Not all cloud is the same cloud
February was not our month. We sent eleven "go" alerts and members marked five of them as bad nights. Four of those five had the same story: the forecast showed 30–40% cloud, which our old score treated as "partly clear, worth a look", and the cloud turned out to be a thin, even veil of high cirrus over the whole sky.
Under cirrus the Moon and Jupiter look fine. Galaxies vanish. The people who complained were all deep-sky observers, and the people who didn't were mostly out looking at Jupiter.
What changed
The score now reads cloud in three layers instead of one total:
- Low cloud (below about 2 km) is weighted heavily for everyone. If it's there, you're not seeing much.
- Mid cloud is weighted almost as heavily, but it tends to arrive in patches, so we look at how it changes hour to hour rather than at a single value.
- High cloud depends on your profile. For deep sky and Milky Way photography it counts nearly as much as low cloud. For Moon and planets it barely counts at all.
That's why each spot now asks what you mostly observe from it. If you use one spot for everything, pick deep sky; it's the strict setting, and a deep-sky night is a good planetary night too.
Rerunning February
We replayed the whole month with the new weights. Deep-sky alerts went from eleven to six, and five of those six were nights people had logged as good. The planetary profile still gets the cirrus nights, which is right.
If your profile is wrong, the alerts will feel wrong. Check it from the link in any alert email.