Atmospheric refraction, the bending of light downwards as it passes through the atmosphere, causes astronomical bodies to appear higher in the sky than they actually are. At high altitudes the effect is slight, but close to low horizons—when light rays have to travel through more of the earth’s atmosphere to reach the observer—the apparent altitude can be as much as half a degree above the actual altitude. Certainly if something appears at the horizon, it is actually below it.
When we are calculating the declination of a celestial body following a field survey, it is necessary to take refraction into account. Because of the extremely high precision of the sightlines he was postulating, Alexander Thom went to pains to estimate atmospheric conditions at the time of supposed observations in order to calculate the appropriate refraction correction. However, as astronomers Bradley Schaefer and William Liller have since pointed out, day-to-day changes in atmospheric conditions effectively rule out sight-lines precise to just a few arc minutes. For most purposes, a mean refraction correction (related only to the horizon altitude concerned) will suffice.
