Validating ALS/TLS LiDAR derived leaf area density profiles against multisource ground truth data and PALM LES microclimate model simulations

This study advances remote sensing and microclimate modeling through the development of VoxPy, a novel approach to derive spatially-explicit leaf area density (LAD) profiles from terrestrial (TLS) and airborne (ALS) LiDAR data for integration into the PALM Large Eddy Simulation (LES) microclimate model. The research was conducted in a managed forest plot in Britz, Brandenburg, Germany, focusing on Sessile Oak (Quercus petraea) stands.

The study compares two methodologies for processing LiDAR point cloud data: our newly developed VoxPy, which implements an efficient TLS proportional leaf area density scaling method, and the established AMAPVox model, which utilizes Free Path Length estimation derived from Beer-Lambert Law. The point clouds underwent preprocessing through noise filtering, crown segmentation, and machine learning-based classification of woody mass and foliage using K-nearest neighbor algorithms.

Initial validation of VoxPy against ground-based leaf area index measurements and the VoxLAD Beer-Lambert model shows strong agreement, demonstrating the method's reliability for LAD estimation. Preliminary results from AMAPVox processing indicate

promising alignment with ground truth data, with full validation ongoing. Statistical analysis reveals platform-specific characteristics, with ALS-derived profiles showing higher sensitivity in upper crown regions and TLS-derived profiles demonstrating stronger accuracy in lower canopy layers.

The first PALM LES microclimate simulation using these LAD profiles during a heatwave period shows encouraging agreement with meteorological tower observations for air temperature through a single diurnal cycle. The simulation employs ICON-D2 mesoscale forcing, with additional three-day simulations planned.

This research establishes a framework for generating high-fidelity canopy structure data essential for microclimate modeling. The preliminary findings underscore the potential impact of precise LAD profiles in simulating forest-atmosphere interactions, with implications for climate-sensitive planning. Future work will focus on comprehensive validation across extended simulation periods.