The Chlorophyll-a Modelling over the Andaman Sea using Bi-Directional LSTM Neural Network

The topography of the south of Thailand is the peninsula between the Andaman Sea and the South China Sea. The global measurement of chlorophyll-a is the NASA satellite image product. 16-year monthly time series (from 2003 to 2018) of the Moderate Resolution Imaging Spectroradiometer (MODIS), the data in both space and time. The approximation and component detailed from the original series of MODIS chlorophyll-a were considered sources of chlorophyll-a stationary level anomalous variability. The paper created BiLong Short Memory (Bi-LSTM) modeling, that kind of neural network. Six positions in the province area, Ranong, Phangnga, Phuket, Krabi, Trang, and Satun, were studied. The model was evaluated using four performances as, mean squared error (MSE), root mean square error (RMSE), the sum of square error (SSE), and correlation coefficient. Phangnga was given a good chlorophyll-a prediction. However, the model has been validated to predict chlorophyll-a using Bi-LSTM.