Taquari Valley: three days above the flood mark, from 25 August to 25 September 2026

“It rained a lot” and “the river rose” get said of almost every September in the South of Brazil. A flood on its own says little. To know whether a year was different takes a reference year read with the same ruler: the same weeks of the calendar, the same basin, the same instruments.

That is what River Pulse does. It reads four signs on the same dates in two years running and asks whether they showed up together.

The four signs follow the path the water takes. It falls in the headwaters, meets soil that takes it in or can no longer cope, comes down the channel and goes out onto the plain. When the four show up in the same period, the flood has gone from a local episode to a movement of the whole basin. That is what we call a river pulse.

In 2026 all four showed up. In 2025, only the soil.

The reading uses three terms the whole way through.

The headwaters in this reading are the basin upstream of Muçum, about 16,000 km². The valley takes in ten municipalities, about 1,560 km². The plain is a strip about 2 km wide along the Taquari inside that valley, about 500 km².

In 2026 the rise shows first upstream, on 21 September, at Passo Tainhas, Passo Carreiro and Linha Colombo. Estrela crosses the flood mark that same day, and the peak arrives on the 22nd, together with Muçum and Encantado. Taquari sees the flood on the 23rd.

In the headwaters, 365.3 mm fell in 2026 and 143.6 mm in 2025. The total counts for less than the concentration: the wettest day of 2026 yielded 131.6 mm, more than three times the wettest day of 2025, which stayed at 36.5 mm. In 2026 there were two days over 50 mm. In 2025, none.

The two maps separate by their pattern, before they separate by intensity. In 2026 the red gathers in the headwaters, around Vacaria and Caxias do Sul, while the valley to the south stays lighter on the same scale. The rain fell on the part of the basin where the river begins. In 2025 the colour is even and the valley keeps pace with the headwaters: it rained in a similar way on every side, and the river stayed in its bed.

Estrela went over 19 m on 21, 22 and 23 September, peaking at 24.46 m on the 22nd. That same day discharge read 10,762 m³/s, about 11 million litres a second. The 2025 peak was 3,839 m³/s, almost three times less, and the stage stayed under the alert.

The chart shows stage and discharge at Estrela through the 2026 period. The alert line sits at 17 m and the flood line at 19 m.

Inside the strip of plain, the area the satellite saw with water went from 19.91 km² in 2025 to 46.96 km² in 2026. The measure is an area: square kilometres of ground covered. It does not say the volume of water.

Part of that area is the river itself, which is always there. What matters is the rest, the water that showed up where a satellite record of many years shows dry ground most of the time. That share went from 2.25 km² to 9.52 km², more than four times. It is the water that reached where it does not usually reach.

Each sign counts when it reaches its mark. On the Taquari, headwaters rain counts at 180 mm or more in the period, or with more rain than the year before and at least one day over 50 mm. Headwaters soil counts at a mean of 0.40 or more, on a scale from 0 to 1. The river counts when Estrela reaches 19 m. The plain counts when the area with water is larger than the year before.

1 of 4 is isolated wetness, no pulse. 2 of 4 leaves the signs mixed, still no pulse. 3 of 4 means one sign is missing, and the reading stays open. 4 of 4 means there was a pulse: headwaters rain, the soil, the river and the plain moved in the same period. Heavy rain alone is not enough, and a high river alone is not enough either.

In 2025 only the soil reached its mark. Wet ground without heavy rain in the headwaters does not move the river. In 2026 all four arrived, and the movement crossed the basin, from the rain at Vacaria to the water in the street at Taquari, in three days.

The count does not measure the height of the flood. A 4 of 4 says the mechanism was complete. The stage in this reading is what says the height: 24.46 m at Estrela, five and a half metres above the flood mark. This valley has seen the river higher. The same count would describe a larger flood. The height is what separates the episodes.

On the plain, unlike the headwaters, the soil was a little drier in 2026: 0.390 against 0.423. The ground was not saturated. The water that covered it came from upstream, down the river.

The reading uses the same day and month in both years, from 25 August to 25 September, with the 25th included.

Rain is measured in millimetres, soil on a scale from 0 to 1, water on the plain in square kilometres, and the river in metres of stage at the gauge. Rain and soil come from the whole area of the headwaters, the valley and the plain, not from one point. Each rain value is the daily mean of cells about 11 km across, summed over the period, and soil is the mean for the period in cells about 9 km across.

Water on the plain comes from satellite radar, with a pixel of about 20 m. The radar passes over this plain about every 12 days. In 2026 the passes were 30 August, 11 September and 23 September. In 2025, 28 August and 21 September. The map shows the greatest extent seen on those passes, not one particular day.

Stage and discharge come from the ANA service, at the Estrela station, code 86879300, with the alert at 17 m and the flood mark at 19 m.

In 2026 there was a pulse in the Taquari Valley: the four signs showed up together, and the flood was of the whole basin. In 2025, with the same wet soil and less rain in the headwaters, there was not.

The practical lesson is in the difference between the two years. What came before the three days of river out of its bed was concentrated rain in the headwaters, upstream of Muçum, a day or two earlier. Someone who looks only at the sky at home is looking in the wrong place.

People will ask if this was El Niño. In the South of Brazil a wet month like this often sits with that pattern, but in the same weeks El Progreso, in Guatemala, went through drought. This reading does not settle the question. The Pacific is not a river in the Taquari, and it is too early in the phenomenon to call it the cause.