Sensitive area in the tropical Indian Ocean for advancing beyond the summer predictability barrier of Indian Ocean Dipole

Using the geophysical fluid dynamics laboratory climate model version 2p1 (GFDL CM2p1), perfect model predictability experiments have been conducted to identify the sensitive area in the tropical Indian Ocean for advancing beyond the summer predictability barrier (SPB) of positive Indian Ocean Dipole (IOD) events. In these experiments, the model is assumed to be perfect, and prediction errors are only caused by initial errors. Initially, the impact of initial error patterns on prediction uncertainties was assessed by comparing dipole pattern initial errors with three sets of spatially correlated noises. The results revealed that dipole pattern initial errors tend to result in larger prediction errors and higher error growth rates in summer, leading to a significant SPB phenomenon. Notably, the large values of these dipole pattern initial errors are concentrated in specific areas. By eliminating initial errors within these areas, the prediction errors in summer are largely reduced, underscoring the sensitivity of prediction uncertainties in summer to initial errors in these areas. Moreover, the prediction errors in summer exhibit a higher sensitivity to initial errors within the subsurface large value area compared to those within the surface large value area. Consequently, the subsurface large value area in the tropical Indian Ocean is the sensitive area for advancing beyond the SPB, aligning with the corresponding location for advancing beyond the WPB. Eliminating initial errors within this area leads to a rapid decrease in prediction uncertainties, with a more pronounced reduction in winter than in summer. Through intensive observations in this sensitive area, significant reductions in prediction errors in both summer and winter can be achieved, thereby greatly improve the forecast skill of IOD events.