S. Purnawan, H. Manik, L. Elson, F. Syahputra, M.R.F. Hidayat
Benthic habitat classification is a key component for managing coastal resources and conserving marine ecosystems. The waters surrounding Lancang Island, part of the Thousand Islands of Jakarta, are characterized by sandbanks and complex reef zones, making them an ideal location for studying the distribution of seabed substrates. This study employed an unsupervised learning approach to classify seabed substrates using single-beam echosounder (SBES) data. Acoustic features, such as Attack, Decay, and BP1, were analyzed using K-means to cluster unlabeled observations, providing an alternative to the supervised methods used in previous research. The analysis identified six dominant substrate classes with spatial distributions that corresponded well to geomorphological structures and depth variations. These findings demonstrate the effectiveness of an unsupervised method for benthic habitat mapping in regions with limited labeled data. Importantly, the generated habitat maps of the Lancang area provide critical information for sustainable coastal resource management and marine ecosystem conservation, supporting decision-making to promote the health and productivity of marine environments. These findings also contribute to SDG 14 by supporting the sustainable management of marine ecosystems. © Published under licence by IOP Publishing Ltd.
Department of Marine Sciences, Universitas Syiah Kuala, Banda Aceh, Indonesia; Department of Marine Science and Technology, IPB University, Bogor, Indonesia; Department of Fisheries Resources Utilization, Abulyatama University, Aceh Besar, Indonesia; Marine Technology Study Program, IPB University, Bogor, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock