Improving the 3D representation of plant architecture and parameterization efficiency of functional-structural tree models using terrestrial LiDAR data.

Functional-structural plant (FSP) models are useful tools for understanding plant functioning and how plants react to their environment. Developing tree FSP models is data-intensive and measuring tree architecture using conventional measurement tools is a laborious process. Light detection and ranging (LiDAR) could be an alternative nondestructive method to obtain structural information about tree architecture. This research investigated how terrestrial LiDAR (TLS)-derived tree traits could be used in the design and parameterization of tree FSP models. A systematic literature search was performed to create an overview of tree parameters needed for FSP model development. The resulting structural parameters were compared to LiDAR literature to get an overview of the possibilities and limitations. Furthermore, a tropical tree and Scots pine FSP model were selected and parametrized with TLS-derived parameters. Quantitative structural models were used to derive the parameters and a total of 37 TLS-scanned tropical trees and 10 Scots pines were included in the analysis. Ninety papers on FSP tree models were screened and eight papers fulfilled all the selection criteria. From these papers, 50 structural parameters used for FSP model development were identified, from which 28 parameters were found to be derivable from LiDAR. The TLS-derived parameters were compared to measurements, and the accuracy was variable. It was found that branch angle could be used as model input, but internode length was unsuitable. Outputs of the FSP models with TLS-derived branch angle differed from the FSP model outcomes with default branch angle. Results showed that it is possible to use TLS for FSP model inputs, although with caution as this has implications for the model variable outputs. In the future, LiDAR could help improve efficiency in building new FSP models, increase the accuracy of existing models, add metrics for optimization, and open new possibilities to explore previously unobtainable plant traits.