Uncovering plant diversity and community structure of dense-canopy and species-rich tropical forests using airborne LiDAR data

Abstract

Tropical forests are widely regarded as the Earth’s most important ecosystems, yet they are severely threatened by anthropogenic disturbances, including over-exploitation and introduction of invasive species. To formulate effective tropical forest conservation and restoration measures, rapid and extensive monitoring of forest structure and biodiversity is essential. However, forest inventory has conventionally relied on field measurements, which are highly labor-intensive and time-consuming, especially concerning the complex canopy architecture and exceptional plant species diversity in tropical forests. Recently, the technological advancements and improved accessibility to airborne light detection and ranging (LiDAR) have enabled efficient monitoring of three-dimensional canopy structure. Harnessing airborne LiDAR data, this thesis aims to investigate multifaceted components of biodiversity in two distinct tropical forest ecosystems in Hong Kong, including the taxonomic diversity of a regenerating secondary forest and the functional diversity of a mangrove forest threatened by biological invasions.

Utilizing a dense-canopy and species-rich secondary forest as a testbed, I proposed and demonstrated a novel community-based method to elucidate the relationship between taxonomic diversity and canopy structural complexity. While traditional individual tree-based and area-based species diversity estimation approaches face substantial challenges in tropical forests, particularly the difficulties associated with segmenting individual trees and identifying understory species, I delved into examining the association between canopy structural metrics derived from LiDAR data and biodiversity indices of plant communities delineated using the watershed segmentation algorithm. By employing machine learning-based random forest regression analysis, I demonstrated the capacity to precisely predict species richness and the Shannon-Wiener index using LiDAR-derived canopy structural metrics. The findings revealed that species richness is predominantly determined by the presence of multi-layered canopy structures. On the other hand, the Shannon-Wiener index, a measure of species evenness, is influenced by both the complexity of canopy layers and canopy shapes.

Then, I investigated the impacts of biological invasions on the functional diversity and ecosystem functioning of a native mangrove forest ecosystem. Leveraging airborne LiDAR data, I retrieved various plant functional traits, including foliage height diversity and leaf area index. Based on multiple functional traits, I estimated grid-based functional diversity indices of the mangrove ecosystem using the multi-dimensional Rao’s quadratic entropy. Various vegetation indices were also derived from a WorldView-2 optical satellite imagery, which served as proxies for ecosystem functioning. My results revealed a contrasting relationship between the functional diversity and ecosystem functioning of native and exotic mangrove species. The exotic species exhibited higher levels of functional diversity, but lower levels of ecosystem functioning compared to the native species, and vice versa.

Overall, this thesis aimed to unleash and demonstrate the immense potential of airborne LiDAR data in exploring various components of biodiversity of dense-canopy and species-rich tropical forests in a spatially explicit manner. By providing insightful ecological and geospatial evidence of tropical forest structure and biodiversity, this thesis can guide effective conservation and restoration practices of the precious tropical forest ecosystems.