This is what happens when two planes don't necessarily get too close together, but could do. I sort of learnt what it meant the hard way, but fortunately not the hardest way.
This conversation recently took place between myself and one of my colleagues. Names have been changed to protect the people who probably don't want their names using. Pretty much everything else is made up too.
INT: An Internet cafe at the COLLEGE OF AIR TRAFFIC CONTROL. DAVID is sitting at one of the computers, browsing some kind of writer death cult website. JAMES walks into the room.
JAMES: "Alright mate."
D: "How was your sim run this morning?"
J: "Mostly alright, few problems here and there. Instructor wasn't very kind about it"
D: "An unkind instructor? No. Anything bad?" </morbid fascination>
J: "Uh, I forgot to tell one of my inbounds to establish on the localiser so he flew straight through it...
J:"...and I got a technical loss of separation near the end."
D: (eyes narrow) "A what?"
J: (smiling) "Technical loss of separation."
D: "At least you didn't send any outside controlled airspace this time. What is it?"
J: "I had a Speedbird going head-on with a Lufthansa—
D: "God, they keep using that sodding Lufthansa. It just won't leave us alone..."
J: (grins) "Yeah. Anyway, they were head-on and I had the Speedbird climb through the Lufthansa. They were like seventy miles apart but [the instructor] still chewed my ear off for it."
I finished this conversation with one of my instructors recently, as I finished what's called a 'summative' in the air traffic simulators I've been on for the past few weeks. A summative is an exam. On a simulator. On a normal simulator 'run' (what's called a 'formative' assessment) you have an instructor sitting beside you, filling out a report form as you go, pointing out your mistakes and occasionally, very occasionally, saying:
...when you (probably accidentally) do something that's expeditious as well as safe. On a summative assessment they are still there but silent, filling out the ominously pink assessment form as you try not to short-circuit the telecomms equipment with the sweat glistening on your fingertips. You work unprompted (and uncorrected), then afterwards the assessor discusses the run with you, asking why you did things the way you did.
Then they tell you all your mistakes.
At the end of my (so far, only) summative run Ian, my instructor and assessor, asked me how I thought I'd done.
"Okay I think," I reply. "I felt like I was on top of it. A few niggles here and there like usual but overall okay."
"I could have given a couple more range checks on the inbounds, I had a couple of wrong level instructions but I fixed them straight away," I continue. "That one that took off I told to climb too high but I fixed that in the same transmission. I didn't move my strips into height order straight away a couple times. That last one coming in to land, I don't know what happened there. I definitely told him to report established on the localiser and he still flew straight through it."
"Anything else?" Ian says, nodding some more.
I cringe inwardly, knowing this almost certainly means I made some huge mistake somewhere and didn't even notice.
"I can't think of anything else."
"Okay." Ian picks up the flight progress strips for a couple of aircraft that have left my screen by that point. "Let's look at these two Babies." Believe it or not, 'Baby' is the ATC callsign for BMI aircraft. "You had this one coming in from the East and this one coming in from the West..."
"...now, the eastbound is at flight level two-nine-zero and the westbound wanted a climb to three-two-zero," he continues. "How many feet is that?"
I look at the 'level box' on the strip where I've written what levels I cleared that aircraft to climb to. "Well, it was at [about 13,000ft] so nineteen thousand feet.
"Right. Nineteen thousand feet, in a BAe 146." What's minimum radar separation?"
"Five miles or a thousand feet [vertically]."
He nods again. "So, eighty miles apart, he's got seventeen thousand feet to climb before he's safely above the other, and is probably climbing less than one thousand feet per minute. Their closing speed is over eight hundred knots, which is how many miles a minute?"
"Yeah," he continues. "You think he's going to make it? Can you prove separation?"
Sighing at myself and shaking my head I reply: "No."
"No. Of course you can't. Now, you did eventually pick up that he wasn't going to make it and stopped the climb one thousand feet below the westbound before it became a problem, which was fine.."
"But I shouldn't have cleared him through the level of the other in the first place," I say.
"No. What would have happened if the radar had failed after you issued the initial climb? You wouldn't have known he wasn't going to make it and he would've lumbered straight through the level of the other, and who knows how close they would've been."
"Okay," I say, somewhat chastened. "They were on opposite sides of the screen though, I thought the climb rate was better than that so he would easily make it. I did consider whether he'd make it but I thought he would."
"But were you sure?"
"I suppose I wasn't," I finally admit.
"No. Whenever there is any doubt, there is no doubt (okay, he didn't really say that, but I really wanted to use that line). You can't climb an aircraft through the level of another it's head-to-head with, even if they're on opposite sides of the screen. It's a technical loss of separation; they could lose separation without further intervention, and you can't work it like that.
"You've got to be cautious and assume they won't make it; climb the lower aircraft so it's a thousand feet below the higher one, maybe get him to report when he's two thousand feet below and see where they are when he gets there. If you can see he would make it you can get him to continue the climb, but if you don't say anything they won't hit each other. If you can see he won't make it, wait until they've passed each other or lock them on headings that won't intersect then finish the climb."
I hope this all makes sense. To quickly recap, the minimum separation for en-route aircraft under radar control in the UK is five miles horizontally or 1,000ft vertically, notwithstanding reduced separation standards permitted at certain airports (Heathrow airport, for example, was recently permitted to reduce separation for aircraft on final approach from three miles to 2.5). These standards have to be maintained for all aircraft inside controlled airspace: you have to have one (horizontal) or the other (vertical). So if you want to climb an aircraft through the level of another, they've got to be at least five miles away when there is less than 1,000ft between their levels. Conversely, if aircraft are less than five miles apart horizontally they must be at least 1,000ft apart vertically.
If you've got aircraft on intersecting tracks (either in opposite directions, converging at an angle or one overtaking another, it doesn't matter) at different levels and the lower aircraft wants to be higher than the other, you've got several choices to meet that request, depending on the aircraft (bear in mind these are given from my current, limited perspective, where an Air Traffic Controller's purview does not include speed control of the aircraft in their airspace).
First, consider where the aircraft are and where they are pointing. If they're going in opposite directions, head-on, the closing speed is the highest and there's little room for guessing, so you're the most cautious. If they're both tracking in the same direction, more or less the same speed, then you've got more options. If one is going significantly faster than the other, you've got more still.
You also have to take aircraft performance into account and the stage of flight they're in. An L1011 that has just taken off, sagging with kerosene and pax is going to climb like a pregnant cow with pages of the Financial Times for wings. I swear the Airbus A340 can only climb because the Earth is a sphere. A Boeing 767 on the end of a transatlantic flight, on the other hand, could comfortably manage 3,000ft/min.
Weather makes a difference. Wind changes at different altitudes. A headwind for one aircraft could work to your advantage... or not, since an aircraft heading directly towards it at a similar level is going to have a tailwind that cancels out any difference. Temperature and altitude also affects climb rate (air that is less dense does not support an aircraft so well).
You still have options if it looks like you can't make the climb in time.
You could temporarily turn the two aircraft so they're no longer pointing at each other, make the climb then put both aircraft back on their original headings when 1,000ft separation has been regained. If one is behind the other and you suspect that a small speed difference may cause you to lose separation you could give one aircraft a dog-leg turn so it has to traverse more distance than the other, to cancel out the speed difference or at least to reduce its effect.
You could climb the lower aircraft so it is 1,000ft below the higher aircraft, then you will be able to make a better judgement about whether that remaining 2,000ft (that the lower aircraft must clear to be safely above the higher one) is doable or if you will have to wait until the two have passed each other. If you're impatient, you could then turn them anyway as before and make the climb!
These options may be combined in various ways and there are other options which I won't delve into at this stage because I don't yet know what they are.
If, taking all of these factors and possibilities into account, there is any doubt whether an aircraft may be able to cross levels with another aircraft and stay more than five miles distant, don't try it on! Notwithstanding a trainee air traffic controller's very feeble wage it's far, far more than the job's worth.