Reading the Sky Before the Rain
Picture a yard in Central Vietnam in early summer, covered in drying lúa (unhusked rice). On the first day a grandmother sits at the edge of the sân (concrete yard) with a rake across her knees and watches the sky all morning before she moves a grain. On the second day the rice is in heaps under a tarp before most of the village has noticed a cloud. Both days end in rain. The clouds told her which kind was coming.
The scene is invented. The cloud sequences in it come from the sources at the bottom, and the vocabulary comes from [[Cloud Study]]: cirro is high, alto is middle, strato is flat, cumulo is puffy, nimbo brings rain.
The slow rain
On the first morning there are a few thin white hooks very high up. That is Cirrus, the hooked kind (uncinus), and by itself it says little. She keeps raking.
By noon the hooks have merged into a pale veil and a ring has appeared around the sun. The WMO describes Cirrostratus as a "transparent, whitish cloud veil" covering the sky and "generally producing halo phenomena" (WMO). She notes the ring and turns the rice once more.
Around three the ring is gone. The sun is still there, blurred, like a lamp behind frosted glass. This is the cue she was waiting for. Altostratus has parts thin enough to reveal the sun "at least vaguely, as through ground glass or frosted glass," and it "does not show halo phenomena" (WMO). A missing ring means the veil has thickened and come down a level. UCAR's front diagrams say rain is usually possible within about six hours once altostratus or low stratus appears (UCAR). So she starts raking the rice into heaps, without hurry. By dusk the sky is a grey lid and the drizzle is steady, which is Nimbostratus doing its job.
Behind the sequence sits a warm front, warm air sliding up over cooler air. The rising air makes cirrus far ahead of the front, cirrostratus as the front gets nearer, then the lower, wetter layers as it arrives (UCAR). Read in order, the clouds are one front walking toward the yard, a step lower each time, and in my reading that is why the prefixes are worth memorising: cirro, alto, strato, nimbo is the countdown. UCAR gives a lead time only for the last step. The sources I found for the earlier steps gave rough, conflicting hours, so I left them out.
The fast rain
The second dawn is clear. By nine a few small puffs with flat bottoms sit over the hills, Cumulus humilis, the fair-weather kind. What bothers her is a row of little turrets standing on a flat mid-level ledge, like battlements. That is Altocumulus castellanus. A newsletter from the NWS office in Albany says that when these clouds are seen in the morning they indicate "high instability and a good likelihood of late afternoon and early evening thunderstorms," while the same turrets in the late afternoon foretell a fair night (NWS Albany, 2017). She spreads half the usual rice and leaves the tarp by the door.
By noon the puffs have grown. An updraft pushes a cumulus upward until it "soon looks like a tower (called towering cumulus)," and at that stage there is little to no rain (NWS Spotter Guide). This is when the neighbours still say it is a lovely day. Later the tower's top turns fibrous and icy (capillatus) and spreads into an anvil (incus). The storm is now mature, which the NWS calls the most dangerous stage, with hail, heavy rain, frequent lightning and strong wind (NWS Spotter Guide).
The last clue comes before the rain does: a low wedge of cloud sliding out ahead of the dark base, the shelf cloud (arcus). NWS Green Bay explains that it forms as "dense, rain cooled air cuts underneath the warm, moist air ahead of the storm," that it "almost always precedes gusty winds," and that it is "not necessarily a precursor to severe weather" (NWS Green Bay). A cool gust reaches the yard first, dust and husks lift off the concrete, and she sweeps the whole yard at a run. Of all the signs in this note it gives the shortest notice, because it is the storm's own leading edge. That part is my reading.
The same NWS guide puts the life of a single thunderstorm cell at 30 to 60 minutes (NWS Spotter Guide). If this is one cell, the rice sits under the tarp for about an hour, not overnight. That part is my inference.
The sky before a typhoon
Central Vietnam gets a third kind of rain, on a much longer clock. NOAA's Hurricane Research Division lists a timeline for one example Category 2 hurricane approaching a coast (a typhoon is the same kind of storm) (NOAA AOML). At 96 hours out a one-metre swell rolls in every 10 seconds under an ordinary sky. At 48 hours the sky is clear, the wind nearly calm and the swell about three metres. At 36 hours the barometer starts to slip and "on the horizon a large mass of white cirrus clouds appear." By 30 hours there is a high overcast, by 24 hours low clouds join it, and by 18 hours the first rain squalls arrive.
In that list the sea speaks first and the cirrus second. NOAA's own timeline has a hurricane watch at 48 hours and a warning at 36, so by the time the sky shows anything, forecasters have already spoken. For this kind of storm the cloud is a confirmation, not the alarm. That is my reading.
Where the sky misleads
Order carries the meaning. A single cirrus streak or a single turret proves little, and the grandmother trusts the sequence. That is my reading too.
The first scene is the textbook warm-front case from the mid-latitudes. I found no source testing how well it holds in Vietnam's rainy season, where afternoon storms like the second day's probably matter more. That is a guess.
Clouds can catch what models miss. In April 2026 the CIMSS satellite blog described Altocumulus castellanus appearing over southern Wisconsin. The model sounding hinted at mid-level instability but "did not forecast convection or precipitation," and the clouds went on to produce a severe thunderstorm (CIMSS). In the other direction, a shelf cloud can bring a gust and no severe weather at all.
So the sky is good for deciding about the rice. For lightning, hail or a typhoon, the official warning wins, and the cloud is a second opinion.
Where to Dig
- The front behind the first scene. UCAR's "Clouds Form Due to Weather Fronts" has the cross-section diagrams. The WMO definitions of Cirrostratus and Altostratus are short and worth reading as written.
- Storm life cycle and spotter vocabulary. The NWS Spotter Guide covers the three stages. Evan Heller's "Ten Neat Clouds" in the Albany newsletter is a friendly tour of the rarer species.
- Shelf clouds and their neighbours. The NWS Green Bay page covers shelf clouds. Wall clouds and funnels, where rotation is the warning, are only one line each in [[Cloud Study]] and need their own read.
- The typhoon clock. NOAA AOML's H5 FAQ gives the hour-by-hour list. I haven't looked for a Vietnam-specific equivalent from the national weather service.
- Folk rules. "A ring around the moon means rain" is everywhere, and I found no study measuring how often it holds.
- The view from above. Geostationary satellite loops should show the cirrus shield spreading and the towers growing. Matching them against a sky I watch myself would test the whole note. I haven't tried.
Author: Ope Tags: #science Links: [[Cloud Study]], [[Core Memories]]