Key points are not available for this paper at this time.
The most important patterns of normalized sea surface temperature (SST) departures in the tropical Atlantic are identified using principal component analysis. The first three eigenvectors of separate monthly and seasonal analyses over the period 1948 to 1972 produce similar patterns of SST variation. In addition, the first three monthly eigenvectors are shown to occur with equal frequency during the earlier part of this century from 1911 to 1939. The second eigenvectors of monthly and seasonal SST departures are significantly correlated with normalized sea level pressure (SLP) departures (also specified using principal component analysis) over the period 1948 to 1972. On monthly time scales, the strongest relationship is found with the SLP field leading the SST field by one month. The second SST eigenvectors are also shown to be significantly correlated with Sahel rainfall variations over the period 1948 to 1972. The SST-rainfall relationship for this recent period appears to be different from that of the early 20th century, when no significant correlation is found. It is suggested that this finding may result from different causes of wet and dry years in the Sahel during the two time periods considered here.
Janice Lough (Sat,) studied this question.