Biases in the Tropical Indian Ocean subsurface temperature variability in a coupled model

Abstract

In this study, the subsurface temperature variability in the Tropical Indian Ocean (TIO) is examined in the Climate Forecast System version 2 (CFSv2) coupled model. The observations and reanalysis show a north–south dominant mode of variability in the TIO subsurface temperature during September–November, the season when the dominant east–west surface mode (Indian Ocean Dipole; IOD) peaks. The nature of the north–south dipole in TIO subsurface temperature is successfully captured by CFSv2. The observations however indicate that this subsurface mode, in general, persists for the next two seasons with stronger signals during December–February, whereas such tenacity is not seen in the model, instead rapid decay of the mode is seen in the model. It is found that the misrepresentation of both equatorial surface wind anomalies and associated Ekman transport as well as the Ekman pumping in the model have close association with the early weakening of the mode in CFSv2. The surface easterlies are generally modulated by the presence of twin anticyclones on both sides of the equator. The model captured these anticyclones with weaker than observed intensity and the northern anticyclone is confined over much smaller region than observed. Association of the subsurface mode with El Nino Southern Oscillation (ENSO) and IOD is further examined in this study. The anomalously prolonged decay phase of El Nino in CFSv2 is found only during the El Nino, IOD co-occurrence years, which was not reported before. This paves way for addressing an important modeling issue which is common in many coupled climate models including CFSv2. The analysis suggests the possible role of coupled air–sea interaction over the TIO on the El Nino cycle in the Pacific. It is also found that the misrepresentation of subsurface variability in CFSv2 during December–February is closely associated with the rapid decay of El Nino forced TIO warming.