A lot of the assumptions of nuclear winter came from the if we had a nuclear exchange in the middle 1980's, the cities in the Northern Hemisphere would burn non-stop for essentially weeks since all firefighting efforts would effectively cease to exist.
The Defense Atomic Support Agency, or maybe the same agency with their earlier name, did not build a whole town but did build many structures, including simulations of a typical suburban street, along with a small "forest" of trees to study the effects of a nuclear detonation. Although you did include the term "perfect simulation" the tests performed were very good at allowing us to estimate the effects.
I do not want to parse words but these tests were above ground, not atmospheric. While we did do atmospheric testing most shots were performed either as ground bursts, air bursts, or above ground bursts.
Ground bursts are designated when the fireball touches the ground, air bursts are when the fireball does not touch the ground, and surface bursts are when the device detonates while itself is on the ground. At a certain altitude (I am too lazy to find my books to give the exact reference) it becomes a high altitude burst.