Delineating Pitch Pine in the Long Island Pine Barrens Using PlanetScope Dove Imagery and MaxEnt

Keywords: Pine Barrens, MaxEnt, Pitch Pine, PlanetScope
Abstract Type: Virtual Poster Abstract

Abstract

Pine Barrens are a critically endangered ecosystem endemic to the Northeastern United States which have undergone rapid decline in recent decades. The Long Island region of New York currently contains approximately 41,000 ha of Pine Barrens (~3% of the total US Pine Barren habitat). Pine Barrens are largely dominated and classified by the existence of Pitch Pine (Pinus rigida), making this species an ideal target for conservation efforts aiming to conserve Pine Barrens more broadly. This research aims to demonstrate a novel remote sensing technique for identifying and cataloging Pitch Pine locations. Specifically, it introduces the application of maximum entropy (MaxEnt) modeling onto the normalized difference vegetation index (NDVI), time series data developed from high spatiotemporal satellite imagery (PlanetScope Dove). MaxEnt is advantageous as it only requires a limited sample of presence locations to be effective. Proof of concept models performed reasonably well with limited data (overall accuracy 95%, Cohen's Kappa 73%) and highlight the potential of this methodology. Currently, the model is restricted to approximately 700 sqkm of eastern Long Island. Future efforts will expand this study area.