On Wednesday, January 27 from 9 a.m. to 12 p.m. (USA Mountain Time), the following data collections may not be available due to planned system maintenance: AMSR-E, Aquarius, High Mountain Asia, IceBridge, ICESat/GLAS, ICESat-2, MEaSUREs, MODIS, NISE, SMAP, SnowEx, and VIIRS.
Temporal gaps in this data set between the start date and the current date are in the process of being filled.
Data Set ID: 

AMSR-E/AMSR2 Unified L2B Global Swath Surface Precipitation, Version 1

This AMSR-E/AMSR2 Unified Level-2B data set provides instantaneous surface precipitation rate and type over land and ocean, and precipitation profiles over ocean. This global data set is generated using the Goddard Profiling Algorithm.

Note: This product uses the Japan Aerospace Exploration Agency (JAXA) AMSR2 Level-1R input brightness temperatures that are calibrated, or unified, across the JAXA AMSR-E and JAXA AMSR2 Level-1R products.

This is the most recent version of these data.

Version Summary: 

Initial release

COMPREHENSIVE Level of Service

Data: Data integrity and usability verified; data customization services available for select data

Documentation: Key metadata and comprehensive user guide available

User Support: Assistance with data access and usage; guidance on use of data in tools and data customization services

See All Level of Service Details

Data Format(s):
  • HDF-EOS5
Spatial Coverage:
N: 89.24, 
S: -89.24, 
E: 180, 
W: -180
Spatial Resolution:
  • 10 km along track x 5 km along scan
Temporal Coverage:
  • 2 July 2012
Temporal Resolution50 minuteMetadata XML:View Metadata Record
Data Contributor(s):Christian Kummerow, Ralph Ferraro, David Randel

Geographic Coverage

Other Access Options

Other Access Options


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.

Kummerow, C., R. Ferraro, and D. Randel. 2020. AMSR-E/AMSR2 Unified L2B Global Swath Surface Precipitation, Version 1. [Indicate subset used]. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. doi: https://doi.org/10.5067/P5MCTDH7674A. [Date Accessed].
24 March 2020
Last modified: 
5 August 2020

Data Description

The parameters included with this data set are described in Tables 1 and 2, and a sample image is provided in Figure 1. Note that the data file does include several ancillary parameters, which are inputs to the AMSR Unified Rainfall algorithm.


Table 1. Data Parameters
Parameter Description Units Fill Value
CloudWaterPath Total cloud liquid water in the atmospheric column. kg/m2 -9999.0
RainWaterPath Total rain water in the atmospheric column. kg/m2 -9999.0
IceWaterPath Total cloud ice in the atmospheric column. kg/m2 -9999.0
ConvectivePrecip The instantaneous convective precipitation rate. mm/hr -9999.0
FrozenPrecip The instantaneous frozen precipitation rate. mm/hr -9999.0
SurfacePrecip The instantaneous total precipitation rate. mm/hr -9999.0
Temp2Meter* Temperature at 2 meters (T2m) in Kelvin. K -999
TotalColWaterVapor* Integrated water vapor in the atmospheric column (TCWV); synonymous with Total Precipitable Water (TPW). mm -99
L1RQualFlag This quality flag is reserved for future use, currently set to '0'. n/a -99
SunglintAngle The relative angle of reflection between the sun and the line of sight of AMSR2. degree -88
ProbabilityofPrecip The fraction of precipitation vs non-precipitation database profiles that make up the final solution. This parameter is used to assess the likelihood that precipitation is occuring and is necessary because of the nature of the precipitation retrieval. percent -99
QualityFlag This quality flag provides an indication of the retrieval quality. See the 'Quality Information' section for additional details.
  • 0: Good
  • 1: Use with caution
  • 2: Use pixel with extreme care over snow covered surface
  • 3: Use with extreme caution
n/a -99

Quality indicator for the pixel. Non valid pixels (values 1 - 5) cause the algorithm to fail.

  • 0: Valid Pixel
  • 1: Invalid geolocation
  • 2: Sensor brightness temperatures out of range
  • 3: Surface code histogram mismatch
  • 4: Missing TCWV, T2m or sfccode (surface code)
  • 5: No Bayesian solution for pixel
n/a -99

Surface type classes.

  • 1: Ocean
  • 2: Sea-ice
  • 3-7: Decreasing vegetation coverage (3 = max, 7 = min)
  • 8-11: Decreasing snow coverage (8 = max, 11 = min)
  • 12: Inland water
  • 13: Land/water boundary - coast
  • 14: Sea-ice ocean boundary
n/a -99
*Indicates an ancillary parameter used as input to the AMSR Unified Rainfall algorithm.

Table 2. Geolocation Parameters
Parameter Description Units Fill Value
Latitude Latitude of the pixel center. degrees north -9999.0
Longitude Longitude of the pixel center. degrees east -9999.0
SCalt Spacecraft Altitude km -9999.0
SClat Spacecraft Latitude degrees north -9999.0
SClon Spacecraft Longitude degrees east -9999.0
scantime Scan time along track in year, month, day, hour, minute and second. n/a n/a
tai93time Scan time along track measured in seconds since 1993-01-01 00:00:00 seconds -9999.0

Sample Data Image

Figure 1. This image shows Total Column Water Vapor data, acquired during a descending orbit pass, on 13 June 2020.

File Information


Data are provided in HDF-EOS5 format and are stored as 8-bit unsigned integers.

Data File

As shown in Figure 2, each data file includes fourteen data fields, seven geolocation fields, and two metadata fields.

Figure 2. This figure shows the AU_Rain fields included in each data file as displayed with Panoply software.

Ancillary Data

A product history file (.ph), quality assurance file (.qa), and science granule metadata file (.xml) are provided with each 50-minute granule. There are approximately 30 granules generated each day; one for each ascending and descending half-orbit.

Naming Convention

Files are named according to the following convention and as described in Tables 3 and 4.

File Name Convention:

Table 3. Valid Values for the File Name Variables
Variable Description
AMSR Advanced Microwave Sounding Radiometer
U Unified
L2 Level-2
X Product Maturity Code (Refer to Table 4 for valid values.)
## file version number
yyyy four-digit year
mm two-digit month
dd two-digit day
hh hour, listed in UTC time, of first scan in the file
mm minute, listed in UTC time, of first scan in the file
f orbit direction flag (A = ascending, D = descending)
ext file extension (.he5 = HDF-EOS5 file, .qa = quality assurance file, .ph = product history file, .xml =  metadata file)

Table 4. Valid Values for the Product Maturity Code
Variable Description
P Preliminary - refers to non-standard, near-real-time data available from the NSIDC DAAC. These data are only available for a limited time until the corresponding standard product is ingested at the NSIDC DAAC.
B Beta - indicates a developing algorithm with updates anticipated.
T Transitional - 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.
V 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.

Spatial Information


This data set coverage extends from 89°24' N to 89°24' S latitudes. A spatial coverage map for a single day is provided in Figure 3.


The pixel spacing is 10 km along track and 5 km along scan. However, the effective resolution is defined by the field of view (FOV), which varies with frequency for the AMSR instrument. Studies using the derived rainfall distribution indicate an effective resolution similar to the 22 GHz channel, or 26 km along track and 15 km along scan.

Figure 3. This spatial coverage map shows one AMSR2 half-orbit descending pass (purple) for path 13D. Note that the numbers listed across the image (above Antarctica) indicate the descending orbit paths flown on 21 April 2020.

Temporal Information


02 July 2012 to present


Each half-orbit swath spans approximately 50 minutes.

Data Acquisition and Processing


The AMSR Unified Rainfall algorithm uses intercalibrated L1R brightness temperatures provided by JAXA for AMSR-E (flown on the EOS-Aqua) and AMSR2 (flown on GCOM-W) to create a consistent precipitation data record from the two satellites. The passive microwave algorithm is designed to take advantage of a previously constructed a priori database of observed precipitation structures and their associated brightness temperature signals from the Global Precipitation Measurement (GPM) mission. This database is used in conjunction with Bayesian inversion techniques to retrieve surface precipitation and integrated liquid and ice water contents.

The main algorithm output parameters include atmospheric profiles (path parameters) and surface precipitation rates. The path parameters are calculated for pixels over ocean only and provide the total amount of cloud water, rain water, and ice water included in the atmospheric column. The surface precipitation parameters are calculated for both ocean and land pixels and provide the instantaneous surface precipitation rate for convective precipitation, frozen precipitation, and total precipitation.


The AMSR2 instrument is a seven-frequency, total-power passive microwave radiometer system. It measures brightness temperatures (Tb) at the following seven frequencies, or channels: 6.925 GHz, 7.3 GHz, 10.65 GHz, 18.7 GHz, 23.8 GHz, 36.5 GHz, and 89.0 GHz. Vertically and horizontally polarized measurements are taken on all channels. These data consist of swath observations from five channels which have been resampled at JAXA; the 6.925 GHz and the 7.3 GHz channels are not used for precipitation. The Tb sensor footprints, also known as the instantaneous field-of-view (IFOV), vary with frequency. The resampling procedure remaps Tb to sets of consistent footprint sizes using the Backus-Gilbert method (Backus et al., 1967). Each resampled set corresponds to the footprint of one frequency and contains that frequency plus all higher-resolution frequencies. Therefore, the number of channels in each resampled set of Tb varies. For more information, see the JAXA Level 1R documentation or refer to Maeda et al. (2016).

AMSR2 scans the entire globe every one to two days. As such, most locations on Earth are imaged at least once per day; locations where swaths overlap, such as near the poles, are sampled more frequently.


This product is derived from Japan Aerospace Exploration Agency (JAXA) AMSR2 Level-1R input Tb observations from the following channels: 10.7 GHz, 18.7 GHz, 23.8 GHz, 36.5 GHz, and 89 GHz.

Note: This AMSR-E/AMSR2 Unified data set presently includes just the AMSR2 portion. In the future, the AMSR-E portion of the AU_Rain data set will be processed using the JAXA AMSR_E Level-1R Tb observations and the GPROF2017 algorithm, thus providing a unified AMSR-E/AMSR2 Rain product.

In addition to the L1R data, several ancillary data products are required by the processing algorithm. These ancillary products are listed in Table 5, and some are also included as parameters in the data file (see Table 1).

Table 5. Ancillary data included in the AMSR Unified Rainfall Algorithm
Product Source Description



Near surface temperature (T2m)



Total column water vapor

Snow/no-snow classification


Snow cover classification

Sea ice coverage


Sea ice cover


Aires et al. (2011)

Goddard profiling algorithm static surface classes

*Global Modeling and Assimilation Office Forward Processing for Instrument Teams (Located at GSFC)


The following section summarizes the approach used to generate the AU_Rain data product. For a more complete description, see the AMSR Unified Rainfall ATBD (Kummerow et al., 2016), the Goddard Profiling (GPROF) literature (Kummerow et al., 2015), and the Global Precipitation Measurement (GPM) Mission ATBD.

As mentioned above, the processing is designed to take advantage of a previously constructed a priori database of observed precipitation structures from the GPM mission. The profiles in the a priori database are created by the GPM Radar/Radiometer 'Combined' algorithm over oceans and by the GPM Dual Frequency Radar (DPR) Ku-band Radar over land. However, an exception is made for snow-covered areas: in these cases, the precipitation profiles are created from ground-based, radar-derived snowfall rates, which are matched directly to AMSR2 Tb to create the precipitation profile database.

Radiative transfer calculations are then applied to GPM observed cloud and precipitation structures to generate a database of simulated Tb; AMSR-E/AMSR2 L1R Tb observations are subsequently compared against the simulated Tb to find the best match in the database from which the estimated rainfall rate is obtained.

Once the profile database and corresponding simulated AMSR-E/AMSR2 Tb are established, the AMSR Unified Rainfall algorithm uses a Bayesian inversion methodology to calculate the probability of observing a profile R for a particular AMSR-E/AMSR2 Tb vector. This probability, Pr(R | Tb), is calculated as the product of two other probabilities: that of observing this particular profile R (for no particular Tb vector) multiplied by that of observing the brightness temperature vector Tb at the same profile R. Equation 1 provides the mathematical calculation for this situation:

Equation 1. Calculate profile probability


Pr(R) = probability of observing a particular profile R (independent of any particular Tb vector)
Pr(Tb | R) = probability of observing a brightness temperature vector Tb at the particular precipitation profile R

The first term on the right-hand side of Equation 1 is derived from the precipitation profiles (Kummerow et al., 2014); the second term on the right-hand side is obtained from radiative transfer computations using the cloud model profiles. The weighting of each model profile in the compositing procedure is based on an exponential factor containing the mean square difference of the sensor-observed Tb and a corresponding set of Tb obtained from radiative transfer calculations through the cloudy atmosphere represented by the model profile. In summary, the retrieval procedure composes a new hydrometeor profile by taking the weighted sum of structures in the cloud structure database that are radiometrically consistent with the observations. The formal solution to the above problem is presented in detail in Kummerow et al. (1996).

The profile retrieval method is an adaptation of the well-known minimum variance solution for obtaining an optimal estimate of geophysical parameters from available information (Lorenc, 1986). While the mechanics of Bayesian inversions are fairly well understood, the AMSR code does not search the entire a priori database, but instead searches only a subset of profiles with coincident near-surface temperature (T2m) and Total Column Water Vapor (TCWV) within 16 distinct surface classes, as described by Prigent and Aires (2011). The surface database (shown in Figure 4) is static, except for the snow/no-snow classification and sea ice classification, which like T2m and TCWV, are provided by the Global Modeling and Assimilation Office Forward Processing for Instrument Teams (GEOS-5 FP-IT) at the Goddard Space Flight Center.

The algorithm runs in two steps because of the ancillary data requirements. First, the pre-processor ingests the AMSR L1R data and the ancillary data to produce an intermediate file for use by the algorithm. Next, the algorithm reads the database profiles and computes the most likely precipitation rate.

Figure 4.  GPROF static surface classes.  If GEOS5 FP-IT suggests that no snow is present in a snow-covered class, the most recent vegetated class for that pixel is used. Conversely, if snow is indicated by GEOS5 FP-IT in a vegetated class, then minimum snow is assigned to that pixel. Sea ice is assigned dynamically based on GEOS5 FP-IT.

Quality Information

Each HDF-EOS5 data file contains core metadata with Quality Assessment (QA) metadata flags that are set by the operational processing code run by the AMSR Science Investigator-led Processing System (SIPS) prior to delivery to the NSIDC DAAC. A separate metadata file in XML format is also delivered to the NSIDC DAAC with the HDF-EOS file. This file contains the same quality assessment (QA) metadata flags as the core metadata contained in the HDF-EOS file. Three levels of QA are applied to AMSR2 files: automatic, operational, and science. Note that if a granule passes automatic QA and operational QA, the granule is forwarded to the NSIDC DAAC for archival and distribution. If not, the granule is reprocessed. Science QA is performed automatically during nominal processing, but only reviewed closely afterward in conjunction with questions that arise after processing is complete.

Automatic QA

Out-of-bounds Level-1R brightness temperatures are screened out prior to product generation.


AMSR2 L1R data are subject to operational QA by JAXA prior to arriving at AMSR SIPS for processing to 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 are 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 was available and processed

Science QA

In the SIPS environment, as part of the processing code, 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), co-located with the SIPS, post-processing science QA involves reviewing the operational QA files and browse images, and performing the following additional 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

Quality Flags

The QualityFlag variable is included in the data file and provides a generalized quality retrieval measure for each pixel, with pixel values ranging from 0 to 3. Descriptions for each of the pixel values are provided below:

  • 0:  'Good.' Pixel has the highest confidence of the best retrieval.
  • 1: 'Use with caution.' Pixels can be set to 1 for the following reasons: Sunglint is present, RFI, geolocate, warm load or for other L1R 'positive value' quality warning flags.
  • 2: 'Use pixel with extreme care over snow-covered surface.' This is a special value for snow-covered surfaces only. The pixel is set to 2 if the probability of precipitation is of poor quality or indeterminate. Use these pixels for climatological averaging of precipitation, but not for individual storm-scale daily cases.
  • 3: 'Use with extreme caution.' Pixels are set to 3 if they have channels missing critical to the retrieval, but the choice has been made to continue the retrieval for the pixel.


Quantifying errors in this data set is complicated because it involves understanding the nature of precipitation. Uncertainties arise when the rain layer thickness is not well understood, or when inhomogeneous rainfall occurs below the resolution of the satellite. Another potential source of error is the non-precipitating component of clouds, which contribute to Tb. Scattering-based retrievals over land also present many uncertainties, most notably the lack of a consistent relationship between frozen rain aloft and liquid rain at lower altitudes. Quantifying the scattering by ice is especially problematic. Ambiguities occur in the data because microwave radiation is scattered not only by rainfall and associated ice but by snow cover and dry land (Kummerow and Ferraro, 2007).


The following limitations have been identified:

  • Shallow orographic precipitation remains a challenge for passive microwave sensors. Without an ice scattering signature, it is very difficult to distinguish shallow warm rain, often associated with orographic precipitation, from noise.
  • High-latitude oceanic precipitation also remains a problem. GPM radars do not have enough sensitivity to detect light drizzle, while the AMSR instrument by itself has difficulties separating cloud water from drizzle. The results are thus uncertain in areas where much of the precipitation falls as drizzle.
  • Light snowfall is also difficult to detect by AMSR, as snow on the ground can look similar to light precipitating snow. This can be remedied with higher frequency channels (e.g. 166 GHz and 183 GHz), but these are not available on AMSR.


Refer to the AMSR2 NRT Anomalies Page for information regarding data anomalies or gaps in coverage. Updates to this page may be forthcoming.


For a detailed description of the AMSR-E and AMSR2 instrument, refer to the AMSRE-E Instrument Description and AMSR2 Channel Specification and Products page.

Software and Tools

For general tools that work with HDF-EOS data, see the HDF-EOS Tools and Information web page or the NSIDC DAAC HDF-EOS web page.

Contacts and Acknowledgments

Christian D. Kummerow
Colorado State University
Fort Collins, CO

Ralph Ferraro
CICS / University of Maryland
College Park, MD

David Randel
Colorado State University
Fort Collins, CO


Aires, F., Prigent, C., Bernardo, F., Jiménez, C., Saunders, R., & Brunel, P. (2011). A Tool to Estimate Land-Surface Emissivities at Microwave frequencies (TELSEM) for use in numerical weather prediction. Quarterly Journal of the Royal Meteorological Society, 137(656), 690–699. https://doi.org/10.1002/qj.803

Kummerow, C. D., Ringerud, S., Crook, J., Randel, D., & Berg, W. (2011). An Observationally Generated A Priori Database for Microwave Rainfall Retrievals. Journal of Atmospheric and Oceanic Technology, 28(2), 113–130. https://doi.org/10.1175/2010jtecha1468.1

Lorenc, A. C. (1986). Analysis methods for numerical weather prediction. Quarterly Journal of the Royal Meteorological Society, 112(474), 1177–1194. https://doi.org/10.1002/qj.49711247414

Reynolds, R. W., Smith, T. M., Liu, C., Chelton, D. B., Casey, K. S., & Schlax, M. G. (2007). Daily High-Resolution-Blended Analyses for Sea Surface Temperature. Journal of Climate, 20(22), 5473–5496. https://doi.org/10.1175/2007jcli1824.1

Kummerow, C. D., Randel, D. L., Kulie, M., Wang, N.-Y., Ferraro, R., Joseph Munchak, S., & Petkovic, V. (2015). The Evolution of the Goddard Profiling Algorithm to a Fully Parametric Scheme. Journal of Atmospheric and Oceanic Technology, 32(12), 2265–2280. https://doi.org/10.1175/jtech-d-15-0039.1

Hou, A. Y., Kakar, R. K., Neeck, S., Azarbarzin, A. A., Kummerow, C. D., Kojima, M., Oki, R., Nakamura, K., & Iguchi, T. (2014). The Global Precipitation Measurement Mission. Bulletin of the American Meteorological Society, 95(5), 701–722. https://doi.org/10.1175/bams-d-13-00164.1

Berg, W., L’Ecuyer, T., & Kummerow, C. (2006). Rainfall Climate Regimes: The Relationship of Regional TRMM Rainfall Biases to the Environment. Journal of Applied Meteorology and Climatology, 45(3), 434–454. https://doi.org/10.1175/jam2331.1

How To

Programmatic Data Access Guide
Data from the NASA National Snow and Ice Data Center Distributed Active Archive Center (NSIDC DAAC) can be accessed directly from our HTTPS file system or through our Application Programming Interface (API). Our API offers you the ability to order data using specific temporal and spatial filters... read more
Filter and order from a data set web page
Many NSIDC data set web pages provide the ability to search and filter data with spatial and temporal contstraints using a map-based interface. This article outlines how to order NSIDC DAAC data using advanced searching and filtering.  Step 1: Go to a data set web page This article will use the... read more