Application of LSTM and MODIS Satellite Imagery for Forecasting Oceanographic Dynamics and Identifying Potential Fishing Zones in the Sunda Strait

Khalifa, Muta Ali ; Yudono, Muchtar Ali Setyo; Prabowo, Nico Wantona ; Santoso, Prakas; Rachmanto, Farhan ; Prasetia, Aditya Teguh

doi:10.26877/asset.v8i3.2815

📅 21 June 2026

Application of LSTM and MODIS Satellite Imagery for Forecasting Oceanographic Dynamics and Identifying Potential Fishing Zones in the Sunda Strait

Khalifa; Muta Ali; Yudono; Muchtar Ali Setyo; Prabowo; Nico Wantona; Santoso; Prakas; Rachmanto; Farhan; Prasetia; Aditya Teguh

Advance Sustainable Science, Engineering and Technology (ASSET)

Universitas Persatuan Guru Republik Indonesia Semarang

📄 Abstract

This study integrates AQUA-MODIS satellite imagery with the Long Short-Term Memory (LSTM) model to forecast oceanographic dynamics and identify Potential Fishing Zones (PFZ) in the Sunda Strait. The dataset spanning from January 2014 to December 2024 was used for model training, while forecasts for January–August 2025 were validated using in-situ observations from six sampling stations. The model predicted sea surface temperature (SST), chlorophyll-a concentration, and ocean current speed, with SST reaching 31°C, chlorophyll-a at 3.5 mg/L, and peak current speeds of 0.4 m/s. The performance metrics for SST (MSE: 1.107, RMSE: 0.994, MAD: 0.794), chlorophyll-a (MSE: 1.609, RMSE: 1.011, MAD: 0.5739), and current speed (MSE: 0.0183, RMSE: 0.1223, MAD: 0.0959) confirmed model accuracy. The PFZ detection algorithm, based on SST, chlorophyll-a, and ocean current data, demonstrated strong spatial agreement with in-situ data, validated using metrics such as MSE and RMSE. This validation approach, employing direct in-situ comparison, supports effective fisheries management by identifying productive fishing areas under varying seasonal and climate conditions. These results underline the operational potential of the LSTM-based forecasting framework for adaptive fisheries decision-making in the Sunda Strait.

🔖 Keywords

#Deep learning; LSTM; time-series forecasting; PFZ mapping; AQUA-MODIS; in-situ validation

ℹ️ Informasi Publikasi

Tanggal Publikasi

21 June 2026

Volume / Nomor / Tahun

Tahun 2026

📝 HOW TO CITE

Khalifa, Muta Ali ; Yudono, Muchtar Ali Setyo; Prabowo, Nico Wantona ; Santoso, Prakas; Rachmanto, Farhan ; Prasetia, Aditya Teguh , "Application of LSTM and MODIS Satellite Imagery for Forecasting Oceanographic Dynamics and Identifying Potential Fishing Zones in the Sunda Strait," Advance Sustainable Science, Engineering and Technology (ASSET), Jun. 2026.

ACM

ACS

APA

ABNT

Chicago

Harvard

IEEE

MLA

Turabian

Vancouver