Get Data


To learn more about Earthdata Login and to 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.


Data Set ID:

TOVS Pathfinder Path-P Daily and Monthly Polar Gridded Atmospheric Parameters, Version 1

The TIROS-N Operational Vertical Sounder (TOVS) Polar Pathfinder (Path-P) data set consists of gridded daily and monthly Arctic and Antarctic atmospheric data derived from various NOAA satellites. TOVS Path-P daily files are available for the Northern and Southern Hemispheres from 60 degrees poleward at 100 km gridded spatial resolution in an Equal-Area Scalable Earth Grid (EASE-Grid). TOVS Path-P monthly files are available for the Northern Hemisphere only from 60 degrees poleward at 100 km gridded spatial resolution in EASE-Grid. Daily and monthly Northern Hemisphere data are available from January 1979 through current processing, and daily Southern Hemisphere data are available from July 1979 through current processing. Data were created from a modified version of the Improved Initialization Inversion Algorithm (3I) (Chedin et al. 1985), a physical-statistical retrieval method improved for use in identifying geophysical variables in snow- and ice-covered areas (Francis 1994).

Geographic Coverage

  • Atmospheric Temperature > Air Temperature
  • Clouds > Cloud Top Pressure
  • Clouds > Cloud Top Temperature
  • Clouds > Cloud Amount/Frequency > Effective Cloud Fraction
  • Atmospheric Radiation > Emissivity
  • Atmospheric Winds > Wind Stress > Geostrophic Drag Coefficient
Spatial Coverage:
  • N: 90, S: 60, E: 180, W: -180

  • N: -60, S: -90, E: 180, W: -180

Spatial Resolution:
  • 100 km x 100 km
Temporal Coverage:
  • 1 January 1979 to 31 December 2005
  • 12 July 1979 to 31 December 2001
(updated 2008)
Temporal Resolution: 1 day, 1 month
Data Format(s):
  • HDF
Platform(s) NOAA-10, NOAA-11, NOAA-12, NOAA-14, NOAA-6, NOAA-7, NOAA-8, NOAA-9
Sensor(s): MSU, TOVS
Version: V1
Data Contributor(s): Jennifer Francis, Axel Schweiger
Data Citation

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.

Francis, J. A. and A. J. Schweiger. 1999, updated 2008. TOVS Pathfinder Path-P Daily and Monthly Polar Gridded Atmospheric Parameters, Version 1. [Indicate subset used]. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. doi: [Date Accessed].

Back to Top

Collapse All / Open All

Detailed Data Description

The NOAA/NASA Pathfinder Program was initiated to provide scientists with global-scale remote sensing data ahead of NASA's Earth Observing System satellite launches. The Pathfinder concept involved careful reprocessing of existing data sets which could be made readily available as high quality products for global change research. Because the polar regions hold special significance in the climate system, the Polar Pathfinders have established a cooperation to maximize the scientific potential of their data sets.

The TIROS TOVS has flown on NOAA polar-orbiting satellites since 1978 and has generated one of the longest and most complete satellite data records in existence. Radiances from the global TOVS data set have been subsetted and processed for the Arctic region, and the retrieved products are presented on a regular grid in a user-friendly format.

The TOVS Path-P provides users with gridded daily Arctic atmospheric soundings. These data were obtained to identify geophysical parameters in snow- and ice-covered areas (Francis 1994). The data set has been designed to address the particular needs of the polar research community including quantities used to compute surface turbulent fluxes and to drive ocean models.

The modified 3I algorithm (Chedin et al. 1985) is applied to the TOVS high-resolution infrared radiation sounder (HIRS) and the microwave sounding unit (MSU) level-1b radiances to generate daily gridded Arctic atmospheric variables. (Radiances were obtained from the National Environmental Satellite Data Service Division and the National Center for Atmospheric Research.) The grids have a spatial resolution of 100 km, and use observations from areas poleward of 60° N and 60° S. The 3I method combines statistical and physical techniques to estimate these geophysical quantities, which are then averaged over a 24-hour period centered at 12:00 Coordinated Universal Time (UTC) to produce one Arctic-wide field per day. Estimates are then organized onto a rectangular grid using a "drop-in-the-bucket" approach.

The data are gridded to the EASE-Grid (Armstrong and Brodzik 1995) equal-area azimuthal projection centered on the North and South Poles. This facilitates combining TOVS data set with other data sets, and opens the door to studies using data from a variety of sources (such as SSM/I and AVHRR data).

The algorithm used to generate these atmospheric variables has been validated directly through comparisons with surface observations from the Coordinated Eastern Arctic Experiment (CEAREX), North Polar drifting station data, and with radiosonde data from Russian ice stations. Comparisons with other TOVS retrieval algorithms provided further validation.


Data are stored in Hierarchical Data Format (HDF), developed by the National Center for Supercomputing Applications (NCSA), and they follow standards and recommendations outlined in the EOSDIS Version 0 Data Product Implementation Guidelines. TOVS Path-P data can be read by analysis tools capable of reading HDF data or using supplied sample programs.

Each daily and monthly file contains a group of 16 HDF Scientific Data Sets. Metadata are implemented as global attributes containing information about the data set dates and times. Much of this information may not be relevant to the user, but it is included for production and archiving purposes. The metadata consist of 25 global attributes.

The grid dimensions of the sixteen Scientific Data Sets for the Northern Hemisphere are 67 rows by 67 columns and for the Southern Hemisphere are 89 rows by 89 columns. Only two Scientific Data Sets are structured as three-dimensional arrays. The three-dimensional field TEMP has 10 levels in the vertical dimension, while WVAPOR has 5 layers. All data sets are represented as 32-bit floating point values. For more information, see the Appendix A: Data Dictionary and Appendix B: Geometry Class Web pages.

Background color on
File and Directory Structure

The TOVS data are available from the FTP site in two directories divided up by region: north and south. Within the north directory, the data are further subdivided into daily and monthly directories. Within the south directory, the data are further subdivided into a daily directory. These directories are in turn subdivided into year directories of the form yyyy. A tools directory is also provided which includes files for viewing and accessing the data. See the Software and Tools section of this document for more information about these tools. Figure 1 displays the directory structure.

FTP directory structure
Figure 1. FTP Directory Structure
Background color on
File Naming Convention

The data files are named according to the following convention and as described in Table 1.

Daily: tpp_Nss_h100_yyyyddd_daily.vx-y.hdf
Monthly: tpp_Nss_h100_yyyymm_monthly.vx-y.hdf


Table 1. File Naming Convention Description
Variable Description
tpp Indicates this is TOVS Path-P data
N Indicates NOAA satellite
ss 2-digit NOAA satellite identifier (06, 07, 09, 10, 11, 12, or 14)
h Hemisphere (n: Northern Hemisphere or s: Southern Hemisphere)
Note: There are no monthly files for the Southern Hemisphere
100 Indicates 100 km gridded data
yyyy 4-digit year
ddd 3-digit day of year (daily files only)
mm 2-digit month (monthly files only)
daily Indicates this is a daily data file
monthly Indicates this is a monthly data file
vx-y two-part version number (for example, v3-4)
.hdf File extension indicates an HDF file

For example, tpp_N06_s100_1987294_daily.v3-3.hdf or tpp_N14_n100_200412_monthly.v3-4.hdf.

Note: The transition of data collection from NOAA 10 to NOAA 11 occurred during the month of September 1991. The file name for the September 1991 monthly data file contains both satellite identifiers as N10N11 indicating that data in this file came from both satellites.

Background color on
File Size

Northern Hemisphere daily files: 506,437 KB uncompressed
Northern Hemisphere monthly files: 506,443 KB uncompressed
Southern Hemisphere daily files: 778,864 KB uncompressed

Background color on
Spatial Coverage

Northern Hemisphere files: 60° N to 90° N
Southern Hemisphere files: 60° S to 90° S

Spatial Resolution

Nominal grid spacing of the TOVS Path-P products is 100.2701 km. The TOVS EASE-Grid spacing is not exactly 100 km because the grid, originally designed for the 25 km Special Sensing Microwave/Imager (SSM/I), required a slightly larger actual cell size (C=25.067525 km) to exactly span the equator. For sake of data product consistency, this cell size was used for all other EASE-Grid products. Of course, few cells actually have these dimensions, but they all have the same area.


The north and south EASE-Grids are based on an equal area azimuthal projection, centered on the North and South Poles, respectively. Grid spacing of the TOVS products is 100.2701 km, resulting in a 67-row by 67-column grid size for the Northern Hemisphere and an 89-row by 89-column grid size for the Southern Hemisphere.

The grid coordinates, r and s, are defined with axes parallel to the rows (s) and columns (r) of the grid and units equal to the sampling interval. The grid sample locations (grid-cell centers) are then the integer coordinate points. The coordinate system starts at the top left corner with r increasing to the right, and s increasing downward. A grid cell (j,i) in the r,s coordinate system is defined as the area between grid coordinates i-0.5 and i+0.5, and j-0.5 and j+0.5. The lower bound is included in the grid cell, while the upper bound is not; i and j are zero-based array indices for this grid cell. This definition means that the grid cells are referenced in r and s by their grid cell center coordinate.

Figure 2 shows the equations to convert between latitude and longitude coordinates and r,s grid coordinates.

Figure 2. Equations to convert between latitude and longitude coordinates and r,s grid coordinates

Note: Latitude and longitude coordinates for each of the Path-P grid cells are included with the data set as well as elevation information and a land mask in ancillary data files: tpp_n100_9999999_ancil.hdf for the Northern Hemisphere and tpp_s100_9999999_ancil.hdf for the Southern Hemisphere. For more information see the Software and Tools section of this document.

Grid Description

The TOVS Path-P EASE-Grids are based on an azimuthal equal-area projection for the Northern and Southern Hemispheres, with a grid resolution of 100 km. Please see All about EASE-Grid for more information on the EASE-Grid projection parameters and grid definitions. For a better understanding of the relationship between the various Polar Pathfinder grids, please see the Summary of NOAA/NASA Polar Pathfinder Grid Relationships.

Background color on
Temporal Coverage

This data set begins in July 1979 and continues through current processing. See the Overview Table at the top of this document for specific dates.

For some days, TOVS Level 1b data were either unavailable or of insufficient quality to ingest into the Path-P processing algorithm. When this occurred, files are available for these days but contain no data. As data become available from alternate satellites, these files will be updated. See Appendix C: Missing Data Files for a list of the missing data files.

Temporal Resolution

Daily data are linear averages of retrievals from all orbits for a single day, defined from 00:00:00 UTC to (excluding) 00:00:00 UTC of the following day. Users will also find files that are very small in size. Data were not available from NOAA for certain orbits but to maintain chronological integrity, files have been added as placeholders to complete the time series.

Background color on
Parameter or Variable

Parameter Description

Table 2 summarizes the parameters in this data set.

Table 2. Summary of Data Set Parameters
Variable Description Units
TEMP Temperatures at pressure levels:  
50, 70, 100, 300, 400, 500, 600, 700, 850 and 900 mb
WVAPOR Precipitable water in layers bounded by:  
300-400, 400-500, 500-700, 700-850, 850-900 mb
SKTEMP Surface Skin Temperature kelvins
HIRS_CLDY Fraction of cloudy pixels per III retrieval box percent
FCLD Total effective cloud fraction percent
CLPRESS Cloud-top pressure millibars (hectapascals)
CLTEMP Cloud-top temperature kelvins
EMISS Emissivity at 50 GHz  none
ISICE Surface Type Flag 0 = Open 
1 = ice 
3 = land 
10 = last orbit showed open water, but changed from orbit to orbit 
11 = last orbit showed ice, but changed from orbit to orbit 
13 = last orbit showed land, but changed from orbit to orbit
SOLZEN Average solar zenith angle (daily only) degrees
PRESS NCEP sea level pressure millibars
PBLSTRAT Boundary layer bulk stratification kelvins
Cg Geostrophic drag coefficient over sea ice (daily only) none
ALPHA Turning angle between geostrophic wind and the surface wind over sea ice (daily only) degrees

The weighting functions of the CO2 absorption bands used by the TOVS retrieval algorithms limit the vertical resolution of the water vapor retrievals to 5 layers between the surface and 300 mb.

Surface classifications may differ from orbit to orbit resulting in examples of daily averaged ice warmer than 273.15 K and water colder than 271.3 K.

The solar zenith angle is the integral from sunrise to sunset divided by the time from sunrise to sunset. This integral is not just a function of solar zenith angle but also a function of day length.

the bulk stratification parameter is defined as the difference in potential temperature between the 1000 and 900 mb levels. It provides a measure of the boundary layer stability. This value tends to be more negative during the winter when surface inversions are common in the Arctic, than in warmer months.

PBLSTRAT is used to compute the geostrophic drag coefficient and a surface stress turning angle based on the parameterization developed by Overland and Davidson (1992) and Overland and Colony (1994). These parameters are useful for estimating near-surface winds, surface stress, and turbulent fluxes.

PBLSTRAT describes the difference in potential temperature between the 1000 and 900 mb levels, and provides a bulk measure of the stratification in the atmospheric boundary layer. In grid boxes with a sea-ice surface, this value is used to compute Cg, the geostrophic drag coefficient, and ALPHA, the turning angle between the geostrophic wind and the surface stress, using an empirically derived relationship.

The geostrophic drag coefficient can vary in space and time by a factor of 2 owing to differences in the strength of near-surface temperature inversion. Thus, large spatial variations in Cg may occur and correspond to surface pressure features.

Surface skin temperature is the radiating temperature of the surface, which may differ from the actual surface temperature if the surface emissivity is less than unity. For more information see Francis and Schweiger 1999.

Two measures of cloud fraction are available within a grid box. The first, FCLD or the effective cloud fraction, is the product of the cloud cover in a grid box and cloud emissivity. Although useful for some studies, it proves difficult to validate with surface observations when clouds are not optically thick in the infrared spectral region. Thin clouds are common in the Arctic, thus FCLD and HIRS_CLDY, the fraction of HIRS fields-of-view within a grid box that are identified as cloudy, can be significantly different from each other.

Most of the products listed above are standard atmospheric variables, but a few, surface skin temperature, cloud fraction and the potential temperature between 1000 and 900 mb, were added specifically for polar research and require some explanation.

Additional Note

Designed so that all parameters are filtered by elevation, the modified 3I algorithm (Chedin et al. 1985) causes footprints with surface elevations of 1000 meters or more to be marked as missing for all parameters. (Consequently, most of Greenland's ice appears as missing.) For elevations less than 1000 m, a daily average is computed across all footprints and orbits for that grid cell. The footprint location changes with each scan, for each pass, although all footprint locations fall within the same grid box. As such, elevations appear to vary greatly from day to day. Users may apply the elevation mask provided to ensure the same area is masked every day.

The reason the retrieval is limited to elevations less than 1000 m is that the weighting functions for some of the channels that peak low in the atmosphere become window channels over high elevations (peak of weighting function at the surface), and the algorithm doesn't take this into account, especially over the high ice sheets of Greenland and Antarctica. Thus, to be on the safe side, retrievals above 1000m were excluded.

The sample file below is composed of sensor data from nine satellite orbits. Each orbit is composed of up to 2500 individual sensor scans, along the orbital track. Each scan contains parameter values and location information for the footprint. Throughout the day, a particular grid cell may be passed over a number of times. For example, four passes occurred in the sample below. Choosing a grid cell location of column 17 and row 45, the following are orbital data for this day, per orbit:

               lat  long  column   row   elevation   orbit   scan number 
               72   -53     17      45     597        2       604 
               72   -54     17      45     428        3       498 
               72   -52     17      45     993        4       143 
               72   -52     17      45     858        5       477 

Sample Data Record

Figure 3 shows sample images for the available data parameters and sample data values from 10 April 1996 (tpp_n100_1996100_daily.v3-3.hdf). For the binary subset version there is one file per day per parameter. The binary file names are prefixed with subset_ and end with an extension that reflects the parameter. For example, the file containing cloud fraction data for 10 April 1996 (day 100) is named: subset_tpp_n100_1996100_daily.v3-3.FCLD.

Geostrophic Wind

Geostrophic Drag

Cloudtop Pressure



Cloud Fraction

HIRS Pixel

Surface Type

Boundary Layer

Temperature at
50 Mb

Temperature at
70 Mb

Temperature at
100 Mb

Temperature at
300 Mb

Temperature at
400 Mb

Temperature at
500 Mb

Temperature at
600 Mb

Temperature at
700 Mb

Temperature at
850 Mb

Temperature at
900 Mb

Water Vapor
at 300 Mb

Water Vapor
at 400 Mb

Water Vapor
at 500 Mb

Water Vapor
at 700 Mb

Water Vapor
at 850 Mb
Figure 3. Sample Images for Available Data Parameters and Sample Data Values from 10 April 1996 (tpp_n100_1996100_daily.v3-3.hdf)

Sample Values from HDF or Binary Data Files -
(Row 32, Column 28 for IDL or C programs;
 Row 33, Column 29 for Fortran programs)

TEMP-50        226.683
TEMP-70        225.740
TEMP-100        224.263
TEMP-300        219.452
TEMP-400        230.288
TEMP-500        240.738
TEMP-600        248.602
TEMP-700        254.337
TEMP-850        258.343
TEMP-900        258.722
WVAPOR-300      0.0600000
WVAPOR-400      0.190000
WVAPOR-500      1.30500
WVAPOR-700        1.63000
WVAPOR-850        1.61500
SKTEMP        251.050
HIRS_CLDY        55.3333
FCLD        106.750
CLPRESS        568.750
CLTEMP        247.292
EMISS       0.751667
ISICE        1.00000
SOLZEN        82.3043
PRESS        1024.21
PBLSTRAT       -16.3351
Cg      0.0265637
ALPHA        26.2023
Background color on
Error Sources

Users should be aware of potential inter-satellite calibration problems. Removal of such systematic biases required the development of an automatic correction scheme. The correction procedure used satellite and radiosonde data sets from NOAA/NESDIS as inputs to the forward model to accurately account for and eliminate biases caused by the radiative transfer model, the instrument or unexpected events, such as the Mt. Pinatubo eruption. However, this method may not fully address all of the inter-satellite calibration problems.

Some of the files provided contain no data when data from the chosen satellite could not be supplied. These files remain as placeholders to be filled using data from alternate satellites as it becomes available. See Appendix C: Missing Data Files for a list of the missing data files.

Other Known Problems

  • Elevation and surface type may vary from orbit to orbit, as reported by the satellites. Because surface classifications may differ from orbit to orbit, grid cells classified as ice may have temperatures warmer than freezing, and grid cells classified as water may have temperatures colder than freezing. Examples:
    Ice temperature warmer than freezing (272°K):
      Date          Row       Column      ISICE       SKTEMP
    -------         ---       ------     -------      -------
    1997172          11         37       1.00000      280.900
    1997190          23         49       1.00000      289.900
    1998199          37         52       1.00000      292.500
    Water temperature colder than freezing (272°K):
      Date          Row       Column      ISICE        SKTEMP
    -------         ---       ------     -------      -------
    1997021          37          26      0.00000      248.700
    1997318          27          20      0.00000      252.040
    1998294          54          21      0.00000      257.500
  • The cloud parameter value (FCLD) may be unrealistically high because no upper bound is imposed to ensure that the data are not artificially biased when averaged. Values of 199 have been observed, which may indicate the occurrence of only one retrieval with a value of 199 percent, or that all retrievals had values near 200 percent.
  • The boundary layer stratification (PBLSTRAT) data may have unrealistically large values. Most, but not all, of these values were filtered out.
  • A few instances have been noticed where the cloud top pressure (CLPRESS) is greater than the surface pressure, sometimes by more than 5mb. Example:
    Surface Pressure less than cloud pressure:
     Date           Row       Column     CLDPRS       SURFACE PRESS
    -------         ---       ------     -------      -------
    1997010          47          46      976.750      969.586
    1998087          64          24      984.000      982.692
    1998351          61          23      984.000      974.951
  • The Southern Hemisphere grids reveal some false sea ice, which is likely caused by poor weather/cloud filtering.
Background color on
Quality Assessment

The algorithm used to generate these grids has been validated directly through comparisons with surface observations from the CEAREX experiment and with radiosonde data from Soviet ice stations. Comparisons with other TOVS retrieval algorithms provided further validation. Additionally, Schweiger et al. (1999) found a strong correlation between the TOVS Path-P data and surface cloud observations obtained from the North Polar drifting meteorological stations, which indicates that the TOVS data effectively represent annual cloud cover. Table 3 provides an initial assessment of the accuracy of some variables included in the Path-P products. Not all variables could be verified and validation of the following Path-P variables has been performed only for sea level.

Table 3. Estimated Accuracy of Path-P Level 3 Variables
Parameter Description Units Estimated Accuracy
TEMP Level temperatures K 3K
WVAPOR Layer Precipitable water mm ~30%
SKTEMP Surface Skin Temperature K 3K
FCLD Effective Cloud Fraction % 30%
CLTEMP Cloud-top temperature K TBD
CLPRESS Cloud-top pressure mb TBD
EMISS Surface Emissivity (50 GHz) 5%
PBLSTRAT Boundary Layer Stratification K 5K
Cg Geostrophic drag coefficient ~30%
ALPHA Turning angle Deg ~10 deg
HIRS_CLDY Fraction of cloudy pixels per III retrieval box Percent 20%

Validation of TOVS data is ongoing. For validation updates or more information please visit the Path-P: TOVS Polar Pathfinder Web site at the University of Washington's Applied Physics Laboratory.

Background color on

Software and Tools

Get Data
Background color on
Software and Tools

Tools for browsing the TOVS Path-P data are available via FTP. To accompany the Polar Pathfinder Product, NSIDC created a TOVS Path-P IDL Tools Tutorial Web page that describes how to use the IDL tools to browse TOVS Path-P data files. Additionally, Fortran and C programs for reading Path-P daily data files have been provided by the University of Washington. Static ancillary HDF data files provide elevation, latitude, longitude, and land mask data. They are described in Table 4.

Table 4. Ancillary HDF Data Files Description
File Name Description
tpp_n100_9999999_ancil.hdf Northern Hemisphere Ancillary File: 
  • Center latitude of grid cell is in 100ths of a degree. Divide by 100.0 to get units of degrees. Range of values is 0 to 90.
  • Center longitude of grid cell is in 100ths of a degree. Divide by 100.0 to get units of degrees. Range of values is -180 to +180.
  • Percent of grid cell covered by land in %. Range of values is 0 to 100.
  • Elevation in 10ths of meters. Divide by 10.0 to get units of meters.
tpp_s100_9999999_ancil.hdf Southern Hemisphere Ancillary File: 
  • Center latitude of grid cell is in 100ths of a degree. Divide by 100.0 to get units of degrees. Range of values is 0 to 90.
  • Center longitude of grid cell is in 100ths of a degree. Divide by 100.0 to get units of degrees. Range of values is -180 to +180.
  • Land Mask: ocean=0, land=1.
  • Elevation in 10ths of meters. Divide by 10.0 to get units of meters.


Users preferring to extract the TOVS Path-P data from HDF will find tools at the NCSA Web site.

Background color on

Data Acquisition and Processing

These data were developed at the University of Washington's Applied Physics Laboratory with funding from Polar Exchange at the Sea Surface (POLES), a NASA EOS interdisciplinary project. For validation updates or more information on POLES or the TOVS Polar Pathfinder project, please see the Path-P: TOVS Polar Pathfinder Web site at the University of Washington.

Source or Platform Description

The TOVS scanner was flown aboard several NOAA Polar Operational Environmental Satellites (POES), as shown in Table 5.

Table 5. TOVS Scanners' Platform and Operational Periods
Satellite Time Period
NOAA-6 12 Jul 1979 to 31 Dec 1982
NOAA-7 01 Jan 1983 to 31 Dec 1984
NOAA-9 01 Jan 1985 to 31 Dec 1986
NOAA-10 01 Jan 1987 to 16 Sep 1991
NOAA-11 17 Sep 1991 to 31 Dec 1994
NOAA-12 01 Jan 1995 to 31 Dec 1996
NOAA-14 01 Jan 1997 to most recent data
Background color on
Sensor or Instrument Description

The TOVS scanner aboard the NOAA polar orbiters includes three sensors, the HIRS/2, MSU and Stratospheric Sounding Unit (SSU). The MSU is a scanning microwave radiometer with four channels in the 50 to 60 GHz oxygen region. The MSU sensors consist of two four-inch diameter antennas, each having an instantaneous field of view (IFOV) of 7.5 degrees.

The 109-km IFOV resolution at nadir creates an underlap, or gap, of approximately 115 km between adjacent scan lines. The MSU data output represents uncorrected brightness temperatures after a 1.84 second integration period (i.e., how long the instrument collects signal from a given position) per step of the scanning antenna.

The MSU has no special calibration sequence to interrupt normal scanning. Calibration data are included in a scan line of data. Each MSU data set normally contains an individual satellite recorder playback. Data with each MSU data set are in chronological order with one record for each MSU scan.

HIRS is a 20 channel scanning radiometer with channels in the 15 micrometer and four micrometer regions. The data are recorded onboard satellite for readout on command. TOVS data are used operationally by NESDIS to produce vertical profiles of temperature and moisture, and to derive other atmospheric variables.

SSU data are not used for this data set.

A summary of the HIRS and MSU instruments' parameters is given in Table 6.

Table 6. HIRS and MSU Instrument Parameters
Instrument Parameters HIRS/2 MSU
Cross-track scan angle (from nadir) 49.5° 47.35°
Scan time 6.4 seconds 25.6 seconds
Ground IFOV at nadir 17.4 km 109.3 km
Ground IFOV at end of scan
58.5 km 323.1 km
Ground IFOV at end of scan
29.9 km 178.8 km
Distance between IFOV centers
42.0 km 168.1 km
Swath width +/-1120 km +/-1174 km
Data precision 13 bits 12 bits
Time between start of each scan line 6.4 seconds 25.6 seconds

For more instrument information see the NOAA Polar Orbiter Data Users Guide.

Background color on
Data Source

TOVS level 1b radiances for the regions poleward of 60° N and 60° S were obtained from the Satellite Data Services Division (SDSD) of the National Oceanographic and Atmospheric Administration (NOAA) and from the National Center for Atmospheric Research (NCAR). Level 1b data files contain raw, quality-controlled radiances as well as calibration and navigation information (Francis 1994).

Background color on
Derivation Techniques and Algorithms

"The 3I algorithm uses HIRS and MSU measurements to deduce the three-dimensional thermal structure of the atmosphere through inversion of the radiative transfer equation. Physical inversion methods consist of solving the radiative transfer equation iteratively until agreement is found between observed and calculated radiances. The 3I algorithm improves upon other iterative methods by including more physics than do purely statistical models. For this application distinctly polar characteristics and the unique physical aspects of snow and ice have been considered. The 3I algorithm also makes use of a library (called TIGR, TOVS Initial Guess Retrieval) of some 1800 atmospheric profiles culled from a global set of more than 150,000 radiosonde measurements. The library acts as a "look-up" table, speeding the computational process. To build the library, a forward radiative transfer model was developed for each of the 1800 profiles and used to calculate brightness temperatures for each HIRS and MSU channel, the Jacobians of the partial derivatives of the radiances B with respect to temperature T and moisture q at each level, and the temperature and radiance means and covariance matrices. Calculations were performed for 10 viewing angles, 19 surface pressures, and two surface emissivities." (Francis 1994)

Background color on
Processing Steps

Below are the steps used in the data retrieval process:

  • Navigate and calibrate the radiances.
  • Detect clouds and remove their effects.
  • Select the first-guess profile from TIGR and retrieve final temperature profile.
  • Retrieve cloud-top pressure and cloud fraction.
  • Retrieve relative humidity, surface temperature, and total precipitable water.

For elevations greater than 1000 meters, all parameters are marked as bad for that orbit. Since each orbital footprint has a unique position, and though several orbits may fall into the same 100 x 100 grid cell, their elevation information is coming from different parts of that cell, and therefore giving different elevations. Thus, regions defined as being higher than 1000 meters vary from day to day due to varying satellite input data.

Background color on

References and Related Publications

Contacts and Acknowledgments

Jennifer Francis
Institute of Marine and Coastal Sciences, Rutgers University
New Brunswick, New Jersey

Axel Schweiger
University of Washington, Applied Physics Laboratory - Polar Science Center
Seattle, Washington

Document Information


September 1999


February 2009 
November 2008 
July 2008 
August 2006

No technical references available for this data set.
No FAQs or How Tos available for this data set.

Access complete Knowledge Base

Questions? Please contact:
NSIDC User Services
Phone: 1 303 492-6199