Cloud detection has been difficult in polar regions. Low-level clouds and snow have similar properties in visible and near-infrared portions of the spectrum., especially for nadir views. Since satellites have historically used a nadir view for the earth’s surface, cloud algorithms for the poles have been prone to more uncertainty than other regions of the earth.

One solution to reduce this uncertainty is to view the earth at different angles to differentiate clouds from the surface. Clouds and surface features vary in the amount of radiation that gets scattered into different angular directions. The Multi-angle Imaging SpectroRadiometer (MISR) on EOS-Terra has the unique capability to measure radiances at multiple angles, allowing for cloud detection based on non-nadir scattering properties.

The Band-Differenced Angular Signature (BDAS) technique utilizes the cameras that measure forward-scattered radiances. The BDAS uses information from the 60 degree and 70.5 degree cameras on MISR to determine differences in the forward scattered radiation between the blue band and near-infrared band. Using this method, clouds generally have a more positive BDAS value, while surface features tend to have lower values, or negative values. Examples of the application of the BDAS for cloud detection will be shown for the Greenland ice sheet and regions of ocean in the subtropics. In addition, examples of cloud detection using near-infrared thresholds will be presented.