From Plinthic Acrisols to Plinthosols and Gleysols: iron and groundwater dynamics in the tertiary sediments of the upper Amazon basin E. F RITSCH a,b , A. J. HERBILLON c , N. R. DO NASCIMENTO d,e , M. GRIMALDI f & A. J. MELFI e a IRD, Institut de Recherche pour le De ´veloppement, UMR 161 CEREGE, Europo ˆle Me ´diterrane ´en de l’Arbois BP 80, 13545 Aix en Provence cedex, France, b IMPMC, Institut de Mine ´ralogie et de Physique des Milieux Condense ´s, UMR CNRS 7590, Universite ´s Paris 6 et 7 and IPGP, Case 115, 4 Place Jussieu, 75252 Paris Cedex 05, France, c Unite ´ des Sciences du Sol, Place Croix du Sud, 2/10, B-1348 Louvain-la Neuve, Belgium, d DEPLAN/IGCE/UNESP, Instituto de Geocie ˆncias e Cie ˆncias Exatas, Rua 10, n°2527, 13500–230 Rio Claro, SP, Brazil, e NUPEGEL-USP, Nu ´cleo de Pesquisa em Geoquı ´mica da Litosfera da Universidade de Sa ˜o Paulo- Alameda das Sibipirunas 44, 13418–900 Piracicaba, SP, Brazil, and f Centre IRD d’Ile de France, UMR 137, 32 avenue Henri Varagnat, 93143 Bondy, France Summary Topography has been reported to be the major factor ruling the spatial distribution of Acrisols, Plinthosols and Gleysols on the seasonally flooded, low elevation plateaux of the upper Amazon basin occupied by Tertiary (Ic xa & Solimo˜es) sediments. In this study, detailed morphological and mineralogical investiga- tions conducted in a representative 25-ha site were combined with hydro-geochemical data to relate the vertical and lateral soil differentiations observed to the hydro-geological history of that part of the basin. As a result of the uplift of the Andes, several cuts in the extensive Tertiary marshlands have formed, at first, slightly incised plateaux of low elevation. There, weathering under hot and humid climates would have generated a reddish, freely drained and bioturbated topsoil layer and the vertical differentiation in subsoil sediments of a plinthite over an iron-depleted mottled clay. The second episode of soil differen- tiation is linked to the replacement of the forest by a savannah under the drier climates of the late Pleistocene, which favours surface runoff and the infill of the incisions by fine particles. This infill, combined with the return to the present humid climate, has then enabled the local groundwater to rise on the plateaux and to generate episaturation at the topsoil/subsoil transition close to the depressions. Nowadays, ferrous iron is released from the partly iron-depleted topsoil weathering front at high water levels during the rainy seasons. It moves from footslope to low-lying positions and from top to bottom in the soil profile according to the groundwater dynamics. The present general trend is thus to the lateral export of iron at high water levels due to subsurface and overland flows, its vertical transfer during the recession of the groundwater and accumulation in a nodular plinthite. In the latter, ferrous iron is adsorbed onto its softest iron masses where it feeds the neoformation of ferrihydrite that rapidly dehydrates into haematite. Introduction As shown on the small-scale soil maps that depict the geograph- ical distribution of the major soil groups recognized by the WRB classification system (Bridges et al., 1998), a large part of the Brazilian upper Amazon basin (i.e. the whole region between the Negro river in the north and the Madeira river in the southeast) is occupied by a soil association whose Acrisols, Plinthosols and Gleysols are the characteristic components. Previously, Sombroek & Camargo (1983) had already repor- ted that on these flat, low elevation plateaux that form the interfluves between all major rivers, topography is the soil for- mation factor ruling the spatial distribution of these three mapping units. These authors further noted that whatever their position in the landscape, their moisture regime and thus the name assigned to them, most of these soils are also charac- terized by the systematic occurrence of a particularly iron-rich subsurface soil horizon, now recognized as a plinthic horizon (FAO, 1998). As these findings summarize the observations reported on the numerous soil maps of the same region drawn Correspondence: E. Fritsch. E-mail: Emmanuel.Fritsch@impmc.jussieu.fr Received 3 March 2006; revised version accepted 2 October 2006 European Journal of Soil Science, October 2007, 58, 989–1006 doi: 10.1111/j.1365-2389.2006.00877.x # 2006 The Authors Journal compilation # 2006 British Society of Soil Science 989