Prediction of Sea Level with Vertical Land Movement Correction Using Deep Learning

Sea level rise (SLR) in small island countries such as Kiribati and Tuvalu have been a significant issue for decades. There is an urgent need more accurate reliable scientific information regarding SLR its trend informed decision making. This study uses the tide gauge (TG) dataset obtained from locations Betio, Funafuti, with sea corrections vertical land movement (VLM) at these data by Global Navigation Satellite System (GNSS) before predictions. The oceanic feature inputs of water temperature, barometric pressure, wind speed, gust, direction, air three lags are considered this modeling. A new decomposition method, namely, successive variational mode (SVMD), employed to extract intrinsic modes each that processed selection Boruta random optimizer (BRO). develops deep learning model, stacked bidirectional long short-term memory (BiLSTM), make (target variable) predictions benchmarked other AI models adaptive boosting regressor (AdaBoost), support vector regression (SVR), multilinear (MLR). With comprehensive evaluation performance metrics, BiLSTM attains superior results 0.994207, 0.994079, 0.988219, 0.899868 correlation coefficient, Wilmott’s Index, Nash–Sutcliffe Legates–McCabe respectively, Kiribati, values 0.996806, 0.996272, 0.992316, 0.919732 case Tuvalu. It also shows lowest error metrics prediction both locations. Finally, analysis linear projection provided GNSS-VLM-corrected average period 2001 2040. rate 2.1 mm/yr 3.9 estimated will 80 mm 150 mm, year 2040 if current trend.