Madrid, Spain
b
Laboratorio de Ecología Microbiana Medioambiental, Microbiología, Facultad de Química y Facultad de Ciencias, Universidad de la República, Uruguay
c
Centro Universitário Regional Este, CURE-Rocha, Universidad de la República, Ruta 9 y Ruta 15 s/n, Uruguay
d
Programa de Pós-Graduação em Oceanografia Física, Química e Geológica, Instituto de Oceanografia, Universidade Federal do Rio Grande, Av. Itália, km 8, Cx.P. 474, 96201-900 Rio Grande, RS, Brazil

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Geochemical study on surface sedi-

ments from five Antarctic lakes, Fildes
Peninsula
• Terrestrial lipid biomarkers were de-
tected in all lakes.
• Aquatic-source indicators were rela-
tively abundant in mid-size lakes near
the glacier.
• Biogenic carbon was dominant in deep
lakes whereas petrogenic carbon in
shallow lakes.
• Distance to Collins Glacier, proximity to
coast & lake depth are determinant fac-
tors in biogeochemical signature.

a r t i c l e i n f o a b s t r a c t

Article history: Lakes are important paleoenvironmental archives retaining abundant information due to their typical high sed-
Received 8 February 2019 imentation rates and susceptibility to environmental changes. Here, we scrutinize the organic matter (OM) com-
Received in revised form 28 March 2019 position, origin and preservation state in surface sediments from five lakes in a remote, warming-sensitive, and
Accepted 29 March 2019
poorly explored region partially covered by the retreating Collins Glacier in King George Island (Antarctica), the
Available online 1 April 2019
Fildes Peninsula. Lipid biomarkers of terrestrial origin (i.e. high-molecular weight n-alkanes, n-alkanoic acids, and
Editor: Damia Barcelo n-alkanols; β-sitosterol, campesterol, and stigmasterol) were detected in the five Fildes Lakes, with the smallest
basin (i.e., Meltwater) showing a particularly strong moss imprint. Aquatic source indicators such as low C/N and
Keywords: terrestrial over aquatic ratios (TAR), or less negative δ13C values were preferentially found in the mid-sized lakes
Lipids biomarkers (i.e., Drake and Ionospheric). Sedimentary carbon in the larger lakes (i.e., Uruguay and Kitezh) displayed a largely
Lakes biogenic origin (i.e., values of carbon preference index, CPI, ≫1), whereas the three lakes close to Collins Glacier
Organic matter (i.e., Drake, Meltwater, and Ionospheric) showed certain contribution from petrogenic sources (CPI ~ 1). The
Maritime Antarctica results suggest that the geochemical signature of the surface sediments in the five Fildes lakes is determined
Isotopes
by factors such as the distance to the retreating Collins Glacier, the proximity to the coast, or the lake depth.
This study illustrates the forensic interest of combining lipid biomarkers, compound-specific isotopic analysis,
and bulk geochemistry to reconstruct paleoenvironments and study climate-sensitive regions.
© 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

⁎ Corresponding author.
E-mail address: [email protected] (D. Carrizo).

https://doi.org/10.1016/j.scitotenv.2019.03.459
0048-9697/© 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
658 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668

1. Introduction lakes is important for understanding the geophysical events and envi-
ronmental processes taking place in this vulnerable ecosystem affected
Lakes are important paleoenvironmental archives of short-term pro- by thermal destabilization. The few geochemical studies in this region
cesses at local to regional scale. They are systems susceptible to environ- are restricted to investigations focused on rocks (Smellie et al., 1984;
mental changes (e.g. climate change, land uses, glacier retreatment, Machado et al., 2005) soils (Mendoça et al., 2013; Michel et al., 2014;
volcanic activity, biota, or human impact) that affect OM delivery and Boy et al., 2016) or marine sediments (Monien et al., 2011), With the
burial, thus imprinting the sedimentary record. Lakes receive abundant only existing on lacustrine sediments describing mineralogical and
land-derived material that enhances lake fertility and aquatic produc- chemical aspects (Alfonso et al., 2015).
tion of OM, both producing high sedimentation rates that allow for The present study addresses the biogeochemical characterization
retaining information from the water basement and catchment area. of recent sediments from five glacial lakes in the Fildes Peninsula by
Lacustrine sediments constitute an integrative record of all local biogeo- combining bulk geochemistry and mineralogy with the use of lipid
chemical and geophysical processes (Cohen, 2003) with a variety of biomarkers and (bulk and compound specific) stable isotopic analysis.
indicators or proxies that can be used to reconstruct past climate and The aim is to provide a geochemical background to reconstruct major
environmental conditions. sources, environmental conditions, and processes affecting the supply
A substantial arsenal of biomarker compounds, together with and burial of OM in the Fildes Lakes. This study contributes to understand
patterns of compound classes and the isotopic compositions of bulk biogeochemical processes affecting the delivery and burial of OM in la-
OM or specific molecules enable to deduce much about the past ecosys- custrine sediments from the Fildes Peninsula, setting the basis for under-
tems and environments in which the sedimentary OM was created and standing future (biological, ecological, and geochemical) alterations in
deposited (e.g., Freeman et al., 1994; Huang et al., 1996; Lamb et al., this highly vulnerable system upon the current climate warming.
2006; Arbi et al., 2018). Lipid biomarkers have been widely used to
reconstruct paleoenvironments and changes in lacustrine systems 2. Materials and methods
(e.g. Arts et al., 2009; Vogts et al., 2009; Castañeda and Schouten,
2011; Ouyang et al., 2015), where different indexes have been 2.1. Study area and sampling setting
developed as indicators of source and alteration of sedimentary OM
(Cranwell, 1973; Simoneit et al., 1991; Meyers, 2003; Pancost and The Fildes Peninsula is a 7 km long peninsula situated in the south-
Boot, 2004). The OM signature is a mixture of compounds from different western end of King George Island (62°08′–62°14′S and 59°02′–
sources with a variable degree of preservation, susceptible to change 58°51′W), in the South Shetland Archipelago, Antarctica (Fig. 1a–b).
upon diagenesis and biological alteration during sinking. Therefore, With an approximate surface of 1400 km2 (Peter et al., 2008), it repre-
biomarkers interpretation has to be always done with caution, where sents one of the largest ice-free areas in the maritime Antarctic region.
the combination of multiple geochemical proxies helps to overcome The peninsula is bordered to the south by the Fildes Strait, to the
these effects and thereby optimize paleoenvironmental and paleoeco- south-east by the Maxwell Bay (a.k.a. Fildes Bay), and to the west by
logical reconstructions. the Drake Pass (Fig. 1c). On its link on the northeast to the rest of King
Antarctica is one of the most susceptible areas to climate and George Island, the peninsula is covered by the Collins Glacier. Geologi-
environmental changes due to the vast extensions of permanent ice, cally, the Fildes Peninsula is a tableland made up of old coastal land-
yet the most remote and unexplored continent on Earth. In particular, forms, with numerous rocky outcrops and an average height of 30 m
Maritime Antarctica is severely affected by climate change with ice above sea level (Antarctic Treaty Secretariat, 2009). Its bedrock is pri-
melting and glaciers retreating in an accelerated way upon the current marily basaltic, with volcanic intrusions and pyroclastic sediments
global warming. Over the last 40–50 years, the western and northern (Xianshen and Xiaohan, 1990).
parts of the Antarctica Peninsula has experienced some of the most The Fildes Peninsula is the most extensively snow-free coastal area in
rapid air temperature increases (~2 °C; Quayle et al., 2002 Thomas and summer on the island, most of which is permanently covered by ice. The
Tetzner, 2018), causing the thawing of large ice volumes and the peninsula ice-free zones have been occupied by glaciers until recent
retreatment of many glaciers. As a consequence, the system is becoming times, mostly since the last glacial maximum (87,000–15,000 yr BP),
more humid, especially in the maritime Antarctica, where small streams when gradual and rapid retreatment events alternated, such as the Fildes
and lakes fed by glacial and snow meltwater arise increasingly during deglaciation between 6,000 and 4,000 yr BP (Barsh and Mausbacher,
austral summer. Most of these lakes have limited outflow and are sup- 1986), or the progressive retreatment since 1500 yr BP (Serrano-
plied by surficial stream inflow as well as groundwater (Malandrino Cañadas, 2003). Since the late Holocene, the surviving ice cover (i.e.
et al., 2009), carrying debris and particulate material from the surround- Collins Glacier) remains stable, with an estimated ice retraction of only
ings. Subaqueous sediments in Antarctic glacial lakes may be considered a few tens of meters (Hall et al., 2010). The stability of the glacier is largely
as integrators of the short-term effects (decades to centuries) of the re- explained by its geographical location (inland) and dome-shaped struc-
cent thermal destabilization. ture, both protecting from marine wave action and climate fluctuations.
Because of the great vulnerability of Antarctica, special protection Nevertheless, recent data suggest a continuous glacier retreatment during
figures were created by the Antarctic Treaty System in 1964 (i.e. Antarc- the last 30 years, due to the increase of mean annual air temperatures and
tic Specially Protected Areas or ASPA). One of these ASPA sites (#125) is the resulting liquid rain (Vieira et al., 2015).
the Fildes Peninsula, a foreland in the south-western end of King George Climate in the Fildes Peninsula is typically cold, moist maritime, with
Island, in the South Shetland Islands of Antarctica. One of the eight ASPA a mean annual temperature of −2.1 °C (Michel et al., 2012). In summer,
125 sites is Collins Glacier (a.k.a. Bellingshausen), a glacial dome defin- mean daily temperature reach up to 2 to 3 °C, whereas in winter long-
ing the ice-cover limit at the southern-most tip of the Fildes Peninsula, term monthly mean temperature can drop below −7 °C (Braun et al.,
south of which ice-free lands extend. Collins Glacier is retreating upon 2004). The northern region of the peninsula is directly influenced by
the current rise of the atmospheric temperature (Lee et al., 2009; the Collins Glacier and contains abundant proglacial lakes (i.e. those
Monien et al., 2011), causing the increase of surficial runoff and the formed by ice melting behind an ice dam or a moraine left by a
ground waterlogging as the ice melts. Erosive effects and chemical retreating glacier; Embleton-Hamann, 2004), as well as recent glacial
weathering impact the ground, and the filling of land holes with melting deposits. In contrast, the central and south region has been longer
water result into glacial lakes. These lakes are the main sinks for water exposed, thus, less affected by glacial processes. Lakes in that region
loading solutes and particulate material from the whole catchment are rather periglacial (i.e. associated to frost action and/or permafrost
area, constituting integrative pictures of the local biological, geochemi- related processes, when the natural drainage of the topography is
cal, and physical processes. The geochemical characterization of these obstructed by an ice sheet, ice cap or glacier; French, 2007), with
 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668 659

Fig. 1. Map of Antarctica (a), showing the South Shetland Archipelago (b), and a close-up of the Fildes Peninsula (c), where the five lakes were studied. Drake, Meltwater, Ionospheric, and
Uruguay are classified as proglacial lakes, whereas Kitezh as periglacial (Vieira et al., 2015). Red stars in c indicate the location of the Antarctic scientific bases from Uruguay (1), Chile (2),
Russia (3), and China (4). The Fildes Peninsula topographical map in c (1:25,000) was obtained from the Australian Antarctic Data Center (https://data.aad.gov.au/aadc/mapcat; Map Nr.
13799). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

influence of disturbing agents such as wind, snow melting, rain, or water turnover. In the five lakes, surface sediments (upper 3 cm
marine action. depth) were collected from the deepest section of the lakes by a scuba
Lichens and mosses are the dominant components of the terrestrial diver using a stainless steel tube (30 cm depth and 66 mm i.d) as a
vegetation in the Fildes Peninsula as in the rest of Antarctica, with dragging tool. The surface sediment samples (~30 g) were transferred
poorly developed tundra-like vegetation covering the foreland. One of to organics-free polyethylene containers with the aid of a solvent-
the greatest diversity of mosses in Antarctica is on King George Island clean (dichloromethane/methanol, 3:1) scoop, and then rapidly closed.
(i.e. 64 species), where the Fildes Peninsula is particularly rich due to Additionally, fresh samples representative of the local vegetation were
the abundance of streams, pools and lakes providing the humid condi- collected from the lakes surroundings. One sample of moss (Sanionia
tions needed by this flora (Li et al., 2009). Deschampsia Antarctica and uncinata), particularly abundant in the northern part of the Fildes
Colobanthus quitensis are two species of higher plants described in Peninsula, was collected from the Drake and Meltwater surroundings;
Maritime Antarctica, where the Antarctic hairgrass D. Antarctica is ex-
tensively propagated over large areas including the Fildes Peninsula Table 1
(Torres-Mellado et al., 2011), whereas Colobanthus quitensis has been Geographical and geochemical features of the five Fildes lakes.
recently recorded on soils from the Byers Peninsula, Livingston Island,
Drake Meltwater Ionospheric Uruguay Kitezh
on South Shetland Islands (Vera et al., 2013).
Latitude 62°10′ 62°10′19″ 62°10′48″ 62°11′ 62°11′36″
16″ 06″
2.2. Sample collection Longitude 58°55′ 58°55′40″ 58°54′39″ 58°54′ 58°57′58″
37″ 40″
Surface sediments were sampled from different lakes on the Fildes Altitude (m.a.s.l.)a 41 42 20 3 1
Peninsula during a geochemical sampling campaign in February– Lake area (m2) 2958 400 16,235 69,781 95,668
Lake maximum 2 1 10 15 15
March 2018, under the support of the Uruguayan IAU (Instituto
depth (m)
Antártico Uruguayo) at the Artigas Base. Five lakes of different typology Distance to 1523 1569 568 147 473
were sampled from the vicinity of the Collins Glacier at different shoreline (m)
locations; a few meters apart (90–210 m) from the glacier margin Distance to Collins 92 120 215 624 3381
(Drake, Meltwater and Ionospheric lakes), or about 0.6 km (Uruguay) Glacier ice
margin (m)
or 3.4 km (Kitezh) apart to the southwest (Table 1, Fig. 1c). The location
TOC (% dw)b 0.62 0.38 0.27 4.0 2.1
of the lakes is shown in Fig. 1, and topographical and hydrographical TN (% dw)c 0.06 0.03 0.03 0.26 0.17
features are summarized in Table 1. The Ionospheric, Kitezh, and C/Nd 10 12 10 15 12
(mostly) Uruguay lakes are located closer to the Maxwell Bay coastline δ13C (‰) −20.7 −19.5 −23.4 −24.2 −23.4
δ15N (‰) −2.0 −0.2 −1.2 −1.4 1.3
relative to the other two lakes. Drake and Meltwater are the smallest
Lake typee Proglacial Proglacial Proglacial Proglacial Periglacial
(b3000 m2) and shallowest lakes (1–2 m depth), whereas Ionospheric,
a
Uruguay, and Kitezh have larger (N16,200 m2) and deeper (10–15 m Meters above sea level.
b
Total organic carbon, percentage relative to total dry weight.
depth) watersheds (Table 1). Unlike all others, the Ionospheric Lake c
Total nitrogen, percentage relative to total dry weight.
has a continuous inlet of water from the leaking Collins Glacier, and d
Ratio of TOC over TN (dimensionless).
an outlet that discharges in the Maxwell Bay, both ensuring a regular e
Classification based on Vieira et al. (2015).
660 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668

two types of lichens (Lecanora spp., and Placopsis contortuplicata) from electron ionization at 70 eV and scanning from m/z 50 to 650. The
the region between the Ionospheric and Uruguay lakes; and one sample analytes were injected (1 μl) and separated on a HP-5MS column
of grass (Deschampsia Antarctica) from the Kitezh surroundings. All sed- (30 m × 0.25 mm i.d. × 0.25 μm film thickness) using He as a carrier
iment and vegetal samples were stored cold (~4 °C) on solvent-clean gas at 1.1 ml min−1. For the non-polar fraction, the oven temperature
polypropylene containers until back at the laboratory, when they were was programmed from 50 °C to 130 °C at 20 °C min−1, then to 300 °C
frozen (−20 °C). They were then freeze-dried before geochemical at 6 °C min−1 (held 20 min). For the acidic fraction, the oven tempera-
analysis. ture was programmed from 70 °C to 130 °C at 20 °C min−1 and to 300 °C
at 10 °C min−1 (held 10 min). For the polar fraction, the oven tempera-
2.3. Mineralogy and geochemistry analysis ture program was the same as for the acidic fraction, but the oven was
held for 15 min at 300 °C. The injector temperature was 290 °C, the
The five surface sediments were analyzed for mineral composition transfer line was at 300 °C and the MS source at 240 °C. Compound iden-
with a Bruker X-Ray diffractometer (AXS D8-Focus, XRD). The samples tification was based on the comparison of mass spectra (NIST library)
were scanned after drying and grinding in the 2·ϴ-diffraction angle and/or the use of reference compounds, which were quantified using
from 5° to 70°, with a scanning step size of 0.01°, at 40 kV and 40 mA external calibration curves. External standards of n-alkanes (C10 to
with a Cu X-ray source (Cu Kα1,2, λ = 1.54056 Å). C40), FAME (C8 to C24), n-alcohols (C10, C14, C16, C18, C20) and branched
Inorganic anions were determined by ion chromatography in the isoprenoids (pristane and phytane) were injected to obtain calibration
soluble phase of the sediment samples. This analysis is fully described curves. Sigma Aldrich supplied all pure chemicals and standards. The
elsewhere (Parro et al., 2011; Sánchez-García et al., 2018). Briefly, analytical series included two procedural blanks (solvents), which
aliquots of 1 g of sample were sonicated (3 × 1 min cycles), diluted in were processed and analyzed to verify method recovery and absence
10 ml of deionized water, and filtered through a 22 μm GFF filter. The su- of undesired contamination. Limits of quantification (LOQ) for n-
pernatants were collected and loaded into a Metrohm 861 Advanced alkanes, n-alkanoic acids and n-alkanols were determined using the
compact ion chromatograph (Metrohm AG, Herisau, Switzerland). The lowest point in the calibration curve; 0.3 ng·g−1 dw for n-alkanes,
column Metrosep A supp 7-250 was used with 3.6 mM sodium carbon- and 0.5 ng·g−1 dw for both n-alkanoic acids and n-alkanols. Recoveries
ate (NaCO3) as eluent. Total major and minor elements were measured of the internal standards averaged 75 ± 11%.
with Ion Coupled Plasma-Mass Spectrometry (ICP-MS). Microwave
digestion using Teflon vessels were used according to previously 2.5. Compound specific isotope analysis
described methods (Nna-Mvondo et al., 2008).
Stable isotope ratios of total organic carbon (δ13C) and total nitrogen The compound-specific isotopic composition of carbon was mea-
15
(δ N) were measured using isotope-ratio mass spectrometry (IRMS), sured on certain lipid families (n-alkanes and n-alkanoic acids as
and following the analytical methods by the USGS (Révész et al., FAME). The compound-specific isotopic analysis (CSIA) was performed
2012). Briefly, the sediment samples were homogenized using mortar coupling the gas chromatograph (Trace GC 1310 ultra) to the isotope-
and pestle. HCl (3 N) was added to the samples to remove carbonates, ratio mass spectrometry system (MAT 253 IRMS, Thermo Fisher Scien-
and, after equilibration for 24 h, adjusted to neutral pH with ultrapure tific). The conditions for the GC analysis were identical to those set for
water. The sample was oven dried (50 °C) until constant weight. Then, the molecular analysis of the polar fraction, whereas those for the
the δ13C and δ15N ratios were measured in a MAT 253 IRMS (Thermo IRMS analysis were as follows: electron ionization 100 eV, Faraday cup
Fisher Scientific) and reported in the standard per mil notation. Three collectors m/z 44, 45 and 46, and temperature of the CuO/NiO combus-
certified standards (USGS41, IAEA-600, and USGS40) were used to tion interface of 1000 °C. The samples were injected in splitless mode,
report an analytical precision of 0.1‰. The content of total nitrogen with an inlet temperature of 250 °C, and helium as a carrier gas at con-
(TN %) and total organic carbon (TOC %) was measured with an elemen- stant flow of 1.1 ml min−1. Isotopic values of the individual lipids
tal analyzer (HT Flash, Thermo Fisher Scientific) during stable isotope separated by GC were calculated using CO2-spikes of known isotopic
measurements. composition, introduced directly into the MS source, three times at
the beginning and end of every run. Reference mixtures (Indiana Uni-
2.4. Lipid extraction, fractionation, analysis and QA/QC versity, USA) of known isotopic composition of n-alkanes (A6) and
FAME (F8) were run every four samples to check accuracy of the isoto-
Lyophilized subsamples of the lacustrine sediments (5–10 g) and pic ratio determined by the GC-IRMS. For the n-alkanoic acids, the δ13C
vegetal materials (~0.5 g) were extracted with a mixture of dichloro- data were calculated from the FAME values, by correcting them for the
methane/methanol (DCM/MeOH, 3:1, v/v) by ultrasound sonication one carbon atom added in the methanolysis (Abrajano et al., 1994).
following the method described elsewhere (Sánchez-García et al.,
2018). Internal standards (tetracosane-D50, myristic acid-D27, and 2- 2.6. Statistical analysis
hexadecanol) were added prior to extraction. The total lipid extracts
(TLE) were concentrated to ~2 ml using rotary evaporation, and ele- A multivariate principal components analysis (PCA) was performed
mental sulfur was removed with activated cooper. The TLE was sepa- using the Infostat software version 2018 (Di Rienzo et al., 2018) to exam-
rated into neutral and acidic fractions using a Bond-elute column ine the relationship between the five lakes and certain biogeochemical
chromatography (bond phase NH2, 500 mg, 40 μm particle size), by variables. The PCA was based on a correlation matrix combining 10 stan-
eluting with 15 ml DCM/2-propanol (2:1,v/v) and with 15 ml of acetic dardized variables; TOC, TN, δ13C, δ15N, terrestrial sterols (i.e. sum of β-
acid (2%) in diethyl ether, respectively. The neutral lipid fraction was sitosterol, campesterol, or stigmasterol), brassicasterol, carbon preference
then further separated into non-polar and polar fractions, by eluting index (CPI) of n-alkanes (C20-C31), high over middle-molecular weight n-
on an alumina column (ca. 0.5 g of Al2O3 powder in a Pasteur pipet) alkanes (HMW/MMW), terrestrial over aquatic ratio of n-alkanes (TAR),
with 4.5 ml of hexane/DCM (9:1, v/v) and 3 ml of DCM/methanol as well as the pristane over phytane ratio (Pr/Ph).
(1:1, v/v), respectively. The acidic fraction was derivatized with BF3
in methanol and the polar fraction with N,O-bis (trimethylsilyl) 3. Results
trifuoroacetamide (BSTFA), respectively producing fatty acid methyl es-
ters (FAME) and trimethylsilyl alcohol derivatives. 3.1. Mineralogical and geochemical composition of the surface sediments
The three polarity fractions were analyzed by gas chromatography
mass spectrometry using a 6850 GC system coupled to a 5975 VL MSD XRD analysis identified the presence of plagioclases (anorthite and
with a triple axis detector (Agilent Technologies) operating with andesine), quartz, and pyroxenes in the surface sediments of the five
 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668 661

Table 2 Similarly, the lowest concentration of TN was found in the sediments

Mineralogical composition (presence/absence) and concentration (μg·g−1) of inorganic from the Ionospheric and Meltwater lakes (0.03%) and the largest in the
and organic anions, as well as major and minor elements, in the surficial sediment samples
from the five Fildes lakes.
Uruguay Lake (0.26%) (Table 1). The Uruguay and Kitezh surface sedi-
ments contained an order of magnitude larger TOC and TN than the
Mineralogy Drake Meltwater Ionospheric Uruguay Kitezh other three lakes. The TOC over TN ratio (C/N) ranged among the five
(presence/absence)
lakes from 10 to 15 (Fig. 2b). The organic carbon δ13C ratio was measured
Plagioclases x x x x x to vary from −24.2‰ (Uruguay) to −19.5‰ (Meltwater), while the total
Quartz x x x x x
nitrogen δ15N from −2.0‰ (Meltwater) to 1.3‰ (Kitezh). For the fresh
Pyroxenes x x x x x
Zeolites x x vegetal material, the stable isotopic analysis on the bulk biomass pro-
Clays x x duced δ13C values from −29.2 to −22.3‰ (lichens), of −27.2‰ (moss),
and − 28.6‰ (grass), and δ15N ratios from −2.2 to −1.6‰ (lichens),
Anions (μg·g−1) Drake Meltwater Ionospheric Uruguay Kitezh −2.0‰ (moss) and 8.2‰ (grass).
Sulfate 23 153 10 1632 430
Chloride 0.47 14 2.8 49 63
Nitrate n.d. n.d. 0.47 1.06 n.d.
3.2. Lipid biomarkers in the Fildes Lakes surface sediments
Bromide n.d. n.d. n.d. 0.32 n.d.
Fluoride n.d. 0.35 0.12 n.d. 0.39
The GC–MS analysis of the lipid extracts detected the presence of
Major and minor elements Drake Meltwater Ionospheric Uruguay Kitezh diverse lipidic families (n-alkanes, n-alkanoic acids, n-alkanols,
(μg·g−1) isoprenoids, and sterols) in the five surface sediments. n-Alkanes
Na 81 95 34 27 29 were found to be present in total concentrations ranging from 0.29 to
Mg 9.5 16 5.8 2.0 3.3 1.5 μg·g−1 of TOC (Table 3). The sediments from Kitezh showed the
Ca 11 19 7.5 2.0 5.5
largest amount of n-alkanes, whereas those from the other four lakes
Si 4.5 4.4 3.0 7.0 6.7
K 2.6 2.7 1.3 1.5 1.5 varied little between each other. This family was composed of linear
Al 0.10 0.06 0.70 1.5 1.0 and saturated alkanes (i.e. n-alkanes) between C12 and C31, with a gen-
S 5.2 36 14 4.0 7.0 eral dominance of the odd-numbered homologues. The five n-alkanes
Fe 0.10 0.14 0.20 1.0 0.10 profiles (Fig. S1) showed a similar pattern with maxima at either C27,
Ti 0.01 0.004 0.04 0.10 0.07
C29 or C31, and minor additional peaks at C23 (Drake, Ionospheric and
x means presence, and n.d. not detected. Kitezh), C23 and C25 (Meltwater), or C21 (Uruguay). The average chain
length (ACL) of the n-alkanes was virtually the same (i.e. 25) in the
five sediments (Table 3), whereas the carbon preference index (CPI)
ranged from 1.3 to 4.5 (Table 3). Source-diagnosis n-alkanes ratios var-
lakes, whereas zeolites (montmorillonite) and clays (kaolinites) were ied from 4.6 to 30 (terrigenous over aquatic ratio or TAR), from 1.1 to 4.4
only observed in the Uruguay and Kitezh lakes (Table 2). (high- over mid-molecular weight n-alkanes or [HMW/MMW]n-alkanes),
Ion chromatography analysis determined a variable concentration of or from 0.23 to 0.61 (Paq = [C23 + C25] / [C23 + C25 + C29 + C31]; Figs. 3
inorganic anions among the lakes (Table 2). Sulfate and chloride were and 4; see Table 3 for ratios definition).
detected in the five sediments, the former at concentration one or two Together with the n-alkanes, the isoprenoids pristane and phytane
orders of magnitude larger (10–1632 μg·g−1) than the latter (0.47–63 were also found in the five surface sediments, at concentrations ranging
μg·g−1). Uruguay and Kitezh were the lakes containing larger amounts from 0.003 to 0.09 μg·g−1 of TOC (Table 3). The ratio of pristane over
of both anions. Nitrate, bromide and fluoride were only detected in phytane (Pr/Ph) was higher in Drake, Meltwater, and Ionospheric
some of the lakes at concentrations close to or lower than 1 μg·g−1 (5.3–8.4) than in Kitezh and, mostly, Uruguay (Fig. 3c).
(Table 2). Neither nitrite nor phosphate was detected in the Fildes The acidic fraction was observed to contain n-alkanoic acids from C10
sediments. to C30 (Fig. S2), at total concentrations ranging from 1.9 μg·g−1
ICP-MS determined the concentration of the major and minor ele- (Uruguay) to 38 μg·g−1 of TOC (Meltwater). Within this family, the
ments showing different trends (Table 2), where the Drake, Meltwater even dominated over the odd n-alkanoic acids, with maximum peaks
and Ionospheric lakes generally showed larger concentration of Na, at C24, C26 or C28 (Fig. S2). The ACL of the n-alkanoic acids ranged from
Mg, and Ca, whereas Uruguay and Kitezh of Si, Al or Ti. 21 to 25 (Table 3).
The TOC content was measured to vary from 0.27% in the Ionospheric The polar fraction contained an n-alkanols homologous series
to 4.0% in the Uruguay Lake (Fig. 2a), relative to dry weight (% dw). spanning from C10 to C30, with dominance of the even compounds

Fig. 2. Geochemical composition of the bulk organic fraction in the surface sediments from the five Fildes Lakes; (a) concentration of total organic carbon (% relative to the total dry weight)
and (b) C/N ratios versus stable carbon isotopic composition of organic carbon (‰ PDB). The five lakes are identified by their initials in panel b.
662 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668

(Fig. S3). The n-alkanols were measured at total concentrations from The fresh samples produced Paq ratios between 0.53 (grass) and 1.0,
0.13 μg·g−1 (Kitezh) to 3.2 μg·g−1 of TOC (Meltwater). The largest and CPI values between 0.94 and 5.8 (Fig. 4).
content of n-alkanols was found in the surface sediments from the
Meltwater (maximum at C26) and Uruguay (maximum at C22) lakes 3.3. Compound specific isotope analysis of lipids biomarkers
(Table 3). The n-alkanol ACL varied from 19 to 24, and the TAR from
0.38 to 15 (Table 3). Phytol was also detected within the polar fraction The compound-specific isotopic composition of the sedimentary n-
at concentrations ranging from 0.02 to 0.11 μg·g−1 of TOC, together alkanes varied widely (from −23.7 to −39.6‰) between the Fildes
with a series of sterols (β-sitosterol, campesterol, brassicasterol and lakes (Table S4). In general, the variation of δ13C was greater within
stigmasterol) of individual concentrations ≤0.21 μg·g−1 of TOC the LMW and MMW n-alkanes (from −23.7 to −39.6‰), than within
(Table 3). Meltwater showed the largest concentration of β-sitosterol those of HMW (from −27.2 to −33.1‰). Among the lakes, Drake and
(0.21 μg·g−1 of TOC), Drake and Meltwater that of brassicasterol (0.09 Ionospheric showed more homogeneous δ13Cn-alkanes ratios than Melt-
and 0.08 μg·g−1 of TOC, respectively), and Uruguay of campesterol water, Uruguay and Kitezh. Sediments from the two latter lakes had
(0.13 μg·g−1 of TOC) (Table 3). the most δ13C-depleted n-alkanes, namely, C19 in Uruguay and C21 in
The fresh vegetal samples contained n-alkanes spanning from C11 to Uruguay and Kitezh (Fig. 5), whereas Meltwater contained the most
C32 with variable profiles depending on the particular species (Fig. S4). δ13C-enriched alkanes (i.e. C23 and C25) together with Kitezh (C17).
The moss sample (Sanionia unciata) exhibited a distinct maximum The δ13C ratios varied less for the n-alkanoic acids (from −21.0 to
at C23. The Antarctic grass (Deschampsia antarctica) had its maximum −31.7‰). In the five lakes, the LMW n-alkanoic acids (C14, C16, and
at C31, with relevant peaks also at C23 and C25. The two lichen samples C18) showed relatively enriched δ13C ratios (from −22.6 to −24.7‰),
(Placopsis contortuplicata and Lecanora sp.) showed a common whereas those of HMW (C24, C26, and C28) were generally 13C-
maximum at C17, and additional peaks at C21, C22, C23, or C27 (Fig. S4). depleted (from −21.0 to −31.7‰), except for the Meltwater sediments
(Fig. 6).

3.4. PCA analysis

Table 3
Concentration (μg·gOC−1) and compositional distribution of lipid biomarkers in the five
lacustrine sediments. The multivariate PCA produced a separation of the Fildes Lakes into
two groups according to the geochemical composition of their surface
Drake Meltwater Ionospheric Uruguay Kitezh
sediments; Uruguay and Kitezh (group I), versus Drake, Ionospheric
n-Alkanes 0.33 0.40 0.48 0.29 1.5 and Meltwater (group II). Two principal components were extracted
ACL n-alkanes (C10-C31)a 25 25 25 25 25 from the multivariate analysis that accounted together for 79.6% of the
CPI n-alkanes (C20-C31)b 1.3 2.5 1.4 4.5 2.8
HMW n-alkanesc 0.27 0.37 0.41 0.27 1.38
total variance (Fig. 7). The first principal component (PC1) explained
HMW/MMW n-alkanesd 1.6 1.7 1.1 3.4 4.4 56.0% of the variance, and the second (PC2) contributed 23.6%. Taking
TAR n-alkanese 4.6 29 9.7 23 30 into account the eigenvectors magnitude, the most important parame-
Paqf 0.49 0.49 0.61 0.33 0.23 ters on the PC1 are the Pr/Ph ratio, TN, TOC, CPI of n-alkanes, and Paq.
Sum of (C27 + C29 + C31) 0.09 0.15 0.11 0.14 0.76
Pr/Ph and Paq are strongly and positively correlated between them
Pristane 0.09 0.02 0.02 0.02 0.02
Phytane 0.01 0.004 0.003 0.03 0.01 and, to a lesser extent, with δ13C, but negatively correlated with TN,
Pr/Ph 8.4 5.3 5.5 0.62 1.9 TOC, CPI of n-alkanes, and TAR (Table S5). As for the PC2, the most im-
n-Alkanoic acids 2.9 38 8.1 1.9 2.5 portant contributors are the sum of terrestrial sterols, brassicasterol
ACL n-alkanoic acids 22 24 21 25 23 and δ15N (Fig. 7), which are negatively correlated (Table S5).
(C10-C30)a
HMW n-alkanoic acids 1.9 32 4.5 1.7 1.7
TAR n-alkanoic acidsg 2.5 9.2 1.7 17 3.9 4. Discussion
Sum of (C24 + C26 + C28) 1.1 21 2.8 1.0 1.2
n-Alkanols 0.36 3.2 0.15 1.9 0.13 4.1. Heterogeneous carbon sources in the surface sediments from the Fildes
ACL n-alkanol (C10-C30)a 23 24 19 23 21
Lakes
HMW n-alkanolsc 0.28 2.9 0.06 1.7 0.08
Sum of (C26 + C28 + C30) 0.06 0.66 0.005 0.19 0.01
TAR n-alkanolsh 3.2 15 0.38 9.2 1.1 Elemental, isotopic and molecular biomarkers show quantitative
Phytol 0.02 0.06 0.02 0.11 0.01 and qualitative differences in the composition of their sedimentary
β-Sitosteroli 0.13 0.21 0.07 0.18 0.02 OM in the five Fildes Lakes. The deepest lakes (i.e. Uruguay and Kitezh),
Campesterolj 0.07 0.10 0.03 0.13 n.d.
Stigmasterolk 0.03 0.06 0.01 0.11 0.01
contain the OM-richest sediments (i.e. highest TOC and TN) and rela-
Brassicasteroll 0.09 0.08 0.02 0.05 n.d. tively depleted δ13C values (Fig. 2), characteristic of lacustrine environ-
a ments. Stable carbon isotopic signatures are commonly used to
ACLi-n average chain length = ∑(i·Xi + … + n·Xn) / ∑Xi + … + Xn), where X is
concentration (van Dongen et al., 2008). investigate sources of OM, where terrestrial carbon sources (i.e. higher
b
CPIi-n, carbon preference index = ½ ∑ (Xi + Xi+2 + … + Xn) / ∑(Xi−1 + Xi+1 + … plants using the C3 pathway) generally have δ13C values around
+ Xn−1) + 1/2∑(Xi + Xi+2 + … + Xn)/∑(Xi+1 + Xi+3 + … + Xn+1), where X is con- −27‰, aquatic carbon sources (i.e. phytoplankton and algae) typical
centration (van Dongen et al., 2008). values around −21‰ (Fry and Sherr, 1984; Meyers, 1994), and mosses
c
High molecular weight n-alkanes, n-alkanoic acids or n-alkanols, sum of the concen-
trations of homologues ≥C21.
around −23‰ (Boy et al., 2016). In the Fildes Peninsula, the intermedi-
d
High over mid-molecular weight n-alkanes = (C27 + C29 + C31) / (C23 + C25). ate δ13C values in the Uruguay (−24.2‰), Kitezh (−23.4‰), and Iono-
e
Ratio of terrigenous over aquatic n-alkanes = (C27 + C29 + C31 ) / (C17 + C19 ) spheric (−23.4‰) lakes suggest a relevant contribution from terrestrial
(van Dongen et al., 2008). carbon sources and are in the range of those measured in soil samples
f
Paq, ratio of emergent and submerged over floating macrophytes = (C23 + C25) / (C23 +
(from −22.3 to −25.6‰) collected from a glacier retreat gradient on
C25 + C29 + C31) (Ficken et al., 2000).
g
Ratio of terrigenous over aquatic n-alkanoic acids = (C24 + C26 + C28) / (C14 + C16 the Fildes Peninsula Meseta (Boy et al., 2016), extensively covered by
+ C18) (van Dongen et al., 2008). mosses, lichens and hairgrass. In contrast, the 13C-enriched ratios in
h
Ratio of terrigenous over aquatic n-alkanols = (C26 + C28) / (C16 + C18) (van Dongen the OM-poorer Drake and Meltwater sediments (−20.7 and −19.5‰,
et al., 2008). respectively) reveal a relatively greater input from aquatic carbon
i
β-Sitosterol, 24-ethyl-cholest-5-en-3β-ol.
j
Campesterol, 24-methylcholest-5-en-3β-ol.
sources. In the five lakes, C/N ratios ≥10 (Fig. 2b) are in the range of
k
Stigmasterol, 24-ethylcholest-5,22E-dien-3β-ol. those typical of lacustrine sediments that is, between values typical
l
Brassicasterol, 24-methyl cholest-5,22-dien-3β-ol. from algae (4–10) and vascular plants (≥20) (Prahl et al., 1980;
 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668 663

Fig. 3. Lipid proxies diagnostic of the origin of organic matter (OM) and the environmental conditions in the five surface sediments; a) carbon preference index of n-alkanes (biogenic vs.
petrogenic origin), b) ratio of terrestrial over aquatic n-alkanes (terrestrial vs. aquatic OM origin), c) pristane over phytane ratio (oxic vs. anoxic conditions), and d) ratio of high molecular-
weight over medium molecular-weight n-alkanes (relative abundance of higher plants vs. other type of vegetation). Please go to Table 2 for details on the ratios calculations.

Meyers, 1994). Bulk δ 15 N ratios slightly higher or lower than zero islands belonging to the South Shetland archipelago (Livingston, Decep-
(i.e. −2.0 to 1.3‰) indicate the participation of N2-fixation sources tion, Penguin, and Barrientos), where Cabrerizo et al. (2016) explained
(Robinson, 2001) such as land plants, algae, and/or cyanobacteria to those biogenic signals in relation to the general presence of mosses
the sediments of the five lakes. Both bulk-element and isotopic compo- and vascular plants (grass and Antarctic pearlwort). The total concen-
sition suggest a miscellaneous aquatic-terrestrial origin of the organic tration of n-alkanes in the surface sediments of the Fildes Lakes
carbon in the surface sediments from the five Fildes Lakes. (0.29–1.5 μg·gOC−1 or 129–3150 ng·gdw−1) were in the range of
Lipid distribution patterns and biomarkers proxies provide addi- those reported in soil samples from the South Shetland Islands
tional insights on the contribution from different carbon sources to sed- (7–3310 ng·gdw−1; Cabrerizo et al., 2016) or from the McMurdo Dry
imentary OM. In the Fildes Lakes, terrestrial sterols such as β-sitosterol, Valleys of Southern Victoria Land (13–2200 ng·gdw−1), in the Antarctic
campesterol, or stigmasterol (Table 3) are ubiquitously detected in the continent (Matsumoto et al., 1990). In addition, the abundance of LMW
five surface sediments. Consistently, molecular distribution patterns in n-alkanoic acids (≤C20) in the Meltwater and Ionospheric sediments
n-alkanes, n-alkanoic acids, and n-alkanols reveal dominance of HMW (Fig. S2) relative to the other lakes suggests certain aquatic origin of
moieties (i.e. ACL of 25, 21–24, and 19–24, respectively). HMW lipids the sedimentary OM in those lakes. Brassicasterol, a sterol commonly
of odd-over-even (n-alkanes) or even-over-odd (n-alkanoic acids and used as biomarker for freshwater and marine diatoms (e.g. Robinson
n-alkanols) predominance are typical for terrestrial plants (Eglinton et al., 1984; Yunker et al., 1995) is present in all lakes except Kitezh, al-
and Hamilton, 1967; Hedges and Prahl, 1993). n-Alkanes distributions though at larger concentration in Drake and Meltwater (Table 3). Still,
similarly dominated by odd HMW moieties (i.e., n-C25, n-C27, n-C29, TAR ratios higher than the unit (Table 3) are observed in all five sedi-
and n-C31) have been also described in soil samples from four nearby ments, pointing to the dominance of terrestrial over aquatic sources

Fig. 4. n-Alkane proxy Paq indicating the type of terrestrial sources in the surface sediments from the five Fildes Lakes and in fresh vegetal material from the surroundings. Paq end-member
ratios (dotted lines) from Ficken et al. (2000), Nott et al. (2000), Mead et al. (2005).
664 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668

Fig. 5. Compound-specific isotopic composition (δ13C) of n-alkanes in the surface sediments from the five Fildes lakes. Vertical dashed lines divide the distribution of n-alkanes (low, mid-
dle, or high molecular weight) according to characteristic organic-matter sources, that is, phytoplankton and/or bacteria, macrophytes and/or mosses, or higher plants, respectively
(Eglinton and Hamilton, 1967; Fahl and Stein, 1997; Ficken et al., 2000). Particularly diverging δ13C values are highlighted with colored ellipses in the Meltwater (δ13C enriched C23
and C25; orange), Ionospheric (δ13C depleted C17: blue), Uruguay (very δ13C depleted C19 and C21; blue), and Kitezh (δ13C-enriched C17; yellow) lakes. (For interpretation of the references
to color in this figure legend, the reader is referred to the web version of this article.)

particularly in the Kitezh, Uruguay, and Meltwater lakes (Fig. 3b). In their n-alkanes distributions (Cabrerizo et al., 2016). Other Antarctica
Kitezh and Uruguay, the elevated TAR values coincide with the depleted vegetation such as the moss Sanionia uncinata has been measured to
δ13C (Fig. 2) and low Paq ratios (Fig. 4). The alkane Paq ratio (Table 3) is a produce Paq values varying from lower values (0.13–0.52) on the
proxy used to distinguish between different groups of aquatic versus South Shetland Islands (Cabrerizo et al., 2016) to a larger value (0.93)
land plants (Ficken et al., 2000; Mead et al., 2005), where land plants in the Fildes Peninsula (Fig. 4). Paq values in Antarctica lichens may
typically produce proxy values of 0.09, emergent macrophytes of 0.25, also vary from 0.24 to 0.42 (Usnea Antarctica in the South Shetland
mosses around 0.63 (Nott et al., 2000), and submerged/floating macro- Islands; Cabrerizo et al., 2016) to 0.92–1.0 (Placopsis contortuplicata
phytes of 0.69 (Ficken et al., 2000). However, certain vascular plants and Lecanora sp., this study). Thus, the range of Paq found in the
may produce Paq values N0.09, such as the grass Deschampsia antarctica, surface sediments from the Fildes Lakes (0.23–0.61) suggests a mixed
which was measured here to produce a Paq of 0.53 or in the South
Shetland Islands in the range from 0.57 to 0.61, as estimated from

Fig. 7. Biplot diagram of the principal component analysis (PCA) for the five surface
sediments and 10 variables studied; TOC, TN, δ13C, δ15N, terrestrial sterols (i.e., sum of
β-sitosterol, campesterol, or stigmasterol), brassicasterol, CPI of n-alkanes, TARn-alkanes,
Paq, and Pr/Ph. Please see Table 2 for details on the ratios calculation. PC 1 and PC 2 explain
Fig. 6. Compound-specific isotopic composition (δ13C) of source-diagnosis n-alkanoic together 79.6% of the system variance. The five sediment samples are represented as blue
acids in the five lacustrine sediments. To facilitate visualization, two separate plots are dots and the geochemical variables as yellow dots with vectors showing the variable direc-
used to display the δ13C composition of a) low molecular-weight congeners related to pro- tion. The loadings (displayed as correlation coefficients) on PC1 and PC2 are shown in
karyotic sources (C14, C16 and C18), and b) high molecular-weight acids typically produced Table S5. (For interpretation of the references to color in this figure legend, the reader is
by higher plants (C24, C26, and C28). referred to the web version of this article.)
 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668 665

contribution from emergent macrophytes, mosses, and vascular plants composition of the five lacustrine sediments. Overall, all n-alkanes
(Fig. 4). The higher Paq values in the Drake, Meltwater, and Ionospheric (Fig. 5) and the HMW n-alkanoic acids (Fig. 6) show δ13C ratios rela-
sediments denote larger contribution from mosses and lichens, whereas tively depleted relative to the bulk OM (i.e. Δδ13C = 5–13‰) as typically
the lower values in Uruguay and Kitezh reflect greater input from emer- observed on mixed lacustrine pools (Meyers, 1997). In the Drake and
gent macrophytes and/or grass. The difference of Paq signatures in the Ionospheric lakes, all n-alkanes display δ13C values varying little be-
Drake, Meltwater, and Ionospheric lakes versus (mostly) Kitezh may tween each other (2–4‰), in contrast to the other lakes showing δ13C
be related to the greater ground humidity in the northern lands close variations up to 10‰ (Fig. 5). For instance, in Uruguay and Kitezh cer-
to the retreating Collins Glacier, where mosses and swamp vegetation tain LMW n-alkanes (i.e. C19 and C21) are particularly 13C-depleted
find optimum conditions to grow. (i.e. more negative δ13C) relative to their homologues. The δ13C deple-
The CPI values reveal qualitative differences on the sedimentary OM tion in the typically phytoplanktonic n-alkanes C19 and C21 (Thiel
from the five Fildes Lakes. CPI is a proxy broadly used to distinguish be- et al., 1997; Allen et al., 2010) may reflect trophic differences in the phy-
tween biogenic and petrogenic sources of hydrocarbons (Harrison, toplankton population of Uruguay and Kitezh relative to the other lakes.
2007), where recent biogenically derived n-alkanes (e.g. plants or In aquatic systems with low primary productivity (i.e. oligotrophic),
biota exudates) tend to favor odd‑carbon numbered chain lengths, competition for carbon is low, therefore primary producers tend to dis-
whereas aged organic material such as petrogenic (i.e. rock-derived) criminate against the heavier isotope (i.e. 13C), thus resulting in more
hydrocarbons that have undergone ageing over geological timescales negative δ13C values (Hollander and McKenzie, 1991). According to
generally do not exhibit this odd‑carbon number preference due to the TAR values, the weight of aquatic carbon sources in Uruguay, Kitezh,
random cleavage of alkyl chains (Killops and Killops, 2005). Therefore, and Meltwater is lower than in the other two lakes. In Uruguay and
environmental samples with a CPI N 1 suggest that the aliphatic Kitezh in particular, this coincides with their greater depth (i.e. 15 m)
hydrocarbons have arisen thorough biogenic input (Rieley et al., 1991; that typically hinders water mixing. Low availability of nutrients in
Hedges and Prahl, 1993), whereas CPI values approaching 1 indicates poorly mixed, deep water columns and a lesser abundance of primary
petrogenic sources (Harrison, 2007). The CPI values in the surface producers (i.e. oligotrophic conditions) may have caused a lower com-
sediments from the five Fildes Lakes (1.3–4.5) denote a general biogenic petition for carbon in the Uruguay and Kitezh lakes, thus resulting in
origin of the sedimentary organic carbon. Indeed the fresh moss, grass the more negative δ13C values of the planktonic C19 and C21 n-alkanes.
and lichen samples showed CPI values similarly varying from 0.94 to In Meltwater in contrast, the MMW n-alkanes C23 and C25 display δ13C
5.8 (Fig. 4), consistent with CPI values reported on identical or similar values relatively enriched (~2–4‰) compared to the rest of n-alkanes
vegetation species on the South Shetland Islands (Cabrerizo et al., 2016) (Fig. 5). This 13C enrichment in n-alkanes particularly abundant in
calculated from a slightly different range of carbon units (C23-C31, rather the analyzed moss and grass samples (Fig. S4b) may reflect greater
than our C21-C31). CPI values similar to those measured in the Fildes contribution of either type of vegetation to this lake. As the Drake and
sediments were observed in soil samples from the South Shetland Islands Ionospheric lakes, Meltwater is located in the surroundings of Collins
(i.e., 1.4–3.4; Cabrerizo et al., 2016) and the Antarctic McMurdo Dry Glacier, where melting streams fed vast extensions of swamp vegeta-
Valleys (i.e., 1.98–2.59; Matsumoto et al., 2010) and attributed to lichens tion, including moss and grass. In such waterlogged ground, vegetation
and/or vascular plant debris. However, in addition to the dominant can uptake carbon from other sources than atmospheric CO2 (e.g. dis-
biogenic origin, of certain input of petrogenic carbon cannot be ruled solved inorganic carbon or (DIC) (Ficken et al., 2000). The use of a car-
out in the Fildes Lakes, mostly in the Drake and Ionospheric sediments, bon source enriched in 13C (δ13CDIC ~0‰) relative to CO2 (δ13C ~−8‰)
where the lowest CPI values (i.e., 1.3 and 1.4, respectively) were may result into OM with less negative δ3C values. The stronger moss
measured (Fig. 3a). Similarly low CPI values (b1.5) in Antarctic settings signal (13C-enriched C23 and C25) in the Meltwater relative to Drake
on King George Island (Prus et al., 2015) or South Shetland Islands and Ionospheric lakes may be the result of the lower alteration of
(Cabrerizo et al., 2016) were attributed to anthropogenic impacts related biomass upon rapid sinking in this extremely small and shallow lake
to proximity to polar scientific stations (e.g. fossil fuel combustion or oil (i.e. 400 m2 and 1 m depth), thus promoting the preservation of original
spills from ships operations). Although such a contamination factor can- organic signatures in the sediments.
not be completely excluded in the Ionospheric Lake, given the relative The CSIA of n-alkanoic acids provides further evidence of the mixed
proximity to the Uruguayan scientific station (Fig. 1), the fact that the carbon origin in the surface sediments of the Fildes Lakes. Overall, LMW
lowest CPI value is found on the furthest lake to scientific polar stations, n-alkanoic acids majorly derived from algae such as C14, C16 and C18 are
whereas the two lakes closest to them (i.e., Uruguay and Kitezh; Fig. 1) observed to be 13C enriched by 3–5‰ (Fig. 6), according to the major use
show the highest CPI values, suggest there must be other factors than of 13C-enriched carbon by primary producers (Cranwell et al., 1987). In
anthropogenic explaining the petrogenic signature observed. Petrogenic contrast, land plant-derived HMW n-alkanoic acids such as C24, C26, and
(i.e. rock-derived) hydrocarbons in the Drake and Ionospheric lakes C28 (Rieley et al., 1991) are relatively 13C depleted in the surface sedi-
may rather come from physical erosion of bedrock (Galy et al., 2015). ments of all lakes but Meltwater (Fig. 6). The δ13C duality generally ob-
Given the proximity of the two lakes to the retreating Collins Glacier, served in the Fildes Lakes illustrates the combined contribution from
we hypothesize that ancient carbon may be released from recently ex- terrestrial and aquatic carbon sources to the surface sediments, whereas
posed surfaces of bedrock where erosion-protective vegetation has not the peculiar 13C enrichment of the HMW n-alkanoic acids in Meltwater
yet developed as in longer-exposed further to the southwest. In addition, supports the greater input of swamp vegetation using 13C-enriched
the low CPI values in the northern lakes may also denote certain contribu- sources of carbon (i.e. DIC) to the sediments of this shallow lake.
tion from bacterial sources, where microbial degradation would result
into OM with lack of carbon-number preference (Killops and Killops, 4.2. Factors influencing the sedimentary organic matter composition in the
2005). In glacial environments, ice is colonized by microbial organisms five Fildes Lakes
adapted to live in such cold conditions (e.g., chemoheterotrophs, anaero-
bic nitrate reducers, sulfate reducers or methanogens) (Skidmore et al., The elemental, molecular, and isotopic results on the five sediment
2000), and sometimes it is covered by thin layers of wind-transported samples allow sorting the Fildes Lakes into two different groups
dark powdered material mainly composed by microbial derivate com- (Fig. 7), according to their geochemical composition. The Uruguay and
pounds called cryoconite (Xu et al., 2010). Upon melting, the glacial ice Kitezh lakes form Group I, which surface sediments are the OM richest
constitutes a potential source of microbial OM to the surrounding lakes (high TOC and TN) and freshest (high CPI), with organic carbon largely
(Skidmore et al., 2000). derived from terrestrial sources (i.e. high TARn-alkanes and depleted bulk
The compound-specific isotopic composition (δ13C) of n-alkanes and δ13C). Sedimentary conditions in these two lakes appear to be more or
n-alkanoic acids also reveal qualitative differences in the molecular less anoxic according to Pr/Ph ratios lower (Uruguay) or slightly higher
666 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668

(Kitezh) than 1 (Fig. 3c). Pristane and phytane are photosynthetic area. However, logistical constrains and hard working conditions
biomarkers (Ishiwatari et al., 1999) majorly originated from phytol (i.e., highly variable meteorology, extremely cold water temperatures,
(Brocks and Summons, 2003), which tends to degrade to either pristane diving time frames of 15–20 min maximum, sharing the scuba diver
or phytane depending on conditions of abundance or absence of oxy- working time, campaign of fixed time-period, etc.) restricted our
gen, respectively (Peters et al., 2005). The Pr/Ph ratio is commonly sampling capability to only one sample per lake. Counting on scuba di-
used for reconstructing paleoredox conditions in sedimentary environ- vers with exclusive dedication to our project or longer diving turns
ments on early diagenetic states (Blumenberg et al., 2012), where would clearly increase the samples retrieving effectiveness, but these
values lower than 1 denote anoxia. In the deepest watersheds of are aspects beyond our reach. Within our capacity of decision, we
Uruguay and Kitezh, mixing along the water column is not favored, could have decided to do more diving immersions in any/some of the
thus promoting anoxic conditions at the lower water layers and surface lakes, but that would have necessarily resulted into a lower number of
sediments, where OM relatively preserved (i.e. fresh) tends to accumu- lakes sampled. On the other hand, working with surface sediments
late. This explains the lower Pr/Ph ratios and greater CPI values limited the interpretation frame to a recent past (i.e. few decades).
observed in the surface sediments from Uruguay and Kitezh relative to Certainly, using sediment cores representing a larger temporal se-
the other lakes. The mineralogy of the sediments in those two lakes quence would increase our interpretation capability to expand the
denotes certain alteration, as zeolites and clays were found together paleoreconstruction to a longer-term frame (at least centuries). The
with the widespread basaltic materials (Table 2). The presence of higher temporal resolution of a core versus surface sediments would
secondary aluminous clay in the Uruguay and Kitezh sediments may offer a great perspective for assessing longer-term processes such as
be explained by the chemical weathering of the original basaltic the effects of climate change on this poorly studied region. Indeed,
bedrock, as previously described in lakes from the extreme southern this is our plan for a near future. The present study was conceived as a
part of the Fildes Peninsula in the region furthest from Collins Glacier first approximation to the sediments biogeochemistry in the Fildes
(Alfonso et al., 2015). Furthermore, a marine fingerprint is observed in Lakes, where the study of surface sediments allowed characterizing
the sediments of these two lakes in form of relative abundance of sulfate recent sediments for their origin, preservation and transport. Future
and chloride anions (Table 2), likely transported from the coast by work is envisioned to investigate the molecular and isotopic profile
marine aerosols and seawater drops. of the long-term sedimentary record of the Fildes Peninsula lakes to
The Drake, Ionospheric, and Meltwater lakes comprise Group II, reconstruct the temporal evolution and response to global warming.
which sediments contain relatively less biomass (low TOC and TN) In a recent sampling campaign (March 2019), sediment cores of
that is less fresh (lower CPI) and largely derived from swamp vegetation 1–1.5 m depth have been already collected from a number of lakes
(high Paq) and/or aquatic (low TAR; just Drake and Ionospheric) carbon (including Uruguay).
sources. Pr/Ph ratios considerably larger than 1 denote a relatively
greater abundance of oxygen in the water column of these three lakes 5. Conclusion
compared with those from Group I (Fig. 3c). Oxic conditions are charac-
teristic of well mixed waters. In the shallow Drake and Meltwater (2 and Bulk geochemistry, mineralogy, isotopic and molecular biomarkers
1 m depth, respectively), high oxygen concentration is likely assured revealed different sources, oxygen conditions, and factors affecting the
thanks to the mixing effect of the frequent and strong polar winds. In supply and burial of sediments in the five Fildes Lakes. Different lipid
the Ionospheric Lake, the continuous water input from the leaking proxies and the dominance of HMW moieties reflect the importance
Collins Glacier and output to the Maxwell Bay likely favor the mixing of terrestrial carbon sources in the five surface sediments, whereas
and oxygenation of water in this deeper watershed (10 m depth). The aquatic carbon sources have relatively greater weight in the Drake and
low TOC and CPI values in the three lakes, together with the monoto- Ionospheric lakes. The amount and composition of the five surface sed-
nous distribution of 13C-depleted n-alkanes in Drake and Ionospheric iments suggest relative larger inputs of petrogenic carbon sources to the
suggest certain contribution of petrogenic carbon sources (Drenzek Drake, Ionospheric and Meltwater lakes, while dominantly biogenic to
et al., 2007), likely linked to erosion of recently-exposed bedrock mate- Uruguay and Kitezh. The inorganic and organic geochemistry of the
rial (i.e. ancient carbon) and glacial meltwater supply (Hood et al., sediments reveals additional information about the location and charac-
2009). This is consistent with the fresh mineralogy (i.e., basaltic with teristics of the lacustrine watersheds. Water mixing and oxygenation is
absence of secondary aluminous clays) and relative abundance of Na, relatively poorer in the deepest Uruguay and Kitezh, where influence
Mg, and Ca cations in the three lakes closest to Collins Glacier. The from the nearby coast seems to be relevant. In contrast, the well-
glacial outflow from the retreating glacier flushes water bearing mixed and oxygenated watersheds of Drake, Ionospheric, and Meltwa-
thaw and eroded particulate material into the Drake, Ionospheric and ter appear to be more influenced by the retreating Collins Glacier that
Meltwater lakes, thus imprinting their sediments with a similar inor- releases ancient carbon and basaltic materials long time preserved
ganic signature. Moreover, the continuous meltwater leaking enhances upon freezing. Our results suggest that the geochemical signature of
the ground humidity and waterlogging that provides the suitable the surface sediments in the five lakes is largely determined by either
edaphic conditions for the swamp vegetation to grow and expand the distance to the retreating Collin Glacier or the proximity to the
nearby Collins Glacier. The relatively greater imprint from mosses in coast, as well as the lake depth. The combined use of bulk, molecular,
Meltwater may be the result of a concentration effect owed to the and compound-specific isotopic biomarkers allowed deciphering rele-
very small size of the lake. vant information for reconstructing the sediments biogeochemistry
In sum, mineralogical, inorganic, isotopic, and lipid biomarker prox- the five Fildes Lakes (origin, preservation, and geochemistry), with rel-
ies reveal compositional differences in the surface sediments from the evance to understand future alterations and geochemical processes in
five Fildes Lakes, where factors such as the distance to the glacial ice this highly vulnerable system upon the current climate warming.
margin, the proximity to the coast, and the lake depth seem to play a
key role in determining their geochemical signature. Acknowledgments

4.3. Study limitations This study was partially founded by the Instituto Antártico Uruguayo
(IAU). Scuba Divers (Rodrigo Toledo and Oscar Correa) are acknowl-
The present biogeochemical study involves certain limitations edged for their field assistance during sampling. All authors thank the
(majorly related with the sampling strategy) that should be mentioned. crew members from the Artigas Base (BCAA). P. M. Sarmiento is ac-
For instance the sampling size; a larger number of subsamples per lake knowledged for the stable isotopes analysis. The authors D. Carrizo
would be desirable for a better statistical representativity of the studied and L. Sánchez-García acknowledge the Spanish Ministry of Science,
 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668 667

Innovation and Universities (MICINN/FEDER) for funding their projects Hood, E., Fellman, J.B., Spencer, R.G.M., Hernes, P.J., Edwards, R.T., D'Amore, D., Scott, D.,
2009. Glaciers as a source of ancient and labile organic matter to the marine environ-
(RYC-2014-19446 and CGL2015-74254JIN, respectively). ment. Nature 462, 1044–1047. https://doi.org/10.1038/nature08580.
Huang, Y., Lockheart, M.J., Logan, G.A., Eglinton, G., 1996. Isotope and molecular evidence
Appendix A. Supplementary data for the diverse origins of carboxylic acids in leaf fossils and sediments from the
Miocene Lake Clarkia deposit ID, USA. Org. Geochem. 24, 289–299.
Ishiwatari, R., Yamada, K., Matsumoto, K., Houtatsu, M., Naraoka, H., 1999. Organic
Supplementary data to this article can be found online at https://doi. molecular and carbon isotopic records of the Japan Sea over the past 30 kyr.
org/10.1016/j.scitotenv.2019.03.459. Paleoceanography 14, 260–270.
Killops, S., Killops, V., 2005. Introduction to Organic Geochemistry. Blackwell Publishing,
Oxford.
Lamb, A.L., Wilson, G.P., Leng, M.J., 2006. A review of coastal palaeoclimate and relative
References sea-level reconstructions using δ13C and C/N ratios in organic material. Earth-Sci.
Rev. 75, 29–57.
Abrajano, T.A., Murphy, D.E., Fang, J., Comet, P., Brooks, J.M., 1994. 13C/12C ratios in indi- Lee, K., Yoon, S.K., Yoon, H.I., 2009. Holocene paleoclimate changes determined using
vidual fatty acids of marine mytilids with and without bacterial symbionts. Org. diatom assemblages from lake Long, King George Island, Antractica. J. Paleolimnol.
Geochem. 21, 611–617. 42, 1–10.
Alfonso, J.A., Vasquez, Y., Hernandez, A.C., Mora, A., Handt, H., Sira, E., 2015. Geochemistry Li, S.P., Ochyra, R., Wu, P.C., Seppelt, R.D., Cai, M.H., Wang, H.-Y., Li, C.S., 2009.
of recent lacustrine sediments from Fildes Peninsula, King George Island, maritime Drepanocladus longifolius (Amblystegiaceae), an addition to the moss flora of King
Antarctica. Antarct. Sci. 27, 462–471. George Island, South Shetland Islands, with a review of Antartic benthic mosses.
Allen, M.A., Neilan, B.A., Burns, B.P., Jahnke, L.L., Summons, R.E., 2010. Lipid biomarkers in Polar Biol. 32, 1415–1425.
Hamelin Pool microbial mats and stromatolites. Org. Geochem. 41, 1207–1218. Machado, A., Lima, E.F., Chemale, J.F., Morata, D., Oteiza, O., Almeida, D.P.M., Figuereido,
Antarctic Treaty Secretariat, 2009. Fildes Peninsula, King George Island. Management Plan A.M.G., Alexandre, F.M., Urrutia, J.L., 2005. Geochemistry constraints of Mesozoic-
for Antarctic Specially Protected Area No. 125: Measure 6, Annex. https://www.ats. Cenozoic calc-alkaline magmatism in the South Shetland arc, Antarctica. J. S. Am.
aq/documents/recatt/Att424_e.pdf Earth Sci. 18, 407–425.
Arbi, I., Liu, S., Zhang, J., Wu, J., Huang, X., 2018. Detection of terrigenous and marine or- Malandrino, M., Abollino, O., Buoso, S., Casalino, C.E., Gasparon, M., Giacomino, A., La Gioia,
ganic matter flow into a eutrophic semi-enclosed bay by δ13C and δ15N of intertidal C., Mentasti, E., 2009. Geochemical characterization of Antartic soils and lacustrine
macrobenthos and basal food sources. Sci. Total Environ. 613, 847–860. https://doi. sediments from Terra Nova bay. Microchem. J. 92, 21–31.
org/10.1016/j.scitotenv.2017.09.143. Matsumoto, G.I., Akiyama, M., Watanuki, K., Torii, T., 1990. Unusual distributions of long-
Arts, M.T., Brett, M.T., Kainz, M.J., 2009. Lipids in Aquatic Ecosystems. Springer, Dordrecht, chain n-alkanes and n-alkenes in Antarctic soil. Org. Geochem. 15, 403–412.
p. 380. Matsumoto, G.I., Honda, E., Sonoda, K., Yamamoto, S., Takemura, T., 2010. Geochemical
Barsh, D., Mausbacher, R., 1986. New data on the relief development of the South Shet- features and sources of hydrocarbons and fatty acids in soils from the McMurdo
land Islands, Antarctica. Interdiscip. Sci. Rev. 11, 211–219. dry valleys in the Antarctic. J. Polym. Sci. 4, 187–196.
Blumenberg, M., Thiel, V., Riegel, W., Kah, L.C., Reitner, J., 2012. Biomarkers of black shales Mead, R., Xu, Y.P., Chong, J., Jaffe, R., 2005. Sediment and soil organic matter source
formed by microbial mats, Late Mesoproterozoic (1.1 Ga) Taoudeni Basin, Mauritania. assessment as revealed by the molecular distribution and carbon isotopic composi-
Precambrian Res. 196–197, 113–127. tion of n-alkanes. Org. Geochem. 36, 363–370.
Boy, J., Godoy, R., Shibistova, O., Boy, D., McCulloch, R., Andrino de la Fuente, A., Aguirre- Mendoça, T., Melo, V.F., Schaefer, C.E.G.R., Simas, F.N.B., Michel, R.F.M., 2013. Clay miner-
Morales, M., Mikutta, Guggenberger, G., 2016. Successional patterns along soil devel- alogy of gelic soils from the Fildes Peninsula, maritime Antarctica. Soil Sci. Soc. Am. J.
opment gradients formed by glacier retreat in the Maritime Antarctic, King George Is- 77, 1842–1851.
land. Rev. Chil. Hist. Nat. 89, 6. Meyers, P.A., 1994. Preservation of elemental and isotopic source identification of
Braun, M., Saurer, H., Goßmann, H., 2004. Climate, energy fluxes and ablation rates on the sedimentary organic matter. Chem. Geol. 144, 289–302.
ice cap of King George Island. Brazilian Antarctic Research. vol. 4, pp. 87–103. Meyers, P.A., 1997. Organic geochemical proxies of paleoceanographic, paleolimnologic,
Brocks, J.J., Summons, R.E., 2003. In: Holland, H.D., Turekian, K.K. (Eds.), Biogeochemistry. and paleoclimatic processes. Org. Geochem. 27, 213–250.
Elsevier, Amsterdam, pp. 63–115. Meyers, P.A., 2003. Applications of organic geochemistry to paleolimnological reconstruc-
Cabrerizo, A., Tejedo, P., Dachs, J., Benayas, J., 2016. Anthropogenic and biogenic hydrocar- tions: a summary of examples from the Laurentian Great Lakes. Org. Geochem. 34,
bons in soils and vegetation from the South Shetland Islands (Antarctica). Sci. Total 261–289.
Environ. 569, 1500–1509. Michel, R.F.M., Schaefer, C., Poelking, E.L., Simas, F.N.B., Fernandes, E.I., Bockheim, J.G.,
Castañeda, I.S., Schouten, S., 2011. A review of molecular organic proxies for examining 2012. Active layer temperature in two Cryosols from King George Island,
modern and ancient lacustrine environments. Quat. Sci. Rev. 30, 2851–2891. Maritime Antarctica. Geomorphology 155, 12–19. https://doi.org/10.1016/j.
Cohen, A.S., 2003. Paleolimnology: The History and Evolution of Lake Systems. Oxford geomorph.2011.12.013.
University Press, New York (500 pp.). Michel, R.F.M., Schaefer, C., Lopez-martinez, J., Simas, F.N.B., Haus, N.W., Serrano, E.,
Cranwell, P.A., 1973. Chain-length distribution of n-alkanes from lake sediments in rela- Bockheim, J.G., 2014. Soils and landforms from Fildes Peninsula and Ardley Island,
tion to post-glacial environmental change. Freshw. Biol. 3, 259–265. maritime Antarctica. Geomorphology 225, 76–86.
Cranwell, P.A., Eblinton, G., Robinson, N., 1987. Lipids of aquatic organisms as potential Monien, P., Schnetger, B., Brumsack, H.J., Hass, H.C., Khun, G., 2011. A geochemical record
contributors to lacustrine sediments. Org. Geochem. 11, 513–527. of late Holocene palaeoenvironmental changes at King George Island (maritime
Di Rienzo, J.A., Casanoves, F., Balzarini, M.G., Gonzalez, L., Tablada, M., Robledo, C.W., 2018. Antarctica). Antarct. Sci. 23, 255–267.
Centro de Transferencia InfoStat, FCA, Universidad Nacional de Córdoba, Argentina. Nna-Mvondo, D., Martin-Redondo, M.P., Martinez-Frías, J., 2008. New application of mi-
http://www.infostat.com.ar (URL, InfoStat version). crowave digestion-inductively coupled plasma-mass spectrometry for multi-
Drenzek, N.J., Montlucon, D.B., Yunker, M.B., Macdonald, R.W., Eglinton, T.I., 2007. Con- element analysis in komatiites. Anal. Chim. Acta 628, 133–142.
straints on the origin of sedimentary organic carbon in the Beaufort Sea from coupled Nott, C.J., Xie, S., Avsejs, L.A., Maddy, D., Chambers, F.M., Evershed, R.P., 2000. n-Alkane
molecular 13C and 14C measurements. Mar. Chem. 103, 146–162. distributions in ombotrophic mires as indicators of vegetation change related to cli-
Eglinton, G., Hamilton, R.J., 1967. Leaf epicuticular waxes. Science 156, 1322–1335. matic variation. Org. Geochem. 31, 231–235.
Embleton-Hamann, C., 2004. Proglacial landforms. In: Goudie, A.S. (Ed.), Encyclopedia of Ouyang, X., Guo, F., Bu, H., 2015. Lipid biomarkers and pertinent indices from aquatic en-
Geomorphology. vol. 2. Routledge, London, pp. 810–813. vironment record paleoclimate and paleoenvironment changes. Quat. Sci. Rev. 123,
Fahl, K., Stein, R., 1997. Modern organic carbon deposition in the Laptev sea and the 180–192.
adjacent continental slope: surface water productivity vs. terrigenous input. Org. Pancost, R.D., Boot, C.S., 2004. The palaeoclimatic utility of terrestrial biomarkers in ma-
Geochem. 26, 379–390. rine sediments. Mar. Chem. 92, 239–261.
Ficken, K.J., Li, B., Swain, D.L., Eglington, G., 2000. An n-alkane proxy for the sedimentary Parro, V., Diego-Castilla, G., Moreno-Paz, M., Blanco, Y., Cruz-Gil, P., Rodríguez-Mandredi,
input of submerged/floating freshwater aquatic macrophytes. Org. Geochem. 31, J.A., Férnandez-Remolar, D., Gómez, F., Gómez, M.J., Rivas, L.A., Demergasso, C.,
745–749. Echeverría, A., Urtuvia, V.N., Ruiz-Bermejo, M., García-Villadangos, M., Postigo, M.,
Freeman, K.H., Wakeham, S.G., Hayes, J.M., 1994. Predictive isotopic biogeochemistry: hy- Sánchez-Román, M., Chong-Diáz, G., Gómez-Elvira, J., 2011. A microbial oasis in the
drocarbons from anoxic marine basins. Org. Geochem. 21, 629–644. hypersaline Atacama subsurface discovered by a life detector chip: implications for
French, H.M., 2007. The Periglacial Environment. John Wiley V.C.H, Chichester (478 pp.). the search of life on Mars. Astrobiology 11, 969–996.
Fry, B., Sherr, E.B., 1984. Delta-C-13 measurements as indicators of carbon flow in marine Peter, H.U., Buesser, C., Mustafa, O., Pfeifer, S., 2008. Risk assessment for the Fildes peninsula
and fresh-water ecosystems. Contrib. Mar. Sci. 27, 13–47. and Ardley Island, and development of management plans for their designation as
Galy, V., Peucker-Ehrenbrink, B., Eglinton, T., 2015. Global carbon export from the terres- Specially Protected or Specially Managed Areas. Research Report 203 12 124 UBA-FB
trial biosphere controlled by erosion. Nature 521, 204–207. 001155e (508 pp.).
Hall, B.L., Koffman, T., Denton, G.H., 2010. Reduced ice extent on the western Antarctic Peters, K.E., Walters, C.C., Moldowan, J.M., 2005. The Biomarker Guide - Part II - Biomarkers
Peninsula at 700–970 cal yr BP. Geology 38, 635–638. and Isotopes in Petroleum Exploration and Earth History. Cambridge University Press,
Harrison, R.M., 2007. In: Harrison, R.M. (Ed.), Principles of Environmental Chemistry. The New York.
Royal Society of Chemistry, Cambridge (UK) (363 pp.). Prahl, F.G., Bennett, J.T., Carpenter, R., 1980. The early diagenesis of aliphatic hydrocarbons
Hedges, J.I., Prahl, F.G., 1993. Early diagenesis: consequences for applications of molecular and organic matter in sedimentary particles from Dabob Bay, Washington. Geochim.
biomarkers. In: Engel, M.H., Mecko, S.A. (Eds.), Organic Geochemistry, Principles and Cosmochim. Acta 44, 1967–1976.
Applications. Plenum Press, New York, pp. 237–253. Prus, W., Fabianska, M.J., Labno, R., 2015. Geochemical markers of soil anthropogenic con-
Hollander, D.J., McKenzie, J.A., 1991. CO2 control on carbon-isotope fractionation during taminants in olar scientific stations nearby (Antarctica, King George Island). Sci. Total
aqueous photosynthesis: a paleo-pCO2 barometer. Geology 19, 929–932. Environ. 518, 266–279.
668 D. Carrizo et al. / Science of the Total Environment 672 (2019) 657–668

Quayle, W.C., Peck, L.S., Peat, H., Ellis-Evans, J.C., Harrigan, P.R., 2002. Extreme responses Thomas, E.R., Tetzner, D.R., 2018. The climate of the Antarctic Peninsula during the twen-
to climate change in Antarctic lakes. Science 295, 645. tieth century: evidence from ice cores. In: IntechOpen (Ed.), Antarctica - A Key to
Révész, K., Qi, H., and Coplen, T.B., 2012. Determination of the δ15N and δ13C of total ni- Global Change https://doi.org/10.5772/intechopen.81507 (18 pp.).
trogen and carbon in solids. RSIL lab code 1832, chap. 5 of Stable isotope-ratio Torres-Mellado, G.A., Jana, R., Casanova-Katny, M.A., 2011. Antarctic hairgrass expansion
methods, sec. C of Révész, Kinga, and Coplen, T.B. eds. Methods of the Reston Stable in the South Shetland archipelago and Antarctic Peninsula revisited. Polar Biol. 34,
Isotope Laboratory (slightly revised from version 1.1 released in 2007): U.S. Geolog- 1679–1688.
ical Survey Techniques and Methods, book 10, 31 p., available only at https://pubs. van Dongen, B.E., Semiletov, I., Weijers, J.W.H., Gustafsson, Ö., 2008. Contrasting lipid bio-
usgs.gov/tm/2006/tm10c5/. (Supersedes versions 1.0 and 1.1 released in 2006 and marker composition of terrestrial organic matter exported from across the Eurasian
2007, respectively.) Arctic by the five Great Russian Arctic Rivers. Global Biogeochem. Cy. 22, GB1011.
Rieley, G., Collier, R.J., Jones, D.M., Eglinton, G., 1991. Biogeochemistry of Ellesmere Lake, https://doi.org/10.1029/2007GB002974.
U.K. I. Source correlation of leaf wax inputs to the sedimentary lipid record. Org. Vera, L., Fernandez-Turiel, T., Quesada, A., 2013. A distribution and reproductive capacity
Geochem. 17, 901–912. of Dechampsia antarctica and Colobanthus quitensis on Byers Peninsula, Livingston
Robinson, D., 2001. δ15N as an integrator of the nitrogen cycle. Trends Ecol. Evol. 16, Island, South Shetland Islands, Maritime Antarctica. Antarct. Sci. 25, 292–302.
153–162. Vieira, R., Marotta, H., Kellem da Rosa, K., Jaña, R., Lorenz-Simoes, C., de Souza-Junior, E.,
Robinson, N., Eglinton, G., Brassell, S.C., Cranwell, P.A., 1984. Dinoflagellate origin for sed- Ferreira, F., Ronquette-Santos, L., dos Santos, J.V., Perroni, M.A., Goncalves, M.,
imentary 4-alpha-methylsteroids and 5-alpha(H)-stanols. Nature 308, 439–442. Farias-Santos, J.P., Rodrigues, R.I., Medieros Galvao, J.C., de Sá Felizardo, J.P., 2015.
Sánchez-García, L., Aeppli, C., Parro, V., Fernandez-Remolar, D., García-Villadangos, M., Análisis Sedimentológico y geomorfológico de áreas lacustres en la Península Fildes,
Chong-Díaz, G., Blanco, Y., Carrizo, D., 2018. Molecular biomarkers in the subsurface of Isla Rey Jorge, Antártica Marítima (in Spanish). Investig. Geogr. Chile 49, 03–33.
the Salar Grande (Atacama, Chile) evaporitic deposits. Biogeochemistry 140, 31–52. Vogts, A., Moossen, H., Rommerskirchen, F., Rullkotter, J., 2009. Distribution patterns and
Serrano-Cañadas, E., 2003. Paisaje Natural y pisos geoecológicos en las áreas libres de stable carbon isotopic composition of alkanes and alkan-1-ols from plant waxes of
hielo de la Antártida marítima: Islas Shetland del Sur. Boletín de la A.G.E. N.° 35, 5–32. African rain forest and savanna C3 species. Org. Geochem. 40, 1037–1054.
Simoneit, B.R.T., Sheng, G., Chen, X., Fu, J., Zhang, J., Xu, Y., 1991. Molecular marker study Xianshen, Z., Xiaohan, L., 1990. Geology of volcanic rocks on Fildes Peninsula, King George
of extractable organic matter in aerosols from urban areas of China. Atmos. Environ. Island. West Antarctica. Chin. J. Polar Res. 1, 8–19.
10, 2111–2129. Xu, Y., Simpson, A.J., Eyles, N., Simpson, M.J., 2010. Sources and molecular composition of
Skidmore, M.L., Foght, J.M., Sharp, M.J., 2000. Microbial life beneath a high arctic glacier. cryoconite organic matter from the Athabasca Glacier, Canadian Rocky Mountains.
Appl. Environ. Microbiol. 66, 3214–3220. Org. Geochem. 41, 177–186.
Smellie, J.L., Pankhurst, R., Thomson, M.R.A., Davies, R.E:.S., 1984. The geology of the South Yunker, M.B., Macdonald, R.W., Veltkamp, D.J., Cretney, W.J., 1995. Terrestrial and marine
Shetland Islands, VI. Stratigraphy, geochemistry and evolution. British Antarctic biomarkers in a seasonally ice-covered Arctic estuary — integration of multivariate
Survey Scientific Report, No. 87 (85 pp.). and biomarker approaches. Mar. Chem. 49, 1–50.
Thiel, V., Merz-preiβ, M., Reitner, J., Michaelis, W., 1997. Biomarker studies on microbial
carbonates: extractable lipids of a calcifying cyanobacterial mat (Everglades, USA).
Facies 36, 163–172.