In general, when using corr_dims, the correlation coefficients are calculated using only the first index along all dimensions other than corr_dims.
corr_dims should only be used when the error correlation matrices do not vary along the other dimension(s).
A warning is raised if any element of the correlation matrix varies by more than 0.05 between using only the first index along
all dimensions other than corr_dims or using only the last index along all dimensions other than corr_dims.
However it is also possible to instead average over the other dimensions (instead of taking first index). An option could be added to do this.
In general, when using corr_dims, the correlation coefficients are calculated using only the first index along all dimensions other than
corr_dims.corr_dimsshould only be used when the error correlation matrices do not vary along the other dimension(s).A warning is raised if any element of the correlation matrix varies by more than 0.05 between using only the first index along
all dimensions other than
corr_dimsor using only the last index along all dimensions other thancorr_dims.However it is also possible to instead average over the other dimensions (instead of taking first index). An option could be added to do this.