I ran through ewok_1004 and compared the current sky brightness values to the ones calculated from the ESO model.
It looks like ESO is predicting significantly fainter sky values. I think the outliers in the lower rights are caused by OpSim using a simple twilight model, where it just sets the sky to a bright value if there is any twilight. There were 19 points that the model couldn’t compute values for because they were above an airmass of 2.5.
Results running on 1/100 of the opsim pointings:
filter, median OpSim, median ESO
u, 22.04, 22.69
g, 21.75, 22.09
r, 21.05, 20.99
i, 19.83, 20.02
z, 17.61, 18.67
y, 17.30, 17.68