Spatial-Temporal Variability of North Pacific Sea Surface Temperature Anomaly Patterns, 1947-1983.

General Circulation Models have been used in the diagnostic mode by several investigators to assess the importance of the underlying initial boundary condition under the hypothesis that North Pacific sea surface temperature anomalies are an important link in the teleconnective specification of downstream atmospheric weather patterns. The objectives of this study are (1) to specify the North Pacific sea surface temperature anomaly patterns over the study period 1947-1983; (2) to determine quantitatively the geographic distribution of the spatial-temporal variability associated with each type of anomaly pattern; and (3) to assess the significance of the resultant type patterns in terms of cloud amount anomaly changes at selected geographic areas. An objective classification of seasonally-composited gridpoint values of sea surface temperature anomalies was performed using linear correlation to identify 18 unique patterns in the North Pacific. The threshold correlation coefficient of 0.45 was used to define the lower limit of type-member similarity. It was determined to be statistically significant. The first five types were found to have comprised more than 62% of the total number of seasonal anomaly composite maps. Of the 147 maps typed, 18.4% were unclassified due to various factors, among them being abnormal downstream advective displacements of boundary currents and strong convergence zones. Resultant preceding/succeeding type sequential analysis led to a dichotomous sorting routine that identified two super type patterns. The cold-pool pattern, centered at 40N, 165W, has a frequency of appearance twice that of the warm-pool pattern (43N, 175E). The chronological sequence of the seasonal anomaly composite maps correlated to the corresponding warm/cold super type was subjected to a runs test for dichotomized data. The observed number of runs was significantly deviant from the expected value and the hypothesis of random arrangement is rejected at the 5% level. Contemporaneous and lagged values of sea surface temperature anomaly and cloud amount anomaly for cold-pool patterns were found to be correlated at an average of 0.51, while the results for the warm-pool case were inconclusive.