<![CDATA[Hidrografi dan batimetri menyediakan informasi kedalaman dan bentuk dasar perairan yang sangat penting bagi navigasi, perencanaan pelabuhan, rekayasa pantai, dan tata ruang laut, terutama pada perairan dinamis dan keruh seperti Kabupaten Cilacap yang berbatasan langsung dengan Samudera Hindia. Penelitian ini bertujuan menganalisis akurasi data batimetri Satellite Derived Bathymetry (SDB) berbasis citra PlanetScope beresolusi tinggi dan membandingkannya dengan hasil pemeruman Single Beam Echosounder System (SBES). Data SBES dikumpulkan menggunakan Garmin 585 C, sedangkan kedalaman SDB diekstraksi dengan algoritma machine learning Generalized Additive Model (GAM). Data dibagi 75% untuk training dan 25% untuk testing, dan kinerja model dinilai menggunakan koefisien determinasi (R²) dan Root Mean Square Error (RMSE). Pengolahan data didukung perangkat lunak SIG dan penginderaan jauh seperti ArcMap, SNAP, ENVI, Global Mapper, RStudio, dan Surfer. Hasil menunjukkan SBES memiliki akurasi tinggi dan tetap menjadi rujukan utama pemetaan dasar perairan Cilacap. Sebaliknya, model SDB–GAM menghasilkan R² sebesar 0,35 dan RMSE 2,80 sehingga belum mampu merepresentasikan variasi kedalaman pada perairan. Nilai ini rendah karena perairan Cilacap yang keruh serta keberadaan sunglint pada citra menyebabkan distorsi reflektansi sehingga kedalaman hasil SDB sulit merepresentasikan variasi kedalaman aktual. Oleh karena itu, SDB – GAM belum direkomendasikan untuk pemetaan operasional di perairan keruh seperti Cilacap tanpa koreksi glint dan peningkatan strategi pemodelan, serta lebih tepat digunakan sebagai pemetaan awal pada kondisi perairan yang lebih jernih dan kedalaman terbatas. Hydrography and bathymetry provide information on water depth and seabed morphology that is crucial for navigation, port planning, coastal engineering, and marine spatial planning, especially in dynamic and turbid waters such as Cilacap Regency, which borders directly with the Indian Ocean. This study aims to analyze the accuracy of bathymetric data Satellite Derived Bathymetry (SDB) derived from high-resolution PlanetScope imagery and compare it with obtained from Single Beam Echosounder System (SBES) sounding. SBES data were collected using a Garmin 585 C, while SDB depths were extracted using the Generalized Additive Model (GAM) machine learning algorithm. Data were split into 75% for training and 25% for testing, and model performance was evaluated using the coefficient of determination (R²) and Root Mean Square Error (RMSE). Data processing was supported by GIS and remote sensing software such as ArcMap, SNAP, ENVI, Global Mapper, RStudio, and Surfer. The results show that SBES provides high accuracy and remains the primary reference for seabed mapping in Cilacap waters. Conversely, the SDB–GAM model produces an R² of 0.35 and an RMSE of 2.80, indicating that it has not yet adequately represented depth variability in waters. This value is low because the murky waters of Cilacap and the presence of sunglint in the image cause reflectance distortion, making it difficult for the SDB results to represent actual depth variatons. Therefore, SDB-GAM is not yet recommended for operational mapping in murky waters such as Cilacap without glint correction and improved modeling strategies, and is more appropriately used for initial mapping in clearer waters and limited depths.]]>
