Dome seeing, in its general meaning of seeing generated by turbulent fluctuations of the index of refraction inside a telescope enclosure, can be distinguished into different phenomena:
From the engineering standpoint, there are two different avenues to prevent the occurrence of dome seeing in a telescope enclosure. One way requires a very tight thermal control of the telescope air volume. Chilling the floor of the telescope volume below the exterior air temperature, with the establishment of a stable stratification, does prevent the occurrence of dome seeing. In the CFHT observatory, where such a cooling system is installed, no seeing related to convection flow inside the dome can be detected, except for mirror seeing due to free convection on the surface of the primary. And even that phenomenon penalizes the telescope only for a small fraction of the observing time, because the chilled floor keeps also the temperature of the primary mirror always at a close range with the dome's ambient air.
The second possibility consists of planning the natural ventilation of the telescope volume by the wind. The NTT, typical example of this configuration, often achieves virtually zero local seeing in this manner. However, when the ventilation is reduced by closing the back louvers a significant seeing reappears. Also, when the mirror is warmer than ambient air, mirror seeing appears present in the NTT.
Thus in conclusion, both the general evidence and the engineering experience point to mirror seeing as the main obstacle toward the removal of local seeing effects in modern observatories. This aspect will therefore be studied in more detail in the next section.