Authors: John Gross*, Central Michigan University, Benjamin W. Heumann, Central Michigan University , Rachel A. Hackett, Michigan Natural Features Inventory, Michigan State University Extension , Anna K. Monfils, Central Michigan University
Topics: Remote Sensing, Biogeography
Keywords: UAS, Remote Sensing, Biogeography
Session Type: Guided Poster
Start / End Time: 8:00 AM / 9:40 AM
Room: Roosevelt 3.5, Marriott, Exhibition Level
Presentation File: No File Uploaded
The ability to accurately map and monitor vegetation biodiversity is critical for understanding the spatiotemporal dynamics of vegetation species distributions and abundances, and is paramount for the development and implementation of effective conservation strategies. One promising method for the rapid assessment of biodiversity is the use of remotely sensed data; specifically, the relationship between biodiversity and the diversity in optical types derived using unsupervised clustering algorithms. Previous studies employing this method have derived optical clusters from hyperspectral data which can pose limitations in terms of data availability and cost, and have focused exclusively on forested ecosystems. This research extends those methods by testing the efficacy of optical diversity for modeling alpha and beta biodiversity in herbaceous, highly biodiverse, wetland ecosystems using both simulated data and UAS imagery. Furthermore we directly consider the impacts of changing spectral resolution by simulating hyperspectral, narrowband, and broadband multispectral signatures. Pearson’s correlations showed strong, positive, statistically significant relationships between optical diversity and the alpha and beta biodiversity, and indicate that optical type diversity is not statistically affected by decreasing spectral resolution. These results are particularity important from a management perspective as they demonstrate that assessments of biodiversity can be made using readily available, cost effective multispectral data without sacrificing accuracy.