Late Runway Change on the Arrival

KLAS→KSFO is 360 nm — about an hour. Short enough that the arrival is programmed early and then left alone, which is exactly the setup a late runway change ambushes. (360 nm is the great-circle figure; you will also see 414 miles quoted, which is statute.)

A normal level for the sector. It matters here only because the FMGS needs a vertical plan before the arrival you are about to select means anything to it.

The plan reads KLAS, a discontinuity, KSFO — the bare skeleton the FMGS builds from FROM/TO alone. Everything else you are about to see, you put there.

DYAMD is the arrival for San Francisco landing 28L/R — the West Plan, which is how SFO runs almost all the time — and I28L is the approach that goes with it. INSERT strings the arrival, the VIA and the final approach into the plan ahead of KSFO. Two things about the page. ARRIVAL> is offered at 1R only at the destination: a LAT REV on an enroute waypoint does not expose it, on the real aircraft or here. And the approach comes first because it carries the runway — there is no runway field on this page, and there never was on the aircraft either.

A waypoint typed onto the flight plan always leaves a discontinuity after it — the Airbus FCOM says a waypoint inserted directly "creates a discontinuity between the new waypoint and the previous one" — because the FMGS will not assume the next fix joins on. Clearing it is you telling the box that it does. INYOE is a published transition onto the arrival, out near Lee Vining, and it sits naturally on a Vegas–San Francisco routing. Two things worth banking. First, a plan with a DISCONTINUITY in it is not a plan the FMGS will fly through — it will not sequence past the gap. Second, INYOE feeds both of the arrivals in this scenario, so the transition you just entered survives the runway change that is about to happen. The enroute half of the plan is rarely what breaks.

A runway change this late is the classic destabiliser — NASA's ASRS has a whole Callback issue on in-close approach changes, and the head-down MCDU time it costs is precisely what you do not want below 10,000 ft. The airmanship answer is to decide early whether you can accept it at all. Having accepted it, the MCDU work starts where it always starts: LAT REV at the destination.

ALWYS is the San Francisco arrival for landing 19L/R — the Southeast Plan, which SFO uses when a winter storm makes the usual westerly configuration unusable. It is genuinely a different named arrival, not a different runway transition on the same one: DYAMD and ALWYS share their track as far as the ALWYS fix and then diverge by landing direction alone. That is why the runway change forces this work at all. Note what you just did in the first press rather than the second: on the Airbus the APPROACH is the selection that carries the runway, and it is what decides which branch of the arrival ends up in the plan — you never type a runway on this page, you pick the approach that lands on it.

Roughly twice the lines there should be. The arrival you were on is still there in full, and the new one has been strung in after it — ALWYS, LAANE and DYAMD each appear twice. Fly this as written and you would track through the whole of the old arrival before setting off for the start of the new one. This is the complaint people post about after changing an arrival in flight: the aircraft heading away from the airport, back toward a fix it has already passed. Whether a given FMS or add-on does exactly this varies, and it is not documented Airbus behaviour — but the habit that protects you is the same everywhere. After any arrival change, read the plan. Do not assume the box tidied up behind you.

DIR TO is the tool that unpicks this. Going direct to a fix on the new arrival cuts out everything between you and it, which is precisely the stale legs of the old arrival, and it puts you back in managed guidance rather than hand-steering an aircraft whose flight plan you no longer trust. Two things about the page itself. The entry executes: 1L is the DIR TO key, and with no RADIAL IN, RADIAL OUT or ABEAM PTS selected it routes you from present position straight to what you typed, with no INSERT to press and no ERASE to fall back on. The page stays on screen afterwards with HEFLY at 1L, but the direct is already flying; the RADIAL IN and RADIAL OUT fields now showing on the right would turn it into a radial intercept, which this clearance does not ask for. And you had to type HEFLY because this page lists only the plan's first four waypoints, so a fix deep in the arrival never appears there; on the aircraft the list shows the waypoints ahead of you, where a line key would have reached it in one press.

What you are confirming is what you can confirm: no DISCONTINUITY anywhere in the plan, and the new arrival's fixes present and running to KSFO. That is the check worth making every time — a plan the FMGS can sequence from end to end, ending at the runway you were actually given. Read what the direct-to actually did: the eighteen lines are gone, because going direct to HEFLY cut out everything between you and it — which was precisely the stale legs of the old arrival. On the line the equivalent check is the ND: one continuous magenta track from the aircraft symbol to the airport, with nothing doubling back on itself.