Poles apart: why has Antarctic sea ice increased, and why can't coupled climate models reproduce observations?

Lead Research Organisation: NERC British Antarctic Survey
Department Name: Science Programmes

Abstract

Due to its pale colour, sea ice reflects much of the incoming solar radiation back into space, keeping local temperatures relatively cold. However, if warming occurs and sea ice melts, it is replaced by darker ocean. This absorbs more solar energy, causing warming, and so the cycle, the so-called 'ice-albedo feedback' loop, continues. Sea ice also modifies the regional surface energy balance by capping the upper layer of the ocean, reducing its loss of heat to the atmosphere. In addition, sea ice is important because it plays a role in the exchange of carbon dioxide between the atmosphere and ocean, thereby affecting how much of this greenhouse gas is in the atmosphere and contributing to global warming. Moreover, sea ice formation is an important element in driving the global thermohaline circulation of heat and salt through the world's oceans. One component of this circulation is the North Atlantic Drift current that carries warm tropical water across the Atlantic and keeps the UK's winter temperatures much warmer than they would be otherwise.

The Intergovernmental Panel on Climate Change (IPCC) assessment reports are an important tool in drivng government policy around the world. However, the present generation of climate models, which are used to predict the future climate scenarios described in these reports, are unable to consistently reproduce the recent increase in Antarctic sea ice. As a result considerable uncertainty must be attached to their predictions of future climate.

This proposal aims to both advance our understanding of the Earth's climate and facilitate improved predictions of its future change to aid policy makers. This will be achieved through the following objectives:

1. To explain the key climate processes involved in the recent Antarctic sea ice increase. We know from observations that changes in the near-surface wind around Antarctica are predominantly responsible for the observed increase in sea ice but we don't know exactly how the wind and the ice interact. Using a state-of-the-art computer model of sea ice and the ocean forced by the latest atmospheric data we will establish the key processes through which changes in the wind are causing the ice to increase.

2. To establish the ultimate driver of the sea ice increase. Policymakers need to know whether we can attribute the observed changes in Antarctic sea ice to human activity. This might happen through changes in the near-surface winds around Antarctica caused by the 'ozone hole' or greenhouse gas increases for example. Alternatively, it may be simply due to natural variations in the Antarctic climate system. If the former is true, we must determine which human activities are responsible. If the latter is correct, we must try to understand connections between the key processes and wider aspects of the climate system.

3. To understand why current climate models fail to simulate the growth in Antarctic sea ice. We will examine the current UK climate model in detail to diagnose which components are to blame and, with our Met Office partner, we will design a development programme to ensure that our findings are transferred into future model improvements in time for the next IPCC report. To help other climate model developers around the world, we will also analyse whether the failings are common to the other models used in the IPCC reports.

Planned Impact

Due to the role of Antarctic sea ice in key components of the Earth's climate system, such as atmosphere-ocean CO2 exchange and the thermohaline circulation, at its highest level the science needs to be communicated to policymakers, the public and to anyone at risk from the effects of climate change. Clarity is required urgently on the issue of the Antarctic sea ice increase because failure to explain the opposing trends in the Arctic and Antarctic has the potential to increase public uncertainty about the validity of climate change: indeed, climate sceptics have already attempted to use the increase in Antarctic sea ice in their arguments. Furthermore, the latest climate models used as the basis for the IPCC reports fail to simulate the increase in Antarctic sea ice, bringing into question the value of their projections of future climate change in the polar regions.

As part of this proposal we will determine (i) whether we can attribute the recent observed increase in Antarctic sea ice to changes to human activity or natural climate variability, and (ii) why the current generation of coupled climate models are unable to reproduce the positive trend in sea ice. These outputs will lead to improved prediction of Antarctic sea ice over the 21st Century, help to inform both policymakers and public about future climate change, and enable the government to rebuff the claims of climate sceptics when being criticised for their current mitigation policies.

In the UK our findings will contribute directly to future climate model development to ensure that (i) the Met Office can continue in its remit to provide up-to-date, robust and traceable scientific evidence to government on climate variability and climate change, and (ii) UK climate modellers maintain their high-standing at the leading edge of climate research, thus allowing the UK to sustain its influential position in future climate negotiations. The two PDRAs will develop professional skills that have applicability to other employment sectors, such as adeptness in communication, gained from courses on communicating science to the public.

Publications

10 25 50
 
Description Based on a novel sea ice budget analysis, we demonstrated that circulation changes generated by the El Niño Southern Oscillation (ENSO), a primary mode of natural climate variability at southern high latitudes, are associated with an east-west dipole of Antarctic sea ice concentration in the South Pacific ocean. This couplet of high-low anomalies propagate eastwards, partly driven by mean sea ice drift. Thus, we showed that natural, tropical atmospheric variability can have a significant impact on Antarctic sea ice variability and is likely to have contributed to the previous increase in regional sea ice extent. A further key finding is that linkages between sea ice anomalies and atmospheric variability can be highly non-local in space and time and, therefore, future attribution studies must consider such temporally and spatially remote connections.
Based on a detailed analysis of a 'state-of-the-art' ocean model, we showed that vertical mixing, freshwater forcing and initial sea ice conditions need to be constrained simultaneously to reproduce accurately the Southern Ocean hydrology and Antarctic sea ice in a model. Given that such processes are often poorly replicated in current coupled models and that such models have significant biases in sea surface temperature, salinity, mixed layer depth and sea ice extent, together with excessive open ocean deep convection, it is unsurprising that these models generally fail to reproduce the observed Antarctic sea ice increase.
Based on a series of chemistry-aerosol-climate model runs, driven by changes in greenhouse gases, ozone, aerosols and different combinations of these, we analysed the relative impact of these various forcings on high-latitude Southern Hemisphere circulation. This suite of experiments was novel in that the model code was modified especially in order to partition the radiation and chemistry schemes as required, so the ozone and aerosols could be allowed to interact or kept separate. A key finding was a non-linearity in the relationship between greenhouse gas and aerosol changes: i.e. that the sum of the changes associated with the two separate forcings did not equal the effect when they interacted. The greatest impact of this non-linearity was observed in the South Pacific, where it had a major impact on the atmospheric circulation. Its exact effect on Antarctic sea ice will be examined in work still to be undertaken. This scenario is relevant to the mid-late 20th Century, when there will be likely to have been further greenhouse gas increases, ozone recovery and with anthropogenic aerosols still relatively high.
Note that although the grant has finished there are still two further papers planned.
Exploitation Route Since the writing of this proposal the increase in Antarctic sea ice has stopped with consecutive years of much lower ice extent than previously seen in the satellite record. Thus, it seems highly likely that the previous increase was predominantly the result of natural variability, particularly as several studies in addition to this one have shown that tropical forcing is an important driver of changes in sea ice variability. The two principal findings (thus far) that are likely to be taken forward by other researchers are: (i) the importance of considering temporally and spatially remote connections between the atmospheric circulation when attributing sea ice variability and change, and (ii) the need to consider the effect of interactions between aerosols and increasing greenhouse gases when making projections of future circulation changes at southern high latitudes.
Sectors Environment,Government, Democracy and Justice

 
Title Global NEMO025-LM 
Description NEMO 3.6 stable + CICE 5.1, with TKE vertical mixing and modified lateral melting scheme due to prognostics ice floes sizes, monthly output of U,V,T,S, W, vertical diffusivity and monthly output of sea ice (Hice Aice, Uice, Vice, Internal stresses, ice tendencies) and wave information (HS, Tp) 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Utilised in a paper in preparation and results given in conferences. 
 
Title Global NEMO025-control 
Description NEMO 3.6 stable + CICE 5.1, with TKE vertical mixing and modified lateral melting scheme with constant ice floes sizes, monthly output of U,V,T,S, W, vertical diffusivity and monthly output of sea ice (Hice Aice, Uice, Vice, Internal stresses, ice tendencies) and wave information (HS, Tp) 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Utilised in a paper in preparation and results given in conferences. 
 
Title Global NEMO1-GLS-wave 
Description NEMO 3.6 stable + CICE 5.1, with modified GLS vertical mixing for wind effects, monthly output of U,V,T,S, W, vertical diffusivity and monthly output of sea ice (Hice Aice, Uice, Vice, Internal stresses, ice tendencies) and wave information (HS, Tp) 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Utilised in a paper in preparation and results given in conferences. 
 
Title Global NEMO1-GLS-wind 
Description NEMO 3.6 stable + CICE 5.1, with modified GLS vertical mixing for wind effects, monthly output of U,V,T,S, W, vertical diffusivity and monthly output of sea ice (Hice Aice, Uice, Vice, Internal stresses, ice tendencies) and wave information (HS, Tp) 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Utilised in a paper in preparation and results given in conferences. 
 
Title Global NEMO1-LM 
Description NEMO 3.6 stable + CICE 5.1, with TKE vertical mixing and modified lateral melting scheme due to prognostics ice floes sizes, monthly output of U,V,T,S, W, vertical diffusivity and monthly output of sea ice (Hice Aice, Uice, Vice, Internal stresses, ice tendencies) and wave information (HS, Tp) 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Utilised in a paper in preparation and results given in conferences. 
 
Title Global NEMO1-control GLS 
Description NEMO 3.6 stable + CICE 5.1, control 2 with GLS vertical mixing, monthly output of U,V,T,S, W, vertical diffusivity and monthly output of sea ice (Hice Aice, Uice, Vice, Internal stresses, ice tendencies) and wave information (HS, Tp) 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Utilised in a paper in preparation and results given in conferences. 
 
Title Global NEMO1-control TKE 
Description NEMO 3.6 stable + CICE 5.1, control 1with TKE vertical mixing lateral melting scheme with constant ice floes sizes, monthly output of U,V,T,S, W, vertical diffusivity and monthly output of sea ice (Hice Aice, Uice, Vice, Internal stresses, ice tendencies) and wave information (HS, Tp) 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Utilised in a paper in preparation and results given in conferences. 
 
Title HadGEM3A-xlaya 
Description pre-industrial control simulation (HadGEM3 + UKCA), 20 years of spin up then 35 year experimental phase simulation. Atmosphere only simulation with fixed SSTs and Sea Ice from HadGEM2-CC CMIP5 historical simulation. SSTs & sea ice were a 12 month average of 1861-1900. All atmospheric chemistry and greenhouse gases set as an average of the same period. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Results presented at AGU 2015 Fall Meeting 
 
Title HadGEM3A-xlayc 
Description GHG simulation. A copy of the xlaya but with present day greenhouse gas concentrations. It ran for 35 years for the experimental phase 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact results presented at AGU 2015 Fall Meeting 
 
Title HadGEM3A-xlaye 
Description El Nino simulation. A copy of xlaya but with an El Nino anomaly added to the SST and Sea Ice fields. It ran for 35 years for the experimental phase 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact None 
 
Title HadGEM3A-xlayf 
Description La Nina simulation. A copy of xlaya but with a La Nina anomaly added to the SST and Sea Ice fields. It ran for 35 years for the experimental phase 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact None 
 
Title HadGEM3A-xlayg 
Description A copy of the pre-industrial control (xlaya), but with a file to include anthropogenic aerosol emissions. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Contribution to a paper in preparation. 
 
Title HadGEM3A-xlayi 
Description All Forcings - A copy of the pre-industrial control (xlaya), but with the CFCs from xlayp, greenhouse gases from xlayp and anthropogenic aerosols from xlayg. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Contribution to a paper in preparation. 
 
Title HadGEM3A-xlayo 
Description A copy of the pre-industrial control (xlaya) but with GHG levels from xlayc and CFCs in the chemistry scheme based on xlayp 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Contribution to a paper in preparation. 
 
Title HadGEM3A-xlayp 
Description Ozone simulation - A copy of the pre-industrial control (xlaya) but with CFCs included in the chemistry scheme to create an ozone hole. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Contributing to a paper in preparation. 
 
Title HadGEM3A-xmxna 
Description Pre-industrial controi HadGEM3 simulation 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Part of a collaboration with Kyle Clem at the Victoria University of Wellington, New Zealand. 
 
Title HadGEM3A-xmxnb 
Description As xmxna but with 2 deg C warmer SSTs in the West Pacific 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Part of a collaboration with Kyle Clem at the Victoria University of Wellington, New Zealand. 
 
Title HadGEM3A-xmxnc 
Description As xmxna but with 2 deg C warmer SSTs in the Indian Ocean 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Part of a collaboration with Kyle Clem at the Victoria University of Wellington, New Zealand. 
 
Title HadGEM3A-xmxnd 
Description As xmxna but with 2 deg C warmer SSTs in the West Pacific and Indian Oceans. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Part of a collaboration with Kyle Clem at the Victoria University of Wellington, New Zealand. 
 
Title ORCA1-005fr4R 
Description NEMO 3.6 alpha + CICE 5.0.4, monthly output As 3T4R, but with rn_ediss = 0.1 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Used in Ocean Modelling paper 
 
Title ORCA1-3T2R 
Description NEMO 3.6 alpha + CICE 5.0.4, monthly output As 3T4R, but runoff according to estimates by Rignot et al. (2013). 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Used in Ocean Modelling paper 
 
Title ORCA1-3T4R 
Description NEMO 3.6 alpha + CICE 5.0.4, monthly rn_ediss = 0.3 and 4000 Gt/yr runoff with concentration along Antarctic coast 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Used in Ocean Modelling paper 
 
Title ORCA1-3T4RE 
Description NEMO 3.6 alpha + CICE 5.0.4, monthly output As 3T4R, but evenly distributed runoff south of 60S. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Used in Ocean Modelling paper 
 
Title ORCA1-3T4RNI 
Description NEMO 3.6 alpha + CICE 5.0.4, monthly output As 3T4R, but without initial sea ice. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Used in Ocean Modelling paper 
 
Title ORCA1-3T4RSSR 
Description NEMO 3.6 alpha + CICE 5.0.4, monthly output As 3T4R, but using freshwater fluxes from time-averaged salt restoring. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact Used in Ocean Modelling paper. 
 
Title PERIANT025-PC_noGM-rnf-rignot 
Description NEMO 3.6 stable + CICE 5.1, 5-day output of U,V,T,S,rho and monthly output of sea ice As PC_noGM-ts-tke1-corr-R, but with runoff following Rignot et al. (2013) 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact None 
 
Title PERIANT025-PC_noGM-rnf-trend 
Description NEMO 3.6 stable + CICE 5.1, 5-day output of U,V,T,S,rho and monthly output of sea ice As PC_noGM-rnf-rignot, but with linear trend in runoff and ice shelf melt rates in Amundsen and Bellingshausen Seas. From 1900 Gt/yr in 1979 to 2300 Gt/yr in 2013. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact None 
 
Title PERIANT025-PC_noGM-ts-tke1-corr-R 
Description NEMO 3.6 stable + CICE 5.1, 5-day output of U,V,T,S,rho and monthly output of sea ice no eddy-induced velocities, time-split free surface, added extra southward volume flux at 30S. First run 1979-1988 and then restarted at 1979. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact None 
 
Description SST model simulations 
Organisation Victoria University of Wellington
Country New Zealand 
Sector Academic/University 
PI Contribution Providing bespoke model simulations based on those produced for the Poles Apart grant
Collaborator Contribution Undertaking analysis of the model results to assess how changes in tropical SSTs correlate with surface air temperature changes over Antarctica.
Impact Paper in preparation.
Start Year 2016
 
Title Mixing modules in NEMO 
Description New ocean mixing modules for the NEMO system 
Type Of Technology Software 
Year Produced 2017 
Open Source License? Yes  
Impact The modelling community is informed on the model development, which will be used as an open source for the scientific research. 
 
Description Answering press questions (CNN) 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact Providing answers to press questions from Brandon Miller (CNN) on Antarctic sea ice minimum
Year(s) Of Engagement Activity 2017
URL http://edition.cnn.com/2017/02/16/world/antarctica-sea-ice-record-low/
 
Description Answering press questions (Climate Central) 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact Skype interview with Andrea Thompson (Climate Central) on Antarctica Sea Ice Minimum (article in preparation for their website).
Year(s) Of Engagement Activity 2017
 
Description Climate change outreach - Manchester museum of science and industry 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact James Pope was part of the BAS 'stall' at this event.
Year(s) Of Engagement Activity 2015
 
Description I'm a scientist, get me out of here 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Schools
Results and Impact PDRA, Dr James Pope took part in the online schools national outreach event, I'm A Scientist, Get Me Out of Here. He was fortunate enough to win the Dysprosium Zone, the winner being decided by the votes of the school students in the zone, who have spent the preceding two weeks asking questions of the scientists in the zone.
Year(s) Of Engagement Activity 2015
URL http://imascientist.org.uk/
 
Description Schools career talk 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Schools
Results and Impact James Pope gave a careers talk to 80 sixth form pupils at Hill House school, Doncaster
Year(s) Of Engagement Activity 2016
 
Description Statement for BAS website 
Form Of Engagement Activity Engagement focused website, blog or social media channel
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact Statement produced on 2017 Antarctic sea ice minimum as a news story for BAS website
Year(s) Of Engagement Activity 2017
URL https://www.bas.ac.uk/media-post/antarctic-sea-ice-extent-lowest-on-record/
 
Description Talk at Scott Polar Research Institute 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Public/other audiences
Results and Impact James Pope gave a presentation on his work on Antarctic sea ice as part of the Poles Apart project
Year(s) Of Engagement Activity 2015