Astro
DEENFR

Learn · Method

Why the Night Decides — Air, Light and Sensor Heat

August 30, 2026 · 9 min read

Between a nebula and its image lie a few thousand light-years of undisturbed vacuum — and then the last hundred kilometers, which decide everything. Atmosphere, weather, stray light and finally the heat inside our own camera housing determine what is left of a night. This article is the methodical engine room of the gallery — including a series of measurements that exists in this form nowhere else.

The last hundred kilometers

Cross section from space down into the garden: airglow at 90 kilometers, the jet stream at 10 kilometers, the weather layer with clouds and dew, near-ground air with roof thermals, the light dome of the cities — and the sensor heat inside the camera itself.spaceup here the sky is perfectthe Moon — the brightest“light pollution” of them allairglow · ~90 km — the air itself glows faintly100 km —stratosphere — usually calmjet stream · ~10 km — fast high-altitude flow, smears the stars10 km —weather layer — clouds, humidity, dewnear-ground air — roof thermals, local seeinglight dome of the citiesgarden · 0 m — and the last source of trouble sits in the camera: sensor heat0 m —
Every layer disturbs in its own way: airglow lifts the sky background, the jet stream smears the stars, the weather layer delivers clouds and dew, the near-ground air shimmers above warm roofs — and the light dome fills the sky with stray light. At the very bottom one more source of trouble is waiting, and it is not in the sky at all.

Seeing: the air as poor optics

That stars twinkle is charming and annoying at once: turbulent parcels of air at slightly different temperatures act like weak lenses in constant flux. In a long exposure nothing twinkles any more — instead every star smears into a little disk whose size is measured in arcseconds and called seeing. Good nights at this site stay below one arcsecond, mediocre ones lie well above — and the single worst culprit is the jet stream: once the high-altitude flow sits over the Taunus, not even the clearest sky helps. That is why every plan here begins with the weather and seeing forecast, not with the choice of target.

The source of trouble inside the housing

The most unexpected insight of this gallery does not come from the sky but from our own image database: the image sensor is noisier the warmer it is — and dramatically so. More than four thousand of our own frames with logged sensor temperature yield +4.7 percent noise per kelvin:

Diagram of our own measurement: image noise rises by 4.7 percent per kelvin of sensor temperature. Between 25 and 36 degrees the noise grows by a factor of 1.66 — for the same result the warm night needs 2.7 times as much exposure time.1.5×2.0×2.5×1.0×25 °C → 36 °C: noise ×1.66= expose 2.7× as long for the same image20 °C25 °C30 °C35 °C40 °Cnoise relative to 20 °Csensor temperature — measured on more than 4,000 of our own frames: +4.7% noise per kelvin
The curve translates directly into hours of your life: because exposure time works quadratically against noise, a muggy August night at 36 °C sensor temperature needs 2.7 times as many hours as the same night at 25 °C. A cool, clear autumn night is literally a gift of exposure time — and cooled astro cameras earn their extra cost precisely here.

A few things worth knowing about this:

  • The clear night is the dew night. A front lens radiates its heat away toward the sky and, under a clear sky, cools several degrees below air temperature — which is why it fogs over before the air reaches the dew point. Of all nights it is the most beautiful ones that end in dew most often; only ground wind from about 4 m/s upward mixes the effect away.
  • Clouds are a gate, not a deduction. A night with dream seeing and a closed cloud deck is worthless — cloud cover multiplies night quality, it does not add to it. The rating logic of this gallery treats it exactly that way: only once the gate is open do seeing, humidity and stability count at all.
  • The Moon is a planning quantity, not an excuse. Planned forward, a full moon is a decision — you fall back on narrowband targets or skip the night. Looking backward it is no measure of punishment for a night that one did use after all. Forecast and assessment are two different questions.
  • Against the light dome, physics helps, not defiance. The stray light of the cities is broadband — the nebula lines are not. That is exactly why the narrowband route from the article on colors works under a suburban sky too; only broadband targets such as the galaxies remain dependent on dark, moonless nights.
  • In the end the bookkeeping is what counts. A deep image in this gallery is rarely one night — it is the sum of the best hours out of several. Frames from poor phases are thrown out before stacking; four good hours beat eight mediocre ones.
AI-generated illustration: a compact red apochromat with a red cooled astro camera on a small red harmonic-drive mount and a carbon tripod on a garden lawn, the Milky Way above, the orange light dome on the horizon.

The workplace: compact apochromat, cooled camera, harmonic-drive mount on the lawn — behind it the light dome, above it the Milky Way, and high up the thin streaks of the jet stream. This image is AI-generated, modeled on the real setup of this gallery — not a photograph.

In our own sky

Two images from which the night's bookkeeping can be read:

IC 1318 panorama · from the Butterfly to the Crescent — four degrees of Cygnus in one frame19 objects — show
Four degrees of Cygnus are not the product of a single night: a panorama of this depth collects its hours over many evenings, each with its own seeing, its own Moon and its own sensor temperature. The finished image is the average of the best moments of a whole season.My photograph · IC 1318 Panorama · view gallery entry
The interior of the ring — faint threads between the famous arcs1 object — show
The faint filaments inside the Veil Nebula are the hard test: such structures stand only a little above the sky background — whether they survive in the image is decided by airglow, stray light and noise, not by the optics. Where they are visible, the night was good.My photograph · NGC 6992 · 1.85° × 3.03° · view gallery entry

Further reading

← All articles