<![CDATA[Aquaculture plays a vital role in addressing global seafood demands and ensuring food security, particularly in tropical regions like the Sunda Strait, Indonesia. However, aquaculture success depends on key environmental parameters, including sea surface temperature (SST), salinity, ocean heat content, and thermocline depth, which exhibit complex spatiotemporal variability. This study applies a Long Short-Term Memory (LSTM) model to predict aquaculture suitability by analyzing five critical oceanographic parameters: depth of the 26°C isotherm (so26chgt), ocean heat content (sohtc300), mixed layer depth (somxl010), sea surface salinity (sosaline), and sea surface temperature (sosstsst). Using the ORAS5 dataset spanning January 2015 to March 2025, the model achieved high accuracy, with R² scores exceeding 0.89 for all parameters. Spatial prediction maps for November 2024 to March 2025 were generated, highlighting regions with optimal environmental conditions for aquaculture. Results indicate that SST and salinity are the most influential factors affecting aquaculture quality, with favorable conditions predominantly observed from December to May. The findings underscore the potential of deep learning models in supporting sustainable aquaculture management through accurate environmental forecasting.]]>
Copyrights © 2022
