Estimating forest biomass from LiDAR data: A comparison of the raster-based and point-cloud data approach

Abstract:

Accurate knowledge of forest biomass and its dynamics is critical for better understanding the carbon cycle and improving forest management decisions to ensure forest sustainability. LiDAR technology provides accurate estimates of aboveground biomass in different ecosystems, minimizing the signal saturation problems that are common with other remote sensing technologies. LiDAR data processing can be based on two different approaches. The first is based on deriving structural metrics from returns classified as vegetation, while the second one is based on metrics derived from the canopy height model (CHM). The CHM is obtained by subtracting the digital elevation model (DEM) that was created from the ground returns, from the digital surface model (DSM), which was itself constructed using the maximum height within each grid cell. The former approach provides a better description of the vertical distribution of the vegetation, whereas the latter significantly reduces the computational burden involved in processing point cloud data at the expense of losing information. This study evaluates the performance of both approaches for biomass estimation over very different ecosystems, including a Mediterranean forest in the Sierra Nevada Mountains of California and a tropical forest in Barro Colorado Island (Panama). In addition, the effect of point density on the variables derived, and ultimately on the estimated biomass, will be assessed.