How Air Traffic Controllers Separate Aircraft
Separation is the gap a controller must keep between aircraft they are responsible for. The size of the gap is a published minimum. It depends on whether the unit can see the aircraft on a surveillance display, on the flight level, and on the wake the aircraft produce.
· by Michal Zembol

A minimum, applied to named aircraft
The controller is responsible for the flights on their position. For each pair that could conflict, they need a form of separation that meets the minimum for that airspace: vertical, or horizontal, or a procedural gap measured in time or distance from the pilots' reports. The minimum is the smallest gap the rules allow. Controllers often keep more, because the aircraft are still moving and the next clearance has to be possible. They are not free to use less.
ICAO publishes the procedures many states use as the starting point, in Doc 4444, the Procedures for Air Navigation Services for Air Traffic Management. A state can prescribe a different minimum, and it can file a difference from the ICAO procedure. The number a trainee learns at one unit is the number in that unit's instructions. The figures below are Doc 4444's baseline, so that the words "5 miles" and "1,000 feet" have a source. They are not a claim that every sector on earth uses them today.
Vertical separation
Vertical separation is the simplest to describe and the one passengers can see on a flight-deck display, if they are shown one. Doc 4444 sets a nominal 300 metres, 1,000 feet, below flight level 290, and a nominal 600 metres, 2,000 feet, at and above that level. In designated airspace, where a regional agreement allows it, the 1,000-foot minimum continues up to flight level 410, or higher if the agreement says so, and 2,000 feet applies above that. The guidance document Doc 4444 points to for this second case is Doc 9574, the manual for a 1,000-foot minimum between FL290 and FL410. In operations that arrangement is called RVSM, reduced vertical separation minimum.
An aircraft has to be approved, and the operator has to be approved, to use the reduced spacing. An aircraft that is not approved is kept on the wider minimum, which is why two aircraft at the same place on the map can still be a problem if one of them cannot use the levels in between. The altimeter setting matters too. Below the transition altitude the levels are altitudes, set on local pressure. Above it they are flight levels, set on a standard pressure. The controller assigns the level. The crew flies it.
Horizontal separation on a surveillance display
Where the unit identifies both aircraft with radar, ADS-B or multilateration, Doc 4444 sets the horizontal minimum at 9.3 kilometres, 5.0 nautical miles, unless a more specific rule applies. The appropriate authority may reduce that, and the document sets floors for the reduction. It may go to 5.6 kilometres, 3.0 nautical miles, where the surveillance system at that place can support it. On the same final approach track, within 18.5 kilometres, 10 nautical miles, of the runway threshold, it may go to 4.6 kilometres, 2.5 nautical miles, and only when a list of conditions is met. Those conditions include runway occupancy, the update rate of the surveillance system, and the tower being able to see the runway and its exits. The reduced final-approach figure is a special case with a published procedure. It is not a general permission to run 2.5 miles everywhere.
Both aircraft have to be identified. A return that has not been identified is not an aircraft the controller may separate to the surveillance minimum. How identification is done, and how the same job is done with pilot reports when there is no picture, is in Surveillance and Procedural Control. Procedural minima are written in time and distance from reports. They are larger than 5 nautical miles, because the information is older and coarser. A controller qualified on a surveillance rating is not, by that rating, qualified to apply the procedural ones.
Wake turbulence is extra
Two aircraft can be 5 nautical miles apart and still be too close if the one behind is in the wake of the one ahead. Wake separation depends on the category, or the group, of each aircraft, and on whether they are departing, arriving, or crossing. Doc 4444 contains tables. States also publish their own, including regroupings that replace the older four categories with finer groups for particular aircraft. The table on the sector is the one the authority has prescribed. It can demand a larger gap than the ordinary surveillance minimum, and it can be different on the approach from the gap used once both aircraft are level in the cruise.
A controller applies the stricter of the rules that fit that pair. When the wake minimum is larger than the surveillance minimum, the pair is kept at the wake minimum. The categories are a property of the aircraft type, not a judgement made from the window.
The runway is a different problem
On the aerodrome, separation includes the runway itself. Only the movements that are allowed to be on it at once are cleared onto it. A landing aircraft, a departure, a crossing vehicle and an aircraft backtracking are all traffic for that purpose. The tower may be using a runway-occupancy rule, a wake gap, and a dependent-approach rule at the same time. The approach unit has usually built the landing order before the flight is transferred. The tower still has to withhold the landing clearance if the runway is not available, and send the aircraft around.
Who works the runway, the taxiways and the circuit is covered in aerodrome control. The airspace around the airport, where the arrival sequence is built, is approach control. The en-route sectors, where the vertical and 5-mile surveillance minima are the daily tool, are area control. What the controller's hour consists of, before the numbers, is in What Does an Air Traffic Controller Do?