Realistic sea ice melt in climate models using field observations and theory

Lead Research Organisation: University of Reading
Department Name: Meteorology


In response to global warming, the ice covers of the Arctic and Antarctic are changing, with a significant, and accelerated, reduction in the summer extent of Arctic sea ice. The observed rapid reduction of Arctic sea ice is more extreme than the predictions of even the most pessimistic of climate models, which suggests that these models do not represent the processes controlling the reduction of sea ice adequately. Satellite observations, field work, and modelling all point to the importance of melt ponds in controlling the summer melt of Arctic sea ice but climate models do not explicitly represent melt ponds.
Melt ponds form when surface snow and sea ice meltwater accumulates on the sea ice surface. Since pond-covered ice has a lower reflectivity (albedo) than bare sea ice, pond-covered ice melts up to 3 times more rapidly than bare ice. This project aims to eliminate fundamental uncertainty in the processes controlling the melt of Arctic sea ice through a combination of recently collected, closely controlled field observations and computer modelling. This project will identify the contribution of melt ponds to the rapid reduction in Arctic sea ice. A result of the research will be a new representation of sea ice melt incorporated into the CICE sea ice model, which is the sea ice model used in many climate models, including the UK Meteorological Office climate model.

Planned Impact

Specific users who will benefit from this proposal

This proposal will generate new scientific knowledge and insight into the thermodynamic and radiative processes, including the role of melt ponds, that determine sea ice melt and thus the heat and mass budget of the Arctic Ocean. In particular, this new knowledge will benefit sea ice geophysicists and polar oceanographers seeking to understand the role of sea ice melt in recent remarkable changes in the Arctic ice cover.

A major practical impact of this proposal is in the generation of a more realistic treatment of sea ice melt in the sea ice component (CICE) of a Global Climate Model (GCM). The CICE sea ice component is used in the HadGEM3 GCM developed in the Hadley Centre UK Meteorological Office (UKMO) and the Community Climate System Model at the (US) National Center for Atmospheric Research.

CICE is also used in many sea ice-ocean models, e.g. within the UK, CICE is used in the National Oceanography Centre Southampton, the National Oceanography Centre Liverpool, the British Antarctic Survey, and the National Centre for Earth Observation.

Engagement with the user group

The PI plays a leading role in organising annual meetings of the UK Sea ice group, which regularly attracts the leading UK climate and polar modelling groups (and experimental and observational groups). This is an informal environment allowing for practical difficulties and issues to be discussed. The project results, and guidance on use of the new sea ice melt module, will be presented in a special session of this forum to the potential UK users.

In order to maximise user take up of the sea ice melt module, several multi-day visits to both NOC sites will be arranged.

The Met Office Hadley Centre climate model HadGEM3 uses a non-standard configuration of CICE to allow close coupling with the atmosphere model, with the surface ice heat budget being incorporated into the atmosphere model. For this reason, we will arrange two mid-project meetings in order to discuss the most practical way for the Hadley Centre to incorporate the new sea ice melt treatment into their model.

The PI convenes an annual session on sea ice processes and models at the European Geosciences Union meeting, and occasional sessions at the American Geophysical Union meeting, both of which would provide ideal forums to make the project work widely known to the international community.

As part of core NCEO activity, the new sea ice melt module will be tested and uploaded to the Los Alamos National Laboratory, who develop and maintain the CICE code, and incorporated into the latest version of CICE. This will maximise the impact of the project, since every user group downloading the latest version of CICE will automatically be using the new sea ice melt module.

We will expand the existing CPOM web site. This web site will provide illustrated information on sea ice melt and melt ponds for a more general audience, and will provide links to documentation of the sea ice melt module and the module itself.

In addition to the various meetings, conferences, and website, the results of the project will be published in high quality journals with a wide readership in the user community.

Wider user interest

The new treatment of sea ice melt will have implications for climate change prediction by affecting atmosphere and ocean coupling and the sea ice heat and mass balance. In particular, the GCMs affected by the new sea ice model are used to provide results for the Intergovernmental Panel on Climate Change, the reports of which are used to guide national and international policy on issues related to climate change.

The demonstrated impact of melt ponds will affect the climate model simulations used to inform government policy and those with financial and societal interest, such as the insurance and oil industries, and Inuit community groups.


10 25 50
publication icon
Feltham D (2015) Arctic sea ice reduction: the evidence, models and impacts. in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences

publication icon
Flocco D (2015) The refreezing of melt ponds on Arctic sea ice in Journal of Geophysical Research: Oceans

publication icon
Sansiviero M (2017) Modelling sea ice formation in the Terra Nova Bay polynya in Journal of Marine Systems

publication icon
Stroeve J (2018) Warm Winter, Thin Ice?

publication icon
Tsamados M (2015) Processes controlling surface, bottom and lateral melt of Arctic sea ice in a state of the art sea ice model. in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences

Description The presence of melt ponds on the surface of Arctic sea ice significantly reduces its albedo, inducing a positive feedback leading to sea ice thinning. While the role of melt ponds in enhancing the summer melt of sea ice is well known, their impact on supressing winter freezing of sea ice has, hitherto, received less attention. Melt ponds freeze by forming an ice lid at the upper surface which insulates them from the atmosphere and traps pond water between the underlying sea ice and the ice lid. The pond water is a store of latent heat, which is released during refreezing. Until a pond freezes completely there can be minimal ice growth at the base of the underlying sea ice.

We have developed a model of melt pond refreezing that accounts for the salinity distribution within the pond and other physical processes. Simulations with this model have demonstrated the significant impact that ponds have in supressing winter ice growth.
Exploitation Route Our results show that climate modelling groups have overestimated ice growth as they do not contain the correct physics of pond refreezing and this overestimate could be significant at leading order.

Before their models can be corrected, significant further work is required that is out of the scope of this project. The detailed model requirements have been identified and it is only lack of resource that prevent further progress.
Sectors Aerospace, Defence and Marine,Education,Environment

Description While climate sea ice models cannot yet account for pond refreezing as they lack a vertical heat balance equation accounting for the separate phases, the degree to which the climate models are overestimating ice growth can be assessed with several model experiments in which certain processes are disabled. This work is in progress and will inform modelling groups and climate prediction centres. In particular we are working with the UK Met Office.
First Year Of Impact 2014
Sector Aerospace, Defence and Marine,Education,Environment
Description Input to Arctic Select Committee of House of Lords - oral and written evidence
Geographic Reach Australia 
Policy Influence Type Gave evidence to a government review
Description NERC standard grant (response mode)
Amount £579,000 (GBP)
Funding ID NE/P001645/1 
Organisation Natural Environment Research Council 
Sector Public
Country United Kingdom
Start 03/2017 
End 03/2020
Description Sea Ice Prediction 
Organisation National Aeronautics and Space Administration (NASA)
Country United States 
Sector Public 
PI Contribution The inclusion of melt pond physics into the UK sea ice model led to improved predictive capability in the model. This led to a collaboration with scientist working in NASA on sea ice prediction (Dr. Alek Petty), who was a former student of the PI.
Collaborator Contribution Collaboration has led to a paper demonstrating enhanced skill in predicting the summer Arctic sea ice minimum extent. This paper has been accepted but not published yet.
Impact Improved sea ice predictability.
Start Year 2016
Title Input to CICE sea ice climate model 
Description Research projects have developed new physics of sea ice processes. Under separate funding, but in collaboration with research projects, this has been turned into new physics modules in the sea ice climate model CICE. 
Type Of Technology Physical Model/Kit 
Year Produced 2017 
Impact The CICE sea ice model is used by climate modelling groups worldwide. In the UK this includes the UK Met Office, NOC and BAS.