TCThe CryosphereTCThe Cryosphere1994-0424Copernicus PublicationsGöttingen, Germany10.5194/tc-10-2721-2016Estimating the extent of Antarctic summer sea ice during the Heroic Age of Antarctic ExplorationEdinburghTomDayJonathan J.j.j.day@reading.ac.ukhttps://orcid.org/0000-0002-3750-649XDepartment of Meteorology, University of Reading, Reading, UKcurrently at: Department of Applied Mathematics and Theoretical
Physics, University of Cambridge, Cambridge, UKJonathan J. Day (j.j.day@reading.ac.uk)21November20161062721273013April201629April201611September201618September2016This work is licensed under a Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/This article is available from https://tc.copernicus.org/articles/10/2721/2016/tc-10-2721-2016.htmlThe full text article is available as a PDF file from https://tc.copernicus.org/articles/10/2721/2016/tc-10-2721-2016.pdf
In stark contrast to the sharp decline in Arctic sea ice, there has been a
steady increase in ice extent around Antarctica during the last three
decades, especially in the Weddell and Ross seas. In general, climate models
do not to capture this trend and a lack of information about sea ice coverage
in the pre-satellite period limits our ability to quantify the sensitivity of
sea ice to climate change and robustly validate climate models. However,
evidence of the presence and nature of sea ice was often recorded during
early Antarctic exploration, though these sources have not previously been
explored or exploited until now. We have analysed observations of the summer
sea ice edge from the ship logbooks of explorers such as Robert Falcon Scott,
Ernest Shackleton and their contemporaries during the Heroic Age of Antarctic Exploration (1897–1917), and in this study we compare these to satellite
observations from the period 1989–2014, offering insight into the ice
conditions of this period, from direct observations, for the first time. This
comparison shows that the summer sea ice edge was between 1.0 and
1.7∘ further north in the Weddell Sea during this period but that
ice conditions were surprisingly comparable to the present day in other
sectors.
Introduction
Understanding the interactions between polar sea ice trends and global
climate change is of key importance in climate science. Amplified Arctic
warming is closely associated with the dramatic reduction in Arctic summer
sea ice, predominately through ice-albedo feedback effects (Manabe and
Stouffer, 1980; Serreze and Barry, 2011). However the reasons for the
unexpected positive trend in Antarctic sea ice extent since the 1970s are not
yet well understood (Parkinson and Cavalieri, 2012; Turner and Overland,
2009), and the climate models submitted to the Fifth Climate Model
Inter-comparison Project do not reproduce this circumpolar satellite-era
increase (Maksym et al., 2012; Turner et al., 2013). This has led the
Intergovernmental Panel on Climate Change to assign low confidence
to future projections of Antarctic sea ice extent in its latest assessment
report (Stocker et al., 2013).
Map of expedition routes taken by ships used in this study. We
only have coordinates for entry and departure of the pack ice for the
1901–1903 Gauss Expedition (Indian Ocean).
This increase in pan-Antarctic sea ice extent is a sum of opposing regional
trends, with large increases in the Indian Ocean, Weddell and in particular
Ross seas dominating over decreases in the Bellingshausen and Amundsen Sea
(Parkinson and Cavalieri, 2012). Both mechanical and thermodynamic forcing by
the atmosphere (Holland and Kwok, 2012) and changes to the Southern Ocean
(Gille, 2002; Jacobs, 2002) are thought to play a role in these overall
trends. It has been suggested that anthropogenically driven changes such as
ozone depletion (Ferreira et al., 2015; Sigmond and Fyfe, 2010; Turner et
al., 2009), and responses to greenhouse gasses via negative sea-ice–ocean
feedback (Zhang, 2007), and ice sheet runoff (Bintanja et al., 2013; Swart
and Fyfe, 2013) all play a role. However, the relative importance of each of
these processes is not well quantified (e.g. Turner et al., 2015). It is also
likely that internal climate variability has played a part in the observed
changes (Polvani and Smith, 2013), with a positive trend in the Southern
Annular Mode thought to be a contributing factor (Lefebvre et al., 2004;
Thompson and Solomon, 2002). Both a lack of sea ice data from the
pre-satellite era and a lack of credibility in climate models restricts
analysis of these processes (Abram et al., 2013). A longer-term context may
give some insight into our understanding of the dominant mechanisms.
Recent years have seen significant efforts in the recovery of historical
meteorological records from ship logbooks (Brohan et al., 2009, 2010). Many
of these logbooks contain detailed descriptions of the sea ice state at
regular intervals and provide an invaluable source of sea ice edge
information, but they require careful interpretation (Ayre et al., 2015). Such
data are available from the earliest Antarctic voyages in the 19th century –
of Cook, Bellingshausen, Ross and others (Wilkinson, 2014) – but data from
this early period are too temporally and spatially restricted for any firm
conclusions to be made (Parkinson, 1990). It is not until the Heroic Age of Antarctic Exploration that a sufficient level of data was collected to make
concrete interpretations about the sea ice cover.
Anomaly between ship-observed ice edge and the 1989–2014 mean
PM Bootstrap algorithm derived ice edge position for the appropriate
calendar day. Anomalies are plotted at logbook position.
The period known as the Heroic Age of Antarctic Exploration began with the
Belgian Antarctic Expedition of 1897–1899 and ended in 1917 with the
conclusion of Shackleton's British Imperial Trans-Antarctic Expedition. This
period saw an expansion of exploration around the continent (Fig. 1),
allowing perhaps the earliest window for pan-Antarctic climate analysis using
observed records. It is worth noting that the difficulty in travelling
through heavy pack ice, much more of a problem during this period than in the
current day, undoubtedly had a large influence on the expedition routes. For
this reason, the sub-expedition voyages were undertaken mainly in the summer
months (November–March), with some notable multi-year exceptions (such as
the Aurora and Endurance, during the 1914–1917 Imperial
Trans-Antarctic Expedition), during which ships drifted, frozen into the pack
ice, throughout the Antarctic winter.
Until now, evidence of Antarctic sea ice conditions during this period has
only been available from proxy sources, such as those derived from ice
cores. Chemical tracers within cores, such as sea salt and
methanesulfonic
acid (Curran et al., 2003), are known to co-vary with the latitude of the
sea ice edge. These interactions are well understood in the Weddell Sea
sector, where the century-long record of ice freeze and thaw dates at South
Orkney Islands (Murphy et al., 2014) has enabled such proxies to be well
calibrated, showing a long-term decline in both spring and autumn ice cover
in this region (Abram et al., 2007). Nevertheless, it is currently unclear
how generalisable these methods are to other regions of Antarctica, due to a
limited set of directly observed sea ice data (Abram et al., 2013). The
results of the present work will provide additional reference points for
such studies.
Twentieth century variations in sea ice cover have been inferred from whale catch
positions, which provide a substantially more abundant, albeit less reliable
and indirect, source of sea ice information (Vaughan, 2000). Previous
studies, using these catch positions as a proxy record for the ice edge,
suggested a 2.8∘ southward shift in the mean latitude of the summer
sea ice edge, equating to a 25 % decline sea ice extent between
1931–1961 and 1971–1987 (Cotté and Guinet, 2007; de la Mare, 1997).
However, there is some disagreement over the magnitude and spatial nature of
results inferred from these records (Ackley et al., 2003; de la Mare, 2009),
in part due to the accuracy of the satellite-derived ice during the summer
melt season (Worby and Comiso, 2004) but also due to the evolution of whaling
practices in the Southern Ocean (Vaughan, 2000). However, the ice decline
documented by de la Mare (1997, 2009) is consistent with a significant warming of the
Southern Ocean between 1950 and 1978 documented recently by Fan et al. (2014).
Whale catch records from earlier periods, in particular the Heroic
Age, have not yet been used in sea ice analysis.
In this study, we use sea ice edge positions recorded in the ship logbooks
during the Heroic Age to estimate the mean summer ice edge latitude,
both regionally and for the whole Antarctic, during the period 1897–1917. We
compare these with modern satellite data in order to determine whether
Antarctic summer sea ice extent was different to the present day and identify
and quantify the possible changes.
Method
We use data collected during 11 expeditions of the Heroic Age for which
frequent observations on the composition and nature of the sea ice were
recorded (Table 1). Many of these logbooks had already been digitised
recently as part of the International Comprehensive Ocean-Atmosphere Data Set
(ICOADS) initiative (Woodruff et al., 2011; Freeman et al., 2016). Others
needed to be digitised specifically for this study, either from photographic
images of the original logbook or in person from the original logbook itself.
These were combined to create a dataset of 191 observed ice edge positions
(included in the Supplement), providing an almost circumpolar picture of the
Antarctic summer sea ice edge (Fig. 2).
Expedition information and source materials used in this study. For
items not digitised by ICOADS (Woodruff et al., 2011),
the source material and archive are listed.
Logbooks from this period typically include meteorological observations taken
at frequent intervals throughout each day. Details of the sea ice cover were
recorded along with a descriptive summary of the sea state and meteorology in
the time period between the quantitative meteorological observations.
Sometimes sea ice remarks were recorded in a specific column of the logbook
describing the ice state but more often included under a general observation
heading that also encompassed additional comments regarding weather, sea
state and other notable features, including wildlife. In some cases, the
times of certain observations or events, such as when the ship passed from a region
of consolidated ice cover into open water, were noted specifically. These
observations provide a clear picture of the presence and composition of the
ice throughout the expeditions. We have only used expeditions for which a
summary of the sea ice conditions was recorded at frequent time intervals,
discarding expeditions which only provide a daily summary for reasons of
precision, and have included a full list of terms used to describe the ice
conditions in Table S1 in the Supplement. We have assumed that the times
against which the observations were recorded reflect the time zone that the ship was
operating in. Whilst this was certainly the practice in Royal Navy logbooks
of the time, it may not be true for all non-British expeditions (S. Woodruff, personal communication, 2015). Logbook times were all converted into UTC
prior to comparison with the PM satellite data. In most logbooks, the
position of each ship was recorded only at midday. Geo-locating remarks about
the ice at other times of day required linear interpolation between the
midday positions of consecutive days based on the time associated with the
ice information.
In order to identify the sea ice edge for the period, we used these remarks
about the sea ice to estimate points in space and time where the ships were
traversing or travelling along the sea ice edge. The aim, and therefore the
route, of an expedition had a significant impact on the number of ice edge
points that can be determined for each particular log and there is a clear
distinction in the type of exploration (see the videos in the Supplement).
Some expeditions, such as the Terra Nova expeditions to the Ross
Sea, were primarily concerned with land exploration or the race to the South
Pole, and therefore the voyage of the ship was mostly a means of transport to
the continent. In this scenario, it is common that the ship only crossed the
ice edge on the journey south to and the return north from the continent – in
which case, the positions of the first and last observations of sea ice on the
respective journeys are taken as the ice edge. However, the focus of other
expeditions, such as the Australasian 1911–1914 expedition to the Western
Pacific sector, on the Aurora, was to explore undocumented
coastlines by ship. Hence, the ship was often travelling along the sea ice
edge for long periods, passing in and out of areas of sea ice frequently, and
the ice edge is often unambiguously recorded in the logbooks using terms such
as “skirting pack ice”.
Interactions with the ice edge, in particular the navigation between open
water and regions of ice cover, can sometimes be complicated by an ambiguity
in the exact time that the ship encountered the ice edge. As an example,
consecutive remarks in the log of the Scotia expedition of 1903–1904, made
on the 1 March at 11:00 UTC and 16 h later on the following day at
05:00 UTC, observe “steaming through loose pack” and “no ice in sight”
respectively. In cases such as this, this ambiguity makes it difficult to
objectively identify the position of the ice edge, so we have taken the
approach of trying to define a northern bound for the ice edge. We believe
that a sensible approach is to utilise the entries recording the transition
between these regions, using only the northernmost point within this pair of
entries if the ship is travelling in a north–south axis and both entries if,
as in the example above, the ship is travelling in an east–west axis. This
approach increases our set of observations and reduces the level of
subjectivity in the inference of the ice edge from the terms used in the
logs.
Before describing how we performed this comparison, it is useful to consider
the relationship between these point data and satellite observed
concentration. In satellite sea ice concentration products, a threshold of
15 % is generally used to define the location of the ice edge. A detailed
study by Worby and Comiso (2004) compared the sea ice edge as recorded by
onboard trained human observers, during expeditions between 1989 and 2000 in
the Western Pacific sector, with the 15 % contour in PM satellite imagery
for the same day. They showed a high level of agreement during the
March–October
ice growth season but noted that during the summer melt season (November–February), the
sea ice edge was systematically further north than the 15 % contour in
both the Nasa team (Cavalieri et al., 1984) and Bootstrap (Comiso, 1986)
algorithms, with the Bootstrap algorithm being a closer match to the in situ
data. They argue that during this time of year, saturated bands of ice and
floes, particularly at the edges of the pack ice, may be very localised,
resulting in ice concentration below the 15 % threshold when averaged
over the 25 km pixel size of the PM dataset. In addition, these bands often
comprise mostly brash ice, the PM signature of which can be almost
indistinguishable from seawater. They estimate the offset between the
observed ice edge and ice edge derived using the Bootstrap algorithm to be
0.75 ± 0.61∘ during the melt season. However, it should be
noted that their analysis has not been extended to other sectors, so we do
not know how representative it is of the ice conditions outside of the
Western Pacific, and also that the observers on-board ship during the
Heroic Age were not trained sea ice observers. Nevertheless, it is
important to take their findings into account in the following analysis,
since an apparent southward shift in the ice edge between the Heroic
Age points and the present-day Bootstrap-derived ice edge may overestimate
the actual change and should therefore best be considered as an upper bound.
Mean differences in ice edge latitude between the ship-observed and
daily mean satellite-derived ice edge position by Antarctic sector. Positive
differences indicate where the ship-observed ice edge is north of the mean
satellite-derived ice edge. Bracketed quantities refer to the same
difference, but with the Worby and Comiso offset subtracted.
Comparison of ship-observed and satellite-derived ice edge latitude,
including a one-to-one line, which indicates no change in position. The
dashed line provides an estimate of the southward offset one would expect
when comparing the satellite-derived ice edge to in situ ship observations,
as calculated by Worby and Comiso (2004).
As there is greater consistency with the in situ ice edge observations, we
have chosen to use the Bootstrap algorithm daily sea ice concentration from
the National Snow and Ice Data Centre (NSIDC) to estimate the present-day sea
ice edge, rather than NASA Team algorithm (Meier et al., 2013a; Peng et al.,
2013). From this, we calculated a daily mean sea ice concentration for the
period 1989–2014, during which daily data were available. Using this field,
we have defined the daily satellite-observed mean ice edge as the contour
joining the midpoints of the northernmost 25 km pixels that do not exceed a
sea ice concentration of 15 %. Then, for each ice edge observation from
the logbooks, we computed the distance from the ship-observed position to
each point on the contour for that particular calendar day using the
Haversine formula and spherical law of cosines and selected the latitude of
the point on the contour with the minimum distance to the ship-observed
position for the paired analysis. These differences were then averaged to
calculate estimates for the mean change in the ice edge latitude for each
sector and for the whole of Antarctica.
We believe the offset between the ice edge recorded by human observers and
the satellite-derived ice edge to be the largest source of uncertainty in
this analysis. Therefore, we use the Worby and Comiso value of
0.75∘ in the discussion of our results in order to address this
source of uncertainty.
Results
The most dramatic change between the Heroic Age and the present day
is in the Weddell Sea, where we have found that the ice edge was
1.71∘ further north during the Heroic Age and 0.96∘
further north with the inclusion of the Worby and Comiso offset, with both
values significant at the 5 % level (Table 2). This change agrees well
with the observed decrease in land-fast ice at South Orkney Islands during the
last century (Murphy et al., 2014).
However, observations in the Weddell Sea are clustered into a small number of
years (1903, 1904 and 1914), and as such may be influenced by natural
year-to-year variability in the sea ice extent. Indeed, the ice edge observed
by the crew of the Scotia in 1903 is particularly far north, even
compared to observations from by the same crew the following year, and the
greater ice extent in this region during the Heroic Age may be an
exaggeration of the actual change between these two periods as a result. El
Niño events often result in negative surface air temperature and positive
sea ice anomalies in the Weddell Sea (Yuan, 2004), and it is possible that the
large El Niño of 1902–1903 may have had some influence on these
anomalies. Nevertheless, our results indicate that many of the recorded ice
edge positions lie further north than has been seen in any year since 1989
(Fig. 3, Supplement Fig. 2, Supplement video 1).
The estimated DJFM Antarctic sea ice extent climatology for the
period 1897–1917, with and without the inclusion of the Worby and Comiso
offset, is plotted alongside time series of DJFM mean sea ice extent
calculated from HadISST2.2, NASA Team and NASA PM Bootstrap sea ice
concentration datasets.
The differences in other sectors appear to be much smaller. We observe
statistically significant but small differences of 0.21∘ in latitude
in the Bellingshausen and Amundsen seas and 0.62∘ the Ross Sea but no
evidence of a significant difference in latitude in the Western Pacific and
Indian Ocean region, which we have merged due to the limited data available
in the Indian Ocean sector (Table 2, Fig. 3). However, as stated in the
previous section, this is an upper bound for the observed change; if we
take into account the observed offset to the satellite data, the ice edge may
actually have been further south in all of these areas (Table 2).
By averaging over all points, we also find the mean circumpolar change in the
location of the ice edge. This pan-Antarctic shift of up to 0.41∘
southwards since the Heroic Age implies at most a 14.2 %
decrease in Antarctic sea ice extent between then and the present day
(Fig. 4). This is much smaller than the 25 % decrease between the 1950s
and 1980s that was inferred by de la Mare (1997) from whaling records and
suggests that, if we accept the results inferred from the whaling records,
the sea ice was less extensive during the period 1897–1917 than it was
during the period 1931–1961.
Although ice edge latitude is a convenient measure in which to analyse the
data, it is more common in climate change assessment reports, such as the
IPCC-AR5, to assess hemispheric sea ice variability and change using sea ice
extent rather than ice edge latitude (Vaughan et al., 2013). We do not have
estimates of the ice edge position for all longitudes and therefore cannot
calculate the ice extent during the Heroic Age from the logbook data
alone. However, we were able to form estimates using the shape of the
satellite sea ice climatology for each day within the DJFM period, by
computing the number of 25 km square grid cells within region enclosed by
two contours: the 15 % ice concentration contour defined above and a
contour radially shifted with respect to the calculated latitude change for
that particular day. We then averaged this over the DJFM period and added the
output to the mean NASA Bootstrap sea ice extent for the period 1989–2016,
to give a loosely approximated estimate of the ice extent during the
Heroic Age. This calculation also included regions enclosed by these
two contours at polynya ice edges. We acknowledge the imprecision of these
estimates that arise from calculating this pan-Antarctic sea ice extent using
only limited individual ice edge position data and suggest that further
research could be done in this using an expanded dataset if possible.
Our estimate of the mean DJFM sea ice extent, based on the mean ice edge
latitude, is 7.4 × 106 km2 (or
5.3 × 106 km2, using the Worby and Comiso offset).
Comparing our DJFM sea ice extent estimate to the Met Office Hadley Centre
sea ice HadISST2.2 dataset (Titchner and Rayner, 2014), we find that our
values are 4.3 × 106 km2 lower (Fig. 4). During this
period HadISST2.2 is based on a climatology for the period 1929–1939,
derived from German Atlas charts (German Hydrographic Institute, 1950).
Data from 185 whaling expeditions (mostly Norwegian but some from English
whaling logbooks) were used as its basis (H. Titchner, personal
communication, 2016).
Summary and discussion
In this study, we have used logbook data recorded by explorers during the
Heroic Age of Antarctic Exploration to estimate an upper bound for the change
between the summer sea ice edge during that period and the present day. The
following summarises our conclusions:
We estimate that the DJFM sea ice edge was at most 0.41∘ further
south between 1989 and 2014 than it was during the Heroic Age
(1897–1917), implying a reduction of 14.2 % in pan-Antarctic extent.
This change is most dramatic and statistically robust in the Weddell
Sea, where the ice edge shifted by 1.71∘ southward between the
two periods.
Our estimate of the change in extent between the Heroic Age
and the present day is small relative to estimates of the change between the
1950s and 1970s, based on whale catch data (Cotté and Guinet, 2007; de la
Mare, 1997; Titchner and Rayner, 2014). This suggests the possibility that
the sea ice was significantly more extensive during the period 1931–1961
than during the Heroic Age.
Outside the Weddell Sea, the mean change in ice edge latitude is small
compared to the known summer offset between in situ and satellite ice edges.
It is therefore plausible that the sea ice extent in these regions has in
fact experienced a small increase. Either way, the climate was much more
similar to the present conditions than one might expect based on climate
model simulations of the early 19th century (e.g. Turner et al., 2013). These
ice edge data, which we make available in the Supplement of this
paper, could be used by climate model developers as a tuning target for
pre-industrial Antarctic sea ice cover.
It has been suggested that the use of whale catch positions for estimating
the sea ice edge will overestimate any changes (Ackley et al., 2003); this
may be responsible for the disparity between our results and those
investigating the whaling data. However, it is also plausible that the
Antarctic ice edge exhibits significant decadal and multi-decadal variability
(Latif et al., 2013), as has been observed in the Arctic (Day et al., 2012;
Divine and Dick, 2006; Miles et al., 2014). If so, it is possible that there
was an increase in sea ice extent between the Heroic Age and the
1930s, followed by the decrease between 1961 and 1971, suggested by de la
Mare (1997). The MSA record from Law Dome, East Antarctica (Curran et al.,
2003), and extensive sea ice in Nimbus satellite data from the early 1960s (Gagné
et al., 2015; Meier et al., 2013b) seem to support this hypothesis.
We have excluded some of the expedition logbooks available to us because of
the lack of ice condition observations, including the Southern Cross
expedition of 1898–1900, which only recorded a daily summary of the sea ice,
and the Hertha and Jason expeditions of 1893–1894, neither
of which recorded a summary of the sea ice. However, there are a few logbooks
from this period, including those from the Norwegian expedition on the
Fram (1910–1912), the Japanese expeditions on the Kainan Maru (1910–1912) and the German expedition on the Deutschland
(1911–1912), that have not yet been imaged and digitised and could
potentially increase our knowledge of the sea ice conditions during this
period. Similarly, although whaling records for the Heroic Age are
less complete than during the later period studied by de la Mare (1997,
2009), whaling records from the Heroic Age could provide additional
validation of our findings.
Further analysis into the processes driving the long-term variations in
Antarctic sea ice described in this study will be an important next step. In
particular, expanding our source of Southern Ocean climatological data by
increasing efforts to digitise ship logbooks from this data-sparse region may
be key to both our understanding recent Antarctic sea ice behaviour and
improving climate model performance.
Data availability
The ice edge points used in this study can be downloaded as as part of the
Supplement associated with this paper. These were derived from the primary
and secondary sources listed in Table 1, as described in Sect. 2.
These data were compared to satellite-derived sea ice concentration data
(Meier et al., 2013), which can be downloaded from the NSIDC website.
The Supplement related to this article is available online at doi:10.5194/tc-10-2721-2016-supplement.
Acknowledgements
Tom Edinburgh was funded by a NERC-SCENARIO-funded summer studentship.
Jonathan Day was funded by an AXA Post-Doctoral Research Fellowship and a
grant from the Walker Institute, University of Reading. We would like to
thank Clive Wilkinson at the Climatic Research Unit, University of East
Anglia, for his advice on logbook sources and for access to digitised
logbook data, Holly Titchener (Met Office) and reviewers Dmitry Divine and
Florence Fetterer for their useful comments and recommendations regarding
the manuscript. Thanks also to the Royal Geographical Society Foyle Reading
Room, for access to original logbooks from the Discovery 1901–1904 and Terra Nova 1903–1904 expeditions.Edited by: C. Haas
Reviewed by: F. Fetterer and D. Divine
ReferencesAbram, N. J., Mulvaney, R., Wolff, E. W., and Mudelsee, M.: Ice core records
as sea ice proxies: An evaluation from the Weddell Sea region of Antarctica,
J. Geophys. Res., 112, D15101, 10.1029/2006JD008139, 2007.Abram, N. J., Wolff, E. W., and Curran, M. A. J.: A review of sea ice proxy
information from polar ice cores, Quat. Sci. Rev., 79, 168–183,
10.1016/j.quascirev.2013.01.011, 2013.Ackley, S., Wadhams, P., Comiso, J. C., and Worby, A. P.: Decadal decrease of
Antarctic sea ice extent inferred from whaling records revisited on the
basis of historical and modern sea ice records, Polar Res., 22, 19–25,
10.1111/j.1751-8369.2003.tb00091.x, 2003.
Arctowski, H.: Les observations météorologiques de l'expédition
du Gauss, Bull. Soc. Belge Astron., 9, 13–17, 1904.Ayre, M., Nicholls, J., Ward, C., and Wheeler, D.: Ships' logbooks from the
Arctic in the pre-instrumental period, Geosci. Data J., 53–62,
10.1002/gdj3.27, 2015.Bintanja, R., van Oldenborgh, G. J., Drijfhout, S. S., Wouters, B., and
Katsman, C. A.: Important role for ocean warming and increased ice-shelf
melt in Antarctic sea-ice expansion, Nat. Geosci., 6, 376–379,
10.1038/ngeo1767, 2013.Brohan, P., Allan, R., Freeman, J. E., Waple, A. M., Wheeler, D., Wilkinson,
C., and Woodruff, S.: Marine Observations of Old Weather, B. Am. Meteor.
Soc., 90, 219–230, 10.1175/2008BAMS2522.1, 2009.Brohan, P., Ward, C., Willetts, G., Wilkinson, C., Allan, R., and Wheeler,
D.: Arctic marine climate of the early nineteenth century, Clim. Past, 6,
315–324, 10.5194/cp-6-315-2010, 2010.
Cavalieri, D. J., Gloersen, P., and Campbell, W. J.: Determination of sea ice
parameters with the Nimbus 7 SMMR, J. Geophys. Res.-Atmos.,
89, 5355–5369, 1984.
Comiso, J. C.: Characteristics of Arctic winter sea ice from satellite
multispectral microwave observations, J. Geophys. Res.-Ocean., 1978–2012,
91, 975–994, 1986.Cotté, C. and Guinet, C.: Historical whaling records reveal major
regional retreat of Antarctic sea ice, Deep Sea Res., 54, 243–252,
10.1016/j.dsr.2006.11.001, 2007.Curran, M. A. J., Ommen, T. D. van, Morgan, V. I., Phillips, K. L., and
Palmer, A. S.: Ice Core Evidence for Antarctic Sea Ice Decline Since the
1950s, Science, 302, 1203–1206, 10.1126/science.1087888, 2003.Day, J. J., Hargreaves, J. C., Annan, J. D., and Abe-Ouchi, A.: Sources of
multi-decadal variability in Arctic sea ice extent, Environ. Res. Lett.,
7, 034011, 10.1088/1748-9326/7/3/034011, 2012.de la Mare, W. K.: Abrupt mid-twentieth-century decline in Antarctic sea-ice
extent from whaling records, Nature, 389, 57–60, 10.1038/37956,
1997.de la Mare, W. K.: Changes in Antarctic sea-ice extent from direct
historical observations and whaling records, Clim. Change, 92,
461–493, 10.1007/s10584-008-9473-2, 2009.Divine, D. V. and Dick, C.: Historical variability of sea ice edge position
in the Nordic Seas, J. Geophys. Res.-Ocean., 111, C01001,
10.1029/2004JC002851, 2006.Fan, T., Deser, C., and Schneider, D. P.: Recent Antarctic sea ice trends in
the context of Southern Ocean surface climate variations since 1950, Geophys.
Res. Lett., 41, 2419–2426, 10.1002/2014GL059239, 2014.Ferreira, D., Marshall, J., Bitz, C. M., Solomon, S., and Plumb, A.:
Antarctic Ocean and Sea Ice Response to Ozone Depletion: A Two-Time-Scale
Problem, J. Clim., 28, 1206–1226, 10.1175/JCLI-D-14-00313.1, 2015.Freeman, E., Woodruff, S. D., Worley, S. J., Lubker, S. J., Kent, E. C.,
Angel, W. E., Berry, D. I., Brohan, P., Eastman, R., Gates, L., Gloeden, W.,
Ji, Z., Lawrimore, J., Rayner, N. A., Rosenhagen, G., and Smith, S. R.:
ICOADS Release 3.0: a major update to the historical marine climate record,
Int. J. Climatol., 10.1002/joc.4775, 2016.Gagné, M.-È., Gillett, N. P., and Fyfe, J. C.: Observed and simulated
changes in Antarctic sea ice extent over the past 50 years, Geophys. Res.
Lett., 42, 90–95, 10.1002/2014GL062231, 2015.
Gille, S. T.: Warming of the Southern Ocean since the 1950s, Science,
295, 1275–1277, 2002.Holland, P. R. and Kwok, R.: Wind-driven trends in Antarctic sea-ice drift,
Nat. Geosci., 5, 872–875, 10.1038/ngeo1627, 2012.Jacobs, S. S.: Freshening of the Ross Sea During the Late 20th Century,
Science, 297, 386–389, 10.1126/science.1069574, 2002.
Kidson, E.: British Antarctic Expedition, 1907–1909, under the command of
Sir E. H. Shackleton. Reports on the scientific investigations, Meteorology,
Government Printer, Melbourne, 118 pp., 1930.Latif, M., Martin, T., and Park, W.: Southern Ocean Sector Centennial Climate
Variability and Recent Decadal Trends, J. Clim., 26, 7767–7782,
10.1175/JCLI-D-12-00281.1, 2013.Lefebvre, W., Goosse, H., Timmermann, R., and Fichefet, T.: Influence of the
Southern Annular Mode on the sea ice–ocean system, J. Geophys. Res., 109,
C09005, 10.1029/2004JC002403, 2004.Maksym, T., Stammerjohn, S., Ackley, S., and Massom, R.: Antarctic Sea
Ice – A Polar Opposite?, Oceanography, 25, 140–151,
10.5670/oceanog.2012.88, 2012.Manabe, S. and Stouffer, R. J.: Sensitivity of a global climate model to an
increase of CO2 concentration in the atmosphere, J. Geophys. Res.-Ocean., 85, 5529–5554, 1980.
Mawson, D.: Part 2: Meteorological Log of the S. Y. “Aurora”, in
Australasian Antarctic Expedition 1911–14: Meteorology, Thomas Henry
Tennant, Sydney, 110–208, 1939.Meier, W., Fetterer, F., Savoie, M., Mallory, S., Duerr, R., and Stroeve, J.:
NOAA/NSIDC Climate Data Record of Passive Microwave Sea Ice Concentration,
Version 2. Boulder, Colorado USA: National Snow and Ice Data Center,
10.7265/N55M63M1, 2013a (updated 2015). Meier, W. N., Gallaher, D., and Campbell, G. G.: New estimates of Arctic and
Antarctic sea ice extent during September 1964 from recovered Nimbus I
satellite imagery, The Cryosphere, 7, 699–705, 10.5194/tc-7-699-2013,
2013b.Miles, M. W., Divine, D. V., Furevik, T., Jansen, E., Moros, M., and Ogilvie,
A. E. J.: A signal of persistent Atlantic multidecadal variability in Arctic
sea ice, Geophys. Res. Lett., 41, 463–469, 10.1002/2013GL058084,
2014.
Mossman, R. C.: Meteorological Results of the Scottish National Antarctic
Expedition, in Report of the scientific results of the voyage of S Y
“Scotia,” vol. 2, Scottish Oceanographical Laboratory, Edinburgh, 1907.Murphy, E. J., Clarke, A., Abram, N. J., and Turner, J.: Variability of
sea-ice in the northern Weddell Sea during the 20th century, J. Geophys.
Res. Ocean., 10.1002/2013JC009511, 2014.
Parkinson, C. L.: Search for the Little Ice Age in Southern Ocean sea-ice
records, Ann. Glaciol., 14, 221–225, 1990.Parkinson, C. L. and Cavalieri, D. J.: Antarctic sea ice variability and
trends, 1979–2010, The Cryosphere, 6, 871–880, 10.5194/tc-6-871-2012,
2012.Peng, G., Meier, W. N., Scott, D. J., and Savoie, M. H.: A long-term and
reproducible passive microwave sea ice concentration data record for climate
studies and monitoring, Earth Syst. Sci. Data, 5, 311–318,
10.5194/essd-5-311-2013, 2013.Polvani, L. M. and Smith, K. L.: Can natural variability explain observed
Antarctic sea ice trends? New modeling evidence from CMIP5, Geophys. Res.
Lett., 40, 3195–3199, 10.1002/grl.50578, 2013.
Rouch, J.: Deuxieme Expedition Antarctique Francaise (1908–10):
Oceanographie Physique, Masson et Cie., (imaged by NOAA), Paris, 1913.Serreze, M. C. and Barry, R. G.: Processes and impacts of Arctic
amplification: A research synthesis, Global Planet. Change, 77, 85–96,
10.1016/j.gloplacha.2011.03.004, 2011.Sigmond, M. and Fyfe, J. C.: Has the ozone hole contributed to increased
Antarctic sea ice extent?, Geophys. Res. Lett., 37,
10.1029/2010GL044301, 2010.
Stocker, T. F., Qin, D., Plattner, G. K., Tignor, M., Allen, S. K., Boschung,
J., Nauels, A., Xia, Y., Bex, V., and Midgley, P. M.: IPCC 2013: Summary for
Policymakers. In Climate Change 2013: The Physical Science Basis, Contrib.
Work, Group Fifth Assess. Rep., Intergov. Panel Clim. Change, University
Press Cambridge, UK, p. 25, 2013.Swart, N. C. and Fyfe, J. C.: The influence of recent Antarctic ice sheet
retreat on simulated sea ice area trends, Geophys. Res. Lett., 40,
4328–4332, 10.1002/grl.50820, 2013.Thompson, D. W. J. and Solomon, S.: Interpretation of Recent Southern
Hemisphere Climate Change, Science, 296, 895–899,
10.1126/science.1069270, 2002.Titchner, H. A. and Rayner, N. A.: The Met Office Hadley Centre sea ice and
sea surface temperature data set, version 2: 1. Sea ice concentrations, J.
Geophys. Res.-Atmos., 119, 2864–2889, 10.1002/2013JD020316, 2014.Turner, J. and Overland, J.: Contrasting climate change in the two polar
regions, Polar Res., 28, 146–164, 10.1111/j.1751-8369.2009.00128.x,
2009.Turner, J., Comiso, J. C., Marshall, G. J., Lachlan-Cope, T. A., Bracegirdle,
T., Maksym, T., Meredith, M. P., Wang, Z., and Orr, A.: Non-annular
atmospheric circulation change induced by stratospheric ozone depletion and
its role in the recent increase of Antarctic sea ice extent, Geophys. Res.
Lett., 36, L08502, 10.1029/2009GL037524, 2009.Turner, J., Bracegirdle, T. J., Phillips, T., Marshall, G. J., and Hosking,
J. S.: An Initial Assessment of Antarctic Sea Ice Extent in the CMIP5
Models, J. Clim., 26, 1473–1484, 10.1175/JCLI-D-12-00068.1, 2013.Turner, J., Hosking, J. S., Bracegirdle, T. J., Marshall, G. J. and Phillips,
T.: Recent changes in Antarctic Sea Ice, Philos. Trans. R. Soc. Math. Phys.
Eng. Sci., 373, 20140163, 10.1098/rsta.2014.0163, 2015.Vaughan, D. G., Comiso, J. C., Allison, I., Carrasco, J., Kaser, G., Kwok,
R., Mote, P., Murray, T., Paul, F., Ren, J., Rignot, E., Solomina, O.,
Steffen, K., and Zhang, T.: Observations: cryosphere. In Climate Change 2013:
The Physical Science Basis, Clim. Change, 317–382, 2013.
Vaughan, S.: Can Antarctic sea-ice extent be determined from whaling
records?, Polar Rec., 36, 345–347, 10.1017/S0032247400016831,
2000.
Wilkinson, C.: Sources of Historic Sea-Ice Observations for the Southern
Hemisphere, Recovery of Logbooks and International Marine Data RECLAIM
Project – 2012 (revised and updated 2014), 2014.Woodruff, S. D., Worley, S. J., Lubker, S. J., Ji, Z., Eric Freeman, J.,
Berry, D. I., Brohan, P., Kent, E. C., Reynolds, R. W., Smith, S. R., and
Wilkinson, C.: ICOADS Release 2.5: extensions and enhancements to the
surface marine meteorological archive, Int. J. Climatol., 31, 951–967,
10.1002/joc.2103, 2011.Worby, A. P. and Comiso, J. C.: Studies of the Antarctic sea ice edge and
ice extent from satellite and ship observations, Remote Sens. Environ.,
92, 98–111, 10.1016/j.rse.2004.05.007, 2004.Yuan, X.: ENSO-related impacts on Antarctic sea ice: a synthesis of
phenomenon and mechanisms, Antarct. Sci., 16, 415–425,
10.1017/S0954102004002238, 2004.Zhang, J.: Increasing Antarctic Sea Ice under Warming Atmospheric and
Oceanic Conditions, J. Clim., 20, 2515–2529, 10.1175/JCLI4136.1,
2007.