ArticlePDF Available

Alkenone paleothermometry: Biological lessons from marine sediment records off western South America

January 2006
Geochimica et Cosmochimica Acta 70(1):101-117

January 2006
70(1):101-117

DOI:10.1016/j.gca.2005.08.023

Authors:

Oregon State University

An empirical global core-top calibration that relates the alkenone unsaturation index to mean annual SST (maSST) is statistically the same as that defined for a subarctic Pacific strain of Emiliania huxleyi (CCMP1742) grown exponentially in batch culture under isothermal conditions. Although both equations have been applied widely for paleoSST reconstruction, uncertainty still stems from two key ecological factors: variability in the details of biosynthesis among genetically distinct alkenone-producing strains, and impacts of non-thermal physiological growth factors on . New batch culture experiments with CCMP1742 here reveal that diverge systematically from the core-top calibration in response to nutrient depletion and light deprivation, two physiological stresses experienced by phytoplankton populations in the real ocean. Other aspects of alkenone/alkenoate composition also respond to these stresses and may serve as signatures of such effects, providing an opportunity to detect, understand, and potentially correct for such impacts on the geologic record. A test case documents that sediments from the Southeast Pacific display the alkenone/alkenoate compositional signature characteristic of cells physiologically stressed by light deprivation. Such an observation could be explained if marine snow provided a major vector of sedimentation for these biomarkers. Late Pleistocene records in the Southeast Pacific yield plausible paleotemperature histories of ice-age cooling, but ice-age alkenone/alkenoate signatures fall outside the range of modern calibration samples of similar . They better match core-top samples deposited beneath waters characterized by much cooler maSST, suggesting key features of ice-age ecology for alkenone-producing haptophytes were different from today, and that the index taken alone may misgauge the total range of ice-age cooling at these locations. Analysis of the full spectrum of alkenone/alkenoate compositions preserved in sediments opens up a new opportunity that may improve the accuracy of paleotemperature estimates based on simple analysis and help resolve longstanding disagreements between various paleotemperature reconstruction methods.

Map on the left shows locations along the continental margin off the west coast of South America where sediment samples were obtained for detailed alkenone/alkenoate analysis. Solid symbols depict sites where surface sediments from core-tops were obtained from 33 multicores, three (of four) gravity cores, and three piston cores (Table 1); open stars depict three piston core sites where sediments were obtained stratigraphically with depth (Table 2). Dashed lines represent isotherms for mean annual sea-surface temperature (maSST) (Ocean Climate Laboratory, 2003). Major oceanographic currents and features in the study region (Strub et al., 1998) are also identified (bold arrows) to provide context for discussion. The graph on the right shows a scatter plot for U K 0 37 values measured in surface sediments versus maSST in waters overlying each multicore site (all data from Table 1). Dotted line depicts the global calibration equation ðU K 0 37 ¼ 0:033T þ 0:044Þ established for surface sediment collected from 60°N to 60°S throughout all ocean basins (Muller et al., 1998). Dashed line depicts the U K 0 37 versus temperature relationship established for a single strain of E. huxleyi grown in laboratory-controlled batch cultures (U K 0 37 ¼ 0:034T þ 0:039: Prahl et al., 1988).

…

Scatter plot of values for (A) %K37, (B) ME/K37, and (C) EE/ME versus U K 0 37 measured in surface sediments (Table 1) and three piston cores (Y69- 71P, Y71-6-12P, and RR9702A-11PC: Table 2) from the continental margin off the west coast of South America (see text for definition of each compositional property). The black lines represent trends defined by data for the same compositional properties measured in a Ônutrient stressÕ (solid) and Ôdark stressÕ (dashed) experiment with E. huxleyi grown isothermally at 15 °C under laboratory-controlled, batch culture conditions. The solid grey lines represent trends defined by time series data obtained from an experiment where Ôdark stressÕ was imposed with a simultaneous 5 °C decrease in growth temperature (i.e., 20–15 °C). The dashed grey lines represent the trend for a control on the latter experiment, i.e., case where the 20–15 °C temperature shift was imposed with no dark stress. The letter E associated with each line depicts the composition of exponentially dividing cells at the start of each experiment. The arrowhead on each line indicates the time series trajectory of compositional change caused by each stress. All culture data used to draw the lines depicted in this figure are given in Table 3. The filled and open symbols for core data correspond to sediments deposited since and prior to the LGM (21 Ka), respectively.

…

Profiles for estimates of U K 0 37 -based sea-surface temperature (SST) with depth in three piston cores from the study area (Fig. 1): (A) Y69-71P; (B) Y71-6-12P; and (C) RR9702A-11PC. SST estimates were calculated using a standard calibration equation (U K 0 37 ¼ 0:034T þ 0:039: Prahl et al., 1988). The depth profile for stable oxygen isotope measurements (d 18 O) on benthic foraminiferal tests from discrete depths in Y79-71P (Clark et al., 2004) and Y71-6- 12P (Feldberg and Mix, 2003) are also plotted in (A) and (B) for stratigraphic reference purposes.

…

Summary of results from batch culture experiments with E. huxleyi strain CCMP1742

…

Stratigraphic profiles for EE/ME values measured in three piston cores from our study area (Y69-71P; Y71-6-12P; and RR9702A-11PC, see stars in Fig. 1). An age model was established for the first two cores using available depth records for d 18 O in benthic foraminiferal shells and radiocarbon dates on foraminiferal carbonate in sediment samples from discrete depths (Pisias and Mix, 1997). For the third core, ages were assigned more arbitrarily, assuming a constant rate of sedimentation and the LGM time horizon (21 Kyr) corresponds to the depth of the observed U K 0 37 minimum at $50 cm (Table 2).

…

Figures - uploaded by Alan C. Mix

Content may be subject to copyright.

Content uploaded by Alan C. Mix

Content may be subject to copyright.

Alkenone paleothermometry: Biological lessons from marine

sediment records oﬀ western South America

Fredrick G. Prahl

, Alan C. Mix, Margaret A. Sparrow

College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331-5503, USA

Received 14 March 2005; accepted in revised form 24 August 2005

Abstract

An empirical global core-top calibration that relates the alkenone unsaturation index UK0

37 to mean annual SST (maSST) is statistically

the same as that deﬁned for a subarctic Paciﬁc strain of Emiliania huxleyi (CCMP1742) grown exponentially in batch culture under iso-

thermal conditions. Although both equations have been applied widely for paleoSST reconstruction, uncertainty still stems from two key

ecological factors: variability in the details of biosynthesis among genetically distinct alkenone-producing strains, and impacts of non-

thermal physiological growth factors on UK0

37. New batch culture experiments with CCMP1742 here reveal that UK0

37 diverge systematically

from the core-top calibration in response to nutrient depletion and light deprivation, two physiological stresses experienced by phyto-

plankton populations in the real ocean. Other aspects of alkenone/alkenoate composition also respond to these stresses and may serve

as signatures of such eﬀects, providing an opportunity to detect, understand, and potentially correct for such impacts on the geologic

record. A test case documents that sediments from the Southeast Paciﬁc display the alkenone/alkenoate compositional signature char-

acteristic of cells physiologically stressed by light deprivation. Such an observation could be explained if marine snow provided a major

vector of sedimentation for these biomarkers. Late Pleistocene UK0

37 records in the Southeast Paciﬁc yield plausible paleotemperature his-

tories of ice-age cooling, but ice-age alkenone/alkenoate signatures fall outside the range of modern calibration samples of similar UK0

37.

They better match core-top samples deposited beneath waters characterized by much cooler maSST, suggesting key features of ice-age

ecology for alkenone-producing haptophytes were diﬀerent from today, and that the UK0

37 index taken alone may misgauge the total range

of ice-age cooling at these locations. Analysis of the full spectrum of alkenone/alkenoate compositions preserved in sediments opens up a

new opportunity that may improve the accuracy of paleotemperature estimates based on simple UK0

37 analysis and help resolve longstand-

ing disagreements between various paleotemperature reconstruction methods.

1. Introduction

Since publication of the seminal work by Brassell et al.

(1986), the alkenone unsaturation index UK0

37 has often been

employed as a paleothermometric tool for Quaternary re-

search. Establishment of a simple, statistically well-be-

haved global calibration (Muller et al., 1998) justiﬁably

supports use of UK0

37 as a proxy for reconstruction of mean

annual sea-surface temperature (maSST). Nonetheless,

paleoreconstructions of maSST change must still be scruti-

nized, as even small systematic biases can change inferences

about mechanisms of climate change.

The relationship between UK0

37 values and maSST is

empirical (Herbert et al., 1998; Muller et al., 1998), not

causative (e.g., Prahl et al., 2000). Sediment trap time series

show that alkenone export from surface waters is seasonal

in many ocean locations (see Prahl et al., 2000 and refer-

ences therein). Also, the euphotic zone in many ocean loca-

tions is often thermally structured and modeled as a two-

layered biological system (Coale and Bruland, 1987; Small

et al., 1987). Thus, at least some portion of alkenone export

could derive from deeper in the euphotic zone where tem-

peratures are signiﬁcantly colder than that at the sea sur-

face (e.g., Bentaleb et al., 1999; Prahl et al., 1993, 2001;

Ternois et al., 1997). Furthermore, work with laboratory

cultures shows UK0

37 values set by cells are species, even

strain, dependent (e.g., Conte et al., 1998; Yamamoto

doi:10.1016/j.gca.2005.08.023

Corresponding author. Fax: +1 541 737 2064.

E-mail address: fprahl@coas.oregonstate.edu (F.G. Prahl).

www.elsevier.com/locate/gca

Geochimica et Cosmochimica Acta 70 (2006) 101–117

et al., 2000), and are sensitive not only to growth tempera-

ture but also to physiological factors such as nutrient and

light availability (e.g., Epstein et al., 1998, 2001). These

observations contribute to uncertainty in absolute temper-

ature estimates derived from UK0

37 using a simple linear cal-

ibration curve, suggesting that the observed ±1.5 C

standard deviation of residuals in the global core-top cali-

bration (compared to typical ±0.5 C or better analytical

reproducibility of single samples) reﬂects systematic eco-

logical eﬀects, rather than just random error. By better

understanding how physiological factors act to shape the

biochemical ﬁngerprint, an opportunity is at hand to im-

prove use of alkenones as paleotemperature proxies.

Here, we document concordance between maSST and

UK0

37 measured in modern sea-ﬂoor sediments accumulating

along the continental margin, in deep basins, and on ridges

oﬀ western South America, that is consistent with the pre-

vious global calibration (Muller et al., 1998). We also char-

acterize the C

, and C

alkenone and related C

alkenoate ÔﬁngerprintsÕpreserved in the surface sediments,

and with depth in three piston cores from sites representa-

tive of the subpolar transition zone oﬀ southern Chile, the

eastern boundary current oﬀ southern Peru, and the equa-

torial ‘‘cold-tongue’’ upwelling zone extending oﬀshore

from South America. Each of these piston cores yields per-

spective on glacial–interglacial changes in oceanographic

conditions over the past 150 Kyrs. Examination of our

collective sediment data set in context with results from

laboratory-controlled batch culture experiments shows that

physiological factors likely have a signiﬁcant impact on the

biomarker signature (and the implied paleotemperature

estimates) preserved in sediments from this region, and that

key features of the ecology of alkenone producers were

potentially quite diﬀerent under ice-age conditions than

at present, with no apparent modern analog in the region.

2. Materials and methods

2.1. Sediment samples

Near-surface sediment was sampled from 33 multicores

collected in 1997 along a north-to-south transect from 13

to 51S oﬀ the Chile–Peru margin, and from four gravity

and three piston cores collected along an east-to-west tran-

sect from 90 to 140W in the eastern equatorial Paciﬁc.

All multicores recovered pristine sediment–water interfac-

es, including soupy ‘‘ﬂuﬀ’’ layers. Upon recovery at sea,

the multicores were stored vertically under refrigerated

conditions (4 C) in the Oregon State University (OSU)

Core Repository (http://corelab-www.coas.oregonstate.

edu/), and after several months of storage were split and

described. These were then sealed in D-tubes and stored

horizontally under refrigeration prior to sampling. The

samples were taken in 1 cm thick sections from within

the bioturbated mixed-layer, as close to the core top as pos-

sible (Table 1). Samples taken at 5–6 cm depth in 13 of the

multicores reﬂect the fact that sediments settled during ver-

tical storage, and measurements of sample depths are in-

dexed to the top of the liner; that is, the true depth of

these samples below the original sediment–water interface

is 2–3 cm. Given the relatively high sedimentation rates

in the region, in some cases of over 100 cm/kyr (Mix

et al., 2003), all of these near-surface sediments are consid-

ered representative of modern conditions, averaged over

the past few hundred to perhaps a thousand years because

of bioturbation eﬀects. Sediments were also sampled with

depth in three piston cores: Y69-71P, Y71-6-12P, and

RR9702A-11PC (Tables 1 and 2), which represent varia-

tions in composition over the last glacial–interglacial cycle.

2.2. Culture experiments

We selected a strain of Emiliania huxleyi (CCMP1742,

aka 55a) (obtained from the Pravasoli-Guillard National

Center for Culture of Marine Phytoplankton; http://

ccmp.bigelow.org/) for culture experiments because its

established UK0

37-growth temperature calibration (Prahl

et al., 1988) provides an excellent match with the empirical

core-top calibration for UK0

37 versus maSST (Muller et al.,

1998).

Three manipulative batch culture experiments were con-

ducted in controlled temperature baths using either f/2 or f/

20 as the nutrient media (see CCMP website for recipes)

and cool-white ﬂuorescent lighting for illumination. Exper-

iments were designed to allow daily sampling over a period

extending from healthy, exponential growth into physio-

logically stressed growth. Inoculations used to initiate

experiments were made using cell stocks acclimated previ-

ously for growth at the experimental temperature. In all

experiments, cells were grown using a 12 h light/dark cycle.

Light intensity during the day was held constant at 50–

60 lEin/(m

s). Each day, approximately 1 h after lights

turned on automatically in the culture room, ﬂasks were

swirled gently to assure homogenization and immediately

subsampled using a sterile pipette in order to monitor

changes in alkenone/alkenoate content and composition,

as well as cell density and nutrient concentration.

Cell density in the culture samples was determined

microscopically using a hemocytometer. Cells for biomark-

er analysis were collected by ﬁltration on pre-combusted

(450 C, 8 h) glass ﬁber ﬁlters (GF/F: 25 mm, Whatman),

which were wrapped in aluminum foil and stored in an

ultrafreezer (80 C) until needed. The ﬁltrate from each

subsample was collected in an acid-washed, plastic polyvial

(30 mL) and stored frozen until analyzed for nitrate and

phosphate by standard autoanalyzer techniques (Strick-

land and Parsons, 1972).

In one paired experiment, inoculated culture ﬂasks

(500 mL) containing f/20 media (100 lM nitrate, 4 lM

phosphate) were maintained isothermally at 15 C and al-

lowed to grow either exponentially until nutrient-depleted

or until cell division was forced to stop by exclusion of

all light. More details about this experimental design are

given elsewhere (Prahl et al., 2003). This experiment evalu-

102 F.G. Prahl et al. 70 (2006) 101–117

Table 1

Alkenone and other ancillary data for surface sediments collected along the continental margin oﬀ the west coast of South America

Code Lat

(S)

Long

(W)

Water depth

(m)

sumSST

(C)

winSST

(C)

maSST

(C)

Core depth

(cm)

(wt%)

%CaCO

(wt%)

RAlk

(lg/gOC)

UK0

37 UK0

37 T

(C)

%K37

%K38

ME/K37

EE/ME

RR9702A-01MC3 50.7 77.0 3964 9.7 7.2 8.7 5.5 0.29 5.6 303 0.370 9.7 52 43 0.253 0.666

RR9702A-08MC2 46.4 76.7 3014 11.9 7.4 10.3 5.5 1.0 5.4 360 0.420 11.2 51 44 0.226 0.729

RR9702A-12MC2 43.4 76.3 3523 13.5 9.2 11.8 5.5 1.3 1.8 348 0.467 12.6 50 44 0.213 0.711

RR9702A-27MC2 40.5 75.9 3850 15.4 10.5 13.2 5.5 1.2 0.4 399 0.470 12.7 50 44 0.185 0.700

RR9702A-44MC2 35.8 73.0 172 16.3 10.7 13.6 5.5 2.5 3.3 294 0.488 13.2 53 43 0.183 0.653

RR9702A-50MC1 23.6 73.6 3396 20.4 16.0 17.8 5.5 0.27 63.4 169 0.640 17.7 53 42 0.118 0.780

RR9702A-60MC1 20.9 81.5 2480 21.4 18.0 19.4 5.5 0.2 90.5 78 0.707 19.7

ggg g

RR9702A-34MC1 36.5 73.4 133 16.7 11.1 13.9 5.5 3.5 1.2 369 0.479 12.9 51 43 0.141 0.725

RR9702A-48MC3 32.6 73.7 3920 17.4 13.1 15.1 5.5 0.56 23.3 387 0.564 15.4 52 43 0.142 0.794

RR9702A-54MC1 21.4 81.4 1323 21.1 17.9 19.2 5.5 0.2 95.6 239 0.749 20.9

ggg g

RR9702A-77MC1 16.1 77.0 2588 22.8 16.6 19.4 5.5 1.6 5.7 679 0.677 18.8 53 42 0.078 0.907

RR9702A-80MC4 13.5 76.9 448 21.7 14.7 18.0 5.5 4.4 0.3 657 0.699 19.4 55 42 0.060 0.779

RR9702A-31MC1 37.7 75.4 3946 17.0 11.7 14.3 5.5 1.3 0.4 505 0.495 13.4 49 45 0.160 0.751

RR9702A-06MC2 46.9 76.6 3298 11.9 7.4 10.3 0.5 0.89 2.7 376 0.402 10.7 47 48 0.239 0.684

RR9702A-14MC4 43.5 76.5 3471 13.5 9.2 11.8 0.5 1.1 1.6 354 0.453 12.2 47 47 0.202 0.711

RR9702A-20MC7 40.0 74.5 1055 15.9 10.1 13.2 1.5 1.0 4.5 452 0.479 12.9 47 48 0.186 0.710

RR9702A-04MC5 47.0 76.6 3354 11.9 7.4 10.3 2.5 0.92 2.7 399 0.410 10.9 46 48 0.241 0.717

RR9702A-10MC2 46.3 76.5 2879 11.9 7.4 10.3 2.0 0.85 6.5 398 0.430 11.5 47 47 0.232 0.696

RR9702A-22MC3 40.0 74.1 430 15.2 9.6 12.9 0.5 1.4 2.0 479 0.487 13.2 47 48 0.189 0.667

RR9702A-25MC2 39.9 75.9 4087 16.1 10.9 13.6 0.5 1.1 0.5 336 0.489 13.2 46 47 0.182 0.691

RR9702A-39MC3 36.2 73.6 510 16.7 11.1 13.9 2.0 2.3 1.0 392 0.507 13.8 49 46 0.174 0.697

RR9702A-42MC1 36.2 73.7 1028 16.7 11.1 13.9 0.5 2.4 4.7 441 0.481 13.0 50 46 0.186 0.668

RR9702A-46MC1 33.3 73.5 3852 16.9 12.6 14.7 2.0 0.94 7.8 397 0.533 14.5 49 46 0.170 0.740

RR9702A-52MC1 23.2 73.3 3418 20.4 16.0 17.8 2.5 0.33 54.2 167 0.633 17.5 51 45 0.125 0.729

RR9702A-62MC1 18.1 79.0 2937 22.5 17.6 19.7 1.5 0.92 75.7 68 0.698 19.4 53 44 0.106 0.889

RR9702A-72MC1 16.5 76.2 3782 22.8 16.6 19.4 2.0 0.84 0.5 349 0.696 19.3 52 44 0.083 0.869

RR9702A-74MC1 16.2 76.2 3476 22.8 16.6 19.4 2.0 1.2 0.1 478 0.675 18.7 53 44 0.088 0.982

RR9702A-66MC1 16.1 77.1 2575 22.9 16.9 19.6 2.0 1.6 6.4 427 0.677 18.8 53 44 0.095 0.819

RR9702A-68MC1 16.0 76.4 3228 22.8 16.6 19.4 1.5 1.6 0.0 513 0.685 19.0 51 45 0.077 0.886

RR9702A-70MC1 16.7 76.0 4124 22.8 16.6 19.4 3.0 0.67 0.0 270 0.661 18.3 53 43 0.098 0.894

RR9702A-64MC1 17.0 78.1 2930 22.9 17.4 19.8 2.0 0.44 55.9 227 0.701 19.5 52 44 0.086 0.982

RR9702A-83MC2 13.2 77.3 1419 22.3 15.9 18.9 2.5 8.9 0.0 571 0.748 20.9 52 44 0.070 0.890

RR9702A-82MC4 13.7 76.7 264 21.7 14.7 18.0 2.0 15.0 4.4 618 0.736 20.5 52 43 0.062 1.035

VNTR01-19PC (7.9) 90.5 3448 27.6 26.8 27.4 3.5

hh h

0.911 25.6

ggg g

VNTR01-9PC 3.0 110.5 3860 26.0 22.1 24.1 2.5

hh h

0.876 24.6

ggg g

VNTR01-12GC 3.0 95.1 3535 26.3 20.3 23.2 3.5

hh h

0.802 22.4

ggg g

VNTR01-10GC 4.5 102.0 3405 25.4 22.2 24.3 3.5

hh h

0.867 24.3

ggg g

VNTR01-8PC 0.0 110.5 3791 25.7 21.8 23.4 0.5

hh h

0.868 24.4

ggg g

VNTR01-21GC (9.6) 94.6 3710 27.8 27.2 27.7 3.5

hh h

0.908 25.6

ggg g

MANOP-14GC (1.0) 140.0 3900 27.0 25.3 26.1 0.5 0.36 80 92 0.957 27.0

ggg g

Numbers in parentheses are north latitude, i.e. N.

sumSST, winSST, mast, summer, winter, and mean annual SST, respectively.

Depths measured from top of core liner. Cores settled during transport, so all samples actually taken within the top 3 cm of original sediment–water interface.

Total C

, and C

alkenone concentration.

Temperature estimated from calibration equation UK0

37 ¼0:034Tþ0:039 (Prahl et al., 1988)

Alkenone/alkenoate compositional properties—see text for deﬁnition.

Chromatograms contained interfering peaks which prohibited reliable quantitation of these compositional properties.

Not determined.

Alkenone paleothermometry oﬀ western South America 103

Table 2

Compositional data for alkenone/alkenoate signatures measured with depth in piston cores collected along the continental margin oﬀ the west coast of South America

RR9702A-11PC Y71-6-12P Y69-71P

Latitude (46.317S); Longitude (76.538W); Water

depth (2825 m)

Latitude (16.443S); Longitude (77.563W); Water depth (2734 m) Latitude (0.083N); Longitude (86.482W); Water depth (2740 m)

Depth (cm) Age (Kyr) UK0

37 ME/K37 EE/ME Depth (cm) Age (Kyr) UK0

37 %K37 %K38 ME/K37 EE/ME Depth (cm) Age (Kyr) UK0

37 %K37 %K38 ME/K37 EE/ME

1.5 0.6 0.463 0.229 0.673 12.5 4.2 0.712 52 44 0.140 0.868 7.3 1.1 0.857 55 42 0.069 0.921

11.5 4.8 0.425 0.269 0.519 17.5 6.5 0.710 54 42 0.102 0.780 10.5 1.5 0.859 54 42 0.066 0.967

21.5 9.0 0.405 0.268 0.448 27.0 10.8 0.683 54 42 0.101 0.746 18.5 2.6 0.854 55 42 0.057 1.077

31.5 13.1 0.322 0.308 0.522 32.0 12.7 0.696 52 43 0.095 0.848 40.5 5.6 0.831 55 41 0.056 1.088

41.5 17.3 0.290 0.321 0.475 45.5 16.8 0.646 53 43 0.113 0.699 50.0 6.9 0.816 54 43 0.055 1.198

51.5 21.5 0.268 0.310 0.504 49.5 18.3 0.616 53 43 0.121 0.645 62.0 8.5 0.797 55 42 0.064 1.026

61.5 0.262 a a 55.5 20.4 0.625 54 42 0.126 0.621 80.5 11.0 0.787 55 42 0.058 0.983

71.5 0.308 a a 60.5 21.9 0.612 53 43 0.131 0.644 100.5 12.8 0.756 54 42 0.056 1.128

81.5 0.365 a a 65.5 23.5 0.625 54 42 0.131 0.658 120.5 14.5 0.767 55 41 0.055 0.928

91.5 0.348 a a 65.5 23.5 0.613 54 42 0.132 0.664 139.5 16.2 0.764 55 41 0.056 0.852

101.5 0.349 a a 70.5 26.0 0.626 54 43 0.124 0.762 149.0 17.0 0.771 55 41 0.054 0.905

111.5 0.347 a a 75.5 28.3 0.624 54 42 0.127 0.707 180.0 19.2 0.777 56 41 0.059 0.719

121.5 0.363 a a 80.5 31.5 0.622 54 42 0.138 0.599 201.0 20.5 0.766 56 41 0.061 0.887

131.5 0.355 a a 85.5 37.8 0.627 54 42 0.138 0.596 218.5 21.5 0.777 57 39 0.070 0.565

141.5 0.349 a a 90.5 44.6 0.639 54 42 0.122 0.596 240.5 22.8 0.779 56 41 0.063 0.570

151.5 0.376 a a 96.5 50.0 0.662 54 42 0.122 0.532 250.0 23.5 0.790 55 41 0.074 0.528

161.5 0.367 a a 100.0 52.9 0.659 55 42 0.117 0.539 260.5 24.2 0.801 55 41 0.062 0.628

171.5 0.367 a a 104.5 56.5 0.683 54 43 0.111 0.533 279.5 26.4 0.814 55 41 0.067 0.651

181.5 0.330 a a 113.0 60.4 0.678 54 42 0.113 0.521 300.5 29.1 0.827 54 43 0.094 0.562

191.5 0.342 a a 117.5 62.1 0.678 53 43 0.117 0.476 322.5 31.8 0.809 54 42 0.069 0.668

200.5 0.337 a a 119.5 62.7 0.674 54 42 0.117 0.513 336.5 33.6 0.820 55 41 0.078 0.514

210.5 0.323 a a 125.5 66.0 0.659 54 43 0.126 0.468 339.0 33.9 0.797 57 41 0.082 0.483

220.5 0.319 a a 130.5 70.9 0.673 54 42 0.132 0.436 348.5 35.1 0.813 55 42 0.078 0.509

230.5 0.335 a a 135.5 75.8 0.712 54 43 0.125 0.448 357.5 36.2 0.808 56 41 0.074 0.530

250.5 0.334 a a 140.5 79.9 0.735 54 42 0.127 0.429 361.0 36.6 0.819 55 42 0.070 0.574

270.1 0.307 a a 145.0 85.0 0.738 53 43 0.123 0.429 377.5 38.7 0.797 56 41 0.077 0.556

290.5 0.300 a a 150.0 91.1 0.695 54 42 0.117 0.440 381.0 39.1 0.785 56 41 0.074 0.568

310.5 0.265 0.320 a 155.0 97.1 0.706 55 42 0.118 0.450 400.5 41.6 0.809 55 41 0.069 0.607

320.5 0.264 0.331 a 160.0 103.1 0.689 54 42 0.121 0.466 420.0 44.0 0.818 56 41 0.074 0.569

330.5 0.259 0.304 a 165.0 106.6 0.712 54 42 0.115 0.424 449.0 47.6 0.833 56 41 0.073 0.531

350.5 0.350 a a 170.0 108.9 0.681 54 42 0.107 0.485 459.0 48.9 0.831 56 41 0.077 0.484

370.5 0.325 a a 175.0 111.7 0.714 53 43 0.104 0.440 482.0 51.8 0.846 56 41 0.076 0.437

391.0 0.347 a a 180.0 115.7 0.734 53 43 0.091 0.485 501.0 54.1 0.819 56 41 0.078 0.565

410.5 0.350 a a 186.5 122.8 0.738 52 43 0.086 0.503 519.5 56.4 0.840 56 41 0.093 0.406

430.5 0.342 a a 194.5 133.8 0.701 51 44 0.100 0.491 538.5 58.9 0.816 56 41 0.077 0.467

450.5 0.332 a a 201.0 138.6 0.689 51 45 0.113 0.541 548.5 60.8 0.799 56 41 0.080 0.428

470.5 0.307 a a 206.0 141.9 0.704 53 43 0.092 0.505 559.5 62.9 0.796 56 41 0.077 0.476

500.5 0.307 a a 210.0 144.8 0.680 52 43 0.102 0.465 577.5 66.3 0.809 55 41 0.091 0.405

520.5 0.326 a a 216.0 149.1 0.683 52 44 0.104 0.520 598.5 70.1 0.806 55 41 0.110 0.346

530.5 0.302 a a 221.0 152.5 0.657 52 44 0.107 0.466 618.0 73.5 0.812 56 41 0.089 0.396

550.5 0.296 a a 226.0 155.7 0.682 52 44 0.092 0.530 637.5 77.0 0.844 55 42 0.087 0.346

570.5 0.302 0.326 a 231.0 158.8 0.660 52 43 0.089 0.489 648.5 79.0 0.869 55 42 0.071 0.413

610.5 0.284 0.311 a 236.0 162.0 0.676 52 44 0.092 0.550 662.5 81.5 0.859 55 42 0.076 0.401

680.5 0.284 a a 241.0 165.9 0.645 53 43 0.097 0.529 678.5 84.8 0.862 55 42 0.086 0.422

700.5 0.306 a a 246.0 170.1 0.673 53 44 0.094 0.560 698.5 89.7 0.864 55 42 0.087 0.397

104 F.G. Prahl et al. 70 (2006) 101–117

ated how nutrient depletion, an environmental stress that

would likely occur during termination of a bloom event,

might aﬀect the alkenone/alkenoate composition transmit-

ted to the sediment record. In a second experiment, a pair

of culture ﬂasks was ﬁlled with f/2 media (900 lM nitrate,

40 lM phosphate), inoculated, and allowed to grow expo-

nentially for several days at 20 C. Growth temperature

was then abruptly decreased to 15 C. When the 5 C tem-

perature decrease was imposed, one ﬂask was covered for a

continuous 5-day period with aluminum foil to exclude all

light while the other, serving as a control, was left other-

wise unchanged. This experiment was designed to evaluate

the extent to which the alkenone/alkenoate composition of

viable cells inadvertently packaged into settling materials

such as marine snow could be modiﬁed physiologically

during transit to the sediment record as viable cells experi-

ence temperature drop and complete darkness during set-

tling. With the exception of higher nutrient

concentrations and a temperature shift, this second exper-

iment was conducted in a manner identical with the ﬁrst

set.

The cultures grown using f/2 media have initial nutrient

concentrations signiﬁcantly higher than typical oceanic val-

ues. We chose this condition, however, to obtain suﬃcient

material to monitor in detail changes in alkenone/alkeno-

ate content through the diﬀerent stages of this experiment

from exponential to stressed growth under conditions that

approximate in key ways known real-world eﬀects. Such

physiological stress eﬀects cannot be simulated using che-

mostats which, by design, focus on nutrient-limited, stea-

dy-state, exponential growth conditions (e.g., Popp et al.,

1998). Given our prior ﬁndings that the direct eﬀects of

high nutrient concentration are negligible on UK0

37 (Prahl

et al., 1988), we believe the use in our culture experiments

of a media with high initial nutrients is most appropriate.

2.3. Lipid biomarker analyses

Each sample, either as wet sediment (1–3 g dry) or as al-

gal cells (P10

) on GF/F, was put into a glass centrifuge

tube (50 mL), spiked with a recovery standard (nonadec-

an-2-one), and ultrasonically extracted using a 3:1 mixture

of methylene chloride and methanol (20 mL, 3·). Extracts

were combined in a separatory funnel and, after addition of

pre-extracted water (10 mL), partitioned into hexane

(20 mL, 3·). The hexane solution was then dried over gran-

ular, anhydrous Na

(24 h) and the solvent was subse-

quently removed by rotary evaporation under gentle

vacuum at room temperature.

From the resultant lipid residue, a fraction enriched in

–C

alkenones and C

alkenoates (Brassell, 1993)

was isolated by silica gel column chromatography. The

concentration and composition of the alkenone/alkenoate

mixtures contained in these fractions were measured quan-

titatively using on-column, capillary gas chromatography

with ﬂame ionization detection. See Prahl et al. (1989) for

further procedural details. All reported concentration data

720.5 0.305 a a 251.0 174.1 0.640 53 43 0.105 0.508 718.5 95.5 0.874 55 42 0.076 0.436

740.5 0.313 a a 737.5 100.6 0.866 55 43 0.074 0.440

760.5 0.321 a a 758.5 105.6 0.850 54 42 0.065 0.489

783.0 111.9 0.814 56 41 0.065 0.502

800.5 116.7 0.874 56 42 0.061 0.396

818.5 121.5 0.899 55 42 0.053 0.498

838.5 125.4 0.905 54 42 0.039 0.853

858.5 129.2 0.872 54 43 0.052 0.772

878.5 132.9 0.847 54 43 0.064 0.642

898.5 136.6 0.840 54 43 0.063 0.604

919.5 140.4 0.855 54 43 0.052 0.648

938.5 143.9 0.857 54 43 0.046 0.650

958.5 147.4 0.824 54 43 0.052 0.654

980.5 151.1 0.827 54 43 0.049 0.751

1002.5 155.0 0.843 54 43 0.046 0.819

a, Gas chromatographic interference prevented reliable quantitation.

Alkenone paleothermometry oﬀ western South America 105

have been corrected for internal standard recovery, which

was typically 80–90%. The routine precision on biomarker

concentration and relative compositional properties and

the unsaturation index UK0

37 was typically better than ±5–

10% and ±0.5–1%, respectively.

2.4. Elemental carbon analyses

Total carbon (TC) and total organic carbon (OC) were

measured in all multicore samples by a high-temperature

combustion method using a Carlo Erba NA-1500 Elemen-

tal Analyzer. Prior to elemental analysis, subsamples of

sediment slotted for OC determination were treated with

an excess quantity of a volatile acid (either concentrated

HCl or H

) to remove inorganic carbon (Hedges and

Stern, 1984; Verardo et al., 1990). Calcium carbonate con-

tent of each sample was then estimated as a weight percent-

age of dry sediment (%CaCO

), assuming all detectable

inorganic carbon, calculated from the diﬀerence between

TC and OC, exists in that mineralogical form.

2.5. Stable isotope analyses and chronostratigraphy

Stratigraphic and chronologic control in the sediment

cores is provided by oxygen isotope analysis of the benthic

foraminfera, Uvigerina peregrina and/or Planulina wueller-

storﬁ, using Finnigan MAT-251 and MAT-252 mass spec-

trometers at Oregon State University. Acid digestion of

carbonate shells was done using an Autoprep Systems com-

mon acid batch system online with the MAT-251, or using

a Finnigan ‘‘Kiel-III’’ separate acid bath device online with

the MAT-252. Calibration to the widely accepted Pee Dee

Belemnite (PDB) scale was accomplished by analysis of

two standards from the National Institute of Standards

and Technology (NIST): NBS-19 and NBS-20. The preci-

sion of replicate d

O analyses for both standards was

±0.06&, with no signiﬁcant oﬀset between results on the

two diﬀerent instruments.

Age models in the sediment cores are based on correla-

tion of down-core d

O variations with global reference re-

cords (e.g., Martinson et al., 1987; Pisias and Mix, 1997)

and linear interpolation between discrete time horizons.

Additional chronologic control comes from published

radiocarbon dates that have been converted to calendar

ages (Clark et al., 2004; Feldberg and Mix, 2003).

3. Results

3.1. Sediment analyses

Near-surface sediment samples from 33 multicores, four

gravity cores, and three piston cores (Fig. 1) were analyzed

for both the content and composition of C

, and C

alkenones and C

alkenoates (Table 1). Both types of com-

pounds, biomarkers unique to a small subset of haptophyte

algae (Green and Leadbetter, 1994), were readily detected

constituents of the solvent extractable lipid fraction isolat-

ed from all samples. The same compositional information

was also obtained from analysis of sediments sampled

stratigraphically in three piston cores (see stars in Fig. 1).

3.2. Alkenone content of surface sediments

Total alkenone concentrations (RAlk), deﬁned by the

sum of all di- and tri-unsaturated C

methyl, C

methyl

and ethyl, and C

ethyl ketone constituents (Brassell,

1993), spanned almost two orders of magnitude in the sam-

ple set, ranging from 0.1 to 92 lg/g dry sediment and aver-

aging 8.0 lg/g. RAlk normalized to total organic carbon

(OC) also spanned a wide range, from 68 to 679 lg/gOC.

RAlk normalized to dry sediment weight covaries positively

with the weight percent OC content of sediments (%OC)

with a high degree of correlation (r= +0.993). Although

this statistical ﬁnding supports the interpretation that the

OC content of these sediments is predominantly of marine

origin, such inference cannot be made unequivocally from

simple correlation alone.

3.3. Calcium carbonate content of surface sediments

%CaCO

estimates varied widely in the set of surface sed-

iments examined, from undetectable to P95% (Table 1).

Since the coccolithophorid E. huxleyi is considered a domi-

nant source of alkenones in the world ocean today (Brassell,

1993), signiﬁcant correlation between RAlk (normalized to

dry sediment) and %CaCO

could be expected. These prop-

erties do covary positively, but the correlation is weak

(r= +0.26). Additional CaCO

contribution from sources

such as foraminiferal tests as well as fundamental diﬀerences

in organic matter decomposition and CaCO

dissolution

mechanisms probably account for this weak correlation.

3.4. Alkenone unsaturation pattern in surface sediments

UK0

37 values measured in the set of surface sediments

spanned 60% of the 0 to 1 range possible for the index

(Table 1). The correlation of UK0

37 with SST in overlying sur-

face waters (Table 1) is high regardless of whether World

Ocean Atlas 2001 data (http://www.awi-bremerhaven.de/

GEO/ODV/data/WOA01/) are viewed as summer (Janu-

ary–March: r= +0.97, n= 40), winter (July–September:

r= +0.96) or mean annual (r= +0.98) values.

All UK0

37 values were converted into algal growth temper-

ature (gT) estimates using the established laboratory cali-

bration equation for the benchmark strain of E. huxleyi

CCMP1742 (i.e., UK0

37 ¼0:034Tþ0:039: Prahl et al.,

1988). gT estimates ranged from 9.7 to 27 C(Table 1),

with values within the seasonal range of SST observed

for waters overlying each coring site in all but two cases

(VNTR01-19PC and 21GC) and generally approximating

maSST. A scatter plot (Fig. 1) shows ﬁndings for surface

sediments from our study region comply well with the

UK0

37–maSST calibration obtained by statistical analysis of

the global data set (Muller et al., 1998).

106 F.G. Prahl et al. 70 (2006) 101–117

3.5. Alkenone/alkenoate signature in surface sediments

Four indices in addition to UK0

37 describe the overall alke-

none/alkenoate signature (Brassell, 1993) preserved in the

surface sediments and its geographic variation throughout

the study region (Table 1). Two indices (%K37, %K38)

quantify the distribution of total C

(or C

) alkenones rel-

ative to alkenones of all three possible carbon chain

lengths, i.e., C

, and C

. A third (ME/K37) quanti-

ﬁes the proportion of total C

methyl ester (ME) concen-

tration relative to total C

alkenone (K37) concentration.

And, a fourth (EE/ME) quantiﬁes the distribution of the

alkenoate in its ethyl ester (EE) relative to its methyl

ester (ME) form.

components are the most abundant alkenones in all

sediments, comprising 46–53% of the total. C

compo-

nents are next most abundant (43–48%), and C

compo-

nents are relatively rare (3–7%). Methyl esters of the C

alkenoic acid are between 5 and 25% as abundant as

alkenones. The relative proportion of the C

alkenoic

acid in its ethyl (EE) and methyl (ME) ester forms ranges

from 0.7–1.0.

Variations in %K37, ME/K37, and EE/ME in modern

(core-top) sediments all follow systematic trends when

plotted versus UK0

37. %K37 values increase gradually in di-

rect proportion to the alkenone unsaturation index

(Fig. 2A) while ME/K37 values decrease (Fig. 2B).

Although signiﬁcant scatter masks any clear trend below

aUK0

37 of 0.6, an increase in EE/ME values is apparent for

higher values of the alkenone unsaturation index (Fig. 2C).

3.6. Down-core stratigraphic biomarker distributions

Alkenone and alkenoate composition was also charac-

terized stratigraphically with depth in cores collected at

three speciﬁc locations along the continental margin oﬀ

maSST (oC)

0.25

0.50

0.75

1.00

multicore tops

Prahl et al. (1988)

Muller et al. (1998)

U K’

5 1015202530

110 100 90 80 70 60

-50

-40

-30

-20

-10

Latitude (+ oN, - oS)

Lon

itude (oW)

Fig. 1. Map on the left shows locations along the continental margin oﬀ the west coast of South America where sediment samples were obtained for

detailed alkenone/alkenoate analysis. Solid symbols depict sites where surface sediments from core-tops were obtained from 33 multicores, three (of four)

gravity cores, and three piston cores (Table 1); open stars depict three piston core sites where sediments were obtained stratigraphically with depth (Table

2). Dashed lines represent isotherms for mean annual sea-surface temperature (maSST) (Ocean Climate Laboratory, 2003). Major oceanographic currents

and features in the study region (Strub et al., 1998) are also identiﬁed (bold arrows) to provide context for discussion. The graph on the right shows a

scatter plot for UK0

37 values measured in surface sediments versus maSST in waters overlying each multicore site (all data from Table 1). Dotted line depicts

the global calibration equation ðUK0

37 ¼0:033Tþ0:044Þestablished for surface sediment collected from 60Nto60S throughout all ocean basins (Muller

et al., 1998). Dashed line depicts the UK0

37 versus temperature relationship established for a single strain of E. huxleyi grown in laboratory-controlled batch

cultures (UK0

37 ¼0:034Tþ0:039: Prahl et al., 1988).

Alkenone paleothermometry oﬀ western South America 107

%K37

ME/K37

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35 Multicores

Y69-71P

Y71-6-12P

RR9702A-11P

Ehux: 15oC, nutrient stress

Ehux: 15oC, dark stress

Ehux: 20 to 15oC shift, dark stress

Ehux: 20 to 15oC shift, control

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

EE/ME

0.25

0.50

0.75

1.00

1.25

UK'

Fig. 2. Scatter plot of values for (A) %K37, (B) ME/K37, and (C) EE/ME versus UK0

37 measured in surface sediments (Table 1) and three piston cores (Y69-

71P, Y71-6-12P, and RR9702A-11PC: Table 2) from the continental margin oﬀ the west coast of South America (see text for deﬁnition of each

compositional property). The black lines represent trends deﬁned by data for the same compositional properties measured in a Ônutrient stressÕ(solid) and

Ôdark stressÕ(dashed) experiment with E. huxleyi grown isothermally at 15 C under laboratory-controlled, batch culture conditions. The solid grey lines

represent trends deﬁned by time series data obtained from an experiment where Ôdark stressÕwas imposed with a simultaneous 5 C decrease in growth

temperature (i.e., 20–15 C). The dashed grey lines represent the trend for a control on the latter experiment, i.e., case where the 20–15 C temperature shift

was imposed with no dark stress. The letter E associated with each line depicts the composition of exponentially dividing cells at the start of each

experiment. The arrowhead on each line indicates the time series trajectory of compositional change caused by each stress. All culture data used to draw

the lines depicted in this ﬁgure are given in Table 3. The ﬁlled and open symbols for core data correspond to sediments deposited since and prior to the

LGM (21 Ka), respectively.

108 F.G. Prahl et al. 70 (2006) 101–117

the west coast of South America (Fig. 1). One (Y69-71P:

0.08N, 86.5W) records the depositional history of equa-

torial waters at the northern edge of our surface sediment

survey area. The second (Y71-6-12P: 16.4S, 77.6W) re-

cords the depositional history of waters underlying a sub-

tropical section of the Peru Current. And the third

(RR9702A-11PC: 46.3S, 76.5W) records depositional his-

tory in the subpolar transition zone, where waters from the

West Wind Drift bifurcate into the north-ﬂowing Peru

Current and the south-ﬂowing Cape Horn Current. In

cores Y69-71P and Y71-6-12P, approximately 150 Kyr of

deposition are represented based on d

O stratigraphy. Be-

cause of the scarcity of calcareous fossils and unavailability

of radiocarbon data, assignment of an equivalent age scale

for RR9702A-11PC is not now possible. For this core, ten-

tative assignment of climate events is made based on corre-

lation of UK0

37 variations here with the other sites.

UK0

37 was reliably determined in all intervals sampled

from the three cores. Values varied signiﬁcantly in each,

spanning a range from 0.756 to 0.905, 0.612 to 0.738,

and 0.259 to 0.463, respectively (Table 2). Down-core vari-

ations generally track benthic foraminiferal records of d

available for both Y69-71P and Y71-6-12P (Figs. 3A and

B) with colder temperature estimates broadly correspond-

ing to glacial intervals (relatively high d

O values). Assign-

ment of the LGM time horizon in RR9702A-11PC is

placed within the 50–60 cm depth interval of the core based

solely on visual inspection of the UK0

37 proﬁle (Fig. 3C).

The correspondence of UK0

37 temperature estimates with

oxygen isotope stages in both Y69-71P and Y71-6-12P is

imperfect. For example, warming associated with the last

deglaciation appears to lag d

O changes at the equator

(Core Y69-71P) but leads to d

O changes in Y71-6-12P.

In contrast with the alkenone-based estimates, tempera-

tures inferred at the equator and in core Y71-6-12P based

on planktonic foraminiferal species assemblages (Feldberg

and Mix, 2003) yield higher-amplitude temperature chang-

es, and the equatorial zone appears to warm earlier during

the deglaciation than the eastern boundary current (Fig. 3).

Reliable stratigraphic measurement of the four other

compositional properties (%K37, %K38, ME/K37, and

EE/ME) analyzed in surface sediments was only possible

20 21 22 23 24 25 26

Core Depth (cm)

250

500

750

1000

3.03.54.04.55.05.5

UK' -T

δ18O

16 17 18 19 20 21

100

150

200

250

δ18O

δ18O3.54.04.55.0

T (oC)T (oC) T (oC)

250

500

750

Y69-71P Y71-6-12P RR9702A-11PC

789101112136

ABC

Fig. 3. Proﬁles for estimates of UK0

37-based sea-surface temperature (SST) with depth in three piston cores from the study area (Fig. 1): (A) Y69-71P; (B)

Y71-6-12P; and (C) RR9702A-11PC. SST estimates were calculated using a standard calibration equation (UK0

37 ¼0:034Tþ0:039: Prahl et al., 1988). The

depth proﬁle for stable oxygen isotope measurements (d

O) on benthic foraminiferal tests from discrete depths in Y79-71P (Clark et al., 2004) and Y71-6-

12P (Feldberg and Mix, 2003) are also plotted in (A) and (B) for stratigraphic reference purposes.

Alkenone paleothermometry oﬀ western South America 109

in Y69-71P and Y71-6-12P. Weak biomarker signal com-

bined with the presence of interfering gas chromatograph-

ic peaks prevented their routine determination in

RR9702A-11PC in all but a few cases for ME/K37 and

EE/ME. Complete down-core records for %K37 and

ME/K37 in Y69-71P and Y71-6-12P are plotted in Figs.

2A and B, respectively. In both cores, the signatures for

these properties fall along trends deﬁned by correspond-

ing surface sediment data. GC traces were of suﬃciently

high quality in the case of only 10% of the samples ana-

lyzed from R9702A-11PC to allow conﬁdent determina-

tion of ME/K37. All acceptable data from this core fall

along the projected surface-sediment trend deﬁned for this

property (Fig. 2B).

EE/ME

0.25 0.50 0.75 1.00 1.25

Age (kiloyears)

100

125

150

175

RR9702A-11PC

Y71-6-12P

Y69-71P

Fig. 4. Stratigraphic proﬁles for EE/ME values measured in three piston cores from our study area (Y69-71P; Y71-6-12P; and RR9702A-11PC, see stars in

Fig. 1). An age model was established for the ﬁrst two cores using available depth records for d

O in benthic foraminiferal shells and radiocarbon dates on

foraminiferal carbonate in sediment samples from discrete depths (Pisias and Mix, 1997). For the third core, ages were assigned more arbitrarily, assuming a

constant rate of sedimentation and the LGM time horizon (21 Kyr) corresponds to the depth of the observed UK0

37 minimum at 50 cm (Table 2).

110 F.G. Prahl et al. 70 (2006) 101–117

In contrast to the case for %K37 and ME/K37, EE/ME

from pre-Holocene samples for Y69-71P and Y71-6-12P

deviate signiﬁcantly from the surface-sediment trend

(Fig. 2C). Closer examination shows that EE/ME varies

systematically through time. EE/ME for all depth intervals

younger than 17 Kyr in Y69-71P and 28 Kyr in Y71-6-

12P are reasonably consistent with the surface-sediment

trends, but decrease at greater ages, until returning to

near-modern values in Y69-71P at ages >120 ka (Fig. 4).

In RR9702A-11PC, GC data quality limited reliable EE/

ME determination to only the six samples examined above

the purported LGM depth at 50–60 cm. Although the

surface-sediment value is consistent with the regional trend

for modern samples, EE/ME appears anomalously low

(Fig. 2C). The shift to lower EE/ME values in RR9702A-

11PC occurs at younger age than that observed in lower

latitude cores Y71-6-12P and Y69-71P, in which EE/ME

values display gradual increase beginning approximately

at the depth of oxygen isotope stage boundary 3/2 and

extending through the LGM to recent times (Fig. 4).

A long-term preservational eﬀect provides one possible

explanation for this EE/ME deviation in older sediments.

That is, the ethyl ester (EE) may over long time spans be

selectively degraded relative to its methyl ester (ME) form.

However, EE/ME in Y69-71P returns down-core to values

within the modern range of values (Fig. 4), arguing against

accumulation of monotonic preservational eﬀects over

time.

An alternative possibility is that the assemblage of

organisms contributing to this biomarker record has chan-

ged over time, i.e., there is no modern analog for alkenone

producers present during some intervals of the last ice age.

If so, this ﬁnding would challenge the accuracy of absolute

ice-age temperatures derived from alkenones in this region.

The relatively low EE/ME signatures in the ice-age samples

are consistent with core-top samples associated with much

lower maSST than is suggested by their UK0

37 signature

(Fig. 2). Clearly, we must better understand physiological

controls on these changing biomarker compositions in or-

der to assess unequivocally the implications for paleocea-

nographic reconstruction of surface water temperatures.

3.7. Physiological controls on biomarker composition

Results from a Ônutrient stressÕand a Ôdark stressÕexper-

iment (Table 3) conﬁrm that the alkenone content of E.

huxleyi can vary widely depending upon the cellÕs physio-

logical status (Conte et al., 1998; Epstein et al., 1998). Sig-

niﬁcant increase in alkenone concentrations occurs under

well-illuminated, nutrient-limited stationary growth condi-

tions while signiﬁcant decrease occurs when cells are ex-

posed to prolonged darkness. Results also conﬁrm that

the metabolic gain and loss of cellular alkenone content

can occur compound selectively and lead to considerable

variation in UK0

37 values independent of temperature. Selec-

tive mobilization of compounds has an opposite eﬀect on

UK0

37, causing values to decrease under nutrient stress condi-

tions and to increase in the case of prolonged dark stress.

The magnitude of the impact on UK0

37 values in the algal

strain studied (CCMP1742) is equivalent to that expected

for a growth temperature change of up to ±3 C when

gauged using the empirical relationship established from

global surface sediments (Muller et al., 1998).

Tabulated results also show that non-thermal physiolog-

ical factors can signiﬁcantly impact compositional proper-

ties of the overall alkenone/alkenoate signature, not just

the C

alkenone unsaturation index UK0

37. However, impacts

of nutrient depletion and continuous darkness on the com-

positional properties %K37, ME/K37, and EE/ME are not

opposite as observed for UK0

37. At least in the paleoceano-

graphic benchmark strain of E. huxleyi, CCMP1742, both ef-

fects are unidirectional for %K37, ME/K37, and nutrient

stress has little quantitative eﬀect on ME/K37.

Fig. 2 places collective compositional data obtained

from these two batch culture experiments in quantitative

perspective with sediment data. %K37, ME/K37, and

EE/ME values measured in exponentially growing (i.e., un-

stressed) E. huxleyi cells all fall outside the compositional

trend deﬁned by sediment data, while those measured in

cells exposed to longer periods of non-thermal stress tend

toward the sediment recorded signature. The latter obser-

vation applies to all properties except ME/K37 in the case

of the nutrient stress experiment. These ﬁndings imply that

the alkenone/alkenoate signatures in these sedimentary re-

cords reﬂect the history of stressed populations, which may

have relationships to the environment diﬀerent from popu-

lations growing exponentially in continuous cultures. Of

particular interest is our observation that populations

stressed by nutrient deprivation may register paleotemper-

atures that are anomalously cold, or that populations

stressed by light deprivation (e.g., in sinking blooms) may

AÔdark stressÕexperiment analogous to the one just de-

scribed, but with a 5 C decrease in growth temperature

(which simulates settling of viable phytoplankton out of

the euphotic zone and into underlying, colder waters),

yielded eﬀects on the cellular biomarker content and the

compositional properties of %K37, ME/K37, and EE/

ME similar to those of the prior experiment (Table 3).

There was one most notable exception, however—virtually

no change in UK0

37 values occurred. In this experiment, the

decrease in growth temperature fortuitously compensated

for the eﬀect of prolonged darkness on UK0

37 values. In a

control for the latter Ôdark stressÕexperiment, growth tem-

perature was shifted but a normal daily 12 h L:D cycle was

maintained. Control results revealed that UK0

37 responds

immediately to the shift in growth temperature, as expected

from prior culture work (Prahl et al., 1988). %K37, ME/

K37, and EE/ME values also all responded and displayed

shifts in the direction expected for cells rapidly adapting

to the new growth temperature (Fig. 2).

Overall ﬁndings demonstrate that the alkenone/alkeno-

ate composition of E. huxleyi grown in culture can be

aﬀected signiﬁcantly by temperature not only when these

Alkenone paleothermometry oﬀ western South America 111

biomarkers are biosynthetically assembled but also when

they are metabolically destroyed by the cell. Assuming

the noted compensatory eﬀect of cooling on the UK0

37 ex-

pressed by cells under prolonged darkness is applicable to

populations growing in the real ocean, such eﬀects might

mitigate some apparent temperature biases depending on

the sinking rate of cells relative to the thermal gradient

experienced with depth.

4. Discussion

Paleoceanographers now employ a variety of methods

to reconstruct glacial–interglacial sea-surface temperature

(SST) change throughout diﬀerent regions of the world

ocean. The classic method for SST reconstruction, involv-

ing statistical analysis of microfossil species assemblages

derived from planktonic organisms such as foraminifera,

radiolaria, and diatoms (CLIMAP, 1981), has been ad-

vanced with a variety of statistical schemes, including

updated transfer functions that circumvent faunal no-ana-

log problems (Mix et al., 1999), several diﬀerent modern

analogous methods (Ortiz and Mix, 1997; Pﬂaumann

et al., 1996; Waelbroeck et al., 1998), neural network pat-

tern matching schemes (Malmgren and Nordlund, 1997),

and correspondence analysis methods that consider multi-

ple environmental parameters (Morey et al., 2005). In spite

of these advances, legitimate questions remain about the

accuracy of faunal temperature estimates based on statisti-

cal calibrations, in part because diﬀerent statistical meth-

ods give slightly diﬀerent temperature estimates in some

Table 3

Summary of results from batch culture experiments with E. huxleyi strain CCMP1742

Sampling day Growth condition Density

(cells/mL)

Nitrate

(lM)

Phosphate

(lM)

RAlk

(pg/cell)

UK0

37 %K37 %K38 ME/K37 EE/ME

Isothermal (15 C), ÔNutrient StressÕExperiment

115, 12:12 L/D 1.93E+05 83.1 11.5 1.2 0.579 68 31 0.070 0.627

215, 12:12 L/D 3.18E+05 74.3 10.7 1.2 0.572 68 31 0.068 0.652

315, 12:12 L/D 4.27E+05 61.6 9.8 1.5 0.559 69 30 0.067 0.617

415, 12:12 L/D 6.16E+05 44.5 8.5 1.3 0.554 69 30 0.073 0.635

515, 12:12 L/D 9.39E+05 18.3 6.6 1.5 0.535 68 31 0.065 0.592

615, 12:12 L/D 1.26E+06 0.1 4.5 1.4 0.511 67 31 0.061 0.621

715, 12:12 L/D 1.48E+06 0.2 2.8 2.0 0.497 63 35 0.062 0.686

815, 12:12 L/D 1.54E+06 0.2 1.6 2.7 0.498 61 37 0.063 0.721

915, 12:12 L/D 1.60E+06 0.1 0.5 3.3 0.496 58 39 0.064 0.809

10 15, 12:12 L/D 1.72E+06 0.0 0.5 3.3 0.492 56 41 0.063 0.820

11 15, 12:12 L/D 1.76E+06 0.0 0.5 3.7 0.493 54 42 0.066 0.797

Isothermal (15 C), ÔDark StressÕExperiment

115, 12:12 L/D 2.04E+05 81.9 11.4 1.5 0.566 68 31 0.066 0.650

215, Dark Shift 3.06E+05 74.3 10.8 1.5 0.553 67 32 0.067 0.608

315, 24 D 3.82E+05 71.4 10.5 1.3 0.528 66 32 0.081 0.610

415, 24 D 4.00E+05 70.4 10.4 1.0 0.554 64 35 0.096 0.675

515, 24 D 4.17E+05 65.5 10.2 0.6 0.567 58 40 0.144 0.760

615, 24 D 4.02E+05 59.8 9.8 0.5 0.587 51 46 0.197 0.859

715, 24 D 3.99E+05 52.4 9.3 0.4 0.608 49 48 0.244 0.841

20–15 C Temperature Shift, ÔDark StressÕExperiment

120, 12:12 L/D 3.64E+05 898 37.4 1.4 0.761 67 31 0.025 a

2 T and Dark Shift 5.59E+05 924 36.0 1.3 0.745 67 31 0.022 a

315, 24 D 5.54E+05 918 36.9 1.3 0.725 65 33 0.031 a

415, 24 D 5.58E+05 924 37.0 0.95 0.749 61 37 0.040 a

515, 24 D 5.41E+05 924 37.5 0.66 0.764 56 42 0.053 a

615, 24 D 5.11E+05 922 37.1 0.62 0.763 54 44 0.063 a

715, 24 D 5.29E+05 929 37.1 0.61 0.765 52 45 0.066 a

Control for 20–15 C Temperature Shift, ÔDark StressÕExperiment

120, 12:12 L/D 1.41E+05 955 34.4 1.8 0.747 67 31 0.028 a

2 T Shift, 12:12 L/D 2.37E+05 790 28.6 1.5 0.749 66 32 0.023 a

315, 12:12 L/D 2.83E+05 931 33.5 1.6 0.687 66 32 0.030 a

415, 12:12 L/D 3.62E+05 924 32.5 1.8 0.645 67 32 0.036 a

515, 12:12 L/D 5.52E+05 902 31.0 1.7 0.599 66 32 0.042 a

615, 12:12 L/D 7.94E+05 871 28.8 1.8 0.571 66 32 0.048 a

715, 12:12 L/D 9.73E+05 842 26.8 1.9 0.556 66 32 0.053 a

815, 12:12 L/D 1.55E+06 749 22.4 2.1 0.529 65 33 0.053 a

915, 12:12 L/D 1.92E+06 818 23.2 1.8 0.530 64 34 0.059 a

10 15, 12:12 L/D 2.86E+06 679 15.9 1.6 0.517 62 35 0.060 a

a, Cannot be calculated; EE not resolved by GC from tri-unsaturated C

ethyl ketone; hence, calculated %K37 and %K38 values are 3% too low and too

high, respectively.

112 F.G. Prahl et al. 70 (2006) 101–117

regions (e.g., Kucera et al., 2005). Examples of promising

geochemical approaches for SST reconstruction include

measurement of magnesium to calcium ratios in planktonic

foraminiferal shells (Elderﬁeld and Ganssen, 2000) and the

alkenone unsaturation index UK0

37 (Brassell et al., 1986).

Paired use of techniques such as these, however, has in

some cases yielded conﬂicting conclusions about the mag-

nitude and timing of SST change (e.g., Niebler et al.,

2003), suggesting that the geochemical methods need to

be assessed for biases just as the faunal methods do.

As a case in point, Feldberg and Mix (2003) argued, based

on results from statistical analysis of foraminiferal species

assemblages, that equatorial waters in the Eastern Tropical

Paciﬁc Ocean (ETP) during the LGM were 3–5 C colder

than at present while those in the north-ﬂowing Peru Current

oﬀ southern Peru were 6–8 C colder. Our stratigraphic anal-

ysis of UK0

37 in a core from the ETP (Y69-71P) and from the

Peru Current oﬀ southern Peru (Y71-6-12P) suggests the

magnitude of SST cooling in the LGM was 3C at both

sites (Figs. 3A and B), and the cooling occurred at slightly

diﬀerent times. Can the noted diﬀerences in assessment be

reconciled objectively? A deﬁnitive answer to this question

is beyond the scope of the present study, but it is now possible

to more objectively advance our sense of uncertainty in UK0

37-

based SST estimates for the Southeast Paciﬁc region, given

results from a detailed analysis of alkenone/alkenoate com-

positions preserved in these sediments coupled with ﬁndings

from our laboratory-controlled batch culture experiments

with E. huxleyi.

4.1. Oceanographic controls on the sedimentary biomarker

signature

The greatest uncertainty in the UK0

37 technique now ap-

pears to stem from yet limited knowledge of the ecology

of alkenone producers in surface waters and of the biogeo-

chemical processes controlling alkenone export production

from the euphotic zone. Two key questions arise. First,

what species contribute to the alkenone/alkenoate signa-

ture preserved in sediments and how does this assemblage

change through space and time? Second, what is the phys-

iological status of cells packaged into settling particulate

material that ultimately forms the long-term sedimentary

record?

Our study shows surface sediments accumulating all

along the continental margin oﬀ the west coast of South

America are consistent with the global UK0

37–maSST calibra-

tion, both in trend and scatter (Fig. 1). On this basis, the

overall alkenone/alkenoate composition in exponentially

dividing cells of this algal strain would be expected to close-

ly match that preserved in surface sediment from our study

region. However, preserved compositions of %K37, ME/

K37, and EE/ME more closely match those in nutrient

and/or dark stressed cells than those in healthy, exponen-

tially dividing cells (Fig. 2). This ﬁnding is not surprising

because biogenic sedimentation is largely controlled by

the settling of particle aggregates (McCave, 1975), which

originate from surface waters as products of feeding pro-

cesses (Turner, 2002). Fecal pellets and marine snow (Pas-

sow et al., 2001; Thornton, 2002) represent two distinct

classes of such biogenic particle aggregates. Assuming dia-

genetic overprinting is an insigniﬁcant process, the alke-

none/alkenoate signatures recorded by sediments

accumulating at the seaﬂoor should reﬂect that encoded

into fecal pellets and marine snow at the time such particle

aggregates are produced or perhaps shortly thereafter.

Bloom events, although particularly renowned in re-

gions of divergence where upwelling occurs, are features

common to most oceanic surface waters (Miller, 2004).

The fate for most bloom-generated organic matter is to

be grazed and remineralized within the euphotic zone.

However, some fraction survives in situ destruction and be-

comes packaged during grazing and other feeding process-

es into aggregates that rapidly settle out of the euphotic

zone, thereby contributing to export production (Turner,

2002). Compositional characteristics of alkenone/alkeno-

ate signatures encoded within export production would de-

pend upon when during blooms that cells are packaged

into these aggregates and speciﬁcally how the packaging

occurred.

Blooms represent periods of unbalanced primary pro-

duction in surface waters and are terminated ultimately

by nutrient depletion (Miller, 2004) with processes such

as lysis, either massive viral or auto-induced (e.g., Bidle

and Falkowski, 2004 and references therein) further facili-

tating the termination. Some of the alkenone/alkenoate

content of export production from blooms certainly derives

from direct grazing of Ônutrient-stressedÕcells and thereby

should carry the biomarker signature of such cells to the

seaﬂoor. A signiﬁcant fraction of the biomarker signature

exported from surface waters is potentially associated with

living cells inadvertently incorporated into marine snow

aggregates generated by other planktonic processes within

the euphotic zone (e.g., Thornton, 2002; Turner, 2002). The

alkenone/alkenoate signature initially imparted to marine

snow by capture of a viable cell, regardless of its growth

stage, could then be physiologically modiﬁed over a period

of days as the aggregate settled out of the euphotic zone

into darkness in transit to the seaﬂoor. In this way, the sed-

iment record would also inherit a signature for ÔstressedÕ

cells.

Literature supports our proposition that the alkenone/

alkenoate composition preserved in marine sediments is

shaped by that most characteristic of stressed cells. Conte

et al. (1998) documented signiﬁcant compositional change

as cells shifted from the exponentially dividing to the sta-

tionary growth phase. Conte et al. (1995) also compared

their laboratory results for ME/K37 with measurements

for this property in samples of suspended particulate mate-

rials (SPM) collected from the euphotic zone in the North

Atlantic and found values measured in SPM collected un-

der non-bloom and late bloom conditions were typically

consistent with those characteristic of cultures in a station-

ary growth phase. Nonetheless, ÔstressedÕsignatures were

Alkenone paleothermometry oﬀ western South America 113

not always observed in their work. SPM collected during

periods of bloom development in mesocosm experiments

displayed ME/K37 values ﬁtting expectation, i.e., those

characteristic of healthy, exponentially dividing cells. Giv-

en the ﬁndings of ConteÕs research group and our current

results, a resemblance of alkenone/alkenoate compositions

in surface sediments throughout our study area and those

in ÔstressedÕcells of E. huxleyi is reasonable (Fig. 2).

Grimalt et al. (2000) reviewed a body of evidence from

ﬁeld study and laboratory experiments that suggest UK0

37 is

insensitive to signiﬁcant alteration once set biosynthetically

within the phytoplankton cell. However, results from more

recent, follow-up studies suggest otherwise (Marchand

et al., 2005; Rontani et al., 2005). Diagenetic eﬀects would

likely be most pronounced on ME/K37, a property based

on the relative abundance of two functionally distinct com-

pound classes with potentially the greatest susceptibility to

diﬀerential degradation (Harvey, 2000). Indeed, experimen-

tal results from laboratory-based studies with microbes

spanning a range of metabolisms show alkenoates are

degraded more rapidly than alkenones (Teece et al.,

1998). However, assuming the order of diagenetic selectiv-

ity deﬁned by these laboratory experiments applied to nat-

ural environments, diagenesis could not possibly transform

the ME/K37 composition of healthy, exponentially divid-

ing cells into one resembling that preserved in sediments

from our study area, because the trend is opposite

(Fig. 2B). Furthermore, ﬁeld evidence is lacking for signif-

icant diﬀerential degradation of alkenones and alkenoates.

In the North Atlantic, comparable rates of loss were appar-

ent for both compound classes at ﬁve bathypelagic sedi-

ment sites (Madureira et al., 1995) and no signiﬁcant

oﬀset was apparent between ME/K37 compositions in

SPM and underlying sediment (Conte et al., 1995). On this

basis, we discount diagenesis as a key cause for composi-

tional diﬀerences noted between alkenone/alkenoate signa-

tures in living cells of E. huxleyi and those preserved in

sediments and submit that ecological and physiological

processes operating in the upper ocean are the controlling

factors.

4.2. Impact of sedimentation process on alkenone-based SST

estimations

Alkenone/alkenoate signatures preserved in marine sed-

iments from the Southeast Paciﬁc suggest that the pre-

served record derives predominantly from physiologically

ÔstressedÕcells. Given that stresses imposed experimentally

on cultured cells are similar to those present in the natural

world and cause signiﬁcant systematic variability in UK0

values at a speciﬁed growth temperature, it is reasonable

to infer that some of the observed scatter in the global

core-top calibration for UK0

37 versus maSST is not just ran-

dom analytical noise. Thus, greater caution is required

whenever down-core UK0

37 measurements are used to recon-

struct absolute changes in maSST. This point is particular-

ly relevant when assessments disagree signiﬁcantly with

those made through analysis of alternative paleoSST prox-

ies, the circumstance now apparent in many regions,

including those represented by two core sites (Y69-71P

and Y71-6-12P) studied in the present work.

How accurately does standard use of alkenone paleoth-

ermometry constrain the magnitude of maSST change be-

tween the LGM and the present in our study area? The

method could either systematically under-estimate or

over-estimate actual SST change, given a relatively minor

change in the mechanical details of marine snow-dominat-

ed sedimentation (Thornton, 2002; Turner, 2002) such as

a simple shift in the timing of biomarker export from sur-

face water blooms. Whether the assessment yielded an un-

der-estimate or over-estimate would depend on how the

minor ecological change related to the past versus present

ocean.

Let us assume that CCMP1742, the strain of E. huxleyi

we cultured, represents the average alkenone/alkenoate

producer living in surface waters throughout our study

area at least since the LGM. Also, let us assume results

from our batch culture experiments accurately gauge how

the conditions of nutrient depletion and continuous dark-

ness would impact the alkenone/alkenoate composition

of these organisms in their natural environment. Given

only the occurrence of nutrient stress and current prospects

for minimal impact from selective diagenetic alteration, a

match between the ME/K37 composition of cells and that

preserved in surface sediments could then not be made

(Fig. 2B) unless most cells contributing to the biomarker

signal experienced dark stress at some point in their depo-

sitional history.

Extrapolating from culture results, dark stress acting

alone would cause an increase in UK0

37 or a perceived Ôwarm-

ing.ÕHowever, waters underlying the euphotic zone become

colder. Results from our temperature shift experiment (Ta-

ble 3) indicate the water temperature decrease encountered

beneath the euphotic zone may partially, perhaps even fully

compensate for the apparent ÔwarmingÕeﬀect of just dark

stress. Consequently, the UK0

37 signature initially incorporat-

ed within sinking marine snow could potentially retain an

approximate record of changing sea-surface temperature,

despite signiﬁcant metabolic loss of the biomarker signal

in the viable cell under continuous darkness conditions.

Of our three piston core sites, the temperature gradient be-

tween the euphotic zone (nominally 100 m) and 250 m

depth is strongest at the equatorial site (10 Conan

annual average), intermediate at the Peru site (5C),

and lowest at the subpolar transition site (2C) as shown

by data from World Ocean Atlas 2001 visualized using

Ocean Data View. On this basis, we predict the net impact

of temperature compensation on any dark stress eﬀect

would be greatest at the equatorial site. However, under

conditions of sea-surface cooling during glacial episodes,

reduced thermal gradients with depth could allow more

of the dark stress eﬀect to be expressed, resulting in a slight

under-estimate of temperature changes by UK0

37. Conversely,

an increase in the vertical thermal gradient during cold

114 F.G. Prahl et al. 70 (2006) 101–117

periods, driven by subsurface temperature changes, could

cause UK0

37 to over-estimate changes in temperature.

No matter how the process of dark stress speciﬁcally im-

pacts the UK0

37 of viable cells captured within sinking marine

snow in natural settings on a quantitative basis, it is possi-

ble that, for a given water temperature, the value incorpo-

rated within marine snow would not necessarily always

start out the same. The initial value would depend upon

the physiological status of the cell captured by the sedimen-

tary process. For example, assuming results from laborato-

ry experiments with CCMP1742 extrapolate to the ﬁeld

(see Popp et al., 2005; Prahl et al., 2005), cells captured un-

der nutrient-replete, mid-bloom conditions would display

UK0

37 values appearing somewhat ÔwarmerÕthan those cap-

tured under nutrient-depleted late-bloom conditions. In

the case of our Southeast Paciﬁc study area, glacial condi-

tions were thought to be more highly productive (Mohtadi

and Hebbeln, 2004). If so, and this productivity was asso-

ciated with less nutrient stress on the populations during

cold events, UK0

37 values could conceivably under-estimate

total temperature changes.

4.3. The paleoceanographic tally

Unquestionably, signiﬁcant glacial–interglacial variabil-

ity in UK0

37 is apparent in down-core stratigraphic records at

three distinct sites in our study area (Table 2). In each case,

values display a relatively smooth decrease back to the

LGM. Standard translation of these records into maSST

estimates suggests the magnitude of LGM ÔcoolingÕwas

not the same everywhere (Fig. 3). It was similar (3C)

at the equatorial site (Y69-71P) and the site within the

northern part of the Peru Current (Y71-6-12P) but two

times greater at the site (RR9702A-11PC) where the West

Wind Drift bifurcates, ultimately forming the Peru Current

(Fig. 1).

Results for Y71-6-12P agree with previous alkenone-

based estimates in the region (TG7, Calvo et al., 2001).

Nevertheless, in view of prospects we document for non-

thermal physiological impacts on UK0

37, and our inference

based on ME/EE signatures that all ice-age samples from

the region may have no modern ecological analogs, there

is now legitimate reason to question the accuracy of

LGM ÔcoolingÕassessed by this method, which is much less

than that inferred from foraminiferal species assemblages

(Feldberg and Mix, 2003; Kucera et al., 2005) and some-

what less than that inferred from radiolarian species assem-

blages (Pisias and Mix, 1997). Furthermore, analysis of

O and Mg/Ca data for foraminifera preserved in cores

from the equatorial region represented by Y69-71P sug-

gests LGM ÔcoolingÕwas much less than that suggested

by any of the faunal estimates, and only 50% of the

UK0

37-based assessment made here (Koutavas and Lynch-

Stieglitz, 2003). Clearly, no consensus yet exists about the

true magnitude of ice-age cooling in this region.

Discrepancies between SST assessments based on analy-

sis of alkenones and foraminiferal species assemblages are

not unique to our study. Niebler et al. (2003) recognized

similar eﬀects in their paleoceanographic study of the equa-

torial and south Atlantic and its associated eastern bound-

ary current system, especially in upwelling systems. They

ascribed the discrepancy to diﬀerences in the life history

of alkenone producers and foraminifera but did not ad-

dress the issue of physiological impacts on UK0

37. In view

of our ﬁndings, non-thermal physiological impacts on the

phytoplankton producing these biomarkers may provide

at least partial explanation for the discrepancy and need

to be considered carefully.

Finding systematic stratigraphic changes in the ME/EE

signature suggests major changes occurred in the phyto-

plankton ecology of surface waters overlying each of our

core sites. This prospect may provide a logical explanation

for some of the observed misﬁts between records for UK0

and various other types of temperature proxies. However,

all fault cannot and should not be cast on the UK0

37 method

as some level of systematic bias almost certainly applies as

well to other biologically based proxy assessments. Unfortu-

nately, there is yet insuﬃcient information available both

from culture experiments with alkenone producers and anal-

ysis of core records from our study area to determine more

speciﬁcally what type of ecological change may have oc-

curred. Pending further study, we cannot determine conﬁ-

dently how our evidence for ecological change impacts the

accuracy of UK0

37-based maSST assessments. Nonetheless,

given EE/ME values display the smallest LGM—Modern

change in RR9702A-11PC (Fig. 4), we now propose that

non-thermal physiological eﬀects on the accuracy of maSST

assessments are lowest and perhaps unimportant in the

southern part of our study area. The stratigraphic behavior

of EE/ME values in the two northern cores suggests non-

thermal physiological factors have not been constant since

the LGM. Consequently, the accuracy of LGM ÔcoolingÕin

these regions currently warrants greatest suspicion.

By careful analysis of the full alkenone/alkenoate com-

positional suite preserved in sediments, understanding of

both past temperature changes and other key elements of

paleocirculation and ecosystems will be improved, yielding

deeper insights into past and plausible future changes in the

global system. Advancement in future development of the

UK0

37 paleothermometer will almost certainly require coordi-

nated laboratory culture and natural surface water studies,

and eﬀorts made to put new biological discoveries in ocean-

ographic context with the fossil sediment record for these

biomarkers.

5. Conclusions

(1) The regional core-top calibration for UK0

37 versus

maSST in the modern Southeast Paciﬁc closely

matches the equivalent global core-top calibration

of Muller et al. (1998).

(2) Batch culture experiments with the paleoceanograph-

ic benchmark strain of E. huxleyi, CCMP1742, dem-

onstrate there can be signiﬁcant non-thermal

Alkenone paleothermometry oﬀ western South America 115

physiological eﬀects on UK0

37 and the alkenone/alkeno-

ate composition. A compelling, although yet unprov-

en, argument can be made that these non-thermal

physiological eﬀects shape the overall biomarker sig-

nature preserved in Southeast Paciﬁc sediments and

likely elsewhere in the ocean.

(3) Analysis of sediment cores collected on the conti-

nental margin oﬀ the west coast of South America

yields plausible regional histories of paleotempera-

ture based on UK0

37, but other aspects of the pre-

served alkenone/alkenoate signature imply that

ice-age samples may have no modern ecological

analog in the core-top calibration set. Further

analysis of the complete set of alkenone/alkenoate

compositional properties in sedimentary records, in

the modern water column, and in carefully

designed culture experiments should shed light on

these ecological issues and eventually yield richer

information about paleoceanographic change than

is now possible through current use of just the

UK0

37 index.

Acknowledgments

We are thankful to Joe Jennings for all nutrient analy-

ses, Dr. R. Schlitzer for making available a powerful way

to visualize data from World Ocean Atlas 2001 using the

software package Ocean Data View (http://www.

awi-bremerhaven.de/GEO/ODV/), and two reviewers

(Jean-Franc¸ois Rontani and one anonymous) for highly

constructive comments on a prior version of this paper.

We are also most grateful to the National Science Founda-

tion funded Core Repository at Oregon State University

for providing all sediment samples analyzed in this

study and to the Ocean Sciences Division of NSF for all

research support (OCE-9730376: both F.G.P. and

A.C.M.; OCE-9986306 and 0350409: F.G.P.).

Associate Editor: H. Rodger Harvey

References

Bentaleb, I., Grimalt, J.O., Vidussi, F., Marty, J.-C., Martin, V., Denis,

M., Hatte, C., Fontugne, M., 1999. The C

alkenone record of

seawater temperatures during seasonal thermocline stratiﬁcation.

Marine Chemistry 64, 301–313.

Bidle, K.D., Falkowski, P.G., 2004. Cell death in planktonic photosyn-

thetic microorganisms. Nature Reviews 2, 643–655.

Brassell, S.C., 1993. Applications of biomarkers for delineating marine

paleoclimatic ﬂuctuations during the Pleistocene. In: Engel, M.H.,

Macko, S.A. (Eds.), Organic Geochemistry Principles Applications.

Plenum Press, New York, pp. 699–738.

Brassell, S.C., Eglinton, G., Marlowe, I.T., Pﬂaumann, U., Sarnthein, M.,

1986. Molecular stratigraphy: a new tool for climatic assessment.

Nature 320, 129–133.

Calvo, E., Pelejero, C., Herguera, J.C., Palanques, A., Grimalt, J.O., 2001.

Insolation dependence of the southeastern Subtropical Paciﬁc sea

surface temperature over the last 400 kyrs. Geophysical Research

Letters 28, 2481–2484.

Clark, P.U., McCabe, A.M., Mix, A.C., Weaver, A.J., 2004. Rapid rise of

sea level 19,000 years ago and its global implications. Science 304,

1141–1144.

CLIMAP Project Members, 1981. Seasonal Reconstructions of the EarthÕs

Surface at the Last Glacial Maximum.

Coale, K.H., Bruland, K.W., 1987. Oceanic stratiﬁed euphotic zone as

elucidated by

234

Th:

238

U disequilibria. Limnology and Oceanography

32, 189–200.

Conte, M.H., Eglinton, G., Madureira, L.A.S., 1995. Origin and fate of

organic biomarker compounds in the water column and sediments of

the eastern North Atlantic. Philosophical Transactions of the Royal

Society of London B 348, 169–178.

Conte, M.H., Thompson, A., Lesley, D., Harris, R., 1998. Genetic and

physiological inﬂuences on the alkenone/alkenoate versus growth

temperature relationship in Emiliania huxleyi and Gephyrocapsa

oceanica. Geochimica et Cosmochimica Acta 62, 51–68.

Elderﬁeld, H., Ganssen, G., 2000. Past temperature and d

O of surface

ocean waters inferred from foraminferal Mg/Ca ratios. Nature 405,

442–445.

Epstein, B.L., dÕHondt, S., Hargraves, P.E., 2001. The possible metabolic

role of C

alkenones in Emiliania huxleyi. Organic Geochemistry 32,

867–875.

Epstein, B.L., dÕHondt, S., Quinn, J.G., Zhang, J., Hargraves, P.E., 1998.

An eﬀect of dissolved nutrient concentrations on alkenone-based

temperature estimates. Paleoceanography 13, 122–126.

Feldberg, M.J., Mix, A.C., 2003. Planktonic foraminifera, sea surface

temperatures, and mechanisms of oceanic change in the Peru and

south equatorial currents, 0–150 ka BP. Paleoceanography 18, 1016.

doi:10.1029/2001PA000740.

Green, J.C., Leadbetter, B.S.C., 1994. The Haptophyte Algae. Clarendon

Press, Oxford.

Grimalt, J.O., Rullkotter, J., Sicre, M.-A., Summons, R., Farrington, J.,

Harvey, H.R., Goni, M., Sawada, K., 2000. Modiﬁcations of the C

alkenone and alkenoate composition in the water column and

sediment: possible implications for sea surface temperature estimates

in paleoceanography. Geochemistry, Geophysics and Geosystems 1,

2000GC000053.

Harvey, H.R., 2000. Alteration processes of alkenones and related lipids in

water columns and sediments. Geochemistry, Geophysics and Geosys-

tems 1, 2000GC000054.

Hedges, J.I., Stern, J.H., 1984. Carbon and nitrogen determinations of

carbonate-containing solids. Limnology and Oceanography 29, 663–666.

Herbert, T.D., Schuﬀert, J.D., Thomas, D., Lange, C., Weinheimer, A.,

Peleo-Alampay, A., Herguera, J.-C., 1998. Depth and seasonality of

alkenone production along the California margin inferred from a core

top transect. Paleoceanography 13, 263–271.

Koutavas, A., Lynch-Stieglitz, J., 2003. Glacial–interglacial dynamics of

the eastern equatorial Paciﬁc cold tongue—Intertropical Convergence

Zone system reconstructed from oxygen isotope records. Paleoceanog-

raphy 18, 1089. doi:10.1029/2003PA000894.

Kucera, M., Kiefer, T., Pﬂaumann, U., Hayes, A., Weinelt, M., Chen, M.-

T., Mix, A.C., Cortijo, E., Duprat, J., Waelbroeck, C., 2005.

Reconstruction of the glacial Atlantic and Paciﬁc sea-surface temper-

atures from assemblages of planktonic foraminifera: multi-technique

approach based on geographically constrained calibration datasets.

Quaternary Science Reviews 24, 951–998.

Madureira, L.A.S., Conte, M.H., Eglinton, G., 1995. The early diagenesis

of lipid biomarker compounds in North Atlantic sediments. Paleoce-

anography 10, 627–642.

Malmgren, B.A., Nordlund, U., 1997. Application of artiﬁcial neural

networks to paleoceanographic data. Palaeogeography, Palaeoclima-

tology, Palaeoecology 136, 359–373.

Marchand, D., Marty, J.-C., Miquel, J.-C., Rontani, J.-F., 2005. Lipids

and their oxidation products as biomarkers for carbon cycling in the

northwestern Mediterranean Sea: results from a sediment trap study.

Marine Chemistry 95, 129–147.

Martinson, D., Pisias, N.G., Hays, J.D., Imbrie, J., Moore, T.C.J.,

Shackleton, N.J., 1987. Age dating and orbital theory of the ice ages:

116 F.G. Prahl et al. 70 (2006) 101–117

development of a high-resolution 0 to 300,000 year chronostratigra-

phy. Quaternary Research 27, 1–27.

McCave, I.N., 1975. Vertical ﬂux of particles in the ocean. Deep-sea

Research I 22, 491–502.

Miller, C.B., 2004. Biological Oceanography. Blackwell Science, Oxford.

Mix, A.C., Morey, A.E., Pisias, N.G., Hostetler, S.W., 1999. Foraminif-

eral faunal estimates of paleotemperature: circumventing the no-

analog problem yields cool ice-age tropics. Paleoceanography 14, 350–

359.

Mix, A.C., Tiedemann, R., Blum, P. (Eds.), 2003. Proc. ODP, Initial

Reports, 202, College Station, TX (Ocean Drilling Program).

Mohtadi, M., Hebbeln, D., 2004. Mechanisms and variations of the

paleoproductivity oﬀ northern Chile (24S–33S) during the last 40,000

years. Paleoceanography, 19. doi:10.1029/2004PA001003.

Morey, A.E., Mix, A.C., Pisias, N.G., 2005. Foraminiferal assemblages

preserved in surface sediments correspond to multiple environmental

variables. Quaternary Science Reviews 24, 925–950.

Muller, P.J., Kirst, G., Ruhland, G., von Storch, I., Rosell-Mele, A., 1998.

Calibration of the alkenone paleotemperature index UK0

37 based on

core-tops from the eastern South Atlantic and the global ocean (60N–

60S). Geochimica et Cosmochimica Acta 62, 1757–1772.

Niebler, H.-S., Arz, H.W., Donner, B., Mulitza, S., Patzold, J., Wefer, G.,

2003. Sea surface temperatures in the equatorial and South Atlantic

Ocean during the Last Glacial Maximum (23–19 ka). Paleoceanogra-

phy 18, 1069. doi:10.1029/2003PA000902.

Ocean Climate Laboratory, 1999. World Ocean Atlas 1998. In: National

Oceanographic Data Center (ed. 15).

Ortiz, J.D., Mix, A.C., 1997. Comparison of Imbrie-Kipp transfer

function and modern analog temperature estimates using sediment

trap and core top foraminiferal faunas. Paleoceanography 12, 175–

190.

Passow, U., Shipe, R.F., Murray, A., Pak, D.K., Brzezinski, M.A.,

Alldredge, A.L., 2001. The origin of transparent exopolymer particles

(TEP) and their role in the sedimentation of particulate matter.

Continental Shelf Research 21, 327–346.

Pﬂaumann, U., Duprat, J., Pujol, C., Labracherie, L., 1996. SIMMAX: a

modern analog technique to deduce Atlantic sea surface temperatures

from planktonic foraminifera in deep-sea sediments. Paleoceanography

11, 15–35.

Pisias, N.G., Mix, A.C., 1997. Spatial and temporal oceanographic

variability of the eastern equatorial Paciﬁc during the late Pleistocene:

evidence from radiolarian microfossils. Paleoceanography 12, 381–393.

Popp, B.N., Kenig, F., Wakeham, S.G., Laws, E.A., Bidigare, R.R., 1998.

Does growth rate aﬀect ketone unsaturation and intracellular carbon

isotopic variability in Emiliania huxleyi. Paleoceanography 13, 35–41.

Popp, B.N., Prahl, F.G., Wallsgrove, R.J., Tanimoto, J., 2005. Seasonal

patterns of alkenone production in the subtropical oligotrophic North

Paciﬁc. Paleoceanography, in press.

Prahl, F.G., Collier, R.B., Dymond, J., Lyle, M., Sparrow, M.A., 1993. A

biomarker perspective on prymnesiophyte productivity in the north-

east Paciﬁc Ocean. Deep-sea Research I 40, 2061–2076.

Prahl, F.G., Herbert, T., Brassell, S.C., Ohkouchi, N., Pagani, M.,

Repeta, D., Rosell-Mele, A., Sikes, E., 2000. Status of alkenone

paleothermometer calibration: report from Working Group3. Geo-

chemistry, Geophysics and Geosystems 1, 1–13.

Prahl, F.G., Muehlhausen, L.A., Lyle, M., 1989. An organic geochemical

assessment of oceanographic conditions at MANOP Site C over the

past 26,000 years. Paleoceanography 4, 495–510.

Prahl, F.G., Muehlhausen, L.A., Zahnle, D.L., 1988. Further evaluation

of long-chain alkenones as indicators of paleoceanographic conditions.

Geochimica et Cosmochimica Acta 52, 2303–2310.

Prahl, F.G., Pilskaln, C.H., Sparrow, M.A., 2001. Seasonal record for

alkenones in sedimentary particles from the Gulf of Maine. Deep-sea

Research I 48, 515–528.

Prahl, F.G., Popp, B.N., Karl, D.M., Sparrow, M.A., 2005. Ecology and

biogeochemistry of alkenone production at Station ALOHA. Deep-sea

Research I 52, 699–719.

Prahl, F.G., Wolfe, G.V., Sparrow, M.A., 2003. Physiological impacts on

alkenone paleothermometry. Paleoceanography 18, 1025. doi:10.1029/

2002PA000803.

Rontani, J.-F., Bonin, P., Jameson, I., Volkman, J.K., 2005. Degradation

of alkenones and related compounds during oxic and anoxic incuba-

tion of the marine haptophyte Emiliania huxleyi with bacterial

consortia isolated from microbial mats from the Camargue, France.

Organic Geochemistry 36, 603–618.

Small, L.F., Knauer, G.A., Tuel, M.D., 1987. The role of sinking fecal

pellets in stratiﬁed euphotic zones. Deep-sea Research I 34, 1705–1712.

Strickland, J.D.H., Parsons, T.R., 1972. A Practical Handbook of

Seawater Analysis. Fisheries Research Board of Canada.

Strub, T.P., Mesias, J.M., Montecino, V., Rutllant, J., Salinas, S., 1998.

Coastal ocean circulation oﬀ western South America. In: Robinson,

A.R., Brink, K.H. (Eds.), The Sea, vol. 11. John Wiley and Sons, New

York, pp. 273–313.

Teece, M.A., Getliﬀ, J.M., Leftley, J.W., Parkes, R.J., Maxwell, J.R.,

1998. Microbial degradation of the marine prymnesiophyte Emiliania

huxleyi under oxic and anoxic conditions as a model for early

diagenesis: long chain alkadienes, alkenones and alkyl alkenoates.

Organic Geochemistry 29, 863–880.

Ternois, Y., Sicre, M.-A., Boireau, A., Conte, M.H., Eglinton, G., 1997.

Evaluation of long-chain alkenones as paleo-temperature indicators in

the Mediterranean Sea. Deep-sea Research I 44, 271–286.

Thornton, D.C.O., 2002. Diatom aggregation in the sea: mechanisms and

ecological implications. European Journal of Phycology 37, 149–167.

Turner, J.T., 2002. Zooplankton fecal pellets, marine snow and sinking

phytoplankton blooms. Aquatic Microbial Ecology 27, 57–102.

Verardo, D.J., Froelich, P.N., McIntyre, A., 1990. Determination of

organic carbon and nitrogen in marine sediments using the Carlo Erba

NA-1500 analyzer. Deep-sea Research I 37, 157–165.

Waelbroeck, C.L., Labeyrie, L., Duplessy, J.-C., Guiot, J., Labracherie,

M., Leclaire, H., Duprat, J., 1998. Improving past sea surface

temperature estimates based on planktonic fossil faunas. Paleoceanog-

raphy 12, 272–283.

Yamamoto, M., Shiraiwa, Y., Inouye, I., 2000. Physiological responses of

lipids in Emiliania huxleyi and Gephyrocapsa oceanica (Haptophyceae)

to growth status and their implications for alkenone paleothermom-

etry. Organic Geochemistry 31, 799–811.

Alkenone paleothermometry oﬀ western South America 117

Upper-ocean temperature characteristics in the subantarctic southeastern Pacific based on biomarker reconstructions

Article

Full-text available

Sep 2023
CLIM PAST

As remnants of living organisms, alkenones and isoprenoid glycerol dialkyl glycerol tetraether lipids (isoGDGTs) are widely used biomarkers for determining ocean water temperatures from the past. The organisms that these proxy carriers stem from are influenced by a number of environmental parameters, such as water depth, nutrient availability, light conditions, or seasonality, which all may significantly bias the calibration to ambient water temperatures. Reliable temperature determinations thus remain challenging, especially in higher latitudes and for undersampled regions. We analyzed 33 sediment surface samples from the southern Chilean continental margin and the Drake Passage for alkenones and isoGDGTs and compared the results with gridded instrumental reference data from the World Ocean Atlas 2005 (WOA05) and previously published data from an extended study area covering the central and western South Pacific towards the Aotearoa / New Zealand continental margin. We show that for alkenone-derived sea surface temperatures (SSTs), the widely used global core-top calibration of Müller et al. (1998) yields the smallest deviation of the WOA05-based SSTs. On the contrary, the calibration of Sikes et al. (1997), determined for higher latitudes and summer SSTs, overestimates modern WOA05-based SSTs in both the annual mean and summer. Our alkenone SSTs show a slight seasonal shift of ∼ 1 ∘C at the southern Chilean margin and up to ∼ 2 ∘C in the Drake Passage towards austral summer SSTs. Samples in the central South Pacific, on the other hand, reflect an annual mean signal. We show that for isoGDGT-based temperatures, the subsurface calibration of Kim et al. (2012a) best reflects temperatures from the WOA05 in areas north of the Subantarctic Front (SAF). Temperatures south of the SAF are, in contrast, significantly overestimated by up to 14 ∘C, irrespective of the applied calibration. In addition, we used the GDGT [2]/[3] ratios, which give an indication of the production depth of the isoGDGTs and/or potential influences from land. Our samples reflect a subsurface (0–200 m water depth) rather than a surface (0–50 m water depth) signal in the entire study area and show a correlation with the monthly dust distribution in the South Pacific, indicating terrigenous influences. The overestimation of isoGDGT surface and subsurface temperatures south of the SAF highlights the need for a reassessment of existing calibrations in the polar Southern Ocean. Therefore, we suggest a modified Southern Ocean tetraether index (TEX86)-based calibration for surface and subsurface temperatures, which shows a lower temperature sensitivity and yields principally lower absolute temperatures, which align more closely with WOA05-derived values and also OH–isoGDGT-derived temperatures.

Upper ocean temperature characteristics in the subantarctic Southeast Pacific based on biomarker reconstructions

Preprint

Full-text available

Dec 2022

Alkenones and Glycerol Dialkyl Glycerol Tetraether lipids (GDGT) as remnants of living organisms are widely used biomarkers for determining past oceans’ water temperatures. The organisms these proxy carriers stem from, are influenced by a number of environmental parameters, such as water depth, nutrient availability, light conditions or seasonality, which all may significantly bias the calibration to ambient water temperatures. Reliable temperature determinations remain thus challenging, especially in higher latitudes and for under-sampled regions. We analyzed 33 sediment surface samples from the Southern Chilean continental margin and the Drake Passage for alkenones and GDGTs and compared the results with gridded instrumental reference data from the World Ocean Atlas 2005 (WOA05), as well as previously published data from an extended study area covering the Central and Western South Pacific towards the New Zealand continental margin. We show that for alkenone-derived SSTs, the widely-used global core-top calibration of Müller et al. (1998) yields the smallest residuals, whereas the calibration of Sikes et al. (1997), adapted to higher latitudes and supposed to show summer SSTs, overestimates modern WOA05-based (summer and annual mean) SSTs. Our alkenone SSTs show a slight seasonal shift of ~1 °C at the Southern Chilean Margin and up to ~2 °C in the Drake Passage towards austral summer SSTs, whereas samples in the Central South Pacific reflect an annual mean signal. We show that for GDGT-based temperatures, a more complex pattern emerges. In areas north of the Subantarctic Front (SAF) the subsurface calibration of Kim et al. (2012a) best reflects temperatures from the WOA05, largely within the margin error of ±2.2 °C. Temperatures south of the SAF instead are significantly overestimated by up to 14 °C, irrespective of the applied calibration. Based on a qualitative assessment of the GDGT [2] / [3]-ratios, which likely indicate water depth of origin, our samples reflect a subsurface (0 to 200 m water depth) rather than a surface (0–50 m water depth) signal. The overestimation of surface and subsurface temperatures south of the SAF highlights the need for a re-assessment of existing calibrations in the polar Southern Ocean, and leads to limitations in reliably both obtaining absolute values and assessing relative changes. Therefore, we suggest a modified Southern Ocean calibration for surface and subsurface GDGT-based temperatures, which shows a lower temperature sensitivity of the TEXL86 and yields principally lower absolute temperatures, which align more closely with WOA05-derived values.

Use of Gas Chromatography-Mass Spectrometry Techniques (GC-MS, GC-MS/MS and GC-QTOF) for the Characterization of Photooxidation and Autoxidation Products of Lipids of Autotrophic Organisms in Environmental Samples

Article

Full-text available

Mar 2022
MOLECULES

Jean-François Rontani

This paper reviews applications of gas chromatography-mass spectrometry techniques for the characterization of photooxidation and autoxidation products of lipids of senescent phototrophic organisms. Particular attention is given to: (i) the selection of oxidation products that are sufficiently stable under environmental conditions and specific to each lipid class and degradation route; (ii) the description of electron ionization mass fragmentation of trimethylsilyl derivatives of these compounds; and (iii) the use of specific fragment ions for monitoring the oxidation of the main unsaturated lipid components of phototrophs. The techniques best geared for this task were gas chromatography-quadrupole-time of flight to monitor fragment ions with very high resolution and accuracy, and gas chromatography-tandem mass spectrometry to monitor very selective transitions in multiple reaction monitoring mode. The extent of the degradation processes can only be estimated if the oxidation products are unaffected by fast secondary oxidation reactions, as it is notably the case of ∆5-sterols, monounsaturated fatty acids, chlorophyll phytyl side-chain, and di- and triterpenoids. In contrast, the primary degradation products of highly branched isoprenoid alkenes possessing more than one trisubstituted double bond, alkenones, carotenoids and polyunsaturated fatty acids, appear to be too unstable with respect to secondary oxidation or other reactions to serve for quantification in environmental samples.

Southern Ocean sea surface temperature synthesis: Part 1. Evaluation of temperature proxies at glacial-interglacial time scales

Article

Full-text available

Oct 2021
QUATERNARY SCI REV

Quaternary interglacial climates are often used as analogues for how the Antarctic Ice Sheet will respond to future climate warming. Southern Ocean marine sediments provide an important paleoclimate archive in this respect. Sea surface temperature (SST) reconstructions in the Southern Ocean depend exclusively on the fossils or geochemical signatures of planktic organisms, but the strengths of these SST proxies remain poorly quantified in this region. To improve confidence in paleoclimate reconstructions, Part 1 of this two-part study evaluates the reliability of Southern Ocean SST proxies employed at Quaternary glacial-interglacial time scales, focusing on three key potential problems: advection/dispersion, seasonality, and non-thermal influences. We find that foraminifera assemblages and long-chain alkenones likely provide the most reliable SST reconstructions in this region. Diatom assemblages and the Globigerina bulloides Mg/Ca ratio are considered to be 'moderately' reliable. Both are subject to potentially significant non-thermal influences, and diatom assemblages are likely modified by species-dependent advection as they sink to the sea floor. Nevertheless, diatoms are valuable at higher latitudes, since alkenones and foraminifera assemblages lose sensitivity below~1 to 2 C. Dinocyst assemblages, radiolarian assemblages, GDGTs and Neogloboquadrina pachyderma Mg/Ca are considered the least reliable in the Southern Ocean, due to weak calibrations, poorly-constrained non-thermal influences, and/or strong advection bias. We note that the seasonality of all proxies remains poorly constrained. Overall, Southern Ocean SST reconstructions using the recommended proxies and calibrations should be robust when averaging across multiple sites and proxy types, but should be treated with caution when analysing spatial variability, a small number of sites, or a single proxy type. Quantifying the effect of advection should be a priority for all planktic groups employed in Southern Ocean paleoclimate reconstructions.

Refining the climate, glacier, and volcanic history of Iceland during the Holocene

Thesis

Full-text available

Jun 2019

David Harning

Iceland’s position at the confluence of major oceanic and atmospheric fronts results in a highly sensitive climate evident in both instrumental and paleo records. However, open questions still remain regarding the pre-instrumental evolution of climate, glacier, and volcanic activity at this hemispherically relevant location. This dissertation capitalizes on and merges a range of analytical techniques in an effort to refine our understanding of Icelandic climate variability, glacier extent, and tephrochronology during the Holocene epoch, with a focus on Northwest Iceland. In order to provide robust age control in our records, this research required the development of a tephrochronological framework for West Iceland, a region that lacks the otherwise widely-dispersed rhyolitic marker tephras. Glacier proxies (threshold lake sediment records and emerging dead vegetation from receding ice margins) provide firm constraints on the Holocene activity of Drangajökull, an ice cap in northwest Iceland, and high-resolution lake sediment proxy records (TOC, δ13C, C/N and biogenic silica) collected adjacent to the glacier elucidate the concomitant climate. Furthermore, we explore two lipid biomarker paleothermometers (alkenones and branched glycerol dialkyl glycerol tetraether, GDGTs) in one of these lakes and its catchment soils for the first time in Iceland to quantify the evolution of Holocene summer temperature. Similar to other Icelandic ice cap histories, our records collectively illustrate that a warm early Holocene (2 to 5 oC above modern) likely resulted in the complete demise of Drangajökull shortly after 9 ka. Subsequent to peak early Holocene summer warmth, lake sediment climate proxies indicate punctuated declines in algal productivity and increases in soil erosion alongside steadily decreasing northern hemisphere (NH) summer insolation. As summers continued to cool, Drangajökull re-nucleated by ~2.3 ka and episodically expanded to its maximum dimension during the Little Ice Age (0.7-0.1 ka), when summer temperatures are estimated to be ~1 oC below modern. Triggers for cold anomalies are linked to variable combinations of freshwater pulses from waning Pleistocene ice sheets, low total solar irradiance, explosive and effusive volcanism, and internal modes of climate variability, with cooling likely sustained by ocean/sea-ice feedbacks. In addition to the lake record, GDGTs were also applied in two other settings: a Holocene soil archive in central Iceland and in the marine realm along the North Iceland Shelf. For the latter, we also developed an Icelandic GDGT-temperature calibration based on marine surface sediment that highlights the reduced uncertainty (± 0.4 oC) achievable for local rather than global calibrations (e.g., ± 4.0 oC). Local calibrations are particularly important for areas where the temperature relationship of GDGTs deviates from the overall linear correlation observed in global calibrations (i.e., cold and warm regions), such as Iceland. Although clearly reflected in the maximum dimensions of Drangajökull, the Little Ice Age cooling is obscured in all lake, soil and marine organic geochemical records investigated in this dissertation. For the former two, the erosion of older soils, nutrients and relic GDGTs likely compromise the records and imply warmth. On the other hand, the development of thick sea ice inferred from highly branched isoprenoid biomarkers on the North Iceland Shelf insulated the subsurface waters during the peak Little Ice Age, likely preventing the ventilation of heat from below the surface layer to the atmosphere. This dissertation provides critical and nuanced observations necessary for evaluating modeling simulations aiming to forecast the poorly constrained climate of the coming century.

Controls on Alkenone Carbon Isotope Fractionation in the Modern Ocean

Article

Full-text available

Dec 2021
GEOCHEM GEOPHY GEOSY

Carbon isotope records from alkenone biomarkers (εp37:2) produced by haptophyte algae are frequently used for atmospheric CO2 paleobarometry, but this method has yielded inconsistent results during periods where CO2 variations are known independently. Recent syntheses of algal cultures have quantitatively demonstrated that εp37:2 indeed records CO2 information: εp37:2 increases as aqueous CO2 concentrations increase relative to carbon demand. However, interpretations of εp37:2 are complicated by irradiance, where higher irradiance yields higher εp37:2. Here we examine the roles of physiology and environment in setting εp37:2 in the ocean. We compile water‐column and sediment core‐top εp37:2 data and add new core‐top measurements, including estimates of cell sizes and growth rates of the alkenone‐producing population. In support of culture studies, we find irradiance to be a key control on εp37:2 in the modern ocean. We test a culture‐derived model of εp37:2 and find that the quantitative relationships calibrated in culture experiments can be used to predict εp37:2 in sediment samples. In water‐column samples, the model substantially overestimates εp37:2, largely resulting from higher irradiance at the depth of sample collection than the integrated light conditions under which the sampled biomass was produced and vertically mixed to the collection depth. We argue that the theory underpinning the conventional diffusive alkenone carbon isotope fractionation model, including the “b” parameter, is not supported by field data and should not be used to reconstruct past CO2 changes. Future estimates of CO2 from εp37:2 should use empirical or mechanistic models to quantitatively account for irradiance and cell size variations.

Evolution of Sea Surface Temperature in the Southern Mid‐latitudes From Late Oligocene Through Early Miocene

Article

Full-text available

Sep 2021

Large Antarctic ice volume changes characterized the middle to Late Oligocene and the first million years of climate evolution during the Miocene. However, the sea surface temperature (SST) evolution over this period remains poorly constrained, as only a few records from contrasting proxies are available. In this study, we present a long‐term alkenone‐derived SST record from sediments drilled by the Ocean Drilling Program (ODP) at Site 1168 in the west Tasmanian Sea spanning 29.8 Ma to 16.7 Ma. The SST record highlight that the long‐term warming in the Late Oligocene linked to the end of the Middle Oligocene Glacial Interval can be recognized also at mid‐to‐high latitudes of the Southern Hemisphere. Warmer average temperatures (25.5°C) characterize the period from 24.6 to 22 Ma; average temperatures then decrease by 1°C–2°C into the Miocene and stabilize by 20.1 Ma. The reconstructed temperatures are highly variable in the warm Late Oligocene waters, and more stable and slightly colder in the Early to Middle Miocene. We confirm that this temperature trend is not an artifact of the latitudinal drift of the site, as the temperature anomaly relative to the modern water temperature at the paleolocation confirms the SST trends of the Oligocene. This is the first alkenone‐derived continuous record to reproduce the long‐term Oligocene climate trend previously interpreted from the benthic δ¹⁸O, which recorded a warming and/or reduction in ice volume from the Middle Oligocene Glacial Interval through the latest Oligocene.

Evolution of Sea Surface Temperature in the Southern Mid-latitudes from Late Oligocene through Early Miocene

Preprint

Jul 2021

Haptophyte algae-based alkenone paleothermometry

Chapter

Jan 2023

Holocene pyrolytic nitrogen compounds by using pyrolysis-GC/MS and its paleoclimatic implication from Jinchuan peatbog, NE China

Article

Oct 2021
QUATERN INT

We present a Holocene record of pyrolytic nitrogen (N) composition from the Jinchuan peat bog within the Longgang Volcanic Field (LVF), NE China. Based on pyrolysis gas chromatography and mass spectrometry (Py-GC/MS) analysis, N-compounds including acetamide, NHetero and NAro have been identified and quantified. Aromatic N-compounds (NAro, benzonitrile and benzamide) are mainly derived from phytoplankton and higher plants, and are resistant to microbial degradation. We propose ‘NHetero/NAro’ as a proxy based on microbial degradation, and has lower values between 10 and 5.8 ka BP, and higher values during the late Holocene. We suggest that temperature is main factor for NHetero/NAro, and sedimentary facies changed from floodplain to peat deposition, variation of effective precipitaion and pH value have limited influence on the performance of the proxy in Jinchuan. Low NHetero/NAro levels, combined with high levels of non-protein derived acetamide, indicate warmer conditions, and vice versa. NHetero/NAro is sensitive to temperature change, and follows a similar pattern to those of previous paleo-temperature time series derived from br-GDGT and long chain alkenones in this region. This Spectral analysis of the NHetero/NAro time series reveals significant periodicities of 103–105, 145 and 482–506 years, suggesting strong influences of solar radiation and monsoon intensity on temperature variations in NE China during the Holocene.

Sea surface temperatures in the equatorial and South Atlantic Ocean during the Last Glacial Maximum (23–19 ka)

Article

Full-text available

Sep 2003

We used planktic foraminiferal assemblages in 70 sediment cores from the tropical and subtropical South Atlantic Ocean (10°N-37°S) to estimate annual mean sea surface temperatures (SSTs) and seasonality for the Last Glacial Maximum with a modified version of the Imbrie-Kipp transfer function method (IKTF) that takes into account the abundance of rare but temperature sensitive species. In contrast to [1981], the reconstructed SSTs indicate cooler glacial SSTs in the entire tropical/subtropical South Atlantic with strongest cooling in the upwelling region off Namibia (7-10°C) and smallest cooling (1-2°C) in the western subtropical gyre. In the western Atlantic, our data support recent temperature estimates from other proxies. In the upwelling regions in the eastern Atlantic, our data conflict with SST reconstructions from alkenones, which may be due to an environmental preference of the alkenone-producing algae or to an underestimation of foraminiferal SSTs due to anomalous high abundances of N. pachyderma (sinistral).

Improving past sea surface temperature estimates based on planktonic fossil faunas

Article

Jun 1998

A Practical Handbook of Seawater Analysis

Article

Jan 1968

Calibration of the alkenone paleotemperature index Uk'37 based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S)

Article

Jan 1998

Calibration of the alkenone paleotemperature index U37K' based on core-tops from the eastern South Atlantic and the global ocean (60oN-60oS) - Present and Past Circulation

Article

May 1998
GEOCHIM COSMOCHIM AC

We have analysed alkenones in 149 surface sediments from the eastern South Atlantic in order to establish a sediment-based calibration of the U37K′ paleotemperature index. Our study covers the major tropical to subpolar production systems and sea-surface temperatures (SST’s) between 0° and 27°C. In order to define the most suitable calibration for this region, the U37K′ values were correlated to seasonal, annual, and production-weighted annual mean atlas temperatures and compared to previously published culture and core-top calibrations. The best linear correlation between U37K′ and SST was obtained using annual mean SST from 0 to 10 m water depth (U37K′ = 0.033 T + 0.069, r2 = 0.981). Data scattering increased significantly using temperatures of waters deeper than 20 m, suggesting that U37K′ reflects mixed-layer SST and that alkenone production at thermocline depths was not high enough to significantly bias the mixed-layer signal. Regressions based on both production-weighted and on actual annual mean atlas SST were virtually identical, indicating that regional variations in the seasonality of primary production have no discernible effect on the U37K′ vs. SST relationship. Comparison with published core-top calibrations from other oceanic regions revealed a high degree of accordance. We, therefore, established a global core-top calibration using U37K′ data from 370 sites between 60°S and 60°N in the Atlantic, Indian, and Pacific Oceans and annual mean atlas SST (0–29°C) from 0 m water depth. The resulting relationship (U37K′ = 0.033 T + 0.044, r2 = 958) is identical within error limits to the widely used E. huxleyi calibrations of Prahl and Wakeham (1987) and Prahl et al. (1988) attesting their general applicability. The observation that core-top calibrations extending over various biogeographical coccolithophorid zones are strongly linear and in better accordance than culture calibrations suggests that U37K′ is less species-dependent than is indicated by culture experiments. The results also suggest that variations in growth rate of algae and nutrient availability do not significantly affect the sedimentary record of U37K′ in open ocean environments.

Application of artificial neural networks to paleoceanographic data.

Conference Paper

Jan 1995
PALAEOGEOGR PALAEOCL

Oceanic stratified euphotic zone as elucidated by 234Th : 238U disequilibria

Article

Jan 1987

Profiles of dissolved and particulate ²³⁴ Th were determined at the VERTEX 2 and 3 stations off Manzanillo, Mexico, and at the VERTEX 4 station about 900 km north of Hawaii. By modeling the disequilibria between ²³⁴ Th and ²³⁸ U in the dissolved and particulate form, estimates of scavenging rates for Th from the dissolved to particulate phases, particle residence times, and the flux of Th via particle removal can be obtained. ²³⁴ Th: ²³⁸ U activity ratio profiles indicate that the euphotic zone can be separated into two layers: An upper oligotrophic layer characterized by low new production values, low net scavenging, and long dissolved ²³⁴ Th residence times; and a subsurface eutrophic layer with higher new production values, more intense scavenging, and shorter dissolved ²³⁴ Th residence times. New production, rather than total primary production may determine net scavenging rates of reactive elements from oceanic surface waters. These results contribute to the emerging descriptions of the layered structure of oligotrophic euphotic zones and support the notion that this may be a general and ubiquitous feature of global stratified oligotrophic regimes. These layered systems can be structured not only in biological and nutrient parameters, but also in terms of the rates of chemical scavenging and elemental transport.

The origin of Transparent Exopolymer Particles (TEP) and their role in the sedimentation of particulate matter

Article

Mar 2001
CONT SHELF RES

Seasonal changes in the concentration of suspended transparent exopolymer particles (TEP) and flux rates of TEP and other particles at 500m were measured at 2-week intervals at a station in the Santa Barbara Channel between May 1995 and June 1997 in order to investigate the hypothesis that the presence of TEP is necessary for the aggregation and subsequent sedimentation of particles from the pelagic zone. During the 2-year period phytoplankton appeared to be the most significant source of TEP. However, in association with phytoplankton, the concentration of TEP was also positively affected by bacteria abundance. Possibly bacteria enhance the production of TEP by phytoplankton. The presence of TEP was necessary for the sedimentation of diatoms. However, only 67% of the peaks in particulate organic carbon flux corresponded to peaks in TEP flux. Lithogenic silica sedimented only when scavenged by marine snow; either by TEP-rich diatom aggregates or by zooplankton-derived snow (larvacean houses). TEP were not involved in the sedimentation of foraminifera. Although sedimentation was the dominant loss processes of TEP out of the euphotic zone, other loss process must have been important at greater depth, as only a small fraction of the standing stock of TEP arrived at 500m.

Origin and Fate of Organic Biomarker Compounds in the Water Column and Sediments of the Eastern North Atlantic

Article

May 1995

This paper focuses upon lipid biomarkers as tracers of the biological carbon cycle and our efforts to derive and validate `molecular tools' for oceanography and palaeoceanography as part of the 1989-1991 UK-JGOFS Biogeochemical Ocean Flux Study (BOFS). Biomarker concentrations and composition in water column particulates and bottom sediments in the North Atlantic show a strong correspondence to seasonal and interannual patterns of productivity. Biomarkers document the rapidity with which vertical flux processes operate in the high latitude North Atlantic: for example, the massive sedimentation of phytodetritus following a coccolithophorid bloom in the Iceland Basin and its subsequent resuspension, and the benthic biological response to this pulse of biologically available carbon. Sedimentary biomarker distributions indicate that organic material decomposition in sediments is dominated by processes at or near the sediment water interface and that downmixing of labile material into the sediments is largely controlled by advective rather than diffusive-like processes.

Mechanisms and variations of the paleoproductivity off northern Chile (24°S–33°S) during the last 40,000 years

Article

Jun 2004

A multiparameter investigation including organic carbon, carbonate, opal, and planktic foraminifera was carried out on five sediment cores from the coastal upwelling area between 24°S and 33°S along the Peru-Chile Current to reconstruct the history of the paleoproductivity and its driving mechanisms during the last 40,000 years. Inferred from our data, we conclude that the Antarctic Circumpolar Current as the main nutrient source in this region mainly drives the productivity by its latitudinal shifts associated with climate change. Simplified, its northerly position during the last glacial led to enhanced productivities, and its southerly position during the Holocene caused lower productivities. At 33°S the paleoproductivity was additionally affected by the southern westerlies and records highest levels during the Last Glacial Maximum (LGM). North of 33°S, several factors (e.g., position and strength of the South Pacific anticyclone, wind stress, continental runoff, and El Niño Southern Oscillation events) supplementary influenced upwelling and paleoproductivity, where maximum values occurred prior to the LGM and during the deglaciation.

Alkenone paleothermometry: Biological lessons from marine sediment records off western South America

Abstract and Figures

Recommended publications

Early Quaternary (Pleistocene) and Precursors

Late Pleistocene calcareous nannofossils in the caribbean and their interoceanic correlation

Patterns of isozyme variation as indicators of biogeographic history in Pilgerodendron uviferum (D....

Ice-quakes and Ice-vibrations Sources of Glacial Seismicity Recorded at Maitri, Antarctica Keywords:...