Testing two cloud removal algorithms for SSM/I

The ability to monitor and map change in tropical forest regions is critical for the study of both carbon dioxide exchange and global climate. Remote sensing provides a very cost-effective and efficient method to monitor and map rainforest extent. In particular, moderate-resolution spaceborne sensors such as the Special Sensor Microwave/Imager (SSM/I) provide the ability to monitor large geographic areas such as the Amazon Basin with great frequency. However, despite their sophistication, passive sensors such as the SSM/I are unable to “see” through heavy clouds. This research was designed to test the effectiveness of two simple algorithms used to remove the effects of cloud cover on SSM/I data. The study area was the Amazon Basin. The approach taken in this research was to subtract original SSM/I imagery from the algorithm-processed imagery on a pixel-by-pixel basis. This was done for each of the SSM/I bands. These difference images were then examined statistically against rainfall data acquired from 321 stations in the Amazon Basin. Based on correlation analysis, it appears that the two algorithms are very effective in removing cloud contamination from SSM/I data. However, their effect varied by SSM/I band and polarization