EDG Data Set Name
Granule Shortname
Data Set Characteristics
Product Description This product will be especially useful for those studying greenhouse forcings in the polar regions where cloud cover exerts its effects on sea ice conditions, the regional ice-albedo feedbacks, the polar heat balance, and surface melting. This product may also be used in cross-validating the MODIS cloud optical thickness and effective particle sizes, both of which impact the Earth's radiative budget. The algorithm is hierarchical and is executed in four stages: In the first stage, the VNIR data are sub-sampled by half (15 m to 30 m) while the TIR data are sampled three times (90 m to 30 m). The SWIR data at 30 m are retained in their original resolution. The VNIR and SWIR DNs are normalized for solar irradiance, solar zenith angle, observation angle, and calibration coefficients. The TIR DNs are converted to temperature. In the second stage, the data are pre-classified to reduce the class ambiguity of a pixel's feature vector by using key features like coastline data, land-water mask etc. The third stage involves the use of a back-propagation neural network to aid in the resolution of feature vectors to one of nine possible classes. In the fourth stage, a spatial consistency test is performed on the classification mask, and pixels deemed spatially inconsistent with their neighbors are re-classified. The final product is a coded pixel map containing the following thematic classes:
[2] snow/ice [3] ice cloud [4] land [5] thin water cloud [6] water cloud over water [8] sea ice [12] water cloud over land [30] bare ground/tundra View a Version Formatted for Printing
Ordering ASTER On-Demand L2 Polar Surface and Cloud Classification Product
Product Information
Contact Information
|