JD, I see that your graphic is SST anomalies, and that brings up a couple of questions.
1. Do the SST measuring methods accurately capture the surface temp of ice in it's various forms, snow, and melt ponds of various depths as compared to open areas of sea water?
2. Is there any significant relationship between SST's and air temps at 2m height? Do they measure SST via satellite? Do they do correlate the buoy readings with satellite data? And how do the measure the surface temps at 2m height? Buoys and interpolation?
My mind is boggled by the complexity of these ocean and climate systems and their interactions, from the micro to the macro, and the apparent high level of chaos in the larger aspects systems.
If I'm getting any of my basic info wrong, please correct me.
I'll preface this by stating I'm not an expert - but rather a systems analyst - but will take a stab at your questions.
In general, the accuracy of readings one gets is only going to be as good as the granularity of the sensors used. There can be huge variation due to local microclimates.
As the nature of the complexity - everything revolves around three factors - the melting point of ice, and the vapor pressure of sea water at given temperatures, and the rate of transfer of sensible heat into phase changes of water. The first two provide the system with symmetry points, the third determines the rate of change that will take place in the system. Regarding temperature, all buffer it. QED, where "cool" currently, the arctic is still within about -2C, +/- a degree or so, and in many cases may reflect the local effect of exposed sea water which runs at around that temperature.
1) - that's a pretty broad question, and the answer I think is best described as "Imperfectly". From remote sensing, it requires making assumptions regarding the energy balance at the surface, as expressed by black body radiation and albedo. It is as one might say, "only skin deep". It doesn't necessarily indicate net heat in what is underneath the surface, nor indicate what the temperature is above the surface.
2) - 2M temperature is mostly observed and then intrapolated directly between stations where they exist, but where they do not, is where various algorithms come into play.
An example of how this is done for the DMI can be found here in a link from skeptical science:
http://www.skepticalscience.com/print.php?r=288"The Danish Meteorological Institute (DMI) Arctic temperature data is an output of the latest operational model as used for meteorological forecasting by the European Centre for Medium-Range Weather Forecasts (ECMWF). This output is an average of all model points at 2m height, currently on a 0.5 degree grid over the most northerly part of the Arctic, above 80N. Because the number of land stations in the World Meteorological Organization (WMO) list above 80 degrees North is very small indeed (a handful), measured data inputs for the model must be supplemented by other sources for high resolution meteorological work. The models assimilate inputs from weather stations, drifting stations and buoys, radiosondes, aircraft, vessels and more recently infra-red and microwave satellite based sensors.
The data is used to create a full three dimensional deterministic model of the global atmosphere which can be run forwards in time so that dynamic atmospheric conditions (weather) and regional climate can be analysed and forecast (or reanalyzed using historical observations as inputs) to the resolution limits of the model.
The resolution of these operational models has continually increased to take advantage of increased computing power and higher spatial resolution satellite data. This means that the DMI Arctic temperature data has had several changes in its history.
Between 1958 and 2002, the output of the ECMWF 40 year reanalysis (ERA-40) is used, (approximately 120km grid horizontal resolution). The ERA-40 re-analysis itself has three main sections of assimilated data, using pre-satellite observations up to 1972, then assimilating some satellite sensor observations, starting with the Vertical Temperature Profile Radiometer (VTPR) on early NOAA satellites, up to 1987/1988, and then using more recent observation types and more satellite sensors, both Infra Red and Microwave, in combined sensor packages such as the Television Infrared Observation Satellite (TIROS) Operational Vertical Sounder (TOVS), Special Sensor Microwave/Imager (SSM/I), Advanced TIROS Operational Vertical Sounder (ATOVS) amongst others, on platforms like the ERS (European Remote Sensing Satellite) and later NOAA series of satellites from 1987 onwards. In 2002 the DMI data switches to a higher resolution T511 model (40km resolution), then in 2006 to T799, (25km) and from 26th Jan 2010, T1279 (16km). These changes could be linked to minor differences in the apparent Summer melt temperature, (there are small differences between the ERA-40 and the T511 outputs in the overlap period in 2002)."
Other sources for 2M temperatures use similar methods and data sources.
