Castellano

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 110, D06114, doi:10.
1029/2004JD005259, 2005
Holocene volcanic history as recorded in the sulfate stratigraphy of the

European Project for Ice Coring in Antarctica Dome C (EDC96)
ice core
E. Castellano,1 S. Becagli,1 M. Hansson,2 M. Hutterli,3 J. R. Petit,4 M. R. Rampino,5
M. Severi,1 J. P. Steffensen,6 R. Traversi,1 and R. Udisti1
Received 21 July 2004; revised 19 November 2004; accepted 30 December 2004; published 25 March 2005.
[1] A detailed history of Holocene volcanism was reconstructed using the sulfate record
of the European Project for Ice Coring in Antarctica Dome C (EDC96) ice core. This
first complete Holocene volcanic record from an Antarctic core provides a reliable
database to compare with long records from Antarctic and Greenland ice cores. A
threshold method based on statistical treatment of the lognormal sulfate flux distribution
was used to differentiate volcanic sulfate spikes from sulfate background
concentrations. Ninety-six eruptions were identified in the EDC96 ice core during the
Holocene, with a mean of 7.9 events per millennium. The frequency distribution (events
per millennium) showed that the last 2000 years were a period of enhanced volcanic
activity. EDC96 volcanic signatures for the last millennium are in good agreement with
those recorded in other Antarctic ice cores. For older periods, comparison is in some cases
less reliable, mainly because of dating uncertainties. Sulfate depositional fluxes of
individual volcanic events vary greatly among the different cores. A volcanic flux
normalization (volcanic flux/Tambora flux ratio) was used to evaluate the relative intensity
of the same event recorded at different sites in the last millennium. Normalized flux
variability for the same event showed the highest value in the 1100–1500 AD period. This
pattern could mirror changes in regional transport linked to climatic variations such as
slight warming stages in the Southern Hemisphere (Southern Hemisphere Medieval
Warming–like period?).
Citation: Castellano, E., S. Becagli, M. Hansson, M. Hutterli, J. R. Petit, M. R. Rampino, M. Severi, J. P. Steffensen, R. Traversi, and
R. Udisti (2005), Holocene volcanic history as recorded in the sulfate stratigraphy of the European Project for Ice Coring in Antarctica
Dome C (EDC96) ice core, J. Geophys. Res., 110, D06114, doi:10.1029/2004JD005259.
1. Introduction [3] Volcanic eruptions emit large amounts of particulate

matter and gases into various levels of the atmosphere,
[2] Polar ice sheets are extensive sources of information
depending on eruption intensity and magma composition
on past atmospheric composition. Snow accumulated layer
[Rampino and Self, 1982]. Explosive eruptions inject ash
by layer and gradually compressed into solid ice in the
and gases (mainly SO2) directly into the stratosphere,
coldest areas of ice sheets contains traces of atmospheric
where they are transported on a regional to global scale
aerosol at the time of deposition. A major record coming
[Coffey, 1996]. Aerosol microdroplets of sulfuric acid,
from ice cores is the history of past explosive volcanism,
formed by atmospheric oxidation of SO2 and gas-to-
reconstructed by electric conductivity measurements (ECM)
particle conversion, have stratospheric lifetimes of months
[e.g., Hammer et al., 1997] and sulfate ice core profiles
to a few years, affecting the Earth’s radiative balance and
[e.g., Zielinski, 2000].
climate and perturbing the chemistry of the stratosphere.
The global distribution of volcanic aerosols was confirmed
1
Department of Chemistry, University of Florence, Florence, Italy. by satellite observation [e.g., Hofmann et al., 2003], and
2
Department of Physical Geography and Quaternary Geology, Stock- its positive (warming) or negative (cooling) climatic
holm University, Stockholm, Sweden. effects have been largely discussed. Robock [2000] and
3
Physics Institute, University of Bern, Bern, Switzerland.
4
Laboratoire de Glaciologie et Géophysique de l’Environnement du
Zielinski [2000] noted that stratospheric sulfate aerosol can
CNRS, Saint-Martin-d’Heres Cedex, France. affect regional or global climate through: (1) reduction of
5
Earth and Environmental Science Program, New York University, New the diurnal temperature cycle, (2) stratospheric heating,
York, New York, USA. (3) tropospheric cooling, and (4) perturbation of seasonal
6
Department of Geophysics, Niels Bohr Institute, University of temperatures (summer cooling and winter warming).
Copenhagen, Copenhagen, Denmark.
Megaeruptions or high-frequency events may speed up
Copyright 2005 by the American Geophysical Union. global climate changes acting as climate driving forces
0148-0227/05/2004JD005259 [Budner and Cole-Dai, 2003].
D06114 1 of 12
21562202d, 2005, D6, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2004JD005259 by CAPES, Wiley Online Library on [10/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
D06114 CASTELLANO ET AL.: HOLOCENE VOLCANIC RECORD AT DOME C D06114
easily distinguished from background values [Udisti et al.,

2000; Cole-Dai et al., 1997].
[6] The volcanic stratigraphy at Dome C (DC, East
Antarctica) was obtained from continuous sulfate measure-
ments carried out on the EDC96 ice core drilled in the
framework of the European Project for Ice Coring in
Antarctica (EPICA). This is the longest paleovolcanic
series from an Antarctic ice core. In a previous paper
[Castellano et al., 2004] we discussed the method used to
discriminate volcanic signatures from background values
and the eruption frequency recorded at DC for the last
45 kyr. Here we focus on the Holocene (0 –11.5 kyr BP)
volcanic record (first 360 m of the EDC96 ice core)
characterizing all the volcanic signatures (dating, maxi-
mum sulfate concentration, and sulfate volcanic flux). The
EDC96 volcanic data set can be used for interhemispheric
and intrahemispheric comparison with other ice core –
based volcanic records both to synchronize ice core
stratigraphies and to study depositional processes of vol-
canic products. In this paper we focus on the comparison
with previously published records from other Antarctic ice
cores in terms of event frequency and intensity of the
Figure 1. Map of Antarctica with location of cited drilling signatures. The spatial variability in the deposition of
sites. Abbreviations are as follows: BY, Byrd; DC, Dome C; single events over the Antarctic ice sheet was evaluated
DML, Dronning Maud Land; DY, Dyer Plateau; LD, Law by comparing normalized sulfate fluxes. Normalization
Dome; PR, Plateau Remote; SI, Siple Dome; SP, South was carried out by dividing volcanic fluxes by the Tam-
Pole; TD, Talos Dome. References are reported in the text. bora flux in different Antarctic sites. We discuss whether
changes in volcanic deposition and variability are possibly
linked to changes in transport pathways driven by climatic
variations.
[4] Volcanic sulfate aerosols and ash are irreversibly
deposited on polar ice sheets [Clausen et al., 1997; Delmas 2. Methods and Procedures
et al., 1992; Hammer et al., 1997; Langway et al., 1995; 2.1. Ice Core Sampling and Analysis
Zielinski et al., 1994]. Sulfate ice core stratigraphy can be [7] DC (75060S; 12324 E; 3233 m a.s.l.) (Figure 1,
used to evaluate the stratospheric load of sulfate and obtain showing all borehole locations cited in the text) was the site
information on the climatic impact of explosive eruptions. chosen in the framework of the EPICA program to obtain an
This is not an easy task. Indeed, the deposition of volcanic ice core of more than 3000 m, which will yield paleocli-
SO24 on ice sheets is affected by the geographic location of matic and paleoenvironmental information on about the last
sources, the pattern of tropospheric/stratospheric exchange, 900 kyr [European Project for Ice Coring in Antarctica
seasonal features of atmospheric circulation [Zielinski, (EPICA) Dome C 2001 – 2002 Science and Drilling Teams,
2000; Coffey, 1996], and depositional processes (such as 2002; EPICA Community Members, 2004]. Some details of
the relative contribution of dry and wet deposition, snow the drilling site and of ice core processing have been
accumulation rate, and snowfall frequency). Finally, post- published previously [Tabacco et al., 1998; Wolff et al.,
depositional effects such as wind scouring can affect 1999; EPICA Dome C 2001 –2002 Science and Drilling
records. To overcome these difficulties in the reconstruction Teams, 2002]. Here we present data from the first 360 m of
of volcanic history, the ice core database on volcanic the EDC96 ice core, covering the entire Holocene (0 –
eruptions must be enlarged as much as possible and should 11.5 kyr BP). Owing to the poor quality of the first 6 m
consider regional features of sites located in the same ice of ice core, the uppermost data were obtained by analyzing
sheet areas. Different transport pathways, also acting on a samples collected in a 7-m-deep snow pit dug during
regional scale (as pointed out by Udisti et al. [2004]), can the 2000 – 2001 field season about 3 km from the main
lead to different depositional fluxes for the same volcanic borehole.
eruption. [8] Ice core sulfate measurements were performed in the
[5] In many cores, volcanic stratigraphy was inferred by field by fast ion chromatography (FIC). FIC was carried out
continuous acidity records from ECM [Hammer et al., by coupling an ion-chromatographic method with a flow-
1980, 1997; Clausen et al., 1997] or dielectric profiling injection analysis apparatus [Udisti et al., 2000]. This
(DEP) [Barnes et al., 2002; Wolff et al., 1999]. However, method was set up to perform one analysis per minute of
the original acidic load can be partially or totally neutralized chloride, nitrate, and sulfate records in samples coming
in the atmosphere or after deposition by buffering effects from a melter device: ice core sections were continuously
(e.g., by dust and ammonia). In contrast, since sulfate is melted after removal of a few mm of ice from core ends and
irreversibly deposited on snow, records are insensitive to possible breaks [Röthlisberger et al., 2000]. The FIC
snow acidity changes; volcanic spikes are therefore more temporal resolution (yr sample1) was calculated by multi-
2 of 12
accumulation rates derived from dD measurements and

calibrated and confirmed by: (1) well-dated time horizons,
such as known volcanic eruptions over the last 700 years;
(2) matching between EDC96 and Vostok volcanic signa-
tures from 1200 to 7100 years BP used to transfer a dated
Vostok ice core sequence (through a link between 10Be in
ice and 14C) to EDC96; (3) the synchronization between the
CH4 event at the end of the Antarctic Cold Reversal (ACR)
and of the Greenland Younger Dryas (YD); (4) a sharp F
signal dated to about 17,500 BP (recovered for the first time
in the Byrd BS68 ice core by Hammer et al. [1997]); and
(5) the estimated depth and age of a 10Be peak (about 41 kyr
BP). The EDC1 age uncertainty for the Holocene is esti-
mated to be ±10 years from 0 to 700 years BP and as much
as ±200 years up to 10 kyr BP. For the first part of the core
(the last millennium), which contains a record of many
historically well-known volcanic events, we compared the
Figure 2. Temporal resolution of FIC data during the EPICA sulfate record with published ice core data sets,
Holocene. Despite some outliers (data corresponding to a paying particular attention to data from ice cores drilled at
greater number of years due to breaks in the core and sites where the annual snow accumulation rate allows
missing ice), the black smoothing curve shows that the stratigraphic dating (Byrd, Law Dome, Siple Dome, Dyer
temporal resolution for the Holocene ranges between about Plateau, and Talos Dome). Unambiguous and well-dated
1 and 1.5 years. volcanic signatures (Agung-1964, Tambora-1816, Huayna-
putina-1601, Kuwae-1460 and an unknown event dated
1259) were used to date the last 1000 years more accurately
plying the measurement rate (1 min sample1) with the through polynomial interpolation of the five temporal hori-
melting rate (4.0 – 7.0 cm min1) and dividing by the zons; this was then used to date other volcanic signatures.
accumulation rate (2.50 – 2.56 cm yr1 water equivalent The new dating was compared with EDC1. Age differences
(w.e.) [Schwander et al., 2001]. The resolution is nearly
annual in the Holocene, with values ranging from 1 to 1.5 yr
sample1. Figure 2 shows the time resolution of FIC mea-
surements versus age. The low-resolution spikes (up to 5 yr
sample1) are numerically insignificant (around 2%) com-
pared to the total sample number and are caused by ice loss
during the decontamination of ice core bar extremities and
possible breaks. The scarce relevance of the low-resolution
spikes is demonstrated by the trend of the 200-year smoothed
curve (dark line). A rather constant time resolution for
superficial or deep ice core sections was obtained by melting
them at different rates (7.0 – 4.0 cm min1) according to
density.
[9] The accuracy of the FIC measurements was assessed
by comparing chloride, nitrate, and sulfate FIC profiles with
data obtained by classical ion chromatographic (IC) analysis
of discontinuous samples at a similar resolution [Littot et
al., 2002]. Figure 3 shows the FIC and IC sulfate profiles
around three volcanic spikes recorded at about 38 and 39 m
(Figure 3a) and at 353 m (Figure 3b). Although background
trends and peak shapes fit very well, there are some differ-
ences in absolute concentrations, especially for the 353 m
spike (the FIC and IC Y scales were shifted for better
comparison). For all the compared peaks, a mean standard
deviation of about 10% was found between IC and FIC
measurements. This uncertainty is very close to the internal
reproducibility (5 – 10%) of FIC and IC. Furthermore, one
must bear in mind that volcanic depositions are revealed by
high sulfate concentrations, for which reproducibility is
better than 5%.
Figure 3. Comparison between FIC and discontinuous IC
2.2. Dating of Volcanic Events measurements in two selected sections of the EDC96 ice
[10] The EDC96 timescale (EDC1) [Schwander et al., core showing three volcanic signatures dated (a) 1259 and
2001] was obtained using a simple ice-flow model, with 1230 AD and (b) 9212 BC.
3 of 12
Table 1. List of Events Recorded in the DC Ice Core for the Last Millenniuma
Year AD Year AD Maximum SO42
Event Depth, m (EDC1 Model) (This Work) DAge Concentration, mg l1 Volcanic Flux, kg km2 Volcanic Eruption Fev/Fb
A 0.64 - 1992 - 313.2 10.7 Pinatubo + Hudson 0.57
B 2.95 - 1964 - 362.5 8.4 Agung 0.71
1 8.00 1894 1887 7 140.0 3.1 Krakatua 0.24
2 8.35 1888 1881 7 289.4 9.3 Cotopaxi 0.72
3 9.54 1868 1861 7 138.8 4.1 Makjan-Cotopaxi-Fuego 0.60
4 12.34 1820 1816 4 605.7 39.3 Tambora 1.96
5 12.68 1814 1807 7 270.5 10.2 unknown 1809 0.71
6 15.36 1764 1758 6 173.6 4.5 Jorullo-Taal 0.87
7 18.62 1700 1696 4 184.6 8.8 Serua 1.07
8 19.70 1678 1675 3 141.8 5.3 Gamkonora 0.61
9 22.20 1625 1624 1 174.6 8.0 Subantarctic 0.61
10 23.20 1603 1601 2 193.6 13.4 Huaynaputina 0.68
11 28.90 1473 1480 7 182.4 9.5 ? 0.47
12 29.77 1453 1460 7 398.6 31.7 Kuwae 1.23
13 34.55 1341 1347 6 211.4 10.4 ? 0.69
14 36.96 1283 1288 5 257.5 22.4 ? 0.93
15 37.64 1267 1271 4 304.2 20.5 ? 1.18
16 38.12 1255 1259 4 637.1 60.4 ? 2.46
17 39.22 1228 1230 2 336.9 25.2 ? 1.28
18 40.79 1189 1190 1 226.8 18.0 ? 0.85
19 41.52 1171 1170 1 310.8 20.8 ? 1.14
a
The age of events was calculated using both the EDC1 dating model [Schwander et al., 2001] and a fitting of major volcanic events (see text for further
details). Volcanic fluxes and Fev/Fb values were calculated as explained in the text. The attribution of sources was made referring to previously published
works, mainly Simkin and Siebert [1994].
(reported in Table 1) are always less than 10 years, Delmas, 1987], the effect of variable snow accumulation
confirming the error estimation of Schwander et al. [2001]. rates can be corrected by considering the sulfate flux
(calculated by multiplying concentration by the accumula-
2.3. Nonvolcanic Background and Detection of Events tion rate expressed in water equivalents) instead of concen-
[11] In order to reconstruct a reliable record of volcanic
events recorded in the EDC96 ice core, it was necessary to
evaluate the nonvolcanic sulfate background and set a
threshold above which spikes can be attributed to volcanic
deposition. The discrimination method used for the EDC96
ice core has already been discussed [Castellano et al.,
2004]; an outline of the method is here reported.
[12] Nonvolcanic sulfate contributions in Antarctic snow
come from sea-salt spray, crustal erosion, and atmospheric
oxidation of biogenic dimethylsulphide (DMS) [Saltzman,
1995; Prospero et al., 1991]. The sea-salt sulfate contribu-
tion to total sulfate budget at DC in the Holocene, evaluated
using Na+ as a specific marker [Legrand and Delmas, 1988;
Röthlisberger et al., 2002], is less than 5%. The Holocene
crustal contribution, calculated by non-sea-salt Ca2+ as a
continental dust marker [Röthlisberger et al., 2002], is even
lower (<0.05%). Since these contributions are of the same
order of measurement reproducibility, we did not correct
sulfate concentrations and will not distinguish between
sulfate and non-sea-salt sulfate in the following discussion.
[13] Discrimination between background and volcanic
peaks was achieved by setting a variable threshold as
follows [Castellano et al., 2004]:
[14] 1. The distribution of sulfate concentrations and
fluxes in the ice is more appropriately described by a
lognormal instead of Gaussian distribution.
[15] 2. The variability of sulfate background concentra-
tions (about 93 ± 34 mg l1) is relatively large even in the
Holocene (Figure 4a). Such variability could be caused by Figure 4. (a) Holocene profiles of sulfate concentrations
slight changes in the accumulation rate and/or in biogenic and (b) the logarithm (neperian) of sulfate flux. The bold
inputs, in turn due to changes in source intensity or in line in Figure 4b is the threshold curve used to
transport pathways of marine air masses. Since DC sulfate distinguish between volcanic spikes and background
depositions are dominated by dry deposition [Legrand and sulfate concentrations.
4 of 12
Table 2. List of Events Recorded in the DC Ice Core for the Period 1000 AD to 9500 BCa
Year AD Age, years BP Maximum SO42
Event Depth, m (EDC1 Model) (EDC1 Model) Concentration, mg l1 Volcanic Flux, kg km2 Fev/Fb
20 51.54 926 1024 173.5 7.8 0.52
21 57.83 765 1185 179.3 7.9 0.40
22 60.42 699 1251 292.6 33.0 1.18
23 64.20 601 1349 324.4 25.1 0.83
24 69.15 467 1483 187.1 7.3 0.35
25 73.58 340 1610 343.4 24.2 0.98
26 74.88 302 1648 200.5 12.0 0.46
27 75.79 276 1674 237.3 16.5 0.62
28 76.56 253 1697 165.9 10.1 0.62
29 76.85 245 1705 168.8 8.7 0.54
30 78.02 210 1740 180.9 6.8 0.45
31 79.49 166 1784 183.9 10.9 0.40
32 85.10 2 1952 201.2 11.4 0.57
33 91.49 200 2150 177.5 14.0 0.60
34 96.04 348 2298 206.0 8.8 0.49
35 97.12 384 2334 497.3 32.2 1.36
36 100.04 482 2432 155.8 5.8 0.52
37 126.57 1351 3301 201.9 6.6 0.34
38 126.85 1361 3311 243.2 7.8 0.40
39 130.78 1496 3446 352.4 26.6 1.36
40 133.84 1598 3548 239.9 17.7 0.95
41 135.60 1657 3607 228.7 11.5 0.71
42 140.01 1804 3754 160.9 5.4 0.41
43 140.30 1814 3764 193.6 8.6 0.64
44 141.56 1858 3808 232.8 13.4 0.74
45 144.93 1975 3925 324.2 28.1 1.06
46 153.91 2288 4238 212.5 11.0 0.68
47 156.78 2389 4339 187.9 8.4 0.47
48 160.18 2509 4459 188.3 10.6 0.61
49 163.16 2615 4565 216.7 7.8 0.55
50 164.16 2651 4601 160.4 7.2 0.58
51 169.44 2842 4792 282.5 17.6 0.89
52 175.24 3049 4999 145.5 10.6 0.73
53 176.14 3081 5031 227.3 14.3 1.15
54 190.26 3593 5543 152.9 6.9 0.58
55 198.76 3892 5842 171.8 7.0 0.49
56 200.32 3949 5899 141.7 5.2 0.28
57 208.75 4255 6205 306.0 41.6 1.18
58 209.90 4297 6247 429.7 36.0 1.99
59 213.19 4420 6370 169.6 9.6 0.52
60 214.00 4450 6400 365.8 24.7 1.55
61 215.72 4514 6464 534.8 37.1 2.36
62 215.94 4523 6473 803.0 61.5 3.27
63 219.00 4639 6589 292.6 16.9 0.96
64 232.75 5161 7111 266.7 15.3 0.69
65 234.45 5228 7178 151.6 5.7 0.49
66 234.74 5239 7189 141.4 5.0 0.43
67 238.99 5389 7339 141.1 7.8 0.57
68 243.33 5538 7488 536.1 52.5 1.29
69 249.29 5738 7688 178.0 6.6 0.55
70 251.40 5809 7759 286.1 18.3 0.98
71 253.93 5896 7846 163.1 3.8 0.64
72 259.44 6094 8044 295.3 20.3 0.86
73 264.16 6265 8215 218.4 14.1 0.64
74 270.59 6490 8440 505.9 38.1 0.94
75 283.15 6919 8869 194.6 10.5 0.47
76 290.18 7155 9105 166.9 8.7 0.70
77 291.37 7194 9144 154.7 7.6 0.39
78 294.14 7283 9233 167.9 10.1 0.67
79 304.39 7609 9559 218.5 14.3 0.76
80 305.70 7650 9600 218.6 19.8 1.24
81 311.08 7822 9772 152.5 6.7 0.32
82 312.86 7878 9828 411.7 33.7 1.63
83 317.52 8024 9974 175.7 9.3 0.40
84 321.15 8141 10091 226.9 10.2 0.78
85 322.05 8171 10121 218.4 14.7 0.57
86 325.31 8277 10227 176.6 9.4 0.79
87 338.68 8719 10669 212.0 9.1 0.78
88 339.40 8743 10693 128.9 5.2 0.25
89 342.82 8861 10811 195.9 11.4 0.66
90 343.74 8893 10843 177.5 10.1 0.47
91 344.97 8935 10885 217.2 24.4 0.81
5 of 12
Table 2. (continued)
Year AD Age, years BP Maximum SO42
Event Depth, m (EDC1 Model) (EDC1 Model) Concentration, mg l1 Volcanic Flux, kg km2 Fev/Fb
92 350.08 9112 11062 164.9 6.4 0.41
93 350.80 9136 11086 156.8 8.0 0.45
94 353.00 9212 11162 383.4 41.0 1.42
a
The age of events was calculated using the EDC1 dating model [Schwander et al., 2001] and BP means before 1950. Volcanic fluxes and Fev/Fb values
were calculated as explained in the text.
trations. Sulfate flux calculation was carried out using mixing processes affect concentrations but not fluxes
highly resolved accumulation rates derived from dD mea- [Barnes et al., 2003]. However, wind redistribution could
surements (resolution of 55 cm, corresponding to about affect both, even leading to the complete removal of snow
15 years in the Holocene) [Schwander et al., 2001] coupled layers recording a volcanic event.
with a simple one-dimensional thinning model (dome [20] A reliable evaluation of sulfate volcanic fluxes
location). The residual variability in the sulfate flux back- allows a more accurate estimation of the relative contribu-
ground is described by a curve obtained by a 3% weighted tion of two possible climatic forcings linked to the global
fit smoothing procedure on the log-flux profile. This sulfur cycle acting on different atmospheric scales: the
smoothing function fits a curve to data using the locally stratosphere (volcanic inputs) and troposphere (biogenic
weighted (in this case 3% of total data) least squared error emissions). Tables 1 and 2 show ratios between volcanic
method. The result is to plot a best fit smooth curve through and background sulfate fluxes (Fev/Fb) for each volcanic
the center of data. This is an extremely robust fitting event. Volcanic events are characterized by concentration
technique nearly insensitive to outliers. The final curve spikes, but their deposition lasts for a short time; as a
used as the threshold between volcanic spikes and sulfate consequence, in most cases, volcanic fluxes are lower than
background concentrations was obtained by adding 2 log s background values (mainly of biogenic origin) deposited in
to the log-flux smoothed profile, as shown in Figure 4b. the same interval of time. Only a few major events show a
[16] 3. In order to reject outliers leading to false positives, Fev/Fb ratio >1. For example, the Tambora eruption pro-
we decided to subjectively evaluate volcanic signature duced a total sulfate deposition at DC only about 2 times
peaks consisting of only one sample above the threshold. higher than the contemporaneous background contribution,
[17] We thus detected 96 volcanic events in the first although its concentration peak is about 6 times higher. The
360 m of the EDC96 core, which covers the Holocene cumulative sulfate flux calculation may be used to evaluate
period (0– 11.5 kyr BP); volcanic signatures are listed in the overall contribution of explosive volcanic activity to the
Tables 1 and 2 along with glaciological parameters (depth, total sulfate budget at DC. The high-resolution EDC96
age, and flux). sulfate record allows a reliable estimation of the total
[18] Different peak detection algorithms have been pre- amount of background sulfate deposition during the entire
sented in previous works [e.g., Cole-Dai et al., 1997; Holocene, about 2.5 104 kg km2. Cumulative explosive
Zielinski et al., 1994]. When comparing volcanic records volcanism contributes about 1.5 103 kg km2, corresponding
from different ice cores, one must bear in mind that the to about 6% of the total sulfate budget.
number of detected events closely depends on the selected
detection method. Nevertheless, differences among the var-
3. Results and Discussion
ious methods are supposed to affect only minor volcanic
signatures just above or below the threshold, while major 3.1. Frequency of Volcanic Events
global eruptions (characterized by higher climatic impact) [21] Figure 5 shows the number of volcanic events per
are insensitive to differences in detection strategy, assuming millennium recorded in the EDC96 ice core (black bars)
that other factors such as regional variability and postdepo- during the Holocene. Volcanic frequencies from other
sitional effects did not influence volcanic deposition. Antarctic and Greenland ice cores, covering at least
1000 years, are also plotted. The average eruption number
2.4. Volcanic Fluxes per millennium recorded at DC during the Holocene is 7.9,
[19] Accurate sulfate fluxes for each volcanic peak (inte- with oscillations between low- and high-frequency periods.
grated net depositional volcanic flux expressed in kg km2) In particular, the last millennium shows the highest eruption
were calculated by subtracting the sulfate background frequency in the whole record (21 eruptions), followed by
concentration from the total sulfate concentration in each 12 volcanic signatures in the 1000 – 0 AD period. As already
sample recording the specific volcanic event. The residual pointed out by Castellano et al. [2004], the last two
was multiplied by the sample length and density; the total millennia are characterized by the greatest number of
flux was calculated by summing single sample contributions volcanic signatures in the whole 45-kyr volcanic record of
[Cole-Dai and Mosley-Thompson, 1999]. The sulfate back- EDC96. This finding is unlikely due to glaciological effects,
ground flux for each volcanic spike was calculated as the such as different resolution or layer thinning; indeed, as
mean short-period background value of 5 samples before mentioned in the methodology section, the temporal reso-
and after the volcanic peak. EDC96 densities in the firn lution of FIC measurements was kept essentially constant
were measured directly. Since sulfate analysis was carried by adjusting the ice core melting rate to density variations,
out in a continuous way, sample lengths correspond to the and the thinning effect in the uppermost 360 m is negligible
mean melting rate of ice core sections in each depth interval. [Barnes et al., 2003]. The high frequency of signatures in
Postdepositional effects due to diffusion or snow-layer the last 2000 years confirms the Cole-Dai et al. [2000]
6 of 12
events recorded in the Greenland GISP2 ice core [Zielinski et

al., 1996], with a mean frequency of 33.3 volcanic events per
millennium during the Holocene (1000-year means ranging
from 18 to 47 eruptions), is much greater than in Antarctic
cores. As already discussed by Cole-Dai et al. [2000] and
Castellano et al. [2004], the difference is not surprising,
considering the relative closeness of Greenland to areas with
a high density of active volcanic systems (e.g., Iceland,
Kamchatka, etc.).
3.2. Comparison of the EDC96 Volcanic Profile With
Other Antarctic Ice Cores
[22] Comparing volcanic fluxes recorded in ice cores
from different sites is essential for reconstructing the spatial
distribution of net deposition from single events, for under-
standing the impact of changes in atmospheric transport,
Figure 5. Frequency of volcanic events in the Holocene, and for evaluating the intensity of past volcanic eruptions
expressed as number of events per millennium, for the DC, and their possible influence on global climate. Furthermore,
PR, PS1, BY, and GISP 2 ice cores. References are reported the synchronization of different ice cores using volcanic
in the text. temporal horizons helps improve ice core dating. We
present here a comparison of volcanic stratigraphies
recorded in different Antarctic ice cores and discuss differ-
ences in depositional fluxes.
measurements on the 4100-year ice core drilled at Plateau 3.2.1. Comparison Among Antarctic Volcanic
Remote (PR). This record lists 33 eruptions in the last Stratigraphies
2000 years, with a millennium distribution similar to that [23] Figure 6 compares the volcanic stratigraphy of the
observed at DC, 19 (last millennium) and 14 (penultimate last millennium (volcanic sulfate flux versus age) recorded
millennium). To minimize the effect of the threshold choice in the EDC96 ice core with that recorded in five other
on moderate volcanic signatures, we considered only Antarctic ice cores on which sulfate measurements were
depositions with fluxes higher than 20 kg km2 (this carried out: BY, SP (PS1 ice core), PR, SI, and DY (see
value was chosen by Cole-Dai et al. [2000] as the limit references in section 3.1). Data related to the EDC96
between large and medium events). For these major signatures are also reported in Table 1. As described in
signatures the two records give the same values: both PR section 2.2, the age of DC events was calculated by
and EDC96 ice cores show 10 and 3 large events recorded interpolating five historically well-known volcanic erup-
in the last 1000 years and in the previous millennium, tions. These events are recorded in all Antarctic ice cores
respectively. The high frequency of volcanic signatures covering the last 1000 years, and their signatures have been
recorded in the Antarctic ice sheet during the last millen- accurately dated in sites characterized by high annual snow
nium is confirmed by other shorter-interval ice core records: accumulation rates [e.g., Cole-Dai et al., 1997; Palmer et
18 events in 550 years at Siple Station (SI) [Cole-Dai et al., 2001] or by using other dating techniques (e.g., den-
al., 1997]; 15 events in 400 years at Dyer Plateau (DY) drochronology and historical reports) [Simkin and Siebert,
[Cole-Dai et al., 1997]; 23 events at the South Pole (SP- 1994]. Thanks to the closeness of temporal horizons, in this
PS1 ice core) [Delmas et al., 1992]; 25 events at Dronning section of ice core, event ages are likely to be more accurate
Maud Land (DML) [Karlöf et al., 2000]; and 25 events in than those obtained from the ice flow model (EDC1).
1150 years at Byrd (BY-NB89 ice core) [Langway et al., [24] Although most of the sulfate spikes in the EDC96 ice
1994]. The larger number of volcanic events recorded at SI core are also recorded in all other ice cores, there are some
and DY (Antarctic Peninsula) could be ascribed to the differences. The most relevant particularity is the number
higher temporal resolution of sulfate measurements (linked of peaks grouped in the so-called ‘‘1259 sequence’’
to the higher accumulation rate), which can potentially (13th century). This sequence, a series of events centered
permit to distinguish short time-spaced volcanic eruptions. on the highest signature dated 1259 AD, is also recorded in
In contrast, global events distributed by long-range trans- Greenland ice cores [Clausen et al., 1997]. Among the four
port (mainly via the stratosphere) are characterized by long Antarctic ice cores covering this time period (EDC96, BY,
(2 – 3 years) residence times of submicrometric sulfuric PS1, and PR), PR and BY show the complete series of five
aerosol; therefore resolutions close to one sample per year volcanic signatures, while PS1 only shows the three central
(as in all Antarctic ice cores here discussed) seem to assure events. In the EDC96 ice core the peak sulfate concentration
consistent results in event computation. Indeed, only very of the signature dated 1276-8 is lower than the threshold
high resolutions (such as those obtained in Greenland and at (see section 2.3) and was not included in Figure 6 and
coastal Antarctic sites) are really able to separate events Table 1. Finally, the highest peak of the ‘‘1259 sequence’’ in
which occurred only a few months apart. For this reason we the PR ice core is the 1277 event. With the exception of the
consider more relevant the relative closeness of SI and DY EDC96 1276-8 event (probably affected by wind scouring),
sites to active volcanic systems (e.g., South Sandwich Island these differences are unlikely due to the choice of different
and Deception Island), highlighting the importance of local detection methods because all signatures are well above the
to regional events. Indeed, Figure 5 shows that the number of threshold.
7 of 12
but synchronous signatures were found in other shallow firn

cores [Legrand and Delmas, 1987; Cole-Dai and Mosley-
Thompson, 1999]. In contrast, two medium-intensity events,
dated 1835-6 and 1593-5 and recorded in all the ice cores
reported in Figure 6 (with the only exception of the oldest
event in the BY core), are not present in the EDC96
stratigraphy. Nevertheless, these signatures show high
fluxes only in ice cores drilled in coastal areas (SI and
DY), while they are recorded as minor events in the PR core
drilled on the East Antarctic Plateau. As discussed in the
next section, differences in net depositional fluxes and the
loss of minor peaks could be due to changes in atmospheric
circulation processes and to local glaciological features
(wind redistribution, relative contribution of wet-dry depo-
sition, accumulation rates, and frequency and timing of
snowfalls), particularly relevant in low-accumulation sites.
[26] Besides DC, the only other Antarctic volcanic stra-
tigraphy extending back to more than 1000 AD comes from
the PR ice core, which covers the last 4100 years. This
volcanic profile is plotted in Figure 7a together with EDC96
stratigraphy (Figure 7b). Although in agreement over the
last 1000 years, the two profiles look quite different in
the older period. In particular, there is no correspondence
between periods characterized by lack of volcanic signals in
the two cores. In the period 500– 1350 BC the DC ice core
records no event, whereas the PR core contains 11 signa-
tures, including three large events. However, in the period
1500– 2000 BC, EDC96 recorded seven events, with two
signatures showing sulfate fluxes higher than 20 kg km2;
in the same temporal range, no volcanic signatures are
recorded in the PR ice core. Another six events are recorded
in the PR core in the 2000 – 2200 BC period without
corresponding signatures in the EDC96 ice core. Differ-
ences are also seen in sections recording a similar numbers
of volcanic events (e.g., in the penultimate millennium),
making it difficult to compare major signatures also. Differ-
ences between PR and EDC96 volcanic signatures cannot
be attributed to the poor quality of ice cores. The compared
sections do not belong to the ‘‘brittle zone,’’ where fragile
ice causes frequent breaks. As for ice-bar quality, previous
works do not show evidence of poor quality in the PR core
[e.g., Cole-Dai et al., 2000]. Breaks in EDC96 affected only
2% of the ice core (see section 2.1), and the ice lost during
decontamination of bar extremities before melting involved
only 1 – 3 mm. In addition, there were no relevant differ-
ences in the number of volcanic peaks in sulfate and DEP or
ECM stratigraphies [Udisti et al., 2000, 2004] (the latter
two were carried out on uncut, 2.2-m-long ice core bars).
Besides changes in atmospheric patterns and glaciological
features, relative shifts in the timescale of the two ice cores
Figure 6. Comparison of volcanic profiles (volcanic flux may account for the observed differences. Indeed, these
versus age) for the last millennium from sulfate measure- cores were drilled in the East Antarctic Plateau, where low
ments in the DC, PR, BY, PS1, SI, and DY ice cores. accumulation rates occur, and were dated with nonstrati-
References are reported in the text. In the PR plot, 133.4 is graphic methods. An accurate peak-to-peak comparison
the out-of-scale flux value of the Kuwae signature. Shaded between the volcanic signatures recorded at DC and VK
areas indicate periods not covered by ice cores. in the last 45 kyr [Udisti et al., 2004] revealed differences
of up to 3500 years, highlighting the importance of
volcanic stratigraphy in synchronizing different ice cores.
[25] Minor signatures in cores are generally in agreement; A preliminary peak-to-peak comparison between volcanic
only three minor volcanic events recorded in the EDC96 signatures in the EDC96 and EDML ice cores (unpub-
core are unmatched in the other records (peaks 3, 6, and 11 lished data) revealed the good fit between volcanic sig-
in Table 1, dated 1861, 1758, and 1480 AD, respectively), natures recorded in the period 1000 – 7000 years BP in
8 of 12
Kuwae signature (1460 AD) is a major example of depo-

sitional spatial variability. This signature is by far the most
intense in the PR volcanic record, while in the other cores
the 1259 event always shows the highest sulfate deposition.
Cole-Dai et al. [2000] tentatively explained the lower
sulfate flux of the ‘‘1259 event’’ in the PR core as possibly
due to a partial loss of sulfate deposition by wind scouring.
However, in the Talos Dome (TD) ice core (see Stenni et al.
[2002]; not reported in Figure 6), Kuwae is again by far the
highest volcanic signature in the last 800 years. It therefore
seems probable that differences in the Kuwae signature can
be ascribed to a real spatial variability in its depositional
fluxes.
[28] In order to objectively compare volcanic flux vari-
ability among the different sites, ice core signatures were
normalized with respect to the 1816 AD Tambora eruption,
as suggested by Cole-Dai et al. [1997], thereby correcting
for regional features. Indeed, with few exceptions such as
the Kuwae event, nonnormalized fluxes are always higher at
coastal sites, where high accumulation rates are observed
and wet deposition strongly contributes to sulfate atmo-
spheric scavenging (see SI and DY in Figure 6). Normal-
ization was carried out by dividing the sulfate flux of each
volcanic event (F) by the Tambora flux (FT) in each core.
Assuming that depositional and postdepositional processes
at individual sites did not differ significantly over a
relatively short interval of time (the last 1000 years),
Tambora-normalized fluxes should be less dependent on
the glaciological features (accumulation rate, snowfall
frequency and annual distribution, wind redistribution,
etc.) of sampling sites [Cole-Dai et al., 1997]. Normalized
fluxes (F/FT) measured in several ice and firn cores are
Figure 7. Comparison of volcanic profiles (volcanic flux reported in Figure 8 along with corresponding standard
versus age) from sulfate measurements in the (a) PR and deviations (s, vertical bars in Figure 8b). In addition to
(b) DC ice cores for the period 1000 AD to 2500 BC and in the previously listed ice cores, the figure also contains the
the (c) DML and (d) DC ice cores for the period 1450 – following Antarctic firn cores: PS14 (South Pole; Delmas
2200 BC. In Figures 7c and 7d, age is calculated by the et al. [1992]), SP96 (South Pole; Cole-Dai and Mosley-
EDC1 dating model (see section 2.2). Thompson [1999]) and TD (Talos Dome; Proposito et al.
[2003]). Tambora-normalized values (Figure 8a) are con-
stantly lower than 1 for all events which occurred after
sites located at opposite sides of Antarctica (data process- 1500 AD, while several events show higher values in the
ing is still in progress). For example, Figures 7c and 7d early last millennium. The variability of the F/FT ratio
compare EDC96 and EDML in the 1450– 2200 BC period, among the different cores (standard deviation in Figure 8b)
when maximum differences where found between EDC96 shows a similar pattern: very low values were observed for
and PR stratigraphies. Indeed, five of six major signatures the most recent events and higher values for the oldest
are contemporaneous in EDC96 and EDML profiles, signatures. Consistent patterns (not shown here) were
sustaining the reliability of the DC record. When EDML obtained through the use of signatures than Tambora as
is finally dated (stratigraphic dating by counting annual normalization reference; this evidence seems to exclude the
layers), the absolute dating of the EDC96 volcanic sig- possibility that F/FT ratio scattering is partly due to an
natures will be assessed. In any case, these comparisons anomalous Tambora signature in some cores. Furthermore,
clearly highlight the need to increase the number of ice the higher variability in the older period does not seem to be
core based volcanic stratigraphies in order to overcome the imputable to higher ratio values as the 1460 event might
lack of single ice cores signatures. suggest. Indeed, events with the same ratios occurred in the
3.2.2. Comparison of Depositional Fluxes 1400 – 1200 AD period, showing a significantly higher
[27] Volcanic sulfate fluxes in the six ice cores (Figure 6) absolute standard deviation with respect to events which
differ considerably over the last 1000 years. As mentioned occurred in the more recent period. In addition, when a
earlier, several factors can account for differences, particu- volcanic event other than Tambora is considered as refer-
larly regional features (e.g., variability in the atmospheric ence for normalization, the Tambora/reference ratios show a
circulation, also related to the period of the year where the low standard deviation in spite of the fact that Tambora is
eruption happens) [Robock, 2000] and glaciological effects the largest signature in the last 500 years. Figure 8b shows
such as wind redistribution processes which are very that normalized ratios s < 0.1 are usual for events that
efficient at low accumulation sites [Petit et al., 1982]. The occurred after 1500 AD. Most volcanic events in the early
9 of 12
periods [Krinner and Genthon, 1998], we support the

hypothesis that the pattern of volcanic deposition intensity
and geographical variability could reflect a warmer climate
of Antarctica in the early last millennium. The re-establish-
ment of colder conditions, starting in about 1500 AD,
reduced the variability of volcanic depositions. This warm/
cold step could be like a Medieval Climate Optimum –like to
Little Ice Age – like transition. Medieval Warming is well
defined in the Northern Hemisphere by several proxies [e.g.,
O’Brien et al., 1995; Wilson et al., 2000], while no clear
evidence has been reported for Austral regions [Dahl-Jensen
et al., 1999]. In a recent paper, Goosse et al. [2004, and
references therein] report evidence from dD and d18O mea-
surements on Antarctic ice cores in support of a Medieval
Warming –like period in the Southern Hemisphere, delayed
by about 150 years with respect to Northern Hemisphere
Medieval Warming. We postulate that changes in the extent
and intra-Antarctic variability of volcanic depositional
fluxes may have been consequences of the establishment
of a Medieval Warming– like period that lasted until about
1500 AD. The re-establishment of mean Holocene climate
conditions (or, perhaps of Little Ice Age – like conditions)
increased the insulation of the Antarctic continent by
increasing the intensity and persistence of the tropospheric
and stratospheric polar vortex. These climatic conditions
could have forced air masses containing sulfuric aerosol to
follow longer pathways around Antarctica, thereby leading
to more homogeneous deposition over the continent.
Figure 8. (a) The ratios between fluxes of events recorded [29] Such changes in atmospheric transport processes
in at least 4 of 9 Antarctic ice cores (PR, SI, DY, PS1, BY, could be similar (but to a lower extent) to that proposed
DC, PS96, PS14, and TD) and Tambora (used as reference by Delmonte et al. [2002] and Udisti et al. [2004] to explain
event) are reported against age for the last millennium. variations in dust deposition and accumulation rates at DC
(b) The standard deviation for the considered events is during the different climatic conditions in the last 45 kyr. In
indicated. References are reported in the text. the coldest periods (especially the Last Glacial Maximum)
the higher intensity and persistence of the polar vortex
prevented the direct advection of air masses to central areas
last millennium (1000 –1500 AD), especially the Kuwae of Antarctica, thereby sorting the size of atmospheric
eruption (F/FT = 2.38 ± 2.10), show higher variability. Why particles and homogenizing accumulation rates.
do events which occurred in the 12th – 16th centuries show
large F/FT ratios and higher variability? Of course, the higher
Tambora-normalized signatures in the older period could be 4. Conclusions
due to larger volcanic emissions [Budner and Cole-Dai, [30] The complete volcanic history of the Holocene was
2003]; they could have been more intense, for example, than reconstructed using the EDC96 high-resolution sulfate
some large modern volcanic events such as the Agung record. A total of 96 eruptions were identified in the last
(1964) and Pinatubo (1991) eruptions, but this does 11,500 years, with a mean of 7.9 events per millennium.
not account for the larger variability. Local glaciological The last 2000 years were characterized by enhanced volca-
features, such as wind redistribution, changes in accumula- nic activity (21 events in the last millennium and 12 events
tion rate, and variations in the frequency and timing of in the penultimate millennium).
snowfalls, could have affected the observed depositional [31] The comparison between EDC96 and other volcanic
variations. Considering that such effects should have been records from Antarctic ice cores highlighted some implica-
minimized by the adopted normalization procedure, we tions in the use of ice core data for reconstructing volcanic
postulate that changes in atmospheric circulation driven by history. Our study revealed two major constraints: (1) for
climate forcing on regional to global scales around Antarc- the oldest sections of the cores (4000 – 2000 BP), where
tica could have affected depositional processes in the 1000 – dating of ice layers is not straightforward, reliable synchro-
1500 AD period. As suggested by Budner and Cole-Dai nization of ice core timescales must be performed before
[2003], the Antarctic polar vortex is involved in the distri- sulfate signatures belonging to the same volcanic event can
bution of stratospheric volcanic aerosols over the continent. be compared; (2) for more recent periods, the comparison
Assuming that the intensity and persistence of the polar among signatures of single volcanic events recorded in ice
vortex (both in the troposphere and stratosphere) mainly cores drilled in different Antarctic sites reveals the high
affect the penetration of air masses to inland Antarctica, variability in sulfate depositional fluxes, even when these
isolating the continental area during cold periods and facil- are normalized with respect to a well characterized eruption
itating the advection of peripheral air masses during warm (e.g., Tambora event, 1816 AD).
10 of 12
[32] Regional features (accumulation rate, frequency and Dahl-Jensen, D., V. I. Morgan, and A. Elcheikh (1999), Monte Carlo in-
annual distribution of snowfalls, wind redistribution) and verse modelling of the Law Dome (Antarctica) temperature profile, Ann.
Glaciol., 29, 145 – 150.
changes in atmospheric transport considerably affect the Delmas, R. J., S. Kirchner, J. M. Palais, and J. R. Petit (1992), 1000 years of
extent and distribution of volcanic deposition on the ice explosive volcanism recorded at the South Pole, Tellus, Ser. B, 44, 335 –
sheet. A Medieval Warming – like period could therefore be 350.
Delmonte, B., J. R. Petit, and V. Maggi (2002), Glacial to Holocene im-
responsible for the observed increase in volcanic flux extent plications of the new 27,000 year dust record from the EPICA Dome C
and dispersion. (East Antarctica) ice core, Clim. Dyn., 18, 647 – 660.
[33] The atmosphere-to-snow exchange process models European Project for Ice Coring in Antarctica (EPICA) Community Mem-
bers (2004)Eight glacial cycles from an Antarctic ice core, Nature, 429,
(formulated in the past based on total beta activity data from 623 – 628.
stratospheric fallout after thermonuclear tests; Clausen and European Project for Ice Coring in Antarctica (EPICA) Dome C 2001 –
Hammer [1988]) have to be improved through the intro- 2002 Science and Drilling Teams (2002), Extending the ice core record
duction of variables linked to changes in atmospheric beyond half a million years, Eos Trans. AGU, 83, 45, 509, 517.
Goosse, H., V. Masson-Delmotte, H. Renssen, M. Delmotte, T. Fichefet,
transport processes. This would improve the reliability of V. Morgan, T. van Ommen, B. K. Khim, and B. Stenni (2004), A late
estimates of the impact of past volcanism on climate, based medieval warm period in the Southern Ocean as a delayed response to
on sulfate fluxes in ice cores. external forcing?, Geophys. Res. Lett., 31, L06203, doi:10.1029/
2003GL019140.
[34] In this context, high-resolution sulfate profiles from Hammer, C. U., H. B. Clausen, and W. Dansgaard (1980), Greenland ice
the two EPICA deep ice cores at Dome C (EDC; 3200 m in sheet evidence of post-glacial volcanism and its climatic impact, Nature,
the 2002/2003 campaign) and Dronning Maud Land 288, 230 – 235.
Hammer, C. U., H. B. Clausen, and C. C. Langway Jr. (1997), 50,000 years
(EDML; 2500 m in the 2003/2004 campaign) and from a of recorded global volcanism, Clim. Change, 35, 1 – 15.
new Italian-French deep drilling project at Talos Dome Hofmann, D., J. Barnes, E. Dutton, T. Deshler, H. Jäger, R. Keen, and
(Ross Sea Region, to begin in the 2004/2005 austral M. Osborn (2003), Surface-based observations of volcanic emissions to
summer) will increase the number and temporal extension the stratosphere, in Volcanism and the Earth’s Atmosphere, Geophys.
Monogr. Ser., vol. 139, edited by A. Robock and C. Oppenheimer,
of Antarctic volcanic stratigraphies, thereby providing new pp. 57 – 73, AGU, Washington, D. C.
tools for investigating the relationship between volcanism Karlöf, L., et al. (2000), A 1500 year record of accumulation at Amundse-
and climate. nisen western Dronning Maud Land, Antarctica, derived from electrical
and radioactive measurements on a 120 m ice core, J. Geophys. Res.,
105(D10), 12,471 – 12,483.
[35] Acknowledgments. This work is a contribution to the ‘‘European Krinner, G., and C. Genthon (1998), GCM simulations of the Last Glacial
Project for Ice Coring in Antarctica’’ (EPICA), a joint ESF (European Maximum surface climate of Greenland and Antarctica, Clim. Dyn., 14,
Science Foundation)/EC scientific program, funded by the European 741 – 758.
Commission and by national contributions from Belgium, Denmark, Langway, C. C., Jr., K. Osada, H. B. Clausen, C. U. Hammer, H. Shoji, and
France, Germany, Italy, the Netherlands, Norway, Sweden, Switzerland A. Mitani (1994), New chemical stratigraphy over the last millennium for
and the United Kingdom. This is EPICA publication 112. Byrd Station, Antarctica, Tellus, Ser. B, 46(1), 40 – 51.
Langway, C. C., Jr., K. Osada, H. B. Clausen, C. U. Hammer, and H. Shoji
(1995), A 10 century comparison of prominent bipolar volcanic events in
References ice cores, J. Geophys. Res., 100, 16,241 – 16,247.
Barnes, P. R. F., E. W. Wolff, R. Mulvaney, R. Udisti, E. Castellano, Legrand, M., and R. J. Delmas (1987), A 220-year continuous record of
R. Röthlisberger, and J.-P. Steffensen (2002), Effect of density on elec- volcanic H2SO4 in the Antarctic ice sheet, Nature, 327, 671 – 676.
trical conductivity of chemically laden polar ice, J. Geophys. Res., Legrand, M., and R. J. Delmas (1988), Soluble impurities in four Antarctic
107(B2), 2029, doi:10.1029/2000JB000080. ice cores over the last 30,000 years, Ann. Glaciol., 10, 116 – 120.
Barnes, P. R. F., E. W. Wolff, H. M. Mader, R. Udisti, E. Castellano, and Littot, G. C., R. Mulvaney, R. Röthlisberger, R. Udisti, E. W. Wolff,
R. Röthlisberger (2003), Evolution of chemical peak shapes in the E. Castellano, M. De Angelis, M. Hansson, S. Sommer, and J. P.
Dome C, Antarctica, ice core, J. Geophys. Res., 108(D3), 4126, Steffensen (2002), Comparison of analytical methods used for mea-
doi:10.1029/2002JD002538. suring major ions in the EPICA-Dome C (Antarctica) ice core, Ann.
Budner, D., and J. Cole-Dai (2003), The number and magnitude of large Glaciol., 35, 299 – 305.
explosive volcanic eruptions between 904 and 1865 A. D.: Quantitative O’Brien, S. R., P. A. Mayewski, L. D. Meeker, D. A. Meese, M. S. Twickler,
evidence from a new South Pole ice core, in Volcanism and the Earth’s and S. I. Whitlow (1995), Complexity of Holocene climate as recon-
Atmosphere, Geophys. Monogr. Ser., vol. 139, edited by A. Robock and structed from a Greenland ice core, Science, 270, 1962 – 1964.
C. Oppenheimer, pp. 165 – 176, AGU, Washington, D. C. Palmer, A. S., T. D. van Ommen, M. A. J. Curran, V. Morgan, J. M.
Castellano, E., S. Becagli, J. Jouzel, A. Migliori, M. Severi, J. P. Steffensen, Souney, and P. A. Mayewski (2001), High-precision dating of volcanic
R. Traversi, and R. Udisti (2004), Volcanic eruption frequency in the last events (A.D. 1301 – 1995) using ice cores from Law Dome, Antarctica,
45 kyr as recorded in EPICA-Dome C ice core (East Antarctica) and J. Geophys. Res., 106(D22), 28,089 – 28,095.
its relationship to climate changes, Global Planet. Change, 42, 195 – Petit, J. R., J. Jouzel, M. Pourchet, and L. Merlivat (1982), A detailed study
205. of snow accumulation and stable isotope content in Dome C (Antarctica),
Clausen, H. B., and C. U. Hammer (1988), The Laki and Tambora eruptions J. Geophys. Res., 87(C6), 4301 – 4308.
as revealed in Greenland ice cores from 11 locations, Ann. Glaciol., 10, Proposito, M., R. Gragnani, O. Flora, L. Genoni, and M. Frezzotti (2003),
16 – 22. Volcanic record at GPS1, GPS2, 31DPT and M2 stations along the
Clausen, H. B., C. U. Hammer, C. S. Hvidberg, D. Dahl-Jensen, J. P. 1998 – 99 ITASE traverse (East Antarctica), Terra Antarct. Rep., 8, 61 –
Steffensen, J. Kipfstuhl, and M. Legrand (1997), A comparison of the 66.
volcanic records over the past 4000 years from the Greenland Ice Core Prospero, J. M., D. L. Savoie, E. S. Saltzman, and R. Larsen (1991), Impact
Project and Dye 3 Greenland ice cores, J. Geophys. Res., 102(C12), of oceanic sources of biogenic sulphur on sulfate aerosol concentrations
26,707 – 26,723. at Mawson, Antarctica, Nature, 350, 221 – 223.
Coffey, M. T. (1996), Observations of the impact of volcanic activity on Rampino, M. R., and S. Self (1982), Historic eruptions of Tambora (1815),
stratospheric chemistry, J. Geophys. Res., 101(D3), 6767 – 6780. Krakatau (1883), and Agung (1963), their stratospheric aerosols, and
Cole-Dai, J., and E. Mosley-Thompson (1999), The Pinatubo eruption in climatic impact, Quat. Res., 18, 127 – 143.
South Pole snow and its potential value to ice-core paleovolcanic records, Robock, A. (2000), Volcanic eruption and climate, Rev. Geophys., 38, 191 –
Ann. Glaciol., 29, 99 – 105. 219.
Cole-Dai, J., E. Mosley-Thompson, and L. G. Thompson (1997), Annually Röthlisberger, R., M. Bigler, M. A. Hutterli, S. Sommer, B. Stauffer, H. G.
resolved southern hemisphere volcanic history from two Antarctic ice Junghans, and D. Wagenbach (2000), Technique for continuous high-
cores, J. Geophys. Res., 102(D14), 16,761 – 16,771. resolution analysis of trace substances in firn and ice cores, Environ.
Cole-Dai, J., E. Mosley-Thompson, S. P. Wight, and L. G. Thompson Sci. Technol., 34, 338 – 342.
(2000), A 4100-year record of explosive volcanism from an East Antarc- Röthlisberger, R., R. Mulvaney, E. W. Wolff, M. A. Hutterli, M. Bigler,
tica ice core, J. Geophys. Res., 105(D19), 24,431 – 24,441. S. Sommer, and J. Jouzel (2002), Dust and sea salt variability in central
11 of 12
East Antarctica (Dome C) over the last 45 kyrs and its implications for Wolff, E. W., I. Basile, J. R. Petit, and J. Schwander (1999), Comparison of
southern high-latitude climate, Geophys. Res. Lett., 29(20), 1963, Holocene electrical records from Dome C and Vostok, Ann. Glaciol., 29,
doi:10.1029/2002GL015186. 89 – 93.
Saltzman, E. S. (1995), Ocean/atmosphere cycling of dimethylsulfide, in Zielinski, G. A. (2000), Use of paleo-records in determining variability
Ice-Core Studies of Global Biogeochemical Cycle, NATO ASI Ser., within the volcanism-climate system, Quat. Sci. Rev., 19, 417 – 438.
vol. 30, Global Environmental Change, edited by R. J. Delmas, Zielinski, G. A., P. A. Mayewski, L. D. Meeker, S. I. Whitlow, S. M.
pp. 65 – 89, Springer, New York. Twickler, M. Morrison, D. A. Meese, A. J. Gow, and R. B. Alley (1994),
Schwander, J., J. Jouzel, C. U. Hammer, J. R. Petit, R. Udisti, and E. W. Record of volcanism since 7000 B.C. from the GISP2 Greenland ice core
Wolff (2001), A tentative chronology for the EPICA Dome Concordia ice and implications for the volcano-climate system, Science, 264, 948 – 952.
core, Geophys. Res. Lett., 28(22), 4243 – 4246. Zielinski, G. A., P. A. Mayewski, L. D. Meeker, S. I. Whitlow, and M. S.
Simkin, T., and L. Siebert (1994), Volcanoes of the World, 2nd ed., Twickler (1996), A 110,000-yr record of explosive volcanism from the
Geoscience Press, Tucson, Ariz. GISP2 (Greenland) ice core, Quat. Res., 45, 109 – 118.
Stenni, B., M. Proposito, R. Gragnani, O. Flora, J. Jouzel, S. Falourd, and
M. Frezzotti (2002), Eight centuries of volcanic signal and climate
change at Talos Dome (East Antarctica), J. Geophys. Res., 107(D9),
4076, doi:10.1029/2000JD000317. S. Becagli, E. Castellano, M. Severi, R. Traversi, and R. Udisti,
Tabacco, I. E., A. Passerini, F. Corbelli, and M. Gorman (1998), Deter- Department of Chemistry, University of Florence, Via della Lastruccia 3,
mination of the surface and bed topography at Dome C, East Antarctica, I-50019 Florence, Italy. (udisti@unifi.it)
J. Glaciol., 44(146), 185 – 191. M. Hansson, Department of Physical Geography and Quaternary
Udisti, R., S. Becagli, E. Castellano, R. Mulvaney, J. Schwander, S. Torcini, Geology, Stockholm University, S-10691 Stockholm, Sweden.
and E. W. Wolff (2000), Holocene electrical and chemical measurements M. Hutterli, Physics Institute, University of Bern, Sidlerstrasse 5,
from the EPICA-Dome C ice core, Ann. Glaciol., 30, 20 – 26. CH-3012 Bern, Switzerland.
Udisti, R., S. Becagli, E. Castellano, B. Delmonte, J. Jouzel, J. R. Petit, J. R. Petit, Laboratoire de Glaciologie et Géophysique de l’Environne-
J. Schwander, B. Stenni, and E. W. Wolff (2004), Stratigraphic corre- ment du CNRS, 54 rue Moliere, BP 96, F-38402 Saint-Martin-d’Heres
lations between the European Project for Ice Coring in Antarctica Cedex, France.
(EPICA) Dome C and Vostok ice cores showing the relative variations M. R. Rampino, Earth and Environmental Science Program, New York
of snow accumulation over the past 45 kyr, J. Geophys. Res., 109, University, 100 Washington Square East, Room 1009, New York, NY
D08101, doi:10.1029/2003JD004180. 10003, USA.
Wilson, R. C. L., S. A. Drury, and J. L. Chapman (2000), The J. P. Steffensen, Department of Geophysics, Niels Bohr Institute,
Great Ice Age: Climate Change and Life, Taylor and Francis, University of Copenhagen, Juliane Maries Vej. 30, DK-2100 Copenhagen,
Philadelphia, Pa. Denmark.
12 of 12

Castellano

Uploaded by

Copyright:

Available Formats

Castellano

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Castellano

Uploaded by

Copyright:

Available Formats

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 110, D06114, doi:10.

Holocene volcanic history as recorded in the sulfate stratigraphy of the

1. Introduction [3] Volcanic eruptions emit large amounts of particulate

easily distinguished from background values [Udisti et al.,

accumulation rates derived from dD measurements and

events recorded in the Greenland GISP2 ice core [Zielinski et

but synchronous signatures were found in other shallow firn

Kuwae signature (1460 AD) is a major example of depo-

periods [Krinner and Genthon, 1998], we support the

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.