Implementing drone imagery as a tool for surveying vegetation in Hawaiʻi

Abstract

Hawaiʻi's forests, essential to its ecological balance and cultural heritage, are key to preserving biodiversity and sustaining water resources. Through thoughtful land management, these vital ecosystems continue to thrive, ensuring the enduring resilience of Hawaiʻi's landscapes. Forest monitoring is important in tracking the presence of threats and biological invasions that persist here. Here in Hawaiʻi, the United States Forest Service Forest Inventory and Analysis (FIA) Program completed its second forest inventory assessment between 2019–2022 using traditional on-ground surveys, providing forest estimates for land management. However, some sites were inaccessible due to the physical challenges faced by field crew members, particularly on sheer cliffs deemed “hazardous” and unsafe. This thesis’ analysis spans two methodological chapters, focusing on innovative methodologies and their applications. Chapter 2 first demonstrates the efficiency and consistency of data collected using drones compared to on-ground surveys, highlighting the practicality of modern technologies when conventional methods were not viable. Results show that there was no difference between the two survey methods for percent cover, vegetation structure, and species composition. Building on this foundation, Chapter 3 applied similar data collection strategies to investigate cliff forest ecosystems on the island of Kauaʻi without the need for a methodological comparison. This approach instead facilitated the safe exploration of inaccessible areas, addressing data gaps and enabling the study of percent cover, species composition, diversity, and their relationships to steep terrain. Results found that slope has a positive effect on native species richness as well as having an influence on the likelihood of locating rare and endangered plant species on areas with increased slope. Moisture also had a positive effect, with increased rainfall increasing the total percent cover of vegetation. Substrate type also had a positive effect on species richness, with rock and similar mixed compositions providing stable places for diverse vegetation to flourish. Finally, mesic forests were found to host higher rates of alpha and gamma diversity of plant species for these areas. Altogether, these methods and datasets establish a critical foundation for forest inventory and monitoring, integrating drone technology to support ongoing conservation efforts in Hawaiʻi and advance methodological innovation in ecological research.