Separate PCAs of the correlation matrices for the Ithaca and Canandaigua data in Table A.1 (after square-root transformation of the precipitation data) yields    with corresponding eigenvalues Ith =...

Separate PCAs of the correlation matrices for the Ithaca and Canandaigua data in Table A.1 (after square-root transformation of the precipitation data) yields

 with corresponding eigenvalues

_Ith
= [1.883, 0.927, 0.190]T and

_Can
= [1.904, 0.925, 0.171]T . Given also the cross-correlations for these data

compute the CCA after truncation to the two leading principal components for each of the locations (and notice that computational simplifications follow from using the principal components), by

a. Computing [SC], where c is the (4
 1) vector [u_Ith, u_Can]^T, and then

b. Finding the canonical vectors and canonical correlations.