Assessing vertical structure of an endemic forest in succession using terrestrial laser scanning (TLS). Case study: Guadalupe Island

Abstract Endemic species comprise 16% of the floral species on Guadalupe Island, including three arboreal species. Almost 96% of the forest coverage was lost due to impacts related to feral goats and wildfires. To date, goats have been eradicated and the restoration of the native vegetation communities is underway. The purpose of this study was to develop a 3D Structural Classification Method (3D-SCM) using Terrestrial Laser Scanner (TLS) for mapping the vertical structure of the forest and to automatically characterize its physical attributes. Several TLS scans were performed in July 2016 to assess this recovery, particularly the succession progression. The 3D-SCM used the Forest Condition Classification (FCC) based on radiometric intensity and height values, and the Individual Shape Index (ISI) based on geometry and intensity parameters. It was designed to classify vegetation by stratum using a trial-and-error model to semi-automatically identify impacted areas and succession. The 3D-SCM was able to differentiate the vegetation strata by each forest community at the stand level with a precision of 93% as well as the tree structure parts (stem, branches, and leaves) with a precision of 97%, it also fits the best characteristic shape for the Guadalupe cypress (decagon) and for the pine (droplet). The resulting 3D-SCM provide precise physical specifications, demonstrates a high correlation R2 = 0.949 for (Diameter Breast Height) DBH, R2 = 0.974 for the crown and R2 = 0.97 for heights between maximum laser pulse of tree heights and crowns with respect to field measurements. Mapping 3D vertical forest measurements are important for quantifying the success of management practices and to assist future restoration actions, as it will allow quantification of forest dynamics and carbon sequestration on this island.