Variable Denudation in the Evolution of the BolivianAndes: Controls and Uplift-Climate-Erosion Feedbacks

Item Type text; Thesis-Reproduction (electronic)

Authors Barnes, Jason B.

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to thismaterial is made possible by the Antevs Library, Department ofGeosciences, and the University Libraries, University of Arizona.Further transmission, reproduction or presentation (such aspublic display or performance) of protected items is prohibitedexcept with permission of the author or the department.

Download date 07/09/2021 09:04:57

Link to Item http://hdl.handle.net/10150/240131

Variable Denudation in the Evolution of the Bolivian Andes:Controls and Uplift- Climate -Erosion Feedbacks

ByJason B. Barnes

A Prepublication Manuscript Submitted to the Faculty of theDEPARTMENT OF GEOSCIENCES

In Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE

In the Graduate CollegeTHE UNIVERSITY OF ARIZONA

2002

1

Variable Denudation in the Evolution of the Bolivian Andes: Controls and

Uplift- Climate -Erosion Feedbacks

Jason B. Barnes and Jon D. PelletierDepartment of Geosciences, University of Arizona, Tucson, AZ, 85721 USA

ABSTRACT. Controls on denudation in the eastern Bolivian Andes are evaluated by

synthesis of new and existing denudation estimates from basin -morphometry, stream -

powered fluvial incision, landslide mapping, sediment flux, erosion surfaces,

thermochronology, foreland basin sediment volumes, and structural restorations.

Centered at 17.5 °S, the northeastern Bolivian Andes exhibit high relief, a wet climate, and

a narrow fold- thrust belt. In contrast, the southeastern Bolivian Andes have low relief, a

semi -arid climate, and a wide fold -thrust belt. Basin -morphometry indicates a northward

increase in relief and relative denudation. Stream -power along river profiles shows greater

average incision rates in the north by a factor of 2 to 4. In the south, profile knickpoints

with high incision rates are controlled by fold -thrust belt structures such as the surface

expressions of basement megathrusts, faults, folds, and lithologic boundaries. Landslide

and sediment -flux data are controlled by climate, elevation, basin morphology, and size

and show a similar trend; short -term denudation -rate averages are greater in the north (1-

9 mm/yr) than the south (0.3 -0.4 mm/yr). Long -term denudation -rate estimates including

fission track, basin fill, erosion surfaces, and structural restorations also exhibit greater

values in the north (0.2 -0.8 mm/yr) compared to the south (0.04 -0.3 mm/yr). Controls on

long -term denudation rates include relief, orographic and global atmospheric circulation

patterns of precipitation, climate change, glaciation, and fold -thrust belt geometry and

2

kinematics. The denudation synthesis supports two conclusions: 1) denudation rates have

increased towards the present 2) an along- strike disparity in denudation (greater in the

north) has existed since at least the Miocene and has increased towards the present.

Denudation rates and controls suggest that Bolivian mountain morphology is controlled by

both its orientation at mid -latitude, and the feedbacks between uplift, kinematics,

orographic effects on precipitation, glaciation, and the increased erosion that accompanies

orogenesis.

INTRODUCTION

The geologic community has recently shown great interest in the feedbacks between uplift,

climate, and erosion. The suggestion that late Cenozoic uplift of the Tibetan plateau may have

influenced global climate spawned intensive study of the causal relationships between global

records of increased late Cenozoic uplift, erosion, and climatic cooling. Molnar and England

(1990) argued that increased erosion could be the result of the global cooling, glaciation, and

higher climatic variability, not only increased tectonic uplift. Raymo and Ruddiman (1992)

proposed that the uplift of the Tibetan plateau altered atmospheric circulation and increased the

drawdown of CO2 by enhanced weathering rates that caused global cooling and additional

erosion. Understanding this uplift- climate relationship requires better knowledge of the

feedbacks between uplift, topography, climate, and erosion in mountain -belt evolution.

Feedbacks between uplift, climate, and erosion shape mountain belts. There are three

common examples. First, tectonic uplift strives to build mountains while erosion tears them

down (for example Willett and Brandon, 2002). The more uplift, the more erosion.

Additionally, erosion influences deformation and consequently the distribution of uplift within

3

an orogenic wedge as described by Coulomb wedge theory and critical taper (Dahlen and Suppe,

1988; Dahlen, 1990; DeCelles and Mitra, 1995). Second, greater relief and elevation increases

precipitation and can eventually induce glaciation, both of which increase erosion (Hallet and

others, 1996; Roe and others, in press). Finally, valley incision can cause the uplift of mountain

peaks through flexural -isostatic compensation induced by the erosional unloading (for example

Montgomery, 1994).

Numerical modeling has been used to investigate the impacts and dynamics of uplift- climate-

erosion feedbacks in orogenic evolution. Beaumont and others (1992) concluded that the

orographic influence of climate on erosion, given enough time ( -5 Myr), could lead to

asymmetries in climate and tectonic style across an orogen. Avouac and Burov (1996) showed

that climatic zones have control on the spatial distribution of intracontinental strain and tectonics.

Using stream -power to estimate fluvial erosion, Willett (1999) demonstrated the positive

feedbacks between asymmetrical exhumation, orographic precipitation, mass removal, average

slopes, and topography across an orogen. Beaumont and others (2001) showed how mid -crustal

channel flow and ductile extrusion can be coupled to focused surface denudation at the

Himalayan front. These and other modeling studies provide important inferences about uplift-

climate- erosion feedbacks that can be applied to the examination of real mountain belts (for

example Hoffman and Grotzinger, 1993; Koons, 1994).

The complex variability of the controls on denudation has limited the application of numerical

models of mountain evolution to real mountain belts (Beaumont and others, 2000; Hovius,

2000). However, quantifying denudation within a mountain -belt through time and space is

important for understanding the role of denudation in orogenesis. An alternative approach to

numerical modeling is to constrain the spatial and temporal variability of denudation rates and

4

their controls using datasets collected from mountain -belt geomorphology and geology. For

example, recent databases and digital topography provide new perspectives on morphometric

controls on basin denudation at variable spatial scales (Hovius, 1998). Furthermore, we can

quantify denudation at different spatial and temporal scales using thermochronology,

cosmogenic and radiometric isotopes, volume estimates from sedimentary basins, fluvial incision

models, stratigraphic and geomorphic field observations, and modern sediment flux.

At the orogenic scale, the competition between uplift and erosion conditions a mountain belt

to strive towards a stable or steady state (Adams, 1980). The steady -state condition requires a

balance between the positive and negative feedbacks associated uplift, climate, and erosion

within orogenic systems. For example, fluvial erosion increases relief, but is mitigated by mass -

wasting processes that tend to reduce relief Therefore, identifying limits and controls on the

topographic boundary conditions of a mountain -belt as well as the internal uplift -climate- erosion

feedbacks are important for describing an orogen's approach to steady state (Willett and

Brandon, 2002). For example, Burbank and others (1996) recognized uniformly steep hillslopes

in the northwestern Himalayas irrespective of denudation or incision rates, implying that a

bedrock landslide threshold process limited relief. Brozovic and others (1997) showed that

glacial processes control the limits to elevation and relief of mountains at mid -latitudes. Meigs

and Sauber (2000) demonstrated that mean topography of the southern Alaska ranges are

controlled by both their latitude and maritime setting, through the feedbacks between uplift,

glacial erosion, and orographic forcing of climate. However, short -term (102 -103 years)

approximation of steady -state topography is difficult because river and hillslope denudation

processes and rates vary dramatically (Bull, 2002). These studies provide a basis for directing

observations in other mountain belts in terms of extrinsic controls, such as subduction and global

5

climatic patterns, and internal feedbacks such as uplift and glacial erosion that might influence

their ability to reach a steady state.

The Andes are an exceptional natural laboratory for exploring uplift- climate -erosion

feedbacks in a mountain system. The Andes possess significant along - strike contrasts in styles

of deformation, uplift, climate, and denudation. For example, the Andes span 60° degrees of

latitude, resulting in a climatic gradient that ranges from tropical to arid. A continental -scale

analysis of Andean topography and mean annual precipitation by Montgomery and others (2001)

suggested that the distribution of crustal mass is controlled by the style and intensity of

denudation in addition to tectonic shortening. Even at a regional scale, contrasts in kinematics,

relief, and erosion can be extreme. The eastern Bolivian Andes exhibits a dramatic along- strike

contrast in climate, topography, and fold -thrust belt geometry and kinematics.

Previous studies have documented strong correlations between the topography, structural

geology, precipitation, orogenic wedge behavior, and erosion in the eastern Bolivian Andes

(Masek and others, 1994; Mugnier and others, 1997; Horton, 1999; Leturmy and others, 2000;

McQuarrie, 2002b). The northern Bolivian Eastern Cordillera possesses high relief, a wet

climate, and a narrow fold -thrust belt that is characterized by out -of- sequence thrusting. The

southern Bolivian Eastern Cordillera possesses low relief, a semi -arid climate, and a wide fold -

thrust belt characterized by more sequential thrusting. Differences in subduction geometry

cannot be responsible for the along - strike contrast in deformation because the top of the slab dips

uniformly at -30° in this region (Tinker and others, 1996). Critical taper theory, analogue, and

numerical modeling have also been applied to the evolution of the Bolivian fold -thrust belt

wedge (Mugnier and others, 1997; Horton, 1999; Leturmy and others, 2000). General

6

conclusions were that erosion has played a controlling role in the kinematic and topographic

evolution of the eastern Bolivian Andes.

Our objective is to build upon the previous studies of Bolivian orogenic -wedge tectonics and

erosion by quantifying denudation across the orocline with real datasets to constrain erosion rates

on multiple temporal scales. This approach will quantify erosional control on the Bolivian fold -

thrust belt and provide values of erosion that may be used to calibrate numerical models of

mountain -belt evolution. To achieve this objective, we synthesized: (1) basin -morphometry data

as a relative index of basin denudation; (2) sediment -flux and landslide mapping data to estimate

short -term denudation rates; (3) apatite fission -track cooling ages, structural cross -sections, basin

fill volumes, and erosion -surface incision to estimate long -term denudation rates, and; (4)

stream- powered fluvial incision along river profiles. We address the following questions:

What are the most direct controls and feedbacks on denudation in this region, and how do

they contrast between the north and south?

What patterns do we see in our denudation rate chronology and what do they suggest

about the feedbacks between uplift, climate, and erosion in the eastern Bolivian Andes?

How does this database elaborate on the idea of erosional control in the evolution of the

eastern Bolivian Andes?

Finally, in order to better understand the feedbacks between uplift, climate, and erosion in the

eastern Bolivian Andes, we use this synthesis of denudation rates and controls to interpret

Bolivian orogenic -wedge behavior with particular attention to its critical taper, kinematic

evolution, and its approximation to steady -state conditions.

7

REGIONAL SETTING: TOPOGRAPHY, TECTONICS, AND CLIMATE

The Andes are the product of Cenozoic crustal shortening and thickening associated with the

Andean fold -thrust belt. Shortening is caused by Nazca plate subduction beneath the western

margin of South America (for example Isacks, 1988). The Andes exhibit great variation in

geomorphology and structural style along their more than 60° of latitude (Kley and others, 1999;

Kerman, 2000; Montgomery and others, 2001; McQuarrie, 2002a). The Bolivian Andes, on a

regional scale, exemplify this physical variability. The Bolivian Andes (14 -22 °S) contain a

striking contrast in topographic relief, structural geometry and kinematics, and climate between

areas north and south of the oroclinal bend at 17.5 °S (fig. 1). A wet climate, high relief, and

narrow fold -thrust belt characterize the north. In contrast, the south is characterized by a semi-

arid climate, low relief, and wide fold -thrust belt (Horton, 1999; McQuarrie, 2002b).

The large -scale topography of the eastern Bolivian Andes is controlled by the structural

geology. Figure 1 shows the four physiographic provinces (from east to west), the latter three of

which are defined by average elevation differences of 1 km: the Altiplano (AP), an internally -

drained high plateau of >3 km elevations bounded on the west by the modern volcanic arc and to

the east by the 6 -km high peaks of the Eastern Cordillera (EC); The EC is 3 km in average

elevation decreasing eastward from 6 -km high peaks; The Interandean zone (IA) is 2 km in

average elevation; and Subandean zone (SA) is 1 km in average elevation (McQuarrie, 2002b).

Through detailed structural cross -section balancing and restoration, Kley (1996, 1999) and

McQuarrie (2002b) inferred the presence of two basement megathrusts. These megathrusts

produce large structural steps controlled by their ramps and terminations. The 6 -km high

structural steps at the terminations of these megathrusts manifest themselves at the surface as 1-

km high topographic steps (McQuarrie, 2002b). The megathrust terminations produce the

8

topographic steps that mark the EC -IA and IA -SA transitions. The regional surface faults

coincident with the structural and topographic steps at the EC -IA and IA -SA boundaries are the

Main Andean Thrust and the Main Frontal Thrust, respectively (Sempere and others, 1988).

The chronology of mountain building in the central Andes is disputed. Contraction began in

the Eastern Cordillera somewhere between Eocene ( McQuarrie, 2002b) and Late Oligocene time

(Isacks, 1988; Gubbels and others, 1993). Eocene mountain building in the EC is constrained by

local thermochronology at c. 40 ± 5 Ma (McBride and others, 1987; Masek and others, 1994).

Late Oligocene contraction in the EC is bracketed by ages (20 -25 Ma) of adjacent foreland basin

deposits (Sempere and others, 1990; Jordan and others, 1997). Active deformation has been

concentrated in the Subandean zone for the last 10 Myr or more (Gubbels and others, 1993;

McQuarrie, 2001).

The Andean fold -thrust belt varies dramatically in kinematics and geometry across the

orocline. In the north, the SA is only 30 -90 km wide, whereas in the south, it is 90 -125 km wide

(fig. 1). The orogenic -wedge decollement slope is twice as steep in the north compared to the

south (2° vs. 4 °), and average surface slopes are up to six times (0.5° vs. 3 °) greater in the north

than in the south (Horton, 1999). Since the Miocene, deformation in the northern SA has been

characterized by out -of- sequence thrusting while the south has been dominated by in- sequence

thrusting (Baby and others, 1992, 1995). In the north, the topographic break between the EC and

the SA has receded to the west by 20 -30 km from the Main Andean Thrust suggesting significant

erosion (Masek and others, 1994). The southern EC possesses a flat -lying late Miocene erosion

surface, the San Juan del Oro surface, which suggests no upper -crustal deformation has occurred

there since -10 Ma (Gubbels and others, 1993). The north is devoid of any similar features

(Horton, 1999).

9

The climate of the eastern Bolivian Andes is correlated to global atmospheric circulation

patterns. Throughout the Cenozoic, the central Andes have been at approximately the same

latitude of -20 °S, and have always been transverse to the major Hadley -cell- driven wind currents

that control global precipitation distribution (Gordon and Jurdy, 1986; Montgomery and others,

2001; Garrison, 2002). The Bolivian Andes are located between the equatorial intertropical

convergence zone (ITCZ) where low pressure and precipitation is abundant, and the horse

latitudes (30° N and S) where high pressures and dry desert conditions persist. Easterly winds

bring precipitation in from the Brazilian shield such that locally, the northeastern Bolivian Andes

receive wetter, northeasterly prevailing winds and the southeastern Bolivian Andes receive drier,

southeasterly prevailing winds.

The regional climate also has strong correlations with topography in the eastern Bolivian

Andes. In the north, the mean annual precipitation is at least twice as great as the south (fig. 1).

In addition, the average stream discharge from the fold -thrust belt is much greater in the north

(Horton, 1999; Montgomery and others, 2001). Masek and others (1994) attributed this contrast

in climate to be orographic effects associated with the greater relief and mountain front slope of

the northeastern Bolivian Andes. Greater relief and wetter climate facilitated by both global and

orographic effects should lead to greater denudation in the north.

DENUDATION

In this section we present denudation information and rates for the eastern Bolivian Andes

from basin -morphometry and drainage network analyses and a synthesis of previous and new

estimates from a variety of sources. We also identify the controls and feedbacks associated with

each denudation estimate. The basin -morphometry and channel network analyses were carried

10

out at 1 -km resolution with the USGS GTOP030 DEM using RiverTools 2.4 (developed by S.D.

Peckham).

Basin -Morphometry and Relative Denudation

Many studies have correlated controls of basin denudation with basin -morphometry and

climatic settings to quantify spatial variations in denudation. Anhert (1970) demonstrated that

mean denudation rate in mid - latitude basins is directly proportional to mean basin relief. Other

studies showed that denudation is directly proportional to drainage -basin area, maximum basin

elevation, mean- trunk -channel gradient, mean annual precipitation, basin -relief ratio (basin

relief/maximum basin length), and local relief (Pipet and Souriau, 1988; Milliman and Syvitski,

1992; Summerfield and Hulton, 1994).

The three major basins of the Bolivian Eastern Cordillera- the Beni, Grande, and Pilcomayo -

exhibit spatial trends in the basin- morphometry and climatic settings indicative of greater

denudation in the north relative to the south. Figure 2 shows the Beni basin draining the northern

EC and the Pilcomayo basin draining the southern EC. In between the Beni and Pilcomayo

basins, the Grande basin occupies the southern EC, but its northwestern headwaters reach into

the northern EC across the orocline. The Beni basin possesses greater maximum basin elevation

and relief, and relief -ratio than either the southern Grande or Pilcomayo basins (table 1). The

Beni basin is approximately the same size as the Pilcomayo basin. In addition, mean annual

precipitation data and modeled mean December-January -February (DJF) precipitation show

precipitation increasing northward (figs. 1 and 3).

Hypsometry (area- elevation) was used to estimate relative denudation between the Beni and

Pilcomayo basins (Masek and others, 1994). Hypsometry of the three major basins -the Beni,

Grande, and Pilcomayo- suggest a northward increase in denudation. The lower the hypsometric

11

integral value, defined by the area beneath the hypsometric curve, the greater the relative basin

denudation (Wohl, 2000). By this criterion, basin denudation increases northward except for the

much smaller Parapeti basin (fig. 4). The Beni and Parapeti basin uplands are well dissected,

and the lowlands possess synclinal valleys of the SA filled with sediment. The Parapeti basin

lowland possesses a large fluvial megafan (Horton and DeCelles, 2001). The Grande and

Pilcomayo uplands are less dissected, with the latter showing significant basin area at high

elevations (fig. 4).

Short -term Denudation -rate Estimates, Controls, and Feedbacks

Previous estimates. - -- Previous short -term denudation -rate estimates for the Bolivian Andes

are quite sparse and vary over four orders of magnitude (fig. 5). Blodgett (1998) mapped

landslides across two northern EC drainage basins (- 10 km2 each). Making assumptions about

area- volume relationships, hillslope storage, and landslide age, he estimated an erosion rate of 4-

14 mm/yr. Landslide age was determined from scar -revegetation times of 10 -35 years according

to the elevation zone. Elevation, climate, and season during which repeat aerial photographs

were taken (and tree -rings) controlled his estimate of re- vegetation time and therefore landslide

age. He assumed a 90% delivery factor of material from the hillslopes to the rivers. The

denudation estimate of 4 -14 mm/yr is similar to estimates from landslide mapping in the

Southern Alps of New Zealand (Hovius and others, 1997). However, the assumption of only

10% hillslope sediment storage estimated from a small number of landslides, compared across a

few repeat aerial photographs taken at different seasons and tree -rings, implies that this estimate

may be an upper limit.

Blodgett (1998) extrapolated these denudation estimates to an area approximately an order of

magnitude larger than the two basins mapped by calibrating Landsat image reflectance

12

spectrums. He concluded that the 4 -14 mm/yr estimate was appropriate to the high and moderate

relief zones of the canyons of the northern EC. Key factors triggering landslides in these two

zones were high relief, relatively moist conditions, and sufficient stream -power (slope and

discharge) to undermine the slopes (Blodgett, 1998). In contrast, the dominant erosional

processes in the highest, previously glaciated valleys of the northern EC were rockfalls, and

avalanching rock debris because they were transport- limited environments where stream -power

was minimal (Blodgett, 1998). In summary, the dominant controls on denudation estimated from

landslide mapping are stream- power, climate, and precipitation, which are all, in turn, partially

controlled by the orographic enhancement of precipitation.

Summerfield and Hulton (1994) reported total basin denudation rate averages from sediment-

flux data for the Amazon basin and Parana basins. The Beni basin rivers are tributaries to the

Amazon river and the Chaco basin rivers are tributaries to the Parana river (fig. 2). Normalized

to basin area, the estimates are 0.093 mm/yr for the Beni basin and 0.014 mm/yr for the Parana

basin (fig. 5). In addition, Blodgett (1998) calculated denudation rates from sediment -flux data

near his landslide study area. These estimates range from 0.0073 -8.1 mm/yr. He noted that the

two lowest - magnitude rates were from formerly glaciated valleys.

New estimates. - -- We calculated denudation -rate estimates from sediment -flux data that

provide additional quantitative evidence for greater denudation rates in the northeastern Bolivian

Andes (fig. 6, table 2). To first order, sediment flux integrates the diverse affects of

geomorphology, climate, lithology, and tectonics that influence erosion and sediment storage in

the basin source area into a denudation rate (Hovius, 1998). However, the short time involved in

data collection (15 years or less in this case), the episodic nature of sediment delivery in

mountains, and the importance of bedload (10% of suspended load or more in rivers) ignored by

13

these measurements suggest that these rates are generally a minimum (Milliman and Meade,

1983; Kirchner and others, 2001). We calculated denudation rate assuming total denudation for

the suspended load (TSS), a 65% denudation rate component for the dissolved load (TDS), and a

rock density of 2.6 g /cm3, after Knighton (1998) (table 2). Other studies use a lower regolith

density, 1.2 -2.2 g /cm3, which further suggests our estimates may be a minimum (for example

Gunnell, 1998). We compiled total suspended load and total dissolved load data from Guyot and

others (1988, 1990, 1993, 1994). Dissolved loads were reported for approximately half of the

Beni and Pilcomayo basin stations, and all Bermejo basin stations. Of those reported stations,

the average TDS /TSS ratio was 13% for the Beni basin and 10% for the other basins. We used

the 13% TDS/TSS average to calculate the TDS for the remaining Beni basin stations and 10%

TDS /TSS for the remaining stations for which the TDS was not reported.

The Beni basin exhibits greater denudation at multiple basin scales and less sign of sediment

storage at the largest basin scales than the south. Denudation -rate averages are generally greater

by a factor of two or three for the Beni basin in the north compared to the Pilcomayo and

Bermejo basins in the south. Beni basin denudation rates are 0.2 -2.5 mm/yr compared to 0.05-

0.8 mm/yr for the Pilcomayo basin (table 2). The average denudation rate is 1.3 mm/yr for the

Beni basin and 0.3 mm/yr for the Pilcomayo basin. The Pirai and Grande basins centered at the

oroclinal bend range from 0.1 -4.7 mm/yr. The Grande basin is unique for several reasons. It

possesses headwaters that reach both north and south of the bend. It drains to the south of the

bend, shares morphometric features intermediate to the Beni and Pilcomayo basins, and has the

highest modeled DJF precipitation (figs. 2, 3, 4, table 4). The Grande basin also possesses the

largest denudation rates. For this reason, further discussion of basin contrasts between the north

and south will be limited to the Beni and Pilcomayo basins.

14

Basin size influences denudation rate in the eastern Bolivian Andes (table 2). For example,

small basins (area = 102 km2) of the Beni and Pirai drainage basins average 0.7 -1.0 mm/yr of

denudation in comparison to 0.4 mm/yr for small basins of the Bermejo drainage basin.

Denudation -rate averages of medium -sized basins (area = 103 km2) of the Beni and Grande

drainage basins are 1.5 -2.5 mm/yr compared to 0.4 mm/yr for medium -sized basins of the

Pilcomayo drainage basin. Denudation -rate averages of the largest -sized basins (area = 104 km2)

of the Beni and Pilcomayo drainage basins are 1.5 mm/yr and 0.3 mm/yr, respectively. The

highest relative denudation -rate averages are in the medium -sized basins. Some denudation rates

decrease with increasing basin size probably reflecting the influence of sediment storage

processes (Milliman and Meade, 1983). This is certainly true when comparing the total basin

denudation -rate averages for the Amazon and Parana basins, which are an order of magnitude

lower than their montane basin components (fig. 5, table 2).

What are the dominant controls on basin denudation in this region? Basin size is the most

important control. In the north, the smallest and formerly glaciated basins possess the lowest

denudation rates. In the south, the average denudation rates for the Pilcomayo and Bermejo

basins are 3 -4 times slower and more uniform than the north. The unique Grande basin

denudation rates are the highest of all coincident with the highest modeled DJF precipitation

suggesting a strong control by seasonal rainfall rates.

Long -term Denudation -rate Estimates, Controls, and Feedbacks

Previous estimates. - -- One previous long -term denudation -rate estimate comes from apatite-

fission -track cooling ages. Low - temperature thermochronometers, such as apatite fission -track

dating, can constrain denudation rates at long time scales through knowledge of crustal thermal

properties (for example Gleadow and Brown, 2000). Apatite thermochronology assumes that a

15

sample age reflects the time since it passed through the 110 °C isotherm, also known as the

closure temperature. The 110 °C isotherm translates to 4 -6 km depths for average geothermal

gradients of 20- 30 °C/km. The resultant exhumation rate is identical to denudation rate (England

and Molnar, 1990). Safran (1998) conducted apatite- fission -track thermochronologic studies on

more than 20 samples of granitoids from the northern EC that included the dataset of Benjamin

and others (1987). She estimated 0.2 -0.6 mm/yr of denudation from samples aged 6 -20 Ma (fig.

5). These exhumation rates decrease with age suggesting an increase in denudation rate towards

the present. These denudation rate calculations incorporated topographic influences on

subsurface temperature gradients maximizing their accuracy (Stuwe and others, 1994). Apatite

and zircon fission -track data from the same location (hence not shown in fig. 5) recorded 4 -8 km

of material eroded over the last 10 -15 Myr, which corresponds to 0.27 -0.8 mm/yr of exhumation

(Benjamin and others, 1987; Masek and others, 1994). More recent work on the Benjamin and

others (1987) dataset suggests that exhumation rates have increased in the last 50 Ma from 0.1 to

as much as 0.75 mm/yr at present in either an exponential or step -wise fashion (Anders and

others, 2002) .

Recent work suggests that glacial erosion can significantly influence the cooling histories of

the northern EC granites (Pelletier, 2002). This implies that exhumation rates inferred from the

granites may be controlled by climate change in the form of global cooling and resultant glacial

erosion. We suggest caution in inferring surface uplift from the EC exhumation rates since

glacial erosion could partially control these exhumation rates, and inferences of surface uplift

from cooling ages are generally questionable (Molnar and England, 1990). More recent climate

change or uplift or both could control denudation estimates from the fission -track ages in the

16

northern EC. The increase in denudation rate towards the present suggested by these datasets

could also be the result of both climate change and uplift.

New estimates. - -- We estimated additional long -term denudation rates from a wide range of

source materials including thermochronology, erosion surfaces and their incision, sediment

volume in the foreland basin, and structural cross - sections (fig. 5). Preliminary apatite- fission-

track cooling ages of sedimentary rocks from southern Bolivia range from 38 +/- 7 to 30 +/- 5

Ma in the central EC, and 19 +/- 13 to 13 +/- 3 Ma in the IA (Ege and others, 2001). For an EC

average age of 34 Ma and an IA average of 16 Ma, we calculated 0.15 mm/yr denudation for the

EC, and 0.31 mm/yr for the IA, assuming closure was reached at 5 km depth (- 25 °C/km

geothermal gradient). These denudation estimates are simply a first -order approximation.

Gubbels (1993) estimated the amount of material removed by denudation from the San Juan

del Oro erosion surface in the southern EC since its late Miocene formation. By DEM analysis

along three transects, he concluded that 110 -240 km2 of material was removed from below the

surface. He approximated the surface by the average between the maximum and average

topographic envelopes, based on locations of the erosion surface remnants. The 110 -240 km2

estimate includes an upper bound in which the maximum topographic envelope approximated the

erosion surface. Over 10 Myr and cross-strike distances of 315, 285, and 260 km, we converted

these areas into denudation rates ranging from 0.038 -0.089 mm/yr (fig. 5). Erosion surface

remnants are bracketed by -3 -10 Ma ages (Kennan and others, 1997). This suggests that the San

Juan del Oro surface formation was time -transgressive. Therefore, using the maximum age in

our rate calculation implies that these denudation -rate estimates are a minimum even though they

incorporate the upper bounds of material area removed. Constrained by K -Ar -dated tuffs

deposits near the top of the erosion surface, further mass balance calculations implied 400 m of

17

denudation since surface incision began at 2 -3 Ma (Gubbels, 1993). We calculated this local

incision rate as 0.13 -0.2 mm/yr over the last 2 -3 Myr. Actual timing of incision into the erosion

surface is more spatially variable and bracketed by tuffs ranging in age from 1.5 -3.5 Ma (Keenan

and others, 1995 and 1997).

We calculated additional denudation -rate estimates from sediment volume data in the

southern foreland basin (fig. 5). Gubbels (1993) estimated the amount of material denuded over

the last 2 -3 Myr as 140 +/- 20 km2 from Tertiary strata imaged by seismic data in the foreland

basin along a single profile at 20 °S. Assuming that the Tertiary sediment was eroded from the

most proximal part of the fold- thrust belt, which is 300 km wide, we calculated a denudation rate

range of 0.13 -0.27 mm/yr. This estimate incorporates an age range of 2 -3 Ma and the error in

the material area. Flemings and Jordan (1989) reported a Subandean basin fill rate of 82 m2 /yr

over the last 5 Myr at 22 °S. With a proximal fold- thrust belt width of 250 km, we estimated 0.33

mm/yr of denudation over the last 5 Myr. Mass balance calculations that suggest the Andean

foreland basin has been overfilled imply that the basin fill denudation estimates may be

minimum rates (Gubbels, 1993). In comparison, although there are no reported sediment volume

data for the north, estimates of modern Tertiary foreland basin thicknesses range from -3 km in

the south to more than twice as much (-4 -7 km) in the north (Horton and DeCelles, 1997). This

suggests that denudation -rate estimates from sediment volumes in the north might be greater than

the south by a factor of two or more.

We also calculated denudation rates from structural cross -sections and their restorations for

both the north and south by estimating the material area removed from above the present Andean

fold- thrust belt topographic surface (fig. 5) (cross - sections from McQuarrie, 2001). We

measured line lengths of the oldest continuous contacts above the present topography, added the

18

appropriate average thickness of pre -Tertiary rocks for each physiographic province estimated

from the restored sections, and added that to additional areas of older rocks above the

topographic surface. For the south at -20 °S, we used average thicknesses of 1.5 km for the

Cretaceous section in the EC, and 1 km in the IA and SA. For the north at -17 °S, we used

average thicknesses of 1.3 km for the Silurian, 800 m for the Devonian, and 800 m for the

Cretaceous sections, respectively. We estimate 1297 km2 of material was removed from the 210

km wide fold -thrust belt in the north, and 1264 km2 from the 380 km wide fold -thrust belt in the

south. We estimate 425 km2 of material removed from the 50 km wide SA in the north

compared to 192 km2 from the 130 km wide SA in the south.

DeCelles and Horton (in press) traced the early to middle Tertiary migration of the Andean

foreland basin from the western Cordillera to its present location. This suggests that Tertiary

foreland basin deposits were present across the entire fold -thrust belt that have since been largely

removed by erosion. Assuming a 3 km thick succession of Tertiary foreland basin deposits

across the entire fold -thrust belt, our estimates of material area removed increases to 2256 km2 in

the north, and 2357 km2 in the south. This adjusts our SA estimates to 627 km2 and 432 km2 in

the north and south, respectively. Favoring the most recent estimates of Andean chronologic

evolution, we assume the EC fold -thrust belt began deforming at 40 Ma, and the Subandean zone

at 15 -20 Ma (McQuarrie, 2001). We estimate 0.08 -0.16 mm/yr of denudation in the south and

0.15 -0.27 mm/yr in the north. The lower values were calculated for just the Paleozoic and

Mesozoic rock areas and the higher values reflect the areas that incorporate the additional 3 km

of Tertiary foreland basin deposits. For the SA, we estimate 0.06 -0.22 mm/yr of denudation in

the south compared to 0.43 -0.84 mm/yr in the north. These SA estimates similarly reflect the

differences in material estimates as well as an age range of 15 -20 Ma. The error in material area

19

was equal to that of the original cross -sections, which was -10% (N. McQuarrie, person.

commun., 2002). Adjustment in age by 5 Myr or less has negligible effect on the denudation

rates. Table 3 summarizes all denudation -rate estimates, their range of values, source, type, and

associated uncertainties.

The most significant controls on the long -term denudation rates from calculated material

volumes are the time and length scales over which they are averaged. Theoretically, these

estimates of eroded material incorporate all of the geomorphic, geologic, and climatic factors

that influence erosion. Fold -thrust belt width is well constrained by map locations, but the

assumption of most proximal width may not be appropriate. Perhaps material volume, or cross -

sectional area alone is a more appropriate unit in which to compare them. This is 140 km2 over

the last 2 -3 Myr, 82 m2 /yr over the last 5 Myr, and - 120 km2 over the last 10 Myr for the

estimated derived from seismic data (Gubbels, 1993), Flemings and Jordan's (1989) estimate of

basin fill rate, and Gubbels' (1993) estimate from the erosion surface and DEM analysis,

respectively. This results in rates of -56 m2 /yr, 82 m2 /yr, and -12m2/yr. For these cases, the

differences in areas are equivalent to the differences in magnitude of the rates. For the structural

cross -sections, the areas of material removed are 1297 -2256 km2 and 425 -627 km2 for the total

fold -thrust belt and Subandean zone in the north, respectively. For the south, the estimates are

1264 -2357 km2 and 192 -432 km2, respectively. In this case, the areas are nearly equal for the

entire fold -thrust belt, but approximately a third larger for the Subandean zone in the north. This

shows a strong control of the fold -thrust belt geometry. The southern denudation rates are half

those in the north (0.08 -0.16 vs. 0.15 -0.27 mm/yr). For the Subandean zone, denudation rates in

the south are a quarter of those in the north (0.06 -0.22 vs. 0.43 -0.84 mm/yr). The material area

estimates imply more similar amounts of material removed over the entire fold -thrust belt than

20

the denudation rates. However, both perspectives display the same trends to different degrees:

greater denudation in the north, denudation increasing toward the present, and the disparity in

denudation between the north and south increasing towards the present.

RIVER PROFILES

Form, Slope, Drainage Area, Channel Width, Lithology, Climate

There are many controls on the morphology of longitudinal river profiles. All of the controls

affect fluvial erosion by influencing channel gradients. In order to investigate regional to local

controls on river gradients and incision, we constructed a longitudinal profile database of the

major rivers that run transverse to the EC fold-thrust belt. This database describes downstream

changes in features important for fluvial incision (fig. 7). Latitude, longitude, elevation, and

distance downstream data were calculated by digitizing channel intersections with topographic

contours (20 m contour interval) of 1:50,000 scale topographic maps and computed the straight -

line distance between these points. We calculated local channel gradients using nearest

neighboring points along these elevation profiles. We calculated drainage area using RiverTools

2.4 to generate channel profiles of drainage area from the 1 km resolution USGS GTOPO30

DEM. To obtain channel widths (cw, measured as floodplain width), we rectified thirty -meter

resolution Landsat TM images using PCl/Geomatica 8.2.1. Channel width uncertainties range

from 5 -30% for wide channel reaches (cw > 100 m) up to 100% or more for the narrow reaches

(cw < 30 m). Channel widths of river reaches masked by cloud cover ( >3% of the database)

were estimated to be the average between the measured values at either end of the covered

region. In order to minimize the errors in channel widths, observations and illustration of their

changes were limited to 5 -point moving averages of the actual data to focus on the larger scale

trends (fig. 7). Mean December -January -February (DJF) precipitation (from 1983 -1991) at each

21

point along profile was taken from a 115th degree resolution grid created by spherical kriging

interpolation between more than 75 precipitation gauging stations (fig. 3). We justify use of DJF

mean precipitation values by the fact that the rainy season (January-March) represents 50 -90% of

the annual stream discharge and 57 -90% of the total annual sediment load of these rivers (Guyot

and others, 1988, 1990, 1994). This more closely represents effective discharge than average

annual precipitation by focusing on rainfall values associated with peak stream flow conditions

as does our measured floodplain widths (Lave and Avouac, 2001). Lithologies, faults, and other

geologic structures were collected from geologic maps (Geobol, 1992; Dunn and others, 1995;

Geobol, 1996a and 1996b; Geobol -YPFB, 1996; McQuarrie, 2001).

The river profile shapes do not match mathematical functions in detail (fig. 7) (Shepherd,

1985; Snow and Slingerland, 1987). However, to first- order, the Coroico river profile fits a

power function (R2 fit = 0.82) and the large southern rivers, such as the Pilcomayo, more closely

fit linear functions (R2 fits = 0.96 -0.99) (fig. 8). Ohmori (1991) recognized a trend in Japanese

rivers where reach -scale profile fitting to exponential, power, and linear functions corresponded

to river -reach evolutional phases of deposition, sediment transportation, and mass equilibrium,

respectively. Applying this system to the entire river profiles suggests that the Rio Coroico is in

a transportation phase whereas the southern rivers, save Rio Parapeti, are in a mass equilibrium

phase. This further supports the other observations of greater denudation in the north.

There are numerous spatial correlations between river gradient, drainage area, channel width,

geologic structures, and lithology (fig. 7). Stream gradient decreases downstream but

knickpoints (slope breaks or interruptions) frequently disrupt this trend. Steeper river reaches

coincide with across structural -strike orientations whereas reaches that parallel strike are

significantly lower in gradient. Increases in drainage area downstream are erratic and dominated

22

by the random distribution and size of tributaries. Large jumps in drainage area sometimes

coincide with profile slope breaks to a reduced average gradient. Channel width generally

correlates inversely with stream gradient and slope. One notable exception is the Rio Grande

reach from 175 -275 km (downstream distance) where the channel width is large in comparison to

the relative gradient suggesting a secondary control. From Landsat image inspection, this reach

corresponds to a zone of massive sediment input by tributaries delivering landslide and debris

material into the main channel. Landslide toes and debris cones control meanders in this river

reach characterized by large channel widths. Perhaps sediment delivery can influence channel

width. Finally, in contrast to most river profile models, channel width does not increase

systematically downstream (for example Lecce, 1997; Knighton, 1999).

All of the river profiles cross the same stratigraphy of Paleozoic marine siliciclastic rocks,

Permian and Mesozoic nonmarine rocks, and Tertiary synorogenic sedimentary rocks

(McQuarrie, 2002b). Dominant lithofacies include quartzites, sandstones, siltstones, shales, and

conglomerates with a general decrease in erodibility with age and concurrent metamorphism (N.

McQuarrie, personal commun., 2002). Mechanical stratigraphy generalizations identify the

Silurian and Cretaceous rocks as relatively weak (McQuarrie and Davis, 2002). Along these

river profiles, steeper river reaches generally coincide with the older, stronger rocks. However,

some reaches over more resistant rocks, such as the Ordovician, are rather low in relative

gradient when coincident with significant, large -scale (1 -10 km) surface deformation (fig. 7).

Profile knickpoints are coincident with lithologic boundaries, channel constriction, folds, and

faults. The largest knickpoints are concurrent with the major surface faults that mark the large

topographic steps (figs. 1 and 7). Rapid increases in drainage area from tributaries often

accompany increases in channel width, reduced gradient, and occur adjacent to the major

23

knickpoints. Mean DJF precipitation does not vary much along profile but maxima are located at

elevations of 700 -2500 m and are coincident with flats above the major southern river

knickpoints.

Our river profile and channel network database shows distinctive spatial patterns between the

northern and southern regions of the Bolivian Andes (figs. 2 and 7, table 1). Relatively, the

north is characterized by higher drainage density, longer low -order channels with greater

elevations drops, lower source density (1St -order channels), and lower sinuosity (table 1). As

already mentioned, river profile- fitting to mathematical functions imply more denudation in the

north than the south. Average profile gradients decrease southward. In the north, Rio Coroico

possesses the largest slopes, mean DJF precipitation, and narrowest channel widths. The Rio

Coroico has an average DJF precipitation value along profile that is 14% higher than the Rio

Pilcomayo and 39% higher than Rios Pilaya and San Juan del Oro. Large channel width, low

channel gradients and mean DJF precipitation characterizes the southern most Rio San Juan del

Oro. The topographic steps that bound the two physiographic province transitions (traced by the

Main Andean and Main Frontal Thrusts) are manifested by increasingly large, better -defined

knickpoints from the northern Rio Grande to the southern Rio Pilaya. The Rio Coroico profile

ends before reaching these major steps but is graded to less than 1% slope for the remainder of

its course to the Beni plain. The southward increase in knickpoint morphology indicates that

there is an equivalent decrease in fluvial erosion because large rivers tend to grade their profiles

quite rapidly (Bull, 2002). This combination of erosion indicators also suggests greater

denudation and relief production in the north.

24

Stream power and Incision Rate

Bedrock -river processes play a key role at the interface between tectonics and denudation by

governing landscape morphology and relief, influencing landscape response times, and

controlling the boundary conditions of hillslope processes (Whipple and Tucker, 1999). The

stream -power model can approximate fluvial incision into bedrock. Stream -power has become

the most widely used parameter for calculating incision in models of landscape evolution (Chase,

1992; Howard, 1994; Howard and others, 1994) and longitudinal profiles of rivers (Seidl and

others, 1994; Hancock and others, 1998; Pazzaglia and others, 1998; Sklar and Deitrich, 1998;

Knighton, 1999; Snyder and others, 2000; Lave and Avouac, 2001; Roe and others, 2002). This

model represents the most important controls on river incision, channel gradient (S) and drainage

area (A, discharge proxy), as a power function;

E= KAI°Sn (1)

where E is the incision rate and K is the erosion coefficient that depends on lithology, sediment

load, and climate (for in -depth discussion of stream -power see Whipple and Tucker, 1999). K,

m, and n values are generally held constant along profile despite local changes in the conditions

embedded in K. Variations on the paired values of m and n result from the emphasis on one of

three controls: total stream -power (m = n = 1), unit stream -power (m = 1/2, n = 1), and shear-

stress (m =1/2, n = 2/3) (Finlayson and others, 2002). Embedded in the unit stream -power model

is the assumption that channel width (cw) scales with effective discharge as cw = aQ". Q is

effective discharge, and a is a coefficient. Assuming steady -state landscape conditions,

calibrated values of K, m, and n result in K varying four orders of magnitude (10- 3- 10-'), and m =

-0.4, and n = -1 (Stock and Montgomery, 1999; Snyder and others, 2000; Whipple and others,

2000; Kirby and Whipple, 2001). These results as well as other studies imply that the unit

25

stream -power and shear -stress models are more accurate for fluvial incision (Lave and Avouac,

2001; Finlayson and others, 2002).

In order to quantify incision along these profiles, we calculated unit stream -power using the

calibrated values of K = 4.32 x10-4, m = 0.4, and n = 1 on the Siwalik Hills of Nepal from Kirby

and Whipple (2001) (fig. 7). The largest error source in the stream- powered incision rates is the

value of K that varies over several orders of magnitude and is dependent on diverse fluvial

conditions. Relative differences in our incision rates are the most important, since stream -power

was applied assuming a uniform K value. The absolute values of incision may vary by an order

of magnitude or more. Our reported incision values may be too high as K decreases with

increasing rock strength and the K value we have used was calibrated on weaker rocks than those

of the eastern Bolivian fold -thrust belt (Stock and Montgomery, 1999). However, the K value

we used is within 25% of the average of all reported mean values (Stock and Montgomery, 1999;

Snyder and others, 2000; Whipple and others, 2000; Kirby and Whipple, 2001). Although some

reaches of these rivers have minor amounts of alluvial fill, the generalization that these rivers are

bedrock rivers on orogenic time scales makes the stream -power model applicable.

Figure 7 shows five -point moving averages of our fluvial incision results along each profile.

Our results show 0.02- 0.3 mm/yr of incision. Averages of profile incision decrease southward

from 0.08 mm/yr for Rio Coroico to 0.02 mm/yr for Rio San Juan del Oro. Average incision

rates on the Rios Grande, Pilcomayo, Pilaya, and Parapeti are 0.04, 0.03, 0.05, and 0.03 mm/yr,

respectively. High rates of local incision (0.05 -0.3 or more mm/yr) are coincident with river

knickpoints, especially the large ones at lower elevations where drainage area is also increased.

Channel gradient most directly influences incision, especially since large tributary influenced

jumps in drainage area are often accompanied by a reduction in grade to less than 1/2 %. Profile

26

averages of these rates are on the lower end of our other calculated rates. They are coincident

with those estimated from sediment flux and the San Juan del Oro erosion surface. The high

values of localized incision, on the order of 0.1+ mm/yr, are similar to the rates estimated from

fission -track cooling ages, and our rates calculated from Tertiary basin deposits and cross -

sections.

The stream -power equation shows that river profiles are controlled by a kinematic wave

equation (Whipple and Tucker, 1999) such as;

ah/at = KAm(ah/ax)n (2)

with h = height, t = time, x = horizontal distance, ah/at = E, ah/ax = S, and n = 1. Describing

incision rate and slope as first derivatives expresses wave migration in only the upstream

direction. Profile knickpoints will translate upstream as a wave with speed directly proportional

to Am;

w = KAmT K, T = constant (3)

with w = knickpoint width and T = total duration of fault activity. The larger the river, the faster

the knickpoint moves upstream and the larger the knickpoint width. Therefore, knickpoint

widths generated by the same source (for example faulting) plotted against rivers of variable

sizes (drainage area) will produce a linear relationship whose slope is equal to m. Figure 9

schematically shows knickpoint width as a function of no incision and two end -member cases of

a small and large river with incision. With no incision, fault motion generates a knickpoint with

a uniform slope and narrow width that would remain intact indefinitely (fig. 9A). With incision,

a fault- generated knickpoint travels as a wave upstream. The stair -step pattern represents the

fact that knickpoints associated with long -lived fault activity (Myrs) are a composite of multiple

small knickpoints, each associated with specific fault motions (fig. 9B and 9C). Both individual

27

knickpoint widths and the composite knickpoint width are wider for the large river case

compared to the small river end -member case (fig. 9C).

We identified six rivers of multiple sizes with knickpoints related to the Main Frontal Thrust.

As discussed above, this fault is coincident with the surface trace of the basement megathrust

that controls the 1 km topographic step at the IA -SA transition (figs. 1 and 7). This basement

thrust has been active for approximately the past 15 Myr (McQuarrie, 2001). We defined

knickpoint width in two ways: 1) by the slope break width on the elevation profile as

schematically shown in figure 9 2) by the width of increased incision as defined by the calculated

stream power curve (figs. 7 and 9). This calibration suggests that m = 0.35 +1- 0.15. The error

of m was estimated by varying the r- squared fit by approximately one half in either direction

(fig. 9D). Although a recent study has shown that the unit stream -power model may be

inaccurate (Tomkin and others, XXXX), this calibration of m and our focus on the relative,

spatial variations in incision affirms the general applicability of it here.

Stream powered Incision: Controls and Feedbacks

The incision of bedrock- rivers controls the morphology and relief of the northeastern Bolivian

Andes (Safran, 1998). Fluvial erosion, as defined by stream- power, is controlled by river profile

curvature and slope (fig. 7). Stream- power's dependence on slope guarantees profile knickpoints

to possess high incision rates. Large topographic steps that are controlled by basement

megathrust terminations, control the largest river knickpoints. Faults, folds, and lithologic

boundaries control the smaller knickpoints. Reach -scale slopes are influenced by rock type,

erodibility, 1 -10 km scale deformation, as well as reach orientation with respect to structural

strike, and tributary addition at lower elevations. As noted by others, these various, identifiable

28

knickpoint controls create non-linear changes in stream -power along the profile of a river (fig. 7)

(Miller, 1990; Lecce, 1997; Knighton, 1999).

DISCUSSION

The combination of observed trends in the denudation data and corroborating evidence for the

feedbacks and controls on denudation in the central Andes suggests that erosion has influenced

the kinematic and topographic evolution of the Bolivian Andes. We explore this idea by

discussing the trends of our denudation -rate synthesis, their implications, and contextualizing

them within the history of uplift and climate in the central Andes. Furthermore, we briefly

discuss the extrinsic controls of subduction and global atmospheric circulation patterns on the

Bolivian orogenic wedge as well as its critical behavior and approach to steady state.

Denudation: Trends and Implications

Despite the uncertainties in the various denudation -rate estimates we have calculated and

synthesized, two general conclusions are evident: 1) denudation rates have increased in time

towards the present 2) a disparity in denudation from north to south has persisted for a long time

and has possibly increased to double or more in recent times. Figure 10 illustrates these two

trends by plotting the ranges of denudation rates versus the time scales over which they are

averaged from the estimates listed in Table 3. Incision rates calculated by stream -power are

excluded because they are independent of time and their absolute values are poorly constrained.

Our short -term denudation rate averages (- 0.5 -1.5 mm/yr) from sediment flux are larger than

our long -term denudation rates (- V0.05 -0.8 mm/yr) (fig. 5). Similarly, Clapp and others (2000)

found that long -term sediment production rates estimated from cosmogenic isotopes were more

than 50% lower than those estimated from short -term sediment gauging data in an arid basin of

southern Israel. They interpreted their results as representative of sediment evacuation following

29

a period of greater weathering and lower sediment transport. This may be the case for the

eastern Bolivian Andes as well. In contrast, Kirchner and others (2001) showed that long -term

average denudation rates from cosmogenic isotopes and fission -track thermochronology were 17

times greater than sediment -flux data collected over 10 -84 years from montane drainages in

Idaho. They concluded that 70 +% of the total long -term sediment delivery from mountain

catchments is not recorded by the decadal- averaged rates from sediment flux data sediment

delivery and is therefore highly episodic. They suggested that sediment -flux data can greatly

underestimate long -term sediment flux in some montane environments. Opposite trends in

denudation rates averaged over multiple time scales suggest that local conditions, such a climate,

must control erosion rate magnitude among similar estimate sources as Hallet and others (1996)

demonstrated for glaciated regions. Climate has been recognized to play a large role in

controlling long -term terrestrial erosion for northern South America by climate and

sedimentation rate correlations in the Amazon fan (Mikkelsen and others, 1997; Harris and Mix,

2001).

With increasing denudation rates through time, the eastern Bolivian Andes appear to be in an

increasing state of sediment evacuation and denudation at the orogen and various montane basin

scales. Our long -term rates are numerous, and are at most -80% of our short -term rates when

compared within their respective north and south regions. The most likely explanation of both

the increasing rates and disparity towards the present is that the numerous positive feedbacks that

enhance denudation in this mountain system are responsible. Assuming they are, we would

certainly expect denudation rates and the north -south disparity to increase through time as the

orogen evolves. Numerical modeling has resulted in similar conclusions when simulating the

windward side of an eroding orogen (Willett, 1999).

30

Synthesis of Basin - Morphometry, Denudation, and Incision: North vs. South

The northern and southern Bolivian Andes possess the same controls on basin denudation yet

contrasts in the distribution of basin -morphometry, channel network statistics, and controls on

river incision. Basin denudation similarities, and channel network and incision contrasts

combine to provide more complete evidence for greater denudation in the north. Synthesizing

the controls and feedbacks that influence basin denudation and fluvial incision provides a

framework for understanding the evolution of the eastern Bolivian Andes. Figure 11

schematically summarizes all of the observations presented in this paper and is the foundation for

the discussion.

In summary (fig. 11), high basin relief, longer channels with larger elevations drops, high

drainage density, and high denudation rates characterize the northern Beni basin. High slopes

and more widespread high incision characterize the northern. Rio Coroico. Low relief, drainage

density, slopes, denudation rates, and high sinuosity characterizes the southern Pilcomayo basin.

More sporadic downstream changes in stream -power and more localized areas of high incision

characterize the southern rivers. Localization of high incision in the south is controlled by large

topographic steps, and to a lesser degree, lithologic boundaries and geologic structures. Areas

above the major knickpoints in the southern EC have yet to "feel" the associated rapid incision

downstream. Small basins are most immediately graded to the trunk- rivers for which they are

tributaries. Basins located on the benches above the topographic steps in the south are graded to

semi -regional flats and hence subject to less denudation as a result of low local stream -power

and relief, in comparison to basins located elsewhere. Furthermore, larger fold -thrust belt width

and greater spacing between anticlines is diagnostic of the southern SA (McQuarrie, 2002b).

The greater spacing of folds in the southern SA facilitates more strike- parallel river orientations

31

in synclinal valleys where slopes are low, fluvial erosion is reduced, and more sediment is stored.

The existence of large knickpoints in the southern rivers, which the stream -power model implies

should be most effective at degrading because of their large discharge (fig. 9), is additional

testament to lower erosion in the south. The syntheses of basin denudation and incision

observations schematically demonstrate feedbacks between uplift, kinematics, climate, and

geomorphology on fluvial erosion in the eastern Bolivian Andes (fig. 11). Although it is difficult

to demonstrate how long the drainage networks we describe today have persisted in time,

evidence from ancient fluvial megafan deposits suggests that the southern drainage systems may

be inherited from the mid -Tertiary (Horton and DeCelles, 2001).

History of Uplift, Climate, and Glaciation in the Bolivian Andes

The history of uplift and climate change in the central Andes provides a broader context

within which to evaluate our two major observations of increasing erosion rates and their north -

south disparity through time. Evidence exists for significant uplift since the late Miocene in the

Eastern Cordillera, a shift to aridity at 15 Ma in the Altiplano, and an increase in erosion at 10 -15

Ma (Gregory -Wodzicki, 2000). Gregory -Wodzicki (2000) concluded that since there was global

evidence for drying at 15 Ma, the aridity shift was probably the result of a climate change and

not an uplift- induced rain shadow effect. She also concluded that the combined evidence of

paleoelevation from fossil floras, the San Juan del Oro erosion surface, and crustal shortening

was enough to demonstrate that significant uplift occurred in the Altiplano and Eastern

Cordillera since the late Miocene. Significant Miocene uplift certainly could explain the

concomitant increase in erosion at 15 Ma inferred from the fission -track data and an increase in

terrigenous flux to the Amazon fan (Gregory -Wodzicki, 2000; Anders and others, 2002).

Regardless of whether it was climate change or uplift or both, average denudation rates

32

calculated in this study that integrate over more than the last 10 -15 Ma should be lower than

estimates that do not. Taken at face value, this is the case for the denudation rates presented,

especially when the comparisons are limited to within the same north or south region (fig. 10).

This phenomenon is most evident for the southern region data where more long -term

denudation -rate estimates exist. Furthermore, Peizhen and others (2001) proposed that a global

increase in sedimentation rates in Pliocene time was the result of increased erosion forced by

increased climatic variability. Though they report no evidence from South America, recognition

that global Plio- Quaternary climate possesses more variability then previous time supports the

trend of our results of increased erosion rates in more recent time.

Strong evidence for the influence of glaciers on relief, erosion, and exhumation (Hallet and

others, 1996; Brozovic and others, 1997; Meigs and Sauber, 2000) warrant additional discussion

here in the context of climate change in the central Andes. Fox (1993) showed that snowline

altitude is very responsive to precipitation in the arid, high- altitude areas of the central Andes.

Since the late Pleistocene glacial maximum, the aridity of the south has resulted in limited

glaciation of even the 6 km summits in southern Bolivia (Fox, 1993). The aridity of the highest

mountain regions may be part of the reason why Quaternary glaciation was much more extensive

in the north relative to the south (Clapperton, 1984). An increase in aridity, as is observed in the

Miocene in the central Andes, would further enhance relief, especially at mid -latitudes (Brozovic

and others, 1997). Numerical modeling of alpine glaciation showed that glaciers with significant

basal freezing can increase relief in tectonically active mountain belts (Tomkin and Braun,

2002). Perhaps this is the case for the northeastern Bolivian Andes where denudation is

enhanced by greater relief (fig. 10).

33

Evolution of the Bolivian Andes

In this section we briefly discuss the evolution of the Bolivian orogenic wedge. We focus on

the external (subduction, global climate patterns), surficial, and internal controls on the orogenic

wedge, as well as critical taper, kinematic models, and steady- state. Figure 11 schematically

illustrates and summarizes the contrasting features of the north and south fold -thrust belt wedges

by integrating our observations from the geomorphology, geology, climate, and kinematics with

those of Horton (1999).

External controls. - -- One external tectonic control on the Bolivian wedge is the subduction

process. As noted above, today the Nazca plate subducts uniformly at -30° below Bolivia

(Tinker and others, 1996). Although the segments of flat -slab subduction to the north and south

evolved from more steeply dipping geometries illustrating the dynamic aspect of subduction, we

assume that the present geometry is representative since the mid -Tertiary (Isacks, 1988). Of

additional concern are the convergence and continental margin dynamics since deformation

began in the Eastern Cordillera at - 40 Ma. In general, the Nazca plate convergence direction

has been consistently oriented northeast- southwest (Pardo -Casas and Molnar, 1987). If the

present Bolivian margin morphology has been consistent back through time, it describes more

oblique subduction to the south and more orthogonal convergence to the north. This implies

more convergence in the north. Although this would suggest more shortening north of the bend,

there is actually more to the south (McQuarrie, 2002b). In the Miocene, the convergence

direction shifted clockwise to a more east -west orientation such that convergence has since been

more uniform across Bolivia. McQuarrie (2002a) used restored amounts of shortening to

reconstruct the western continental margin back through time to show that the bend we see today

only began to take shape by -20 Ma. Previous to that time, the Bolivian continental margin

34

morphology was much straighter, striking northwest- southeast at Bolivia. This might have

implications for Miocene versus older stages of central Andean orogenesis. Synchronous

acceleration in shortening with a slowing of convergence rates at about the same time elucidated

from geologic and geodetic data suggests increased plate coupling towards the present as would

be expected by continued growth of the Andes (Hindle and others, 2002). It also implies that an

increasing amount of the convergence is being taken up by mountain building. The subduction

geometry and plate coupling is uniform enough across Bolivia that the dynamic effects to the

Bolivian orogenic wedge have presumably been spatially uniform and therefore have not

differentially affected the wedge.

Global atmospheric circulation. - -- The other important external control on the eastern

Bolivian Andes is the global atmospheric circulation patterns that influence precipitation. As

mentioned above, the Bolivian Andes have been at approximately the same latitude since

deformation began in the Eastern Cordillera (Gordon and Jurdy, 1986). This establishes the

current atmospheric circulation patterns and their influence on precipitation as initial conditions

for deformation east of the Altiplano (Montgomery and others, 2001). Therefore, the orographic

influence on precipitation aside, the north has always received more precipitation from the wetter

northeasterly prevailing winds. Global climate changes (climatic cooling and increased

variability) are assumed to occur uniformly across the entire region, but as shown with the

example of glaciation, the effects on denudation across the region can be quite different. Since

precipitation is enhanced in the north and reduced in the south both by global atmospheric

circulation patterns and by orographic effects it is difficult to separate them. Instead, we

conclude that both play a role in influencing the disparity in denudation across the orocline.

Surface Controls on the Bolivian Orogenic Wedge

35

What might be the important controls on the surface boundary conditions of the Bolivian

orogenic wedge? The relief production of glaciers forced by climatic cooling both enhances

exhumation and longer -term denudation, but might also inhibit subsequent sediment delivery

from their basins and reduce subsequent short-term fluvial basin denudation (Blodgett, 1998;

Tomkin and Braun, 2002). In essence, the glaciers scour down to resistant bedrock, produce

large amounts of sediment during their existence and retreat, followed subsequently by little

erosion and sediment delivery after most of the glacial deposits have been removed. The lack of

erosion is a result of minimal stream -power controlled by both the aridity at such high elevations

(which controls glaciation) and the exposure of dominantly resistant substrate with little to no

regolith on the hillslopes. Previously glaciated valleys and cirques are then controlled by some

threshold for rockfalls and rock debris avalanches as observed by Blodgett (1998). This is

reminiscent of the bedrock landsliding threshold that maintains regionally uniform slopes

independent of glaciers as suggested by Burbank and others (1996) for the Himalayas. The

limits to relief and erosion imposed by glaciers (Brozovic and others, 1997), and rockfall

thresholds are the important controls of the high topography of the Bolivian EC. The remaining

surficial controls imposed on the wedge are the previously discussed feedbacks between the

kinematics, stream- powered incision, geomorphology, and basin denudation (fig. 11).

Critical Taper of the Bolivian Orogenic Wedge

Critical taper theory asserts that a fold-thrust belt wedge tends towards a geometrically stable

(critical) condition where the horizontal compressive force and gravitational body force sum to

balance the basal friction resisting the fold- thrust belt wedge's advance (Dahlen and Suppe,

1988). The topographic slope and basal decollement form angles with respect to horizontal that

sum to define wedge taper. The critical condition is dependent on the mechanical properties of

36

the wedge, including internal strength and strength of the basal decollement which may be

affected by fluid pressure. There are three basic wedge conditions. A subcritical condition

consists of internal deformation and no wedge advance as the result of insufficient taper. A

critical condition involves wedge advance by accretion of new material at the base in self - similar

form. A supercritical wedge has potential for stable sliding along a single basal thrust and is

likely to return to critical conditions by frontal accretion or extension upwedge.

Horton (1999) applied the theory of critical taper to the Bolivian Andes. Using GPS

measurements, he concluded that the northern Bolivian orogenic wedge is subcritical, whereas

the southern wedge is in a critical condition. This was an unexpected result because greater

relief, larger topographic front and decollement angle should facilitate conditions closer to

criticality in the north (fig. 11). He hypothesized that erosion controlled the kinematic evolution

of the Bolivian Andes. In the north, long -term subcritical conditions were promoted through

continuous, rapid erosion. In the south, low denudation assisted in maintaining critical wedge

conditions.

Horton (1999) concluded that the lower threshold for critical taper of the southern Bolivian

orogenic wedge was probably the result of greater internal strength relative to the north (fig. 11).

He suggested that two factors imparted more strength to the southern wedge: (1) the greater

involvement of strong crystalline basement in the wedge and (2) a larger portion of high -grade

(stronger) metasedimentary Paleozoic rocks. In contrast, the north contains low -grade (weaker)

Paleozoic rocks as suggested by surface rock exposures in both regions (fig. 11). If this were

true, the weaker rocks in the northern wedge would facilitate more erosion because of their

reduced strength. This suggests that rock strength and erodibility are important for orogenic

wedge scale erosion and kinematics as well as at the river profile scale as we have observed.

37

Reduced rock strength of similar lithologies resulting from lower metamorphic grade (weaker)

would contribute to the enhancement of fluvial, glacial, and basin denudation in the north as

observed.

Kinematic Models and Erosion

Mugnier and others (1997) constructed a best -fit analogue model of the southern Subandean

zone in which the thrust system was influenced by sedimentation but no erosion. Leturmy and

others (2000) combined analogue and numerical models to evaluate the contrasting kinematics of

the northern and southern Bolivian Andes. They observed a more complicated evolution that

demonstrated that erosion restrains thrust wedge propagation. They showed that high

sedimentation rates associated with basin subsidence promoted piggyback basin formation that

first induces a shift forward in the deformation front, followed by out -of- sequence thrusting. The

northern wedge possesses larger piggyback basins, testifying to this potential story of earlier high

rates of erosion, sedimentation, and subsidence now characterized by back -thrusting, high

erosion, and little sedimentation. Furthermore, the inference of sediment evacuation of the

wedge following a period of sediment production from our similar denudation rate trends to that

of Clapp and others (2000) supports this story. This suggests that the negative feedbacks

associated with the orographic localization of precipitation, erosion, and relief may be

contributing to the hindrance of the northern wedges' advance. For the southern wedge, low

erosion, but high synorogenic sedimentation rates demonstrated by growth strata and less out -of-

sequence thrusting support a simpler story of more continuous wedge propagation eastward

(Baby and others, 1995; Horton and DeCelles, 1997).

All of the qualitative and quantitative information on denudation synthesized in this paper

provides significant evidence of substantially more sediments being evacuated from the north

38

than the south, and larger regions of material mass at greater elevations in the south (fig. 11).

The geomorphology and kinematic style of the south are both consistent with more current

sedimentation on the wedge contributing to its propagation. In contrast, little sediment storage is

taking place on the northern wedge, but the larger amount of denudation is certainly reflected in

a much deeper foredeep basin outboard to the east (fig. 11). However, retroarc foredeep basin

sediment accommodation is generally the result of flexural subsidence due to the topographic,

sedimentary, and dynamic (subduction) loads (DeCelles and Giles, 1996). The relatively rapid

reduction in hinterland topographic load by upland denudation suggests that the sediment basin

load may be controlling the foredeep subsidence in the north. In addition, more localized regions

of high incision in the south controlled by megathrusting, in turn controls the geomorphic

benches (fig. 11). The lower stream -power in the south detracts from the fluvial systems' ability

to remove sediment from hillslope processes (as seen by the Rio Grande river reach 175 -275 km

downstream distance) such as landsliding. In the north, according to Blodgett (1998), 90% of

landslide material is essentially moved immediately by the fluvial system due to its higher

stream- power, which is in turn controlled by the climate and affected by the kinematics and the

geomorphology.

Steady state and the Bolivian Orogenic Wedge

Our longer -term denudation -rate estimates cluster around 0.1 -0.4 mm/yr (fig. 5). Estimates of

rock and/or surface uplift over the last 25 Myr from numerous sources for the Altiplano and

Eastern Cordillera are remarkably similar at 0.1 -0.4 mm/yr (summarized by Gregory- Wodzicki,

2000). This suggests that over the last 25 Myrs, a steady-steady may characterize the eastern

Bolivian fold- thrust belt where uplift and denudation have been approximately balanced. The

39

uplift estimates are sparse enough that they do not permit comparison in terms of the same trends

we see in the denudation rates.

In the south at -20 °S, current rock uplift rates from space -geodetic data are 0.01 -0.1 mm/yr in

the EC and 0.5 -2 mm/yr in the SA (Lamb, 2000). Our modern denudation -rate averages are

-0.3 -0.4 mm/yr in the south and 1 -9 mm/yr in the north. Assuming the uplift estimates are

indicative of the entire orogen, uplift has decreased towards the present in the EC and may now

be outpaced by denudation (fig. 5). As suggested by Bull (2002), topographic steady -state is

difficult to reach over shorter time scales (102 -103 years), and this may be the case for the EC

where active deformation has long since moved eastward yet significant topography and relief

remains. Instead, positive feedbacks between orographic precipitation, isostatic peak uplift and

erosion (Masek and others, 1994) may now be causing a runaway from the long -term steady-

state condition. Another possibility suggested by the denudation rates is that steady -state balance

of the Bolivian orogenic wedge is maintained non -uniformly. For example, high uplift in the SA

where erosion is low due to limited topography and relief, is balanced by denudation of the same

order of magnitude in the hinterland EC where landslides, fluvial erosion, and glaciation are

outpacing uplift.

CONCLUSIONS

1. Our denudation synthesis suggests two conclusions for the Tertiary evolution of the eastern

Bolivian Andes: A) Average denudation rates have increased towards the present. B) A

concomitant north -south spatial disparity in denudation rate has existed since the Miocene and

has potentially increased to a factor of two or more in recent times.

40

2. River knickpoints and high stream- powered incision rates are controlled by the fold -thrust

belt kinematics at the orogenic scale, and more locally by the geology. In the south, localizing

most of the high incision near the EC -IA and IA -SA boundaries hinders fluvial denudation. In

the north, more widespread high incision enhances fluvial denudation and relief.

Central Andean mountain building has been influenced by its orientation at mid -latitudes,

wedge strength contrasts, and the feedbacks between uplift, climate, and erosion that are

associated with increased topography. The denudation synthesis presented here certainly begs

the question concerning to what greater precision can we extract denudation rate information

from mountain morphology and geology spanning multiple spatial- temporal scales. The most

obvious line of future research would be to produce a database of time -space denudation with

thermochronometric and cosmo- isotopic techniques in an active orogenic environment where the

structural evolution is well -constrained, strath terraces are preserved to constrain incision rates,

and comprehensive seismic, sediment volume, precipitation, topography, and sediment -flux data

exists.

ACKNOWLEDGEMENTS

This project was partially supported by two Chevron summer student research grants. Bryan

Isacks generously provided the elevation profile data for all of the rivers save Rio Coroico in

addition to the Landsat TM images. Mathius Vuille provided the mean DJF precipitation data.

We thank Nadine McQuarrie for access to her original structural restorations and guidance with

the method for estimating the amount of material removed by erosion. Brian Horton's work

inspired this project. We thank Brian Horton, Nadine McQuarrie, and Peter DeCelles, for

41

valuable discussions, preprints, insights, and training in Bolivian geology. Stephen DeLong,

Clem Chase, and Peter DeCelles provided comments on earlier versions of the manuscript.

42

REFERENCES CITED

Adams, J., 1980, Contemporary uplift and erosion of the southern Alps, New Zealand:Geological Society of America Bulletin, v. 91, p. 1 -114.

Anders, M.H., Gregory -Wodzicki, K.M., and Spiegelman, M., 2002, A critical evaluation of lateTertiary accelerated uplift rates for the eastern cordillera, central Andes of Bolivia: Journal ofGeology, v. 110, p. 89 -100.

Anhert, F., 1970, Functional relationships between denudation, relief, and uplift in large mid -latitude drainage basins: American Journal of Science, v. 268, p. 243 -263.

Avouac, J.P., and Burov, E.B., 1996, Erosion as a driving mechanism of intracontinentalmountain growth: Journal of Geophysical Research, v. 101, p. 17747 -17769.

Baby, P., Herail, G., Salinas, R., and Sempere, T., 1992, Geometry and kinematic evolution ofpassive roof duplexes deduced from cross -section balancing: Example from the forelandthrust system of the southern Bolivian Subandean zone: Tectonics, v. 11, p. 523 -536.

Baby, P., Moretti, B., Guillier, R., Limachi, R., Mendez, E., 011er, J., and Specht, M., 1995,Petroleum system of the northern and central Sub -Andean zone, in Tankard, A.J., Suarez, R.,and Welsink, H.J., editors, Petroleum Basins of South America: American Association ofPetroleum Geologists Memoir 62, p. 445 -458.

Beaumont, C., Kooi, H., and Willett, S., 2000, Coupled tectonic-surface process models withapplication to rifted margins and collisional orogens, in Summerfield, M.A., editor,Geomorphology and global tectonics: New York, Jon Wiley, p. 77 -105.

Beaumont, C., Fullsack, P., and Hamilton, J., 1992, Erosional control of active compressionalorogens, in McClay, K.R., ed., Thrust Tectonics: London, Chapman and Hall, p. 1 -18.

Beaumont, C., Jamieson, R.A., Nguyen, M.H., and Lee, B., 2001, Himalayan tectonics explainedby extrusion of a low- viscosity crustal channel coupled to focused surface denudation: Nature,v. 414, p. 738-742.

Benjamin, M.T., Johnson, N.M., and Naeser, C.W., 1987, Recent rapid uplift in the BolivianAndes: evidence from fission -track dating: Geology, v. 15, p. 680 -683.

Blodgett, T.A., 1998, Erosion rates on the NE Escarpment of the Eastern Cordillera, Bolivia,derived from aerial photographs and Thematic Mapper Images: Ph.D. thesis, CornellUniversity, Ithaca, 153 p.

Brozovic, N., Burbank, D.W., and Meigs, A.J., 1997, Climatic limits on landscape developmentin the northwestern Himalaya: Science, v. 276, p. 571 -575.

43

Bull, W.B., 2002, Steady -state landscapes -a convenient myth: Geological Society of AmericaAbstracts with Programs, v. 34, no. 6, paper p. A115 -A116.

Burbank, D.W., Leland, J., Fielding, E., Anderson, R.S., Brozovic, N., Reid, M.R., and Duncan,C., 1996, Bedrock incision, rock uplift and threshold hillslopes in the northwest Himalayas:Nature, v. 379, p. 505 -510.

Chase, C.G., 1992, Fluvial landsculpting and the fractal dimension of topography:Geomorphology, v. 5, p. 39-57.

Clapp, E.M., Bierman, P.R., Schick, A.P., Lekach, J., Enzel, Y., and Caffee, M., 2000, Sedimentyield exceeds sediment production in arid region drainage basins: Geology, v. 28, p. 995 -998.

Clapperton, C.M., 1984, The glaciation of the Andes: Quaternary Science Reviews, v. 2, p. 83-155.

Dahlen, F.A., and Suppe, J., 1988, Mechanics, growth, and erosion of mountain belts, in ClarkJr., S.P, Bruchfiel, B.C., Suppe, J., editors, Processes in continental lithospheric deformation:Geological Society of America Special Paper 218, p. 161 -178.

Dahlen, F.A., 1990, Critical taper model of fold- and -thrust belts and accretionary wedges:Annual Reviews of Earth and Planetary Sciences, v. 18, p. 55 -99.

DeCelles, P.G., and Horton, B.K., Early to middle Tertiary foreland basin development and thehistory of Andean crustal shortening in Bolivia: Geological Society of America Bulletin, v.XX (in press).

DeCelles, P.G., and Giles, K., 1996, Foreland basin systems: Basin Research, v. 8, p. 105 -123.

DeCelles, P.G., and Mitra, G., 1995, History of the Sevier orogenic wedge in terms of criticaltaper models, northeast Utah, and southwest Wyoming: Geological Society of AmericaBulletin, v. 107, p. 454 -462.

Dunn, J.F., Hartshorn, K.G., and Hartshorn, P.W., 1995, Structural styles and hydrocarbonpotential of the Subandean thrust belt of southern Bolivia, in Tankard, A.J, Suarez, S.,Welsink, H.J. editors, Petroleum basins of South America: American Association ofPetroleum Geologists Memoir 62, p. 523 -543.

Ege, H., Jacobshagen, V., Scheuber, E., Sobel, E., and Vietor, T., 2001, Thrust relatedexhumation revealed by apatite fission track dating, central Andes (southern Bolivia):Geophysical Research Abstracts, EGS XXVI General Assembly, Nice, France, v. 3, p. 624.

England, P., and Molnar, P., 1990, Surface uplift, uplift of rocks, and exhumation of rocks:Geology, v. 18, p. 1173 -1177.

44

Flemings, P.B., and Jordan, T.E., 1989, A synthetic stratigraphic model of foreland basindevelopment: Journal of Geophysical Research, v. 94, p. 3851 -3866.

Finlayson, D.P., Montgomery, D.R., and Hallet, B., 2002, Spatial coincidence of rapid inferrederosion with young metamorphic massifs in the Himalayas: Geology, v. 30, p. 219 -222.

Fox, A.N., 1993, Snowline altitude and climate in the central Andes (5 -28 °S) at present andduring the late pleistocene glacial maximum: Ph.D. thesis, Cornell University, Ithaca, 504 p.

Garrison, T., 2002, Oceanography: Pacific Grove, Thomson Learning, 554 p.

Geobol [Servicio Geologico de Bolivia], 1992, Mapas Tematicos de Recursos Minerales deBolivia, Tarija y Villazon (serie II- MTB -1B): Le Paz, Geobol, scale 1:250,000.

Geobol [Servicio Geologico de Bolivia], 1996a, Mapa Geologico de la Region de Vallegrande:Le Paz, Geobol, scale 1:250,000.

Geobol [Servicio Geologico de Bolivia], 1996b, Mapas Tematicos de Recursos Minerales deBolivia, Sucre (serie II- MTB -8B): Le Paz, Geobol, scale 1:250,000.

Geobol -YPFB, 1996, Mapa Geologico de Bolivia, scale 1:1,000,000.

Gleadow, A.J.W., and Brown, R.W., 2000, Fission -track thermochronology and the long -termdenudational response to tectonics, in Summerfield, M.A., editor, Geomorphology and globaltectonics: New York, Jon Wiley, p. 57 -75.

Gordon, R.G., and Jurdy, D.M., 1986, Cenozoic global plate motions: Journal of GeophysicalResearch, v. 91, p. 12389-12406.

Gregory -Wodzicki, K.M., 2000, Uplift history of the central and northern Andes: a review:Geological Society of America Bulletin, v. 112, p. 1091 -1105.

Gubbels, T., 1993, Tectonics and Geomorphology of the Eastern Flank of the Central Andes, 18degrees to 23 degrees S Latitude: Ph.D. thesis, Cornell University, Ithaca, 211 p.

Gubbels, T.L., Isacks, B.L., and Farrar, E., 1993, High -level surfaces, plateau uplift, and forelanddevelopment, Bolivian central Andes: Geology, v. 21, p. 695 -698.

Gunnell, Y., 1998, Present, past and potential denudation rates: is there a link? Tentativeevidence from fission -track data, river sediment loads and terrain analysis in the South Indianshield: Geomorphology, v. 25, p. 135 -153.

Guyot, J.L., Bourges, J., Hoorelbecke, R., Calle, H., Cortes, J., and Barragan Guzman, M.C.,1988, Sediment discharge from the Andes to the Amazonia along the Beni River, in SedimentBudgets: IAHS Publication, n. 174, p. 443 -451.

45

Guyot, J.L., Calle, H., Cortes, J., and Pereira, M., 1990, Transport of suspended sediment anddissolved material from the Andes to the Rio de la Plata by the Bolivian tributaries of the RioParaguay (Rios Pilcomayo and Bermejo): Hydrological Sciences, v. 35, p. 653 -665.

Guyot, J.L., Jouanneau, J -M., Quintanilla, J., and Wasson, J -G., 1993, Dissolved and suspendedsediment loads exported from the Andes by the Beni river (Bolivian Amazonia), during aflood: Geodinamica acta, v. 6, p. 233 -241.

Guyot, J.L., Bourges, J., and Cortez, J., 1994, Sediment transport in the Rio Grande, an Andeanriver of the Bolivian Amazon drainage basin, in Variability in stream erosion and sedimenttransport: IAHS Publication, n. 224, p. 223 -231.

Hallet, B., Hunter, L., and Bogen, J., 1996, Rates of erosion and sediment evacuation by glaciers:A review of field data and their implications: Global and Planetary Change, v. 12, p. 213 -235.

Hancock, G.S., Anderson, R.S., and Whipple, K.X., 1998, Beyond power: bedrock river incisionprocess and form, in Tinkler, K.J., Wohl, E.E., editors, Rivers over Rock: Fluvial processes inbedrock channels: Geophysical Monograph 107, p. 35 -60.

Harris, S.E., and Mix, A.C., 2002, Climate and tectonic influences on continental erosion oftropical South America, 0 -13 Ma: Geology, v. 30, p. 447 -450.

Hindle, D., Kley, J., Klosko, E., Stein, S., Dixon, T., and Norabuena, E., 2002, Consistency ofgeologic and geodetic displacements during Andean orogenesis: Geophysical ResearchLetters, v. 29, p. 29 -1 - 29 -4.

Hoffman, P.F., and Grotzinger, J.P., 1993, Orographic precipitation, erosional unloading, andtectonic style: Geology, v. 21, p. 195 -198.

Horton. B.K., Hampton, B.A., and Waanders, G.L., 2001, Paleogene synorogenic sedimentationin the Altiplano and implications for initial mountain building in the central Andes:Geological Society of America Bulletin, v. 113, p. 1387 -1400.

Horton, B.K., and DeCelles, P.G., 2001, Modern and ancient fluvial megafans in the forelandbasin system of the central Andes, southern Bolivia: Implications for drainage networkevolution in fold -thrust belts: Basin Research, v. 13, p. 43 -63.

Horton, B.K., and DeCelles, P.G., 1997, The modern foreland basin system adjacent to theCentral Andes: Geology, v. 25, p. 895 -898.

Horton, B.K., 1999, Erosional control on the geometry and kinematics of thrust belt developmentin the central Andes: Tectonics, v. 18, p. 1292 -1304.

Hovius, N., 2000, Macroscale process systems of mountain belt erosion, in Summerfield, M.A.,editor, Geomorphology and global tectonics: Jon Wiley, NY, p. 77 -105.

46

Hovius, N., 1998, Controls on sediment supply by large rivers, in Shanley, K.W., McCabe, P.J.,editors, Relative role of eustacy, climate, and tectonics in continental rocks: Society forSedimentary Geology (SEPM) Special Publication 59, p. 3 -16.

Hovius, N., Stark, C.P., and Allen, P.A., 1997, Sediment flux from a mountain belt derived bylandslide mapping: Geology, v. 25, p. 231 -234.

Howard, A.D., 1994, A detachment- limited model of drainage basin evolution: Water ResourcesResearch, v. 30, p. 2261-2285.

Howard, A.D., Dietrich, W.E., and Seidl, M.A., 1994, Modeling fluvial erosion on regional tocontinental scales: Journal of Geophysical Research, v. 99, p. 13971- 13986.

Isacks, B.L., 1988, Uplift of the central Andean plateau and bending of the Bolivian orocline:Journal of Geophysical Research, v. 93, p. 3211 -3231.

Jordan, T.E., Reynolds, J.H., and Erikson, J.P., 1997, Variability in age of initial shortening anduplift in the central Andes, in Ruddiman, W.F., editor, Tectonic uplift and climate change:New York, Plenum, p. 41 -61.

Kennan, L., 2000, Large -scale geomorphology of the Andes: Interrelationships of tectonics,magmatism and climate, in Summerfield, M.A., editor, Geomorphology and global tectonics:New York, Jon Wiley, p.167 -199.

Kennan, L., Lamb, S.H., and Rundle, C.C., 1995, K -Ar dates from the Altiplano and Cordilleraoriental of Bolivia: implications for the Cenozoic stratigraphy of and tectonics: Journal ofSouth American Earth Sciences, v. 8, p. 163 -186.

Kennan, L., Lamb, S.H., and Hoke, L., 1997, High altitude paleosurfaces in the Bolivian Andes:evidence for late Cenozoic surface uplift, in Widdowson, M., editors, Paleosurfaces:Recognition, reconstruction, and Interpretation, Geological Society of London, SpecialPublication, n. 120, p. 307-324.

Kirby, E., and Whipple, K.X., 2001, Quantifying differential rock -uplift rates via stream profileanalysis: Geology, v. 29, p. 415 -418.

Kirchner, J.W., Finkel, R.C., Riebe, C.S., Granger, D.E., Clayton, J.L., King, J.G., and Megahan,W.F., 2001, Mountain erosion over 10 yr, 10 k.y., and 10 m.y. time scales: Geology, v. 29, p.591 -594.

Kley, J., 1999, Geologic and geometric constraints on a kinematic model of the bolivianorocline: Journal of South American Earth Science, v. 12, p. 221-235.

Kley, J., 1996, Transition from basement -involved to thin -skinned thrusting in the CordilleraOriental of southern Bolivia: Tectonics, v. 15, p. 763 -775.

47

Kley, J., Monaldi, C.R., and Salfity, J.A., 1999, Along -strike segmentation of the Andeanforeland: causes and consequences: Tectonophysics, v. 301, p. 75 -94.

Knighton, A.D., 1999, Downstream variation in stream power: Geomorphology, v. 29, p. 293-306.

Knighton, A.D., 1998, Fluvial forms and processes: New York, Oxford University Press, 383 p.

Koons, P.O., 1994, Three- dimensional critical wedges: tectonics and topography in obliquecollisional orogens: Journal of Geophysical Research, v. 99, p. 12301 -12315.

Lamb, S., 2000, Active deformation in the Bolivian Andes, South America: Journal ofGeophysical Research, v. 105, p. 25627- 25653.

Lave, J., and Avouac, J.P., 2001, Fluvial incision and tectonic uplift across the Himalayas ofcentral Nepal: Journal of Geophysical Research, v. 106, p. 26561- 26591.

Lecce, S.A., 1997, Nonlinear downstream changes in stream power on Wisconsin's Blue River:Annals of the Association of American Geographers, v. 87, p. 471 -486.

Leturmy, P., Mugnier, J.L., Vinour, P., Baby, P., Colletta, B., and Chabron, E., 2000, Piggybackbasin development above a thin -skinned thrust belt with two detachment levels as a functionof interactions between tectonic and superficial mass transfer: The case of the Subandean zone(Bolivia): Tectonophysics, v. 320, p. 45 -67.

Masek, J.G., Isacks, B.L., Gubbels, T.L., and Fielding, E.J., 1994, Erosion and tectonics at themargins of continental plateaus: Journal of Geophysical Research, v. 99, p. 13941 -13956.

Mikkelsen, N., Maslin, M., Giraudeau, J., and Showers, W., 1997, Biostratigraphy andsedimentation rates of the Amazon fan, in Flood, R.D., Piper, D.J.W., Klaus, A., Peterson,L.C., editors, Proceedings of the Ocean Drilling Program, Scientific Results, v. 155, p. 577-592.

McBride, S.L., Clark, A.H., Farrar, E., and Archibald, D.A., 1987, Delimitation of a crypticEocene tectno -thermal domain in the Eastern Cordillera of the Bolivian Andes through K -Ardating and 40Ar -39Ar step- heating: Journal of the Geological Society of London, v. 144, p.243-255.

McQuarrie, N., 2001, The Making of a High -Elevation Plateau: Insights from the CentralAndean Plateau, Bolivia: Ph.D. thesis, University of Arizona, Tucson, 255 p.

McQuarrie, N., 2002a, Initial plate geometry, shortening variations, and evolution of theBolivian orocline: Geology, v. 30, p. 867 -870.

48

McQuarrie, N., 2002b, The kinematic history of the central Andean fold -thrust belt, Bolivia:Implications for building a high plateau: Geological Society of America Bulletin, v. 114, p.950 -963.

McQuarrie, N., and Davis, G.H., 2002, Crossing the several scales of strain- accomplishingmechanisms in the hinterland of the central Andean fold -thrust belt, Bolivia: Journal ofStructural Geology, v. 24, p. 1587 -1602.

Meigs, A., and Sauber, J., 2000, Southern Alaska as an example of the long -term consequencesof mountain building under the influences of glaciers: Quaternary Science Reviews, v. 19, p.1543 -1562.

Miller, J.R., 1990, The influence of bedrock geology on knickpoint development and channel -bed degradation along downcutting streams in south -central Indiana: Journal of Geology, v.99, p. 591 -605.

Milliman, J.D., and Meade, R., 1983, World -wide delivery of sediment to the oceans: Journal ofGeology, v. 91, p. 1 -21.

Milliman, J.D., and Syvitski, J.P.M., 1992, Geomorphic /tectonic control of sediment discharge tothe ocean: the importance of small mountain rivers: Journal of Geology, v. 100, p. 525 -544.

Molnar, P., and England, P., 1990, Late Cenozoic uplift of mountain ranges and global climatechange: Chicken or egg ?: Nature, v. 346, p. 29 -34.

Montgomery, D.R., 1994, Valley incision and the uplift of mountain peaks: Journal ofGeophysical Research, v. 99, p. 13913 -13921.

Montgomery, D.R., Balco, G., and Willett, S.D., 2001, Climate, tectonics, and the morphologyof the Andes: Geology, v. 29, p. 579 -582.

Mugnier, J.L., Baby, P., Colletta, B., Vinour, P., Bale, P., and Leturmy, P., 1997, Thrustgeometry controlled by erosion and sedimentation: A view from analogue models: Geology,v. 25, p. 427 -430.

Ohmori, H, 1991, Change in the mathematical function type describing the longitudinal profileof a river through an evolutionary process: Journal of Geology, v. 99, p. 97 -110.

Pardo -Casas, F., and Molnar, P., 1987, Relative motion of the Nazca (Farallon) and SouthAmerican plates since Late Cretaceous time: Tectonics, v. 6, p.233 -248.

Pazzaglia, F.J., Gardner, T.W., and Merritts, D.J., 1998, Bedrock fluvial incision andlongitudinal profile development over geologic time scales determined by fluvial terraces, inTinkler, K.J., K.J., Wohl, E.E., editors, Rivers over Rock: Fluvial processes in bedrockchannels: Geophysical Monograph 107, p. 35 -60.

49

Pelletier, J.D., 2002, Late Cenozoic glacial erosion and exhumation: implications for thepaleoelevation of the central Andes and Lhasa: Geological Society of America AnnualMeeting, paper n. 22 -14.

Peizhen, Z., Molnar, P., and Downs, W.R., 2001, Increased sedimentation rates and grain sizes2 -4 Myr ago due to the influence of climate change on erosion rates: Nature, v. 410, p. 891-897.

Pinet, P., and Souriau, M., 1988, Continental erosion and large scale relief: Tectonics, v. 7, p.563 -582.

Raymo, M.E., and Ruddiman, W.F., 1992, Tectonic forcing of late Cenozoic climate: Nature, v.359, p. 117 -122.

Roe, G.H., Montgomery, D.R., and Hallet, B., 2002, Effects of orographic precipitation on theconcavity of steady -state river profiles: Geology, v. 30, p. 143 -146.

Roe, G.H., Montgomery, D.R., and Hallet, B., Orographic climate feedbacks on the relief ofmountain ranges; Journal of Geophysical Research, v. XX (in press).

Safran, E., 1998, Channel network incision and patterns of mountain geomorphology: Ph.D.thesis, University of California, Santa Barbara, 327 p.

Seidl. M.A., Dietrich, W.E., and Kirchner, J.W., 1994, Longitudinal profile development intobedrock: An analysis of Hawaiian channels: Journal of Geology, v. 102, p. 457 -474.

Sempere, T., Herail, G., and 011er, J., 1988, Structural and sedimentary aspects of the Bolivianorocline: International Geological Congress, 28th Washington, D.C. Abstracts, v. 3, p. 72 -73.

Sempere, T., Herail, G., 011er, J., and Bonhomme, M.G., 1990, Late -Oligocene -early Miocenemajor tectonic crisis and related basins in Bolivia: Geology, v. 18, p. 946 -949.

Shepherd, R.G., 1985, Regression analysis of river profiles: Journal of Geology, v. 93, p. 377-3 84.

Sklar, L, and Dietrich, W.E., 1998, River longitudinal profiles and bedrock incision models:stream power and the influence of sediment supply, in Tinkler, K.J., Wohl, E.E., editors,Rivers over Rock: Fluvial processes in bedrock channels: Geophysical Monograph 107, p.237-260.

Snow, R.S., and Slingerland, R.L., 1987, Mathematical modeling of graded river profiles:Journal of Geology, v. 95, p. 15 -33.

Snyder, N.P., Whipple, K.X., Tucker, G.E., and Merritts, D.J., 2000, Landscape response totectonic forcing: Digital elevation model analysis of stream profiles in the Mendocino triple

50

junction region, northern California: Geological Society of America Bulletin, v. 112, p. 1250-

1263 .

Stock, J.D., and Montgomery, D.R., 1999, Geologic constraints on bedrock river incision usingthe stream power law: Journal of Geophysical Research, v. 104, p. 4983 -4993.

Stuwe, K., White, L., and Brown, R.L., 1994, The influence of eroding topography on steady -state isotherms: Earth and Planetary Science Letters, v. 124, p. 63 -74.

Summerfield, M.A., and Hulton, N.J., 1994, Natural controls of fluvial denudation in majorworld drainage basins: Journal of Geophysical Research, v. 99, p. 13871-13884.

Tinker, M.A., Wallace, T.C., Beck, S.L., Myers, S., and Papanikolas, A., 1996, Geometry andstate of stress of the Nazca plate beneath Bolivia and its implication for the evolution of theBolivian orocline: Geology, v. 24, p. 387 -390.

Tomkin, J.H., and Braun, J., 2002, The influence of alpine glaciation on the relief of tectnicallyactive mountain belts: American Journal of Science, v. 302, p. 169 -190.

Tomkin, J.H., Brandon, M.T., Pazzaglia, F.J., Barbour, J.R., and Willett, S.D., XXXX,Quantitative testing of bedrock incision models, Clearwater River, WA: Journal?

Tucker, G.E., and Slingerland, R., 1996, Predicting sediment flux from fold and thrust belts:Basin Research, v. 8, p. 329 -349.

Willett, S.D., 1999, Orogeny and orography: The effects of erosion on the structure of mountainbelts: Journal of Geophysical Research, v. 104, p. 28957- 28981.

Willett, S.D., and Brandon, M.T., 2002, On steady states in mountain belts: Geology, v. 30, p.175 -178.

Whipple, K.X., and Tucker, G.E., 1999, Dynamics of the stream -power river incision model:Implications for height limits of mountain ranges, landscape response timescales, and researchneeds: Journal of Geophysical Research, v. 104, p. 17661 -17674.

Whipple, K.X., Snyder, N.P., and Dollenmayer, K., 2000, Rates and processes of bedrock riverincision by the Upper Ukak River since the 1912 Novarupta ash flow in the Valley of TenThousand Smokes, Alaska: Geology, v. 28, p. 835 -838.

Wohl, E., 2000, Mountain Rivers: Water Resources Monograph 14, American GeophysicalUnion, Washington, DC, 320 p.

51

FIGURE CAPTIONS

Figure 1. Tectonics, topography, and climate of the Bolivian Andes. (A) Major

physiographic provinces are defined by average elevation as follows: Altiplano (AP),

> 3 km, Eastern Cordillera (EC), 3 km, Interandean zone (IA), 2 km, and Subandean

zone (SA), 1 km. The faults coincident with the EC -IA and IA -SA boundaries from

are the Main Andean Thrust and the Main Frontal Thrust, respectively. Inset shows

location of the Bolivian Andes in western central South America (simplified from

Horton and others, 2001). (B) Topography and precipitation profiles. Topography

data is based on a 100-km wide bin along each profile. Maximum -minimum

elevation is shaded, average is black line. Width of shaded zone is relief.

Precipitation data (circles) are projected from sites within 200 km of section line

(from Horton, 1999). High relief, narrow SA, and wet climate north of the oroclinal

bend contrasts with the low relief, wide SA, and a semi-arid climate to the south.

Figure 2. Regional hydrology of the eastern Bolivian Andes. The Beni, Pirai, and

Grande basins drain north into the Beni plain, whereas the Parapeti, Pilcomayo, and

Bermejo basins drain south into the Chaco plain. The major river basins are outlined

(solid black lines). SJDO = San Juan del Oro. The river traces (gray lines) designate

the location of the longitudinal river profiles displayed in figure 7. Note the

unnaturally truncated boundaries near the Parapeti and Pilcomayo basin headwaters.

This is an artifact of the DEM and channel network analysis. The natural boundaries

are outlined in the short dashed lines for comparison (from Guyot et al., 1990, 1994).

52

Figure 3. Mean December -January -February precipitation of the central Andes.

(A) Precipitation data sites. (B) Mean December - January -February (DJF)

precipitation (1983 -1991) grid calculated by interpolation via spherical kriging

between more than 75 sites. Note the northward increase in precipitation. Grid

resolution is 1 /5th of a degree.

Figure 4. Hypsometry of the major basins draining the eastern Bolivian Andes. See

figure 2 for locations.

Figure 5. Denudation rate estimates. AFTT = apatite fission -track thermochronology.

*Denotes time period over which that rate is averaged. Parentheses identify the

source from which the estimate was taken or from material that we have modified to

create the estimate (see text).

Figure 6.

1990,

Index map of sediment gauging stations (compiled from Guyot et al., 1988,

1993, 1994). Station codes match those in table 2.

Figure 7. Longitudinal profile data of the major transverse central Andean rivers. (A)

Elevation and stream power profiles, the latter calculated with K, m, n values of 4.32

x 10-4, 0.4, and 1, respectively (from Kirby and Whipple, 2001). See figure 2 for

locations. Regional -scale lithology codes: O = Ordovician, S = Silurian, D =

Devonian, C = Carboniferous, TR = Triassic, J = Jurassic, K = Cretaceous, T =

Tertiary, P = Paleogene, E = Eocene, 01= Oligocene, N = Neogene, Q = Quaternary.

53

Provinces, faults, structures, degree of deformation (1 -10 km scale), and reach -scale

geomorphic features are also included. The Main Frontal Thrust knickpoint widths as

defined by the longitudinal profile (kpt W) and stream power curve (SP kpt W) used

in the calculation of m are displayed as horizontal lines adjacent to the profiles (Fig.

8). (B) River profile data of elevation (LP), upstream drainage area (DA), slope

(SLP), mean December -January- February (DJF) precipitation (P), and channel width

(CW). To emphasize the regional trends, profiles shown are 2 -point (SLP, LP, P) and

5 -point (CW and stream power in (A)) moving averages of the actual data. Note the

different scales of both axes and the staggered vertical axes.

Figure 8. Power and linear function- fitting of the Rio Coroico and Rio Pilcomayo

profiles.

Figure 9. Schematic diagram showing evolution of knickpoint width from continuous

fault activity as a function of A`". (A) River profile and knickpoint with no bedrock

incision. (B) Small versus (C) large river knickpoint width with incision. (D)

Drainage area verses knickpoint width defined two ways: 1) from the longitudinal

profile (squares and solid line) and 2) by the width of the zone of increased stream

power (circles and dashed line) (figure 7). Knickpoint widths are from Rios Pilaya,

Pilcomayo, Parapeti, Azero, and Limon (see figure 2 for locations). The width used

in both cases for the two smallest rivers, Azero and Limon, are from the profiles as no

stream -power curve was calculated.

54

Figure 10. Logarithmic plot of denudation rates vs. time scale over which they are

averaged for the Bolivian Andes. Circles (north) and squares (south) are averages

and bars are their range of values, if available. Sediment -flux data are from only the

Beni and Pilcomayo basins. Data are slightly staggered over their relevant time scale

for clarity. Diagonal bar for Safran's (1998) fission -track data symbolizes that the

sample ages range over a 14 Myr time frame and the rates decrease with age (average

rate is centered over median age). Brackets represent the size of the north -south

disparities in average rates across similar source data. Gently inclined dashed lines

across the time scales connect the more accurate average rate estimates (Table 3)

showing an increase in denudation rate and north (N) - south (S) disparity towards

the present.

Figure 11. Schematic diagram showing the contrasting attributes of the northern and

southern Bolivian Andes. Solid lines are rivers, large dashed lines are anticlinal

ridges, and small dashed lines are small basin drainage boundaries. Shaded gray

regions represent areas of high incision. The two, stacked bodies in oblique cross-

section are the megathrusts (from McQuarrie, 2002b). Angles depicted are the

average angle of the topographic front and basal decollement (from Horton, 1999). D

= denudation. Abbreviations same as figure 1.

Table 1. Basin morphometry and channel network statistics.

Feature Beni Grande Pilcomayo Parapeti

Area (km2) 85336 65837 87980 7149

Relief (km) 5.60 4.70 5.23 2.73

Relief ratio 1.57 1.44 1.46 1.62Drainagedensity (km-1)

1,290 1.282 1.285 1.265

Sourcedensity (km-2)

0.267 0.423 0.469 0.350

Sinuosity 1.14 1.25 1.23 1.15

Number ofchannels *:1st -3rd order 28758 34115 49869 3081

4th -5th order 241 257 380 226th order 11 10 16 1

7th -8th order 3 4 4 N/A

Along channelslope *:

1st -3rd order .0742 .0945 .0850 .04054th -5th order .0285 .0233 .0227 .01536th order .0052 .0111 .0086 .00277th -8th order .0018 .0044 .0050 N/A

Along channellength* (km):1st -3rd order 4.5 3.5 3.3 4.04th -5th order 23.4 19.2 18.5 30.86th order 86.6 38.2 37.6 82.67th -8th order 225.6 331.5 368.5 N/A

Elevationdrop* (m):1st -3rd order 258 226 206 1244th -5th order 536 373 318 2766th order 623 333 338 2307th -8th order 536 927 1330 N/A

*order = Strahler order N/A = not applicable

Table 2. Sediment gauge data and denudation rates.See Figure 6 for station locations, Figure 2 for basin locations.

Station Elevation (m) DA (km2) TSS (106 t/yr) Denudation* (mm /yr)

Beni BasinSI 1800 272 1.4 2.150UN 1200 359 0.7 0.813TQ 1200 595 0.3 0.210TM 1185 954 1.6 0.700VB 1050 1440 8.6 2.491SR 435 4700 5.1 0.453AQ 600 10600 48 1.889Al 420 29600 103 1.451AB 284 67200 191 1.121

Pirai BasinEL 650 64 0.03 0.192ES 550 203 0.42 0.847BE 900 480 0.06 0.512AN 650 1420 3 0.865TA 600 1590 1.3 0.335LB 350 2880 2.3 0.327PE 280 4160 1.1 0.108

Grande BasinPZ 1080 4360 2.1 0.197AM 1850 9200 106 4.719MI 950 10800 26 0.986

HU 1600 11200 24 0.878

PA 1500 23700 136 2.351

PN 950 31200 203 2.665

AP 450 59800 136 0.932

Parapeti BasinSA 550 7500 19 1.038

Pilcomayo BasinNU 2300 1600 1.1 0.282SL 3100 4200 0.5 0.049AT 2500 6340 12 0.775VQ 2000 13200 22 0.687EP 2300 20100 2.4 0.049CH 2100 42900 14 0.134SJ 800 47500 31 0.258VI 340 81300 72 0.350

Bermejo BasinPJ 1200 220 0.11 0.215

CM# 2100 230 0.05 0.087

ER# 1200 290 0.06 0.097TO 1900 460 1.5 1.265OB 1900 920 0.4 0.170

Averages

J

1

7

1.0

1.5

1.5

0.7

0.4

2.5

1.6

0.4

0.3

0.4

*Station code changed from source to avoid duplication.'Estimated TDS/TSS ratio is 13% for the Beni basin, 10% for the other basins when TfS data not available. Assumes a 65%denudation component for the dissolved load and an average rock density of 2.6 g/cm (Knighton, 1998, p.81).

Table 3. Denudation -rate estimates and uncertainties.Symbols below values correspond to Figure 5. AFT = apatite fission -track

Rate Range(mm /yr) Method Type Uncertainties Interpretation

4 -14

Alandslidemapping

area assumed area -volumerelationship, age,and 10% basin storage

upper limit

0.2 -0.6 AFTcoolingages

pointsource

15% from ages,corrected for topography

accurate

0.15 -0.31o AFTcoolingages

pointsource

20% from ages,assumed closure at 5 kmdepth

first -orderapproximation

0.049 -4.7Table 2

sedimentflux

basinareaaverage

short time -scale,assumed rock densityand some TDS/TSS

minimum,accurate

0.038 -0.089.._ erosionsurface/DEManalysis

area time -transgressive (3 -10Ma) age -used 10 Ma,includes upper bound ofmaterial area

minimum

0.13 -0.2 erosionsurfaceincision

massbalance

timing of erosion surface,incision error -25 %+

0.13 -0.27- - --

basinfill

area 15% error in area,proximal source,age range 2 -3 Ma

minimum?

0.33 strat.-sections

pointsources

proximal source only,averaged fill rate

minimum?

0.08 -0.84 cross-sections

area averaged stratigraphicthicknesses, proximalsource, present dayfold- thrust belt lengthscale

accurate?

- 14°

9RAZ/L

MAP KEY \Peru

BENIPLAIN

A'-2a s

thrust fault &physiographicprovinceboundary

internationalborder

- 16°1

\%

) LaPaz C

chd6

')?60

city /town

Figure 1

rA

1

\\

18°Nr, Bolivia!

I 1

i \ AP

/, B

L

-30°S

PACIFICOCEAN

ARGENTINA

71ir60'W

1

20°Chile

- 2201

J1

70°I

EC

Potosi

IA SA

Tupiza

200 km

CHACOPLAIN

\ I

` Argentina ``66° 64° 62°

I I I

B

Oc 4co

2

0

North A'

II -2

41 'Nit

6B

E4...

c

> 2Ta

0

100 200 300cross -strike distance (km)

South

1

0

MAP KEYza)

-0(DA-0iff precipitationó

- elevation

3

100 200 300 400 500cross -strike distance (km)

(

r

14°BENIPLAIN

Peru\\- 16°

/ Lai Paz

Coroico1. Eastern Cordillera

- 2. Interandean Zone3. Subandean Zone ARGENTINA

80°w r 70°V1l 60°W

Cochabamba

GrandePira

Bolivia

Azero

1

1 /

Chile \\1

- 22°ii

L\ 70°

Figure 2

,..

I

1

j: -_,1_

Parapeti

m

CHACOBermejo ' PLAIN

,\ Argentina''

ina `\,

/ .68° 66° 64° 62°

I i I I I

Peru

- 140 11 PLAINli

/k

1

- 16° 7 La Paz

/(

_.18°

- 20°

)

- 22°1 Chile1 Dl70°

APrecipitation gauge

sitesI1

200 km

Cochabamba

r- D

CO

1D DBolivia

::REI

at, D Dm (ti

1

ii

68°

MUM

--ID

0 13 D CHACOPLAIN_/ - ---

ID 1r \ i

/ .ti

k.---) \ Argentina \-8 / 66° 64° 62°

! I I I

Figure 3

PeruBENI

- 14 350,1 PLAINB

Mean DJ F Preipr ation350 contour interval: 50 mm /yr !`§

200 km

16°100

_.

18°50

1

S

-20°i1

- 22°¡ Chile

170°

300

300

Cochabamba

350

ti

11

1_ 150° 100

68i 50 166°

Bolivia

300

250

CHACOPLAINf ---`

200 -NtArgentina \.\64 62°

I I

= 0.8E

ó E 0.6c ó>

w á 0.4

76

0.2

Pilcomayo

\ Grande

-,- BeniParapeti

Hypsometry

Figure 4

0.2 0.4 0.6 0.8Area (normalized to maximum)

4 -14 mm /yrlandslide mapping

*last 35 years0.0073 -8.1 mm/yrsediment gauges

*last 10 years(Blodgett, 1998)

o A) 0.093 mm/yrB) 0.014 mm/yr

basin averages fromsediment flux data

*last 10 years?(Summerfield and

Hutton, 1994)0.2 -0.6 mm /yr

AFTT*last 6 -20 Myr

(Safran, 1998)

o EC) 0.15 mm /yrAFTT *last 34 Myr

IA) 0.31 mm/yrAFTT *last 16 Myr(Ege et al., 2001)

X) 0.038 -0.076 mm/yrY) 0.039 -0.077 mm/yrZ) 0.044 -0.089 mm /yr

erosion surface*last 10 Myr

(Gubbels, 1993)0.33 mm/yr

basin fill*last 5 Myr

(Flemings & Jordan, 1989)

0.13 -0.27 mm /yrTertiary strata*last 2 -3 Myr

( Gubbels, 1993)

r

- 140 i.Peru

- 16° \--/ [6La

) PazCochabambai

18 r BoliviaI I

\4.

EC

-2001

Chile

/- 22°

L 70°

N &Nb

Potosi

Y

X

Tupiza

1 r\ L

\/ Argentina \-\,,66° 640 62°

I I l

Nb) 0.43-0.84 mm/yrSb) 0.06-0.22 mm/yr

cross-sectionsSubandean zone only

*last 15-20 Myr(McQuarrie, 2001)

N) 0.15 -0.27 mm /yrS) 0.08 -0.16 mm/yr

cross -sections*last 40 Myr

(McQuarrie, 2001)

Figure 5

i1

I) BENI

- 14° li AB, PLAIN\

Peru /' AQ+ -+-AIk + SR

- 160 ,\,- SI UNVBi ''' TMj La TQ//

rrL.

1- 8° \ ri Bolivia

1 \1

_20°I

Cochabamba

2201I

/

¿7;Oo

1<I

Chile

L\

200 km

PEt-LBA BE4kELS

PA ANTAHU + P'ÑMI _LAP

AT+Potosi NU VQ

SLE+PZ +

SA

EP±c-r. +VII Tupia CM PJ` OB ER

- I

¡ ( IArgentina 1 ,o `'^ 66° 64° 62° \-`.68°1 i

I I I

CHACOPLAIN

Figure 6

o5000 granitoids N

4500

4000

3500-.-3000-

2500-

2000-

i 500 -

1000-

500

0

1

.9 -

.8-

.7 -

ci).6-

aÓ .5_co

.4 -

.3 -

.2 -

o

.250 -

.225 -

.200 -

.175 -

.075 -

.050 -

.025 -

0

0

5000

4500

4000 -

3500 -

3000 -cO 2500-cd

Ñ 2000W

1500

1000

500

olo

Rio Coroico

O

Al Stream powerl : E = KAmSn

A

.50

- .45

- .40

- .35

- .30

25

Rio Coroico-longitudinal profile data 1

LP __I DA

7

50 75 100

Distance downstream (km)

V

- -'---/

fYI-

I

SLP

N

5000 -volcanics

4500

4000 -'

3500 -Ec 3000 -'of/ 2500 ,mW 2000 -

1500 -

1000

500 -

0tAr-1'.J-`P

i

25

S, O, minor J

5Jd

CW...,/-- - - --

P

50 75 100

Distance downstream (km)

125

-,

B

.25

.20

.15

.10

.05

o150

125 150

800 - 700

700 Ñ - 600:v

600 á - 500

500 - 400 *.á400 ' - 300

w 3300 ctoD - 200"

>200 - 100 o

c-.100 Z -0 T

O -a0 ñ

NL a... 300

33

"5--

O - .50D

,tightly folded, faulted

debris -chokedtributary zone

Stream powerE = KAmSn

P, E, OI, To

Subandean fold- thrustInterandean beltZone

7/1=r

mS -0 p cÉá 2

Ir)1

Ts`EE.

1kpt W -

SPkptW-

A

Rio Grande

5000

4500

4000

-. 3500 -

E3000 -c

-1.-ii.- 2500 -

ILI2000 -

1500 -

1000 -

500 -

o0

- .45

- .40

- .35

- .30

- .25

- .20

-.15

-.10

- .05

100 200 300 400

Rio Grande

500

Distance downstream (km)

,

-longitudinal profile data 1II

II I iI PLP 1 iiI

j ' DA l --(` ;

/11 `V I, II ; i/ ' liI I 2t/ ICWI /-,r----Ì- I¡ I i/d v

600

B

r-%I.

I I I I.II I

I , `I

....-Jr ,..-----

SLP

r1 i r--- \ ,%, \/

P

100 200 300 400 500

Distance downstream (km)

Figure 7

600

0700

- 800 - 700 n-700-p -600

=- 600 - 500 m

- 500 v - 400 á- 400 w. - 300 ^

5 3...- 300 crQ - 200

CD

- 200 a-100

-0 00

-a3 200N

3700 `C\"..

Ec 3000-O

>2500 -

a)l.L.l 2000-

5000

4500-

4000 -

3500 -

1500 -

1000 -

500 -

0

IgnimbritesO,S,Ktightlyfolded

O

tightlyfolded

K Rio PilcomayoO

DJ, K, T. .

InterandeanZone

Stream powerE = KAmSn

kpt W -SPIpt

Subandeanfold- thrust belt

A

Q

Chacoplain

- .50

- .45

- .40

- .35

- .30

- .25

- .20

- .15

- .10

- .05

0 100 200 300 400 500

Distance downstream (km)

.250 - 5000-Rio Pilcomayo

.225 - 4500-

200 - 4000--longitudinal profile data

.175 - 3500- LPE

.150 - c 3000a)8-.125-25O0- IN

CO /t /(n.100 - 2000 - j ; I ¡ 1 I,

111 /A, ,lJ k.075 -

.050 -

.025 -

1500 -

1000 -

500

o- oo

-"'_

-i

600 700o

800

"'' r'.- - soo 700 n.. - I - 700 -v - 600

B .I

j - 600 w - 500 NI / 3

I I 1`I 1 t1 DA

- 500 0- 400 R.:

i1`j ; LI 1

I - 400 . - 300

/ Nil 1 I i / Du

/ 1 1 / - 300 - 200

Li----------- 'cw , -200 m -100 p\--_- -ng)-100 X -o

- 0 oo 300

A--.

3 200 3N3

SLP

-.P

D

5000 r"

4500 -

4000 .1

3500 -

E'--' 3000

-t s'

cÌ:. 2000

.250 -

.225 -

.200

.175 -

.075 -

.050 -

.025 -

o

1500 -

1000 -

500 -

oo

3500 -

3000 -

2500

2000 -

1500 -c

1000cd

W 500

o

100 200 300 400 500 600Distance downstream (km)

mostly D, some K, Tr, C

Rio Parapeti

InterandeanZone

mostly P, some K, Tr, C

Subandean fold -thrust belt

óC ÿÉ

2

IptW

A

700 800

K,P Tr,C P

1-1-' Chaco

r--, plain

.c

fd

.50

- .45

- .40

- .35

- .30

- .25

`o_ - .20

-- - .15wmó

1 -.10

Stream power -.05

- SP IV W E = KAmSno

50 100 150 200 250 300 350 400

Distance downstream (km)

- 800

Rio Parapeti-700

-longitudinal profile data-

- 600

DA/ B - 500LP ,- iß'p- % - 400

- 300

SLP /=; ____ -200------ cw -1 00----- .------

P -o

50 100 150 200 250 300Distance downstream (km)

Figure 7

350 400

- 700

-600

- 500 cD

- 400 ás- 300

3- 200

100 p

0 m

[300g200 313.

3

3

5000

4500 -

4000 -

E3500 -

c 3000 -o

2500 ->Lll 2000 -

1500 -

1000 -

500 -

0

.250 - 5000

.225 - 4500

.200 - 4000

.175 - 3500.-.

a).150 -

E3000

Ó .125 - o 2500

.100 - ai 2000W

.075 - 1500

.050 - 1000

.025 - 500

o-

.250 -

.225-

.200 -

.175 -

a).150 -

Ó .125 -(!)

.100

.075 -

.050 -

.025 -

0

0

minor Ko

folded, faulted

Rio Pilayao

AD, minor S

InterandeanZone

K, C,

'P,=T

Sub-andeanfold-thrust belt

Stream power

E = KAmSn

0 100 200 300 400

Distance downstream (km)

Rio Pilava

500

.50

.45

.40

.35

- .30

- .25

- .20

- .15

- .10

Plt w - 05

0600

-800 -700 o- soo

-longitudinal profile data B -700

- 600 - 500fv

- 500 0 - 400 áDA 5

- - ---- ---r .__.- 400 ?. - 300

/ I^ h po `.I

i- 300 - 200

j CW_/ -200 D -100 0/ (D c_-100 -0 -n2 -11

LP

o

P

-0 0 300[

N3

100 200 300 400 500 600 `5Distance downstream (km) v

5000 -o

Tcgltightly folded ^ o

4500 - I Ktightly folded,'

4000 -faulted

3500 -

TK

east flank of Camargosyncline

3000 -C

2500 -ct3

w 20 ° °- Rio San Juan del Oro1500 -

1000 -Stream power

soo - E KAmsn

0

.50

.45

.40

- .35

- .30

- .25

- .20

- .15

- .10

- .05

,

50 100 150 200 250 300 350Distance downstream (km)

5000 - -1000 - 700

4500 - - 900 - 600 C

4000 - Ì t -800 _500 w m

35°°- LP\CW

Rio San Juan del Oro -700 ? -400^ oE3000 - \ -600 -300 x

I m 32500 -

I- 500 g -200 v m

a>i

I DA -- - . _ -- -- áW 2000- I r t------- -400 ^ -100 0..-- -"f1500 -- ---/- s-i i. r`_ -300 -0 -n

r 1 j 1 / -, I ~1000 - l ----- -- -...._ 200 r 300 ñ

/ t I `--'I I

500 -/ IP

-100 L 200SLP 3

0 - ~ - -- .0.

0 50 100 150 200 250 300 350 -51

Distance downstream (km)

Figure 7

0

5000 -I

4500

4000 -

3500 -

E 3000 -C

2500 -

Ñ 2000 -w

1500 -

1000 -

500 -

o

I

I

1

1

1

1

\ Rio Coroico

y = 4566.8x-0.381

I/ R2

0 25 50 75 100

Distance downstream (km)

Figure 8

5000 -

4500 -

4000 -

. 3500 -

c 3000 -ó

2500 -

aiw 2000 -

1500 -

1000 -

500 -

0

Rio Pilcomayo, ,. y = -6.3x + 4244.. ..

,.., ,, .

125 150 0 100 200 300 400 500 600Distance downstream (km)

i700 800

1

t

noincision

knickpoint

Fault 1

Distance

Bsmall river

incision

Knickpoint width

Fault 1

5-

D4.5-

Distance

y=0.34x+2.9 00-R2=0.90 -

,o

0 y=0.35x+2.53.5- R2=0.54

3-o m=0.35+/-0.15

2.5-

22 3 4 5

Log of Drainage Area (km2)

Figure 9

1

Clarge river

Knickpoint width

incision

1

Fault 1

Distance

102-

101 -

10°-

10 -1-

10-2-

10 -310° 101

Figure 10

1

,

102 103 164 165 106

Time scale (yr)

,

107

102

- 101

-10°

- 10-1

-10-2

10 -3108

wetNORTH o

glaciatedBasins: high relief, high drainage denisity,

longer low -order streams with greaterelevation drops, high D rates

Rivers: high slopes and widespreadincision (gray), low sinuousity,narrow channel widths, high DJF

precip. along profile

Sq

Kinematics: steep decollement, more taper, back -thrusting,less basement/more low -grade (weak) Paleozoic rocks =low internal wedge strength, closely spaced anticline ridges, largepiggyback basins, subcritical condition (internal deformation)

Figure 11

SOUTHi- semi -arid Basins: low relief, low drainage density, small

basins graded to benches above the majorknickpoints, low D rates

Rivers: low slopes and morelocalized high incision (gray) atmegathrust controlled steps,high sinuosity, wide channelwidths, low DJF precip.

along profile

A

Kinematics: shallow decollement, less taper, in- sequence thrusting,more basement/more high -grade (strong) Paleozoic rocks = high internalwedge strength, widely spaced anticlinal ridges, small piggyback basins,critical condition (wedge advance)