A scanning lidar, a relatively new type of sensor which INTRODUCTION explicitly measures canopy height, was used to measure
Structural descriptions of forests are crucial to understructure of conifer forests in the Pacific Northwest.
standing how forest ecosystems function. In particular, SLICER (Scanning Lidar Imager of Canopies by Echo Reinformation on broad-scale patterns of mass and vertical covery), an airborne pulsed laser developed by NASA canopy structure would help advance studies of the global which scans a swath of five 10-m diameter footprints along C cycle (Post, 1993), forest productivity (Ryan and the aircraft's flightpath, captures the power of the re-Yoder, 1997), use of forest canopy habitats by birds flected laser pulse as a function of height from the top of (MacArthur, 1958), arboreal mammals (Carey, 1996), and the canopy to the ground. Ground measurements of forarthropods (Schowalter, 1995), interactions between forest stand structure were collected on 26 plots with coinciests and streams (Gregory et al., 1991), and prediction of dent SLICER data. Height, basal area, total biomass, and the behavior of wildfires in the canopy (Rothermel, 1991). leaf biomass as estimated from field data could be pre-Although fine-scale studies have demonstrated the indicted from SLICER-derived metrics with r 2 values of fluence of structural characteristics on function, applying 0. 95, 0.96, 0.96, and 0.84,this knowledge at broad scales has been problematical ships were strong up to a height of 52 m, basal area of because information on broad-scale patterns of vertical 132 m 2 /ha and total biomass of 1300 Mg/ha. In light of canopy structure has been very difficult to obtain. Pasthese strong relationships, large-footprint, airborne scansive remote sensing tools such as the Thematic Mapper ning lidar shows promise for characterizing stand struccannot provide detailed height, total biomass, or leaf bioture for management and research purposes. ©Elsevier mass estimates beyond early stages of succession in for-Science Inc., 1999 ests with high leaf area or biomass. Using Landsat TM data for example, Cohen et al. (1995) could only distinguish two structural classes for forests older than 80 years. Past efforts to characterize the vertical dimension of