Geodynamic Environment of Quaternary Morphostructures of the Subandean Foreland Basins of Peru and Bolivia: Characteristics and Study Methods

Home , Andean foreland basins, Beni River

GEODYNAMIC ENVIRONMENT OF QUATERNARY MORPHOSTRUCTURES OF THE SUBANDEAN FORELAND BASINS OF PERU AND BOLIVIA: CHARACTERISTICS AND STUDY METHODS

J.F. Dumont” and M. Fournier? ’ 1 *ORSTOM, and Colorado State University,Department of Earth Resources, ‘FortCollins, CO 80523, U.S.A. TORSTOM, 74 route d’Aulnay, 93143 Bondy, France .

The Quaternary evolution of westem Amazon basins is considered using morphostructural analysis of fluvial belts and lakes. Meandering rivers from the basin surface are analyzed using three scales of elements: the ridge and swale pattem, the mosaic pattem and the fluvial belt. Lake patterns are differentiated according to their shape, the hydrological environment and the structural context. Influences of basement structuresand neotectonic deformations, whenknown, are considered.The morphostructural patterns are discussed in the context of a balance of intemal (tectonic) and extemal (hydraulic and sedimentologic) processes. Special attention is focussed on the effect of neotectonics, which is generally neglected, on the formation and the location of Quaternary morphostructures. The adaptation of study methods and sampling to the characteristicsof the different morphostructures is considered.

INTRODUCTION covering about 25% of the westem Amazonian lowlands. They represent the areas of active subsidence related to Foreland tectonics and their control on surface flexural basins. .” deformation and deposition of Quaternary deposits in Flexural basins are generated by compressive tectonics in western Amazon are poorly understood, largely because few front of the eastern margin of the Andean range, and by direct elements and methods are available for the observation this overloading of the Brazilian Craton border, which is down- of such processes. The objective of paper is to present an warped as a result. The basins are filled by asymmetrical overview of the most characteristic lake and river patterns sedimentary prisms, thick on the range side and thinning observed on the-surface of subandean foreland basins, in toward the Brazilian Craton (Plafker, 1964; Pardo, 1982; order to determine how they can help to understand the Mégard, 1984). A bulge effect of the continental crust is Quaternary evolution of these regions. This paper will focus observed on the eastem margin of the basin (Lyon-Caen et the extended floodplains of the basins, which are on al., 1985). The formation and development of these basins is presently subsiding. related to the rising of the Andes as a result of the Quechuas We challenge here the idea commonly held that fluvial phases, during Mio-Pliocene time (Lavenu and Marocco, dynamics and avulsions act randomly on the surface of 1984; Sempere et al., 1990). Two main basins, the Marañón subsiding basins. Our demonstration is based on and the Beni Basins, are located respectively to the north and considerations of basement structures as well as neotectonic to the south of the Peru-Bolivia Andean segment (Fig. 1). and seismotectonic data from the basin margins. Climates, as Between these two large basins, in central and south Peru well as sedimentary and hydrological processes, are also (Ucayali and Madre de Dios Basin, respectively), the very important, and can play a‘major effect in some cases, foreland basins are narrow and elongated parallel to the but these external processes are not determinant in most range. In the center of the segment the Fitzcarrald Arch is a cases. main structural and interfluvial axis trending ENE-WSW (Oppenheim, 1975), which connects the Andes (to the west) to the Brazilian Craton (to the east) without the occurrence 1 GEOLOGICAL BACKGROUND ’ of any active basin. The Andean foreland basins are characterized by Up to 5000 m of late Cretaceous and Cenozoic sediments extended and flat Quaternary deposits. Three types of are deposited in the Marañón Basin (Seminario and Guizado, landscape are observed (i) very flat extended floodplains at 1976; Sanz, 1974), as well as in the Beni Basin (Sempere et elevations between 100 and 300 m; (ii) poorly dissected al., 1990). Unpublished seismic reflection data suggest that uplands; and (iii) fairly uplifted areas with hills up to 700 m the Upper Ucayali and Upper Madre de Dios basins have in elevation, called the ‘Montaña’ by geographers and early regihered lower amounts of subsidence. Before the travelers. Most developed in central and eastem Peru (Sierra Miocene, the area which extends from the Altiplano to the deMoa), the Montaña was formed during the Pliocene and west and the Iquitos Arch to the east was characterized by the Quaternary, when some partsI of the former and more deposition of the well known ‘red beds’ of late Cretaceous to I continuous subandean basins were uplifted (Fig. 1). The Eocene-Oligocene age (Mégard, 1984; Marocco, 1978). Marañón andBeni basins-ye th4 main extended floodplains These widespread deposits, which are not restricted to 129 WinSPIRS 2.1 Usage is subject to the terms and conditions of the subscription and License Agreement and the applicable Copyright and intellectual property protection as dictated by the appropriate laws of your country and/or International Convention. No. Records Request 1 210 FOURNIER* 2 159 FOURNIER* in AU 3 189 DUMONT" 4 110 DUMONT* in AU *5 3 #2 and #4 Record 1 of 1 - GeoRef Disc 4: 1/96-12/97 TI: Geodynamic environment of Quaternary morphostructures of the Subandean foreland basins of Peru and Bolivia; characteristics and study methods. AU: Dumont-J-F; Fournier-M œ> BK: In: Quaternary of South America. -3 BA: Iriondo-Martin (editor) SO: Quaternary International. 21; Pages 129-142. 1994. PB: Pergamon. Oxford, United Kingdom. 1994. PY: 1994 AN: 96-80881 J

J.F. Dumont and M. Foumier 130- - ,I

* FIG. 1. Structuralscheme of the Subandes of Peru and Bolivia. (1) Main areas of Quaternarydeposits. (2) Subandean zone of block tectonics , @I central Peru. (3) Subandean Thrust and Fold Belt. (4) Indications of uplift tendencies. (5) Indications of block or limited area of active , subsidence. (6) Ria lakes related to floodplain aggradation. (7) Ria lakes related to tilting. (8) Main reverse (a) or vertical @) faults. (9) ".% Estimated faults. . .. . ' foreland regions, have been deposited in relatively smaller Andes. Classically, concavity of a thrust belt (toward the tectonic basins, sÜrrounded by mountains of probably low foreland) increases shortening and the overloading effect, elevation (Lavenu and Marocco, 1984). Before the Miocene, and increases the subsidence as a result. On the contrary, foreland basins were probably not flexural basins as they'are foreland areas in front of a convex thrust belt are less presently, and the Andean drainage pattern at that time subsident, with periods of relative uplift. The northwest probably has little similitude with the present fluvial (concave) and southeast (convex) ends of the Peru-Bolivia network. A rough rising of the Andes during the Miocene in Andean segment are concerned here: the trend of the STFB relation to the formation of the Subandean Thrust and Fold changes from NNW-SSE in Peru to "E-SSW in south Belt (STFB) (Mégard, 1984) bracketed the previous Ecuador, and the concavity fronts the Pastaza depression intraforeland basins in a line of flexural basins located (West Marañón Basin). On the contrary, in south Bolivia the between the Andean Piedmont to the west and the bulge of STFB trend changes from NW-SE to NS, convex towards the overweight Brazilian Craton to the east. As a result, the the Amazonian plain, and there fronts the uplifted area of the drainage of the Subandes was collected along the Piedmont Bañados del Izozog and the Amazon-Parana basin in the flexural basins, finding an outlet eastward to the interfluve, with outcrops of the Brazilian Craton only some Atlantic Ocean, through the Brazilian Craton. . kilometers from the Andean Piedmont. The model of flexural basins suggests a development Regarding the second point (ii), it is now generally parallel to the mountain belt. In fact, subandean flexural accepted that 'flat' subduction segments, such as occur basins are rather discontinuous. In the Andes, three elements between 3" and 13"s (central Peru) and 27" and 33"s (central appear to determine important anomalies in the development Chile), determine a wide zone of foreland deformation and the location of the flexural foreland basins: (i) planar characterized by block tectonics (Jordan et aZ., 1983), and curvatures of the foreland belt, (ii) the geometry of the related to a stronger coupling between the subducted plate subducted plate beneath the Andes and (iü) the inheritance and the crust. As a result, the effect of overweight is weaker. of old structures, especially the Paleozoic ones. Such block tectonics have developed in central Peru since The first point (i) is relative to the main curvatures of the the Pliocene, leading to the formation of the present day ‘

’- Morphostructures of the Subandean Foreland Basins --131-

1: )per Ucayali Basin, located on the downwarped side of the Ucamara Depression, to the floodplain of the Tapiche River. < L :rra de Moa (Dumont et al., 1991). Extensive clay deposits observed away from the river belts The effect of old structures (iii) is probably the most may be explained, by these particular flows established important and also the most difficult to determine. The effect through the floodplain during higher water stages, in some of Paleozoic structures on the shape of the present structural cases more than 150 km before the normal confluence of the zones of the Andes has been recognized: the Abancay rivers. Because precise topography and hydraulic data are Deflection (Marocco, 1978), which determines eastward the lacking, hypotheses are speculative. Fitzcarrald Arch, the Iquitos Arch (Radambresil, 1977) and Internal processes are evidenced by active subsidence, the western (Grand Pajonal unit) and eastern (Sierra de Moa) especially during the Quaternary. More than 500 m of parts of the Upper Ucayali Basin (Martin and Paredes, 1977; Quaternary sediments have been deposited in the Marañón Benavides, oral commun., 1989). Other structures are less Basin (Seminario and Guizado, 1976; Sanz, 1974), and about known and their effect is more speculative, like the 800 m of late Tertiary and Quaternary sediments in the Beni Panafrican lineaments which extend westward to the Andes. Basin (Plafker, 1964). To give an idea of the importance of The Marañón Fault Zone of Laurent (1985), which has a the subsidence, note that we observe the same mean ratio of determining effect on the development of the south Marañón subsidence during the Quaternary (0.25 “/year) to the ratio Basin, is an example (Dumont, 1992). Another is the SW end of uplift of the Andes since the late Oligocene. The high rate of the Tapajos Fault (CERESIS, 1985), which determines a of subsidence, especially with respect to the output of the significant limit in the Beni Basin. In the remainder of this basin, means that internal processes cannot be neglected, and paper these structures will be mentioned if necessary, but it probably compensate more or less for the effect of fluvial is our opinion that most of the explanation of most of the aggradation. Large rivers on a very flat plain will probably structural anomalies must be attributed to inherited basement be very sensitive to any change in the position of the structure from Hercynian or older times. subsidence axis and depressed blocks. Internal and external processes will combine to give the present pattern of the floodplain, but we hypothesize that internal processes are BALANCE OF INTERNAL AND EXTERNAL more important and can expliin most of the special PROCESSES morphological pattern observed on the surface of subsiding External processes are dominated by a wet tropical areas. climate generating dynamic fluvial systems. At the exit of the subandean basins toward the Brazilian Craton, the upper IDENTDFICATION OF THE Amazon River in Iquitos and the Upper Madeira River in MORPHOSTRUCTURAL ELEMENTS Guajá Mirim are already among the largest rivers on earth (Gibbs, 1967; Guyot, 1992). These rivers contribute 15% of Meandering Rivers the total discharge of the Amazon River. Running down from Studies of fluvial patterns - active and relict -are very the eastern Andes these rivers are charged with traction and numerous. Nevertheless, from a practical point of view, the suspended sediments, and are called ‘white water rivers’. interpretation of samples collected from a river belt is not Sediment loss in the foreland basins is high: according to a easy, because several short term and local changes (like point detailed hydrological survey, about 60% of the sedimentary bar deposits) are combined with longer term changes such as charge of the Mamore River is trapped in the Beni Basin those due to climatic changes and tectonic activity. So, most (Guyot, 1992). authors remain in the field of general fluvial processes and A wet tropical climate in the Amazonian plain results in leave aside the questions dependent upon regional or specific extensive wetlands, swamps and lakes, fed by local influences. In fact, the problem of the evolution of a precipitation and drained by small ‘black water’ rivers, rich meandering river belt may be handled from a structuralpoint in organic acids. The surface of the foreland basins has a of view, and mapping the successive elements in space and mean slope lower than 10 cdkm. Swamps and lakes time. This work is easy and relatively accurate with satellite represent 3.4% of the surface of the Marañón Basin imagery like TM or Spot. (Kaliola et al., 1991) and 3% of the Beni Basin (Plafker, Three spaceithe related elements constitute the active 1964). fluvial belt (Fig. 2), from the smaller to the larger: third order The rainy season occurs in the foothills and in the (small) elements are represented by the ridge and swale flatlands at the same time, between December and May. morphology (scroll pattern) observed on Spot images as well thin Flooding of the flatlands by local precipitation can contain contrasted and curved lines; second order (intermediate) and limit water coming from the rivers and remote elements are represented by the mosaic pattern which precipitation in the Andean Piedmont. Thus, levees cannot represents areas of continuous growth of scroll pattern; and develop on the margins of the rivers, and only suspended the first order (large) element is represented by the belt of fines reach remote parts of the floodplain. Unfortunately, the recent occupation by the river. These elements will be hydraulics during floods is quite unknown, but may be presented in more detail in the same order. important for the understanding of the filling of the basin. For example, it is said by settlers in the south Marañón Basin Ridge and swale elements . that during extensive floods, overbank flows from the These are the most elementary components of fluvial Marañón and Ucayali Rivers tend to drain S-SE, through the morphology of meandering rivers (see Walker and Cant,

, 1 I

132 J.F. Dumont and M.Fournier RIDGE AND SWALE PATTERN , -Scale: up to 2OOm spaced and 3km long. cycle: 6 to 10 years. Influence: Climate, hydrology.

Interest: short term managing of river bank, navigation, minor risks. ! FLUVIAL MOSAIC

-Scale: elements of up to 50 km2.

Cycle: 40 years to 150 years. Influence: Climate,'hydrology, local tectonic. Interest: long term river bank managing,

1 major risks, placer deposits. MEANDER BELT -Scale: Tenth of kilometers wide, hundred of kilometers long.

Q&: 1to 3 thousand years. Innuence:sedimentary Neotectonics, processes. hydrology,

CL),, ,-I Interest: Land use planification and .- -@ recovering of materials, major risks. .. FIG. 2. Scale classificationof fluvial landforms in western Amazon.

ording to Nanson (1980), a single scroll bar generally result in neck or chute cut-off (during 198211983 fine sediments develops on a point bar plat- for the Beni River and 199111992 for the Mamore River, latively coarse sediments. With continuous according to P. Cholet, Expert for the COTESU, Swiss n the scroll bar grows, becoming a floodplain Technical Cooperation). On the contrary, it is probable that low to moderate levels during high water stages gently Large ridge and swale morphology in the western increase the deposition without drastically changing the Amazon is related to rivers with a high migration rate. Along position of river banks, allowing temporary settlements on the lower part of the Ucayali River, the scroll elements are the river banks. Change from a wet to a drier climate may 50-100 m apart, and up to 3 km long. Multistory image generate stability of the river channel due to underfit fluvial studies of the Ucayali River have shown a migration rate of conditions (Baker, 1978). These conditions result in paleosol about 40 dyear (1972-1988) from the Ucayali River in the occupation of the lower part of the previous scroll central part of the basin (Dumont, 1992), and of 60 &year topography (Dumont et al., 1992). - (1972-1975) from the lower Ucayali (Dumont, in press). Relatively smaller values of around 10 dyear have been Mosaic elements ' obtained from the Manu River in the Madre de Dios Basin An element of fluvial mosaic represents an area of by Rasänen et al. (1991a). continuous meander evolution, until a neck cut-off or an Some data dependent upon historical testimony exist on abandonment of the channel occurs. This is a timehpace the rate of construction and abandonment of the scroll related structural element (Fig. 2): an element crosses and pattern. Building and abandonment of an element occur with partly or completely erases the previous ones. The mosaic an average period of 4-6 years (Lamotte, 1990) in the lower elements from the lower Ucayali cover areas up to 20 km*. Ucayali River. A similar value is indicated by Fawcett Clearness of morphological structures, more affected by the (1953) from a survey of the rivers of the Peru-Bolivia erosion when older, may be compared with the structural border. Critical flow conditions are necessary to initiate a succession of the mosaic elements to define the successive new scroll bar construction, but it is unclear how exceptional stages of mosaic development. The map which is obtained floods contribute to the process. Catastrophic floods constitutes a basic document for the study of the structure of I

'. '. Morphostructuresof the SubandeanForeland Basins - -- (1) the present one, (2) 140-190 BP, (3) 340-520 BP, (4) 990-1310 BP and (5) 2100-2310 BP. The second period is differentiated from the present active development of meanders because it is preserved in concave river banks, and predates the present meander migration. Also, the present stage (1) is probably near its end, because two of the present meanders are close to being cut off (Fig. 3). The oldest stage (5) is in fact a complex one and includes at least two stages of mosaic construction. The oldest mosaic elements have not been sampled and dated directly, but their age is probably close to the age obtained for the floodplain outside of the mosaic belt, which is 2700 BP. Due to the limited number of available ages, the separation in successive periods remains relatively arbitrary. We must only consider this as a practical means of presentation, supported by the morphological aspect. Nevertheless, the relatively small number of successive meander constructions observed on the image (three to four stages before the present one) contrasts with the ages obtained, and is older than what would be expected from a regularly constructed mearider belt. The successive stages of mosaic development are FIG. 3. Mosaic pattem of the Ucayali River (see Fig. 4 for location). Age separated by periods of time of 265, 720 and 1055 years, of the mosaic elements: (1) ridges and swales of the present stage of mosaic increasing in length towards the past. It can hardly be argued construction; (2) 140-190 BP; (3) 340-520 BP; (4) 990-1310 BP; (5) BP. Numbers: points of collection of samples, age in years BP that the time span for the development of a mosaic element 2100-2310 10 (laboratory number in parentheses): 9 (850): 990 f50; (825): 2720 i40; was up to 10 times longer in the past. Two phenomena can 13 (832): 140i40; 16(834): 19Oi40; lS(863): 1310i40; 19(862):2100 be suggested to explain this phenomenon. The first is related f 40; 20 (848): 520 * 50; 21 (858): 340 i50; 22 (860): 2320 f 40. Dating to the lateral migration of the Ucayali River due to made in the ORSTOM Laboratory in Bondy. neotectonic faulting along the Brazilian Craton Border (Dumont et al., 1988). The result is a unidirectional the meander belt, and for determination of where sampling migration of the river channel leading to an asymmetrical of river bank deposits should be done. river belt, with abandoned river scarps and oxbows, all The rate of scroll bar construction obtained from concave toward the direction of migration (criterion defined historical relations suggests the fidl development of a mosaic by Von Bandat, 1962). It may be suggested that the element in a span of time ranging between 65 and 160 years. successive periods of preserved mosaic elements are related A comparative value is determined from historical reports to tectonic quiescence. Tectonic crisis reactivated river for the Amazon River in Iquitos. During Hispanic times, the mobility toward the downwarped margin of the floodplain. river has successively threatened to destroy the city or As suggested by Baker (1978), fluvial systems adjust to migrate to the opposite side of the river belt, leaving the climate in tropical regions. A dry period leads to a lower harbor without fluvial access, with a mean period of 60 years discharge of the river, a reduction of the mobility or even a (from 40 to 80 years) (Garcia and SHNA, 1987). According stability of the channel due to temporary underfitness. If a to Räsänen et al. (1991a), the Manu River (a small tributary wet period occurs again, the discharge increases and of the Madre de Dios River) has an average lateral migration reactivates the fluvial migration toward the more depressed rate of 12 &year, and a life expectancy of meander loops of areas determined by the downwarped edge of the faulted about 160 years. . zones. According to Servant et al. (1981), Holocene climate The map of mosaic succession has permitted a coherent period fluctuated with a periodicity of about 1000-2000 sampling of the river bank. Radiocarbon dating has given years in the southern part of the Beni Basin. Such fluctuation ages from 140 to 2720 BP (Fig. 3), which are consistent with has been registered in the south Marañón Basin (Dumont et the respective position of the samples with respect to the a¿., 1992). The ages which are available are not enough to mosaic structuralmap. Only one age is not consistent (no. 22, correlate climate fluctuations with the stages of mosaic Fig. 3), but is too old with respect to the relatively recent formation, but periodicity in both cases is similar in scale. position of the mosaic element from which it comes. 2 Reworking of fossil wood is possible, as observed by Fluvial belts Rasänen et al. (1991a). Nevertheless, from more than 20 Fluvial meander belts are the most prominent analyses performed in the area and published elsewhere, this morphological features on the surface of large subsiding is the only case of an inconsistent age. basins. Such pattems have been identified in the Beni Basin The ages obtained for the mosaic elements, combined (Plafker, 1964; Allenby, 1988; Dumont and Hanagarth, with the surface aspect (more or less erased morphology), 1993), the Marañón Basin (Dumont and Garcia, 1990; suggest the development of five successive major stages of Räsänen et aL, 1991b) and in the European Pannonic Basin mosaic construction, which are successively dated (Fig. 3): (Mike, 1975). According to Salo et al. (1986), the areas of 134 J.F. Dumont and M. Fournier

FIG. 4. Morphostructural scheme of the south Marañdn Basin (Ucamara Depression). (1) Subandean low foothills. (2) Uplands of the (4) Brazilian Craton. (3) Elongated lakes. Fluvial scars. (5) Basement faults. (6)Surface lineaments interpreted as faults. Anticline (7)and syncline (8) in the basement. (9) Statistical trends of rivers: (A) west of the depression; (I3) Ucayali and south of the depression; (C and D) central part superimposed on the Marañdn Fault Zone. (10) Basement faults reported to the Marañdn Fault Zone. (11) Fault lines and direction of extension reported to normal faulting in Quaternary sediments. (12) Paleozoic uplift of Santa Elena. recent erosion and deposition represent 26.6% of the Precise timing of successive river shifting is lacking, Peruvian lowlands and 12% of them are present floodplains. because sampling of abandoned meander belts is difficult. A common characteristic of these basins is the occurrence Data from the basin border suggest that no river belts are of both active (present) and inactive (fossil) meander belts. older than the late Pleistocene/early Holocene in the In most cases, the meander belt has not been totally Marañón Basin (Dumont, 1992). Three principal stages of abandoned, but reduced drainage remains with a narrow river belt shifting suggest there was a mean river belt channel. This phenomenon is known as an undefit pattern. occupancy of about 2500 BP. Data from the Great Hungarian Defined by Dury (1970), underfhess is related to a river Plain suggests shorter stages of about 1500 years (up to where geometry, especially sinuosity and wavelength, does seven stages during the Holocene, from Mike, 1975). not fit with its present channel width, according to the usual The widespread and rather unidirectional tendencies of parameters of a graded river. The inactive belts are the river shifting on the surface of large basins has suggested interpreted as previous courses of the active rivers. The the influence of large scale deformations of the surface of the observation of upstream and downstream links between the basin (Mike, 1975; Allenby, 1988; Dumont, 1992). portions of underfit channel and the remaining active Nevertheless, this aspect is probably less evident than the sections testify to the abandonment by shifting of the main effect of local structures. Before discussing the effect of trunk of the river (Dumont, 1992). Erosion progressively neotectonics on river shifting, we will present briefly the destroyed the scroll pattern, making it unclear. Thus, older case of the south Marañón Basin (Fig. 4). inactive meander belts are characterized by more unclear and The southeastern part of the Marañón Basin is a vast erased scroll patterns, and successive shifting of a river from floodplain of about 25,000 km2, which is common to the the oldest position to the present one may be deduced (Mike, Marañón, Ucayali and Tapiche white water rivers (called 1975; Dumont, 1992). Ucamara depression by Villarejo, 1988). The floodplain is t

Morphostructures of the Subandean Foreland Basins 135

FIG. 5. Neotectonic scheme of the central and northem Peruvian Andes. (1) Subandean Thrust and Fold Belt. (2) Subandean Tilted Block zone. (3) Uplands of the Brazilian Craton. (4) Ria lakes, enlarged in (A). (5) Elongated lakes. (6)Structural axis of the north Marañ611 Basin. Stereonets from focal mechanisms indicated by index number as the authors: AS9, from Assumpçao and Suarez (1988), A890504 from Assumpçao (1992) and SU5 and SU12 from Suarez et al. (1983).

drained by small rivers (Samiria and Pacaya Rivers basin to the other, and how tectonics are implicated remains principally) interpreted as abandoned courses of the controversial, and will be discussed later. Marañón and Ucayali rivers (Dumont and Garcia, 1990; Räsänen et al., 1991b; Dumont, 1992). Respective multi- Deltas and ‘LukeAmazonas’ staged shifting of the Marañón River to the north and of the Fossil delta patterns have been identified in the northern Ucayali River to the south resulted in a separation of about part of the Beni Basin by Campbell (1990), who interprets 100 km apart from a previous common river trunk. The them as sedimentary deposits from an Andean river entering Ucayali River, as well as its previous Samiria and Pacaya a huge ‘Lake Amazonas’ during late Quaternary (Campbell stages, are parallel to the main structural lines and faults et al., 1985; Frailey et al., 1988; Campbell, 1990). Lake (mainly of pre-Cretaceous age) in the basement (Fig. 4, nos Amazonas is a controversial hypothesis contested by 9 and 10) (Dumont, 1993). These faults, which belong to the Tuomisto et al. (1992), who believe that an extensive fluvial Marañón Fault Zone identified by Laurent (1985), are environment may have led to confusing lithostratigraphical supposed to have induced preferred flow direction on the correlations, and they suggest an alternative explanation basin surface. This is supported by neotectonic data showing based on tectonically controlled fluvial deposition. normal faults in the Quaternary fluvial deposits from the One of the strong arguments presented to support the upland, to the east of the basin (Fig. 4, no. 11) (Dumont et Lake Amazonas hypothesis is the occurrence of numerous al., 1988; Dumont, 1993). These faults trend N55E, parallel ‘bird’s foot’ fossil deltaS.in the northem margin of the Beni to the main portions of the Ucayali and the other underfit Basin (Campbell, 1990). In fact, all these deltas are not streams in the basin. This direction is also parallel to P-axes fossil, and the discovery of an active delta will give a more from focal mechanisms observed east and west of the realistic basis to discuss the sig”e of these Quaternary depression (Assumpçáo and Suarez, 1988; Assumpçáo, morphostructures. 1992) (Fig. 5, stereograms AS.9 and A 890504). Puinahua and Punga lakes are elongated lakes trending parallel to this Active delta from the Beni River direction, the second one (Punga lake) showing historical The occurrence of an active delta morphology has been subsidence (Dumont and Garcia, 1991; Dumont, 1993). discovered about 35 km north of Rurrenabaque (Piedmont The case of the south Marañón Basin shows that local border), on the western margin of the Beni River (Fig. 6). basement structures and neotectonic faults give a relatively The delta structure is about 5 km wide, invading a lake of good basis for explaining the main river trends. However, about 35 kmz. From the west, the lake is supplied by small these local phenomena give no explanation for the rivers of black or clear water draining the low Piedmont and unidirectional shifting of meander belts from one part of the the floodplain. To the east, it is connected to the Beni River 136 J.F. Dumont and M. Fournier

FIG. 6. Morphostructuralscheme of the northwest Beni Basin, from an October 25 1975 Landsat image. (1) Subandean Thrust and Fault Belt. (2) Lake invaded by a delta. The enlarged part (lower right) has been drawn from aerial photos taken in 1961 (note, the meander to off). the north has not yet been cut A-E are the successive positions of the Beni River, from the oldest to the recent stages. through the delta, which has been constructed from overbank Rurrenabaque and to the west of the Beni River, the upper outflow during upper water stages. It is clear that the fluvial terraces (probably of late Quaternary age) are cut by delta system is a lateral construction of the Beni River, north-south trending normal faults, with a downward located between the river belt and the margin of the movement of the eastern compartment (Fig. 6). The northern floodplain. It is not located at the mouth of a large river as is extension of the fault fits with the separation between the hypothesized for the fossil occurrences, and the relatively extensive floodplain to the east and upland to the west, along large size of the delta contrasts with the relatively small size the western side of the Beni River. Lakes and delta patterns of the lake. are located on the downwarped block (the exact fault The eastern margin of the Beni River is a wide floodplain position is not clear on the images), between the fault line which extends up to the Mamore River. This area is and the Beni River (Fig. 6). characterized by several underfit rivers and dry channel To the nokh of the area where lake and delta are segments, abandoned from previous courses of the Beni developed, river banks are low (less than 2 m). There, the River (Allenby, 1988; Dumont and Hanagarth, 1993). The forest which previously covered the river banks is dead, and course of the Beni River was northeast from Rurrenabaque, this phenomenon has been interpreted as an evidence of along the Tapado and probably (this is less clear) the increasing flooding due to active subsidence (Dumont et aZ., Yakuma Rivers (Fig. 6), from where it gradually swung 1991b). The sudden flooding of the forest occurred in 1983 westward onto its present northerly course. (informationfrom P. Cholet), and up to 1 m of sediments has The area located to the west of the Beni River (Pando) is been deposited since this time. The phenomenon probably a relative upland covered with forests and pampas. Near resulted from the convergence between active subsidence Morphostructures of the Subandean Foreland Basins 137

FIG. 7. Morphostructural scheme of the Beni Basin. (1) Rectangular lakes. (2) Ria lakes. (3) Direction of wind according to sand dunes (from Servant et d.,1981). (4) Reyes (R) Puerto Siles (PS) structural axis. A. (1) estimated faults; (2) rapids; (3) swamps. B: same ’ symbols as in A and on the main figure (left).

and exceptional floods due to the ENS0 (El Niño) fossil delta was active up to 150 km to the north of the phenomenon. Piedmont.

Fossil deltas Ria Lakes:‘ Where, why? Fossil deltas have been described from the Pando region, A ria lake is a clear water lake formed where the lower between the Beni and Madre de Dios Rivers, and far to the part of a tributary is ponded at the margin of the floodplain north in the Rio Branco region of Brazil (Campbell, 1990). of the trunk river by more rapid aggradation (Holz et al., These regions are uplands covered by pampas and forests, 1979). Defined originally in central Amazon as ‘ria fluvial’ with a low elevation (probably no more than some tens of by Gourou in 1949 (in Tricart, 1977), the term ‘ria lake’ is meters) over the surrounding floodplains.of the Beni and presently more commonly used. In central Brazil this type Madre de Dios Rivers. Underfit streams and reversed of lake is commonly related to postglacial eustasy (Irion, drainage along some delta arms postdate the formation of the 1984), but the influence of neotectonics on the main trend paleodeltas (Campbell, 1990). In all the cases, deltas are in of these lakes was observed early (Sternberg, 1950). A ria the position of lateral construction with respect to river belts, lake, formerly described in east Africa, and mentioned by and are driven towards a lake of very limited extension in Schumm (1977, 1986), resulted from a valley which was comparison to the scale of the deltas. Regarding the lake and flooded as a consequence of tilting. This constitutes a deltas described on the surface of the Pando Block, it is not second type that may be differentiated from the common sure whether a local fault (not observed) or a main tilting or type which results from a valley flooded by damming. curvature of the surface of the pampas may originate local Occurrences of the two types of ria lake have been observed subsidence and lakes where deltas were formed. in western Amazon. Ria lakes are frequent at the margin of Fossil deltas similar to the active one have been observed large rivers in the western Marañón Basin, in the central on the surface of the uplifted Pando Block, up to 200 km to Ucayali Basin (Rasänen, et al., 1987; Dumont, 1993), and the north (Fig. 1). In the light of the active delta we suggest in the south Beni Basin (Dumont, 1992; Dumont and Guyot, that fossil deltas characterize limited areas of active 1993). subsidence, before or during the onset of the uplift of the The structural axis of the western Marañón Basin is Pando Block. As a result, the area of active subsidence of the located 100-150 km away from the Piedmont, parallel to the north Beni Basin is probably more restricted now, limited to front of the subandean foothills, and following the same a narrow belt about 60 km wide, than it was before, when the curvature concave to the east (Sanz, 1974). An elongated 138 J.F. Dumont and M. Foumier cluster of ria lakes is superimposed on the axis of the basin, characterized by "E-WE late Paleozoic strike-slip faults, with only an exaggerated curvature (Fig. 5). which are related to the Marañón Fault Zone, a belt of high In the central Ucayali basin, ria lakes are located on the fault density. The NE orientationof the Puinahua lake is sub- western margin of the Subandean Tilted Block Zone of parallel to the 'en échelon' system of the Marañón Fault central Peru (STBZ, Fig. 5) (Dumont et al., 1991a). The Zone (Fig. 5). The Punga Lake is superimposed over the occurrence of a cluster of large ria lakes in specific areas of structure of the Santa Elena uplift (Dumont and Garcia, the foreland basins which are superimposed over axes of 1991), interpreted as a crystalline horst surrounded by active subsidence suggests a structural control for the Paleozoic sedimentary strata (Laurent, 1985). The NE development of ria lakes. elongation of the lake is parallel to a few of the structural Ria lakes in the Beni Basin are located on incised features northward and northwestward from the Santa Elena topography of the craton margin, east of the Mamore River high zone, according to Laurent (1985). The lakes trend (Fig. 7B). The alluvial plain which is responsible for the parallel to P-axis orientation of focal mechanisms observed damming of the tributaries is presently drained by small on both sides of the depression (Fig. 5, stereonet A.890504 streams which have no connection with the foothills. For this and AS.9). This trend is compatible with normal faulting reason, they have little sedimentary aggradation, and the showing a "W-SSE extension observed in the upland ponded lakes in this region are probably fossil. The fluvial border (Fig. 4, no. 11; Dumont et al., 1988). network of the area suggests that the Rio Grande, which has a high sedimentary charge, previously flowed through this Rectangular Lakes plain and is responsible for the alluvial aggradation which Hundreds of rectangular lakes with the same orientation ponded the tributaries. The shifting of the Rio Grande constitute the most striking pattern related to recent towards the west, away from the craton border (as well as a Quaternary activity on the surface of the Beni Basin, and related western shifting of the Beni River) possibly resulted until now their very origin is an enigma. The purpose here is due to the onset of a tectonic event and a rising of the bulge to present the most striking aspects of these lakes, and to on the craton margin. suggest that complex processes, both internal and external, Ria lakes related to tilting occur in the central eastern part probably combine in their formation. of the Beni Basin, north of Puerto Siles (Fig. 7A). This area Plafler (1964) described these lakes from aerial photos is away from the influence of sediment charged rivers. Some and flights. He gives a structural interpretation, based on of these lakes are derived fromrectangular lakes (see below). similar trends of the lake patterns and of the lineamentsin the Hard rock outcrops and rapids occur where the Mamore craton margin in Brazil. On the one hand, this interpretation, River crosses the area and, as a consequence, high which supposed that neotectonic blocks are as numerous as sedimentary aggradation cannot be involved here. Rapids the lakes, is hardly acceptable, for two main reasons: (i) there located on the upstream border of the area suggest the effect are a lot of small lakes, which implies tectonic blocks of only of block tectonics, as proposed by Allenby (1988), and a some hundreds of meters with a remote basement control slight tilting that has resulted in a flooded topography. Block through thousands of meters of unconsolidated alluviums, tectonics along a SW-NE axis have been suggested by and (ii) all the lakes, small and large, are very shallow, with Plafker (1964) and Allenby (1988). This structural axis is the depths less than 3 m. On the other hand, the structural southwestern extension of the Tapajo fault of central Brazil, influences cannot be totally dismissed, especially regarding suspected of neotectonic activity (CERESIS, 1985). the main fluvial trend and lake alignments like the Rurrenabaque-Puerto Siles-Tapajo fault line mentioned Elongated Lakes before. Recently, Plafker's hypothesis has been challenged We use the term elongated lakes in a descriptive sense, by a new hypothesis based on external processes, one related applied to lakes which are only approximately rectangular or to anthropic pre-Columbian agriculturaltechniques (Barbery exposing a more or less geometrical shape. Nevertheless, et al., 1991), and the other derived from fluvial morphology these lakes appear to be closely related to tectonics, like oxbows (Guyot et al., 1991, oral commun., 1992). according to the cases of the Puinahua and the Punga Lakes, The lakes are well delimited, rectangular with slightly observed in the south Marañón Basin. bent borders in detail (Figs 6 and 8A). Plafker (1964) and The Puinahua (1324 kmz) and Punga (341 €an2)Lakes are Allenby (1988) gave detailed statistical analyses of these both located in the south part of the Marañón Basin, called lakes: the cumulative length-azimuths of lineaments defined the Ucamara Depression (Villarejo, 1988). They are roughly by lake margins are N50E (long side) and N3 15E on the short rectangular, slightly arrow-shaped and wider on the foothill side (Fig. 8C). Lake margins are remarkably parallel all over side than on the craton side. Their border is not sharply the Beni Basin, which is about 500 km NS and WE. defined on satellite images, because of floating vegetation as The largest lakes (over 20 km long) are observed in a well as progressive change to swampy areas. backswamp position with respect to abandoned river belts These two lakes have been interpreted as the surface which follow the Reyes-Puerto Siles structural axis (Fig. 7). expression of tensional stress superimposed over reactivated Two lakes have been studied: the Laguna Suarez of basement structures due to the onset of Andean tectonics Trinidad and the Laguna Colorada of Santa Rosa. The (Dumont and Garcia, 1991). Morphological evidence comes characteristics of the two lakes are very similar. Their depth from the observation of flooded forest areas previously is about 1.3 m during the dry season, and the water level rises located out of reach of floods. According to Laurent (1985), less than 1 m during the rainy season. Before roads were the Ucamara Depression extends mainly over an area built, people used to cross the lakes by horse. The NW Morphostructures of the Subandean Foreland Basins 139

FIG. 8. Upper part (from Plafker, 1964): outline of oriented rectangular lakes from the Beni Basin. Dashed lines are dry lakes. Lower part: al. aerial photo from the Holocene sand dunes of the Santa CNZ area (location shown in Fig. 7); data from Servant et (1981), document from M. Servant. Inside (from Plafker, 1964): cumulative length-azimuth diagram of lineaments defined by lake and dry lake margins. U

140 J.F.Dumont and M. Fournier

margin of the Laguna Suarez is a swamp of small depth (0.7 processes, after aperiod of formation of incised morphology. m). The limit between the lake and the swamp is underlined The interpretation of rectangular lakes remains unclear, and by a flooded shallow area, and floating vegetation on the our opinion is that direct tectonic influences are not obvious, swamp side makes the limit visible. The basement around the while external processes such as drier paleoclimate lakes consists of a 2-3 m thick layer of indurated gray-green are pro6ably determinant. From a stratigraphic and clay, with occurrences of iron nodules. Because the bottom sedimentologicpoint of view, these rectangular lakes do not of the lakes is relatively hard, it may be supposed that the represent a good target (i.e. for the sampling of Quaternary layer of indurated clay extends below the lakes. North of deposits). These shallow lakes probably export more Santa Rosa this layer is eroded and weathered, giving over suspended sediments than they accumulate, and their bottom hundreds of kilometers a hard cap of iron concretions called of indurated clay can hardly be pierced. ‘cascarosa’. The local drainage of the Laguna Suarez Lake Basement structures and neotectonic faults control the comes from the swamp to the south and goes to the Ibare local trend of rivers, which is a confirmation of rather River to the north. During strong wind periods (which are classical ideas. In a particular case, asymmetric fluvial frequent in the region), the waves remove clay particles from mosaics and river belts reflect local deformations, generally the borders and the bottom of the lake, and we have observed the tilting of the floodplain. during a visit with M. Tejada (from SEMENA) that entering These considerations of lakes and rivers suggest that water is free of suspended sediment, although the water study methods must be adapted to the characteristics of the leaving the lake is charged with this sediment. This morphostructural pattern which is studied, and to the phenomenon is probably able to maintain and even increase objective. Especially, random sampling from lake bottoms the size of the lakes, by removing clay particles. Because of without consideration of the history of the lake may be the limited effect of waves in depth, the depths of lakes inefficient,just as in the case of river bank sampling without remain relatively constant and the bottom very flat. knowing the structure of the fluvial mosaic. Similar lakes in other regions have been related to the effect of wind (Carson and Hussey, 1962). Elongation of Evaluation of Large Scale Deformations: Are Fluvial lakes should be at right angles to the winddirections (Carson Shiftings Representative ? and Hussey, 1962; Livingstone, 1954). Plafker (1964) and Large scale river shifting is probably the most evident Allenby (1988) have dismissed the wind effect because of an fluvial process observed on the surface of a subsiding basin. .inaccurate correlation between the present direction of Several authors have interpreted the effect of active tectonics strong wind and lake orientation. Servant et al. (1981) have on river mobility on the basis of extended unidirectional described elongated sand dunes trending N240E in the area shifting in the context of a subsiding flexural basin (Mike, of Santa Cruz, to the SE of the area where rectangular lakes 1975; Allenby, 1988; Dumont, 1992). are developed. The dunes are related to a period of decayed Besides the academic question of what determines the forest and drier climate, which occurred after 3300 BP, and avulsion process, we think that the question is worth special probably between 3400 and 1400 BP (Servant et al., 1981). attention because the disposal of the successive fluvial belts These dunes trend exactly at right angles with respect to the reflects the Holocene sedimentary provinces and large scale long side of the lakes (Fig. 8). The cap of indurated clay on evolution during this period. The methods used for the study which lakes are developed probably also resulted from a of the fluvial mosaic are basically the same here, and drier climate. As suggested by Iriondo (pers. commun., successive deposition of fluvial belts can be considered as 1992), lakes may have developed on previous deflation spacehime related elements of Quaternary deposits, and areas. Relatively precarious hydrological and ecological constitute a basic document for the study of Quaternary conditions maintain the shape and the depth of these formations and deposits of the floodplain. depressions after they have been flooded due to the present wet climate. Sampling Methods and 14CDating Implications The last point and the conclusion will be addressed to DISCUSSION AND CONCLUSIONS sampling and interpretation that may result from 14C dating of samples which have been randomly taken from a river Evaluation of Local Tendencies bank, or not accurately located on the map. The case of the The different types of lake constitute a panel of Ucayali River shows that a range of 3000 years may be morphostructures to analyze local tendencies of the surface obtained from samples taken from the present river banks, of the basins. Elongated lakes are the most closely related to sometimes from two very close points. Only a previous study neotectonics. Nevertheless, it may be difficult to identify of the fluvial mosaic can help to point out the significanceof such cases of fossil pattern if the structural context, including the ages which are obtained and, with this precaution, we basement structures and syn-subsidence tectonics, is not have obtained relatively coherent 14Cdating. The question is considered. Ria lakes combine both subsidence or tilting and the same, or even more difficult, regarding the dating of contrasted hydrological conditions, with charged water on fossil river belts. Mapping from remote sensing images, the one side and clear water on the other (the effect of eustasy is successive river belts must precede any sampling operation, not considered here). Study of sedimentary deposits in a ria just as a structural map constitutes the framework of the lake may provide valuable information on the environment geological study of a region, which allows not only a as well as on the evolution of the subsidence, because the chronology, but also a relative position of the event in the formation of the lake postdates the onset of tectonic general history of the region. I i.

Morphostructuresof the Subandean Foreland Basins 141 ACKNOWLEDGEMENTS Dumont, J.F., Hérail G. and Guyot J.L. (1991b). Subsidencia, inestabilidad fluvial y reparticion de los placeres distales de oro. El caso del Río Beni This work was completed as part of the Cooperative Research (Bolivia). Symposium International sur les Gisements Alluviaux d’Ur, Programme between the Institut Français de Recherche Scientifiquepour le 3-5 June 1991, La Paz, Bolivie. Développement en Coopération (ORSTOM) and the Instituto Geofísico del Dumont, J.F., Garcia, F. and Foumier, M. (1992). Registros de cambios Peru (IGP, Lima), the Instituto para la Investigacidn de la Amazonia climaticos por los depositos y las morfologias fluviales en la Amazonia Peruana (IIAP, Iquitos), the Universidad Major San Andres (UMSA, La Occidental. In: Ortlied, L. and Machare, J. (eds), Proceedings of the Paz) and the Servicio para el Mejoramiento de la Navigacidn Amazonica International Symposium on Pale0 Enso Records, Lima, abstracts, pp. (SEMENA, Trinidad). We especially thank M. Iriondo for helpful 87-92. discussion and comments prior to the revision of the paper, and M. Servant Dury, G.H. (1970). General theory of meanderhg valleys and underfit for complementary data and comments. The field work has been facilited in streams. In: Dury, G.H. (ed.), River and River Terraces, pp. 264-275. Peru by H. Poupon, F. Fahn (ORSTOM), J. Beuzeville (IIAP), J. Lopez- Macmillan, London. Parodi (CIJH), and in Bolivia by G. Da Silva and M. Tejada (SEMENA), Fawcett, P.H. (1953). Expedition Fawceit. Ed. Hutchinson. and P. Chollet. Finally, we are indebted to E. Wohl for the correction of the Frailey, D., Lavina, E.L., Rancy, A. and Souza, J.P. (1988). A proposed English text. Pleistocene Holocene lake in the Amazon basin and its significance to Amazonian geology and biogeography. Acta Amazonica, 18(3-4), 119-143. REFERENCES Garcia, J.S. and SHNA (1987). El Rio que se aleja: Cambio del curso del Amazonas, historia y estudio tecnico. Centro de Estudios Teol6gicos de Allenby, R.J. (1988). Origin of rectangular and aligned lakes in the Beni la Amazonía, Iquitos, 57 pp. Basin of Bolivia. Tectonophysics, 145, 1-20. Gibbs, R.J. (1967). The geochemistry of the Amazon River System: Part I. Assumpçao, M. (1992). The regional intraplate stress field in South The factors that control the salinity and the composition and America. Journal of Geophysical Research, 97(B8), 11889-1 1903. concentration of suspended solids. Geological Society of America Assumpçao, M. and Suarez, G. (1988). Source mechanisms of moderate size Bulletin, 78, 1203-1232. earthquakes and stress orientation in mid-plate South America. Guyot, J.L. (1992). Hydrogéochimie des fleuves de l’Amazonie bolivienne. Geophysical Journal, 92,253-267. These de Doctorat d‘université en Géologie Géochimie, Université de Baker, V.R. (1978). Adjustment of fluvial system to climate and source Bordeaux I, 362 pp. terrain in tropical and subtropical environments. In: Miall, A.D. (ed.), Guyot, J.L., Corbin, D., Quintanilla, J. and Calle, H. (1991). Hydrochimie Fluvial Sedimentology, Canadian Society of Petroleum Geologists, des lacs dans la région de Trinidad (Amazonie Bolivienne). Influence Memoir no. 5, pp. 211-230. d’un fleuve Andin. Revue d’Hydrobiologìe Tropicaie, 24(1), 3-12. Barbery, E., Solanilla, V. and Barba, J. (1991). Proyecto Moxos. Holz, K.R., Baker, V.R., Sutton, S.M. and penteado-Orellana, M.M. (1979). Unpublished report from Universidad Autonoma Gabriel Rene Moreno South American river morphology and hydrology: NASA Special (Santa Cruz) and Universidad Technica del Beni (Trinidad), Bolivia, Publication 412, Apollo-Soyuz Test Project Summary Science Erport, 152 pp. Vol. II, pp. 545-594. Campbell, K.E. (1990). The geologic basis of biogeographic pattems in Irion, G. (1984). Sedimentation and sediments of Amazonian rivers Amazonia. In: Peters, G. and Hutterer, R. (eds), Vertebrates in the and evolution of the Amazonian landscape since PliocFne times. In: Tropics, pp. 3343. Museum Alexander Koenig, Bonn. Sioli, H. (ed.), The Amazon, Limnology and Landscape Ecology of a Campbell, K.E., Frailey, D. and Arellano, J.L. (1985). The geology of the Mighty Tropical River and its Basin, pp. 201-213. W. Junk, Rio Beni: Further evidence for Holocene flooding in Amazonia. Natural Dordrecht. History Museum of Los Angeles County, Contribution in Science, no. Jordan, T.E., Isachs, B.L., Allmendinger, R.W., Brewer, J.A., Ramos, V.A. 364,18 pp. and Ando, C.J. (1983). Andean tectonic related to geometry of Carson, C.E. and Hussey, K.M. (1962). The oriented lakes of Arctic Alaska. subducted Nazca plate. Bulletin of the Geological Society of Amerìca, Journal of Geology, 70,417-439. 94,341-361. CERESIS (1985). SISRA Project, Preliminary Neotectonic Map of South Kaliola, R., Puhakka, M., Salo, J., Tuomisto, H. and Ruokolainen, K. America. CERESIS, Lima. (1991). The dynamics, distribution and classification of swamp Dumont, J.F. (1992). Rasgos morfoestructurales de la llanura Amazdnica vegetation in Peruvian Amazonia. Ann. Bot. Fennici, 28,225-239. del Ped efecto de la Neotect6nica sobre los cambios fluviales y la Lamotte, S. (1990). Fluvial dynamics and succession in the lower Ucayali delimitaci611 de las provincias morfol6gicas. Bulletin de l’Institut River Basin, Peruvian Amazonia. Forest Ecology and Management, Français d’Etudes Andines, 21(3), 801-833. 33/34, 141-156. Dumont, J.F. (1993). Lake pattems as related to neotectonks in subsiding Laurent, H. (1985). El pre-Cretaceo en el oriente peruana: su distribuc6n y basins: The example of the Ucamara Depression, Peru. Tectonophysics, sus rasgos estructurales.Boletín de la Sociedad Geológica del Peni, 74, 222,69-78. 33-59. Dumont, J.F. (in press). The Upper Amazon River. In: Schumm and Laurent, H. and Pardo, A. (1975). Ensayo de interpretacion del basamento Winkley (eds), Engineering Problems Associated with the Natural del Noronente Peruano. Boletín de la Sociedad Geológica del PerÚ, 45, Variability of Large Alluvial Rivers. The American Society of Civil 25-48. Engineers. Lavenu, A. and Marocco, R. (1984). Continental sedimentation and Dumont, J.F. and Garcia, F. (1990). Neotectonics of the Andean foredeep tectonics in a fold belt related to a subduction zone: the Central Andes basin (Marañ611 basin) in north-eastem Peru. In: International (Peru-Bolivia) during the Tertiary. Bull. Centres Rech. Explor.-Prod. Symposium on Andean Geodynamics, pp. 107-110. Institut Français de Elf-Aquitaine, 8(1), 57-70. Recherche Scientifique pour le Développement en Coopération, Livingstone, D.A. (1954). On the orientation of lake basins. American ORSTOM, Paris. Journal of Science, 252,547-554. Dumont, J.F. and Garcia, F. (1991). Active subsidence controlled by Lyon-Caen, H., Molnar, P. and Suarez, G. (1985). Gravity anomalies and basement structures in the Marañ611 Basin of northeastem Peru. Land flexure of the Brazilian Shield beneath the Bolivian Andes. Earth and Subsidence, Proceedings of the Fourth International Symposium on Planetary Science Leiters, 7581-92. Land Subsidence, Intemational Association of Hydrological Sciences, Marocco, R. (1978). Géologie des Andes Péruviennes. MBmoire ORSTOM, no. 200, pp. 343-350. no. 94, 195 pp. Dumont, J.F. and Guyot, J.L. (1993). Ria lac: oú, pourquoi? Proceedings of Martin, C. and Paredes, J. (1977). Données nouvelles sur le Paléozoique du the Third International Conference on Geomorphology, 23-29 August Pérou Central. Comptes Rendus de l’Académie des Sciences de Paris, 1993, Hamilton, Canada. 284, 1647-1650. Dumont, J.F. and Hanagarth, W. (1993). River shifting and tectonics in the Mégard, F. (1984). The Andean orogenic period and its major structures in Beni Basin (Bolivia). Proceedings of the Third International central and northem Peru. Journal of the Geological Society, London, Conference on Geomorphology, 23-29 August 1993, Hamilton, 141,893-900. Canada. Mike, K. (1975). Utilization of the analysis of ancient river beds for the S. Dumont, J.F., Lamotte, and Foumier, M. (1988). Neotectonica del Arco detection of Holocene crustal movements. Tectonophysics, 29, de Iquitos (Jenaro Herrera, Peni). Boletín de la Sociedad Geológica del 359-368. Perú, 77.7-17. Nanson, G.C. (1980). Point bar and floodplain formation of the meandering Dumont, J.F., Deza, E. and Garcia, F. (1991a). Morphostructuralprovinces Beaton River, northeastem British Columbia, Canada. Sedimentology, and neotectonicsin Amazonian lowlands of Peru. South American Earth 27,3-29. Science, 4(4), 287-295. Oppenheim, V. (1975). The first (1944) geological exploration of the upper 142 J.F. Dumont and M. Foumier

Amazon valley in Peru. Boletín de la Sociedad Geológica del PerÚ, 45, Nororiente peruano. Boletín de la Sociedad Geológica del Pení, 44, 83-94. 106-127. Pardo, A. Caracteristicas estructuralesde la faja subandina del norte Schumm, S. (1977). The Fluvial System. John Wiley, New York, 337 pp. (1982). S. in del Peru. Symposium ‘Exploracion Petrolera en las Cuencas Schumm, (1986). Alluvial river response to active tectonics. Studies Subandinas de Venezuela, Colombia, Ecuador y Peru ’, Bogota. Geophysics, Active Tectonics, pp. 80-94. National Academy Press, Plafker, G. (1964). Oriented lake and lineaments of northeastem Bolivia. Washington, DC. Bulletin of the Geological Society of America, 75,503-522. Seminario, S.F. and Guizado, J.J. (1976). Sintesis bioestratigrafka de la Radambresil (1977). Levantamento de recursos naturais, Vol. 15, Folha region de la Selva del PerÚ. Secondo Congreso Latinoamericano de SB.19 JURUA. Ministerio das Minas e Energia, Projeto Radambresil. Geologia, Caracas, 1973, t.II, pp. 881-886. Räsänen, M.E., Salo, J.S. and Kaliola, R.J. (1987). Fluvial perturbance in Sempere, T., HBrail, G., Oller, J. and Bonhomme, M.G. (1990). Late the westem Amazon river basin: Regulation by long term sub-Andean Oligocene-early Miocene major tectonic crisis and related basins in tectonics. Science, 238,1398-1401. Bolivia. Geology, 18,946-949. R&änen, M.E., Salo, J.S., Jungnert, H. and Pittmann, R. (1990). Evolution Servant, M., Fontes, J.C., Rieu, M. and Salibge, J.F. (1981). Phases of Westem Amazon Lowland relief: Impact of Andean foreland climatiquesarides Holoche dans le sud-ouest de l’Amazonie. Comptes dynamics. Terra Nova, Rendus de l’Académie des Sciences de Paris, 292,1295-1297. Salo, 2,320-332. RLänen, M.E., J.S. and Jungnert, H. (1991a). Holocene floodplain Sternberg, H.O’.R. (1950). Vales tectbnicas na planicie amazonicas? lake sediments in the Amazon: ’*C dating and palaeoecological use. Revista Brasilera de Geografia, 4(12), 511-534. Quaternary Science Reviews, Suarez, G., Molnar, P. and BurcMield, B.C. (1983). Seismicity,fault plane Salo, 10,363-372. Räsänen, M.E., Neller, R., J. and Jungnert, H. (1991b). Recent and solution, depth of faulting, and active tectonics of the Andes of Peru, ancient fluvial depositional systems in the Andean forelands of the Ecuador and Southem Colombia. Journal of Geophysical Research, 88, Peruvian Amazon. In: History of the Fluvial and Alluvial Landscapes of 10403-10428. the Western Amazon Andean Forelands, Annales Universitatis Tricart, J. (1977). Types de lits fluviaux en Amazonie brbsilienne. Annales Turkuensis, Ser. A, no. part 3, pp. de Géographie, 473, 1-5 1. 75,Salo, 1-22. Räsänen, ME, Neller, R., J. and Jungnert, H. (1992). Recent and Tuomisto, H., Ruokolainen, K. and Salo, J. (1992). Lago Amazonas: Fact ancient fluvial depositional systems in the Andean forelands of the or fancy? Acta Amazonica, 22(3), 353-361. Peruvian Amazon. Geological Magazine, 129(3), 293-306. Viarejo, A. (1988; first edition 1943). Así es Ia Selva. Centro de Estudio Salo, J., Kalliola, R., Hwnen, I., Mwnen, I., Niemela, P., Puhakka, M. Teologicos de la Amazonia, Iquitos, 330 pp. and Coley, D. (1986). River dynamics and the diversity of Amazon Von Bandat, H.F. (1962). Aerogeology. Gulf, Houston, TX, 350 pp. lowlands forest. Nature, 322,254-258. Walker, R.G. and Cant, D.J. (1984). Sandy fluvial systems. In: Walker, R.G. Sanz, V.P. (1974). Geologia preliminar del area Tigre-Corrientes en el (ed.), Facies Models, pp. 71-89. Geoscience Canada, Reprint series 1.