Assessing the Accuracy of Spectral Indices for Wetland Habitat Differentiation Using Open-Source Satellite Data

Funding disclosure: Massachusetts Space Grant Consortium (MASGC), Bridgewater State University Bartlett College of Science and Mathematics, Bridgewater State University Office of Undergraduate Research

Abstract

Wetlands are critical ecosystems that support biodiversity, improve water quality, mitigate flooding, and offer a plethora of other ecological services. However, these vital areas are facing extensive degradation worldwide due to human activity and climate change. In response to this degradation, wetland restoration efforts have emerged in an attempt to recover ecosystem functions and services. Monitoring these restoration projects is crucial for understanding the ecological recovery of the ecosystem and the success of restoration. Traditional field surveys, though effective, are often costly, time-intensive, and labor-intensive. Remote sensing provides an efficient alternative, that allows for large-scale, repeatable, and detailed observations of wetland conditions over time. By using satellite-derived indices, remote sensing can reveal valuable information about habitat health, structure, and change with a fraction of the resources required for on-ground surveys. This study aims to evaluate the effectiveness of various spectral indices in distinguishing key wetland habitat types, open water, floodplain, and upland forest. Using open-source satellite imagery, we assess the performance of indices in identifying distinct habitat features. Identifying indices that excel at capturing specific wetland characteristics is expected to improve monitoring capabilities and provide more targeted insights.