Advancing Animal Ecology through LiDAR-Based 3D Visibility Analysis

Xin Zong*

*Corresponding author for this work

Research output: ThesisPhD Thesis - Research UT, graduation UT

90 Downloads (Pure)

Abstract

The visual perception of an animal plays a variety of roles in its life such as habitat selection, food detection, species recognition, predation avoidance, and reproduction, as it provides an animal with information that enables it to distinguish an object from its background and to locate a detected target in the environment with a high accuracy. The transmission of visual information is particularly sensitive to the spatial arrangement of physical elements in the environment because the sender and receiver of visual cues need to be directly spatially linked. Visibility refers to the habitat property that enables animals to access visual information. The visibility of habitats to animals profoundly influences their behavior and ecology. Thus, quantifying and considering environmental visibility would greatly enhance our understanding of animal behavior.
Animal behavior occurs at multiple spatial scales. For a wide range of animal behaviors such as escaping from predation and determining movement path, the choice of visibility is made by individuals at fine scales. To understand the impacts of visibility on animal behavior occurring at fine scales, highly spatially resolved visibility information and a high spatial resolution (spatial grain) are required. In forest landscapes, visual obstruction at close ranges is frequently dominated by the refined three-dimensional (3D) plant structure. However, existing conventional methods for visibility estimation are labor-intensive or unable to account for 3D plant structure. Therefore, there is a need for a more effective approach for fine-scale visibility estimation that can characterize the 3D environmental structure effectively and efficiently, particularly the 3D vegetation structure in a forested environment.
This thesis aims to measure fine-scale visibility (i.e., 3D viewsheds) at the plot and landscape level using the LiDAR technique and explore the application of LiDAR-based fine-scale visibility in the understanding of animal spatial behavior across landscapes, with red deer as an example. The research in the thesis firstly presented a novel method to evaluate 3D viewsheds in forests using terrestrial LiDAR (TLS) data. The results showed that TLS can serve as an appropriate tool to rapidly estimate fine-scale visibility in forests. Secondly, the thesis examined the performance of forest structure metrics derived from airborne LiDAR (ALS) data for the prediction of 3D cumulative viewsheds derived from TLS at the ground level. 3D cumulative viewsheds below the canopy could be contiguously measured by integrating terrestrial and airborne LiDAR. To demonstrate the application of the concept of 3D cumulative viewshed in the study of animal spatial behavior at the landscape level, the thesis consequently employed LiDAR-based fine-scale visibility to investigate how visibility in forests influences the habitat selection and movement rate of red deer in the Bavarian Forest National Park, Germany, coupled with GPS locational data. Red deer were found to select intermediate habitat visibility and move faster in high-visibility areas across the landscape. Then, the thesis moved on to examining how visibility affects the behavioral response of the deer exposed to four types of human disturbance: recreational activities, forest roads, hiking trails, and hunting. In areas with lower visibility, red deer tended to tolerate a higher intensity of human recreational activities and use areas closer to forest roads or hiking trails.
This thesis showed that 3D viewsheds can effectively characterize the obstruction effects of the refined 3D environmental structure on the sightlines of individual animals. In forest landscapes, below-canopy 3D viewsheds of animals can be accurately estimated using LiDAR data at the plot and landscape level. The successful application of LiDAR-derived 3D viewsheds in the understanding of habitat selection and movement behavior of red deer highlights its advantages compared to the conventional methods for visibility estimation. The concept of 3D viewsheds and the LiDAR-based methods for the estimation of 3D viewsheds developed in this thesis holds the potential to inform a broader range of questions in the fields of animal ecology and behavior.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • University of Twente
  • Faculty of Geo-Information Science and Earth Observation
Supervisors/Advisors
  • Skidmore, Andrew , Supervisor
  • Wang, Tiejun, Co-Supervisor
Award date10 Jul 2023
Place of PublicationEnschede
Publisher
Print ISBNs978-90-365-5660-6
Electronic ISBNs978-90-365-5661-3
DOIs
Publication statusPublished - 10 Jul 2023

Fingerprint

Dive into the research topics of 'Advancing Animal Ecology through LiDAR-Based 3D Visibility Analysis'. Together they form a unique fingerprint.

Cite this