The papers in which the tables were presented have discussed the accuracy of the data, however some comments are warranted in this context.
While every effort has been made within the context of the model to obtain cross-sections of high computational accuracy, the physical accuracy of the data is difficult to judge. As laboratory data for line broadening by hydrogen are almost non-existent, the only means of testing the data is by application to the solar spectrum.
The most that can be said about the accuracy of the data computed using this method, is that within the accuracy of the f-values and synthesis method used, the data (i.e. the tables) have been shown in previous papers to be consistent with the observed broadening in the solar spectrum if one makes the assumption that the solar abundances are equal to the meteoritic ones. We are not aware of any discrepancies. Anstee et al. ([1997]), in their paper examining the solar Fe abundance using strong lines, did report a discrepancy between the observed broadening and the theoretical broadening for transitions with an upper e5D state. However this discrepancy has since been found to be attributable to strong Stark broadening of this energy level. The previous applications of the data to strong lines of a number of elements with well determined f-values shows a consistency of derived abundances with meteoritic values that would indicate that the uncertainty may well be as low as 5%.
Perhaps the most likely source of gross error in the list would be mistaken identification of transition types (i.e. angular momentum quantum numbers l), and incorrect identification of the parent configuration, in the automated processing of the electron configurations. However, comparison with the manually produced list, of which a reasonable fraction of the lines involve double electron excitations, with lines computed automatically revealed no discrepancies.
In any case, all data used in the calculation of the line broadening data, such as angular momentum quantum numbers, energy levels, and series limits, are provided in Table 1 (although this information is not in VALD). Hence, if one suspects an error, the input data can be checked.
Copyright The European Southern Observatory (ESO)