Skip to the story
mariuscomper.uk Română
today

Bucharest · what was here

The meandering Dâmbovița

There was a river that looped through Bucharest: 9.7 kilometres of meanders through the city, 49 mills along the Dâmbovița, islands and side arms. Between 1880 and 1883 they cut it straight, and between 1985 and 1988 they poured it in concrete — two rivers stacked: clean water on show, sewage underneath. Scroll back through the river's three lives, then type your address and find out if you live on the old riverbed.

Scroll: you go back to 1789, on the free river.

Marius Comper · 3 October 2026

The free river, 1789

On Purcell's 1789 drawing, the Dâmbovița enters the city at Ciurel and leaves at Văcărești, winding between neighbourhoods, with mill weirs at every turn. The drawing is not a survey — it is a town sketch, and that is how we show it: as a drawing, not a measurement.

The river you see moving is drawn from the 1871 Pappasoglu plan, the best map of the wild river: 9.7 kilometres of main bed through the city, plus the Tanners' Gârlița and the mill leat. The fight against the floods had already begun around 1775.

  • the 1871 bed, after Pappasoglu
  • today's streets

Mills and islands, 1852

On Borroczyn's 1852 plan the river is still wild — but the reproduction we work from is small and faded, so we show it as a reconstruction, with a question mark attached. What is certain: the city lived between waters. The quarter between the arms was called “între gârle” (between the brooks), and at the city's exit, by the slaughterhouse, there was still an island in 1852.

On I. Teofilu's plan, 49 mills are fixed along the Dâmbovița — as many as stood before 1862. After the floods of 1864–65, the law for abolishing the river mills was to be decreed, with all their dams and weirs. The free river had a decade and a half left.

The cut

The canal, 1880–83

On 21 November 1880 (old style) they began digging a deep trench from Ciurel to below the Abator, so the Dâmbovița could never flood the city again. The works lasted until 1883 (some sources say 1885, under the French engineer Boisguerin): the bed deepened by 6 metres, the bottom lined with oak planking, consolidated banks, 20-metre quays and two water drops, at Grozăvești and Vitan, for a major bed with a flow of 220 cubic metres per second.

The straight line you see is engineer Cerkez's 1880 plan — the project, not the survey. Where it cuts, it cuts without mercy: some plots move from one bank to the other, the side arms and eyots vanish, and the drained meadow makes no more ponds. In the window covered by both plans, the canal is 15.9% shorter than the river: 4.304 against 5.119 kilometres. Bridges were built over the sloped banks: in 1889 there were 12, seven of stone and five of iron.

What came after

The canal did not stay as they dug it. In 1900 they removed the Grozăvești drop and trained the bed upstream, to Ciurel. Between 1900 and 1950 they built the great sewers: six on the right bank, eighteen on the left. In 1942 a new rectification began, between the Ciurel weir and Roșu commune.

From Ciurel (91.43 metres) to the city's exit, below the Abator (63.10 metres), the right bank drops 28.33 metres — the slope this whole story runs down.

The double river

Concrete, 1985–88

The third great stage, between 1985 and 1988: upstream of the Grozăvești node they built Lacul Morii, the Argeș–Dâmbovița–Ialomița basins were interconnected, and through the city the river became two stacked courses — a concrete tub of clean water at the surface, a sewage collector box underground, connected to the sewerage.

Two rivers stacked where one winding river had been.

Today

Today it runs straight, between concrete walls. The meanders stayed underneath: under asphalt, under blocks, under courtyards. Somewhere, under your street, the 1871 water may still run — on paper.

Search your address below: the map knows 8,625 numbers on 534 streets, on a 500-metre strip along the old bed.

Look for yourself

Pick the year, then turn the map with the mouse or a finger. The wheel, the two-finger pinch or the + and − buttons zoom in and out.

We cover the meander strip: 8,625 numbers on 534 streets, at most 500 metres from the 1871 bed or the canal line. The old bed's position has a median error of about 127 metres — take the verdict as a neighbourhood, not a survey peg.

The river in figures

Every figure comes from our measurements on the plans or from the sources below; method and errors are in the next section.

MeasureValueSource
Main bed, 18719.7 kmPappasoglu trace
Tanners' Gârlița + mill leat1.4 + 2.1 kmPappasoglu trace
Projected canal, 18805.7 kmCerkez trace
Shortening in the shared window15.9%4.304 vs 5.119 km
Mills on the Teofilu plan49Cebuc, BMIM
Bridges in 188912 (7 stone, 5 iron)Giurescu 1966
Ciurel–Abator right-bank fall28.33 mGiurescu 1966
Major bed capacity220 m³/s (1880–85 canal)ASTR 2012

How we rebuilt the river

The 1871 bed

Major Pappasoglu's 1871 plan is the best map of the wild river. We fitted it to today's map on 21 churches, with an affine transform: the median error is 127 metres, and 80% of points fall within 215 metres. We also tried an elastic warp (thin-plate spline), but it came out worse — so we publish the affine with its residuals. On the fitted plan we traced the main bed (9.716 km), the Tanners' Gârlița (1.373 km) and the mill leat (2.070 km).

The 1880 canal

Engineer Cerkez's 1880 plan, scale 1:10,000, is the canal project: in blue, the new cut; in green, the abandoned meanders. We fitted it on 6 churches (median error 88 metres) and traced 15 points along the cut: 5.743 km on the raw points, 5.729 km after resampling at 80 metres. The line's ends do not touch the 1871 bed — the canal abandons the loops, in the west by over a kilometre. The line shows the project; what was actually dug between 1880 and 1883 is confirmed by texts, not by our geometry.

1789 and 1852: drawings, not measurements

For 1789 we have Purcell's drawing (sheet 9), and for 1852 a small reproduction after Borroczyn. Both are sketches, not surveys: we fitted them roughly to today's map, with errors of hundreds of metres, and we show them as reconstructions. In the 3D motion, the years 1789 and 1852 use the measured 1871 bed — the closest honest measure of the free river — and the drawings appear as documents, not geometry.

Today's river and the streets

Today's bed and the streets come from OpenStreetMap (ODbL licence). The 1985–88 training — the surface concrete tub and the underground collector box — is described by engineers Stematiu and Teodorescu (ASTR, 2012); only the tub is drawn in 3D, the underground box is not. River, canal and streets are drawn wider than in reality, so they read from above.

What we do not know: the date of the Teofilu plan with the 49 mills — a mills map drawn by eng. Teofil is dated June 1864 (Duțu, n.37, MIB archive, inv. 42464), but Cebuc gives no date, and the post-1862 demolitions should postdate the sheet, to be checked against the file; the exact year the canal was finished — 1883 after Giurescu, 1885 after ASTR; and the meaning of the southern blue line on the Cerkez plan, which we did not use.

Sources