Vertical Predictions

Every time, fuel figure and pseudo-waypoint on every FMS page comes out of one prediction engine. This article covers its three concerns: what it assumes you will fly, where it draws its answers, and how its foundation — the performance model — is calibrated. (The phase machine feeding it is article 22; optimisation and fuel policy are article 25.)

1. One engine, two sets of assumptions

Per FCOM DSC-22_20-10-40:

The predictions displayed on the MCDU assume that the FMGES will guide the aircraft along the preplanned lateral and vertical flight plans. The predictions displayed on the ND assume that the aircraft will continue to operate in the modes (selected or managed) that are currently active. As long as the aircraft is flying the flight plan under managed guidance, the predictions on the MCDU will match those on the ND.

The MCDU assumes you will return to the plan (the ideal world); the ND assumes you keep flying as you are (the real world). Fully managed, the worlds coincide; go selected, and the two screens begin telling different stories. The engine's input list: both axes of the plan and their revisions, the cost index, current wind/temperature, position versus plan, the active guidance modes, the cabin rate of descent, FUEL PRED-page GW/CG, engine anti-ice state in descent, the PERF APPR QNH, and the speed-control method — with dashes shown while a field is being recomputed (dashes = thinking, not broken). And the official usage rule, one sentence worth a chapter:

The flight crew should rely on the ND for short-term predictions. It indicates what the aircraft will do under the currently engaged modes (selected or managed). The flight crew should rely on the MCDU for long-term predictions when managed guidance is active or about to be reengaged.

2. The idealist's assumptions — the MCDU side

Per FCOM DSC-22_20-10-40:

‐ If the aircraft is not guided along the flight plan (selected guidance), the MCDU predictions assume that it will return immediately to the flight plan, intercepting at a predetermined angle, and will then proceed under managed guidance. ‐ If the aircraft does not fly the managed speed profile (ECON, SPD CSTR...), the MCDU predictions assume that it will remain at the present selected speed/Mach until it reaches the next SPD CSTR or SPD LIM or enters the next flight phase.

Off the route → assumed to return at once. On a selected speed → assumed to hold it until the next constraint, limit or phase, then resume managed. Every MCDU ETA is built on the prodigal's return. The geometry of that return:

If the aircraft is not on the lateral flight plan, the FMGEC assumes (for prediction) that it will return immediately to the active lateral leg with a 45 ° convergence angle or that it will fly directly to the 'TO' waypoint whenever the required convergence angle is greater than 45 °.

A 45° convergence back onto the active leg — or direct to the TO waypoint when more than 45° would be needed. So under radar vectors the MCDU's EFOB and ETA are not wrong — they are conditional, and the condition is written in that 45°: the longer and further the vectoring, the shorter the imagined homeward path compared with the real one, and the more optimistic the numbers (article 30's monitoring calibration).

3. The realist's symbol system — the ND pseudo-waypoints

The ND lays its predictions along the plan (in NAV) or along the track line (otherwise), with colour as grammar:

• Level-off arrows: reaching the FCU altitude = blue arrow; reaching a constraint altitude (when stricter and the modes suit) = magenta arrow — blue is what you wound, magenta is what the database owns (article 06's colours, extended to the map).
• T/D: always white — it is merely a geometric fact (article 08: nothing happens there by itself); the continue-descent symbol runs white/blue with DES armed or not.
• I/P (intercept point): in DES with a vertical deviation, where the aircraft is predicted to regain the profile — the graphic answer to article 08's "deviation without visible correction" (the N+1 rule).
• Speed-change dot (magenta): where automatic acceleration/deceleration will meet a SPD LIM/CSTR/hold speed — and on selected speed the symbol simply vanishes (no theatre for a play that won't run).
• DECEL — per FCOM DSC-22_20-10-40: ‐ Magenta, if in managed speed and NAV or approach mode is engaged ‐ White, if in selected speed or HDG/TRK mode ‐ Automatic deceleration only occur when displayed in magenta. A white DECEL will not slow the aircraft (and only a magenta DECEL sequences the approach phase — article 22's ticket).
• Constraint circles: magenta = predicted satisfied; amber = predicted missed; white = guidance is not even honouring it — NAV engaged but flying OP/V-S, per the FCOM: the system disregards any altitude constraints and puts white circles around the waypoints (article 09's chain, on the map).
• The green arc (energy circle): centred on the aircraft, facing along the track line, its radius = the required landing distance — from present position and speed down to VAPP at touchdown. Shown only with a HDG/TRK lateral mode in cruise/descent/approach phase (the newer standard adds "within 180 NM of destination") — it disappears in managed lateral, because with a profile you don't need an energy ruler.

As a general rule, the ND indicates what the aircraft will fly, with the current active FG modes.

The solid green line is the trajectory being flown (the plan dims to dashes in HDG), and every symbol re-seats itself at each prediction cycle.

4. The MCDU's ledger — and the stars

The type table: pseudo-waypoints T/C, T/D, S/C, S/D, I/P, SPD LIM, DECEL; F-PLN A carries per-fix TIME/SPD/ALT plus destination ETA/DIST/EFOB; F-PLN B carries per-fix EFOB and forecast wind (article 22's display page); VERT REV serves the miss amount and per-fix EFOB/EXTRA; PROG serves REC MAX, VDEV, required distance to land, and the direct distance. The stars:

When a speed or an altitude constraint is at a waypoint, a star symbol appears adjacent to the speed or altitude prediction. If the star is magenta, the constraint is predicted to be matched. If the star is amber, the constraint is predicted to be missed. … Note: If an altitude constraint is predicted as missed, the system tells you what will be the error at the specific waypoint.

Magenta star = the promise will be kept; amber = it won't — and for a missed altitude constraint the VERT REV page states by how much. No mental arithmetic required. VDEV appears in three views — the green brick beside the PFD altitude scale in descent, the number on PROG, the I/P on the ND — together answering which side of the profile, how far, and where I rejoin.

5. Calibrating the foundation — PERF and IDLE factors

Per FCOM DSC-22_20-10-40:

With time, the real aircraft drag and engine performance can deviate from the nominal model. The airline Flight Operations should periodically revise the value of these factors to adapt FMGES predictions to actual aircraft performance.

Airframes age; the model doesn't — unless the operator fits the reading glasses. The PERF factor corrects the predicted fuel flow (predicted FF = model FF × correction), effective for the whole flight and reaching even the ECON speeds; +1.5 means this airframe burns 1.5 % more than book. The IDLE factor corrects the assumed thrust of the descent profile (idle + delta): a positive factor makes the profile shallower and T/D earlier (more thrust margin in the model — recoverable when the wind shifts), a negative one steeper and later; the FCOM's own example: an IDLE factor of +4 lengthens the computed descent by 11 NM. This is the origin of article 08's "idle plus margin" assumption — and of article 04's three-layer answer to "the predictions are off". Changing the factors: on the ground only, behind a change code (default ARM, AMI-settable), authorised personnel.

DPO (descent profile optimisation — an option, not installed on the baseline airframe): the idea is to shave the idle margin and the approach speed margins for fuel — a steeper profile, later T/D, shorter deceleration segments; the price is flexibility against wind error — the forecast must be right, VDEV watched, speedbrakes ready. For a non-equipped fleet its teaching value is the mirror: it explains why the standard profile deliberately keeps its margins.

6. The descent profile — built once, rebuilt twice

From the profile clauses (same chapter): the descent profile is computed before the descent phase begins, from both axes of the plan, the DES WIND and PERF APPR winds, and the maximum cabin rate of descent; its goals — economy, constraint satisfaction, and 1000 ft at VAPP. In descent it is recomputed in exactly two cases: the plan is revised, or the approach data (wind/VAPP/landing configuration) is changed. And the ECON climb/descent speeds freeze at phase activation (article 25) — changing the cost index mid-descent does not change the speed you are descending at.

7. Operating the two screens

The two-screen reconciliation habit (both rules quoted in section 1): under radar vectors — the ND tells the truth (green arc for energy, white DECEL as a reminder that slowing is now your job), and the MCDU's ETA/EFOB carry a 45° discount; back in NAV with managed guidance — the screens merge and the MCDU regains its authority. Prediction weirdness: no reset — the three-step triage plus the SEC-minus-STAR trick (article 19); dashes mean wait. Interfaces: T/D's passivity (article 08) is why it draws white here; DECEL's magenta condition meets the phase machine (article 22); ECON/OPT/REC MAX and fuel policy are article 25.

[!warning]- Four misconceptions this article corrects (1) MCDU predictions under radar vectors are not wrong — they are conditional on an immediate 45° return to the plan; their credibility decays with the vectoring. (2) A white DECEL will not decelerate the aircraft — only magenta acts (and only magenta sequences the approach phase); on selected speed the speed-change dots vanish entirely. (3) A white circle around a constraint fix means guidance is ignoring it right now (OP/V-S with NAV engaged) — not that the constraint is satisfied. (4) The green arc is not decoration — it is the required landing distance from the present energy state, and it exists precisely (and only) when you are off managed lateral guidance.

Self-test

[!note]- Q1. The two screens' assumptions, one sentence each?

MCDU: the aircraft will return to and fly the pre-planned lateral and vertical plans. ND: the aircraft will continue under the currently active modes. Fully managed, they match.

[!note]- Q2. The two branches of the 45° assumption — and its effect on EFOB credibility under vectors?

Return to the active leg at a 45° convergence, or direct to the TO waypoint when more than 45° would be required. The imagined path shortens relative to the real one as vectoring extends — the fuel/time numbers grow optimistic.

[!note]- Q3. Blue level-off arrow versus magenta level-off arrow?

Blue: reaching the FCU-selected altitude. Magenta: reaching a constraint altitude (when stricter and the modes honour it) — yours versus the database's.

[!note]- Q4. What does a white DECEL mean — and when do the speed-change dots disappear?

Selected speed or HDG/TRK: no automatic deceleration will occur (magenta only). The dots vanish whenever speed is selected.

[!note]- Q5. A white circle around a constrained waypoint with NAV engaged — what is happening?

The vertical guidance is OP or V/S: constraints are disregarded, and the system marks them white to say so.

[!note]- Q6. What does the green arc measure, when is it displayed, and why is it absent in managed lateral?

The required landing distance from present position/speed to VAPP at touchdown. Displayed with HDG/TRK lateral in cruise/descent/approach (newer standard: within 180 NM). With managed lateral there is a profile — no energy ruler needed.

[!note]- Q7. Beyond the amber star, what else does the system tell you about a missed altitude constraint?

The magnitude of the predicted error at that waypoint, on the VERT REV page.

[!note]- Q8. What do the PERF and IDLE factors each correct — and the FCOM's +4 example?

PERF: predicted fuel flow (whole flight, touches ECON speeds). IDLE: the assumed descent thrust — positive = shallower profile, earlier T/D; an IDLE factor of +4 lengthens the computed descent by 11 NM.

[!note]- Q9. When is the descent profile computed, and in which two cases is it recomputed during descent?

Before the descent phase begins (inputs include DES WIND/PERF APPR winds and the cabin rate). Recomputed only on a plan revision or a change of approach data (wind/VAPP/landing configuration).

[!note]- Q10. Short-term and long-term — which screen for each?

Short-term: the ND (what the aircraft will do under current modes). Long-term: the MCDU — provided managed guidance is active or about to be re-engaged.

Key takeaways

References

The two assumption sets, input list, usage rule, MCDU return assumptions and the 45° geometry, ND pseudo-waypoint behaviours (arrows, T/D, I/P, speed dots, DECEL colours, constraint circles, energy circle), MCDU prediction tables, star grammar with miss-amount note, VDEV views, and the PERF/IDLE factor clauses (including the +4/11 NM example and change-code protection) per FCOM DSC-22_20-10-40 (predictions section); descent-profile construction, recomputation conditions and the DPO description per its profile sections. DPO is not installed on the baseline airframe of this series — presented as the mirror that explains the standard profile's margins. The two-screen habit is an integrative synthesis.

Independent study material, not an Airbus publication and not endorsed by the manufacturer. Always defer to the current operator FCOM, FCTM, and QRH for operational use.