The river under the city

Inside the city the Dâmbovița made many meanders, which braked the water and silted the bed. Mills, vegetable gardens, improvised bridges and commercial buildings crowded the banks and strangled the course — and since the city had no sewerage, the river had become the collector of all household and industrial residues.

After the 1774 deluge, Alexandru Ipsilanti starts the first works in 1775: a wooden dam at Lungulețu and an 8 km derivation canal to the Ciorogârla, described by Sulzer in a work published in 1782. The Ipsilanti ditch is considered by specialists Romania's first hydrotechnical work — but it tamed only the upper Dâmbovița.

People threw rubbish into the ditches, and the waters broke out again: 1801, 1804, 1814. With no human victims, nobody took the floods seriously — when the water rose, Bucharesters climbed on roofs with what they could grab. Mills pressed the water surface toward the banks, and in 1831 the outhouses, stables and kiosks on the banks were ordered cut, mills included.

Plague and cholera epidemics were frequent until the second half of the 19th century, amid poor hygiene and no sanitary system; one contemporary blamed cholera on “the unhealthiness of the breathed air and bad water”. Only after the devastating floods of 1862–1865 did Cuza order the river rectified and channelled.

Then came the years of small works: after 1864–1865 the Arcuda canal is dug, in 1868 the bed is cleaned with small rectifications between Foișor and Manolache commune, in 1871 the campaign pushes hard upstream from Foișor, and in 1870–1872 the main wooden bridges are replaced with metal ones. Rectification inside the city had run since 1850, removing 182 properties on the left and 192 on the right.

Floods did not vanish after the canal either: in 1893, at 125 m³/s, the dikes broke upstream of Conțești, and in July 1975 the highest known discharge was recorded, 650 m³/s. The small river still knows anger.

The real rectification starts in 1879–1880: the bed is deepened by 6 m, floored with oak on beams and piles, walled below and grassed above, with two waterfalls — at Grozăvești, with an electric plant, and at Vitan — holding the levels. The 1880–1885 works are carried out by the French engineer Alexandre Boisguerin.

The canal's figures, after the engineers of the time: the minor bed carries 8 m³/s, the major up to splai level 220 m³/s, and the splais run 20 m each, paved and tree-lined on both sides. From a wild river, the Dâmbovița becomes a city axis.

In 1885 the Brezoaiele hydro node moves the river onto the Ipsilanti derivation: high waters leave for the Ciorogârla, while small and medium flows go through a 10 m³/s canal to Arcuda, for drinking water. In 1900 the Grozăvești fall is removed and the bed is arranged upstream to Ciurel.

Modern drinking water is born alongside: the 1879 project, filtering the Dâmbovița, finishes in 1884 (Culmann and Ziegler), and from 1888 the Arcuda station sends water through the 15.5 km aqueduct to Grozăvești. The river became first a canal, then a drinking-water source.

Compare the two plans: in 1871 the river loops deep past Curtea Veche and snakes toward Văcărești; in 1911 the canal cuts straight through the centre, and meanders survive only west and south, outside the channelled reach.

Cișmigiu park sits on a vast pond: the 1791 Austrian plan shows marshy ground, and the pond — “merchant Dura's”, donated between 1671 and 1691 to the Sărindar monastery — linked through a ditch to the Dâmbovița floods. The garden's name comes from Dumitru Suilgi Bașa, the “great Cișmigiu”, who walled here one of the two fountains ordered by Ipsilanti.

In 1830 Kiseleff orders Baron Borroczyn to drain the pond and make a public garden. An 1840 report finds Cișmigiu's water below the Dâmbovița level — in rains the river drained into the pond — and proposes regulating the river and deepening the ditch. In May 1844 the pond is surveyed (Râmniceanu), Bibescu calls in Meyer, creator of the Kiseleff garden, and a commission finds the pond princely land except for “Fântâna Boului”. Princely order no. 71, of 15 February 1845, seals the garden; the lake cleaning starts in 1859.

Filimon, quoted after a modern work (Vătămanu, 1973), paints the pre-1843 pond in strong colours: endless water, rot, frogs, “venomous reptiles”, epidemic vapours — and an island with a tavern of vices in the middle. Take the picture as literature, not as minutes.

To the north, the Colentina earned its lake chain only in the 20th century: in 1933 works start on the Buftea reservoir, Herăstrău lake — the most representative — finishes in 1936, when Floreasca and Băneasa begin, and by 1970 all landscaped lakes look as today. Downstream of the Buftea dam run 15 lakes: Flămânzeni, Buciumeni, Mogoșoaia, Chitila, Străulești, Grivița, Băneasa, Herăstrău, Floreasca, Tei, Plumbuita, Fundeni, Pantelimon I, Pantelimon II and Cernica.

And south of the old city, on the river's meadow, lay Ochiul Boului and Tătarului lakes — gone today. Bucharest's water map is far richer than the concrete canal suggests.

Between 1930 and 1932 the silt is dredged between the Abator and Ciurel, over 8,438 m. Then, between 1934 and 1936, the central course is covered with concrete slabs between the former Senate bridge and Piața Unirii — 610 m of unseen river, poured “to hide it from passers-by”.

In 1942 a new rectification starts between the Ciurel dam and Roșu commune, continued after 1944, draining Ciurel, Crângași and Roșu. Also in the century's first half, 6 main collectors join on the right bank and 18 on the left, all over 150 cm wide.

There were grander dreams too: one project wanted a navigable Dâmbovița, linked to the Bucharest–Danube canal, with the bed widened to 50–60 m and a lake at Ciurel. It stayed on paper.

After 1975 — less green space, more stormwater, and the metro changing the groundwater drainage — the complex 1985–1988 arrangement was imposed: a stepped artificial basin for the primenire flow and part of the floods, plus the collector box beneath the clean-water basin.

The system's figures, from July 1985, over 10.4 km from Lacul Morii to Vitan-Bârzești: the basin runs 20–60 m wide and 1.5–4.5 m deep and transits 55–120 m³/s; the box gathers 12 main and 11 secondary collectors, 17,238 m in plan and 44,632 m developed, to the Glina treatment plant, draining by gravity. In heavy rains the box carries 20 to 80 m³/s.

The level is held by six hydro nodes with flap gates: Grozăvești, Eroilor bridge, Piața Națiunilor Unite, Mărășești bridge, Mihai Bravu and Vitan-Bârzești. The upstream shield is Lacul Morii: 19 million m³ total volume, 220 ha, 15 million m³ permanent retention.

Defence runs in several lines: Ilfov reservoirs, the Brezoaiele node, the 1.8 km Bolovani derivation for 50 m³/s, the non-permanent Dragomirești and the Giulești polder — 295 ha and 5.64 million m³, partly filled by the Sector 6 mayoralty, hence unusable. And the Văcărești lake at the south intake, meant for microclimate and leisure, was never filled: unfinished works and studies showing it would flood basements by raising the water table.

Do not mix them up: another, upstream Văcărești belongs to the defence scheme with reservoir and polder. The city one, at the south intake, stayed an empty plan — a nature park today.

We started from two peer-reviewed engineering histories (ASTR 2012 and PESD 2008), two studies from the Bucharest City Museum journal (Cișmigiu; fires–epidemics–floods) and a study on the Colentina's anthropic lakes. The river history on this page is only what these sources say; the full list, with quotes, is in the Sources section below.

The old course comes from the 1871 Pappasoglu plan (public domain, Wikimedia Commons scan): we traced it by hand with 79 points, following the river band at full resolution, then fitted the sheet over the city with an affine transform on 12 control points — churches and buildings that never moved (Antim, the Patriarchate, Colțea, Domnița Bălașa, Radu Vodă, Slobozia, Manu Cavafu, Cărămidarii de Sus, Foișor, Cotroceni palace, Filaret station).

We measured the error honestly, dropping each point from the fit in turn: median 114 m, maximum 199 m. Every course point inherits the error of its nearest control, grown with distance. That is why lookup results are indicative: the old plan and its fit allow no exact address-level measurement. The 1911 plan (1:5,000, Army Geographic Service) served as witness: its canal sits on the same alignment as today's, checked at Eroilor, Unirii and Mihai Bravu.

What we could not prove stayed out: the oft-repeated crediting of the channelling plans to Grigore Cerchez has no solid source we could find (ASTR puts the 1880–1885 works on the Frenchman Alexandre Boisguerin), nobody gives the mills' number, the 1865 flood's victims turned up in no reliable source, and the Dâmbovicioara — the tributary encyclopaedias mention — is absent from the 1871 sheet where indicated.

Today's addresses and canal geometry are OpenStreetMap (© OpenStreetMap contributors, ODbL licence); church coordinates, Nominatim. Lookup sends nothing anywhere: distances are precomputed at build, the browser only displays them.