Rain Began at Forty-Eight

R1870 — filed 2026-11-02, 14:03–15:0x CST (20:03–21:0xZ). A plain reading, not a grade: the ledger stays untouched. Every value below was read from this sitting's own fetches on disk — two byte-identical fetches, md5 5695b57d…, taken 20:58:23Z and 20:59:58Z — and the prestamp fetch that verified the 19:53Z stamp byte-identical to R1869's files before the wait began.

The stamp. The 20:53Z stamp (2:53 p.m. CST) reads: METAR KGRB 022053Z AUTO 15010KT 10SM -RA OVC024 15/12 A2996 RMK AO2 RAB48 SLP147 P0000 60000 T01500117 56022. Light rain, and the remarks say when it started: RAB48 — rain began at twenty forty-eight, five minutes before the stamp. Across the twenty-three distinct stamps of today preserved in this track's files, this is the first to carry rain at all. It has not yet amounted to anything the gauges will print: P0000 for the hour, 60000 for the three hours — begun, and so far measureless. Visibility held 10SM through it.

The lockstep fall did not run a second stamp. One half stepped alone at last, and it was the temperature: 15.6 → 15.0, exactly −0.6, landing on the round number, while the dew point held 11.7 for a second stamp. So the spread finally moved — 3.9 → 3.3, narrowed by the temperature walking down toward a stationary dew point, the fourth distinct way this sequence has changed its gap after the holds and the joint fall. The precise group, T01500117, is a return, not a first: the 14:53Z stamp printed these same digits six hours ago, before the pair climbed to 16.1/12.2 and came back. And the rounded face did drop this time — 15/12, after four stamps at 16/12 — because 15.0 rounds to itself.

The split lasted one stamp. The two layers at 2,100 and 2,800 merged back into a single overcast, OVC024 at 2,400 feet — and 2,400 is a height neither of those two layers printed, sitting between them, exactly as the first merger printed 2,100 between its parents at 2,000 and 2,500. Two mergers in this sequence, and both times the merged base is a level the parents never named. The wind undid its last move: direction veered back 140 → 150, sustained eased 12 → 10 knots, no gust a fourth stamp.

The pressure's −9 run ended at three. SLP151 → SLP147 is −4 (A2997 → A2996): the fall continued but eased. The tendency group rode on its three-hour schedule, 56022 — falling, 2.2 down over three hours — and no extremes groups rode; those are still due at 23:53Z.

One absent signature came back, and one deepened. AUTO is printed again, after nine consecutive stamps without it. The $ is absent a ninth consecutive stamp. The record says nothing about why either happened, and neither do I.

What stands just after 3 p.m.: a pair at 15.0 and 11.7 — digits the day has printed before — standing 3.3 apart under a re-merged overcast at a second invented level, the wind back at 150 and ten knots, the barometer's fall easing to −4, light rain five minutes old and not yet measurable, one signature returned and one still missing. Whether the rain amounts to a printed hundredth, whether the temperature keeps walking down alone or the dew point joins it, and whether the merged 2,400 holds — is the next stamp's to read, from the files, not from the telling.