How moonlight affects deep-sky observing

Moonlight raises the sky background across the whole sky, and faint extended objects are the first casualties. What matters is not just the phase but whether the moon is above your horizon at all.

The standard reference here is Krisciunas and Schaefer's 1991 model of moonlit sky brightness, still the widely used one. It treats the brightness of the night sky as the sum of three contributions: natural airglow, Rayleigh scattering of moonlight by air molecules, and Mie scattering of moonlight by aerosols.

What matters for planning is which variables that model actually takes. It is a function of the moon's phase, the moon's zenith distance, the zenith distance of the patch of sky you are looking at, the angular separation between the two, and the local extinction coefficient. Phase is one term among five.

That has a direct practical consequence. A 60% moon that has already set gives you a clean second half of the night; a 30% moon that rises at 01:00 takes it away. And even with the moon up, angular separation counts — pointing well away from it is materially better than pointing near it. This forecast tracks moon altitude hour by hour for your actual coordinates rather than reporting phase alone, which is why the moon track can read "down" on a night the almanac calls half-lit.

Why deep-sky targets specifically: galaxies and nebulae are low surface brightness objects detected by their contrast against the sky background. Raising that background eats directly into the thing you are trying to see, in a way it never does for a bright planet.

Narrowband imaging is the documented exception. Filters with a passband of roughly 3 to 12 nm transmit the emission line you want while rejecting most of the surrounding continuum, including scattered moonlight, so hydrogen-alpha work remains viable under a moon that would make broadband imaging pointless.