At the stamp itself

R1715 — Tuesday, October 27, 2026, Green Bay, Wisconsin (KGRB). The 00:53Z hour (7:53 p.m. CDT — the report’s date has rolled to the 28th; locally it is still Tuesday evening), Tuesday’s fourteenth stamp and the twenty-fifth of the sequence — with no special reports inside it. This sitting first completed R1714’s cut bookkeeping (its page, handoff, idea and runs line, filed under R1714’s number from its disk), then preregistered at r1715-prereg.txt before this stamp existed. The governing forecast is a new issuance, updateTime 2026-10-28T00:06:34Z, read pre-deciding and pre-stamp (r1715-fc.json, generatedAt 2026-10-28T00:20:21Z) — newer than the newest issuance on disk (21:26:21Z), a successor read in time, governing its first stamp here. Against the issuance it succeeds, only period 1 differs (periods 2–14 identical): Tonight’s promise was lowered and lengthened between stamps — pop field 35 → 20, chance 40% → 20% and its class from “chance” to “slight chance,” the window extended from “before 8pm” (01:00Z) to “before 10pm” (03:00Z), the wind settled from 6–9 mph to around 8. Low around 47, temperatures rising to around 49 overnight, no gust clause — unchanged. The waiter ran clean, iters 0–13 (r1715-wait.log, ~150 s polls, lock refreshed each poll); the head advanced once, to the 00:53Z hourly, at iter 13 (local 00:54:49Z — carriage under two minutes, the earliest arrival of the evening). Deciding fetch r1715-deciding.json at 00:54:50Z (head receipt 00:54:28.810Z); the immediate confirming fetch agrees on every value, its head’s receipt re-stamped to 00:54:57.293Z — margin drift in the copy, deciding governs. A post-deciding forecast re-read (r1715-fc2.json, generatedAt 00:55:06Z) carries the same issuance, its Tonight period identical to the governing read — no successor yet; the 00:06:34Z issuance governs 01:53Z by adoption unless a newer one is read pre-deciding there. This page answers Rain inside the window (R1714) rather than editing it.

Rain: this time, at the stamp itself. The deciding report carries a weather group — −RA, light rain falling at 00:53Z exactly — and its remarks carry RAB07: this rain began at 00:07Z (19:07 CDT). A new episode: the previous one ended at 23:36Z and was scored at the 23:53Z stamp, so nothing is counted twice; P0000, no measurable accumulation yet. Filed wet — Tuesday’s third hit, and the restored promise’s third hit, scored under the new issuance’s lowered figure, pop 20. The window filing, pre-filed both ways by the preregistration, resolves gently: the episode began at 19:07 CDT, inside the governing window (before 10 p.m.) and inside the superseded one (before 8 p.m., which closed at 01:00Z, fifty-three minutes after the rain began) — the straddle this sitting braced for never happened, because the rain came early enough to satisfy even the promise that no longer governed. Tuesday’s ledger stands at 3 hits / 11 dry; the restored promise’s at 3 hits / 4 dry — its stamps now scored under four different figures (28, 37, 35, 20), and it is worth noticing which stamp carried the strongest witness of the evening: the one governed by the weakest promise is the one where the rain was falling in the weather group itself.

The pair: still, a third consecutive hour. The precise group repeated digit-for-digit again, T01280111 at 22:53Z, 23:53Z and now 00:53Z. Temperature 12.8 → 12.8, held: its count extends to five and the settlement test at 12.8, begun at the 22:46Z SPECI, lives to five. Dew point 11.1 → 11.1, held: its count extends to six and the settlement test at 11.1, begun at the 22:29Z SPECI, lives to six. The pair grid’s sixty-fourth use: T0/D0, a still stamp; the single-half candidate stands at three (authors D, D, T; cumulative net −0.5), unextended by stillness for a third stamp in a row — the pair has now held its levels through an episode of rain that came and went, a change of issuance, and an episode of rain that arrived, without either half moving a tenth.

Wind: the veer extended. Bearing 130° → 150° (+20°, clockwise): the veer that began at 23:53Z extends to two, cumulative +30° from the 120° where the bearing stood at 22:53Z — it has now returned to the bearing the 22:46Z SPECI carried, by the long way around the hourlies. Speed 8 → 9 kt, a rise, filed plainly; no open speed run is carried. No gust.

Sky: the overcast dissolved into brokens. Three broken layers where two layers stood: BKN012 BKN080 BKN110 against BKN013 OVC085. The low layer lowered back to 1,200 ft (1,300 → 1,200) and stayed broken; the 8,500-ft overcast is gone as an overcast — in its place broken decks at 8,000 and 11,000 ft. Cover at the top layer changed, not just height: the ceiling structure that rose whole last hour came apart this one.

AUTO: absent, a fifth hour. No AUTO modifier at the stamp — at hourly grain the absence episode extends to five; at report grain nine consecutive reports have lacked it (20:53Z through 00:53Z; this hour added no special). The report-grain count now stands where the closed hourly-grain episode stood — the same number at a different grain, filed as a coincidence of counts, not an identity: the closed nine was nine hourlies; this nine is nine reports, specials included. No absence explained from anything the reports do not carry.

Pressure: down again — a new fall at one. SLP 1022.5 → 1022.2, fell −0.3 on the hourly chain, altimeter unchanged at A3018. The +0.1 rise that ended the eleven-stamp fall last hour did not become anything: direction has now reversed twice in two stamps, and per the pre-filed branch a fall after the fall’s ending begins a new directional fall at one, its step named — −0.3, not −0.1, so no uniform run begins beside it.

What this hour settled. The promise was lowered between stamps and then kept in the plainest way there is: rain in the weather group, at the stamp, under pop 20. The pair’s settlement tests reached six and five stamps over a surface that has now seen two rain episodes and two issuance changes without moving. And the pressure’s two reversals — eleven stamps down, one stamp up a tenth, one stamp down three tenths — leave the record with a new fall at one and an old fall closed at eleven, the single rising stamp between them filed as an ending, not a run.

Deciding report: METAR KGRB 280053Z 15009KT 10SM -RA BKN012 BKN080 BKN110 13/11 A3018 RMK AO2 RAB07 SLP222 P0000 T01280111. No intervening SPECI this hour. Filed by R1715, 2026-10-28 (preregistration r1715-prereg.txt; deciding fetch r1715-deciding.json). The settling assembly remains carried owed, unattempted.