Digital Data in Biodiversity Research Conference

In the face of increasing ecological threats posed by invasive species, accurate detection and modeling are essential for effective management and mitigation. This presentation details a groundbreaking study focused on the invasion of non-native tree species in the Greater Cape Floristic Region(GCFR) of South Africa, a biodiversity hotspot, where these species threaten the native fynbos vegetation. Traditionally, hyperspectral techniques have been used in local spatial settings (Durán et al., 2019; Schneider et al., 2017) so this is not only exploring the questions of how we can model invasion using hyperspectral data and multiscale imagery at a large scale, but also how effective it is at predicting low density invasion when using validated, ground truth data? Utilizing an integrated remote sensing approach, this research leverages data from the Sentinel Satellite, alongside advanced airborne imaging sensors mounted on NASA’S G3 and G5 research aircraft. Notably, this project marks NASA's inaugural Biodiversity field campaign that integrates airborne imaging spectroscopy, lidar, and field observations, setting a new standard for ecological monitoring and biodiversity conservation using remote sensing technologies.

The study employs innovative analytical techniques, including band dimensionality reduction and principal component analysis to enhance the understanding of spectral signatures that differentiate invasive from native plant species. The spectral data is used to calculate three diversity metrics: alpha, beta, and gamma diversity to understand diversity in and between communities. By analyzing these diversity metrics, the research identifies specific areas most at risk and provides crucial data to inform conservation strategies.


Co-authors: Peter Adler, Alanna Rebello, Elisa Van Cleemput, Katie Suding, Ben Poulter, Ayia Lindquist, Meghan Hayden, Cibele Hummel do Amaral, Laura Dee