Predictability of sea surface temperature anomalies at the eastern pole of the Indian Ocean Dipole—using a convolutional neural network model

Abstract

In this study, we train a convolutional neural network (CNN) model using a selection of Coupled Model Intercomparison Project (CMIP) phase 5 and 6 models to investigate the predictability of the sea surface temperature (SST) variability off the Sumatra-Java coast in the tropical southeast Indian Ocean, the eastern pole of the Indian Ocean Dipole (IOD). Results show that the CNN model can beat the persistence of the interannual SST variability, such that the eastern IOD (EIOD) SST variability can be forecast up to 6 months in advance. Visualizing the CNN model using a gradient weighted class activation map shows that the strong positive IOD events (cold EIOD SST anomalies) can stem from different processes: internal Indian Ocean dynamics were associated with the 1994 positive IOD, teleconnection from the equatorial Pacific was important in 1997, and cooling off the Australian coast in the southeast Indian Ocean contributed to the 2019 positive IOD. The CNN model overcomes the winter prediction barrier of the IOD, to a large extent due to the frequent transition from a warm state of the Indian Ocean to a negative IOD condition (warm EIOD SST anomalies) over the boreal winter to the following spring period. The forecasting skills of the CNN model are on par with predictions from a coupled seasonal forecasting model (ACCESS-S2), even outperforming this dynamic model in seasons leading to the IOD peaks. The ability of the CNN model to identify key dynamic drivers of the EIOD SST variability suggests that the CMIP models can capture the internal Indian Ocean variability and its teleconnection with the Pacific climate variability.

Harnessing the Power of Graph Representation in Climate Forecasting: Predicting Global Monthly Mean Sea Surface Temperatures and Anomalies

“…Ratnam et al (2020) proposed an FCN to forecast SSTAs and the IOD oscillation over the Indian Ocean. Also, the IOD forecasts have been made using an LSTM (Pravallika et al, 2022) and a CNN (Feng et al, 2022). Moreover, a CNN was developed to forecast SSTs and marine heatwaves around Australia (Boschetti et al, 2022).…”

Deep learning for skillful long-lead ENSO forecasts

“…The heatmaps are extracted from the best ensemble member among the top ten. These heatmaps have been proposed by Selvaraju et al (2020) and were further found useful for estimating the relative contribution of the predictors in a few recent past studies (Liu et al, 2021;Feng et al, 2022). Equation (3) elaborates on the estimation of activation maps for specific convolutional layers (Krizhevsky et al, 2012) and Equation ( 4) elaborates on the extraction of gradients from a specific layer (Selvaraju et al, 2020).…”

“…Therefore, in the current study, we propose a forecasting scheme for ENSO at very long lead times (up to 3 years) based on CNN by additionally taking into account the varying parameters of CNN models for each season and training them with a customized loss function which considers extreme ENSO events separately. This is a kind of a novel attempt compared to previous studies of deep learning applied to climate predictions (Ham et al, 2019(Ham et al, , 2021bGeng and Wang, 2021;Hu et al, 2021;Liu et al, 2021;Mu et al, 2021;Feng et al, 2022). We train the CNN models using the past monthly global sea-surface temperature anomalies (SSTA) and vertically averaged subsurface ocean temperature anomaly (VATA) (averaged over 0-300 m depth, a proxy for heat content) fields.…”

Predicting extreme floods and droughts in East Africa with deep learning

“…The larger the gradients from a specific region the stronger the control it has over the variability 32 . The gradient heatmaps 32 are different from that used in several past studies 33,34 where the activation values are multiplied by gradients to produce heatmaps, however such heatmaps are prone to contamination by large predictor values and thus misleading the importance of specific regions 32 . Eq.…”