Introduction

Soil Temperature

Alaska

Russia

Snow Cover

Snow Off Day

Sea Ice Index

Browse & Compare Images

Archived Data & Images

Image Derivation

Interpretation Resources

Sea Ice Index: Image Derivation

Introduction
Data Sources
Processing Steps
Monthly Mean Ice Concentration
Ice Extent Values
Average Ice Edge and Ice Extent Anomalies
Trends in Ice Extent
Anomalies in Concentration
Trends in Concentration
Spatial Coverage: The "Pole Hole"
References
Citing the Sea Ice Index

Introduction

Records of sea ice extent and concentration from satellite passive microwave data are available for October 1978 through June 1987 from the Nimbus-7 Scanning Multichannel Microwave Radiometer (SMMR) and since July 1987 from its successor, the Special Sensor Microwave/Imager (SSM/I). Sea ice concentration can be estimated from brightness temperature data because sea ice and water have differing passive microwave signatures.

Data Sources

The Sea Ice Index is produced primarily from Sea Ice Concentrations from Nimbus-7 SSMR and DMSP SSM/I Passive Microwave Data, produced at the Goddard Space Flight Center (GSFC). This product contains daily and monthly grids of sea ice concentrations derived to provide a consistent long-term time series beginning with the SMMR instrument in October 1978. These data are produced and are periodically updated by GSFC investigators using brightness temperatures from NSIDC. Sea Ice Index fields are produced from the final versions of the GSFC product.

The most recent data used in the Sea Ice Index are processed from NSIDC’s Near Real-Time SSM/I Polar Gridded Sea Ice Concentrations (NRTSI). The NRTSI fields are also processed in a similar manner as the GSFC product, but the input brightness temperature data have a lower quality-control level and there may be missing data.

Each Sea Ice Index monthly update uses the highest processing level available for that given month. Thus, fields produced from NRTSI data will later be replaced with values from GSFC data. These changes in data source can result in slight changes in the total extent, concentration, and anomaly values; the changes are generally less than 10,000 km² (15-20 grid points).

To create the images for the Sea Ice Index, daily data are used to obtain each month’s average sea ice concentrations. The averages are then appended to a time series of that month’s means before estimating trends, and differenced with the long term (1979 - 2000) average for that month to estimate anomalies.

Consistency between NRTSI and GSFC-derived data was checked by processing NRTSI data for 2002 using both brightness temperature data from NSIDC and near real-time brightness temperature data from NASA Marshall Space Flight Center (MSFC) as input, and then differencing monthly Sea Ice Index grids with monthly grids derived from GSFC daily grids for 2002. We found that the brightness temperature source made little difference in overall extent and area. The differences in the NRTSI and GSFC-derived monthly area and extent values for the year 2002 are at most 1.6% (the area difference in June with RSS brightness temperatures) with most differences much lower or negligible. Yearly average differences in ice extent of 0.38% and area of 0.42% are on the same order as reported by the GSFC series. See Section 2 of the Sea Ice Concentrations from Nimbus-7 SSMR and DMSP SSM/I Passive Microwave Data documentation. Note that while these small differences show that overall summaries of ice extent within the Sea Ice Index data set are consistent over time, significant regional differences in ice concentration may be present across changes in instrumentation within the GSFC data set, and across the GSFC/NRTSI change in the Sea Ice Index data set.

Processing Steps

Monthly Mean Ice Concentration

GSFC daily grids are averaged at NSIDC into monthly grids, to create a monthly mean concentration series and to obtain standard deviations. (Note that there may be differences in these monthly mean concentration grids derived from daily data and the monthly mean concentration fields available as part of the GSFC data set. These result from differences in processing discussed in Data Sources.) For NRTSI, monthly averages are computed from gridded daily products. The nominal pixel size of the gridded products is 25 km. Occasionally data from one or more days within the month is missing. Averages are computed only if there are at least 20 days of data. SMMR only operated every other day, resulting in half as much data as for SSM/I, thus a period of 10 days is required each month for SMMR.

In constructing monthly mean ice concentrations, we are conflating a temporal and spatial average. That is, a mean concentration of 50% for a pixel in September could be the result of 100% concentration for 15 days and 0% concentration for 15 days, or of 50% concentration for 30 days, or some other combination. The meaning of "average" concentration can therefore be ambiguous, especially near the ice edge, where wind can move ice rapidly into and out of the area covered by a pixel.

The "Total area" at the bottom of the ice concentration image is the sum of the area covered by ice, not including the open water area within the margins of the pack or the area near the pole that is not imaged by the sensor (the “pole hole”). The total area is always less than total extent.

Ice Extent and Ice Area Values

The values for ice extent are obtained by summing the area covered by all pixels that have 15% or greater ice concentration. We assume that the area not imaged by the sensor at the North Pole is entirely ice-covered. Each pixel’s area is calculated individually, and is obtained by multiplying the nominal pixel size (625 km²) by the square of the map scale at the center of the pixel. Pixel areas range from 382 to 664 square kilometers for the Northern Hemisphere and 443 to 664 square kilometers for the Southern Hemisphere, under the polar stereographic projection and grid used for the input data sets.

The extent values are useful in a temporal series but caution should be used citing the numbers apart from the time series or comparing with values derived from other studies. Ice concentrations are sensitive to the algorithm used, and resulting numbers for extent depend not only on algorithms but on other processing steps as well. The extent values have uncertain significance when taken individually. For example, the 15% concentration cutoff for extent is somewhat arbitrary. Using a 20% or 30% cutoff will give different numbers, although similar trends, for extent (for examples, see Parkinson et al. 1999).

The values for ice area are obtained by summing the concentration of ice within each pixel over the entire ice extent. For example, if a pixel's area was 600 km² and its ice concentration was 75%, then the ice area, for that pixel, would be 450 km².

Note that unlike ice extent, the Arctic values for ice area do not include the area near the pole not imaged by the sensor (the “pole hole”). This area is 1.19 million square kilometers for SMMR (from the beginning of the series through June 1987) and 0.31 million square kilometers for SSM/I (from July 1987 to present). Therefore, there is a discontinuity in the "area" data values in this file at the June/July 1987 boundary.

"Average" Ice Edge and Ice Extent Anomalies

On monthly mean images, the ice edge is where concentration drops below 15%, although a mean ice edge is a somewhat ambiguous concept. How does this edge position compare with what is typical for the month, based on the January 1979 - December 2000 portion of the data set? To help answer this question, we computed a median edge for the month by showing those pixels for which there is a 50% probability of ice occurring at 15% concentration or greater. Alternatively, we could have computed an average edge by averaging the mean concentrations, and using the 15% cutoff in the average concentration image as the "mean" edge. However, this method results in a mean edge that is unlikely to resemble any typical ice edge, because the location of the edge varies considerably from year to year.

Images of anomalies in ice extent show the difference between the location of the median ice edge for the month, as described above, and the ice edge location for a particular month. The total extent of sea ice for that month is also shown.

Trends in Ice Extent

Ice extent anomalies are plotted as a time series of percent differences between the total extent for the month in question, and the mean for that month, where the mean is based on the January 1979 - December 2000 portion of the data set. The trend, in percent change per decade, is obtained using least squares regression, and a 95% confidence interval for the resulting slope is given.

Anomalies in Concentration

To produce concentration anomaly images, monthly concentration images are subtracted from an image of the mean for the month in question from the 1979 - 2000 portion of the data set. The color bar shows, in percent, how much the ice concentration for the month differs from the mean calculated for that month over the 1979 - 2000 portion of the data set.

Note that an area may have a positive anomaly for a given month while at the same time showing a negative trend in concentration over time.

Trends in Concentration

A least squares regression is performed on the time series of concentrations at each pixel. The time period covered by the trend analysis is from November 1978 to the present month. Pixels that have zero concentration are left out of the time series when calculating trends. For the most part, these pixels are near the ice edge. The slope of the linear fit gives the trend in concentration for that pixel. A 95% confidence interval is used for significance (the null hypothesis that the slope of the fit line is zero is rejected with 95% confidence). Trends are shown in percent change in concentration per decade. If the null hypothesis of no slope (no trend in concentration) cannot be rejected, the pixel is shown in white.

Spatial Coverage: The "Pole Hole"

The SMMR and SSM/I instruments do not image a circular sector over the poles, due to orbit inclination. In the GSFC data set, the area of this "pole hole" is 0.31 x 106 km² for SSM/I and 1.19 x 106 km² for SMMR. In calculating northern hemisphere ice extent, we assume that the entire area is ice covered. The northern hemisphere ice concentration anomaly and trend images have no data for the area covered by the larger SMMR hole, because the time series upon which these derived values are based includes the SMMR instrument.

References

Cavalieri, D.J., P. Gloersen and W. J. Campbell. 1984. Determination of sea ice parameters with the NIMBUS-7 SMMR. Journal of Geophysical Research 89 (D4): 5355-5369.

Cavalieri, D.J., K.M. St. Germain and C.T. Swift. 1995. Reduction of weather effects in the calculation of sea ice concentration with the DMSP SSM/I. Journal of Glaciology 41: 455-464.

Cavalieri, D., P. Gloerson, and J. Zwally. 1990. DMSP SSM/I daily polar gridded sea ice concentrations. Edited by J. Maslanik and J. Stroeve. Boulder, CO: National Snow and Ice Data Center. Digital media.

Cavalieri, D., C. Parkinson, P. Gloerson, and H.J. Zwally. 1997, updated 2005. Sea ice concentrations from Nimbus-7 SMMR and DMSP SSM/I passive microwave data, June to September 2001. Boulder, CO: National Snow and Ice Data Center. Digital media and CD-ROM.

Cavalieri, D., P. Gloerson, and J. Zwally. 2002, updated daily. Near Real-Time DMSP SSM/I Daily Polar Gridded Sea Ice Concentrations. Edited by J. Maslanik and J. Stroeve. Boulder, CO: National Snow and Ice Data Center. Digital media.

Eppler, D.T., L.D. Farmer, A.W. Lohanick, M.A. Anderson, D. Cavalieri, J. Comiso, P. Gloersen, C. Garrity, T.C. Grenfell, M. Hallikainen, J.A. Maslanik, C. Mätzler, R.A. Melloh, I. Rubinstein, and C.T. Swift. 1992. Passive microwave signatures of sea ice. In Microwave remote sensing of sea ice. Edited by F.D. Carsey. Washington, D.C.: American Geophysical Union. 47-71.

Hollinger, J.P., J.L. Peirce, and G.A. Poe. 1990. SSM/I Instrument Evaluation. IEEE Transactions on Geoscience and Remote Sensing 28 (5): 781-790.

Parkinson, C.L., D.J. Cavalieri, P. Gloersen, H.J. Zwally, and J.C. Comiso. 1999. Arctic sea ice extents, areas, and trends, 1978-1996. Journal of Geophysical Research 104 (C9): 20,837-20,856.

Steffen, K., J. Key, D. Cavalieri, J. Comiso, P. Gloersen, K.S. Germain, and I. Rubinstein. 1992. The estimation of geophysical parameters using passive microwave algorithms in microwave remote sensing of sea ice. Edited by F.D. Carsey. Washington, D.C.: American Geophysical Union. 201-231.

Citing the Sea Ice Index

Please use this citation when referencing the Sea Ice Index:

Fetterer, F., K. Knowles, W. Meier, and M. Savoie. 2002, updated 2007. Sea ice index. Boulder, CO: National Snow and Ice Data Center. Digital media.