Characterizing Seedling Change Along Seismic Lines in the Boreal Forest of Alberta with Airborne Laser Scanning

Abstract

Seismic lines – linear disturbance corridors constructed for petroleum exploration – contribute to shifts in wildlife populations, species interactions, and other ecosystem functions in Alberta’s boreal forest. As a result, characterization of post-disturbance vegetation growth along these lines is important for forest management. However, peer-reviewed work aimed at mapping vegetation growth along seismic lines with Airborne Laser Scanning (ALS) is limited. Here, I demonstrate how ALS data employed to map conifer seedling growth along seismic lines is influenced by choice of remote-sensing unit of analysis and ecosite type. I compared point-cloud-based and raster-based approaches to estimate height-change for 194 field-measured seedlings located in upland and wetland ecosite types in a 7 km 2 study area over a five-year monitoring period. Results show that point-cloud models produced more accurate estimates of seedling height change than raster models, with a point-cloud 60th percentile metric model performing best (RMSE = 25.70 cm, MAE = 15.05 cm). Upland sites produced more accurate results (RMSE = 20.55 cm, MAE = 11.27 cm, MEDAE = 7.03 cm) than wetland sites. Mapping seedling growth across the study area revealed that 33.24 km of seismic lines were in a state of minimal or no growth, while 4.21 km and 3.65 km were in a state of moderate and high growth, respectively. I found that factors such as overhanging trees or adjacent forest noise influenced the accuracy of height-change estimates. Overall, this study demonstrates the applicability of ALS technologies for mapping seedling growth across large areas and supports further research using these approaches.