Climatic Evolution Characteristics Of The Principal Modes Of The Sea Surface Temperature And Its Coupling With Net Sea Surface Heat Flux In The Tropical Indian Ocean

Based on the sea surface temperature and net sea surface heat flux data fromSODA (Simple Ocean Data Assimilation) and ECMWF (European Center forMedium-range Weather Forecast), changes of the spatial structure of the tropicalIndian Ocean (TIO) sea surface temperature anomaly (SSTA) is studied. Comparedthe spatial structure changes of the principal modes of the TIO SSTA and its couplingwith net sea surface heat flux anomalies pre-and post-1976/1977’s climate shift, theclimatic evolution characteristics of the principal modes of the TIO SSTA and itscoupling with net sea surface heat flux anomalies are studied taking moving mannerof the traditional EOF (Empirical Orthogonal Function) decomposition and SVD(Singular Value Decomposition) method, respectively. Not only the ideas that thespatial structure of one field and the relationship between two ones may changesignificantly are verified during the study, but also the time and place are detected.The main results are as follows:TIO SSTA show a basin range warming trend from the almost unanimouslynegative anomalies evolving into relatively obvious positive ones, and the lineartrends of the SSTA show certain periodicity in different basins. Western equatorialTIO is the region with large possibility of abrupt change occurring.There are notable evolution characteristics of structural features of the TIO SSTAEOF’s main mode, in particular the anomalies change abruptly between two adjacentfields of significantly wider range. Largest variation occurs in the south of20°S TIOand the basin does not always show the same sign. The spatial pattern of the IndianOcean dipole mode (IODM) is not always the well-know distribution of the east-westdipole, but has an evolution from south-north dipole to an east-west one, and the mostobvious variations locate in the southeast and western equatorial IO where ocean dynamics plays an important role.The method of moving EOF can effectively detect significant changes of the TIOSSTA, including the spatial structure. There are three notable variations in the mainmode occurring in1974,1976, and1998respectively, with obvious signals occurringin the eastern IO south of20°S, TIO warm pool and coastal seas of Sumatra-Java.For the second mode, there are four significant changes in1976(overlap with themain mode),1987,1994and2005, the first two are in the western and southeasternTIO while the last two are in the western Australia and costal seas of Sumatra-Java.The first two modes of moving EOF can fit the notable changes of the structuralfeatures of the TIO SSTA in evolution. There is a complementary trend between thevariance rate of the second mode and that of the first mode. Near1976/1977, variancerates of the two modes is the closest and the cumulative rate become minimum.In the first coupling mode from moving SVD, net sea surface heat flux anomaliescan well explain the basin wide changes of the SSTA itself. Affected by it, SSTA varylargest in the coastal seas of Sumatra-Java and south of20°S TIO. The SSTA fieldforms a dipole mode in most time in the second coupling mode with local sitesvarying significantly, and can also well explain its dipole-type changes itself inevolution. Relationship of the each mode has a turning change near1976/1977. Thereis a complementary trend between the variance rate of each field explained by thesecond mode and that of the first one in moving SVD, and also between two fields ineach mode.Moving SVD of SSTA and net sea surface heat flux anomalies can detectsignificant changes of coupling structural features: in the first mode, spatial patternsof the two fields both have only one evident quasi-decadal variation occurring in1971with different locations. In the second mode, obvious variations appear in1988and1994for two fields. Besides, the SSTA field and net sea surface heat flux anomalyfield shows significant change in1964,1997and1971, respectively.Moving SVD of the time rate of SSTA and the net sea surface heat fluxanomalies indicates that: the spatial pattern of the first and second mode shows thebasin and dipole-type distribution respectively, and the correlation coefficient between two fields are high for each mode. There are complementary trends between the twomodes, for both the squared covariance fraction and the variance rate of time rate ofSSTA, and decadal variations occur near1976/1977and early1990s.