RECONSTRUCTION AND FORECASTING SEA SURFACE TEMPERATURE IN KARIMATA STRAIT AND NATUNA SEA USING MACHINE LEARNING
Reconstruction and forecasting of sea surface temperatures in the Karimata Strait and the Natuna Sea were carried out using machine learning. The data used to evaluate and predict sea surface temperature is Aqua-MODIS data in the period January 2007 - June 2019 with a spatial resolution of 4 km. Dat...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/44101 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Reconstruction and forecasting of sea surface temperatures in the Karimata Strait and the Natuna Sea were carried out using machine learning. The data used to evaluate and predict sea surface temperature is Aqua-MODIS data in the period January 2007 - June 2019 with a spatial resolution of 4 km. Data is divided into 10 years (2007-2016), 5 years (2012-2016) and 3 years (2014-2016) input data. The data received is divided into three, namely; Normal Year (2017), El Nino Year (July 2018 – June 2019), and La Nino Year (July 2017 – June 2018). This research was conducted at several points as a stationary station which was divided into two groups, namely; offshore station and station near the beach. Reconstruction is good for 10-year data at offshore stations with an average RMSE of 0.91 oC, an average error of 2.42%, and an average correlation coefficient of 0.795. Overall, the prediction results for 2020 have the same pattern as the Normal Year because the data provided have more data patterns in normal conditions. |
|---|