The σ0 of the empirical model is 0.0108, which is only twice the spread in the intensity values. The physically induced and the empirical model are compared graphically in Fig. 5. It shows a scatter plot of the observed intensity values against the predicted ones. An ideal model would show a 45° line. Both models do not only show a spread of values, which could be interpreted as random measurement errors, but also systematic deviations from this line.