Noer, Kayla Rashieka (2025) Estimasi Suhu Permukaan Perairan di Laguna De Bay, Filipina Serta Hubungannya Dengan Indeks ENSO Menggunakan Model Artificial Neural Networks. Masters thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 6016241016-Master_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
6016241016-Master_Thesis.pdf Restricted to Repository staff only Download (2MB) | Request a copy |
Abstract
Sebagai danau terbesar di Filipina, Laguna de Bay menunjukkan sensitivitas tinggi terhadap perubahan iklim lokal dan global. Suhu permukaan danau ini dipengaruhi oleh faktor internal dan eksternal, termasuk fenomena iklim global seperti El Niño–Southern Oscillation (ENSO). Perubahan suhu akibat ENSO dapat berdampak langsung maupun tidak langsung terhadap stabilitas ekosistem danau maupun kondisi lingkungan di sekitarnya. Oleh karena itu, pemantauan hubungan antara suhu permukaan air danau dan indeks ENSO menjadi penting untuk mengantisipasi potensi dampak yang ditimbulkan. Penelitian ini menggunakan data citra satelit Sentinel-3 yang diolah menggunakan machine learning berbasis ANN melalui platform Google Colab untuk memperoleh estimasi suhu permukaan air di Laguna De Bay. Hasil penelitian menunjukkan bahwa ANN secara konsisten unggul dibandingkan RF, dengan performa terbaik pada band 8 (MAE 0,045°C, RMSE 0,051°C, R 0,970, R² 0,941). Sebaliknya, RF menunjukkan akurasi terendah pada band 7 (MAE 0,174°C, RMSE 0,293°C, R² 0,661). Band 8 menjadi kanal paling optimal secara numerik untuk kedua algoritma, sedangkan band 9, meskipun kurang stabil secara kuantitatif (MAE 0,101°C, RMSE 0,174°C), menunjukkan keunggulan visual signifikan (MAE visual 0,0001°C, RMSE 0,011°C), sehingga sangat efektif untuk mengidentifikasi dinamika spasial suhu mikro dan anomali termal di perairan tropis.Selain itu, analisis klimatologis menunjukkan bahwa pengaruh El Niño–Southern Oscillation (ENSO) terhadap variasi suhu permukaan perairan di Laguna de Bay bersifat lemah, dengan korelasi positif namun rendah (R = 0,34, R² = 0,12). Nilai Oceanic Niño Index (ONI) lokal di danau ini jauh lebih kecil (0,019°C–0,105°C) dibandingkan ONI global (0,71°C–0,81°C), meskipun tetap menunjukkan pola searah selama fase El Niño. Respons yang lemah ini diduga dipengaruhi oleh lokasi geografis Laguna de Bay yang berada di wilayah transisi antara Samudra Pasifik dan Asia Tenggara.
====================================================================================================================================
As the largest lake in the Philippines, Laguna de Bay exhibits a high sensitivity to both local and global climate changes. Its surface water temperature is influenced by a combination of internal dynamics and external drivers, including global climatic phenomena such as the El Niño–Southern Oscillation (ENSO). Temperature fluctuations associated with ENSO can directly or indirectly affect the lake’s ecological stability as well as the surrounding environmental conditions. Therefore, monitoring the relationship between lake surface temperature and the ENSO index is crucial for anticipating potential impacts. This study utilizes Sentinel-3 satellite imagery, processed through an Artificial Neural Network (ANN) model via the Google Colab platform, to estimate the surface temperature of Laguna de Bay. The findings reveal that ANN consistently outperforms the Random Forest (RF) algorithm, achieving its best performance on Band 8 with MAE of 0.045°C, RMSE of 0.051°C, R of 0.970, and R² of 0.941. In contrast, RF recorded its lowest accuracy on Band 7, with MAE of 0.174°C, RMSE of 0.293°C, and R² of only 0.661. Band 8 emerged as the most numerically optimal channel for both algorithms, while Band 9, despite being less stable in quantitative terms (MAE 0.101°C, RMSE 0.174°C), demonstrated significant visual advantages (visual MAE of 0.0001°C, RMSE 0.011°C), making it highly effective for identifying microthermal spatial dynamics and thermal anomalies in tropical waters. Furthermore, climatological analysis indicates that the influence of ENSO on surface temperature variation in Laguna de Bay is weak, with a positive but low correlation (R = 0.34, R² = 0.12). The local Oceanic Niño Index (ONI) ranged from 0.019°C to 0.105°C—substantially lower than the global ONI (0.71°C to 0.81°C)— although both exhibited similar trends during El Niño phases. This weak response is likely due to the lake’s geographic position within the transitional zone between the Pacific Ocean and Southeast Asia.
Actions (login required)
 |
View Item |