Here's some interesting discussion from Judith Curry.
Historic Variations in Arctic sea ice. Part II: 1920-1950 Among a number of other things, it basically states that the ice extents data prior to the satellite record is shite.
Actually that's not what they say at all. Color me unsurprised that you lie about it.
Liar
Historic Variations in Arctic sea ice. Part II: 1920-1950
How reliable are the sea ice observations during the first half of the 20th century?
During the period 1920-1940, the Arctic Ocean was regarded as relatively unexplored territory. A 1934
National Geographic article discusses the Lindbergh’s flights over the Arctic:
“National Geographic, Sept. 1934, pp. 261-304, “Flying Around the North Atlantic” by Anne Morrow Lindbergh. She and her husband crossed the icecap twice, followed the coastline, from Disko Bay area to Clavering Island, visited Dr.Lauge Koch’s research area, and literally ”redrew” the map of some areas of Greenland.”
The lack of data from Russia and other areas because its sheer remoteness made accurate data gathering problematic, large gaps in knowledge due to WW2 i.e no DMI maps (Danish Meteorological Institute) during this period, a hiatus in cooperation during the Cold war, uncertainty over final summer amounts as DMI did not usually record data after August, and estimates by climatological methods means historic sea ice data is fragmented and incomplete. This is not helped by the different sea ice data bases often being at considerable variance with each other and changes in the way in which ice concentrations were calculated as described below:
“In 1968, The United States began reporting ice concentration in eights rather than tenths, then returned to tenths reporting in 1980. Canada retained the tenths format throughout the period. […….] SIC summary analyses in eights frequently overlay reconnaissance data which is plotted in tenths. Since SIC produced these charts for in-house, not public use, which of these formats is used is seldom noted on the charts….There is also a period when U.S. recon data receives a slightly different treatment from chart to chart…… With a good knowledge of codes and the history of code changes and a dash of intuition and care, it is possible to roam the data set without great fear of misinterpretation. Others entering without some preparation may find it hazardous.” [
link]
Of course, professional arctic sea ice researchers are aware of these factors and adjust data accordingly but the preceding does raise the question as to whether X amount of ice in the satellite era (1979 onwards) is really the same as X amount in the period prior to that, derived through climatological or physical observations in often difficult conditions by such as whalers, which brings us to the thorny question of what the definition of ice extent actually is.
This concern led me to ask the following question of NSIDC:
“ …..how did pre satellite researchers estimating sea ice extent tell the difference between water, water floating on ice, and solid ice, and how can satellites differentiate between the three states? I was struck by Russian reports from the 1950’s at The Scott Polar institute in Cambridge when staff at the floating research stations commented about using Wellington boots in order to walk around the station, and how little dry ice islands eventually formed by the end of the summer surrounded by water on top of ice.”
I received the following reply from Julienne Stroeve (reproduced with permission):
“ … using passive microwave data it is very easy to tell the difference between ice and water as the dielectric constant differs quite a bit and this is reflected in large differences in the microwave emission. The main advantage of using passive microwave is that it can see the ice even if it’s cloudy or dark. There is a problem however in summer when melt ponds form on the ice since the sea ice algorithms then underestimate how much ice there really is (they think it’s open water). That’s one reason why we focus on extent rather than true ice area for the NSIDC sea ice news and analysis web site.
Visible and thermal imagery provides higher spatial resolution but is often hampered by clouds. Trying to do this work using earlier visible and thermal imagery requires the scientists to go through each image and manually filter out the clouds and determine where the ice is.
The Arctic in those pre satellite days was simply too large to be effectively and continually monitored. Observers getting close enough to the ice edge to make physical observations might be deterred from proceeding further by apparently impenetrable ice although better, more open conditions, might lie beyond. Data from such as the Russian sector – where much warming occurred, was not always taken into account. (However, the reader should be aware that, as Larsen noted, ice did sporadically return whilst reports from 1939-45 are sparse for obvious reasons.)
In trying to determine the true extent of sea ice during the period we run the risk of comparing apples-physical observations, and oranges- satellite altimetry, and the different methods employed over the years creates uncertainties over whether each accurately picked up what is ice, what is water covered ice and what is open water. This makes it difficult to determine how modern ice extent compares to the past with any certainty.
