[A-TT30] Machine Learning Techniques in Weather, Climate, Ocean, Hydrology and Disease Predictions

[ATT30-P05] Skillful prediction of the Indian Ocean Dipole using machine learning techniques

*Venkata Ratnam Jayanthi¹, Swadhin Behera¹, Masami Nonaka¹, Kalpesh Ravindra Patil¹ (1.Application Laboratory, JAMSTEC)

Keywords:Artificial neural network, Deep learning

In this study, we used several machine learning models to predict the Indian Ocean Dipole (IOD) index at long leads of 8-10 months. The input attributes were derived based on a lag correlation analysis between the observed IOD index and sea surface temperature, sea surface height, vertically averaged (0-100m) salinity, and soil moisture. Of all the models, we found the artificial neural network model to perform better with a correlation coefficient of 0.7 at an 8-month lead time. The results also demonstrate that the simple machine learning model’s performance is comparable to that of deep learning models in predicting IOD at long leads.