The first step is to develop the procedure for a systematic fit of a function that best describes the data for Adelaide – seemingly data of reasonable quality. We want to fit a linear function of the form
Уі — A) + Pixi + £i (8.3)
to the data. This requires that the errors, ei, be independent and identically distributed. This involves transforming the data into a form wherein there is homogeneity of the data as the clearness index varies. As can be seen in Figure 8.1, there is a much greater spread for the diffuse fraction for middle values of the clearness index.