Figure Legend:
Each image shows the average sea-ice concentration for the western Antarctic Peninsula (WAP) region for the month and year in question. The Palmer LTER large-scale study is outlined by the white box. Each monthly image also has solid and dotted contours in addition to the grayscale gradient showing the percent ice coverage. The solid white contour delineates the 15% ice concentration contour (i.e., what is typically defined as the ice edge), while the dotted white contour is the mean 15% contour (for the defined time period). Similarly, the solid black contour (within the ice field) delineates the 75% ice concentration contour while the dotted black contour is the mean 75% contour.
Data Info:
The WAP sea-ice concentrations were derived from the GSFC Bootstrap algorithm made available by the National Snow and Ice Data Center (NSIDC). The images from 1/1991 to 12/1999 are from the SMMR-SSMI time-series while the images from 1/2000 to 6/2001 are from the SSMI time-series
NOTE:
There are two different time-series one could use for 1990's data: the combined SMMR(1980s)-SSMI(1990s) intercalibrated time-series or the SSMI time-series. The intercalibrated SMMR-SSMI time-series has many other improvements besides the intersensor calibration: an improved landmask, better error checking, and better masking of false ice retrievals due to weather or coastal contamination. However, it takes longer for these data to appear at NSIDC, so currently the improved SMMR-SSMI time-series is only available from 10/78 to 12/99.
That unfortunately means that the 2000 and 2001 data shown on these web pages have not been intercalibrated with the rest of the time-series. The difference between the SSMI data from the SMMR-SSMI time-series versus the unadjusted SSMI time-series is illustrated under the third bullet in item #4. Those figures show that the intercalibration (and other adjustments) mostly affects the summer estimates whereby the SSMI unadjusted summer estimates are sometimes significantly higher (as for year 1999) than the intercalibrated estimates. The decrease in summer estimates in the SMMR-SSMI time-series is most likely due to the improved masking of false ice retrievals near coastal boundaries - a contamination problem particularly pertinent to the LTER study region.
One should note, however, that the SMMR-SSMI intercalibrated time-series has its own caveats (e.g. the intercalibration between the SMMR and F8 SSMI sensor is based on a very short period of overlap in the austral winter of 1987), but in general it is the preferred dataset to use, especially given the other improvements.