Use of Terrestrial Laser Scanning to Model Fuel Characteristics in Shrub-Steppe

Biological invasion, climate change, and other anthropogenic and non-anthropogenic factors are altering ecosystem function of arid shrublands in the western U.S., with notable effects including changes in community composition and increased incidence and severity of wildfires. Wildfire itself contributes to replacement of native flora communities with fire-prone invasives (prominently cheatgrass, Bromus tectorum), a positive feedback loop which threatens long-term degradation of burned areas. Efficient methods of vegetation inventory over large areas are essential to study and manage changes in ecological paradigms, and furthermore to anticipate and control wildfire. However, the application of remote sensing information from aerial or satellite platforms to shrub-steppe ecosystems is limited by spectral signal mixing and coarseness of data relative to low-stature vegetation. Terrestrial laser scanning (TLS) technology provides rapid collection of high-resolution structural information at ranges up to hundreds of meters, offering an opportunity to efficiently record vegetation characteristics in large swaths. We tested the ability of TLS to quantify abundance and biomass of different vegetation stem diameter classes in shrub-steppe plots in southwestern Idaho, with classes selected to emulate timelag fuel classes commonly used in fuel inventories and fire modeling. We used data from destructively-sampled reference quadrats within scans for training and evaluation of TLS-derived estimates. We demonstrate TLS as an effective standalone tool for shrubland vegetation inventory, while future applications of these methods include collecting training data for interpretation of coarser remote sensing information, and providing accurate 3D simulations of fuel beds to spatially explicit wildfire models.