Return to the Research Projects at NSIDC page

Merging of Precursor and EOS Data Sets: Resampling of AMSR-E Swath TBs to EASE-Grid

NASAThis project is funded by NASA grant NNG03GN38G


To resample AMSR-E swath brightness temperatures to EASE-Grid


Richard Armstrong

Project Summary

For temporal investigations including the production of CDRs, original satellite swath data must be resampled to a fixed Earth grid in such a way that brightness temperature histories refer to specific areas on the ground that are stationary in time and common to all channels of the sensor. Because the passive microwave data are inherently of low spatial resolution, the resampling scheme should minimize loss of spatial information. Resampling to an effective common spatial resolution implies resampling all data to the lowest spatial resolution of any channel. For SSM/I, this corresponds to the 40 by 70 kilometer spatial resolution (-3dB footprint) of the 19 GHz channels. AMSR-E offers many improvements over SSM/I. In addition to more frequencies and higher spatial resolution at the equivalent SSM/I frequencies, the 6.9 GHz channel, the channel having the lowest spatial resolution (50 kilometers), is oversampled by a factor of five. This significant oversampling offers possibilities of synthesizing more ideal footprints with reduced elongation and deconvolving antenna temperatures for limited improvements in spatial resolution.

Investigators have

  • completed test versions of the software and ancillary data needed to implement the Backus-Gilbert resampling for the AMSR-E data
  • defined and tested the five and 10 kilometer versions of the EASE-Grid
  • modified NSIDC's AMSR-E processing system to handle the latest format changes to the Level-2A data by building “scaffold” software to do timing tests for various approaches to implementing the resampling
  • chosen a method and completed the software design for the production system.

The resampling of AMSR-E swath TBs to EASE-Grid is currently being undertaken using a basic inverse distance method but we will implement the new method based on the actual antenna pattern during the coming year.

Within this project, work continued on an optimal procedure to blend the EOS MODIS and AMSR-E snow products. Work on this blended product involves building a turnkey software system, whereby the investigators “blend” global SWE derived from passive microwave with snow extent derived from the MODIS CMG product. Given an 8-day MODIS product file, one command currently produces a blended product using Pathfinder SSM/I data, up to and including the most recent “near-real-time” (within the last 24 hours) SSM/I data. With slightly more manual intervention, investigators can create the blended product using the standard AMSR-E daily snow product.

This development includes generating gridded AMSR-E TBs from the standard Level-2A TB product, and modifying the current software that produces SWE from SSM/I or SMMR data to also accept AMSR-E gridded TBs as input. With some operator intervention (retrieving and staging Level-2A TB files for the desired time period), investigators can produce Northern Hemisphere gridded AMSR-E TBs. The SSM/I SWE algorithm software is being modified to accept AMSR-E TBs as input.

Related Resources

For more information, see the NSIDC AMSR-E/Aqua Daily EASE-Grid Brightness Temperatures product page.


Contact NSIDC User Services for more information.