Kansai (KIX) Observed Traffic '−78%' — Anatomy of a Coverage Shortfall, September 2026 (closed)
- Kansai did not shut down on 6 September. Our own count dropped to 105; an independent ADS-B network counted 478 movements there the same day and has not dipped since.
- The shortfall was in our collection: our raw position feed over the Kansai terminal area thinned to about a third between 03:00 and 13:00 UTC that day and read ordinary every day after. Our count read 450 on 7 September and 455 on the 9th. This page is now closed.
- Our published KIX numbers for 5–8 September were revised twice as the aggregation settled, and both revisions are shown on this page. 8 September is an aggregation gap in our table, not a second drop. Nothing here says anything about the operating condition of the airport — for that, check your airline and the airport on the day.
Advisory timeline — how active is this episode
Kansai International had been one of our steadiest counts in Asia: between 393 and 497 observed movements a day for 28 consecutive days, a 468/day 30-day median with no gap day. Then 6 September read 105 — the day that tripped the anomaly flag — and 7 September read 450, back inside the normal band, with 455 on the 9th. 8 September is marked partial: it is an aggregation gap on our side, not traffic — our worldwide table still holds only 116,126 movements for that date three days later, against 247,484 on the 7th and 247,790 on the 9th, and our raw position feed over Kansai read ordinary all through the 8th. 10 September was still aggregating when this page was closed and is not plotted. The daily series over the window:
Fig. 1 — Observed movements per day at Kansai International (KIX), 9 Aug – 9 Sep 2026 · * partial day · Data as of 11 Sep 2026, 15:48 UTC · Source: FlySafe monitoring (crowdsourced ADS-B receiver network).
Airports in the exposure zone
| Airport | Size | Observed signal | FIR |
|---|---|---|---|
| Kansai International Airport (KIX) | large | — · 6 Sep: 105 observed movements (66 dep / 39 arr) vs a 468 baseline; independent network 478 the same day and never below 471 since; raw terminal-area sample thinned to ~a third 03:00–13:00 UTC; 450 on 7 Sep, 455 on 9 Sep | RJTG |
| Osaka Itami International Airport (ITM) | large | 43 km · 6 Sep: 155 vs a 191 baseline — modestly low, the nearest neighbour to Kansai; 173 on 7 Sep, 162 on 9 Sep | RJBG |
| Kobe Airport (UKB) | medium | 23 km · 6 Sep: 60 vs a 94 baseline — modestly low; 93 on 7 Sep, 90 on 9 Sep | RJBG |
| Tokyo Haneda International Airport (HND) | large | — · 6 Sep: 1,326 vs a 1,366 baseline — ordinary; 1,345 on 7 Sep, 1,333 on 9 Sep | RJTG |
| Narita International Airport (NRT) | large | — · 6 Sep: 630 vs a 632 baseline — ordinary; 622 on 7 Sep, 608 on 9 Sep | RJTG |
| Fukuoka Airport (FUK) | large | — · 6 Sep: 436 vs a 400 baseline — ordinary; 405 on 7 Sep, 409 on 9 Sep | RJBG |
Distance alone does not determine impact — wind direction and the flight levels the ash reaches decide which airports and routes are affected on a given day. FIR = Flight Information Region.
What our monitoring shows: our number fell, Kansai's traffic did not
FlySafe tracks observed departures and arrivals at 493 airports from crowdsourced ADS-B receiver data. Observed counts are a floor, not a total — receiver coverage varies — but an airport measured against its own baseline is a reliable shape. Kansai International's last 30 days:
Fig. 2 — Observed daily movements at KIX, 11 Aug – 9 Sept (latest complete UTC day). Data as of 11 Sep 2026, 15:48 UTC · Source: FlySafe monitoring (ADS-B receiver network).
- Two low bars, two different things. The 6 September bar is the event. The 8 September bar (54) is an aggregation gap on our side, not traffic — our worldwide table still holds only 116,126 movements for that date, against 247,484 on the 7th and 247,790 on the 9th, and the raw position feed over Kansai read ordinary all through the 8th. Do not read it as a second drop.
- 30-day baseline in our own feed: 468 observed movements/day (median), range 393–497 across 9 Aug – 5 Sep with no gap day.
- 6 Sep: 105 — 66 departures, 39 arrivals, 78% below the baseline. 7 Sep: 450. 9 Sep: 455. Both back inside the band.
- The independent ADS-B network read 478 on 6 Sep — inside its own 475–516 band for late August, with 477 on the 5th — and then 471, 489, 496 and 488 on the 7th through the 10th. It never dipped. On the day our count collapsed, the other network saw an ordinary Sunday at Kansai.
- Settled, the hole was Kansai's alone. On 6 September our feed read Tokyo Haneda 1,326 against a 1,366 baseline, Narita 630 (632) and Fukuoka 436 (400) — ordinary. The two airports closest to Kansai ran modestly low: Osaka Itami 155 (191) and Kobe 60 (94). Japan-wide we logged 3,996 movements against 4,451 on the 5th and 4,438 on the 7th, and worldwide 221,095 against 247,484 on the 7th — about 11% short. A mild global softness, and one deep local hole on top of it.
- The raw feed dates the window. Distinct aircraft seen below 10,000 ft over the Kansai terminal area: 65 across 03:00–12:59 UTC on 6 September, against 193 and 184 in the same hours on the 5th and 4th — a sample thinned to about a third. The overnight block either side read normal (75 for 00:00–02:59, against 81 and 75), and the 7th through the 10th read 182, 168, 187 and 190 in the 03:00–12:59 hours — four ordinary days in a row.
- Numbers on this page were revised twice as the aggregation settled. First version (8 Sep): 364 for 5 Sep, 100 for 6 Sep, a partial 32 for the 7th. Second (9 Sep): 460, 45, 430, and Japanese neighbours reading low. Fully settled: 460, 105, 450 — and the neighbours ordinary after all. The verdict never changed; the evidence under it moved twice, and both moves are on this page.
Our counts are a floor, not a total: they are what our receivers saw and what our aggregation has finished writing, and a thinner sample prints a smaller number even when nothing at the airport changed. Live view: KIX airport brief.
What it means for flights
A crowdsourced count can fall without any traffic falling. Observed movements are reconstructed from ADS-B receivers run by volunteers, then aggregated on our side. When receivers near an airport go offline — power, internet, a failed antenna, a feeder on holiday — or when our own collection has a thin day, the flights still operate, but our reconstruction stops seeing the low-altitude arrival and departure legs it needs. The number drops; the runway does not. This is why we publish observed counts as a floor, not a total, and why a −90% reading opens an investigation rather than a headline.
The one-airport test, run twice, on two different states of the data. Whether the neighbours fell with Kansai decides where to look. On 8 September they read normal; on the 9th, with the aggregation half-settled, they read badly short and this page said so. Fully settled, they read ordinary: on 6 September Haneda 1,326 against a 1,366 baseline, Narita 630 against 632, Fukuoka 436 against 400, with the two airports nearest Kansai modestly low — Itami 155 against 191, Kobe 60 against 94. Japan-wide we logged 3,996 movements against 4,451 on the 5th and 4,438 on the 7th; worldwide we were about 11% below the 7th. So the final picture is a mild general softness in our feed that day, plus one deep hole centred on Kansai and fading with distance from it — the signature of receiver coverage around one airport, not of anything happening on its runways.
Two independent counts disagreeing is the strongest single check we have. Our feed and the independent ADS-B network we cross-check against are built from different receiver populations. Through late August they agreed closely on KIX — ours 393–497, theirs 475–516. On 6 September ours read 105 while theirs read 478; on the 7th ours read 450 against their 471, and their count has stayed between 471 and 496 every day since without a single dip. When two networks with different receivers disagree that hard on one day and agree on every day either side of it, the disagreement is about the receivers and the pipeline. If the airport had actually stopped, both would have fallen.
The raw positions date the window — and it never reopened. Below the daily aggregate sits the position feed itself. Counting distinct aircraft under 10,000 ft over the Kansai terminal area: the 00:00–02:59 UTC block on 6 September read 75, in line with 81 and 75 on the 5th and 4th — then 03:00–12:59 UTC fell to 65 aircraft against 193 and 184 in the same hours on the two previous days, a sample thinned to about a third. From the 7th to the 10th those same hours read 182, 168, 187 and 190. One window, a clean start and a clean end, and four ordinary days after it. That is not the shape of an operational disruption, which builds and decays with the schedule and leaves cancellations behind it. It is also why 8 September is a gap in our derived table and not in reality: the positions for that day are there, the daily aggregation for it is not.
What the airspace record says — and what it does not. No closure event registered for either FIR covering the Kansai area (RJTG or RJBG) at any point in the window, and no NOTAM series appeared in our active set for either FIR. This is a FIR-level observation and does not describe conditions at the airport itself; we publish no operational verdict on Kansai, and go/no-go decisions belong to the airport, the airlines and the Japanese authorities.
Why the numbers on this page moved — twice. The first version, published 8 September, was written on an aggregation that had not finished: 364 for 5 September, 100 for 6 September, a partial 32 for the 7th, near-baseline neighbours. The second, on the 9th, read 460, 45 and 430 with the neighbours badly short, and concluded the shortfall spanned Japan. Fully settled, those days read 460, 105 and 450 — and the neighbours were ordinary after all, so the second version's Japan-wide reading was an artifact of a half-written table and is retracted here. We publish the revisions rather than quietly overwriting them, because a page about the difference between a data hole and a disruption has no business hiding one of its own. The lesson we are keeping: a partly-aggregated day can look like evidence in either direction, and neither reading should be trusted until the table stops moving. What never moved across all three versions: the verdict, the independent network's 478, and the raw position feed.
What the data showed — closing entry, 11 September 2026. The signal is over and this page is closed. The recovery number: our observed count at Kansai read 450 movements on 7 September and 455 on 9 September — 96% and 97% of the 468/day baseline, both inside the 393–497 band the airport held for the four weeks before the event. The deficit lasted one day and one window inside it. The full ledger, as it finally settled: What the data did not show, at any point: cancellations, a closure, or any operational verdict on Kansai. We publish no such verdict — that belongs to the airport, the airlines and the Japanese authorities. What it did show is a detection layer firing correctly on its own feed and an automated multi-signal cross-check identifying the cause as coverage within hours, twice over a moving table. Closing state for the record: coverage shortfall, single day, recovered. Current numbers for this airport continue on the KIX airport brief; this page is a settled archive and will not be updated further.
Passenger questions, answered from the data
Our data does not show an operational disruption. Our own observed count fell to 105 movements against a 468/day baseline, but an independent ADS-B network counted 478 at KIX the same day and has read 471–496 every day since, and our raw position feed over the Kansai terminal area thinned for about ten hours that day and read ordinary before and after. We read this as a loss of coverage on our side, and our own count read 450 on 7 September and 455 on the 9th. For the status of a specific flight, check your airline and the airport on the day.
Because that is what our receivers and our aggregation produced, and we publish what we see. Our detection layer flags any tracked airport that falls 20% or more below its own rolling average; KIX fell far below its baseline on 6 September and the flag fired correctly — 90% below on the unsettled figures it fired on, 78% below once the aggregation finished. The investigation behind the flag is what determines whether a drop is operational or a coverage artifact — here, the cross-checks said coverage, and the page says so.
Treat them as a floor, not a total: they are the movements our crowdsourced receiver network observed and our aggregation has finished writing, which is always at or below the real figure and varies with coverage. That is why we cross-check against an independent network, against neighbouring airports and the global total, and against the raw position feed before calling any drop operational — why days with known coverage or aggregation gaps are labeled on the chart rather than quietly plotted as low traffic, and why this page shows the figures we revised when the aggregation settled.
For media — the full kit
Everything on this page — charts, CSV, quotes — is published under CC BY 4.0: free to reuse with attribution to FlySafe.zone. Charts come in dark (social/web) and light (print/light layouts) versions; attribution and license are embedded in every file.
“Observed traffic at Osaka's Kansai International fell to 105 movements on 6 September 2026 in FlySafe's own crowdsourced ADS-B feed — 78% below its 468-a-day baseline — while an independent ADS-B network counted 478 movements at the same airport that day and has counted between 471 and 496 every day since, and FlySafe's Kansai count returned to 450 the following day and 455 on 9 September, indicating a loss of receiver and aggregation coverage rather than a disruption to flights, according to flight-monitoring service FlySafe.zone, which has closed the event.”
Boilerplate — about the source (for your article footer)
FlySafe.zone monitors airspace risk and observed airport activity worldwide: 334 flight-information regions scored daily, GPS-interference signals, volcanic-ash exposure and daily observed movements at 493 airports from the crowdsourced ADS-B receiver network. Methodology: flysafe.zone/methodology.
Need a different cut of the data, a comment, or numbers for another airport? Leave an email — we answer fast during active events.

