Using a deep-learning approach to infer and forecast the Indonesian Throughflow transport from sea surface height

The Indonesian Throughflow (ITF) connects the tropical Pacific and Indian Oceans and is critical to the regional and global climate systems. Previous research indicates that the Indo-Pacific pressure gradient is a major driver of the ITF, implying the possibility of forecasting ITF transport by the sea surface height (SSH) of the Indo-Pacific Ocean. Here we used a deep-learning approach with the convolutional neural network (CNN) model to reproduce ITF transport. The CNN model was trained with a random selection of the Coupled Model Intercomparison Project Phase 6 (CMIP6) simulations and verified with residual components of the CMIP6 simulations. A test of the training results showed that the CNN model with SSH is able to reproduce approximately 90% of the total variance of ITF transport. The CNN model with CMIP6 was then transformed to the Simple Ocean Data Assimilation (SODA) dataset and this transformed model reproduced approximately 80% of the total variance of ITF transport in the SODA. A time series of ITF transport, verified by Monitoring the ITF (MITF) and International Nusantara Stratification and Transport (INSTANT) measurements of ITF, was then produced by the model using satellite observations from 1993 to 2021. We discovered that the CNN model can make a valid prediction with a lead time of 7 months, implying that the ITF transport can be predicted using the deep-learning approach with SSH data.

Automated summary

The Indonesian Throughflow (ITF) is a crucial link between the tropical Pacific and Indian Oceans, and understanding its transport is important for climate systems. In this study, a deep-learning approach using a convolutional neural network model was used to reproduce the ITF transport based on sea surface height data. The model showed high accuracy in predicting ITF transport, indicating the potential of using deep-learning methods for forecasting.