Spectral and spatial requirements of remote measurements of pelagic Sargassum macroalgae
Remote detection of pelagic Sargassum is often hindered by its spectral similarity to other floating materials and by the inadequate spatial resolution. Using measurements from multi-spectral satellite sensors (Moderate Resolution Imaging Spectroradiometer or MODIS), Landsat, WorldView-2 (or WV-2) a...
Gespeichert in:
Veröffentlicht in: | Remote sensing of environment 2015-09, Vol.167, p.229-246 |
---|---|
Hauptverfasser: | , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Remote detection of pelagic Sargassum is often hindered by its spectral similarity to other floating materials and by the inadequate spatial resolution. Using measurements from multi-spectral satellite sensors (Moderate Resolution Imaging Spectroradiometer or MODIS), Landsat, WorldView-2 (or WV-2) as well as hyperspectral sensors (Hyperspectral Imager for the Coastal Ocean or HICO, Airborne Visible-InfraRed Imaging Spectrometer or AVIRIS) and airborne digital photos, we analyze and compare their ability (in terms of spectral and spatial resolutions) to detect Sargassum and to differentiate it from other floating materials such as Trichodesmium, Syringodium, Ulva, garbage, and emulsified oil. Field measurements suggest that Sargassum has a distinctive reflectance curvature of ~630nm due to its chlorophyll c pigments, which provides a unique spectral signature when combined with the reflectance ratio between brown (~650nm) and green (~555nm) wavelengths. For a 10-nm resolution sensor on the hyperspectral HyspIRI mission currently being planned by NASA, a stepwise rule to examine several indexes established from 6 bands (centered at 555, 605, 625, 645, 685, 755nm) is shown to be effective to unambiguously differentiate Sargassum from all other floating materials Numerical simulations using spectral endmembers and noise in the satellite-derived reflectance suggest that spectral discrimination is degraded when a pixel is mixed between Sargassum and water. A minimum of 20–30% Sargassum coverage within a pixel is required to retain such ability, while the partial coverage can be as low as 1–2% when detecting floating materials without spectral discrimination. With its expected signal-to-noise ratios (SNRs~200:1), the hyperspectral HyspIRI mission may provide a compromise between spatial resolution and spatial coverage to improve our capacity to detect, discriminate, and quantify Sargassum.
•Spectral bands and step-wise rules to fingerprint Sargassum determined•Multi-resolution sensors compared in their ability to detect and quantify Sargassum•Detection limit due to spatial resolution quantified through simulations•Potential of the HyspIRI mission in detecting and quantifying Sargassum discussed |
---|---|
ISSN: | 0034-4257 1879-0704 |
DOI: | 10.1016/j.rse.2015.05.022 |