Embedded Files
Most other institutions use "sea ice extent" instead of "sea ice area". So why does CryosphereComputing not follow other institutions? The most important reason for extent is probably tradition. The first passive microwave measurements were taken in the 1970s when the world wasn't concerned about climate change. During that time the bigger concern was ships getting stuck in sea ice. With passive microwave measurements scientists were able to produce daily ice maps for the whole Arctic, which is a big step forward from aircraft or ground based reports. To get a reasonable ice edge (dangerous zone) a grid cell concentration of 15% was chosen. For the aim of providing safe passage for sea faring vessels sea ice extent is a reasonable metric, but for climate change analysis it is not an accurate description of the ocean surface.
The articifcal area increase hides how close to an ice-free state the Arctic really is.
The 2016 September daily minimum was anounched as 4.14 million km2, but the true sea ice area was only 2.34 million km2.
A good description of the minimum would be: The 2016 September minimum ice cover was 2.34 million km2 spread out over an area of 4.14 million km2.
If we have a completely ice free ocean then the additional area soaking up solar energy is just 2.34 million km2 and not 4.14 million km2.
If this picture were a "grid cell" it would be counted as an intact rainforest using the extent metric. (23% rainforest concentration)
If this kind of measurement is not acceptable for a rainforest, why should it be ok for sea ice?
Sea ice extent is a binary system, stating either there is ice in a grid cell or there isn't ice in a grid cell. The concentration threshold at which a grid cell is counted as ice covered is typically 15%.
One grid cell in NSIDC 25km grid has an area of 625 km2
At 15% concentration the true ice area is 93.75 km2
So the extent metric can overestimate the true area by 666% (625 km2 / 93.75 km2 )
This is rarely the case, but in summer and autumn 200% overestimations are not untypical.
Both grid cells fulfill the 15% threshold and are counted as completely ice covered using the extent metric.
partial ice coverage
dispersed ice
NSIDC Area is calculated as the sum of all pixels with sea ice concentration between 15% and 100%. The actual area of a pixel is included in the calculation. Lake ice is ignored to reduce noise from coastline measurement errors.
In winter, new sea ice is very thin and not a solid ice pack. Sometimes it forms so-called Pancake_ice, which is very prone to melt in stormy conditions. In recent years this type of ice cover has increased dramatically over the Arctic. You don't need to be a climate scientist to see that this type of ice cover has different thermodynamic properties (heat exchange between bodies) than a solid ice pack. The albedo (absorption of sunlight) and energy exchange between ocean-atmosphere is somewhere between an ice free ocean and a solid ice pack. So thermodynamically 50% ice concentration is quite accurate even though visually the ice concentration is close to 90%.
Special Summer Scenario
23 Jul 2012, Arctic Ocean, NASA (EOSDIS) Worldview
Special Winter Scenario
26 Jan 2017, Sea of Okhotsk, NASA (EOSDIS) Worldview
During the melting season melt ponds lower the sea ice albedo and change thermodynamic interactions with the atmosphere. The measured sea ice concentration reflects this change.
The low concentration areas have high melting potential, either from melt ponds' low albedo, a very dispersed ice pack or both combined. Their high surface-area to volume ratio results in a high energy transfer from ocean, atmosphere or direct sunlight. (law of surface-area-to-volume)
For the whole ice-pack the compaction ratio (Area / Extent) gives a very good indication of how quickly the sea ice will melt.
Page updated
Google Sites
Report abuse