The key wavelength for moonlight is 395nm, half-way between the U and B filters. The critical spectral type for this wavelength is G2V, where a 395nm - V colour is 0.8 magnitudes. The brightness of the night sky can come from several papers, e.g. Krisciunas and Schaefer (1991) , Sanchez (2008) , Gemini's Website or NOAO's website. The most conservative number comes from Gemini, with an estimate of 16 magnitudes per square arcsec for a wavelength between U and V. The NOAO website appears too optimistic, with ~17.5 magnitudes per square arcsec. If we take the Sanchez value for extinction at 395nm of 0.4 magnitudes at zenith and approximate scattering as isotropic, we arrive at 17.4 magnitudes per arcsec. It therefore seems that with the following assumptions:
  • The air-mass limit is 1.5 when observing at full moon (less than 5 days from full moon) and 2.0 otherwise.
  • Observations are never made within 30 degrees of the moon, where the sky becomes significantly brighter,
we can assume 17 magnitudes per square arcsec, and a key stellar (395nm) - V color of 0.8. For starlight to be twice as bright as the sky, the limiting bright-time magnitude becomes V=13.2.