Air Canada Flight 190 at 36600 experienced severe wake turbulence due to a Boeing 747-400 United 896 Flight at 37000. Assume the worst possible combination of civilian aircraft under standard speeds and air conditions, if they are separated by 1000 can the smaller plane in rear still suffer severe wake turbulence and at what sort of distance separation?
Summary: 1000 ft vertical separation is not enough for wake turbulence from a heavier aircraft flying above to decay. The wake turbulence will however be significantly less strong by then. A lateral distance of 25 NM is required to avoid all wake turbulence.
If both aircraft are level, the situation looks something like this:
En-route, the vortices evolves in altitudes at which the rate of decay leads to a typical persistence of 2-3 minutes (min), with a typical sink rate of about 400 ft/min. Wakes will also be transported by wind.
Considering the high operating air speeds in cruise and the standard 1000 ft vertical separation in Reduced Vertical Separation Minima airspace, wake can be encountered up to 25 nautical miles (NM) behind the generating aeroplane. The most significant encounters are reported within a distance of 15 NM.
If the wake sinks at 400 ft/min, it will reach the lower altitude 2½ minutes later. The typical persistence time is listed as 2-3 minutes, so a strong wake still exists, although at lower intensity.
For a large difference in aircraft weight, this can still result in severe turbulence. An example is the incident involving a Bombardier Challenger CL604 and an Airbus A380 in 2017 (as mentioned by Digital Dracula in the comments):
The analysis of the flight data of both aircraft showed that at 0838:07 UTC the A380 had passed the CL604 overhead with a vertical distance of 1,000 ft. [...]
Within one second the bank angle increased to 42° to the right. At the same time the aileron deflection to the left increased to 20° and the vertical acceleration to 1.6 g. In the following second vertical acceleration changed to -3.2 g. [...]
At the time of the accident [the flight attendant] had been standing in the middle of the cabin preparing the service. Four of the six passengers had also not been seated. In her recollection the airplane had turned three times around its longitudinal axis, during which the occupants had been thrown against the ceiling and the seats. Several of the passengers suffered injuries, some of which were bleeding.
The situation is more dangerous, if one or both of the aircraft are climbing or descending. This can result in a situation, where the aircraft are much closer vertically than 1000 ft:
In the en-route phase of flight, three major factors contribute to increase the likelihood of wake turbulence:
Crossing traffic situation: If crossing traffic is climbing or descending in proximity (either the generating or following aeroplane), the wake generating aeroplane might cross the follower’s trajectory with minimum time for decay, so stronger wake turbulence might be encountered.
Thermal tropopause altitude: Wake vortex decays more slowly below the tropopause, if there is an increased risk of encountering severe wake turbulence.
Weight of the generating aeroplane: Heavier aeroplane types generate stronger wake vortices and are likely to induce more severe wake turbulence encounters, especially for smaller aeroplane types.
(EASA Safety Information Bulletin 2017-10R1, emphasis mine)
This exact situation existed for the Air Canada 190 incident you mentioned:
It was established that the aircraft had been climbing in accordance with its ATC clearance from FL350 to cruise altitude FL370 behind a United Airlines operated Boeing 747-400 en route from Hong Kong to Chicago and level at FL 370 and that the incident had occurred as the A319 had passed FL366 at a distance of 10.7 nm in trail behind the Boeing.
The vertical separation was only 400 ft and at a distance of 10.7 NM the wake turbulence had not decayed much yet.