This website uses cookies, which are small text files that are used to make websites work more effectively. In order to continue using this website, you will need to accept the use of cookies.

In a study recently published by the EGU journal The Cryosphere, and led by Dr Ed Blockley, manager of the Polar Climate Group within the Met Office Hadley Centre, the impact of including sea-ice thickness information within the initial conditions of a global, coupled seasonal forecasting system has been investigated for the first time.

Although model-based forecasting systems have been used successfully for seasonal prediction of mid-latitude weather and climate for some time, now their application to Arctic sea-ice prediction is much less mature. Although the use of these dynamical models for seasonal sea-ice prediction is still in its relative infancy, several studies have demonstrated skill in retrospective forecasts of September Arctic sea ice extent made from spring. However, these studies were not able to match the potential skill suggested by so-called “perfect model” studies, where the initial state of the model is known precisely. This implies that there is potential for further improving the skill of these models by performing seasonal forecasts from a more realistic initial state. Furthermore, several studies over the last decade have found that winter sea-ice thickness provides important preconditioning for the evolution of Arctic sea ice through the summer melt season. This suggests that seasonal predictions of Arctic summer sea ice, made by these forecast models, could be improved by specifying a more realistic representation of sea-ice thickness when initialising these models.

“We used the Met Office’s Global Seasonal coupled ensemble prediction system and incorporated satellite sea-ice thickness measurements, derived from the European Space Agency’s CryoSat-2 satellite by the Centre for Polar Observation and Modelling, within the initialisation process” explains Dr Blockley. “We compared forecasts from a large ensemble of GloSea seasonal predictions, performed over the period 2011-2015, against a control experiment that did not include any observations of sea-ice thickness. We show a significant improvement in forecasts of September Arctic sea-ice edge location made from May, with a reduction of Integrated Ice Edge Error of around 37%.”

Human activity in the Arctic is increasing in response to declining sea-ice cover and access to the Arctic Ocean is becoming more important for socio-economic reasons. Human activities in the Arctic include commercial ventures like tourism, mineral and oil extraction, fishing and shipping, as well as traditional activities such as subsistence hunting and fishing, search and rescue, and community re-supply. Accurate forecasts of Arctic sea ice on all timescales are thus becoming increasingly important for the safety of human activities in the Arctic. In particular, improved knowledge of sea ice on seasonal timescales (approx. 90 days) allows for improved planning which should lead to a reduction in risk and operational costs for human activities in the Arctic Ocean.


Blockley, E. W. and Peterson, K. A.: Improving Met Office seasonal predictions of Arctic sea ice using assimilation of CryoSat-2 thickness, The Cryosphere, 12, 3419-3438,, 2018.

September forecast probability of Ice