Low, mid and high cloud, and why we split them
A single "60% cloud" figure hides the thing you need to know. Sixty per cent low cloud is a wall. Sixty per cent high cirrus is a veil you might still image through, and it will quietly ruin your data while looking fine to your eye.
Low cloud sits below roughly 2 km. It is opaque, it has hard edges, and it either blocks your target or it does not. It is also the layer most likely to burn off after sunset as the ground cools, which is why an overcast afternoon does not always mean a lost night.
Mid-level cloud, from about 2 to 6 km, is the awkward middle. Altocumulus breaks into patches and drifts. It gives you gaps, and whether those gaps are usable depends on how long they last relative to your exposure length.
High cloud is ice crystal cirrus above about 6 km. This is the layer that catches people out. You can look up, see stars, and conclude the sky is clear. Meanwhile the cirrus is scattering enough light to lift your background, halo every bright star, and flatten the contrast in every sub-exposure you take. Visual observers often barely notice it. Imagers lose the session.
That asymmetry is the whole reason the forecast breaks the layers out, and why the wide-field and long-exposure modes in the window finder penalise high cloud far more heavily than the deep-sky visual mode does.
REVIEWED 2026-08-12 · SOURCES: MET Norway, sun and moon positions computed on our own servers