DOWNLOADING DATA VIA HTTPS
To learn more about Earthdata Login and register for an account, please see How to Register with Earthdata Login. Once you have logged in, data can be downloaded via a Web browser, command line, or client. For help with downloading data, please see Options Available for Bulk Downloading Data from HTTPS with Earthdata Login.
NOTE: Reverb will be decommissioned in the coming months and replaced with Earthdata Search. All links to Reverb will be removed at that time.
Near-Real-Time SSM/I-SSMIS EASE-Grid Daily Global Ice Concentration and Snow Extent, Version 5
The Near-real-time Ice and Snow Extent (NISE) data set provides daily, global maps of sea ice concentrations and snow extent. These data are not suitable for time series, anomalies, or trends analyses. They are meant to provide a best estimate of current ice and snow conditions based on information and algorithms available at the time the data are acquired. Near-real-time products are not intended for operational use in assessing sea ice conditions for navigation.
- The NISE snow and sea ice algorithm both use near-real-time brightness temperature observations from the SSMIS instrument on DMSP-F18.
- The coefficients for the NISE snow algorithm were updated to better match results from DMSP-F13 as a result of an inter-calibration between F13 and F17 during an overlap period from 3/27/2008-3/26/2009 and between F17 and F18 during an overlap period from 7/1/2014-6/30/2015.
- ESDT metadata was updated to reflect the change in data set version number.
The NISE Version 5 data record begins 12/01/2016. A two-month overlap with NISE Version 4 is planned: 12/01/2016-01/31/2017
|Spatial Resolution:||Not Specified|
|Temporal Resolution:||1 day|
|Platform(s)||DMSP 5D-2/F13, DMSP 5D-3/F17, DMSP 5D-3/F18|
|Data Contributor(s):||J Stewart, Mary Brodzik|
|Metadata XML:||View Metadata Record|
As a condition of using these data, you must cite the use of this data set using the following citation. For more information, see our Use and Copyright Web page.Brodzik, M. J. and J. S. Stewart. 2016. Near-Real-Time SSM/I-SSMIS EASE-Grid Daily Global Ice Concentration and Snow Extent, Version 5. [Indicate subset used]. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. doi: http://dx.doi.org/10.5067/3KB2JPLFPK3R. [Date Accessed].
Detailed Data Description
The NISE product is updated daily using the best available data from the past five days. Data are in Hierarchical Data Format - Earth Observing System (HDF-EOS) format and browse files are in Joint Photographic Experts Group (JPEG)
.jpg format. HDF-EOS data files are available from 01 December 2016 through the present. Data in both formats are updated daily.
Daily data are provided in a single HDF-EOS file containing four grid objects: one data grid and one age grid each for both the Northern and Southern hemispheres. The data grids contain snow extent, sea ice concentration, and coastal (mixed) pixels. The age grids contain the age of input data in days (from day of data acquisition to map production) relative to the date of the daily file. Values are stored in the data and age grids as binary arrays of unsigned 1-byte (8-bit) data ranging in value from 0 to 255.
Files are named according to the following conventions and as described in Table 1.
||Near-real-time Ice and Snow Extent|
||Special Sensor Microwave Imager/Sounder: sensor|
||2-digit month of year|
||2-digit day of month|
|h||Hemisphere (1: Northern, 2: Southern)|
HDF-EOS data files: 2.1 MB each
HDF browse files: 156 KB each
Metadata files: 6 KB each
Northern Hemisphere GIF browse files: 49 KB each
Southern Hemisphere GIF browse files: 37 KB each
Spatial coverage is shown in Figures 1 and 2, and is global except for a gap of three degrees latitude from each pole (87 to 90 degrees latitude). The application of monthly-varying masks limits the mapped extent of snow and sea ice in both hemispheres.
Spatial Coverage Maps
The spatial resolution for this data set is 25 km.
Projection and Grid Description
Sea ice concentration and snow extent maps are provided in two azimuthal, equal-area projections: the Southern Hemisphere 25 km low resolution (SL, indicating Southern Low) and Northern Hemisphere 25 km low resolution (NL, SL, indicating Northern Low) Equal-Area Scalable Earth-Grids (EASE-Grids). See All About EASE-Grid for more information about the equal-area projections used for this product. Grids are 721 columns by 721 rows. The respective pole is aligned with the center of the pixel at the center of the grid.
For each 24-hour period, NISE is updated using the most recent data for a given grid cell. The frequency of updates varies as a function of latitude. Grid cells representing latitudes above 55 degrees or below -55 degrees, for which multiple satellite passes are available each day, are usually updated every 24 hours. Due to the orbital geometry of the DMSP satellite and the swath width of the SSM/I sensor, the time interval between successive observations at low-latitude locations (-20 to 20 degrees) can be up to five days (Hollinger et al. 1987). Problems are not anticipated with this low-update frequency given the absence of sea ice and very limited snow extent at these low latitudes. During occasional periods when input data are unavailable or unobtainable, the NISE product age values at any location may be older than five days. An age grid indicates the number of days since each grid cell was last updated. Refer to the Data Acquisition Methods section for more information.
Each HDF-EOS file represents daily data comprising the last available snow extent or sea ice concentration data for each pixel.
Table 2 lists data values and the following parameters used in the data grids:
- Snow extent: Presence or absence of snow
- Sea ice concentration: Sea ice concentration (percent)
- Coastal pixel: 100-km-wide area of grid cells comprising the mixed pixel regions along a coastline
|1-100||Sea ice concentration (%)|
|101||Permanent ice (Greenland, Antarctica)|
|252||Mixed pixels at coastlines
(unable to reliably apply microwave algorithms)
|253||Suspect ice value|
Table 3 lists data values and the following parameter used in the age grids:
- Age: Age of input data used to derive a corresponding data grid value, in days relative to the date of the daily file
|0-254||Age in days since date of file|
|255||Filler value for corners (off-Earth) and undetermined data pixels|
Cell values range from 0 to 255 for snow extent, sea ice concentration, and age grids.
Figures 3 and 4 show snow extent, sea ice concentration, and age layers for 03 February 2002 for the Northern and Southern Hemispheres, respectively.
Software and Tools
The EASE-Grid Geolocation Tools Web page provides files containing arrays of latitude and longitude values for each grid cell. Fortran and C source code are available for converting grid cell locations to latitude and longitude values, and vice-versa. An Interactive Data Language (IDL) program is available for converting latitude and longitude values to grid column and row coordinates.
NSIDC's Hierarchical Data Format - Earth Observing System (HDF-EOS) Web site provides information about the HDF-EOS format, tools that extract HDF-EOS objects into ASCII or flat binary formats, and links to other HDF-EOS resources. In addition, example code for access and visualization of NISE data in NCL, Matlab, and IDL is provided on the HDF-EOS Comprehensive Examples Web page.
Quality control for this product is performed by the National Environmental Satellite, Data, and Information System (NESDIS) when converting Temperature Data Records to brightness temperatures. NSIDC visually inspects daily NISE browse files.
Data Acquisition and Processing
The Fleet Numerical Meteorology and Oceanography Center (FNMOC) first receives SSMIS raw antenna temperatures. Data are then sent to NESDIS where they are processed into swath brightness temperatures (TBs). NSIDC obtains these swath TBs via FTP once per day from the NESDIS Comprehensive Large Array-data Stewardship System (CLASS), typically within two to four days of the satellite overpass.
The NISE product provides a best estimate of current ice and snow conditions based on information and algorithms available to NSIDC at the time the product is created. If new input data from CLASS are unavailable or unobtainable, then NSIDC has no new information with which to update the NISE product. The current day's data layers will be identical to those of the previous day, and the age layers will be incremented by one day.
Problems obtaining the input data are usually resolved within one business day; however, if new input data are unavailable or unobtainable for five days in a row, NISE data production is halted until the problem is resolved. The NISE product is then reprocessed from the beginning of the interruption, and the new data are released.
All algorithms are subject to change in order to provide the best possible snow and ice mapping capabilities. Snow extent derived from passive microwave satellite data is a product that is constantly undergoing revision and improvement at NSIDC.
Sea ice concentration percentage for the NISE Version 4 product is derived using the NASA Team total sea ice (first year ice plus multiyear ice) algorithm (Cavalieri et al. 1992). For sea ice, all SSMIS (19, 22, and 37) GHz data from a given 24-hour period are binned to the EASE-Grid using a "drop-in-the-bucket" interpolation method. For snow extent mapping, TB values from the satellite's morning pass only are used as input to the nearest neighbor interpolation. Regions for which the respective interpolation algorithm will be used are defined using a land/ocean/ice cap mask.
Snow extent is mapped separately using an algorithm developed for Scanning Multichannel Microwave Radiometer (SMMR) data as described in Chang, Foster, and Hall (1987), and modified for use with SSM/I data as described in Armstrong and Brodzik (2001). NSIDC modified the snow extent mapping algorithm in March 2002, based primarily on a recent study by Armstrong and Brodzik (2002). One goal of this study was to determine when the differences between microwave algorithm output and the validation data are random and when they are systematic; therefore, this study made use of larger and more comprehensive validation data sets that could provide a full range of snow/climate conditions, rather than limited data that might only represent a snapshot in time and space.
Armstrong and Brodzik evaluated snow extent derived from passive microwave data through comparison with over ten years of the NOAA Northern Hemisphere snow charts, which are based on visible-band satellite data. Results clearly indicated time periods and geographic regions where the two techniques agreed and where they tended to consistently disagree. While not always an exact representation of the actual snow extent, the NOAA snow maps are highly accurate and are the product of a well-understood analysis procedure, and as such they were considered truth in these comparisons. The algorithms compared represented examples that included both mid- and high-frequency channels, vertical and horizontal polarizations, and polarization difference approaches, thus allowing an evaluation of the relative merits of these different approaches at the hemispheric scale.
This study demonstrated that both the new NSIDC NISE algorithm and the original Goodison (1989) algorithm underestimated snow extent in the presence of shallow snow; however, during winter and spring the Goodison algorithm tended to consistently overestimate the snow extent (both wet and dry snow) in various locations -- in particular, over such regions as the high-elevation deserts of Central Asia. Armstrong and Brodzik concluded that this was most likely due to an enhanced spectral gradient of the vertical polarization channel in the presence of frozen desert soils. Similarly, the 37 GHz polarization difference that drives the wet snow algorithm often responded to the soil types in this region such that it caused a false snow signal.
In summary, this study indicated that horizontal-polarization-based algorithms, while apparently underestimating snow extent during early winter, appear to provide the best overall estimates of snow extent at the continental to hemispheric scale through the period of maximum snow extent and into the melt season. Vertical-polarization-based algorithms (Goodison 1989) provide similar results but with a consistent tendency to falsely identify snow-free desert soils and/or frozen ground as snow-covered.
Climatologies mask out spurious data caused primarily by weather effects on sea ice and snow extent passive microwave data. Currently the sea ice climatology is a monthly ocean mask derived from historical SMMR data (1979-1987) and SSM/I data (1987-2003).
The snow extent climatology for the Northern Hemisphere is a monthly mask derived from the Northern Hemisphere EASE-Grid Weekly Snow Cover and Sea Ice Extent from October 1966 to May 2005. Beginning 01 July 2005, the snow extent climatology for the Southern Hemisphere is a monthly mask derived from the SSM/I period data from 1987 to 2003, identifying by month those parts of South America and New Zealand that may be snow-covered during that month. The original Southern Hemisphere snow climatology was a static file (not monthly) that served as a spatial representation of the expected snow line in the Andes as a function of latitude and elevation to determine snow extent (Schwerdtfeger 1976). NISE files dated 30 June 2005 and earlier use the original mask.
Northern Hemisphere Monthly Snow Extent Climatologies
Southern Hemisphere Monthly Snow Extent Climatologies
Northern Hemisphere Monthly Sea Ice Climatologies
|January - June||July - December|
Southern Hemisphere Monthly Sea Ice Climatologies
|January - June||July - December|
Physical conditions affecting the accuracy of the sea ice concentration algorithm include atmospheric water content, ocean roughening and spray, presence of thin ice, and formation of melt ponds on the sea ice. Errors become greatest during mid- to late summer, resulting primarily from melt ponds on the ice surface, and also from atmospheric- and weather-related effects over open ocean. To minimize the error over open ocean, a filter is applied to detect these atmospheric effects. False ice concentration estimates may also occur along coastlines due to mixed pixels. Mixed pixels contain signals from both land and water in unknown proportions. For the NISE product, such errors are minimized by designating these pixels as coastal pixels.
The snow extent mapping algorithm maps a grid cell as snow-covered when it has a computed snow depth greater than 2.5 cm.
The presence of dense coniferous and deciduous forests presents problems for mapping snow extent with SSMIS because the vegetation canopy obscures snow on the ground. The best conditions for accurate snow extent mapping using SSMIS are in areas of little or no vegetation such as in the prairies and tundra.
Known problems typical of SSM/I data include occasional missing data, mislocated scan lines, and out-of-bounds data values. Such errors usually result in no NISE observations at affected locations on those days, so the most recent observation at that location is used. The age field indicates when the pixel's observation was made.
Errors may also be introduced when the brightness temperature data that the algorithm uses become unreliable. For instance, from April 5-May 25, 2016, the 37V channel of the SSMIS onboard DMSP-F17 was affected by changes in a solar panel position on the satellite. NISE data from this time period should be used with caution. Largely as a result of this incident, NISE began using data from DMSP-F18 on Nov 1, 2016.
NSIDC creates the current NISE product using passive microwave data from the Special Sensor Microwave Imager/Sounder (SSMIS) on board the Defense Meteorological Satellite Program (DMSP) F18 satellite. The SSMIS instrument is the next generation Special Sensor Microwave/Imager (SSM/I) instrument.
The SSMIS sensor is a conically-scanning passive microwave radiometer that harnesses the imaging and sounding capabilities of three previous DMSP microwave sensors, including the SSM/I, the SSM/T-1 temperature sounder, and the SSMI/T-2 moisture sounder. The SSMIS sensor measures microwave energy at 24 frequencies from 19 to 183 GHz with a swath width of 1700 km. Refer to the SSMIS Instrument Description Web page for more details.
Table 4 outlines the processing and algorithm history for this product.
This revision is an update to NISE Version 4. All Version 4 data (17 August 2009 - present) have been reprocessed with this system.
This revision is designated NISE Version 4. All data from 17 August 2009 to the present have been processed with this system. The NISE Version 2 product from F13 has been produced through 31 August 2009. (No NISE Version 3 product is available).
|V3||N/A||NA; no NISE Version 3 product is available.|
|V2.3||2008-10-06||Ported NISE processing system from SGI to linux. No significant changes in output.|
|V2.2||2006-04-27||New Northern Hemisphere snow climatologies with data from 1966-2005, and new Northern and Southern Hemisphere ice climatologies with data from 1979-2003.|
|V2.1||2005-07-01||Static Southern Hemisphere snow climatology limiting possible snow to the Andes region was replaced with a monthly climatology that now includes the Andes and New Zealand.|
|V2||2005-06-10||Data from the start of the SSM/I F13 mission (04 May 1995) to 31 December 1999 were processed to NISE Version 2.|
This revision is designated NISE Version 2. All data from 01 January 2000 to present have been reprocessed with this system. NISE Version 1 files will be deleted at a future date.
|V1.9||2002-03-20||New snow extent algorithm|
|V1.8||2000-06-01||New ice climatologies through 1999|
|V1.7||2000-03-02||Changed metadata field for PRODUCTIONDATETIME from local time to UTC|
|V1.5||1999-10-07||First modifications for managing data in the EOSDIS Core System (ECS)|
|V1.4||1999-04-09||Fixed geolocation errors in NISE-to-HDFEOS program|
|V1.1||1997-12-31||Fixed 2-digit year problem in time for new year|
|V1||1997-10-31||Initial operational release|
This data set was originally designed to provide NASA EOS researchers with near-real-time daily, global snow extent and sea ice concentration data. The following NASA EOS instrument teams use the NISE data to generate their products:
Multi-angle Imaging SpectroRadiometer (MISR). The MISR instrument is part of a suite of sensors on NASA's EOS Terra satellite. The NISE product will be used as ancillary data for the MISR Top-of-Atmosphere/Cloud product, which requires near-real-time daily, global snow and sea ice extent data.
Clouds and the Earth's Radiant Energy System (CERES). CERES requires both daily and monthly averaged global snow and ice extent maps for several of their Earth radiation budget products. CERES is currently part of the Tropical Rainfall Measuring Mission (TRMM) aboard the EOS Terra platform.
Moderate Resolution Imaging Spectroradiometer (MODIS). The MODIS Atmosphere Discipline Group uses NISE data to produce their cloud mask.
NSIDC anticipates additional use of the NISE product to:
Produce sea surface temperature maps from the MODIS instrument.
Provide ancillary data to the Global Land Ice Monitoring from Space project, which uses the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER).
Compare with Advanced Microwave Scanning Radiometer (AMSR) sea ice concentrations and snow extent maps.
The accuracy of the sea ice concentration estimates is within approximately five percent in most areas during the majority of the year. For the sea ice component of the current NISE Version 5 product, NSIDC has done preliminary inter-calibration between F17 and F18 and found that the sea ice algorithm coefficients yield similar ice extents during an overlap period of 01 March 2015 through 29 February 2016. F18 sea ice estimates should be reasonably consistent with F17 estimates, although differences of up to approximately 28,000 sq km may be possible in daily total extents. The differences are primarily near the ice edge, where shifts of one to two grid cells (25-50 km) may be seen.
Version 5 of NISE incorporates a spillover correction (Cavalieri et al. 1999) that reduces or eliminates sea ice concentrations near coastlines when there is open water present. Due to the relatively large microwave footprint size, passive microwave emissions from adjacent land masses contaminate the signal for coastal pixels, producing spurious sea ice extents in coastal regions that are actually void of sea ice, especially during summer months
Armstrong and Brodzik (2001) demonstrate that the snow extent algorithm can provide daily global snow extent maps to an accuracy of approximately 50 km, except in areas of wet snow or dense forest cover. In the snow extent product, when the snow is wet -- when liquid water is present on the snow grain surface -- the snow pack becomes predominantly an emitter and much of the scattered portion of the ground signal is lost, greatly limiting algorithm accuracy. To reduce the frequency of observations over wet snow, only data from the early morning (descending) orbits are used as input to the algorithm.
References and Related Publications
Contacts and Acknowledgments
Mary Jo Brodzik
National Snow and Ice Data Center
CIRES, 449 UCB
University of Colorado
Boulder, Colorado USA 80309-0449
J. Scott Stewart
National Snow and Ice Data Center
CIRES, 449 UCB
University of Colorado
Boulder, Colorado USA 80309-0449
NASA provided funding for the production of this data set through the NSIDC Distributed Active Archive Center (DAAC).
DOCUMENT CREATION DATE
December 02, 2016