From night 2456000 to 2456007 we have excellent data. Here they are (sorry about the sideways plot) for each band and each night. I show the ratio of the DS intensity in EFM-corrected images to the total image flux. Alfa was always near 1.7 so flux is not expected to be lost off the edge of the image. We have scatter at the 1% level and up.
We have a fairly close relation between observed and expected ratio if we chose the simple Schönberg phase law (blue curve), while we seem to be contradicting the more advanced Lommel-Seeliger phase law (red curve). Hmm? Moon and Earth modeled as Lambert spheres in this argument.
The larger scatter on the first and fifth nights in some filters indicate observing problems and not geophysics!
I think issues in the EFM method are: centering of the Moon in order to extract DS intensity; and choosing the pixels to use as source. On the first night problems are due to very low earthshine at this phase.
The phases or Sun-Earth-Moon angles covered by the 7 nights extend from 97 to 34 degrees. Unfortunately skipping the interesting 42-degree angle. Except for the brevity of the observing window on the last night I see no strong indication that we are getting particularly bad data at high airmass – some of these data points are at airmass 3-7!
Now then – the above to be repeated with the BBSO and FFM methods: all for the poster at the EGU.