Characterization of ICESat/GLAS waveforms over terrestrial ecosystems: Implications for vegetation mapping

ICESat/GLAS laser altimetry data have become increasingly utilized for vegetation mapping and canopy characterization. Waveform shapes are dependent upon complex relationships between several factors on the illuminated surface including topography, brightness, clouds, satellite pointing, laser energy, footprint size, shape and orientation, and vegetation height and position within the footprint. However, the understanding of these factors is presently unclear. We first examine the simple case introduced into the GLAS waveform by laser retro-reflectors placed at White Sands, New Mexico, as a proxy for vegetation height detection. We observed that the 1/e(2) energy distribution was only an approximation and that strong reflectors contribute to the returned energy beyond the estimated parameters. The precise position of the corner cube reflector within each footprint coupled with laser pointing angle is an important factor on the estimated vegetation height. Next we examine the implications of vegetation structure and surface topography on the waveform shape and derived elevations, compared with data from an airborne lidar system at Freeman Ranch, Texas, which has been targeted by ICESat since 2005. Small-footprint waveforms were combined to synthesize the energy distribution within a GLAS footprint, and they compared well (> 90% correlation) to the GLAS waveforms. The GLAS-estimated canopy heights compared well (mu = 0.21 m) to the airborne lidar estimates during leaf-on conditions. However, GLAS-estimated ground elevations are biased by similar to 1 m compared to airborne lidar in vegetated regions. In this landscape, the GLAS-energy ratio (canopy-to-ground energy) was a good indicator (R-2 = 0.74) of the amount of woody cover within the footprint.