Modeling the sea-surface pCO2 of the central Bay of Bengal region using machine learning algorithms

The present study explores the capabilities of advanced machine learning algorithms in predicting sea-surface pCO2 (partial pressure carbon dioxide) open oceans Bay Bengal (BoB). We collect available observations (outside EEZ (Exclusive Economic Zone)) from cruise tracks and mooring stations. Due to paucity data BoB, we attempt predict based on Sea Surface Temperature (SST) Salinity (SSS). Comparing MLR, ANN, XGBoost algorithm against a common dataset reveals that performs best for BoB. Using satellite-derived SST SSS, using model compare same with in-situ observations. satisfactorily, having correlation 0.75 RMSE ±12.23?atm. Further this model, emulate monthly variations central BoB between 2010–2019. satellite data, show is warming at rate 0.0175 °C per year, whereas SSS decreases -0.0088 PSU year. modeled shows declination ?0.4852 ?atm perform sensitivity experiments find contribute ? 41% 37% declining trends last decade. Seasonal analysis pre-monsoon season has highest decrease pCO2.