When this parameter is set to YES then weights are modified to include uncertainty in sky fit optimal variances equivalent to maximising:
sum over I [W(I)**2 * V(I)] + sum over I,J [W(I)*W(J)*V(I,J)]
------------------------------------------------------------ ( sum W(I) * F(I) )**2
where:
W(I) is the weight on pixel I
V(I) is the variance on the value of pixel I
V(I,J) is the covariance of the skyfit between pixels I and J
F(I) is the normalised profile at pixel I
This reduces to solving:
A(I,J) W(J) = F(I)
where:
A(I,J) = V(I,J) I not equal to J and
A(I,I) = V(I) + V(I,J)
This option is slow as it is computationally intensive.
Martin Clayton