Mapping the invasive palm species Arenga obtusifolia using multiple endmember spectral mixture analysis (MESMA) and PRISMA hyperspectral data in Ujung Kulon National Park, Indonesia

• Main Authors: Damarjati, Safri, Nugraha, Wahyu Ananta, Arjasakusuma, Sanjiwana
• Format: Article PeerReviewed
• Language: English
• Published: Taylor and Francis Ltd. 2022
• Subjects: Geography and Environmental Sciences
• Online Access: https://repository.ugm.ac.id/281887/1/Mapping%20the%20invasive%20palm%20species%20Arenga%20obtusifolia%20using%20multiple%20endmember%20spectral%20mixture%20analysis%20%20MESMA%20%20and%20PRISMA%20hyperspectral%20data%20in%20Ujung.pdf; https://repository.ugm.ac.id/281887/; https://www.tandfonline.com/doi/full/10.1080/10106049.2022.2138981
• Institution: Universitas Gadjah Mada

Arenga obtusifolia or “langkap” is an invasive palm that interferes with the growth of native vegetation in Ujung Kulon National Park. This study aims to map A. obtusifolia distribution in Ujung Kulon using PRISMA hyperspectral imagery and MESMA (Multiple Endmember Spectral Mixture Analysis). 38 spectra samples for the A. obtusifolia, sand, wetland vegetation and other vegetation were collected, and optimized using the IES (Iterative Endmember Selection), Multiple Signal Classification (MUSIC) and Automated MUsic and Spectral Separability based Endmember Selection (AMUSES) algorithms. Here, different schemes of MESMA were tested by using spectral weighting (SPEC) and stable zone unmixing (STABLE). Our results demonstrated that MESMA and PRISMA data was able to identify A. obtusifolia with the accuracy ranging from 55.36 (AMUSES-STABLE) to 83.93 (MUSIC-STABLE). Our results indicated the potential of MESMA and PRISMA hyperspectral data using optimized spectra from MUSIC and stable zone unmixing for mapping invasive A. obtusifolia. © 2022 Informa UK Limited, trading as Taylor & Francis Group.