Site testing is the process used by astronomers in deciding where to build a new ground-based telescope. It characterises a geographic location's dew point, range of wind speed, temperature and humidity variation, and the proportion of nights with cloud cover that it has in a year. The site's darkness, or lack of light pollution, will control the background brightness of the observable sky, so remote locations are normally considered. Particularly important is assessment of the site's astronomical seeing, which is a measure of how much points of light in the sky are enlarged into circles by the turbulence of the atmosphere.

Excellent sites, such as Mauna Kea in Hawaii, Cerro Paranal in Chile, and Roque de los Muchachos on La Palma in the Canary Islands are located on isolated mountain tops, places which combine the factors of laminar airflow, low humidity, great distances from the lights of population centres and sub-arcsecond seeing. Site testing at Dome C on the Antarctic Plateau suggests it has better seeing than any of these mid-latitude sites, but telescopes have not yet been constructed there. Site testing can have different requirements depending on the wavelength of light that the telescope is designed to observe: for example, the Atacama Large Millimeter Array's site in the Atacama Desert in Chile is exceptionally dry. This aids its observation of millimetre and submillimetre radiation, which is otherwise absorbed by water vapor in the atmosphere.

The process of site testing frequently takes several years. Since new telescopes of medium or large aperture cost upward of tens of millions of dollars, a number of sites are normally assessed and several shortlisted before a decision is made. Road accessibility will also affect the cost of construction of the telescope, so site testing is frequently carried out at locations adjacent to existing observatories. Telescope domes must be air-conditioned to keep the telescope as close as possible to the night-time temperature. A site that has little variation from its mean temperature on diurnal and seasonal timescales requires less dome temperature control.

Site testing uses a variety of instruments. Standard would be an automated digital weather station, for recording windspeed, humidity and similar environmental and weather factors, and a webcam for sky images. Measuring turbulence in different levels in the atmosphere can be done in several ways. An acoustic radar can measure the lower several hundred metres of atmosphere, including the boundary layer, which will directly affect the airflow around the telescope's dome and the telescope itself. The atmospheric opacity must also be measured. (need mention of DIMMs, SCIDAR)...

Observations at wavelengths longer than two microns are affected by the surrounding thermal background heat: colder sites therefore have less inherent noise.

[1] [2]