About the MMAB Ice Concentration Grids[Updates to be filled in 2001 to present. Major updates include August 2004, change to high resolution algorithm, 2006 revised land masks, 2009 change to using AMSR-E for concentrations] Last modified 12 April 2001
Skip down to full description
12 April 2001Land masks updated. See http://polar.ncep.noaa.gov/seaice/support/land.shtml for details.
1 July 1996The land mask for the polar stereographic ice grids was revised to exclude more of the coastal points. The previous grid permitted too many points which were contaminated due to land being within the pixel to be given sea ice concentrations. The new masks, with 157 being land and 195 being coast, are available in the ftp directory in files nland.map and sland.map (raster) or nland.psg and sland.psg (grib).
8 April 1996A global sea ice grid is now available for viewing. The global grid is a half-degree latitude-longitude grid. Unlike the hemispheric grids (described below), the global grid has data values at every point. Values for which there is no observation from the current day are filled in from yesterday's global map. The other significant difference from the hemispheric maps is that an additional filter is applied -- if the current sea surface analysis says that the temperature is greater than 2 C at a point, ice concentrations are reset to zero.
20 March 1996The sea ice fields were approved for operational implementation.
20 February 1996A new flag, 177, has been added to the data files. This flag is for weather. The web display of the data has been changed as well. First, weather is now shown as its own color (faint purple). Second, 'no data' is now displayed as a purple, rather than the fairly dark green it used to be. Third, areas which are marked as 'coastal' in the land mask are now flagged in gray as opposed to formerly, when they were marked in black as land.
13 February 1996At a few points, every few days, the weather filter is erroneously calling points weather when they are in fact ice covered. The short term response is to flag weather points as being 'no data'. This is the cause of the increase in 'no data' on the images. A separate flag for weather will be added once other software is prepared for the change.
ftp://polar.ncep.noaa.gov/pub/ice in files named northpsg.yymmdd, southpsg.yymmdd, where yymmdd is a 6 digit date. While the file name date should be correct, rely only the date encoded in the file.
If there have been no glitches in the system, keeping in mind that this is an experimental system, the images are updated by about 9:30 Eastern Time (Standard or Daylight, as prevails in the metropolitan District of Columbia area). The data used to generate the grids span plus or minus 12 hours of 00 UTC. Due to transmission times and the like, the most recent orbit or two (of the 14 which are available for a day) may be absent from the gridded field.
False ice concentrations are reported due to weather and land contamination. This field cannot be used directly as input to a numerical model, but it can be, and is, used as a starting point for further analysis.
The data source for the algorithm is the SSMI instrument on the DMSP F-13 satellite.
The algorithm used is the NASA Team Algorithm [Cavalieri, 1992], with the improved weather filter [Gloersen and Cavalieri, 1986], using brightness temperatures referenced to the F-8 satellite [Abdalati, et al., 1995] from F-11. The ice concentration field shows little effect from F-13 to F-11 differences, but the weather filter and brightness temperatures do need recalibration for the new instrument. The grid used is a 25.4 km polar stereographic grid, true at 60, with 80 W at the bottom of the image in the Northern Hemisphere and at the top in the Southern Hemisphere.
Beyond the above, the polar gap (latitudes poleward of 87.5, which are unobserved by the SSMI) is filled in by a laplacean if there are data available surrounding the unobserved gap. The weather filter is applied to adjacent pairs of points, rather than only to the isolated points. A land mask derived from the NESDIS 1/16th degree mask is applied for graphic output, but not internally to the basic data grids. The masking grid is currently generous about defining a point as ocean, so that much of the coastal region shows up as being ice covered. This is really due to land signals which mimic sea ice cover. When working with the data, it is best to apply a land mask suitable to your use rather than to rely on the one included here. If no data are available at a point, that point is flagged.
Full documentation of the data processing, timing, and algorithms is available in the document "Automated Passive Microwave Sea Ice Concentration Analysis at NCEP" by Robert W. Grumbine.
Send comments or questions to Robert.Grumbine@noaa.gov
Return to Analysis page
Return to Sea Ice Home page
Return to Marine Modeling and Analysis Branch Home Page