I found this, which looks to have just been published this year. Thinking about it more since commenting, I did remember one of the owls, I want to say the Dusky Eagle Owl, only nests in swan nests in one particular species of trees, or something along those lines as well.

Roost Selection and Behavior of the Long-eared Owl (asio Otus) Wintering in New Jersey

ABSTRACT - Roosting Long-eared Owls (Asiotus) selected conifers with dense foliage that concealed all or most of the main trunk with no apparent regard to tree species. Roosts were established only in clumps of 2 or more closely-spaced conifers (3-15 m in height), always near a variety of open habitats. Communal roosts of 2-4 owls were significantly preferred to solitary roosts. Strong fidelity for a single roosting tree was observed within each winter, although the owls shifted to a new main roost site each yr. Owls concealed themselves indense foliage; when approached, they would hide or freeze and flush only at close distances. Evidence indicated that these owls had habituated to remarkably close human activity, although they were readily able to detect an intruder. The 2 most frequented roosts were within 8 m of large buildings which may have provided wind protection and increased shade for hiding. The owls remained at roosts well into darkness and when flushed during the day, showed strong aversion to daylight activity.

Roosting Behavior. – In most cases owls concealed themselves completely in a dense portion of the conifer and were not visible until flushed or an intention movement was made. Consequently, it was not always possible to accurately note information such as roosting height, distance from trunk, or individual distances. On one occasion, a Long-eared Owl was perched on a completely exposed branch, but when I approached, it hopped along several branches and hid behind the tree trunk while keeping a continuous watch in my direction. Similar hiding behaviors were observed on 3 other roost counts. On 2 other occasions, owls were seen trying to avoid detection by elongating their posture, erecting their ear tufts and closing their eyelids nearly completely. The resultant motionless form was maintained unless I approached closer than 3-4 m; then the owls flushed. This concealing posture was identical to the “broken branch” appearance described in Bent (1938:163). Another time, I observed a person, unaware of the owl roost, walk within 4 m of an owl that was roosting on an exposed branch, but the owl remained undisturbed. However, when I approached within 9 m and looked directly at the same owl, it began staring intently, quickly rotated its head back and forth, and then flushed. These contrasting observation suggesthat Long-eared Owls will habituate to nearby human traffic, but are readily able to discriminate when they are being watched. Such selective attention to a predator’s eyes (in this case, the author’s) can have considerable survival value (Suarez and Gallup 1983) in that prey species can monitor the direction of a predator’s visual focus and may be able to take advantage of better escape opportunities (Gagliardi et al. 1976). Randie and Austing (1952) reported a similar ability of Long-eared Owls to discern scattered members of a searching party and change the direction of their escape flight accordingly.

Also came across this study on GHOs changing roosting height to avoid biting flies. So it seems there is as much going into the thought about where they live as we put into where we choose to spend our time.