Integrasi α-β Spectral Diversity dengan Linear Spectral Unmixing (LSU) sebagai Proxy Spectral Variation Hypothesis (SVH) pada Ekosistem Mangrove

Ainiyah, Khusniatul (2026) Integrasi α-β Spectral Diversity dengan Linear Spectral Unmixing (LSU) sebagai Proxy Spectral Variation Hypothesis (SVH) pada Ekosistem Mangrove. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6016241006-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Penurunan biodiversitas global menjadi salah satu tantangan lingkungan utama, termasuk pada ekosistem mangrove yang berperan penting dalam penyimpanan karbon dan perlindungan pesisir. Studi ini mengevaluasi penerapan Spectral Variation Hypothesis (SVH) melalui integrasi α-β spectral diversity dan Linear Spectral Unmixing (LSU) untuk merepresentasikan biodiversitas mangrove di pesisir timur Surabaya (Pamurbaya). Menggunakan citra Sentinel-2A, α-β spectral diversity diekstraksi dengan BiodivMapR, sedangkan fraksi spesies dipetakan menggunakan Fully Constrained LSU. Data lapangan digunakan sebagai ground truth untuk menghitung α-β taxonomic diversity, serta memvalidasi hasil melalui confusion matrix, korelasi Pearson, RMSE, dbRDA, Mantel test, dan Procrustes analysis. Hasil penelitian menunjukkan bahwa α-β spectral diversity mampu membedakan komposisi komunitas mangrove secara spasial, di mana kawasan konservasi KRM (Zona 2) yang memiliki komposisi vegetasi lebih heterogen menunjukkan biodiversitas spektral yang lebih tinggi dibandingkan kawasan mangrove alami (Zona 1) yang didominasi oleh A. marina. Analisis LSU mengidentifikasi dominansi A. marina sebesar 53,20%, diikuti R. mucronata (7,17%), R. apiculata (6,94%), S. alba (0,02%), dan kelas Unidentified (32,67%), dengan akurasi keseluruhan sebesar 71,05% yang menunjukkan kemampuan yang baik dalam memetakan komposisi vegetasi pada piksel campuran. Indeks Shannon spektral menjadi prediktor terbaik biodiversitas taksonomi (r = 0,754), sedangkan dbRDA (F = 3,5997; p = 0,001) dan Procrustes (r = 0,4315; p = 0,002) mengonfirmasi bahwa variasi spektral mampu merepresentasikan struktur komunitas mangrove, meskipun hubungan langsung dengan komposisi taksonomi masih terbatas (Mantel r = 0,121; p = 0,077). Temuan ini menegaskan bahwa citra Sentinel-2A berpotensi sebagai proxy biodiversitas mangrove dalam kerangka SVH untuk mendukung pemantauan dan pengelolaan konservasi pesisir.
===============================================================================================================================
Global biodiversity loss has become one of the most pressing environmental challenges, particularly in mangrove ecosystems, which play a vital role in carbon sequestration and coastal protection. This study evaluated the application of the Spectral Variation Hypothesis (SVH) by integrating α-β spectral diversity and Linear Spectral Unmixing (LSU) to represent mangrove biodiversity in the eastern coastal region of Surabaya (Pamurbaya). Using Sentinel-2A imagery, α-β spectral diversity was extracted with BiodivMapR, while species fractions were mapped using Fully Constrained LSU. Field survey data were used as ground truth to calculate α-β taxonomic diversity and to validate the results using a confusion matrix, Pearson correlation, RMSE, distance-based redundancy analysis (dbRDA), Mantel test, and Procrustes analysis. The results showed that α-β spectral diversity effectively distinguished the spatial variation in mangrove community composition, with the Kebun Raya Mangrove (KRM) conservation area (Zone 2), characterized by more heterogeneous vegetation, exhibiting higher spectral biodiversity than the natural mangrove area (Zone 1), which was dominated by A. marina. LSU identified A. marina as the dominant species fraction (53,20%), followed by R. mucronata (7,17%), R. apiculata (6,94%), S. alba (0,02%), and the Unidentified class (32,67%), achieving an overall accuracy of 71,05%, indicating good performance in mapping vegetation composition within mixed pixels. Spectral Shannon diversity was the best predictor of taxonomic biodiversity (r = 0,754), while dbRDA (F = 3,5997, p = 0,001) and Procrustes analysis (r = 0,4315, p = 0,002) confirmed that spectral variation effectively represented the spatial structure of mangrove communities, although its direct relationship with taxonomic composition remained limited (Mantel r = 0,121, p = 0,077). These findings demonstrate that Sentinel-2A imagery has strong potential as a proxy for mangrove biodiversity within the SVH framework, supporting biodiversity monitoring and the conservation management of coastal ecosystems.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	α-β spectral diversity, Biodiversitas, Linear Spectral Unmixing (LSU), Mangrove, Spectral Variation Hypothesis (SVH), α-β spectral diversity, Biodiversity, Linear Spectral Unmixing (LSU), Mangrove, Spectral Variation Hypothesis (SVH)
Subjects:	G Geography. Anthropology. Recreation > G Geography (General) > G70.5.I4 Remote sensing
Divisions:	Faculty of Civil, Planning, and Geo Engineering (CIVPLAN) > Geomatics Engineering > 29101-(S2) Master Thesis
Depositing User:	Khusniatul Ainiyah
Date Deposited:	24 Jul 2026 04:08
Last Modified:	24 Jul 2026 04:08
URI:	http://repository.its.ac.id/id/eprint/136890

Actions (login required)

View Item