Analysis of the potential of in situ hyperspectral optical data and high spatial resolution satellite imagery to map bottom types in the Abrolhos Coral Reef Bank (Brazil)

• Ano de defesa: 2014
• Autor(a) principal: Maria Laura Zoffoli
• Orientador(a): Milton Kampel, João Antonio Lorenzzetti
• Banca de defesa: Cláudio Clemente Faria Barbosa, Carlos Alessandre Domingos Lentini, Ricardo Coutinho
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Instituto Nacional de Pesquisas Espaciais (INPE)
• Programa de Pós-Graduação: Programa de Pós-Graduação do INPE em Sensoriamento Remoto
• País: BR
• Link de acesso: http://urlib.net/sid.inpe.br/mtc-m21b/2014/10.17.18.30

Human activity and natural climate trends constitute a major threat to coral reefs worldwide. Models predict a significant reduction in reef spatial extension together with a decline in biodiversity in the relatively near future. In this context, monitoring programs to detect changes in reef ecosystems are essential. In recent years, coral reef mapping using remote sensing data has benefited from instruments with better resolutions and computational advances in storage and processing capabilities. However, the water column represents an additional complexity when extracting information from submerged substrates by remote sensing, which demands a correction of its effect. Another limitation arises when focusing in complex benthic substrates, with a mixture of different taxonomic and/or functional types with the aim to detect alterations. In the first part of this work a deep study about water column correction methods was proposed, showing some inter-comparisons of different algorithms and evaluating the uncertainties associated with the determination of bottom reflectance from above water radiometry. An inter-comparison of three methods was carried out using simulated spectra and remote sensing data of the Worldview-2 (WV02) sensor over the Abrolhos Coral Reef Bank (ACRB), Brazil. The results showed that the available methods were not able to completely eliminate the water column effect, but they minimized its influence. Uncertainties in retrieving the bottom reflectance from above-water spectra increase with depths and the concentration of optically water constituents. In clear waters with low colored dissolved organic matter (CDOM) concentration, uncertainties augmented towards the red spectral bands. While in environments with high CDOM concentration, uncertainties increased both towards the blue and the red spectral ends. Choosing the best method depends on the marine environment, available input data and desired outcome for the scientific application. In the second part of this study, hyperspectral above-water radiometry sampled over natural benthic communities in the ACRB were analyzed to: (i) evaluate the ability of a water column algorithm to retrieve accurate bottom reflectance spectra from above-water radiometric measurements; and (ii) assess the separability of submerged benthic communities with subtle differences based on their reflectance spectra. Despite of the good performance to retrieve bottom reflectance from above-water radiometric measurements, benthic communities in the ACRB, whose waters were characterized as Case-2, were not spectrally separable. This implies that effective monitoring programs in the ACRB able to detect insightful alterations in the benthic coverage, necessarily have to include both complementary approaches: remote sensing and field campaigns. In the last part of this work, an area belonging to the ACRB was mapped for the first time into benthic classes using high spatial resolution orbital imagery of the WorldView-2 (WV02) sensor. This sensor has 6 bands in the visible region, but only one band at 478 nm was used because of the low performance to correct the water column effect, which represented one of the main challenges in this work. The produced map had an accuracy of 88.2\% and a Kappa index of 0.81 using four classes (Reef, Sand, Macroalgae, Inter-reef areas), however new validation points are required to validate the entire scene. The isolated reefs covered an area of approximately 22 $km^{2}$ that correspond to 12.4\% of the total area. \emph{Chapeirões} were distributed almost in all the studied area, formed by patches with different sizes. Macroalgae were the most abundant class in the scene and was concentrated around the Archipelago and inter-reef areas. Further works regarding extensive in situ data collection to properly describe the benthic communities and their spectral behavior, as well as the water column constituents and depths surveys are recommended to improve satellite mapping of the bottoms types in the Abrolhos Coral Reef Bank.