Large scale patterns in sea level

The purpose of Empirical Orthogonal Functions (EOFs) is to find spatial patterns that capture the temporal variability in data. EOFs decompose the signal in the data into different modes, each of which captures a portion of the variability. Below are the first modes of the EOFs of monthly Sea Surface Height (SSH, or sea level) in different regions of the Pacific Ocean.
The time-series show anomalies (deviations from the long-term average). The map shows the dominant spatial pattern. Red colors represent regions where SSH goes up when the anomalies are positive, blue colors represent regions where SSH decreases when the anomalies are positive.

North Pacific Ocean (1st Mode : 12% of variability)

Equatorial Pacific Ocean (1st Mode : 39% of variability)

South Pacific Ocean (1st Mode : 20% of variability)