ARISE (Antarctic Remote Ice Sensing Experiment) in the East 2003: Validation of satellite-derived sea-ice data products.

摘要

Preliminary results are presented from the first validation of geophysical dataproducts (ice concentration, snow thickness on sea ice ( ) and ice temperature ( ) fromthe NASA EOS Aqua AMSR-E sensor, in East Antarctica (in September-October 2003). Thechallenge of collecting sufficient measurements with which to adequately validate thecoarse-resolution AMSR-E data products was addressed by means of a hierarchicalapproach, using detailed in situ measurements, digital aerial photography and other satellitedata. Initial results indicate that, at least under cold conditions with a dry snowcover, thereis a reasonably close agreement between satellite- and aerial photo-derived iceconcentrations i.e., 97.2 ±3.6% for NT2 and 96.5 ±2.5% for BBA algorithms versus 94.3±10% for the aerial photos. In general, the AMSR-E concentration represents a slightoverestimate of the actual concentration, with the largest discrepancies occurring in regionscontaining a relatively high proportion of thin ice. Although the AMSR-E concentrations fromthe NT2 and BBA algorithms are similar on average, differences of >5% occur on a point-by-point basis, again related to thin ice distribution. The AMSR-E ice temperature ( ) productagrees with coincident surface measurements to within approximately 0.5o C. Regardingsnow thickness, the AMSR retrieval is a significant underestimate compared to in situmeasurements weighted by the percentage of thin ice (and open water) present. For thecase study analysed, the underestimate was 46% for the overall average, but 23%compared to smooth ice measurements. An encouraging factor is that the spatialdistribution of the AMSR-E product follows an expected and consistent spatial pattern,suggesting that the observed difference may be an offset (at least under freezingconditions). Areas of discrepancy are identified, and the need for future work highlighted.