Sea-surface temperature (SST) and ice concentration

Sea-surface temperature is taken from various providers, who process the observational data in different ways.  Each provider uses data from several different observational sources.  Satellites measure sea-surface temperature (SST) in a layer a few microns thick in the uppermost mm of the ocean, drifting buoys measure SST at a depth of about 0.2-1.5m, ships sample sea water down to about 10m while the vessel is underway. Diurnal variations are not seen at deeper measurements. At ECMWF, sea-surface temperature are initialised using analyses received daily from the Met Office (OSTIA, 5 km resolution).  Sometimes sea-surface temperature is taken from a forecast made by coupling the NEMO ocean model to the ECMWF Integrated Forecasting System.  In this case, the SST is the average temperature of the uppermost metre of the ocean and does exhibit diurnal variations.

Sometimes sea-surface temperatures derived from satellites are shown over lakes or over model a land areas near the coast.  These are set to that of the Flake model mixed layer temperature.  This gives a much smoother field than Cy48r and earlier, helping to give much more realistic values around coastlines. 

Additionally NEMO, and the LIM2 subprogram within it, forecast changes in the sea-surface temperature (SST) and sea ice evolution.  The Louvain-la-Neuve Sea Ice Model (LIM2) is a prognostic sea-ice model that deals with the dynamic and thermodynamic evolution of the sea surface so that sea-ice cover evolves dynamically.  It is incorporated into the dynamic ocean model (NEMO).    Output from these programs are used interactively by all IFS atmospheric models.  Ensemble Control Forecast (ex-HRES) uses the same initial ice extent as medium range ensemble.  

Note: ECMWF uses LIM2 which is an earlier version of the Louvain-la-Neuve sea ice model currently available (Version 3.6).

Throughout the forecast period the changing extent of sea-ice and the variation of the ice shelf with time have important effects upon the energy and moisture balance at the atmosphere/surface boundary. The ice extent will change through the forecast period in response to sea temperatures and air temperatures, ocean currents and wind.  Melt ponds existing or forming on otherwise extensive sea ice can reduce albedo locally.

The impacts of differently-evolving SST and ice cover distributions should be considered when comparing different forecasts, even when they are from the same data time.  Ice cover can be extensive in fjords and inlets but may be too small to be shown on ice charts.  The local ice cover may influence forecast parameters nearby (e.g. 2m temperature, fog, etc.).  The Baltic Sea and Bay of Bothnia can be particularly affected.

The impact of waves on the sea ice is not yet at completely represented in ECMWF models.  There is particular uncertainty when the wave height at the sea ice edge is intense.

Sea ice cover

Sea ice cover evolves according to the sea ice model LIM2.  But in general the LIM2 sea ice model tends:

  • to melt sea ice too slowly, particularly where the analysis tends to have a bias towards greater sea ice thickness (e.g. in the Beaufort Sea).  
  • to form sea ice too slowly, broadly where sea temperatures are slightly warmer and where there is a negative bias (e.g. off northeast Greenland). 

The averages hide a lot of features and the bias in ice thickness in LIM2 is not uniform.

Incorrect ice analysis and/or forecast near coasts can give anomalous T2m temperatures at nearby land locations.  Too much ice tends to make land temperatures too cold, too little tends to keep land temperatures nearer sea temperatures.

Fig2.1.4.8-1: General bias of ice thickness from LIM2.  The scale is in m.  Red too thick, blue too thin.  The averages hide a lot of features and the bias in ice thickness in LIM2 is not uniform.


Sea ice extent 

The climate location of the ice edge is based on a near 30year periods but the area of winter ice cover has been seen to be reduced over recent years.  Maximum area of antarctic sea ice in the southern winters in 2023 and 2024 (not shown) are also lower than the available long term satellite records.  The climate of ice cover, particularly during recent years, probably generally covers too great an area.   This can have an impact on forecast temperatures and the derivation of Extreme Forecast Index in polar regions. Users should inspect the observed sea ice extent and compare with that of the the model climate.


Fig2.1.4.8-2: Sea ice concentration in early January 2025 compared with the Median ice edge over the period 1981-2010 (orange line).  Observed arctic sea ice extent is taken from Sea Ice Today from the National Snow and Ice Center (NSIDC) at University of Colorado Boulder.   

Fig2.1.4.8-3: Observed arctic sea ice extent has become less than the general extent during previous years.  Over recent years the extent of ice has been less than the median winter extent.

Ice charts

The extent of ice cover is shown with different thresholds on Opencharts and ecCharts..

Ice cover on Opencharts

Fig2.1.4.8-2: Example of sea temperatures and sea ice fraction shown by Opencharts.  Sea ice fraction (cyan hues, %) is shown where ice coverage is greater than 50% but most of the marginal ice zone (15%-50%) is not plotted.  This gives the impression that there is no sea ice outside this area.  However, patchy ice and bergs are likely in the surrounding waters.Ice in smaller discontinuous patches or as bergs can occur elsewhere mainly in purple coloured areas but particularly where sea temperatures are sub-zero.  Current sea temperatures and ice cover chart on Opencharts.

The climatological location of the sea ice edge is shown for reference as an overlaid contour (magenta) which shows where sea ice fraction is on average >50%, and so can be compared with the edge of the turquoise sea ice areas; stippling highlights climatological ice cover above 50%.  The climate location of the ice edge is based on an earlier period and may show too great an area of climate ice cover.


Sea temperatures anomaly on Opencharts.

Fig2.1.4.8-3: Example of sea temperature anomaly Openchart.  Temperature anomalies are shown only where the current sea ice fraction is <50%.  Sea ice fraction (cyan hues, %) is shown where ice coverage is greater than 50% but most of the marginal ice zone (15%-50%) is not plotted.  In areas of large sea ice depletion relative to climatology (i.e. inside the magenta contour), the sea surface temperature anomaly may be unreliable (or zero (white)) due to a lack of past sea surface temperature data.  Current sea temperatures anomaly on Opencharts.

The climatological location of the sea ice edge is shown for reference as an overlaid contour (magenta) which shows where sea ice fraction is on average >50%, and so can be compared with the edge of the turquoise sea ice areas; stippling highlights climatological ice cover above 50%.  The climate location of the ice edge is based on an earlier period and may show too great an area of climate ice cover.

Ice cover on ecCharts

Fig2.1.4.8-4: Example of ecChart display of sea ice cover as fraction of a grid box. - only one range of coverage (0.2 to 1.0 fraction of grid box).  The extent of ice is larger than Opencharts (which only shows ice cover to 50%) and includes broken ice cover but without any detail.


Fig2.1.4.8-5: Example of ecChart display of ensemble mean of sea ice cover as fraction of a grid box. - range of coverage (0.1 to 1.0 fraction of grid box).  The extent of ice is larger with more detail including probability of the broken ice coverage.


Fig2.1.4.8-6: Example of ecChart display of sea ice cover as fraction of a grid box and ensemble mean for sea ice cover as fraction of a grid box displayed together.  The green fringe shows approximate areas between 0.1 and 0.2 sea ice coverage (fraction of grid box). 

Example of sea ice forecast sequence

Fig2.1.4.8-7: Sequence of sea-ice and sea-surface temperatures from the ENS CTRL run data time 00 UTC 27 April 2017.   T+0hr (00UTC 27 April 17), T+120hr (00UTC 02 May 17),  T+240hr (00UTC 07 May 17), and  T+360hr (00UTC 12 May 17).  On such plots the climatological average sea ice cover is shown in pink (contour and stippling, for >50%), just discernible in the northern Gulf of Bothnia and in the White Sea.   Dark purple areas (SST between 0C and -2C) are prone to ice formation if not already in existence.   Areas of sea ice are shown as turquoise. 

Note:  

  • Movement of ice (turquoise) in the northern Gulf of Bothnia  due to the winds.
  • Steady rise of sea-surface temperatures in the Black Sea, and especially in the shallow waters of both the Sea of Azov and the northern Caspian Sea.  In the White Sea (east of Finland, top of plot) sea ice cover is less than the climatological average for this time of year.  Using these plots, the user can assess where sea ice cover is above/below average.