Índice de vegetação por sensoriamento remoto para identificação de espécies de mangue no Nordeste brasileiro.

Abstract

This thesis aimed to develop and validate a new vegetation index for identifying and differentiating mangrove species on the northeastern coast of Brazil, integrating leaf anatomical and spectral attributes in an innovative approach. The research was structured into four interconnected and complementary chapters. The first chapter presents a bibliometric analysis of the use of vegetation indices in mangrove monitoring. The review highlighted the relevance of vegetation indices in remote sensing and their role in mapping and monitoring the dynamics of these ecosystems, but also highlighted their limitations in specific species identification due to the high spectral similarity between them. The second chapter addressed the leaf anatomy of the three main mangrove species in the region – Avicennia schaueriana, Laguncularia racemosa, and Rhizophora mangle – along the Gramame River estuary (Paraíba). Intra- and interspecific variation was observed in leaf tissues, demonstrating structural plasticity associated with species functional adjustment processes. These results provided essential information for anatomical differentiation and subsequent integration with spectral data. The third chapter integrated field spectroradiometry measurements and images from the Landsat 9 and Sentinel-2 satellites, comparing spectral bands and traditional vegetation indices (NDVI and SAVI) for their ability to distinguish species. Statistical analyses indicated that the near infrared was the most discriminating spectral band, and that SAVI outperformed NDVI, reducing ground interference in spectral responses. Landsat 9 stood out for its better radiometric resolution and temporal consistency. The fourth chapter proposed the Mangrove Anatomical Spectral Index (MASI), developed from the integration of anatomical and spectral data, with a 95% confidence level. MASI was applied to the Landsat and Sentinel series and tested in four mangroves: Gramame and Camaratuba (Paraíba), and Itapessoca and Timbó (Pernambuco). Because it was specifically calibrated for mangrove vegetation, MASI should be used exclusively in this type of vegetation, a necessary condition to ensure anatomical and spectral coherence. The index demonstrated high accuracy in separating species, reflecting zonation patterns and identifying local ecological changes. MASI represents a methodological advancement for the monitoring and conservation of Brazilian mangroves.