Each eigenvalues is proportional to the percentage of the original data that is accounted for by mode . This percentage of Variance is calculated as : show equation 4
And ,the principal component time series or PC can be derived by projecting the original data series Z^' onto the eigenvectors of matrix E as follows equation 5
Where A is the principal component matrix. showing the time series of coefficient of each EOF mode in the time series of anomalies.