Pipeline for integrating lidar data with structured demographic models for ecological inference. (a) Forest populations are surveyed at appropriate time intervals with a lidar scanner to produce high-resolution 3D point clouds. Point clouds are processed into canopy height models (CHMs) and digital terrain models (DTMs). (b) Individual tree segmentation is performed using a CHM-based algorithm (e.g. Dalponte & Coomes, 2016). (c) Relevant environmental variables are derived. For example, TWI is an indicator of soil moisture accumulation (Kopecký et al., 2021) and is calculated from the DTM. Local competition is inferred from the canopy density surrounding each tree, using the CHM. (d) Trees are matched across time points using a crown-matching algorithm (Battison et al., 2024; Olsoy et al., 2024), before measuring changes in size and classifying mortalities. (e) Continuous traits (e.g. height and crown area) are evaluated for their ability to predict survival and growth across the life cycle. (f) Selected traits and environmental parameters are used to model the vital rates using regressions. (g) The integral projection model (IPM) is constructed based on the selected vital rate models and key life history traits (e.g. mean life expectancy, longevity, generation time) are calculated. (h) Finally, IPM outputs are analysed to test ecological hypotheses.
The first paper from my PhD is now out in the world 🌲📖 Modelling forest dynamics using IPMs and repeat lidar doi.org/10.1002/rse2...
