Researchers from Columbia University and the University of Western Ontario have combined Landsat TM data with European Remote Sensing 2 radar data and lidar-derived elevation data to map a remote boreal landscape in northern Alberta, Canada at a 6 meter resolution with 98% accuracy. Their technique uses discrete wavelet transformations to find relationships between the three data types and then classify the land cover based on those relationships. This work is not landscape-specific and could be used for many different land cover types. The application goal of this work is to allow forest managers to better assess land under their stewardship with minimal disturbance.
Kaheil, Y.H., I.F. Creed (2009). “Detecting and Downscaling Wet Areas on Boreal Landscapes.” IEEE Geoscience and Remote Sensing Letters. vol. 6, no. 2, pp. 179–183.
“Mapping wet landscapes in more detail;” environmental research web [external link]