Documentation for AMSR-E/Aqua L2B Surface Soil Moisture, Ancillary Parms, & QC EASE-Grids, Version 2

Back to Top

Collapse All / Open All

Detailed Data Description


Level-2B soil moisture data are unique to AMSR-E products. It consists of point data in Hierarchical Data Format - Earth Observing System (HDF-EOS) format where the resulting grid is in a table format rather than a grid that image processing programs can easily visualize. Files contain core metadata, product-specific attributes, and data fields. The data fields are summarized in Table 1. Table 1 uses the following notations to describe the data types:

Float64: 64-bit (8-byte) floating-point integer
Float32: 32-bit (4-byte) floating-point integer
Int16: 16-bit (2-byte) signed integer

Note: The number of records per granule depends on the number of gridded points over land.

Table 1. Data Fields

Field Name

Data Type



Fill Value

Time Float64 Scan start time in International Atomic Time in seconds with 01 January 1993 00:00:00 as the zero base (TAI93). n/a n/a
Latitude Float32 Latitude (-90.0 to 90.0) n/a 99 / 98
Longitude Float32 Longitude (-180.0 to 180.0) n/a 999 / 998
Row_Index Int16 EASE-Grid row index (1-586) n/a -9999
Column_Index Int16 EASE-Grid column index (0-1382) n/a -9999
TB_QC_Flag Int16

Brightness temperature (Tb) quality control flag. A non-zero value indicates a given channel is out of limits for a given pixel as the following values indicate. These values indicate the first bad channel detected, though more than one channel may be bad.


Good Tb in all channels


Bad Tb in 89H GHz


Bad Tb in 89V GHz


Bad Tb in 36.5H GHz


Bad Tb in 36.5V GHz


Bad Tb in 23.8H GHz


Bad Tb in 23.8V GHz


Bad Tb in 18.7H GHz


Bad Tb in 18.7V GHz


Bad Tb in 10.7H GHz


Bad Tb in 10.7V GHz


Bad Tb in 6.9H GHz


Bad Tb in 6.9V GHz
n/a -9999
Heterogeneity_Index Int16 As part of the Level-2B processing, a heterogeneity index is computed as the standard deviation of the 36.5H GHz, 11 km resolution data points within each 25 km EASE-Grid cell. The index is used as an output data quality flag. A value of -9999 implies bad Tb data in any channel (TB_QC_Flag). Multiply data values by 0.01 to obtain units in Kelvins (K). -9999
Surface_Type Int16 Indicates surface type classification. n/a -9999
Soil_Moisture Int16 Soil moisture at 10.7 GHz resolution. A value of -9999 indicates no retrieval due to bad Tbdata in the retrieval channels (TB_QC_Flag), or screening by land surface classification (Inversion_QC_Flag_1). Multiply data values by 0.001 to obtain soil moisture in g cm-3. Range: 0 to 500. -9999
Veg_Water_Content Int16 Vegetation and surface roughness parameter at 10.7 GHz resolution. This term incorporates effects of vegetation and surface roughness together. See the Derivation Techniques and Algorithms section of this document. A value of -9999 indicates no retrieval. Multiply data values by 0.01 to obtain vegetation water content in kg m-2. Range: 0-1000. -9999
Land_Surface_Temp Int16 Land surface temperature is not calculated because of Radio Frequency Interference (RFI) contamination in the 6.9 GHz channels. The field contains only fill values (-9999). Multiple data values by 0.1 to obtain land surface temperature. -9999
Inversion_QC_Flag_1 Int16 Inversion quality control flag. Values are as follows:

10: Good retrieval using empirical algorithm
12: Bad retrieval using empirical algorithm
14: No retrieval
20: Good retrieval using iterative algorithm
22: Questionable retrieval using iterative algorithm
24: Bad retrieval using iterative algorithm
26: No retrieval
n/a -9999
Inversion_QC_Flag_2 Int16 Not currently used. n/a -9999
Inversion_QC_Flag_3 Int16 Not currently used. n/a -9999


Sample Data Record

Figure 1 and Figure 2 are an example of the data fields for Level-2B soil moisture data.

Figure 1. Data Fields Table Columns 1 - 10

Figure 2. Data Fields Table Columns 11 - 15

Background color on
File Naming Convention

This section explains the file naming convention used for this product with an example. The date and time correspond to the first scan of the granule.

Example file name: AMSR_E_L2_Land_T06_200706062353_D.hdf


Refer to Table 2 for the valid values for the file name variables listed above.

Variable Description
Table 2. Valid Values for the File Name Variables


Product Maturity Code (Refer to Table 3 for valid values.)


file version number


four-digit year


two-digit month


two-digit day


hour, listed in UTC time, of first scan in the file


minute, listed in UTC time, of first scan in the file


orbit direction flag (A = ascending, D = descending)


HDF-EOS data format


Table 3. Valid Values for the Product Maturity Code

Product Maturity Code



Preliminary - refers to non-standard, near-real-time data available from NSIDC. These data are only available for a limited time until the corresponding standard product is ingested at NSIDC.


Beta - indicates a developing algorithm with updates anticipated.


Transistional - period between beta and validated where the product is past the beta stage, but not quite ready for validation. This is where the algorithm matures and stabilizes.


Validated - products are upgraded to Validated once the algorithm is verified by the algorithm team and validated by the validation teams. Validated products have an associated validation stage. Refer to Table 4 for a description of the stages.


Table 4. Validation Stages

Validation Stage


Stage 1

Product accuracy is estimated using a small number of independent measurements obtained from selected locations, time periods, and ground-truth/field program efforts.

Stage 2

Product accuracy is assessed over a widely distributed set of locations and time periods via several ground-truth and validation efforts.

Stage 3

Product accuracy is assessed, and the uncertainties in the product are well-established via independent measurements made in a systematic and statistically robust way that represents global conditions.

Table 5 provides examples of file name extensions for related files that further describe or supplement data files.

Table 5. Related File Extensions and Descriptions
Extensions for Related Files Description
.jpg Browse data
.qa Quality assurance information
.ph Product history data
.xml Metadata files

Background color on
File Size

Each half-orbit granule is on the average 0.61 MB. The actual size depends on the number of gridded points over land.

Background color on
Spatial Coverage

Spatial Coverage Map

The following map shows a typical day of coverage with 28 half-orbits.

Coverage is global between 89.24°N and 89.24°S, except for snow-covered and densely-vegetated areas. See AMSR-E Pole Hole for a description of holes that occur at the North and South Poles. The swath width is 1445 km.

Spatial Resolution

Input brightness temperature data at 10.7 GHz, corresponding to a 38 km mean spatial resolution, are resampled to a global cylindrical 25 km EASE-Grid cell spacing. The effective spatial resolution is thus slightly higher than the inherent 38 km resolution.


Input brightness temperature data at 10.7 GHz, corresponding to a 38 km mean spatial resolution, are resampled to a global cylindrical 25 km EASE-Grid cell spacing. The effective spatial resolution is thus slightly higher than the inherent 38 km resolution.

Grid Description

Level-2A brightness temperatures are resampled to a global cylindrical EASE-Grid (1383 columns by 586 rows) with a nominal grid spacing of 25 km (true at 30°S). Level-2B data are unique to AMSR-E products. They consist of HDF-EOS point data where the resulting grid is in table format, rather than a grid that image processing programs can easily visualize. In the case of the Level-2B soil data, each geophysical variable value has a corresponding EASE-Grid row and column index.

Please refer to All About EASE-Grid for more information on related products and tools.

Background color on
Temporal Coverage

Temporal coverage is from 19 June 2002 to 3 October 2011. See AMSR-E Data Versions for a summary of temporal coverage for different AMSR-E products and algorithms.

Temporal Resolution

Each swath spans approximately 50 minutes. The data sampling interval is 2.6 msec per sample for the 6.9 GHz to 36.5 GHz channels, and 1.3 msec for the 89.0 GHz channel. A full scan takes approximately 1.5 seconds. AMSR-E collects 243 data points per scan for the 6.9 GHz to 36.5 GHz channels, and 486 data points for the 89.0 GHz channel.

The number of satellite passes per day is a function of latitude as shown in AMSR-E Observation Times.

Background color on
Parameter or Variable

Surface soil moisture (g cm-3): soil moisture in the top ~1 centimeters of soil, averaged over the retrieval footprint

Vegetation water content (kg m-2): equivalent water content (including surface roughness) in the vertical column of vegetation, averaged over the retrieval footprint

The following are the estimated accuracy of the output variables:

  • Surface soil moisture: 0.06 g cm-3
  • Vegetation/roughness parameter: 0.15 kg m-2
Background color on

Software and Tools

Software and Tools

For tools that work with AMSR-E data, see the Tools for AMSR-E Data Web page.

For general tools that work with HDF-EOS data, refer to the NSIDC: Hierarchical Data Format - Earth Observing System (HDF-EOS) Web site.

Background color on

Data Acquisition and Processing

Data Source

AMSR-E/Aqua L2A Global Swath Spatially-Resampled Brightness Temperatures (Tb) are used as input to calculating soil moisture variables. Static input databases are used for surface classification and to identify valid grid points for retrieval.

Surface topography data are derived from the United States Geological Survey (USGS) GTOPO30 global digital elevation model. Horizontal grid spacing is 30 arc seconds. Preprocessing of these data enables screening out points over ocean, mountains, and areas where the topographic variability within a grid cell is likely to degrade geophysical retrievals.

Sand and clay fraction are derived from a 1 degree x 1 degree latitude/longitude global soil type database that estimates soil dielectric properties as a function of soil moisture content.

A mask of permanent ice and snow is used to screen out these areas over land.

Vegetation type is derived from the USGS 1 km global land cover characteristics database. These data estimate the dependence of vegetation type on the model coefficient that relates vegetation water content to vegetation opacity.

Finally, precipitable water and surface air temperature are derived from National Center for Environmental Prediction (NCEP) or European Centre for Medium-Range Weather Forecasts (ECMWF) global reanalysis climatologies, or from real-time forecast model outputs. These data are used for estimating atmospheric contributions in the geophysical retrieval algorithm (Njoku 1999).

Background color on
Derivation Techniques and Algorithms

The soil moisture algorithm uses Polarization Ratios (PR), which are sometimes called normalized polarization differences of the AMSR-E channel brightness temperatures. PR is the difference between the vertical and horizontal brightness temperatures at a given frequency divided by their sum. This effectively eliminates or reduces surface temperature effects, which is necessary since no dynamic ancillary surface temperature data are input to the algorithm. The algorithm first computes a vegetation/roughness parameter g using PR 10.7 GHz and PR 18.7 GHz, plus three empirical coefficients. Soil moisture is then computed using departures of PR 10.7 GHz from a baseline value, plus four additional coefficients. The baseline values for PR 10.7 GHz are based on monthly minima at each grid cell over an annual cycle.

The vegetation/roughness parameter g incorporates effects of vegetation and roughness together, because both have the same functional form (exponential) in their influence on the normalized polarization differences in the simplified model used in the retrieval algorithm. The parameter g may be interpreted as an equivalent vegetation water content with units of kg m-2. In a desert with no vegetation, any value of g greater than zero is due to roughness only. The value of g reflects the influence of roughness on the polarization ratio as if equivalent vegetation of amount g (kg m-2) were present. If the surface were smooth everywhere, then g would equal the vegetation water content in kg m-2 since the roughness contribution would be zero. Vegetation water content and roughness cannot be determined independently from g, and it is computed primarily as a lumped correction factor for the soil moisture retrieval (Njoku and Chan 2006).

Refer to Njoku et al. (2004) for an assessment of calibration biases over land, and methods used to correct these biases.

Processing Steps

AMSR-E/Aqua L2A Global Swath Spatially-Resampled Brightness Temperatures are resampled to the EASE-Grid using a drop-in-the-bucket method, where all the samples that fall within a grid cell are averaged together. Gridding the data allows brightness temperatures to be combined with ancillary data for classification and retrieval, and allows statistics of brightness temperatures and retrieved variables to be calculated at fixed grid locations from successive orbits.

Surface type classification of gridded brightness temperature identifies and screens grid cells that include major water bodies, dense vegetation, snow, and permanent ice, for which retrievals will not be possible. Other tests are performed, such as for excessive relief, precipitation, frozen ground, and Radio Frequency Interference. However, since the reliability of these tests, as well as their influence on the retrievals, is not well characterized, detection of these conditions does not prevent retrievals from being made and values being written into the output array. The user must decide whether to do further screening on the data based on the surface type flags.

The Surface_Type field in the Level-2B land surface product is a 16-bit integer in which nine possible surface classes are represented by nine individual bits. A bit is assigned with a value of one if the surface class that the bit represents is flagged by the classification subroutine. The sequence begins with the least significant bit, for example, the rightmost bit, in order to keep the resulting word small in decimal values.

       16  15  14  13  12  11  10   9   8   7   6   5   4   3   2   1
      | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ? | ? | ? | ? | ? | ? | ? | ? | ? |

bit 1: Permanent ice sheet
bit 2: Mountainous terrain
bit 3: Snow
bit 4: Frozen ground
bit 5: Precipitation
bit 6: RFI
bit 7: Dense vegetation
bit 8: Moderate vegetation
bit 9: Low vegetation

Surface_Type = (bit 1 * 20) + (bit 2 * 21) + (bit 3 * 22) + (bit 4 * 23) + (bit 5 * 24) + (bit 6 * 25) + (bit 7 * 26) + (bit 8 * 27) + (bit 9 * 28)

For example, if the tests for mountainous terrain, snow, and precipitation were all true and the other tests were all false, the value of Surface_Type would be 0000000000010110 binary, or 22 decimal.

Finally, Level-2B data are output to HDF-EOS data format. The number of records per granule depends on the number of gridded points over land.

Version History

See AMSR-E Data Versions for a summary of algorithm changes since the start of mission.

Error Sources

AMSR-E measurements of soil moisture are directly sensitive only to the top 1 cm of soil averaged over approximately 60 km spatial extent. Significant uncertainty may therefore arise when these measurements are compared against point-derived in-situ data, due to differences in sampling depth and spatial extent between satellite and in-situ observations.

Measurements of soil moisture are most accurate in areas of low vegetation. Attenuation from vegetation increases the retrieval error in soil moisture (Njoku et al. 2003). Surface type classification is assigned to indicate low and moderate vegetation, and retrievals are not performed in dense vegetation.

The retrieval algorithm does not explicitly model effects of topography, snow cover, clouds, and precipitation. Other potential error sources include anomalous inputs from bad radiometric data and low-level processing errors. The processing algorithm includes checks to identify these and other anomalies and assign appropriate flags (Njoku 1999).

Soil moisture retrievals represent averages over the horizontal footprint area and vertical sampling depth in the top ~1 cm of soil. The actual sampling depth varies with the amount of moisture in the soil. Soil moisture deeper than ~1 cm below the surface may not be sensed by AMSR-E.

The 6.9 GHz channel is shared with mobile communication services; therefore, retrievals using this frequency are subject to Radio Frequency Interference (RFI), particularly near large urban land areas. The soil moisture algorithm uses the 10.7 GHz channel to alleviate the RFI problem. For more information, see the Derivation Techniques and Algorithms section of this document.

Refer to Njoku et al. (2004) for an assessment of calibration biases over land, and methods used to correct these biases.

Quality Assessment

Each HDF-EOS file contains core metadata with Quality Assessment (QA) metadata flags that are set by the Science Investigator-led Processing System (SIPS) at the Global Hydrology and Climate Center (GHCC) prior to delivery to NSIDC. A separate metadata file in XML format is also delivered to NSIDC with the HDF-EOS file; it contains the same information as the core metadata. Three levels of QA are conducted with the AMSR-E Level-2 and -3 products: automatic, operational, and science QA. If a product does not fail QA, it is ready to be used for higher-level-processing, browse generation, active science QA, archive, and distribution. If a granule fails QA, SIPS does not send the granule to NSIDC until it is reprocessed. Level-3 products that fail QA are never delivered to NSIDC (Conway 2002).

Automatic QA

Surface type classification screens out invalid grid cells, including major water bodies, permanent ice, dense vegetation, and snow.

Quality control is monitored by convergence or limit checks in the retrieval algorithms. In off-line QC, global fields of soil water content, vegetation water content, and brightness temperature are created by averaging the output Level-2 products onto daily and/or monthly grids. The number of samples, means, and standard deviations are examined for missing data and spatial and temporal coherence.

Operational QA

AMSR-E Level-2A data arriving at GHCC are subject to operational QA prior to processing higher-Level-products. Operational QA varies by product, but it typically checks for the following criteria in a given file (Conway 2002):

  • File is correctly named and sized
  • File contains all expected elements
  • File is in the expected format
  • Required EOS fields of Time, Latitude, and Longitude are present and populated
  • Structural metadata is correct and complete
  • The file is not a duplicate
  • The HDF-EOS version number is provided in the global attributes
  • The correct number of input files were available and processed

Science QA

AMSR-E Level-2A data arriving at GHCC are also subject to science QA prior to processing higher-Level-products. If less than 50 percent of a granule's data is good, the science QA flag is marked suspect when the granule is delivered to NSIDC. In the SIPS environment, the science QA includes checking the maximum and minimum variable values, and percent of missing data and out-of-bounds data per variable value. At the Science Computing Facility (SCF), also at GHCC, science QA involves reviewing the operational QA files, generating browse images, and performing the following additional automated QA procedures (Conway 2002):

  • Historical data comparisons
  • Detection of errors in geolocation
  • Verification of calibration data
  • Trends in calibration data
  • Detection of large scatter among data points that should be consistent

Geolocation errors are corrected during Level-2A processing to prevent processing anomalies such as extended execution times and large percentages of out-of-bounds data in the products derived from Level-2A data.

The Team Lead SIPS (TLSIPS) developed tools for use at SIPS and SCF for inspecting the data granules. These tools generate a QA browse image in Portable Network Graphics (PNG) format and a QA summary report in text format for each data granule. Each browse file shows Level-2A and Level-2B data. These are forwarded from Remote Sensing Systems (RSS) to GHCC along with associated granule information, where they are converted to HDF raster images prior to delivery to NSIDC.

Please refer to AMSR-E Validation Data for information about data used to check the accuracy and precision of AMSR-E observations.

Background color on
Sensor or Instrument Description

Please refer to the AMSR-E Instrument Description document.

Background color on

References and Related Publications

Contacts and Acknowledgments

Investigator(s) Name and Title

Eni G. Njoku
NASA/Jet Propulsion Laboratory
M/S 300-233
4800 Oak Grove Drive
Pasadena, CA 91109 USA

Technical Contact

NSIDC User Services
National Snow and Ice Data Center
University of Colorado
Boulder, CO 80309-0449  USA
phone: +1 303.492.6199
fax: +1 303.492.2468
form: Contact NSIDC User Services

Document Information


March 2004


October 2007