COMMENT ENVIRONMENT Across the sea of POLAR SCIENCE Eerie remains of CONSERVATION Social shifts HISTORY Underground bone plastic waste in search lost Arctic expedition hint are causing bush fires in plan made brain mapper a of solutions p.278 at what went wrong p.280 Zimbabwe p.281 war hero p.281 MARIO TAMA/GETTY

Sea ice off the coast of West Antarctica starts to melt, as seen from NASA’s Operation IceBridge aircraft in October 2016. Solve Antarctica’s sea-ice puzzle John Turner and Josefino Comiso call for a coordinated push to crack the baffling rise and fall of sea ice around Antarctica.

ifferent stories are unfolding at for decades3. Record maxima were recorded the mean annual minimum. the two poles of our planet. In the in 2012, 2013 and 2014 (ref. 4). Researchers are struggling to understand Arctic, more than half of the sum- So it came as a surprise to scientists when these stark differences5. Why do Antarctica’s Dmer sea ice has disappeared since the late on 1 March 2017, Antarctic sea-ice cover marked regional and seasonal patterns of 1970s1. The steady decline is what global cli- shrank to a historic low. Its extent was the sea-ice change differ from the more uniform mate models predict for a warming world2. smallest observed since satellite monitoring decline seen around most of the Arctic? Why Meanwhile, in Antarctic waters, sea-ice began in 1978 (see ‘Poles apart’) — at about has Antarctica managed to keep its sea ice cover has been stable, and even increasing, 2 million square kilometres, or 27% below until now? Is the 2017 Antarctic decline a

©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts r20eser vJULYed. 2017 | VOL 547 | NATURE | 275

COMMENT

brief anomaly or the start of a longer-term (thermodynamics). Such knowledge under- El Niño–Southern Oscillation, suggesting shift6,7? Is sea-ice cover more variable than we pins climate models; these need to better that sea ice is sensitive to conditions far thought? Pressingly, why do even the most represent the complex interactions between from the poles. highly-rated climate models have Antarctic sea ice and the atmosphere, ocean and ice Another issue is how the balance between sea ice decreasing rather than increasing in sheet. What’s required now is a focused and dynamics and thermodynamics drives the recent decades? We need to know whether coordinated international effort across the advance and retreat of ice. The thickness crucial interactions and feedbacks between scientific disciplines that observe and model and volume of ice depend on many factors, the atmosphere, ocean and sea ice are miss- global climate and the polar regions. including the flow of heat from the ocean ing from the models, and to what extent to the atmosphere and to the ice. Sea ice human influences are implicated6. LIMITED RECORDS influences the saltiness of the ocean. As the What happens in the Antarctic affects the Satellites provide the best spatial informa- ocean freezes, salt enters the water; as the ice whole planet. The Southern Ocean has a key tion on sea ice around Antarctica. Regular melts, fresh water returns to the sea. Such role in global ocean circulation; a frozen sea observations reveal how ice cover varies over processes are difficult to measure precisely, surface alters the exchange of heat and gases, days, years and decades8. Weather, especially having uncertainties to within 50–100% of including carbon dioxide, between ocean storms with high winds, has a daily influ- the signal. And they are hard to model. and atmosphere. Sea ice reflects sunlight and ence, as well as a seasonal one. Longer-term Satellite altimeters can measure the dis- influences weather changes are driven by larger patterns in the tance between the surfaces of the sea ice and systems, the forma- “Sea-ice temperature and circulation of the atmos- ocean accurately, and this distance can be tion of clouds and coverage might phere and oceans. used to calculate the ice thickness. But it is precipitation pat- have been up But near-continuous satellite observations hard to interpret these data without know- terns. These in turn to 25% greater only reach back about four decades. Longer ing how much snow is on the ice, its density affect the mass of the in the 1940s to records are essential to link sea-ice changes and whether the snow’s weight pushes the ice Antarctic ice sheet 1960s.” to trends in climate. Information from ships’ surface below sea level (as is often the case). and its contribution logbooks, coastal stations, whale-catch Calibrating and validating satellite data are to sea-level rise. Sea ice is also crucial to records, early satellite imagery and chemical crucial, as is developing algorithms to merge marine ecosystems. A wide range of organ- analyses of ice cores hint that sea-ice cover- and analyse information from a variety of isms, including krill, penguins, seals and age might have been up to 25% greater in the sources. whales, depend on its seasonal advance and 1940s to 1960s6. Ice, ocean and air must be sampled at retreat. Collecting more ice cores and historical appropriate intervals over a wide enough area It is therefore imperative that research- records, and synthesizing the information to establish how they interact. Research ice- ers understand the fate of Antarctic sea they contain, could reveal local trends. breaker cruises are essential for collecting in ice, especially in places where its area and These would help to identify which climatic situ observations; one such was the US PIP- thickness are changing, and why. This factors drive Antarctic sea-ice changes6. For ERS (Polynyas, Ice Production and seasonal requires bringing together understandings instance, in 2017, the area most depleted Evolution in the Ross Sea) campaign in 2017. of the drivers behind the movement of the of sea ice was south of the Eastern Pacific But ships travel only along narrow routes and ice (through drift and deformation) as well Ocean. This region has strong links to for a short time, typically 1–3 months. as those that control its growth and melt the climate of the tropics, including the Increasingly, autonomous instruments and vehicles — underwater, on-ice and airborne — are providing data throughout POLES APART the year and from inaccessible or dangerous Arctic sea ice has been disappearing steadily since satellite imaging began in the late 1970s. Around regions. These robotic systems are giving Antarctica the sea ice has been stable or growing — until 2017 when it plummeted to a record low. revolutionary new information and insights 3 into the formation, evolution, thickness Arctic Antarctic sea ice and melting of sea ice. Sensors mounted on Antarctica reached a record high in marine mammals (such as elephant seals), or 2014 before decreasing 2 dramatically. on floats and gliders, also beam back data on 12-year temperature, salinity and other physical and running mean biogeochemical parameters. But to operate continuously, these instruments need to 1 be robust enough to withstand the harsh SOURCE: US NATIONAL SNOW AND ICE DATA CENTER AND ICE DATA SNOW SOURCE: US NATIONAL )

2 Antarctic marine environment.

0 IMPROVE MODELS Current climate models struggle to simu- (million km late the seasonal and regional variability –1 seen in Antarctic sea ice. Most models have biases, even in basic features such as the size and spatial patterns of the annual cycle of

Sea-ice extent* relative to 1981–2010 mean relative Sea-ice extent* Monthly The Arctic's cap of Antarctic sea-ice growth and retreat, or the –2 average permanent sea ice has shrunk by about 13% amount of heat input to the ice from the per decade. ocean. The models fail to simulate even gross changes2, such as those driven by tropical 9 –3 influences on regional winds . Because ice 1979 1984 1989 1994 1999 2004 2009 2014 and climate are closely coupled, even small

*Total area of satellite pixels with sea-ice concentration 15% or higher. errors multiply quickly. Features that need to be modelled more

276 | NATURE | VOL 547 | 20 JULY ©20172017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved. ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved.

Leopard seals live, hunt and breed among the pack ice in Antarctic waters.

accurately include the belt of strong westerly longer than the satellite record. This brings together diverse communities winds that rings Antarctica, and the Amund- Some gaps in our knowledge can be filled with an interest in Antarctic climate to assess sen Sea Low — a stormy area southwest of through nationally funded research. More the performance of models. ■ the Antarctic Peninsula. Models disagree, demanding cross-disciplinary work must

PAUL NICKLEN/NGC PAUL for example, on whether persistent westerly be supported through international col- John Turner leads the polar climate and winds should increase or decrease sea-ice laboration. Leading the way are organiza- prediction group at the British Antarctic coverage around Antarctica. Simulations tions such as the Scientific Committee on Survey, Cambridge, UK, and is the of clouds and precipitation are also inade- Antarctic Research, President of the International Association quate. These cannot currently reproduce the “Because ice the Scientific Com- of Meteorology and Atmospheric Sciences. correct amounts of snowfall or sea surface and climate are mittee on Oce- Josefino Comiso is a senior research temperature of the Southern Ocean (the lat- closely coupled, anic Research, the scientist at NASA’s Goddard Space Flight ter is widely overestimated by the models). even small World Climate Center, Greenbelt, Maryland, USA. Climate models must also include the errors multiply Research Pro- e-mail: [email protected] mixing of waters by surface winds and quickly.” gramme’s Climate the impact of waves on the formation and and Cryosphere 1. Serreze, M. C. & Stroeve, J. Phil. Trans. R. Soc. Lond. A 373, 20140159 (2015). break-up of sea ice. Precipitation and melt- project and the Past Global Changes pro- 2. Turner, J., Bracegirdle, T. J., Phillips, T., Marshall, water from ice sheets and icebergs influence ject. But essential work remains to be done, G. J. & Hosking, J. S. J. Clim. 26, 1473–1484 the vertical structure of the ocean and how including: more detailed model comparisons (2013). it holds heat, which also affects the growth and assessments; more research cruises; and 3. Comiso, J. C. & Nishio, F. J. Geophys. Res. 113, C02S07 (2008). and decay of sea ice. Researchers need the continuity and enhancement of satellite 4. Reid, P. & Massom, R. in State of the Climate in to develop models of the atmosphere– observing programmes relevant to sea ice. 2014 (ed. Blunden, J. & Arndt, D. S.) Spec. Suppl. ocean–sea-ice environment at high spatial These organizations should partner with Bull. .Am. Meteorol. Soc. 96, S163–S164 (2015). resolution. funding agencies to make that happen. 5. Turner, J. & Overland, J. E. Polar Res. 28, 146–164 (2009). Better representations of the Southern 6. Antarctic Sea Ice Variability in the Southern CONNECT RESEARCH Ocean and its sea ice must now be a prior- Ocean–Climate System (National Academies Explaining the recent variability in Antarctic ity for modelling centres, which have been Press, 2017); available at go.nature.com/2trez1a sea ice, and improving projections of its focused on simulating the loss of Arctic sea 7. Hobbs, W. R. et al. Global Planet. Change 143, 228–250 (2016). future in a changing climate, requires pro- ice. Such models will be crucial to the next 8. Jones, J. M. et al. Nature Clim. Change 6, jects that bridge many disciplines. For exam- assessment of the Intergovernmental Panel 917–926 (2016). ple, the research communities involved in on Climate Change, which is due around 9. Holland, P. R. & Kwok, R. Nature Geosci. 5, ice-core analysis, historical data rescue 2020–21. A good example of the collabora- 872–875 (2012). and climate modelling need to collaborate tive projects needed is the Great Antarctic A list of co-signatories accompanies this Comment to track sea-ice variability over timescales Climate Hack (see go.nature.com/2ttpzcd). online (see go.nature.com/2tjeebi).

©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved. ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts r20eser vJULYed. 2017 | VOL 547 | NATURE | 277