{"id":572904,"date":"2026-01-21T15:28:30","date_gmt":"2026-01-21T18:28:30","guid":{"rendered":"https:\/\/revistapesquisa.fapesp.br\/?p=572904"},"modified":"2026-01-21T16:22:39","modified_gmt":"2026-01-21T19:22:39","slug":"rivers-alter-landscapes-by-incorporating-parts-of-other-rivers","status":"publish","type":"post","link":"https:\/\/revistapesquisa.fapesp.br\/en\/rivers-alter-landscapes-by-incorporating-parts-of-other-rivers\/","title":{"rendered":"Rivers alter landscapes by incorporating parts of other rivers"},"content":{"rendered":"

The more erratic the course of a river, with seemingly illogical deviations, the greater the interest it holds for geographer Andr\u00e9 Salgado, of the Federal University of Goi\u00e1s (UFG). Over the past 15 years, using satellite imagery, soil analysis, and many hours of walking and observation, he has identified advances and retreats in several river networks in Brazil, expanding areas of research begun in the 1950s.<\/p>\n

His studies suggest that the course of the S\u00e3o Francisco River was once 250 kilometers (km) longer, before its headwaters were taken over by the Grande River, the main tributary of the Paran\u00e1 River. Conversely, the course of the Poti, the main tributary of the Parna\u00edba, between the states of Cear\u00e1 and Piau\u00ed, has extended since taking over the majority of the upper and middle stretch of the Acara\u00fa. In Rond\u00f4nia, the Branco River absorbed tributaries of the Essequibo, the largest river in Guyana, and in Santa Catarina the Itaja\u00ed-A\u00e7u took possession of tributaries of the Uruguay and Igua\u00e7u. \u201cRiver basins are constantly changing,\u201d says Salgado.<\/p>\n\n \n \n \n <\/picture>Alexandre Affonso\/Pesquisa FAPESP<\/span>\n

Through what is known as \u2018river capture,\u2019 rivers act like cannibals, engulfing parts of others which will shrink or disappear. In an article from May 2013 in Science<\/em>, Sean Willett and other geologists from the Swiss Federal Institute of Technology in Zurich called rivers participating in this slow-moving phenomenon \u2018aggressors\u2019 and \u2018victims,\u2019 reflecting their contrasting roles. The rate of surface or underground changes to water flow depends on the type of soil, the nature of the rocks (the less resistant they are, the easier water can erode them), and the slope of the terrain, since water tends to flow toward lower areas.<\/p>\n

Climate is also a factor, with the level of rainfall directly influencing the erosive power of rivers, as are tectonic movements, such as the uplifting of the Earth\u2019s outermost layer of soil and rock. \u201cRiver capture is like a cake being eaten from both ends. The one that eats faster reaches the other side first,\u201d compares Salgado.<\/p>\n

\"\"

Raphael dos Passos Borba \/ Wikipedia<\/span>The Itaja\u00ed-A\u00e7u River crosses Blumenau (Santa Catarina) after capturing tributaries of the Uruguay and Igua\u00e7u riversRaphael dos Passos Borba \/ Wikipedia<\/span><\/p><\/div>\n

By changing paths and merging segments of their courses, rivers increase or reduce the amount of water that irrigates a region. As a consequence, the climate, types of vegetation, and human occupation may change. \u201cThe Parna\u00edba captured the Poti, at some point between 23 million and 2.5 million years ago, leaving the current state of Cear\u00e1 drier than it would have been without this diversion of water toward Piau\u00ed,\u201d says Salgado, one of the authors of the May 2024 article in Revista Brasileira de Geomorfologia<\/em> which describes these episodes.<\/p>\n

US Geographer William Morris Davis (1850\u20131934) began studying this phenomenon when describing, in 1896, the capture of a river in northeastern France. Other analyses have explained the diversions of the Ohio and James rivers in the USA, the Thames River in London, the Nile in Egypt, and the Yellow River in China, among others. In Brazil, starting in the 1950s, geographers Aziz Ab\u2019Saber (1921\u20132012), of the University of S\u00e3o Paulo (USP), and Antonio Christofoletti (1936\u20131999), of S\u00e3o Paulo State University (UNESP), Rio Claro campus, proposed some possibilities of river capture, although still with limited arguments due to the less refined analytical techniques available at the time compared to today.<\/p>\n

\"\"

Tulio F\u2009\/\u2009Wikimedia | Andr\u00e9 Salgado \/ UFG <\/span>The Branco River (in the larger photo<\/em>), in Caracara\u00ed (Roraima), swelled as it gained waters from the Essequibo River, leaving behind traces of an ancient river on the boundary between the two basins (above<\/em>)Tulio F\u2009\/\u2009Wikimedia | Andr\u00e9 Salgado \/ UFG <\/span><\/p><\/div>\n

Over time, the reshaping of rivers took on political significance in terms of defining territorial boundaries, as happened with the Branco River. Its main tributaries were once part of the Essequibo River basin, which rises in a mountain range on the border between the states of Roraima and Par\u00e1, crosses Guyana, and flows northward into the Atlantic Ocean. In December 2021, in the Journal of South American Earth Sciences<\/em>, Salgado and colleagues from the Federal University of Minas Gerais (UFMG), where he taught until 2022, argued that the Branco River captured the Uraricoera, one of the tributaries of the Essequibo, and later the Tacutu, likely due to subsidence of the Earth\u2019s crust in southern Roraima, which would have changed the river\u2019s course and expanded its drainage area.<\/p>\n

\u201cIf the Branco had continued in a northeastern direction instead of shifting its course to the southwest, Brazil would have lost all of central and northern Roraima to Guyana, because the borders between the countries in this region were essentially defined by the boundaries of the river basins,\u201d says Salgado. \u201cAs the European colonizers entered the Amazon by the rivers, if the capture hadn\u2019t happened, the first to reach the interior of Roraima would have been the English.\u201d<\/p>\n

Analyses of sediments from the riverbanks indicate that this change likely occurred during a relatively recent geological period, between 18,000 and 10,000 years ago. Using other methodologies, fellow geographer F\u00e1bio Alves, of the Federal University of Western Bahia (UFOB), studied the region and confirmed the findings of the teams from UFG and the National Institute for Space Research (INPE), concluding that the capture of the Branco River must have started around 18,700 years ago.<\/p>\n

\u201cThe Amazon basin formed and expanded through successive river captures,\u201d says Alves, based on an article from October 2018 in Neotropical Ichthyology<\/em>. In that study, biologist James Albert, from the University of Louisiana at Lafayette, in the USA, and geologists Carina Hoorn, from the University of Amsterdam, in the Netherlands, and Pedro Val, from the City University of New York (CUNY), argue that the network of rivers crossing the Amazon rainforest is the result of river capture that occurred between 5.6 million and 4.9 million years ago, accelerating the diversification of plant and animal species in the region (see<\/em> Pesquisa FAPESP issue n\u00b0 334<\/em><\/a>).<\/p>\n

\"\"

Fellipe Abreu \/ Getty Images <\/span>Serra da Canastra and the upper stretch of the S\u00e3o Francisco River, just after its sourceFellipe Abreu \/ Getty Images <\/span><\/p><\/div>\n

Losses and diversions<\/strong>
\nResearch in this field has increased the interest in certain tourist attractions. Managers of Itatiaia National Park, for example, could place a sign near the trickle of water that emerges at an altitude of 1,980 meters, in the municipality of Bocaina de Minas, in the state of Minas Gerais, where the Grande River originates, noting that around 5 million years ago this was the source of another river: the S\u00e3o Francisco. Today, the Velho Chico<\/em>, the affectionate nickname given to the S\u00e3o Francisco River, begins in the Serra da Canastra mountain range, also in Minas Gerais, about 250 km away.<\/p>\n

Salgado became interested in this river capture after a remark by geographer \u00c9ric Andrade Rezende, who was doing his PhD at the time. \u201cIn 2015, \u00c9ric was on a road going to Pimenta, in Minas Gerais, and noticed something strange,\u201d he says. \u201cThere, the Grande River, the main headwater of the Paran\u00e1 River, made a curve on a flat area. He went back and said: \u2018What if this used to be the S\u00e3o Francisco and not the Grande River, like it is today?\u2019\u201d Salgado, who until then had studied smaller and less prominent rivers, decided to pursue the idea.<\/p>\n

Rezende, currently working in the municipal government of Contagem, in Minas Gerais, says: \u201cThe older the process is in geological time, the harder its evolution is to reconstruct, because the evidence gets erased by erosion or tectonic activity.\u201d Salgado, a PhD student at the time, and his advisor, geologist Paulo de Tarso Amorim Castro, published the hypothesis in an article in June 2018 in Revista Brasileira de Geomorfologia<\/em>, based on sediment analysis, geological structures, and surface movements in the region.<\/p>\n<\/div>

\n \n \n \n <\/picture>Alexandre Affonso \/ Pesquisa FAPESP<\/span><\/div>
\n

Jos\u00e9 C\u00e2ndido Stevaux, a geologist from the Federal University of Mato Grosso do Sul (UFMS) who studies changes in the Paran\u00e1 River over the past 10,000 years, got excited at a congress in Cear\u00e1 in 2019 when he heard a lecture in which Salgado presented these arguments. \u201cIn the 1980s, when I was at the University of Maring\u00e1, in Paran\u00e1, my biology colleagues would ask me if there had been any contact between the Paran\u00e1 and the S\u00e3o Francisco, because they shared many fish species, and I didn\u2019t know what to say,\u201d he recalls.<\/p>\n

Supporting the hypothesis, the group from UFG studied the dispersal of fish species common to both river networks, such as the dourado (Salminus brasiliensis<\/em>) and the pintado (Pseudoplatystoma corruscans<\/em>). The conclusions, detailed in January in the journal Geomorphology<\/em>, indicate that there was a temporary connection between the rivers and the headwaters of the S\u00e3o Francisco gradually became those of the Grande River between 5 million and 6 million years ago.<\/p>\n

As a result, the Grande River engulfed a thousand kilometers of the course of the S\u00e3o Francisco. Including its tributaries, the headwaters of the Paran\u00e1 River would have incorporated around 50,000 km\u00b2, slightly more than the area of the state of Rio de Janeiro. \u201cThe S\u00e3o Francisco is the big loser, because it lost water to the Paran\u00e1, to the Paragua\u00e7u, which originates in Chapada Diamantina, in Bahia, and to the Jequitinhonha, in Minas Gerais,\u201d notes Salgado (see infographic on page 47<\/em>).<\/p>\n<\/div>

\n \n \n \n <\/picture>Alexandre Affonso\/Pesquisa FAPESP<\/span><\/div>
\n

An expert on the transformations of the Velho Chico<\/em> over the past 90,000 years (see<\/em> Pesquisa FAPESP issue n\u00b0 331<\/em><\/a>), geographer Genisson Panta, a PhD student at the Federal University of Pernambuco (UFPE) and high school teacher at a public state school in Macei\u00f3, Alagoas, comments: \u201cAfter the upper stretch of the S\u00e3o Francisco was reshaped by means of captures, the river carved into rock and deepened its canyon, 2,000 km away, between 1.5 million and 3.5 million years ago.\u201d<\/p>\n

In Cabrob\u00f3, in the sert\u00e3o<\/em> (badlands) of Pernambuco, the S\u00e3o Francisco makes an abrupt turn to the east, also called an elbow, a classic sign of river capture, indicating that it was incorporated by a smaller and less voluminous river. Panta says that a sign of this river capture is an abandoned channel, known as a dry valley, on the border between Bahia and Piau\u00ed, where the river must once have run, as proposed in the 1990s by UFPE geographer Jannes Mabessone (1931\u20132007).<\/p>\n

\u201cThe Parna\u00edba may have been the ancestor of the S\u00e3o Francisco,\u201d he suggests. Civil engineer Lu\u00eds Flores de Morais R\u00eago (1896\u20131940), from the Polytechnic School at USP, had already proposed in a 1945 book that, millions of years ago, the S\u00e3o Francisco must have flowed north and emptied into the sea between Maranh\u00e3o and Piaui.<\/p>\n

Pollen<\/strong>
\n\u201cA strictly geological focus is limiting,\u201d concluded USP geographer D\u00e9borah de Oliveira after working in the Serra do Mar mountain range. Earth tremors reshaped the landscape and altered the courses of two rivers in the region, the Para\u00edba do Sul and the Tiet\u00ea. The former turned back toward the coast, and the latter, which had also once emptied into the sea, reversed its course and now flows inland through the state of S\u00e3o Paulo (see<\/em> Pesquisa FAPESP issue n\u00b0 77<\/em><\/a>).<\/p>\n

Oliveira studied the abrupt deviations of the Guaratuba River, in the Serra do Mar, which once flowed toward the Upper Tiet\u00ea and now empties into the sea in Bertioga, S\u00e3o Paulo. In 2013, she noticed that the abandoned channel of the captured river did not form a dry valley, rather a swampy one, with histosols\u2014soils rich in organic matter that could contain pollen or plant remains. Three years later, while walking to a conference at La Vilette Park, in Paris, she told geographer Nat\u00e1lia Nunes Patucci, one of her PhD students who was accompanying her: \u201cInvestigate the dry valleys. There\u2019s something important there.\u201d<\/p>\n

Patucci sought help from pollen specialists at the Center for Nuclear Energy in Agriculture (CENA), in Piracicaba, and at the Institute of Geology (IGC), both at USP. She learned how to perform analyses and eventually found pollen from plants typical of cold and humid climates, such as Podocarpus <\/em>sp., Ilex<\/em> sp., and Symplocos<\/em> sp. \u201cWhen the climate shifted to hot and dry, around 25,000 years ago, it must have rained more and erosion increased, facilitating the capture of rivers in the region,\u201d says Oliveira. Her plan is to search for pollen in other rivers she and her team are studying in the state of S\u00e3o Paulo.<\/p>\n

\u201cWhen it rained more, the rivers and lakes joined together,\u201d agrees Salgado, who has also done research in the region. \u201cAs the erosion continues, other rivers will be captured, once again reshaping the landscape that we see today.\u201d As an example, he cites the Amazon basin, which is gradually capturing the entire upper stretch of the Orinoco River Basin, in Venezuela, via the Casiquiare River. This aquatic cannibalism has advanced, but has not finished. As this connection grows, the Amazon River, the largest in the world, will appropriate 40,000 km\u00b2 of the Orinoco River\u2019s drainage area, the fourth largest globally\u2014equivalent to the land area of Switzerland.<\/p>\n

The story above was published with the title “Fluvial cannibalism<\/strong>” in issue 355 of September\/2025.<\/p>\n

Projects<\/strong>
\n1.<\/strong> Study of the relationships between relief \/ lithology \/ soil \/ drainage at various scales of analysis (n\u00b012\/19048-0<\/a>); Grant Mechanism<\/strong> Regular Research Grant; Researcher responsible<\/strong> D\u00e9borah de Oliveira (USP); Investment<\/strong> R$47,985.52.
\n2.<\/strong> Geochronology for evaluating river capture processes: Application in sediments from the dry valleys of Guaratuba (S\u00e3o Paulo) (n\u00b0 16\/06654-0<\/a>); Grant Mechanism<\/strong> Doctoral Fellowship; Supervisor<\/strong> D\u00e9borah de Oliveira (USP); Beneficiary<\/strong> Natalia Nunes Patucci; Investment<\/strong> R$155,123.41.<\/p>\n

Scientific articles<\/strong>
\nALBERT, J. S. et al<\/em>. The changing course of the Amazon river in the Neogene: Center stage for Neotropical diversification<\/a>. Neotropical Ichthyology<\/strong>. Vol. 16, no. 3, e180033. Oct. 18, 2018.
\nCASEMIRO, F. A. S. et al.<\/em> The timing of large drainage rearrangement in South America: A study based on morphological and ecological evidence<\/a>. Geomorphology<\/strong>. Vol. 468, 109457. Jan. 1, 2025.
\nREZENDE, E. A. et al<\/em>. Evolu\u00e7\u00e3o da rede de drenagem e evid\u00eancias de antigas conex\u00f5es entre as bacias dos rios Grande e S\u00e3o Francisco no Sudeste brasileiro<\/a>. Revista Brasileira de Geomorfologia<\/strong>. Vol. 19, no. July 3, 2018.
\nRODRIGUES, W. F. et al.<\/em> Subterranean river captures in siliciclastic rocks in a semiarid climate: The case of the Poti River Canyon, Brazilian Northeast<\/a>. Revista Brasileira de Geomorfologia<\/strong>. Vol. 25, no. 2. May 15, 2024.
\nSALGADO, A. R. et al.<\/em> Large rivers, slow drainage rearrangements: The ongoing fluvial piracy of a major river by its tributary in the Branco river basin \u2012 Northern Amazon<\/a>. Journal of South American Earth Sciences<\/strong>. Vol. 112, 103598. Dec. 2021.
\nSTOKES, M. F. et al.<\/em> Ongoing river capture in the Amazon<\/a>. Geophysical Research Letters<\/strong>. Vol. 45, no. 11, pp. 5545\u201352. June 16, 2018.
\nWILLETT, S. D. et al<\/em>. Dynamic reorganization of river basins<\/a>. Science<\/strong>. Vol. 343, no. 6175. Mar. 7, 2014.<\/p>\n

Book<\/strong>
\nR\u00caGO, L. F. de M. O vale do S\u00e3o Francisco: Ensaio de monografia regional<\/strong>. S\u00e3o Paulo: Editora Renascen\u00e7a. 1945.<\/p>\n","protected":false},"excerpt":{"rendered":"Rivers alter landscapes by incorporating parts of other rivers","protected":false},"author":17,"featured_media":572917,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_exactmetrics_skip_tracking":false,"_exactmetrics_sitenote_active":false,"_exactmetrics_sitenote_note":"","_exactmetrics_sitenote_category":0,"footnotes":""},"categories":[159],"tags":[209,224,200,231],"coauthors":[5968],"class_list":["post-572904","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-science","tag-biology","tag-ecology","tag-environment","tag-evolution"],"acf":[],"_links":{"self":[{"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/posts\/572904","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/users\/17"}],"replies":[{"embeddable":true,"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/comments?post=572904"}],"version-history":[{"count":6,"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/posts\/572904\/revisions"}],"predecessor-version":[{"id":577167,"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/posts\/572904\/revisions\/577167"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/media\/572917"}],"wp:attachment":[{"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/media?parent=572904"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/categories?post=572904"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/tags?post=572904"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/revistapesquisa.fapesp.br\/en\/wp-json\/wp\/v2\/coauthors?post=572904"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}