Low-frequency modes of variability of the Indian Ocean and their connection with the Indian Ocean Dipole

The thesis investigates the low-frequency variability of the Indian Ocean, based on a number of independent observational and reanalysis data sets. These data sets include altimetry sea level anomaly (SLA) gridded field, subsurface temperature from selected Expendable Bathythermograph (XBT) lines, NOAA Optimum Interpolation Sea Surface Temperature (SST) and Extended SST V2 (Kaplan) data sets, tide gauge and reconstructed sea level data, surface winds from National Center for Environmental Prediction (NCEP) and Dipole Mode Indexes from HadISST, as well as coral core records used as a long term proxy climate signal. The analysis is primarily based on identifying the major temporal scales by means of spectral analysis. It is found that in most regions of the Indian Ocean, the low-frequency variability (corresponding to periods of six months and longer) is concentrated in five spectral bands: semi-annual, annual, 18-20 months, 3 years, and 4 years and longer; at least during the last two decades. For each of the identified low-frequency variability modes, their spatial distributions as well as the temporal behaviour are investigated; the corresponding results are presented by means of spatial maps of power spectral density and movies. Despite the extensive literature on the variability of the Indian Ocean, to the best of our knowledge no systematic analysis of this kind has been conducted so far. While the existence of semi-annual and annual signals is well known, the presence of such a pronounced and well defined 18-month signal has not been previously identified. Its discovery is one of the main results of this study. Both the 18-month and 3-year signals are found to be directly related to the Indian Ocean Dipole (IOD) mode. The IOD events in the last two decades seem to predominantly occur as a result of their constructive interference. The connection established between these two quasi-periodic signals with the IOD allows us to suggest that the IOD phenomenon has oscillatory character. This suggestion opens up possibilities to improve the prediction of IOD, which is of a major importance for the seasonal weather prediction in Australia. In particular, it is shown that the status of the IOD in the last three years could be predicted two seasons ahead based on the estimate of the state of the 18-month signal.