The Empirical Orthogonal Function (EOF) method, first introduced to geophysics by E. Lorenz [2], has subsequently been used to enable analysis of data with complex spatial/temporal structures. Using this method, the most efficient decomposition of the data into representative modes is determined by empirically finding the eigenfunctions that best describe the information. It can be proven that the EOF method describes the data in the most compact form in a sense to be described below. The EOF eigenmodes can be ordered in terms of the percentage of the total variance to be described by each mode.