返回
IOD-related predictability barriers induced by initial errors in the tropical Indian Ocean in a CGCM
冯蓉
中国科学院大气物理研究所LASG
Using a Geophysical Fluid Dynamics Laboratory Climate Model version 2p1 coupled model, the effects of initial sea temperature errors on the predictability of the Indian Ocean Dipole (IOD) are explored. When initial temperature errors are superimposed on the tropical Indian Ocean, a winter predictability barrier (WPB) and a summer predictability barrier (SPB) exist in IOD predictions. The existence of the WPB has a close relation with El Ni?o–Southern Oscillation (ENSO) in the winter of the growing phase of positive IOD events. That is, when ENSO exists in winter, no WPB appears in IOD predictions, and vice versa. In contrast, there is no inherent connection between the existence of the SPB and ENSO. The WPB is closely related with the initial temperature errors in the tropical Indian Ocean, where two types of WPB-related initial errors display opposite patterns and a west–east dipole. The dominant spatial pattern of SPB-related initial errors is similar to that of WPB-related initial errors. Furthermore, the potential observing locations (i.e., the sensitive areas) of IOD events to advance beyond the predictability barriers are explored. The results show that the areas where the large values of the dipole pattern initial errors are located have great effects on prediction uncertainties and provide useful information regarding the sensitive areas. Further, the prediction uncertainties are more sensitive to the initial errors in the subsurface large value areas than to those in the surface large value areas. The results indicate that the subsurface large value areas are sensitive areas for advancing beyond the WPB and SPB of IOD predictions and if we carry out intensive observations across these areas, the prediction errors in winter and summer may be largely reduced. This will lead to large improvements in the skill of IOD event forecasts.