This project is a continuation of the AMSR-E sea ice algorithm validation and refinement work that was started in 2003 under NASA NRA-03-OES-02. We propose to continue to validate the three AMSR-E standard sea ice products (sea ice concentration, snow depth on sea ice, and sea ice temperature), to identify the sources and magnitudes of error, and to refine the corresponding algorithms. The objective of this proposal is to address the outstanding validation issues detailed below These will be addressed largely through the analysis of the coincident in situ, aircraft and satellite sea-ice measurements collected in the Alaskan Arctic during March 2006 . While the proposed cost is equivalent to two separate proposals, we think that this unified five-investigator approach best meets NASA�s objectives by having the same team of scientists, who worked together in the field, analyze the data.
The March 2006 campaign was the second of two coordinated Arctic field campaigns. The first was successfully completed in March 2003 and provided an initial set of data which were used to start the validation studies of the AMSR-E sea ice products. Results from the analyses of these data will be published as a collection of 11 papers in a special section of IEEE Transactions on Geoscience and Remote Sensing in September 2006. The objective of this proposal is to use the data collected in March 2006 and data from other satellite sensors to address the validation issues that remain.
For the sea ice concentration algorithm we will (a) validate the weather correction scheme and the retrievals under cloudy conditions through a comparison with radar data and with atmospheric retrievals from TOVS and AMSU-B, (b) investigate the generation of retrieval precision errors on a pixel-by-pixel basis, and (c) investigate the retrieval accuracy for different stages of new ice growth using high-resolution visible and radar satellite data over dynamic sea ice areas, such as polynyas and marginal ice zones.
For snow depth on sea ice, the March 2006 campaign provided the first comprehensive data set for validating this product. Our approach will utilize a two-step upscaling process. The first involves a comparison of the airborne and surface measurements made near Barrow to validate the airborne sensor retrievals. These sensors include a passive microwave radiometer with similar frequencies as the AMSR-E, a laser altimeter, a radar altimeter, and a wide-band radar, which measures snow thickness directly. The second step involves a direct comparison of the airborne sensor measurements with AMSR-E retrievals covering large areas in the Chukchi and Beaufort seas under a variety of snow depth and sea ice conditions.
For the sea ice temperature product, the validation approach will also follow a two-step process. First, we will compare the aircraft passive microwave data with the in situ surface and snow-ice interface temperatures measured near Barrow. We will then use the airborne IR surface temperatures with the validated snow depth retrievals to validate directly the AMSR-E temperature retrievals in areas of the Chukchi and Beaufort seas. Consistency between the ice temperature and snow depth retrievals will be an important factor in refining these algorithms.
