Variability in Terrestrial Laser Scanning (TLS) derived allometric equations for northern Sierra Nevada forests

Abstract:

Allometric equations form the basis for rapid estimation of tree biomass from easy-to-measure structural parameters such as species, diameter-at-breast-height, and height. These equations have historically been derived via destructive sampling, and have been shown to result in high uncertainties and a lack of portability across different site conditions. These uncertainties largely result from low sample sizes and poor sampling when using destructive sampling techniques. Terrestrial Laser Scanning (TLS) may provide an alternative basis for a non-destructive sampling method to improve aboveground biomass estimates. This approach uses techniques such as quantitative structural modeling (Raumonen et al. 2013) to determine per-tree volume which, when multiplied by the wood density, results in an estimate of the tree biomass. Per-tree volume estimates using these approaches have been found to be relatively accurate, but the impact on variability in wood density between and within species, and across different site conditions, has not been well established. As such, we are asking: how does wood density affect variability in TLS-derived allometric equations? To accomplish this, we developed species-and regionally-specific allometric equations based on species, diameter-at-breast height, height and crown spread, climate factors, and stand characteristics for all major tree species found in northern Sierra Nevada forests. Uncertainties in both TLS-derived volume estimations as well as wood densities were propagated into the allometric models. Our results provide a new set of allometric equations for northern Sierra Nevada tree species with improved estimates of uncertainties as well as the source of these errors.