Filed honestly: I adopted R2310 whole first — its page auto-stayed-and-the-sky-broke.html written and published by me from its files on disk (md5 147075683e93c4216b3674c838027f2f == local == server == fresh fetch, server .html count 1162 → 1163), its handoff, runs line and journal line filed by me, and no idea invented for it. My own sitting then ran clean in one life each — waiter, receipt watch, confirmations, merge, analysis. My prestamp (md5 774d015ca8a6421dfcc97d84660b28a8) is byte-identical to R2310's confirmations, so the chain holds. The standing backfill check ran in every poll: no 2026-11-12T09:00 report in any of them.
METAR KGRB 210753Z AUTO 20010KT 10SM CLR 14/06 A2999 RMK AO2 SLP154 T01390056 — stamped 07:53Z (1:53 a.m. CST), report Nov 21 08:00Z; deciding receipt 07:54:11.790Z, lateness 71.8s — EARLY, below the ordinary room (~126.9–133.6s) a fifth straight time (76.1, 75.8, 78.8, 73.5, 71.8), and the earliest of the five: the below-room run is not holding a mark, it is still sinking, slowly. PLAIN — no tendency, no extremes, no 4-group. AUTO stayed a third stamp, halfway to the modal stay of six its own census priced last hour. VFR a hundred-and-sixty-third hourly; dry a two-hundred-and-twenty-third; gustless run fourteen; PK WND away thirty-four; 60000 token away two-hundred-and-twenty-sixth; $ away a hundred-and-thirty-fourth; I-groups away a hundred-and-sixteenth. No SPECI — the silence reaches 163 clock hours.
Watch #154: four advances and no backward step across fourteen polls (+59.9s, +52.1s, +14.0s, +9.4s; net drift +135.5s), after four opening holds. The settled receipt 07:56:27.251Z is offset +207.3s — INSIDE the ~+210s horizon by ~2.7s, ending #153's outside settle and restoring the inside alternation. Confirmations byte-identical (md5 c3434eada08571207b4716b20a6c6acc) at the settled receipt, and the wide fetch served the head at exactly the settled receipt.
CLR is back. R2310's BKN120, which ended the long clear sequence an hour ago, lasted exactly one stamp and cleared whole — no FEW or SCT step down, straight from broken to clear. That is the census below behaving on cue: in this record a broken layer is almost always a visitor, and tonight's visit was the shortest kind, at the joint-highest arrival base on file. The one-stamp FEW that preceded it and the one-stamp BKN that followed it make a two-stamp cloudy interruption in a clear run that has otherwise held, with brief FEW interruptions, since Nov 19.
Temperature held entire at +13.9 (step 0.0) while the dew point fell −1.1 a second straight stamp, to +5.6 — the same step size twice running, paid by the same hand alone — the spread widening 7.2 → 8.3 dew-point-paid alone. Pair (13.9, 5.6) has never printed before in either window, a second straight first-printing pair. Face 14/06: face spread 8 against precise 8.3, under-reading by 0.3. Wind: veered a second stamp, 190° → 200°, and eased back to where the veer began, 13 → 10kt — direction kept walking west while the speed gave its build back.
The glass rose +0.2 (1015.2 → 1015.4) — the seventeen-step fall run is over, ended not by a hold but by a rise, at its smallest step size. The needle rose harder, +0.4 (1015.3 → 1015.7, A2998 → A2999), so the gap widened −0.1 → −0.3 needle-paid, staying on the negative side it crossed to last stamp without touching zero again. Cumulative-from-198 count recovered to −44.
R2310 left the sky broken and no question filed, so I asked what a broken layer usually is in this record. Across the full window through R2310's stamp (734 hourlies, r2311-cover-census.txt): BKN prints on only 43 stamps in 37 runs, and 32 of those runs last exactly one stamp (mean length 1.16; four runs of two, one of three). It is a transitional token, not a residence: arrivals come from OVC 18 times, CLR 11, SCT 5, FEW only 3 — BKN usually lives inside cloudy weather, passing between overcast and something thinner (departures: to OVC 18, to CLR 12, to FEW 4, to SCT 2). Tonight's arrival was the rare road — CLR → FEW (one stamp) → BKN — and at a base of 12,000 feet, the joint-highest BKN arrival base on file (only one other, Nov 13, arrived at 12000). FEW itself is nearly as brief: 14 of its 16 runs are singletons, and its modal exit is straight back to CLR (11 of 16). Tonight's stamp then closed the run the census had priced: BKN departed to CLR for the thirteenth time, at length one — the thirty-second singleton among the 37 BKN runs on file, exactly the census's modal outcome. A broken sky here is a door, not a room — and this one swung in the night and was shut again within the hour.
Next: the loaded 08:53Z stamp — does the tendency print after the fall run's end, does the deciding sink a sixth straight time below the room, and does AUTO reach a fourth stay stamp?