Key points are not available for this paper at this time.
Principal component analysis of tropical surface data over the period 1951–1960 suggests the Southern Oscillation can be represented by the first components of pressure, temperature, and rainfall. The patterns obtained are generally consistent with previous definitions of the oscillation but show another center of action over northeast Brazil in phase with the Indian Ocean. This center is associated with a modulated Walker cell over Brazil and the Atlantic which appears to interact with the sea surface in a similar way to that which Bjerknes (1969) has described for the Pacific.
John W. Kidson (Sat,) studied this question.