Seasonal dynamics of UAV-LiDAR derived canopy transmittance in a mixed-forest ecosystem

Forest vegetation regulates carbon and water fluxes and mediates the exchange of energy between the land surface and the atmosphere. However, quantitative information on how structural changes alter light penetration and their intra/inter‑annual dynamics is still limited. Phenological shifts are usually inferred from spectral indices such as the NDVI, which provide only indirect estimates of canopy cover. In this contribution we present a comprehensive UAV‑LiDAR (DJI Zenmuse L2) time‑series that records (bi‑monthly) overflights (25 flights) throughout two vegetation seasons over the ECOSENSE field site in Germany, covering 7 ha mixed‑temperate forest dominated by F. sylvatica and P. menziesii.

A dedicated processing chain was implemented to extract transmittance from the LiDAR point clouds. First, LiDAR beam trajectories are reconstructed and traced through a voxel grid. Second, the transmittance is calculated for a voxel size of 25-50cm resolution with an efficient implementation of AMAPVox developed in Python. Third, unseen and undersampled voxels are identified via occlusion mapping and the quantification of explored voxel volume to drive uncertainty estimates. 

Across the 24‑month record the resulting transmittance maps display phenological patterns. The dataset, together with the newly created Python implementation for transmittance calculation and tight integration of occlusion mapping, enables quantitative analysis of structural canopy changes and provides a robust framework for linking these changes to eco‑physiological and hydrological variables that were measured concurrently on the ECOSENSE field site.