Notes. As is explained in Section 3, we fitted two different trends to the geometric albedo data: a constant and a linear trend, considering uncertainties in both the albedo and free parameter. We calculated the BIC difference between these two models for each free parameters and for each bandpass dataset. Typically, a ΔBIC < 10 is considered not significant. In general, we find no significant trends. There is a slight preference for a linear model across the different variables; however, since the ΔBIC is mostly below 10, this is not significant. If the model cannot choose between a constant or a linear trend, then we conclude that there is little evidence to suggest that the data shows a correlation. The only combination where there is a ΔBIC > 10 is the CKC albedos as a function of log(g). Physically, this could be interesting; however, looking at the data, it is clear that the trend is mostly driven by the few outer points at the low and high end of the log(g) space. At this point, more precise data across the full range of the free parameter is needed to confirm this correlation.