The operational rule (CAT.POL.A.215) offers two ways to be legal past the terrain, both judged on the net path and both within 9.3 km (5 NM) either side of the intended track:
Clear the terrain climbing — the net path has a positive gradient at 1,000 ft above all terrain and obstructions along the route.
Drift down past it — the net path clears all terrain and obstructions vertically by 2,000 ft, continuing to an aerodrome where a landing can be made, with the net path showing a positive gradient at 1,500 ft above that aerodrome.
What moves the net path is everything that moves thrust or drag:
Mass — heavier means a steeper drift and a lower level-off, which is why fuel jettison can be part of the procedure.
Temperature — a warm day robs MCT of the thrust the drift-down is built on.
Anti-ice bleed — the escape routes over high terrain often carry an icing penalty on the data.
Wind — changes nothing about the path through the air; it only decides where over the ground the altitudes happen.
Descending on Mach, descending on IAS
The everyday descent runs the climb schedule backwards: constant Mach from the cruise level down to the crossover altitude, constant IAS below it, and 250 kt below FL100. The kinematics mirror the climb:
Constant Mach (below the tropopause) — the warming air raises the speed of sound, so the TAS rises, and the IAS rises faster, climbing toward VMO; above the tropopause the TAS holds constant while the IAS still rises.
Constant IAS — the TAS and the Mach number both fall as the air thickens.
That geometry names the limits: MMO guards the top of the descent, and the constant-Mach segment presses the IAS toward VMO, which is precisely why the reference changes to IAS at the crossover. The buffet chart relaxes on the way down: descending at constant Mach, the corridor widens and the margin to both the low-speed and the high-speed buffet increases: the rising IAS pulls away from the low-speed boundary while the boundary Mach numbers spread apart below the aerodynamic ceiling.
Energy management: the descent as a ledger
An aeroplane at the top of descent holds three currencies:
Chemical energy — the fuel in the tanks, the only currency that costs money.
Potential energy — the altitude, banked on the way up.
Kinetic energy — the speed the aeroplane carries.
The descent is the art of spending the second and third so that almost none of the first is needed: a continuous descent starts at the computed top of descent and rides idle thrust all the way, trading altitude for distance with no level segments burning fuel and making noise on the way. Every early level-off converts a cheap descent mile into an expensive cruise mile.
The profile is planned against the ground, so the wind edits it: an increasing tailwind stretches the ground distance the descent covers and leaves the aeroplane high and fast on the planned profile; an increasing headwind does the opposite, dropping it below profile unless the descent starts later. The Mach-to-IAS conversion matters too: through the constant-IAS segment the TAS is steadily falling, a built-in deceleration that flattens the path the aeroplane actually needs, and the FMS bends the planned profile around it.
And when it goes wrong: a shortcut, a late descent clearance or a tailwind burst leaves the aeroplane high or fast, with more energy than the remaining track can absorb. Energy in hand must go somewhere, and the only exits are drag and distance:
Speedbrakes — drag on demand at almost any speed, the first honest tool.
Earlier configuration — gear and flap buy serious drag once inside their limit speeds, steepening the achievable path.
Track miles — an extension or a hold from ATC adds the distance the energy needs.
Slowing early — sheds the kinetic share of the excess while track remains and unlocks the configuration drag; note that a clean jet above its best lift-to-drag speed actually glides further per thousand feet when it slows, so slowing works through what it enables, not by steepening the clean path.