Climatic Influences on Sedimentology & Geomorphology of the Rio Ramis Valley, Peru - JSR, 2005

7/25/2019 Climatic Influences on Sedimentology & Geomorphology of the Rio Ramis Valley, Peru - JSR, 2005

http://slidepdf.com/reader/full/climatic-influences-on-sedimentology-geomorphology-of-the-rio-ramis-valley 1/17

JOURNAL OF SEDIMENTARY RESEARCH, VOL. 75, NO. 1, JANUARY, 2005, P. 12–28Copyright 2005, SEPM (Society for Sedimentary Geology) 1527-1404/05/075-12/$03.00 DOI 10.2110/jsr.2005.003

CLIMATIC INFLUENCE ON SEDIMENTOLOGY AND GEOMORPHOLOGY OF THERIO RAMIS VALLEY, PERU

RENEE L. FARABAUGH1 AND CATHERINE A. RIGSBY2

1 Idaho State University, Department of Geoscience, Pocatello, Idaho 83201, U.S.A.2 East Carolina University, Department of Geology, Greenville, North Carolina 27858, U.S.A.

email: r [email protected]

ABSTRACT: Fluctuations in regional precipitation and in base-level(the level of Lake Titicaca) trigger changes in fluvial erosion and de-position in the largest tributary valley of Lake Titicaca, the Rio Ramis,Peru. Relationships between fluvial sedimentation and terrace forma-tion demonstrate that large-scale aggradational and downcuttingevents are associated with specific regional climatic events that havebeen previously documented in the Lake Titicaca basin and elsewhereon the Altiplano. Five laterally extensive terrace tracts (A through E)are present within the valley. Downcutting of the Ramis valley is cor-relative with base-level (Lake Titicaca level) fall caused by a decreasein precipitation. Aggradation is correlative with base-level rise duringincreasing precipitation. Sedimentary facies underlying the terraces re-cord deposition in a large paleolake followed by deposition of both

meandering and braided fluvial sediments. The highest terrace in thevalley—E—is underlain by a thick sequence of lacustrine strata thatwas deposited during an extremely wet period in the Lake Titicacaregion that terminated at about 40,000 cal yr BP. These lacustrinestrata were subsequently downcut, likely at a time of large drop inlake-level, resulting in a deeply incised valley. The main phase of fluvialaggradation in the valley (sediments underlying the D through B ter-races) was slow, spanning about 25,000 years. It occurred before, dur-ing, and after the Last Glacial Maximum (from before 35,140 790cal yr BP until about 10,421 181 cal yr BP), a time period that wasmostly characterized in the Lake Titicaca basin by high precipitationrates and high base-levels. The aggradation resulted in almost 40 m of sediment accumulation in the valley and was followed by rapid (about2,000 years), but episodic, incision into the underlying fluvial sedimentsto form the C and B terraces, as well as the erosional surface beneath

the modern fill terrace (A). This erosional phase corresponded to sev-eral periods of local base-level (Lake Titicaca) lowering between 10,241 181 and 1,668 76 cal yr BP. The modern fill terrace has aggradedabout 2 m above modern river level, likely within the last 1,500 years.

INTRODUCTION

The presence of high—commonly 25 m above modern river level—fluvial terraces in most of the tributaries to the Lake Titicaca (LT) basinindicates that landscape evolution in these watersheds is controlled primarilyby fluvial processes. Because riparian resources formed the basis for thesubsistence economy of the early inhabitants of the basin (Aldenderfer 1989,1998), and for many of the present-day inhabitants as well, understandingthe fluvial history of these watersheds is particularly important.

From the perspective of process sedimentology, the study of this high-altitude, semiarid, monsoon-dominated watershed leads to important newinsights into the geomorphic and sedimentary responses of a major riversystem to climate change and other forcings. Whereas the magnitude andduration of climate change that creates preserved changes in a fluvial sys-tem is often unclear (Vandenberghe and Maddy 2001), this study demon-strates the formation of terraces by aggradation and incision triggered byprecipitation-forced base-level fluctuation. This study particularly benefitsfrom the unusually well known late Quaternary history of climate and base-level in the Lake Titicaca watershed.

The most important variable in the fluvial history of the Ramis valley

appears to have been climate-triggered base-level change. Most workersagree that climate is important in the evolution of fluvial systems (Porteret al. 1992; Bridgland 2000; Schulte 2002; and many others), and it isgenerally accepted that major changes in the fluvial environment occurduring times of climatic transition (e.g., Knox 1972; Rose and Boardman1983; Bull 1991; Vandenberghe 1995; Bridgland 2000; Reneau 2000). Pre-vious studies in the Altiplano region of South America documented basin-scale climatic cycles of decadal to centennial to orbital timescales (e.g.,Melice and Roucou 1998; Thompson et al. 1998; Baker et al. 2001a; Bakeret al. 2001b) and their effects on Altiplano fluvial and lacustrine systems(e.g., Baucom and Rigsby 1999; Rigsby et al. 2003). Here we systemati-cally examine the relationships among the histories of precipitation, base-level, and fluvial aggradation versus incision in the Rio Ramis valley. We

suggest that large-scale aggradational and downcutting events are associ-ated with specific regional climatic episodes, specifically, that wetter pe-riods (times of higher precipitation and increasing lake-level) correspondto aggradational events and drier periods (times of lower precipitation andfalling lake-level) correspond to downcutting events.

BACKGROUND AND SETTING

In general, antecedent topography, lithologic variations (e.g., grain size),vegetative cover, and base-level exert primary control on the geomorphol-ogy and sedimentology of fluvial landscapes (Schumm 1993; Fuller et al.1998; Maddy et al. 2001; Jain and Tandon 2003). In turn, large-scale pro-cesses such as tectonics and climate control these variables (Bridgland2000; Nott et al. 2002; Starkel 2003; Vandenberghe 2003). The likelihoodthat neither active volcanism nor significant alterations in vegetative cover

(both of which could significantly alter sediment loads) greatly impactedthis part of the Titicaca watershed during the Holocene (Rigsby et al. 2003)simplifies the task of deducing the causes of landscape evolution in theRamis valley. The youngest macroscopic ash found in Lake Titicaca coresdates at 27,000 cal yr BP (Baker et al. 2001b), and there are no Quaternaryvolcanic centers in the watershed (Newell 1946). Likewise, pollen studiesfrom Lake Titicaca cores, and from cores to the north, indicate that byabout 13,700 cal yr BP, near-modern vegetation was established in thewatershed and subsequent changes during the Holocene were relativelyminor (Hansen et al. 1984; Paduano et al. 2003). Thus, it is unlikely thatsediment cohesion was greatly affected by vegetation changes. Further-more, there is no field evidence of active tectonics in the watershed, andthe Holocene tectonic history of the Altiplano region appears to have beenbenign enough (Clapperton 1993; Bills et al. 1994) that responses of theRamis valley to tectonic deformation are restricted to the addition of sed-iment load from the erosion of preexisting topographic highs or the con-finement of the channel due to antecedent bedrock constrictions.

In this cold, semiarid, high-altitude region, changes in precipitation andbase-level seem likely to dominate the forcing of fluvial and geomorphicprocesses. Variability of effective moisture (P-E) is recorded by changesin river discharge and changes in LT water levels (e.g., Abbott et al. 1997;Seltzer et al. 1998; Baker et al. 2001b). Such changes are generally re-corded in river valleys by the formation of terraces, caused by either down-cutting or aggradation. Rivers such as the Rio Ramis, which dischargedirectly into lacustrine basins, respond immediately to changing base (lake)levels and record their response to those changes by aggradation or down-



13SEDIMENTOLOGY AND GEOMORPHOLOGY OF THE RIO RAMIS VALLEY, PERU

FIG. 1.—Site map showing the subdivisions of the Central Andes Plateau: the Altiplano and thePuna (after Allmendinger et al. 1997). Thelocation of the Lake Titicaca basin is outlinedand enlarged to show the lake’s two subbasins(Lago Grande and Lago Huinamarca), the majortributaries to the lake, the Rio Desaguadero, andthe Rio Ramis study area in southwestern Peru.

cutting in the river valley. The result is the formation of traceable anddatable terrace tracts. We are at a great advantage studying the fluvialresponse to climate history and base-level in this region, relative to manyother fluvial geomorphic studies, because we know the detailed late Qua-ternary history of both climate (especially effective moisture) and base-level (the level of Lake Titicaca) from previously published paleoclimaticresearch on the Lake Titicaca watershed.

Geologic Setting

The Rio Ramis is the largest tributary to LT in the northern Altiplano.The Altiplano is part of the Central Andes plateau, which consists of twoprovinces—the Altiplano to the north and the Puna to the south (Allmen-dinger et al. 1997). The Titicaca basin (Fig. 1) lies within the northernportion of the Altiplano, straddling the Bolivia–Peru international bound-ary, and is bordered by the Cordilleras Vilcanota and Carabaya to the north,the Cordillera Oriental (including the Cordilleras Aricoma, Apolobamba,and Real) to the east, and the Cordillera Occidental, the active volcanicarc, to the west. The surface of LT is approximately 3,810 m above sealevel. The lake, which is subdivided into two main basins (Lago Grandein the north and Lago Huinaimarca in the south), is about 175 km long by50 km wide (Wirrmann 1992). The modern lake is a nearly closed basinthat loses most of its water (91%) by evaporation (Roche et al. 1992). Theremainder is discharged southward via the Rio Desaguadero, which flowsto the shallow, saline Lago Poopo. Below Lago Poopo is the large saltpan,the Salar de Uyuni.

The lake’s catchment area spans the region from 140906 to 170829south latitude and 680334 to 710142 west longitude (Wirrmann1992). The drainage network contributing to the LT basin contains six

major rivers: Rio Ramis, Rio Ilave, Rio Coata, Rio Catari, Rio Huancane,and Rio Suchez (Fig. 1). Of the six rivers, the Rio Ramis accounts for themost discharge, with a monthly average of 160 m3s–1 (data from ServicioNacional de Meteorologıa e Hidrologıa del Peru). The river drains snow-fields in the Cordillera Occidental from elevations up to 6000 m and hasa main watercourse that is 283 km long.

Climatic Setting

The modern Altiplano has a high-altitude semiarid climate, with meanannual temperatures between 7 and 10C (Roche et al. 1992) and puna(grassland or shrubland) vegetation (Paduano et al. 2003). Average annual

precipitation in the LT watershed is about 800 mm; about three-quartersof this takes place in the December–January–February wet season (Rocheet al. 1992). During these austral summer months, an upper-troposphereanticyclone, the Bolivian High, forms over the Altiplano, causing low-leveleasterly winds to strengthen and precipitation to increase on the Altiplano(Lenters and Cook 1997; Garreaud et al. 2003), part of the South Americansummer monsoon (SASM; Zhou and Lau 1998). Interannual to decadalvariability of precipitation on the Altiplano, caused by variability in theintensity of the SASM, has been ascribed to various external controls in-cluding ENSO (Aceituno and Montecinos 1993; Vuille et al. 2000; Gar-reaud et al. 2003) and tropical Atlantic variability (Baker et al. 2001a;Baker et al. 2001b). On longer (orbital) timescales, it has been suggestedthat variability of Altiplano precipitation is tied to SASM intensity (Bakeret al. 2001a; Baker et al. 2001b), Pacific sea-surface temperatures (Bradley

et al. 2003), or the variability of tropical Atlantic sea-surface temperaturegradients (Baker et al. 2001a; Baker et al. 2001b).Modern instrumental records demonstrate that fluvial discharge in the

Titicaca basin responds rapidly to the summer precipitation peak and thatlake-level rises rapidly in response to the seasonal discharge peaks (Fig.2). The 10-year interval shown in Figure 2 illustrates the relationship be-tween Ramis discharge, Juliaca (just south of the Ramis valley) precipita-tion, and the level of Lake Titicaca. An increase in rainfall on the Altiplanoresults in increased discharge to the lake from the Ramis and in increasedlake-level. During the subsequent dry season, a gradual decline in lake-level is attributed to decreased riverine input and continual high rates of evaporation. The El Nino event of 1982–1983, for example, witnessed verylow precipitation, little river discharge, and a large drop in lake-level. Thelake highstand (about 4 m above the long-term average) resulted fromsuccessive wetter than average years including 1986, which correspondedwith a well-developed La Nina event.

Annual and decadal lake-level fluctuations are also related to changingprecipitation on the Altiplano. Paleoclimatic records, including lake-levelhistories from LT (Roche et al. 1992; Abbott et al. 1997; Baker et al.2001b) and the Salar de Uyuni (Baker et al. 2001a), ice cores from thenearby mountain glaciers of Sajama (Thompson et al. 1998), Illimani(Hoffmann et al. 2003), and Quelccaya (Thompson et al. 1986; Melice andRoucou 1998), and fluvial histories of the Rio Ilave (Rigsby et al. 2003)and Rio Desaguadero (Baucom and Rigsby 1999) valleys, demonstrate thatprecipitation variability affected lake level and fluvial discharge on muchlonger timescales as well. Sediment cores from LT (Baker et al. 2001b)



14 R.L. FARABAUGH AND C.A. RIGSBY

FIG. 2.—Lake-level change, Ramis discharge,and precipitation (from Juliaca). Lake Titicacalevel has been recorded in Puno, Peru, since1914. Interannual wet and dry periods are alsorecorded. All data are from Servicio Nacional deMeteorologıa e Hidrologıa del Peru(SENAMHI).

and the Salar de Uyuni (Baker et al. 2001a) record changes in effectivemoisture on multi-centennial to orbital timescales that are apparently cor-

relative to changes in tropical Atlantic sea-surface temperature gradients(Baker et al. 2001a; Baker et al. 2001b). Similarly, data from the Quelccayaice core suggests that decadal wet periods correspond to higher levels of LT and to decadal variability in tropical Atlantic sea-surface temperatures(Melice and Roucou 1998). In the Rio Desaguadero and Ilave valleys, themulticentennial- to orbital-scale fluctuations between wet and dry episodesare recorded by changing fluvial and lacustrine landscapes (Baucom andRigsby 1999; Rigsby et al. 2003).

All of these studies document basin-wide correlative changes during extremewet and dry events on the Altiplano. LT was much higher and large paleolakesexisted on the central Altiplano during the last glacial maximum (LGM). Adrill core from the Salar de Uyuni (Baker et al. 2001a; Fritz et al. 2004)documents three major paleolake phases—Coipasa, Tauca, and Minchin, withages of 11,500–13,400 cal yr BP, 16,000–25,000 cal yr BP, and 36,000–46,000cal yr BP, respectively (Fritz et al. 2004). These central Altiplano highstandsalso correspond to high lake-levels in LT (Baker et al. 2001b) and depleted 18Oice values in the Sajama and Illimani ice cores (Thompson et al. 1998;Hoffmann et al. 2003). At these times of higher precipitation, sediments of theRio Desaguadero valley document an overflowing LT and widespread lacus-trine sedimentation (Baucom and Rigsby 1999; Rollins 2001; Warren 2002).In addition, much of the sedimentation and formation of depositional terracesin the Rio Ilave valley may also have taken place during these intervals of high precipitation (Rigsby et al. 2003).

Morphology of the Modern Rio Ramis

The longitudinal profile of the Rio Ramis valley drops about 20 m inelevation over about 60 km distance (Fig. 3). The gradient gently decreasesdownstream. Locally, the profile varies between bedrock constrictions,commonly flattening upstream of constrictions and steepening rapidly with-in the constrictions. Figure 3 illustrates that the presence and absence of these topographic highs near the modern river is controlled by the occur-rence of Devonian bedrock strata (the Cabanillas Formation). Transitionsin river systems are commonly caused by increased sediment supply andchanges in channel gradient (Leopold and Wolman 1957; Schumm andKhan 1972; Smith and Smith 1984). Typically, the gradient in the Ramisvalley is about 0.00047. The slope between the Paccha Baja and Lulicuncaconstrictions is gentle and has a value of 0.00025. The slope steepens toabout 0.000374 between the Lulicunca and San Sebastian bedrock constric-tions because the constrictions are close together (Fig. 3). It is evident thatboth the longitudinal profile and the fluvial planform are related to prox-

imity to constrictions. Upstream, where the valley slope is higher, the flu-vial sediments are coarser-grained, typically either braided-river sediments

or coarse-grained meandering-river sediment. Downstream of the constric-tions, the profile flattens again and the river transitions to dominantly finer-grained meandering-river planforms and sediment types.

Overall, the modern Rio Ramis valley is transitional between a mean-dering and braided fluvial system. Although the sinuosity index (2.0 fromPuyutira to Cacochico, 1.6 from Cacochico to the confluence with the RioAzangaro, and 1.6 between this confluence and the lake) reflects the dom-inant meandering planform, multi-channel reaches are also evident in sev-eral locations (Fig. 4A). In general, the valley is most constricted to thenorth of Puyutira and, although it has a meandering planform in this region,it lacks wide terraces (Fig. 4B), oxbow lakes, and wetlands. The centralreach of the river valley in the study area has a typical meandering channelpattern with wide terraced surfaces (Fig. 4C), abundant oxbow lakes, andfloodplain wetlands. Below the confluence with the Rio Azangaro, sedimentload increases, the valley is less constricted, the gradient decreases (Fig.3), and the river develops a multi-channel planform (Fig. 4A). Within thevalley, differences in channel type correspond to differences in topographyand sediment load. Where bedrock constrictions occur adjacent to the mainchannel (Figs. 3, 4D), coarse-grained material is transported downstreamand piled up behind subsequent bedrock constrictions. In addition, whereabundant alluvium is being supplied to the river from adjacent topographichighs—at Pucara, for example—braiding occurs downstream of the high.

METHODOLOGY

Our interpretation of fluvial sequences in the Ramis valley is based ondetailed field analysis of the sedimentology and geomorphology of the val-ley. Terraces and modern cutbank exposures were surveyed, using a GPS(error 5 m) and a total station, to create the longitudinal profile (Fig. 3)and seventeen river-perpendicular profiles. Analysis of these profiles al-lowed terrace classification and a determination of the sequence of terraceformation. Sediments beneath the terrace surfaces were measured and de-scribed using standard sedimentological methods of facies analysis to gen-erate fifteen stratigraphic sections (Fig. 3). The focus of these descriptionswas on lateral and vertical facies changes as indicated by variations in grainsize, sedimentary structures, bedding morphology, and the nature of stratalcontacts. From these data, we were able to differentiate between majorfluvial depositional environments preserved in the terrace sediments, toidentify smaller-scale environments (such as channel, point-bar, crevassesplay, wetland, and floodplain) in some locations, and to reconstruct thedepositional history of the river valley.




FIG. 3.—Geologic map and longitudinal profileof the study area in the Rio Ramis valley. Themap shows locations of transverse profiles(dotted lines), profiles near bedrock constrictions(dashed lines), and stratigraphic sections (filledcircles). The longitudinal profile displaysdistance as measured in meters downstream fromthe first reference point. Gray lines show theaverage GPS elevation (diamonds) and elevationerror ( 5 m; triangles). Geologic units fromNewell (1946).

Small amounts of datable material, such as roots and soil carbonate, arepresent in the slightly sandy fine-grained lithofacies. Samples were sub-mitted for AMS radiocarbon analysis to the National Ocean Sciences Ac-celerated Mass Spectrometer (NOSAMS) Facility, Woods Hole Oceano-graphic Institution. The resulting 14C dates were converted to calendaryears before present (cal yr BP) using Calib version 4.2 (Stuiver et al.1998) for dates 20,000 14C years and CalPal: The Koln RadiocarbonCalibration and Paleoclimate Research Package (Weninger et al. 2002) fordates 20,000 14C (Table 1). The dates used in the text are the 2-sigmacalibrated ages.

Because of the dynamic nature of fluvial systems, we realize the limi-tations of relying on 14C dates on sediments and paleosols for terrace cor-relations. Nevertheless, as in similar studies of this type, the use of suchdates is valid, if done carefully. The terraces in the Ramis valley occur as

extensive, visually traceable surfaces throughout the valley (Fig. 5). Ourinterpretations of the terraces (and our use of the radiocarbon dates forthose interpretations) are based upon the physical correlation of key visu-ally traceable terrace and stratal surfaces. In several locations, especiallywhere physical correlations are unavailable, it is possible that dated samplesreflect reworked fluvial material from remnant terraces within the valley.This is taken into consideration when using the dates to relate sedimento-logic and geomorphic changes to LT climatic base-level fluctuations.

SEDIMENTOLOGY

The terraced strata of the Rio Ramis valley contain 13 distinct lithofacies(Table 2), grouped into three facies associations (FA) that record depositionin meandering-river (FA1), braided-river (FA2), and lacustrine (FA3) en-




FIG. 4.—A) The modern Rio Ramis valley exhibits both braided and meandering morphologies. Vegetated islands, typical of multi-channel flow, as well as point-bardeposits, typical of meandering systems, are present. Within both braided and meandering morphologies, valley width affects terrace morphology. B) North of the Puyutira

measured section (Fig. 3), for example, the valley is constrained by mountains, terrace formation is limited, and only the A through C surfaces are evident. C) In thecentral reach of the study area, the valley widens and terraces form vast, high surfaces. D) Locally, outcropping Devonian shale and quartzite form bedrock constrictionsthrough which the river channel is narrow and terraces are poorly developed.

vironments (Fig. 6, Table 3). Specific sediment types and sedimentarystructures characterize each lithofacies, and distinct assemblages and ver-tical sequences of lithofacies characterize each facies association (Table 3,Fig. 6).

Both the meandering-river and braided-river facies associations are char-acterized by fining-upward sequences (Fig. 6). The meandering-river strata(FA1) contain meter-thick, fine-grained units and well-developed, laterallydiscontinuous, fining-upward sequences (Fig. 7A). The braided-river strata(FA2) are coarser and have less well-developed, laterally continuous, fin-ing-upward sequences (Fig. 7B). The lacustrine strata (FA3) comprise lo-cally fossiliferous, thinly laminated silt and clay deposits that unconform-ably underlie the fluvial strata.

Although both the meandering-river and braided-river deposits preservedin the terraced strata of the Rio Ramis valley have modern analogs in theadjacent river valleys (allowing convenient, detailed comparisons), someuncertainty exists in the distinction between braided and meandering end-member fluvial types. This is the result of the complexity of both themodern and the ancient fluvial systems and suggests that, although bothbraided-river and meandering-river strata are preserved in the Rio Ramissequences, at least some of the lithofacies described here may be the prod-uct of non-end-member systems. Indeed, the modern Rio Ramis valleyexhibits transitional reaches characterized by well developed braid bars in

tightly meandering reaches. The scarcity of geomorphically distinct lateral-accretion deposits also makes the definitive identification of meanderingsystems difficult. Point-bar deposits are rare in the modern river system aswell—even in clearly meandering reaches that contain no braid bars, mostpoint-bars are small and poorly developed. Such complexities are commonin arid-region river systems. Within the obvious and common limitationsassociated with such complexity, we are confident in our interpretations.

Meandering-River Systems (FA1)

Meandering-river deposits constitute the most abundant strata in the RioRamis valley. They are present in all reaches of the river valley, exceptPucara, and contain sediments of both coarse- and fine-grained lithofacies(Table 3). Vertically well-developed, laterally discontinuous, fining-upwardstratigraphic sections of the meandering-river strata (Figs. 6A, 7A) containcross-bedded gravel or sand at the base of sequences that fine upward tothick accumulations of silt and clay. These fining-upward sequences sug-gest sediment accumulation typical of meandering channel formation,point-bar migration, and floodplain accretion.

Description of Lithofacies.—Strata of the meandering-river facies as-sociations in the Rio Ramis valley comprise well developed fining-upwardsequences that typically overlie sharp, scoured basal contacts. Sand or grav-




T A B L E 1 . — S a m p l e n u m b e r s ,

l o c a t i o n n a m e ,

t e r r a c e n u m b e r , r a d i o c a r b o n a n d 2 - s

i g m a c a l i b r a t e d a g

e s f o r a l l t h e s a m p l e s u s e d i n t h e s t u d y .

S a m p l e N a m e

A M S L a b I D #

M a t e r i a l D a t e d

1 3 C

V a l u e s ( ‰ )

R a d i o c a r b o n A g e

( 1 4 C y r B P )

2 - S i g m a c a l i b r a t e d A g e

( c a l y r B P )

L o c a t i o n

C c a l u y o 4

O S - 3 7 1 3 6

B u l k o r g a n i c m a t e r i a l i n c l a y

2 3 . 8

1

1 , 5 8 0

5 5

1 , 4 5 7

1 1 2

A b o u t 0 . 5 m

a b o v e u n c o n f o r m i t y s u r f a c e i n t e r -

r a c e A

C a l a p u j a 2 - 1

O S - 3 7 1 3 7


2 3 . 7

7

1 , 7 7 0

4 0

1 , 6 6 8

7 6

F r o m t e r r a c e

A

I q u i l o C - 1

O S - 3 7 4 7 3

B u l k o r g a n i c m a t e r i a l i n p a l e o s o l

2 3 . 3

4

7 , 2 2 0

5 5

8 , 0 2 6

8 7

F r o m b o t t o m

o f p a l e o s o l u n i t i n t e r r a c e E s e q u e n c e

Y u c a j a c h e 2

O S - 3 6 9 3 8

C a r b o n a t e m a t e r i a l i n c l a y

6 . 4 5

7 , 5 2 0

3 0

8 , 3 3 9

5 0

C l a y - s i z e c a r b o n a t e d e p o s i t n e a r u n c o n f o r m i t y s u r -

f a c e i n t e r r

a c e D

S a l s a p a t a 5

O S - 3 6 9 3 9

C a r b o n a t e m a t e r i a l i n p a l e o s o l

7 . 6 1

8 , 2 0 0

3 0

9 , 0 9 5

6 6

F r o m p a l e o s o

l p r e s e n t a t 7 m i n S a l s a p a t a m e a -

s u r e d s e c t i o n

Y u c a j a c h e 3

O S - 3 7 4 7 4

B u l k o r g a n i c m a t e r i a l i n s i l t

2 3 . 5

5

9 , 2 9 0

8 0

1 0 , 4

2 1

1 8 1

F r o m fl u v i a l s e d i m e n t a b o v e u n c o n f o r m i t y s u r f a c e

i n t e r r a c e D

C a m a r a j a 2

O S - 3 6 9 3 4

B u l k o r g a n i c m a t e r i a l i n p a l e o s o l

2 . 4 6

1 5 , 4

5 0

4 0

1 8 , 4

9 7

5 4 9


l p r e s e n t a t 1 0 m i n C a m a r a j a m e a -


C o l q u e 2

O S - 3 7 2 9 5


( s m a l l s a m p l e )

2 1 . 1

5

2 0 , 4

0 0

4 3 0

2 1 , 9

7 0

5 0 0 *


D s e d i m e n t s i n C o l q u e t r i b u t a r y s e c -

t i o n

P u y u t i r a 8

O S - 3 6 9 4 0


6 . 1 1

2 9 , 9

0 0

1 5 0

3 1 , 5

1 0

5 1 0 *


l p r e s e n t a t 1 1 m i n P u y u t i r a m e a -


C a m a r a j a 1

O S - 3 6 9 3 6


1 . 2

3 2 , 8

0 0

1 4 9

3 5 , 1

4 0

7 9 0 *


l p r e s e n t a t 7 m m i n C a m a r a j a m e a -


I q u i l o A - 2

O S - 3 6 9 3 7

C a r b o n a t e m a t e r i a l i n c l a y

6 . 5

3 5 , 1

0 0

3 4 0

3 7 , 7

0 0

6 0 0 *


D i n s e d i m e n t s d i r e c t l y o v e r l y i n g

l a k e s e d i m e n t s

C c a l u y o 1

O S - 3 7 1 3 5


2 4 . 7

3 9 , 5

0 0

6 1 0

4 0 , 1

7 0

4 3 0 *

F r o m l a c u s t r i n e c l a y s b e l o w t h e u n c o n f o r m i t y s u r -

f a c e

P u y u t i r a 2

O S - 3 7 1 3 2

P l a n t / W o o d m a t e r i a l i n s e d i m e n t

2 6 . 2

1

4 8 , 0

0 0

1 , 4 0 0

4 7 , 6

0 0

1 , 7 4 0 *

R e w o r k e d fl u

v i a l s e d i m e n t s a t 2 m i n P u y u t i r a

m e a s u r e d s

e c t i o n

S a m p l e s a r e l i s t e d f r o m y o u n g e s t t o o l d e s t . D a t e s m a r k e d w i t h * d e n o t e c a l i b r a t i o n u s i n g C a l P a l ( W e n i n g e r e t a

l . 2 0 0 2 ) . A l l o t h e r s c a l i b r a t e d u s i n g C a l i b ( S t u i v e r e t a l .

1 9 9 8 ) .

el (Sp, Smg, Fsil, Gp, or Gt,) fines upward to ripple-bedded silt and finesand (FsSr), then to meter-thick, locally laminated clay deposits (Fsmtmand FORl). Both coarse- and fine-grained sequences are present. Large-scale planar and small-scale trough cross-bedded gravels (Gp and Gt) atthe base of fining-upward units characterize coarse-grained sequences.Fine-grained sequences of this facies association typically have basal unitsof small-scale ( 1 m) planar cross-bedded (Sp; Fig. 8A) to massive (Smg)sand, interbedded clay and sand (FSil), or ripple-bedded silt and sand (FsSr;

Fig. 8B). Both coarse- and fine-grained sequences are capped by massive(Fsmtm; Fig. 8C) to laminated (FORl; Fig. 8D) clay and silt.

In this facies association, the large-scale planar cross-bedding (Gp and Sp)at the bases of coarse fining-upward sequences are typically from 10 cm to 1m in amplitude and contain interbedded, decimeter-thick layers of sand andgravel. The Gp or Gt (Fig. 8E) deposits typically fine upward to large- and/orsmall-scale Sp (Fig. 8A), to FsSr (Fig. 8B), and finally, to Fsmtm (Fig. 8C)and FORl (Fig. 8D). In systems dominated by fine sediments, small-scale Spgenerally begins the sequence, and the deposits are typically arranged as deci-meter-thick packages. These cross-stratified units exhibit reactivation surfacesand locally contain angular, centimeter-size clay clasts. The small-scale Spdeposits fine upward to Smg, to FsSr, then to Fsmtm.

Locally, isolated meter-scale erosional scours (Se; Fig. 8C) cut into theclay- to silt-size sediments. The scours are filled with both trough and

planar cross-stratified medium to coarse sand deposits (St, Sp); laterally,the scours grade to units of silt and clay exhibiting soft-sediment defor-mation and clay containing centimeter-size lenses of medium to fine sandripples. At the Colque measured section, up to boulder-size angular mudblocks (Fig. 8F) composed of Fsmtm clay are present.

Paleosols (P) occur both at the top of and within (Fig. 8G) these fining-upward sequences. They are typically formed in clay- to silt-size sedimentand characteristically contain root material and disseminated organics. Lo-cally, the pedogenic surfaces buried within sediment sequences display ironstaining and gray-white calcium carbonate horizons, and, rarely, pisolitesare present.

Interpretation.—A meandering river deposited these sediments. Evi-dence for this interpretation includes the presence of well-developed fining-upward sequences, abundant ripple-bedded sand, and thick, laterally con-tinuous clay to silt deposits cut by channelized erosional scours. Most of

the strata in FA1 likely record the migration of point bars and the accu-mulation of floodplain silt and clay that is typical of meandering riversystems. The finer-grained meander deposits occur where coarse materialis absent, such as near channel banks where point-bar and floodplain stratadominate. As is common in other river valleys (Grams and Schmidt 2002),the coarser-grained meander deposits occur where abundant coarse alluvi-um directly enters the system or when the river discharge is able to trans-port the coarse material.

Sets of trough cross-bedded gravel (Gt) at the bases of fining-upwardsequences in the Ramis valley strata are similar to the migrating 3D duneforms described by Khadkikar (1999) in the Mahi River of western India.Small-scale sandy planar cross-stratification (Sp) and ripple-bedded silt tosand (FsSr) are point-bar deposits. The interlayered clay and sand deposits(FSil) record lateral-accretion of laminated fine-grained sands and silts ontops of the point-bars and in proximal floodplain regions during overbankflooding. Fine-grained and organic rich lithofacies, such as the massive,mottled silt (Fsmtm) and the horizontally laminated clay (FORl), common-ly accumulate on the floodplain, typically by vertical accretion adjacent tothe channel (Grams and Schmidt 2002). Root mottling, abundant organicmaterial, and local thin laminae record stabilization of the floodplain anddevelopment of floodplain ponds. The discrete, laterally discontinuouschannel scours into laterally and vertically fine-grained sediments (Figs.6C, 7A) argue for the meandering nature of the deposits. The large mudblocks within these deposits suggest that cohesive, fine-grained banks con-fined the channel of the paleo-Ramis, at least locally. Soil development atthe top of outcrop sections is from modern weathering processes occurring




FIG. 5.—Photograph of the C terrace near theIquilo B measured section. This terrace is wideand longitudinally traceable throughout most of the river valley.

TABLE 2.—Summary of lithofacies used in this study.

Lithofacies Code Description Sedimentary Structure Interpretation

FORlFORmlFsmtmFSilFsSr

fine clay (F) with organics (OR)fine clay (F) with organics (OR)silt (Fs) with mottlingclay (F) to sand (S)silt (Fs) and sand (S)

horizontal lamination (l)massive (m) to laminated (l)massive (m)interlayered (il)ripple-bedded

floodplain pondslacustrine sedimentsfloodplain depositsoverbank depositspoint-bar deposits

SmgSpStGm

sand (S) with interspersed gravel (g)sand (S)sand (S)gravel (G)

massive (m)planar cross beds (p)trough cross beds (t)massive (m)

channel depositspoint-bar and sand-flat depositschannel and bar depositschannel deposits

GpGtPSe

gravel (G)gravel (G)paleosolscour (S)

planar cross beds (p)trough cross beds (t)

erosional (e)

cross-channel-bar depositschannel and bar depositsequilibrium periodserosion by new channel

on the exposed terrace surfaces. Where paleosols occur within the sections,they may preserve paleosurfaces formed when vertical aggradation wasslow (Daniels 2003) and the floodplain was stable (exposed to weathering)for prolonged intervals (Fang et al. 2003).

Braided-River Systems (FA2)

Braided-river strata are common throughout the central reaches of theriver valley, specifically in outcrops at the Pucara, Colque, Cacomayo, Ni-casio, Salsapata, and Calapuja locations (Fig. 3). In contrast to the well-developed fining-upward sequences in FA1, the braided systems are char-acterized by laterally continuous, but poorly developed, fining-upward se-quences of large-scale planar and trough cross-bedded sand and graveloverlain by massive to mottled silt and clay (Fig. 7B). They contain thecoarsest deposits in the river valley (sandy matrix-supported pebble tosmall-cobble gravels), suggesting an abundance of gravel bedload in theriver at the time of deposition, and generally indicate high or variableprecipitation and runoff.

Description of Lithofacies.—Gravelly and sandy lithofacies dominatethis facies association. Typically, the fining-upward sequences begin withplanar or trough cross-bedded gravel (Gp, Gt) and sand (Sp, St) depositsand, less commonly, massive gravel (Gm) or sand (Smg) deposits. Over-lying units fine upward to ripple-bedded sand and silt (FsSr), and finally,to massive mottled silt and clay (Fsmtm).

Trough cross-bedded gravel (Gt) and sand (St) are most abundant at thebase of fining-upward sequences (Fig. 8E). They occur within the sequenc-es as flat- to shallowly concave-based, laterally continuous, meter-thickpackets (Fig. 7B). Small- and large-scale planar sand (Sp) and gravel (Gp)

also appear at the bases of fining-upward sequences. Large-scale planarcross-bed sets are usually 10 cm but no more than 1 m thick; small-scale planar cross-bedded sets are generally 10 cm thick. Within eachlarge-scale gravelly or sandy planar cross-bed (Gp or Sp) set, the bedsconsist of alternating layers (15–20 cm) of pebbles and sand. Where small-scale gravelly or sandy planar cross-bed (Gp or Sp) sets occur, they typi-cally consist of multiple decimeter-thick packages (Fig. 8A). The fining-upward sequences may also begin with massive gravels (Gm) that containangular centimeter-size clay clasts.

Massive very fine- to coarse-grained sand with interspersed pebbles(Smg; Fig. 7A) and massive to mottled clay to silt (Fsmtm) overlie thesedeposits. The massive sand deposits vary in thickness from about 10 cmto 1 m. The clay to silt deposits are up to 1 m thick, typically red-brown,commonly contain disseminated root traces, and locally contain centimeter-size silt and fine- to medium-grained sand lenses with lenticular ripplelamination and soft sediment deformation. These clays are rarely disturbed

by soil-forming processes within stratigraphic sections. Paleosols (P; Fig.8G) in these strata are up to 1 m in thickness, weakly developed and iron-stained, and typically contain calcareous layers.

Interpretation.—A braided river deposited this facies association. Generally,the coarse-grained laterally continuous, crudely developed fining-upward se-quences of braided systems result from irregular discharge and the transport of gravelly and sandy bedload in a multi-channel system (Spaliviero 2003). In theRamis valley, FA2 deposits are typically found near areas of bedrock constric-tion or at river confluences, where sediment supply is abundant. However, thebraided facies association is not restricted to these specific environments. Braid-ed-stream sediments are also found in the Ramis valley wherever rapid channel




FIG. 6.—Typical stratigraphic sections from the Rio Ramis valley, showing the three dominant types of facies associations (FA) in the river valley. A) Superimposedfine and coarse meandering-river deposits (FA1) from the Puyutira measured section. B) Braided-river deposits (FA2) from the Pucara measured section. C) Thick lacustrinestrata (FA3) overlain unconformably by meandering-river strata at the Rio Iquilo B measured section. (See Figure 3 for section locations and Tables 2 and 3 lithofaciesand facies-association codes).

TABLE 3.— Descriptions, dominant lithofacies, and depositional environments of the facies associations preserved beneath the Rio Ramis terraces.

Facies Association Dominant Lithofacies Description Depositional Environment

FA1 cross-bedded gravel and sand crudely fining to thin clay andsilt units

Braided Fluvial

FA2 Gt, Gp, Gm, Sp, Smg, FsSr, Fsi l, Fsmtm, FOR1 cross-bedded gravel and sand fining to r ipple-bedded sand tomassive, mottled to laminated clays

Meandering Fluvial

FA3 FOR1 thick units of thinly laminated clay Lacustrine

aggradation has occurred. This aggradation results in the formation of the poor-ly developed fining-upward sequences that are characteristic of this facies as-sociation. Similar braided-stream deposits occur in the Rio Ilave valley, whereRigsby et al. (2003) attribute their formation to rapid deposition of sedimentsduring flood conditions that are related to annual- and millennial-scale precip-itation changes.

Coarser components of the fining-upward sequences represent the activemulti-channel tract; as the channels migrate and become inactive, the de-posits fine upward. The small-scale planar cross-bedded lithofacies (Sp, Gp)and regular repetitions of trough cross-bedded sand and gravel (St, Gt)indicate the migration of dunes within braid channels (Siegenthaler andHuggenberger 1993). Large-scale planar cross-beds may represent lateral-

accretion within a pool of a braided channel. In an aggrading braided river,such deposits have high preservation potential (Siegenthaler and Huggen-beger 1993).

The massive and mottled fine-grained deposits (Fsmtm) represent waningflow on bar tops and in interchannel areas of the braided river system,which receive sediment only during floods. However, deposits such as thesemay also be formed by deposition in abandoned channels (Bentham et al.1993). When waning flow occurs, channel reactivation controls renewedbar migration. When channel abandonment occurs, vegetation and soil-forming processes stabilize and also disturb the sediments. The presenceof paleosols (P) in stratigraphic sections of these braided-river depositsindicates subaerial exposure and local decrease or lack of sediment supply.They may have formed in a manner similar to those observed in the me-andering systems.

Lacustrine Systems (FA3)

The lacustrine facies association is characterized by meter-thick accu-mulations of thinly laminated gray to blue-gray clay (FORml; Fig. 8H).




FIG. 7.—Both meandering-river and braided-river systems are characterized by fining-upward sequences. A) Meandering-river sequences are typically arranged as well-developed in stratigraphic column but lack lateral continuity in outcrop. B) Braided-river sequences are typically arranged as crudely developed sequences in stratigraphiccolumn, and display lateral continuity in outcrop.

The deposits are found at the Ccaluyo and Iquilo measured sections as wellas in river-level outcrops near Pucara and in the Rio Iquilo tributary (Fig.3). In all locations, fluvial deposits directly overlie the lacustrine strata(Figs. 6C, 8H). The fluvial and lacustrine strata are separated by a high-relief and laterally traceable erosional unconformity.

Description of Lithofacies.—The lacustrine FA is composed entirely of lithofacies FORml: thinly laminated to mottled, gastropod- and ostracode-bearing blue-gray clay with abundant organic material. Whole and frag-mented gastropods and ostracods are present as both discrete laminae andas isolated shells. The strata contain local root casts, and mudcracks arepresent near the tops of the deposits. The deeply scoured contact betweenthe fluvial and lacustrine sediment is locally characterized by a rooted andiron-stained surface. The root casts extend at least 10 cm into the under-lying clay and are deeply oxidized, but preserved in decimeter-thick, al-ternating layers of clay and silt.

Interpretation.—These strata were deposited by settling from suspen-

sion in a lacustrine environment. Evidence for this interpretation includesthe laminated and fine-grained nature of the sediment as well as the abun-dance of fresh-water gastropod and ostracode fragments and the absenceof coarse material. The lamination, lack of coarse material, and lack of large-scale sedimentary structures suggest a quiet-water, and possibly deep-water, environment. The densely rooted clay cap and associated mudcracksrecord a shallow-water transitional environment, suggesting that the lake-level was falling before fluvial activity resulted in the deep erosion of thepreviously deposited lake strata.

TERRACE MORPHOLOGY

The fluvial and lacustrine strata just described are preserved beneath fivedistinct terrace tracts (A through E) that range in height from 1.7 to 52.7m above the modern river level and are up to 4 km wide (Figs. 4C, 9).

The terraces generally decrease in height and increase in width downvalley.They are physically traceable throughout most of the study area but areinfluenced by their proximity to three prominent bedrock constrictions(Figs. 9, 10). Both the height and the extent of the terraces and the faciesassociations that comprise the terraced strata vary with proximity to theconstrictions.

A through E

Description.—Terrace E, with a height about 50 m above river level, ispresent in the northern sites of Pucara, Colque, and Cacochico (Figs. 3, 9).The E terrace surface is underlain by lacustrine strata that were deposited

at least 40,170 430 cal yr BP (Table 1) and are separated from thefluvial sediments of terraces D through B by an unconformity surface. Thisunconformity crops out locally along the river valley (at Ccaluyo and Pu-cara 2) and forms a traceable surface in the Rio Iquilo tributary, along thePucara profile line (Fig. 11), from 40 m above and 4.5 km away from themodern river level in the terrace E level to low-flow river level at the RioIquilo–Rio Pucara confluence (Fig. 3). Small, isolated remnants of fluvialstrata are present immediately above the unconformity in some locations.Reworked woody material and clay from fluvial remnants yielded dates of 47,600 1,740 cal BP at Puyutira and 37,700 600 at Iquilo A.

With heights ranging from 20 to 40 m above river level, the D terraceis present everywhere except Nicasio. This terrace typically occurs only onthe east side of the modern river. The surface is often dissected and hasan average width of 1,900 m, extending to the base of either terrace E orthe alluvial fans at the base of nearby mountains. Sediments underlying theterrace consist of FA1 and FA2 fluvial deposits. Carbonate sediments and

organic-rich sediments in this sequence are dated at 31,510 510, and21,970 500 cal yr BP (Table 1). Both carbonates and organics in buriedpaleosols from Camaraja and Yucajache yield ages of 35,140 790,18,497 549, 10,421 181, and 8,339 50 cal yr BP (Table 1, Fig.12).

With a height of 14 m above modern river level, terrace C is typicallythe height of the cutbank surface at the locations of the measured sections.It is the widest and most extensive terrace in the river valley. It is physicallytraceable throughout the study area (Fig. 5) and has an average width of 2,100 m. North of Colque, it is typically narrow, but downstream (south)of Colque the terrace surface becomes very wide on the east side of theriver (Fig. 9). At all locations, chiefly FA1 meandering-river sedimentsunderlie the C surface. Paleosols are common near the height of the Csurface in continuous sections. At Salsapata, a paleosol between the C and

B terrace heights records a date of 9,095 66 cal yr BP (Table 1).The B terrace surface, with a height of about 9 m and a width of about

400 m, is present in all profile lines except Nicasio. The B terrace formedfrom incision and is commonly separated from the other terraces by anoxbow lake or a cutoff meander (Fig. 4C). Predominantly braided-riversediments of FA2 underlie the B surface; in several locations, the terraceheight is marked in continuous sections by a paleosol.

Terrace A is present in all reaches of the river valley and has an averageheight of 2 m and an average width of 280 m. The terrace A sedimentsare meandering-river deposits and differ from deposits at the same heightwithin continuous sequences on higher cutbank exposures. At Calapuja2




FIG. 8.—Typical lithofacies structures appear in both meandering-river and braided-river sequences. A) Planar cross-bedded sand (Sp), such as in this photograph, istypical of active channel environments in both meandering and braided systems. B) The meandering units commonly fine upward to ripple-bedded silt and sand deposits(FsSr) that are usually associated with tops of point-bars or proximal floodplain deposits. C) Distal-floodplain clay deposits (Fsmtm) or D) the thinly laminated clay of floodplain wetlands or oxbow lakes (FORl) may overlie these strata. E) The braided-river sequences commonly begin with trough cross-bedded gravel (Gt) and commonlyfine upward to fine-grained bar-top deposits. In meandering-river systems, the fine-grained deposits are commonly cut by erosional scours (Se, photograph C). F) At Colque,mudblocks and boulder-size clay rip-ups composed of floodplain clay deposits are found adjacent to a channel scour. G) Paleosols (P) are also found in stratigraphicsections of both meandering-river and braided-river sediments. H) An unconformity surface below the fluvial sediments is traced from below the modern river level to theE terrace surface. This surface separates the fluvial sediments from the underlying lacustrine deposits.




FIG.9.—River-perpendicular profiles from the Ramis valley. Profiles are shown

from north (top) to south (bottom) along the river valley (refer to Figure 3 for exactlocations), with the location of bedrock constrictions and the Ramis–Azangaro con-fluence indicated with arrows. Note the five distinct terrace surfaces and the generaldownstream decrease in height and increase in width of the terrace C surface. (Seetext for discussion.)

and Ccaluyo, sediments beneath terrace A yielded dates of 1,668 76 and1,457 112 cal yr BP, respectively.

History of Terrace Development.—The E terrace was formed by dis-section of underlying lacustrine strata that accumulated in the valley priorto the large-scale incision. The small remnants of fluvial strata that overliethe local unconformity surface are the deposits of the river system thatincised the lacustrine strata. The E terrace surface is closer to the river inupstream stretches; downstream, the valley widens and the surface is farther

away from the river and very dissected. The lacustrine strata beneath theE surface were downcut (Fig. 12) to at least the modern river level sometime after 40,170 430 cal yr BP but before 35,140 790 cal yr BP(the oldest reliably dated fluvial strata in the A through D sequence; Table1). The downcutting probably started before 37,700 600 cal yr BP—theage of the reworked clay in the fluvial strata at Iquilo A (Fig. 12). Theincision resulted in the rooted and iron-stained erosional surface betweenthe terrace E lacustrine strata and the overlying (D through A) sequenceof fluvial strata. This irregular unconformity surface marks at least 40 mof incision across the 10.1-km-wide valley that was previously occupiedby an extensive freshwater lake (FA3). The braided-river and meandering-

river strata of the D–A terraced sequences were deposited in the valleycreated by this erosional event.

Erosional events of such a valley-wide scale, even when triggered bybase-level lowering, are typically episodic (Schumm et al. 1987; Womackand Schumm 1977). Erosion into cohesive clays (e.g., the lacustrine sedi-ments in the Ramis valley) is typically fairly rapid (McFadden and Mc-Auliffe 1997), depending on the rate of base-level fall (Schumm 1993).Because the unconformity surface is traceable throughout the Ramis valley,

base-level fall must have been large and rapid, allowing the incision topropagate upstream.

The D terrace formed by valley aggradation. Dates from the Camarajafill sequence suggest that the surface aggraded between 35,140 790 and18,497 549 cal yr BP (Table 1) and possibly extended to 10,421 181cal yr BP (Fig. 12). The 35,140 790 and 18,497 549 cal yr BP datesare from paleosols in the stratigraphic sequence at Camaraja. Buried pa-leosols bound by pedogenically unaltered sediment imply slow aggradation(Daniels 2003) and record equilibrium periods during the stepwise depo-sition in the Ramis.

The sediments under the D terrace consist of both FA1 and FA2; how-ever, this change in the fluvial system is not laterally consistent among themeasured sections throughout the Ramis valley. Changes from meanderingto braided or braided to meandering may reflect episodic erosion and ag-

gradation occurring at the same time in different parts of the river valley(Vandenberge 1995). In this case, longitudinal variability in the river val-ley, especially the presence of bedrock constrictions, is the likely cause of the lateral variability of facies associations. Preservation of such complexfacies architecture may be the result of periods of rapid aggradation (Skel-ley et al. 2003).

The sediments beneath the D depositional terrace were downcut to form theC and B terraces as well as the erosion surface beneath the A modern terrace(Fig. 12). This sequence of events was initiated sometime after 10,421 cal yrBP (the age of the erosion surface at the top of D) and sometime before 1,668 76 cal yr BP (the oldest date from within the terrace A sediment sequence;Table 1). Paleosols present at or near the heights of the terraces imply periodsof stasis in the downcutting. The 9,095 66 cal yr BP date of carbonatematerial in a paleosol in the C/B sediment sequence (Table 1) records one suchtime of soil development. However, the number of individual stabilization

events and their precise timing is ambiguous. The presence of two poorlydeveloped paleosols at the level of the C and B surfaces at Puyutira and Sal-sapata (Fig. 2) imply at least two separate downcutting events, separated bybrief periods of soil formation followed by renewed incision to form the surfacebelow the modern river level.

The modern depositional terrace A has accumulated in, at least, the past1,497 112 cal yr BP (from a date in Ccaluyo fluvial sediments). It isdistinguished as a depositional terrace because the underlying sedimentsonlap the D aggradational deposits (Fig. 12). Aggradation of fine-grainedsediments reflects modern overbank sedimentation in response to a domi-nantly meandering-type system.

Bedrock Constrictions

Three major bedrock constrictions are present in the study area (Figs. 3,10). These constrictions, formed by large outcrops of Devonian shale andquartzite bedrock (Fig. 3; Newell 1946), influence the habit of the adjacentriver terraces and sediments. In the areas of maximum constriction, terraceformation is rare. Only where the cross-river distance between topographichighs is large, as at Lulicunca (Fig. 3), do terraces form in areas of max-imum constriction. Immediately adjacent to constrictions (to the north andsouth), terraces are present but are narrower and less well developed thanin nonconstricted reaches of the river valley (Fig. 9).

Because bedrock constrictions restrain transport, sediments build up andterrace heights increase behind (upstream of) the constrictions. For exam-ple, although terrace heights generally decrease downstream, the heights of




FIG.10.—River-perpendicular profiles through the three major bedrock constrictions in the Ramis valley (refer to Figure 3 for exact locations and Figure 8 for locationsrelative to other profiles). The vertical line through each set of profiles marks the position of river level (zero meters) on the east side of the channel. The narrowest profile

at each locality (marked with an arrow) represents the area of maximum constriction.

FIG. 11.—Schematic cross section showing thetraceable unconformity surface exposed in theRio Iquilo tributary along Pucara profile. SeeFigure 3 for profile and section locations.

the terrace C, and, to a lesser extent, the B and D surfaces increase upstreamof constrictions (Fig. 13). Also, above the constrictions, coarser-grainedmeandering-stream or braided-stream facies associations are present (Fig.14). Downstream from the constrictions, the C surface decreases to lessthan or close to its height in the previous reach and, in most cases, thefacies associations become finer grained (dominantly meandering). Also,during times of higher precipitation, more sediment is derived from topo-graphic highs of this type (Pratt et al. 2002; Grams and Schmidt 2002),adding to an already taxed sediment transport system. Because of decreasedstream competence, the coarse sediment cannot be transported and is de-posited upstream of the next constriction (Figs. 13, 14).

FLUVIAL HISTORY AND RELATIONSHIP TO CLIMATE

We reconstruct the fluvial history of the Ramis valley on the basis of sedimentologic and geomorphic field data, along with radiocarbon dates.Ultimately, changes in fluvial landscapes depend partly on the magnitude,rate, duration, and direction of base-level changes (Schumm 1993), whichare commonly controlled by climatic variations (Ritter et al. 1995). Because

LT is base-level for the Rio Ramis, changes in LT water level have asignificant influence on aggradation and incision in the river valley. Asdiscussed earlier and illustrated in Figure 2, changes in lake-level are con-trolled largely by fluctuations in precipitation.

Studies of sediment cores from LT have yielded long-term records of changes in Ramis base-level (lake-level) and regional precipitation (Bakeret al. 2001b; Baker et al. unpublished data). Contemporaneous lacustrinestrata in the Salar de Uyuni (Baker et al. 2001a; Fritz et al. 2004) and theRio Desaguadero valley (Baucom and Rigsby 1999; Rollins 2001; Warren2002), combined with the record from the Sajama, Huascaran, and Illimaniice cores (Thompson and Mosley-Thompson 1995; Thompson et al. 1998;

Hoffmann et al. 2003) and the LT sediment cores, show that the precipi-tation changes recorded in the LT sediments were regional in nature (alsosee Bradley et al. 2003; Vuille et al. 2003a; Vuille et al. 2003b). In theRio Desaguadero valley (the hydrologic connection between LT in thenorth and the Salar de Uyuni in the south) lacustrine sediments record atleast four distinct wet periods of paleolake formation (Rollins 2001; Warren2002) that are also recorded in outcrops and in cores from the Salar de




FIG. 12.—Model for the history of terrace development in the Rio Ramis valley. Arrows show aggradational (up) and erosional (down) events. Numbers show thesequence of events and correspond to those in Figure 15.

FIG. 13.—Graph showing the variation interrace heights along the river valley. Heightsare partially controlled by proximity to bedrockconstrictions.

Uyuni and LT. The youngest of the Altiplano paleolakes, ‘‘Coipasa’’ (Syl-vestre et al. 1999), has been radiocarbon dated to between 13,400 and11,500 cal yr BP (Servant et al. 1995; Baker et al. 2001b). The much largerpaleolake ‘‘Tauca,’’ with a maximum depth of 140 m (Bills et al. 1994),has been dated between 25,000 and 16,000 cal yr BP (Baker et al. 2001a).And, an older large paleolake ‘‘Minchin’’ (Servant and Fontes 1978),which had a minimum elevation of 40 m above modern-day lake Titicaca,has been dated to 38,200 14C yrs BP in the northern Rio Desaguaderovalley (Rigsby, unpublished data) and between 36,000 and 46,000 cal yrBP (Fritz et al. 2004) in the Salar de Uyuni. Intervening dry episodes arerecorded by fluvial sedimentation and/or erosion in the Rio Desaguaderovalley (Baucom and Rigsby 1999; Rollins 2001; Warren 2002), depositionof salt in the Salar de Uyuni (Baker et al. 2001a), and lowered water levels

in LT (Baker et al. 2001b). Less dramatic changes in effective moistureare also recorded in many of these proxies.As illustrated in Figure 15, we hypothesize that aggradation and down-

cutting events in the Rio Ramis valley occurred in response to the regionalchanges in effective moisture that are recorded throughout the Altiplano—

changes that resulted in precipitation-controlled base-level fluctuations. De-creases in precipitation resulted in decreased base (LT) level. Increases inprecipitation resulted in increased base (LT) level and, during the majorevents, the formation of large lakes on the central Altiplano. Of course, asin any fluvial system, topography, runoff, and sediment supply complicatethe response of the river system to large-scale changes of base-level andprecipitation. The following sections describe the major events recorded bythe sedimentology and geomorphology of the river valley and the relation-ship of those events to regional-scale climate change. Event numbers cor-respond to those used in Figures 12 and 15.

Event 1

The lacustrine deposits (FA3) beneath the E terrace record the presenceof an extensive lake that covered a large portion of the northern Altiplanoat and before 40,170 430 cal yr BP—the age of lacustrine strata justbelow the unconformity at Ccaluyo (Table 1, Fig. 12). In the Ramis valley,these strata are present at a maximum elevation of 40 m above river level




FIG. 14.—Graph showing the distribution of facies associations along river valley and therelationship between facies location and bedrock

constrictions.

FIG. 15.—Hypothesized relationship between downcutting and aggradational periods and base-level change triggered by changes in climate (precipitation) in the RioRamis valley. Lake-level curve is from Baker et al. (2001b) and uses benthic diatoms as a proxy for water depth. An abundance of benthic diatoms indicates dry conditionsand implies a lowering of base-level; lack of benthic diatoms indicates wet conditions and implies a rise of base-level. The Salar de Uyuni gamma curve is from Baker etal. (2001a)—salts (dry intervals) have low natural gamma radiation and muds (wet intervals) have high natural gamma radiation. Blank intervals in the curve mark locationsof volcanic ash. The numbers correspond to the sequence of events illustrated in Figure 12. Asterisks indicate radiocarbon dates from the Ramis strata.

(in the Rio Iquilo tributary)—about 150 m above modern Lake Titicaca.The sediments are lithologically distinct from the older and tilted lake sed-iments of the Azangaro Formation (Fig. 3; Newell 1946), which they over-lie unconformably. Strata of similar age have been noted in outcrops of thenorthern Rio Desaguadero valley, where they crop out at an elevation of 3,865 m (Rigsby and Baucom 1998), and along the Rio Catari (Argolloand Mourguiart 2000). Drill cores from LT reveal that the lake was con-tinuously deep and fresh from prior to 60,000 to about 15,000 cal yr BP(Baker, unpublished data). The lacustrine sediments correspond to an ex-

tremely wet period, between 46,000 and 36,000 cal yr BP (Clapperton1993; Clapperton 1997; Fritz et al. 2004), when paleolake ‘‘Minchin’’ cov-ered a large part of the central Bolivian Altiplano (Baker et al. 2001a).

Apparently, at the same time that paleolake ‘‘Minchin’’ flooded the cen-tral Altiplano, LT rose at least 40 m above its modern lake-level (Baucomand Rigsby 1999). This high lake stand found in sediments south of themodern LT now appears to have been coeval with lacustrine sedimentsfrom the Ramis valley north of LT. A rise of LT to 150 m above its presentlevel would have resulted in a huge expansion of its total area and would




have necessitated a much-higher-than-modern outlet height. The subsequentdowncutting of this hypothesized outlet would have catastrophically low-ered lake-level, resulting in rapid downcutting of the T5 surface.

Event 2

The lacustrine strata deposited during Event 1 were downcut and themajor unconformity surface beneath the Rio Ramis fluvial strata was

formed as base-level decreased, some time after 40,170 430 cal yr BPand prior to the start of accumulation of sediments in terrace D begansometime before 35,140 790 cal yr BP (Fig. 14). Reworked fluvialsediment from terrace remnants at Iquilo A and Puyutira suggest that thedowncutting began before 37,700 600 cal yr BP. This event documentsa major drop of LT level, perhaps coincident with the desiccation of pa-leolake ‘‘Minchin,’’ marking a major dry episode on the Altiplano. Becausethe substrate consisted of fine-grained cohesive clays, a large, rapid change(spanning about 2,000 years) in base-level had the potential to propagateincision upstream throughout the entire valley (Schumm 1993). As is com-mon in incision triggered by base-level lowering (Schumm et al. 1987;Womack and Schumm 1977), erosion in the Ramis valley was episodic.The rates of desiccation and subsequent downcutting are ambiguous; lake-level fall may have been extremely rapid if controlled by erosion of theUlloma/Callapa topographic constriction in the northern Rio Desaguadero

valley (Clapperton 1993; Rigsby and Baucom 1998).

Event 3

A large-scale and long-lasting aggradation to the D terrace level (Fig.12) began sometime before 35,140 790 cal yr BP. Aggradation occurredas a result of high effective moisture (increased precipitation) and base-level increase in the LT region between about 30,000 and 20,000 cal yrBP (Fig. 15). Sediments in the Rio Ilave valley, 100 km to the south, mayalso have accumulated during this same aggradational event (Rigsby et al.2003). Because the aggradation was relatively slow, paleosols formed atCamaraja during stable periods at 35,140 790 and 18,497 549 cal yrBP. It is likely that the aggradation had ended and the valley had alreadybegun to be formed by incision prior to 10,421 181 cal yr BP.

In addition to base-level increase, upper reaches of the river valley alsoreceived more runoff (because of increased precipitation), which likely in-creased sediment delivery to the river (Maddy et al. 2001 and many others).Some of this increased sediment supply must have been the result of wide-spread expansion of glaciers in the higher-altitude portions of the watershedcoincident with the global LGM (at around 21 ka), a period of cold and wetconditions on the Altiplano. During this period vegetation was sparse (Paduanoet al. 2003) and sediment erodibility likely high. Deglaciation began in thewatershed at about 20,000 cal yr BP (Seltzer et al. 2002). As is common (Bull1991; Schumm 1993; Vandenberghe 1995; Pratt et al. 2002; Spaliviero 2003;Skelley et al. 2003), the increased sediment supply and the inability of theriver to transport the load caused the valley to aggrade.

Event 4

LT fell below its outlet at about 11,500 yr BP (Baker et al. 2001b), andbase-level may have dropped rapidly for the next few hundred years (Fig.15). In the Ramis valley, several downcutting episodes occurred between10,241 181 and 1,668 76 cal yr BP. As LT fell periodically duringthe early Holocene, the river valley underwent several rapid erosionalevents. This erosion was a response to episodic drying in the river valley,interrupted by wetter intervals when soil formation occurred. The down-cutting event resulted in the formation not only of terrace C (Event 4a)and terrace B (Event 4b), but also the erosional trace beneath the terraceA depositional surface (Event 4c; Fig. 12).

The downcutting likely corresponded to dry periods on the Altiplano at11,500 to 10,000 cal yr BP, 8,500 to 7,500 cal yr BP, and 7,000 to 5,000

cal yr BP (Fig. 15). The dry periods decreased lake levels, and the rivervalley eroded to accommodate this base-level change. Erosion to the topof the D terrace and subsequently to the C surface may have occurredduring the drop in the level of LT between 11,500 and 10,000 cal yr BP.The B surface may have formed as LT levels decreased between 8,500 and7,500 cal yr BP (Fig. 15). Between 7,000 and 5,000 cal yr BP, LT sedi-ments record the lowest levels of the lake during the past 60,000 cal yr BP(Fig. 15), as the level of LT reached about 85 m below modern (Cross et

al. 2000). This base-level fall may account for the formation of the erosionsurface beneath the A terrace surface.

Most likely, each individual erosional event followed a, sometimes brief,wet period of soil formation and sediment aggradation. These wet periodsare seen in the LT watershed at about 10,000 to 8,500 cal yr BP and 7,500to 7,000 cal yr BP (Fig. 15) and are likely times of the paleosol formationnear the tops of the D, C, and B terrace surfaces. The 8,339 50 cal yrBP paleosol at the D terrace level at Yucajache records one such wet period.The erosional terraces in the Ilave valley formed in a similar manner anddisplay gravel lags and paleosols in terrace tops that record similar periodsof equilibrium (Rigsby et al. 2003).

Event 5

In the Ramis valley, the terrace A depositional surface suggests that base-level rose and/or precipitation, runoff, and sediment load increased enoughto allow aggradation for the past 1,457 112 cal yr BP. The terrace Aaggradation is coincident with a phase of lake-level rise and overflow atabout 2,000 cal yr BP, as noted by Baucom and Rigsby (1999). Similardeposits are found in the Rio Ilave basin, where between 4,000 and 1,600cal yr BP two onlapping depositional terraces formed in the valley (Rigsbyet al. 2003). The terraces in both valleys may also have been fed by ag-ricultural runoff (and the resultant increased sediment load) as human pop-ulations rose dramatically in the LT watershed after about 4,000 cal yr BP.

DISCUSSION

The sedimentology and geomorphology in the Rio Ramis valley were

controlled dominantly by base-level changes of LT, which are attributed toclimatic episodes recorded throughout the LT basin. Although it is typicallydifficult to correlate terrace development with climate change (Bull 1991;Schumm 1993; Ritter et al. 1995), reliable paleoclimate records and gooddating can make the task easier. The terraces and sediments of the RioRamis are traceable throughout the valley, and radiocarbon dates generallysubstantiate the hypothesized correlation between fluvial history and base-level change (Fig. 15). Differences in terrace height and extent, as well asvariations in the depositional environment of the terraced sediment, seemto have been caused by smaller-scale variations that resulted from localtopographic or lithologic controls.

Transitions in river systems are often caused by addition of sedimentsupply or changes in channel stream gradient and are commonly associatedwith topographic influence or areas of river confluences (Freeman 2000).In the Ramis valley, differences in both river planform and lateral distri-

bution of facies associations are highly dependent on the location of thebedrock constrictions. For example, the abundant gravel in facies associa-tions near the Ramis constrictions is likely the result of increased supplyof coarse sediment from the Devonian topographic highs. Because the ad-dition of coarse material often causes a river to transition from meanderingto braided (Smith and Smith 1984), the coarse multi-channel environmentsseen in the constricted regions are not unexpected. Changes in channelplanform can also be the result of changes in gradient (Smith and Smith1984; Simpson and Smith 2001). In the Ramis valley, changes in gradientare coincident with bedrock constrictions (Fig. 3). Coarser-grained braided-river sediments are present near the constrictions where gradients are high-




er; finer-grained meandering-river sediments are more common away fromthe constrictions where gradients are more gentle (Fig 3).

Terraces form during periods of aggradation or incision. Terrace aggra-dation occurs as a result of inhibited bedload transport (Bull 1991). In thesparsely vegetated Ramis valley, periods of increased precipitation allowincreased erosion of adjacent hillslopes and deposition in the river valley.Rising base-level (a result of increased regional precipitation) also enhancesthe likelihood of aggradation. Conversely, downcutting generally occurs as

sediment supply to the river declines and precipitation or base-level, orboth, decrease. The fact that the fluvial terraces are traceable throughoutthe Ramis valley, except in areas of bedrock constriction, suggests thatsimilar conditions were occurring throughout the watershed at the time of terrace formation.

LT (base) levels fluctuate with changes of effective moisture, drivenlargely by changes in precipitation and temperature. Periods of lacustrinesedimentation and fluvial aggradation or incision relate directly to changinglake phases. The lacustrine phase at around 40,000 cal yr BP and the ag-gradational period between about 35,140 790 and 10,421 181 cal yrBP in the Ramis correspond to the ‘‘Minchin’’ and ‘‘Tauca’’ lacustrinephases. The downcutting event to form C and B terraces between about10,421 181 and 1,668 76 cal yr BP occurred as a result of episodicdrying of the LT watershed. In the last 2,000 years, the terrace A surface

aggraded in response to LT rise and, possibly, human land-use practices.In the Ramis valley, aggradation caused by precipitation-forced base-levelchange occurs slowly whereas incision occurs fairly rapidly.

The Ramis valley fluvial landscapes are similar to those reported fromthe Ilave valley (Rigsby et al. 2003). Both valleys contain thick sedimentarysequences that resulted from aggradation starting before the LGM cold andwet conditions and extending to 8,250 cal yr BP in the Ilave valley. Alsosimilar to the Ramis, erosional terraces formed in the Ilave valley by epi-sodic downcutting between 6,000 and 4,500 cal yr BP, and low inset ter-races formed within the last 4,000 years. Interestingly, the Ramis valleycontains only one inset terrace (A), whereas the Ilave contains two (oneformed between approximately 4,000 to 2,500 cal yr BP, another formedbetween approximately 2,200 and 1,600 cal yr BP). This difference mayresult from differences in topography and lithology (specifically, the abun-dance of easily eroded and transported pyroclastic sediment in the Ilave

watershed may have allowed the aggradation of its two late Holocene de-positional terraces between the two river valleys) or from differences inthe occupation (agricultural) history of the two river valleys.

The fact that the Ramis and Ilave valleys record similar responses suggeststhat changes in regional climate (effective moisture) and base-level were thedominant controls on both fluvial landscapes. During periods of high or risingbase-level, the Rios Ramis and Ilave aggraded. Aggradation during these pe-riods was generally slow, as is common in arid regions with abundant sedimentsupply (Bull 1991), allowing the buildup of thick sedimentary sequences andthe formation of fill terraces in the river valleys. Downcutting in response tobase-level change commonly occurs in a rapid, episodic manner (Schumm etal. 1987; Womack and Schumm 1977). The downcutting in the Ramis valleyshows such a response and forms fill-cut terraces in rapid and episodic pulsesthat correspond to periods of falling lake-level.

SUMMARY

Three main facies associations exist in the Rio Ramis valley: braided,meandering and lacustrine. Braided-river sediments are commonly presentnear bedrock constrictions, whereas meandering-river sediments are mostcommon in the longer, unencumbered reaches.

There are five terrace tracts in the valley. The terrace E surface is un-derlain by lacustrine strata deposited in an 40,000 cal yr BP paleolake.The terrace D depositional surface was deposited between 35,140 790and 10,421 181 cal yr BP and is separated from these lacustrine strataby a high-relief erosional unconformity. The terrace C and B surfaces are

erosional terraces that formed during two pulses of erosion between 10,241 and 1,668 76 cal yr BP. Terrace A is the modern fill terrace thatformed from aggradation beginning before 1,668 76 cal yr BP to present.The terraces are traceable throughout the river valley, but they increase inheight (and the terraced sediments increase in grain size) upstream of bed-rock constrictions.

Periods of aggradation and downcutting in the Ramis valley (and in otherriver valleys in the LT region) correspond to changes in base-level. When

LT was high or rising, the rivers in its watershed aggraded; when LT waslow or falling, the rivers in its watershed incised. Inasmuch as these stagesof LT have been related to multi-centennial to orbital scale changes inprecipitation and near-global-scale climate forcings, the fluvial systems of the LT watershed are themselves hypothesized to have a coherent responseto the same large-scale climate forcing.

ACKNOWLEDGMENTS

This work was supported by a National Science Foundation grant (to Rigsby,#ATM-9709035). We thank the personnel of Instituto Arqueologıa de Peru for help-ing to facilitate our work in the Ramis valley; the Autoridad Autonomo de LagoTiticaca for their continued support of our Titicaca basin research; Albino PilcoQuispe, Honorato Ttacca, Jennifer Garland, Ashley Ballentine, and Nathaniel Bakerfor their valuable field support; and Michael D. Blum, David S. Leigh, Martin H.Trauth, and Paul A. Baker for their critical reviews of the manuscript.

REFERENCES

ABBOTT, M.B., BINFORD, M.W., BRENNER, M., AND KELTS, K., 1997, A 3500 14C yr high-reso-lution record of water-level changes in Lake Titicaca, Bolivia/Peru: Quaternary Research,v. 47, p. 167–180.

ACEITUNO, P., AND MONTECINOS, A., 1993, Circulation anomalies associated with dry and wetperiods in South American Altiplano: American Meteorological Society, Fourth InternationalConference of Southern Hemisphere Meteorology and Oceanography, p. 330–331.

ALDENDERFER, M., 1989, The Archaic period in the south-central Andes: Journal of WorldPrehistory, v. 3, p. 117–158.

ALDENDERFER, M., 1998, Montane Foragers: Asana and the South-Central Andean Archaic:Iowa City, Iowa, University of Iowa Press, 344 p.

ALLMENDINGER, R.W., JORDAN, T.E., KAY, S.M., AND ISACKS, B.L., 1997, The evolution of theAltiplano–Puno Plateau of the Central Andes: Annual Review of Earth and Planetary Sci-ences, v. 25, p. 139–174.

ARGOLLO, J., AND MOURGUIART, P., 2000, Late Quaternary climate history of the Bolivian Al-tiplano: Quaternary International, v. 72, p. 37–51.

BAKER, P.A., RIGSBY, C.A., SELTZER, G.O., FRITZ, S.C., LOWENSTEIN, T.K., BACHER, N.P., ANDVELIZ, C., 2001a, Tropical climate changes at millennial and orbital timescales on the Bo-livian Altiplano: Nature, v. 409, p. 698–701.

BAKER, P.A., SELTZER, G.O., FRITZ, S.C., DUNBAR, R.B., GROVE, M.J., TAPIA, P.M., CROSS, S.L.,ROWE, H.D., AND BRODA, J.P., 2001b, The history of South American tropical precipitationfor the past 25,000 years: Science, v. 291, p. 640–643.

BAUCOM, P.C., AND RIGSBY, C.A., 1999, Climate and lake-level history of the northernAltiplano,Bolivia, as recorded in Holocene sediments of the Rio Desaguadero: Journal of SedimentaryResearch, v. 69, p. 597–611.

BENTHAM, P.A., TALLING, P.J., AND BURBANK, D.W., 1993, Braided stream and flood-plain de-posits in rapidly aggrading basin: The East Scanilla Formation, Spanish Pyrenees, in Best,J.L., and Bristow, C.S., Braided Rivers: London, Geological Society, 419 p.

BILLS, B.G., DE SILVA, S.L., CURREY, D.R., EMENGER, R.S., LILLQUIST, K.D., DONNELLAN, A., AND

WORDEN, B., 1994, Hydro-isostatic deflection and tectonic tilting in the central Andes: Initialresults of a GPS survey of Lake Minchin shorelines: Geophysical Research Letters, v. 21,p. 293–296.

BRADLEY, R.S., VUILLE, M., HARDY, D., AND THOMPSON, L.G., 2003, Low latitude ice cores recordPacific sea-surface temperatures, Geophysical Research Letters, v. 30, no. 4, 1174, doi:1029/ 2002L016546.

BRIDGLAND, D.R., 2000, River terrace systems in north-west Europe: an archive of environ-mental change, uplift and early human occupation: Quaternary Science Reviews, v. 19, p.1293–1303.

BULL, W.B., 1991, Geomorphic Response to Climatic Change: New York, Oxford UniversityPress, 326 p.

CLAPPERTON, C.M., 1993, Quaternary Geology and Geomorphology of South America: Am-sterdam, Elsevier, 779 p.

CLAPPERTON, C.M., 1997, Late Quaternary glacier advances and paleolake high-stands in Bo-livian Altiplano: Quaternary International, v. 34, p. 39–49.

CROSS, S.L., BAKER, P.A., SELTZER, G.O., FRITZ, S.C., AND DUNBAR, R.B., 2000, A new estimateof the Holocene lowstand level of Lake Titicaca, central Andes, and implications for tropicalpalaeohydrology: The Holocene, v. 10, p. 21–32.

DANIELS, M.J., 2003, Floodplain aggradation and pedogenesis in a semiarid environment: Geo-morphology, v. 56, p. 225–242.

FANG, X., LU, L., MASON, J.A., YANG, S., AN, Z., LI, J., AND GUO, Z. 2003, Pedogenic response




to millennial summer monsoon enhancements on the Tibetan Plateau: Quaternary Interna-tional, v. 106–107, p. 79–88.

FREEMAN, A.K.L., 2000, Application of high-resolution alluvial stratigraphy in assessing thehunter-gatherer/agricultural transition in the Santa Cruz River Valley, southeastern Arizona:Geoarchaeology: An International Journal, v. 15, p. 559–589.

FRITZ, S.C., BAKER, P.A., LOWENSTEIN, T.K., SELTZER, G.O., RIGSBY, C.A., DWYER, G.S., TAPIA,P.M., ARNOLD, K.K., KU, T-L., AND LUO, S., 2004, Hydrologic variation during the last170,000 years in the southern hemisphere tropics of South America: Quaternary Research,v. 61, p. 95–104.

FULLER, I.C., MACKIN, M.G., LEWIN, J., PASSMORE, D.G., AND WINTLE, A.G., 1998, River response

to high-frequency oscillations in southern Europe over the past 200 k.y.: Geology, v. 26, p.275–278.GARREAUD, R., VUILLE, M., AND CLEMENT, A., 2003, The climate of the Altiplano: observed

current conditions and mechanisms of past changes: Palaeogeography, Palaeoclimatology,Palaeoecology, v. 194, p. 5–22.

GRAMS, P.E., AND SCHMIDT, J.C., 2002, Streamflow regulation and multi-level flood plain for-mation: channel narrowing on the aggrading Green River in the eastern Uinta Mountains,Colorado and Utah: Geomorphology, v. 44, p. 337–360.

HANSON, B.C.S., WRIGHT, H.E., AND BRADBURY, J.P., 1984, Pollen studies in the Junin area,central Peruvian Andes: Geological Society of America, Bulletin, v. 95, p. 1454–1465.

HOFFMANN, G., RAMIREZ, E., TAUPIN, J.D., FRANCOU, B., RIBSTEIN, P., DELMAS, R., DURR, H.,GALLAIRE, R., SIMOES, J., SCHOTTERER, U., STIEVENARD, M., AND WERNER, M., 2003, Coherentisotope history of Andean ice cores over the last century: Geophysical Research Letters, v.30, no. 4, 1179, doi:10.1029/2002GL014870.

JAIN, M., AND TANDON, S.K., 2003, Fluvial response to Late Quaternary climate changes, west-ern India: Quaternary Science Reviews, v. 22, p. 2223–2235.

KHADKIKAR, A.S., 1999, Trough cross-bedded conglomerate facies: Sedimentary Geology, v.128, p. 39–49.

KNOX, J., 1972, Valley alluviation in southwestern Wisconsin: Association of American Ge-

ographers, Annals, v. 62, p. 401–410.LENTERS, J., AND COOK, K., 1997, On the origin of the Bolivian high and related circulation featuresof the South American climate: Journal of Atmospheric Sciences, v. 54, p. 656–677.

LEOPOLD, L.B., AND W OLMAN M.G., 1957, River channel patterns: braided, meandering andstraight: Geological Survey, Professional Paper 282-B, p. 39–73.

MADDY, D., BRIDGLAND, D., AND WESTAWAT, R., 2001, Uplift-driven valley incision and climate-controlled river terrace development in the Thames Valley, U.K.: Quaternary International,v. 79, p. 23–26.

MCFADDEN, L.D., AND MCAULIFFE, J.R., 1997, Lithologically influenced geomorphic responsesto Holocene climatic changes in the Southern Colorado Plateau, Arizona: A soil-geomorphicand ecologic perspective: Geomorphology, v. 19, p. 303–332.

MELICE, J.L., AND ROUCOU, P., 1998, Decadal time scale variability recorded in the Quelccayasummit ice core 18O isotopic ratio series and its relation with the sea-surface temperature:Climate Dynamics, v. 14, p. 117–132.

NEWELL, N.D., 1946, Geology of the Lake Titicaca region, Peru and Bolivia: Geological Societyof America, Memoir 36, 111 p.

NOTT, J., PRICE, D., AND NANSON, G., 2002, Stream response to Quaternary climate change:evidence from the Shoalhaven River catchment, southeastern highlands, temperateA ustralia:Quaternary Science Reviews, v. 21, p. 965–974.

PADUANO, G., BUSH, M., BAKER, P., FRITZ, S., AND SELTZER, G., 2003, The deglaciation of LakeTiticaca (Peru/Bolivia): a vegetation and fire history: Palaeogeography, Palaeoclimatology,Palaeoecology, p. 259–279.

PORTER, S.C., AN, Z., AND ZHENG, H., 1992, Cyclic Quaternary alluviation and terracing in anonglacial drainage basin on the northern flank of Qinling Shan, China: Quaternary Re-search, v. 38, p. 157–169.

PRATT, B., BURBANK, D.W., HEIMSATH, A., AND OJHA, T., 2002, Impulsive alluviation during earlyHolocene strengthened monsoons, central Nepal Himalaya: Geology, v. 30, p. 911–914.

RENEAU, S.L., 2000, Stream incision and terrace development in Frijoles Canyon, BandelierNational Monument, New Mexico, and the influence of lithology and climate: Geomor-phology, v. 32, p. 171–193.

RIGSBY, C.A., AND BAUCOM, P.C., 1998, Fluvial terraces record late Pleistocene highstands onthe Bolivian Altiplano: evidence that Lake Titicaca had multiple effluent streams at 30,000BP [abstract]: EOS, Transactions, American Geophysical Union.

RIGSBY, C.A., BAKER, P.A., AND ALDENDERFER, M. S., 2003, Fluvial history and human occu-pation of the Rio Ilave valley, Peru: Palaeogeography, Palaeoclimatology, Palaeoecology,p. 165–185.

RITTER, D.F., KOCHEL, R.C., AND MILLER, J.R., 1995, Process Geomorphology: Boston, WCB/ McGraw-Hill, 546 p.

ROCHE, M.A., BOURGES, J., AND M ATTOS, R., 1992, Climatology and hydrology of the LakeTiticaca Basin, in Dejoux, C., and Iltis, A., eds., Lake Titicaca: A Synthesis of LimnologicalKnowledge: Dordrecht, the Netherlands, Kluwer Academic Publishers, p. 63–85.

ROLLINS, S.M., 2001, Quaternary lacustrine and fluvial history of the central Bolivian Altiplanoas recorded in subsurface strata of the Rio Desaguadero valley, Bolivia [unpublished M.S.thesis]: East Carolina University, Greenville, North Carolina, 152 p.

ROSE, J., AND BOARDMAN, J., 1983, River activity in relation to short-term climatic deterioration:Quaternary Studies in Poland, v. 4, p. 189–198.

SCHULTE, L., 2002, Climatic and human influence on river systems and glacierfluctuations in southeastSpain since the Last Glacial Maximum: Quaternary International, v. 93–94, p. 85–100.

SCHUMM, S.A., 1993, River response to baselevel change: implications for sequence stratigra-phy: Journal of Geology, v. 101, p. 279–294.

SCHUMM, S.A., AND K HAN, H.R., 1972, Experimental study of channel patterns: GeologicalSociety of America, Bulletin, v. 83, p. 1755–1770.

SCHUMM, S.A., MOSLEY, M.P., AND WEAVER, W.E., 1987, Experimental Fluvial Geomorphology:New York, John Wiley & Sons, 413 p.

SELTZER, G.O., BAKER, P., CROSS, S., DUNBAR, R., AND FRITZ, S., 1998, High-resolution seismicreflection profiles from Lake Titicaca, Peru–Bolivia: Evidence for Holocene aridity in thetropical Andes: Geology, v. 26, p. 167–170.

SELTZER, G.O., RODBELL, D.T., BAKER, P.A., FRITZ, S.C., TAPIA, P.M., ROWE, H.D., AND DUNBAR,R.B., 2002, Early Warming of Tropical South America at the Last Glacial–InterglacialTran-sition: Science, v. 296, p. 1685–1686.

SERVANT, M., AND FONTES, J.C., 1978, Les lacs quaternaires des hauts plateaux des Andesboliviennes, premieres interpretations paleoclimatiques: Institut Francais de RechercheScientifique pour le developpement en Cooperation (ORSTOM), Cahiers, Serie Geologie10, p. 9–24.

SERVANT, M., FOURNIER, M., ARGOLLO, J., SERVANT-VILDARY, S., SYLVESTRE, F., WIRRMANN, D.,AND Y BERT, J.P., 1995, La derniere transition glaciaire/interglaciaire des Andes tropicalessud (Bolivie) d’apres l’etude des variations des niveaux lacustres et des fluctuations glaci-aires: Academie des Sciences (Paris), Comptes Rendus, t. 320 serie Iia, p. 729–736 .

SIEGENTHALER, C., AND HUGGENBERGER, P., 1993, Pleistocene Rhine gravel: deposits of a braidedriver system with dominant pool deposits, in Best, J.L., and Bristow, C.S., eds., BraidedRivers: London, Geological Society, 419 p.

SIMPSON, C.J., AND S MITH, D.G., 2001, The braided Milk River, northern Montana, fails theLeopold–Wolman discharge-gradient test: Geomorphology, v. 41, p. 337–353.

SKELLEY, R.L., BRISTOW, C.S., AND ETHRIDGE, F.G., 2003, Architecture of channel-belt depositsin an aggrading shallow sandbed braided river: the Lower Niobrara River, northeast Ne-braska: Sedimentary Geology, v. 158, p. 249–270.

SMITH, N.D., AND SMITH, D.G., 1984, William River: An outstanding example of channel wid-ening and braiding caused by bedload addition: Geology, v. 12, p. 78–82.

SPALIVIERO, M., 2003, Historic fluvial development of Alpine foreland Tagliamento River, Italy,and consequences for floodplain management: Geomorphology, v. 52, p. 317–333.

STARKEL, L., 2003, Climatically controlled terraces in uplifting mountain areas: QuaternaryScience Reviews, v. 22, p. 2189–2198.

STUIVER, M., REIMER, R.J., BARD, E., BECK, J.W., BURR, G.S., HUGHEN, K.A., KROMER, B.,MCCORMAC, F.G., PLICHT, J.V.D., AND SPURK, M., 1998, INTCAL98 radiocarbon age cali-bration 24,000–0 cal BP: Radiocarbon, v. 40, p. 1041–1083.

SYLVESTRE, F., SERVANT, M., SERVANT-VILDARY, S., CAUSSE, C., FOURNIER, M., AND YBERT, J.-P.,1999, Lake-level chronology on the southern Bolivian Altiplano (18–23S) during late-glacial time and the early Holocene: Quaternary Research, v. 51, p. 54–66.

THOMPSON, L.G., AND MOSLEY-THOMPSON, E., 1995, Late Glacial stage and Holocene tropical icecore records from Huascaran, Peru: Science, v. 269, p. 46–50.

THOMPSON, L.G., DAVIS, M.E., MOSLEY-THOMPSON, E., SOWERS, T.A., HENDERSON, K.A., ZAGO-RODNOV, V.S., LIN, P.-N., MIKHALENKO, V.N., CAMPEN, R.K., BOLZAN, J.F., COLE-DAI, J., AND

FRANCOU, B., 1998, A 25,000-year tropical climate history from Bolivian ice cores: Science,v. 282, p. 1858–1864.

THOMPSON, L.G., MOSLEY-THOMPSON, E., DAANSBARG, W., AND GROOTES, P.M., 1986, The LittleIce Age as recorded in the stratigraphy of the tropical Quelccaya ice cap: Science, v. 234,p. 361–364.

VANDENBERGHE, J., 1995, Timescales, climate and river development: Quaternary Science Re-views, v. 14, p. 631–638.VANDENBERGHE, J., 2003, Climate forcing of fluvial system development: an evolution of ideas:

Quaternary Science Reviews, v. 22, p. 2053–2060.VANDENBERGHE, J., AND M ADDY D., 2001, The response of river systems to climate change:

Quaternary International, v. 79, p. 1–3.VUILLE, M., BRADLEY, R., AND KEIMIG, F., 2000, Interannual climate variability in the Central

Andes and its relation to tropical Pacific and Atlantic forcing: Journal of Geophysical Re-search, v. 105, p. 12,447–12,460.

VUILLE, M., BRADLEY, R.S., WERNER, M., HEALY, R., AND KEIMIG, F., 2003a, Modeling 18O inprecipitation over the tropical Americas, part 1. Interannual variability and climatic controls:Journal of Geophysical Research, v. 108, (D6), 4174, doi:10.1029/2001JD002038.

VUILLE, M., BRADLEY, R.S., HEALY, R., WERNER, M., HARDY, D.R., THOMPSON, L.G., AND KEIMIG,F., 2003b, Modeling 18O in precipitation over the tropical Americas, part 2. Simulation of the stable isotope signal in Andean ice cores: Journal of Geophysical Research, v. 108, (D6),4175, doi:10.1029/2001JD002039.

WARREN, M., 2002, Timing and magnitude of lacustrine phases on the Bolivian Altiplano asrecorded in the subsurface strata of the Rio Desaguadero valley, Bolivia [unpublished M.S.thesis]: East Carolina University, Greenville, North Carolina, 232 p.

WENINGER, B., JORIS, C., AND DANZEGLOCKE, U., 2002, CalPal Cologne Radiocarbon Calibration

and Paleoclimate Research Package: Universitat zu Koln, Institut fur Ur- and Fruhgeschi-chte, Radiocarbon Laboratory, http:\www.calpal.de.

WIRRMANN, D., 1992, Morphology and bathymetry, in Dejoux, C., and Iltis, A., eds., LakeTiticaca: A Synthesis of Limnological Knowledge: Dordrecht, The Netherlands, KluwerAcademic Publishers, p. 170–176.

WOMACK, W.R., AND SCHUMM, S.A., 1977, Terraces of Douglas Creek, northwestern Colorado:An example of episodic erosion: Geology, v. 5, p. 72–76.

ZHOU, J., AND LAU, K.-M., 1998, Does a monsoon climate exist over South America?: Journalof Climate, v. 11, p. 1020–1040.

Received 2 May 2003; accepted 26 April 2004.

Documents

Climatic Influences on Sedimentology & Geomorphology of the Rio Ramis Valley, Peru - JSR, 2005