Files
Abstract
Forest inventory, an essential component of forestry and natural resource management, heavily relies on the precise measurement of tree height. Nevertheless, the accuracy of this measurement is often compromised by instrumental inaccuracies and human subjectivity. To address these challenges, this thesis explores the potential of Light Detection and Ranging (LiDAR) technology as an alternative solution, offering a reduction in instrument-related errors and the elimination of human biases. In the initial chapter, we investigate the pivotal role of tree height in developing growth and yield models. Through comprehensive analysis and comparisons between various LiDAR data height extraction methods and field inventory height measurements, we establish that LiDAR-derived heights, specifically from 'pixel_metrics,' closely align with field data. We leverage these findings to incorporate LiDAR-derived height projections into yield prediction equations, unveiling LiDAR's superior performance over traditional field measurements in several metrics. When contrasting projected field volume with actual volume in the second year, LiDAR-estimated volume (Adj. R2 = 0.85) exhibits a slight advantage over field projections (Adj. R2 = 0.85). Similarly, in green weight calculations, LiDAR-predicted green weight in year 2 (Adj. R2 = 0.85) slightly outperforms field-derived green weight (Adj. R2 = 0.84).In the third chapter, our exploration expands to encompass Unmanned Aerial Vehicle (UAV) derived LiDAR data characterized by high point density. This expansion introduces enhanced resolution and examines its correlation with forest volume and green weight using PMRC 2014 equations. Our LiDAR metrics model for volume gives an R-squared (Adj. R2) value of 0.79, while the green weight model exhibits an Adj. R2 value of 0.81. Although this approach shows immense potential, we also observe minor discrepancies in volume and biomass estimates, which can be attributed to factors such as the lack of data for lower yield values and variations in data variability between LiDAR and PMRC estimates. Nonetheless, this study highlights the potential of UAV LiDAR data as a valuable tool for precise forest inventory at the plot level. The innovative methods explored in this thesis, combined with data analysis, hold the potential to transform forest management, providing more accurate assessments and contributing to the sustainability of natural resource management.