Pemetaan Vegetasi Mangrove Menggunakan Data Satelit Sentinel-2A dengan Algoritma MVI dan NDVI di Pesisir Cirebon

Abstract

Kondisi ekosistem mangrove yang telah menurun di Desa Mundupesisir, Kecamatan Mundu, Kabupaten Cirebon, Jawa Barat akibat dari kegiatan antropologis. Hal tersebut memunculkan urgensi pemantauan ekosistem mangrove agar kerusakan tidak memburuk. Penelitian ini bertujuan untuk memetakan sebaran, kerapatan, dan luasan mangrove di Desa Mundupesisir menggunakan citra Sentinel-2A. Terdapat 2 algoritma yang digunakan, yaitu algoritma Mangrove Vegetation Index (MVI) dan Normalized Difference Vegetation Index (NDVI). Algoritma MVI digunakan untuk membedakan vegetasi mangrove dan bukan mangrove dengan nilai ambang batas 2.374–19.409. Algoritma NDVI digunakan untuk memetakan sebaran kerapatan mangrove. Klasifikasi menggunakan metode maximum likelihood menghasilkan 3 kelas kerapatan mangrove dengan total luas 23.07 hektar. Kelas kerapatan tersebut terdiri dari kelas rapat (21.31 ha), sedang (0.67 ha), dan jarang (1.07 ha). Uji akurasi menggunakan metode confusion matrix terhadap kerapatan mangrove yang menghasilkan nilai akurasi keseluruhan sebesar 94.52% dengan nilai koefisien kappa sebesar 0.97. Uji korelasi statistik dari persentase tutupan kanopi mangrove antara nilai NDVI dan data lapang menghasilkan nilai koefisien korelasi (r) yaitu 0.882 serta nilai koefisien determinasi (R2) yaitu 0.778.

The condition of the mangrove ecosystem had declined in Mundupesisir Village, Mundu District, Cirebon Regency, West Java due to anthropological activities. This raised the urgency of monitoring the mangrove ecosystem so that damage did not worsen. This research aimed to map the distribution, density, and extent of mangroves in Mundupesisir Village using Sentinel-2A imagery. There were 2 algorithms used, namely the Mangrove Vegetation Index (MVI) algorithm and the Normalized Difference Vegetation Index (NDVI). The MVI algorithm was used to differentiate mangrove and non-mangrove vegetation with a threshold value of 2.374–19.409. The NDVI algorithm was used to map the distribution of mangrove density. Classification using the maximum likelihood method produced 3 classes of mangrove density with a total area of 23.07 hectares. These density classes consisted of dense (21.31 ha), medium (0.67 ha), and sparse (1.07 ha). The accuracy test using the confusion matrix method for mangrove density resulted in an overall accuracy value of 94.52% with a kappa coefficient value of 0.97. The statistical correlation test of the percentage of mangrove canopy cover between NDVI values and field data produced a correlation coefficient (r) value of 0.882 and a coefficient of determination (R2) value of 0.778.