Climatic control on eastern Andean denudation rates (Central Cordillera from Ecuador to Bolivia)

Climatic control on Eastern Andean denudation rates (Central Cordillera from Ecuador to Bolivia) E. Pepin* JL. Guyot E. Armijos H. Bazan P. Fraisy J.S. Moquet L. Noriega J. Ordonez R. Pombosa P. Vauchel GET-UNIVERSITE DE TOULOUSE-CNRS-IRD-OMP, 14 Av. E. Belin, F-31400 Toulouse France *Corresponding author. E-mail: [email protected]

1

INTRODUCTION

Specific suspended sediment yield can be defined as the total mass of sediment transported as suspended load by the river at the outlet of its drainage basin, per unit of catchment area and per unit of time (expressed in units of t.km−2 .year−1 ). The ratio between sediment yield and sediment density is providing the catchment mean denudation rate (mm.yr−1 ). Sediment yield (and thus denudation rate) is rather representative to mechanical erosion in contrast with solute fluxes linked to the chemical erosion. For this reason, many studies have focused on sediment yield fluctuations in order to figure out the controlling factors of landscape mechanical erosion. Correlations found between denudation rate and topographic, lithological, climatic, ecological or tectonic factors from around the world have been divergent especially according to the area and the time scale of study (e.g. Pinet and Souriau, 1988; Louvat and Allegre, 1997; Hovius et al., 1998; Hicks et al., 1996; Verstraeten and Poesen, 2001; Reiners et al., 2003; Dadson et al., 2004; Vanacker et al., 2007b). For instance, Safran et al. (2005) highlighted correlation between milennial-scale erosion rate and channel steepness (i.e. channel gradient corrected for drainage area) contigous catchments from Bolivia whereas Aalto et al. (2006) found strong relationships between current sediment yield, slope and lithology in embedded Bolivian catchments. Moreover, Dadson et al. (2003) proved the role of sismicity and the variability of climate on current denudation rate for a set of various little catchments from Taiwan and Vanacker et al. (2007b) demonstrated how vegetation cover is controlling current denudation rates of a single catchment of Ecuador. Therefore, understanding what factors are controlling denudation rate is still an open debate that is why it appears crutial to study more kinds of sedimentary systems .

Preprint submitted to Elsevier

14 December 2011

For these reason, we focus for the first time on eight large catchments extending along the Andean range and covering an area of ∼ 674 000 km2 from Ecuador to the south of Bolivia. These catchments are large enough to provide a first order tends of suspended sediment yields and runoffs representative of the Andes Cordillera range. The repartition of the current mean denudation rate has been thus determined along the Eastern side of the Cordillera as well as the mean current runoff and its seasonal variability. Several topographic parameters as well as lithologic and vegetation cover indexes have also been computed. This database allows to specify the main controlling factors of the current denudation rate at large spacial scale. Our results assess the role of climate-driven erosion and especially they provide evidence for the predominant control of the seasonal variability of rainfall.

2

STUDIED AREA AND METHODS

8 mountainous catchments located on the eastern side of the Andes, from Ecuador to the south of Bolivia; between latitudes 00,00’ S and 22,55’ S and longitudes 79,37’ W and 63,11’ W (figure 1, table 1) have been studied. Their outlets correspond to hydrological stations located at the fringes of the mountain range, so that 97% of the studied area are located above an altitude of 400 m that we arbitrarly defined as the Andes boundary. This mountainous nature prevents significant sediment sinks and attests the catchments to be purely denudational basins. Therefore, our catchment set is suitable for a mesoscale erosional processes study. Moreover, the erosive behaviour of the studied area is considerable. For instance, the seven catchments of our set included in the Amazonian basin, are known to account for only 11% of its total surface area but to product the quasi totality of the Amazonian suspended sediment (Meade et al., 1985; Guyot, 1993; Goulding, 2003; Filizola, 2003).

2.1

Suspended sediment

Our study is based on mean denudation rates values computed from the suspended sediment time series on each catchment of the area of study. The entire water and suspended sediment discharges database is coming from the HYBAM program that is working with several governmental and scientific organizations to provide water and sediment discharge data throughout the Amazonian basin (http://www.ore-hybam.org/). HYBAM is using a specific procedure for water discharge and sediment concentration sampling (Callede et al., 2000; Filizola and Guyot, 2004; Guyot et al., 2005; Cochonneau et al., 2006) that involved daily gauge height and rating curves, regular water discharge measurement, daily surface samples and depth integrated sediment 2

concentration samples. Calibrated water and sediment discharges data are thus coming from Armijos et al. (????b,a) (FRS, BOR, CHA station), Guyot et al. (2011); ? and Malbrunot (2006) (ATA, RUR station), (Guyot et al., 1996) (ABA and SAN stations). In Ecuador, the FRS station is a fictive station created by adding together the hydrological contributions from Rio Napo in Francisco de Orellana and Rio Coca in San Sebastian (Laraque et al., 2004). It allows to obtain a catchment comparable in size to the other ones. In Peru, rain-discharge simulations has been computed to reconstruct unavailable water discharges time series of Rio Ucayali in Atalaya (ATA) Lavado et al. (submitted); ?. The length of the suspended sediment time series (from 2 to 43 yr) and the number of suspended sediment measurements vary according to each station’s history, which going from North to the South include: FRS (408 samples, 2000-2010), BOR (335 samples, 2004-2010), CHA (571 samples, 2004-2010), ATA (74 samples, 2004-2006), RUR (970 samples, 1969 2010), ABA (851 samples, 1976-1990), SAA (642 samples, 1973-1983), VIL (1334 samples, 1977-2005). Because of its shorter period of measurement, the data obtained at the Atalaya station are only considered to be estimates. Discharge data and suspended sediment concentration have been combined as daily totals or monthly totals to produce average annual sediment flux (Kt.yr−1 ) (see table 2). Mean annual yields (t.kg −1 .km−2 ) are obtained by dividing the suspended sediment flux by the catchment area. The modern denudation rates Er are computed by dividing the mean annual yields by the dry sediment density (2500 kg.m − 3). Because bed load is difficult to estimate and poorly documented, denudation rates are estimated only with the suspended sediments contribution. The solute fluxes have also been neglected. However, in the Andean catchments of Bolivia, suspended sediment load seems to account for ∼ 85 % of the total sediment fluxes (Guyot, 1993; Malbrunot, 2006). This ratio is fluctuating according to the catchment aridity and size (90% for large and humide basins, 75% for aride or small basins) (Collins and T.Dunne, 1989; Aalto et al., 2006). Besides of these limitations, our denudation rate estimation is submitted to uncertainty because the suspended sediment load could represent only a fraction of the sediments exported from a basin and can miss large, episodic events within the measuring period. However, catchment size and suspended sediment sampling frequency (∼ 10 days) ensure that the sampling time step is closed to the catchment response time to drastic climatic events.

2.2

Climate

In the Amazonian catchment, annual pluviometric maximums are observed in the Andean Piedmont, whereas the annual minimum was recorded in some semi-arid valleys of the Andes (Roche et al., 1990; Ronchail et al., 2005; Laraque et al., 2007). There is a variable relationship between the rain, the 3

altitude and the orientation of the basins (Johnson, 1976; Pulwarty et al., 1998; Ronchail et al., 2005; Buytaert et al., 2006; Espinoza et al., accepted). The hydrological cycles are strongly conditioned by the precipitation regime. Characteristic hydrographs of mountainous areas such as the Andes are defined by many peaks that merge downstream of the mountain range to create large discharges of water in the tropical zones. The annual cycle of water discharges is thus increasingly defined as moving from upstream to downstream in the catchment area (Roche and Fernandez, 1988). Both water discharge and catchment-average rainfall are generally used to describe the climate. Therefore, interannual catchment-average rainfall has been determined for each catchment as well as the rainfall seasonal variability that details each catchment precipitation regime and supplement our climate description. An exceptional rainfall database sampled and calibrated as part of the HYBAM program has been used (Espinoza et al., accepted). This rainfall database extends from 32 to 41 yr and includes more than 360 rainfall stations in our study area. For each catchment, a mean runoff (catchmentaverage rainfall) has been calculated from monthly-average rainfall records of the stations distributed throughout the catchments using a Kriging method. An index of rainfall seasonal variability subsequently referred to as CvP , and defined by the standard deviation of the monthly runoff, normalized by the inter-annual average runoff has also been computed. A strong correlation exists between the runoffs computed from the rainfall database and the average water discharges (figure 2A: R2 =0.84) as well as between the variability index of precipitation and discharges (figure 2B: R2 =0.99). These results suggest that the same monthly tendencies are recorded by the precipitation and discharges measured. They validate the databases since the two types of data are acquired independently and with different techniques, and do not produce the same type of errors. In the following, only the rainfall data (average runoff and variability) is compared to the denudation rate, as (1) both values are available for all the catchments, (2) Rainfall data are not involved in the sediment flux calculation. (3) Rainfall time series are longer than water discharge ones.

2.3

Topographic and lithologic parameters

Various parameters, presented in the following paragraphs, allow to detail the topographic characteristics of the studied area (see table 1). However, the study of such parameters with regard to erosive behavior is based on the tectonic context knowledge. Indeed, tectonics is an important external forcing for erosion processes as well as for topographic features. At our study spacial scale, previous contributions are referring to two main period of uplift. A first uplift phase of the Andes seems to have taken place at the end of Oligocene and the beginning of Miocene (Cooper et al., 1995; Gregory-Wodzicki, 2000; 4

Gillis et al., 2006; Barnes and Pelletier, 2006; Ege et al., 2007). A second major uplift phase is recorded around 8-10 M yrin both sides of the Central Andes (Thouret et al., 2007; Garzione et al., 2006; Schildgen et al., 2007) as well as sourthward of the study area, between southern and central Chile (Riquelme et al., 2007; Far´ıas et al., 2008). Thermochronologic data from the basins in the north of Bolivia also suggests two periods of uplift at around 40-25 M yr and 15 M yr in Bolivia, presenting an acceleration of exhumation after 15 M yr (Barnes and Pelletier, 2006; Barnes et al., 2006). Periods of exhumation inferior to 10 M yr have been recorded with significant variations in the average exhumation rates from one valley to another Safran et al. (2006) and along the sub-Andean rivers (Barnes et al., 2006). Finally, the beginning of the sub-Andean deformation front seems synchronous from Bolivia to Colombia towards 10 M yr (Roddaz et al., 2005). Therefore, any notable uplift gradient has been described from North to South at the spatial scale of study. Because of the mesoscale setting of the catchments studied, the SRTM topographic database for South America appears accurate enough to perform the basin-scale morphometric analyses as its absolute vertical uncertainty is estimated to be near 6 m (Farr et al., 2007) and its spatial resolution is around 90 m in the area of study. Various topographic parameters known as potentially important for erosion control have been computed for each catchment using semi-automatic procedures from ARGIS software. Slope and drained surface (directly related to the transport laws) are very commonly used (Whipple and Tucker, 1999). Aalto et al. (2006) showed a strong correlation between these parameters and the erosion rate in a set of 47 Bolivian catchments. Maximal and average altitudes have been defined as in the Aalto et al. (2006) study. Moreover, this parameters are directly linked to the catchment hyspometric curve and are thus indirect markers of the catchment evolution states. However, the presence of isolated volcanoes in some catchments increases artificially their maximal altitude values. For this reason, we chose to not study the maximal altitude effect in the following correlation study. To enhance the topographic database, a lithologic index has been defined. The compilation of 1/1000000 lithologic maps of Moquet et al. (2011) has been used to document the main rock type distribution on each catchment except for the VIL catchment where the SOTERLAC lithologic database (ref?) replaces the laking data. Catchment lithologic index (PLI) has been computed combining the rock type distribution with the Probst (1990) contribution that classify rock type according to their susceptibility to mechanical erosion as previously suggested by Guyot (1993); Aalto et al. (2006). These autors have defined the lithologic index to be one of the most powerful control for sediment yield evolution in Bolivia. Table 3 summarizes the PLI assigned for each rock type defined by Moquet et al. (2011)’s lithologic compilation as well as the final lithologic index assigned to each catchment. 5

Finally, the fraction of soil covered by vegetation is determined using the Fcover index (c.f. http://postel.medias-france.org). The vegetation effect on erosion has previously been proved by Vanacker et al. (2007a). This index is delivred with a grid resolution of 1 km. Data from summer XX were used to minimized the snow cover.

3 3.1

RESULTS AND DISCUSSION Denudation rate

The distribution of the denudation rates in the Atlantic hillslopes of the Andes is not homogeneous (figure 3 and table 2). It is characterized by a North-South gradient, with low values in the north (0.22-0.5 mm.yr−1 ) between Ecuador and Peru, and with higher values in Bolivia (0.8-1.2 mm.yr−1 ). The average denudation rate for the area of our study is around 0.6 mm.yr−1 .

3.2

Climatic controls on modern denudation rate

Compiled precipitation data shows a strong N-S gradient for the average annual runoff and for the seasonal regimes on the scale of the catchments (figure 3). The northern Andes, exposed to wet winds, is characterized by abundant rains of up to 2000 mm.yr−1 (3515 mm.yr−1 in the Ecuadorian basin). In Ecuador and northern Peru, the rainfall regime is equatorial with a weak seasonal variation. Some pluviometric stations located in the same catchment show inverse seasonal regimes (Espinoza et al., 2006; Laraque et al., 2007), which explains the weak seasonal variability on the scale of the catchments (figure 3). The south (Bolivia, southern Peru) presents lower levels of rainfall (from 750 to 1310 mm.yr−1 for the 4 catchments, see table 2 for the exact values). This entire area is characterized by the same tropical-austral regime, with a well-marked seasonal variability. In the four Bolivian catchments, the rainy season represents between 67% and 87% of the average annual rainfall. The central zone in Peru is made up of various intermediary regimes between the equatorial regime and the tropical regime.

In addition, the table 4 shows a Pearson’s correlation matrix that allows us to evaluate possible correlations between all the different variables (i.e. Restrepo et al., 2006; Aalto et al., 2006; Pepin et al., 2010). The last line of the matrix shows the correlations between the different variables with the current denudation rate. It appears that the seasonal variability of precipitation is the principal factor (correlation of 0.73) that explains the denudation rate, followed by 6

a negative correlation with the average annual rainfall. The denudation rate is also strongly correlated with the vegetation cover parameter (correlation of 0.73). However, the vegetation cover appears to be significantly dependent of both climate parameters as high correlations are recorded between these variables (correlation of -0.97 with CVp and 0.83 with the Runoff). The other parameters are playing a less significant role in denudation rate control. The high correlation between denudation rate and CVp shows the importance of taking climatic variability into account for both erosional study and modelization of climatic effect. The effect of climate variability on sediment yield have already be highlighted by Dadson et al. (2004). These authors focused on current denudation rates of Taiwan catchments shown that after a first large earthquake responsible of numerous landslides, (1) subsequent typhoons have increased the landsliding and (2) large amount of sediment have been transported within the chanel network during later storms. Thus, the high variability of rainfall is increasing both the mechanical erosion and the sediment transport within the catchment. Moreover, because no strong tectonic gradient is documented at the scale of our study area, the climate variability has certainly another effect than to amply the erosional power of drastic tectonic event. We suggest that at the Andean range scale, the high seasonnal climate variability combined to a significant aridity on Bolivian catchments is not farouring perenial and constant vegetation cover contrary to the steady -humide climate of northern catchments. Therefore the northern catchments appear protected by constant luxuriant vegetation to mechanical erosion contrary to the southern catchments where the climate is responsible of less vegetation cover, vegetation seasonnal variation and more extreme climatic events. The shielding effect of high vegetation cover on denudation rates has already been suggested by Pepin et al. (2010) that focused on Chilean catchments. Finally, the significant negative correlation between Er and the mean runoff contradicts the intuitive idea of major erosion in the areas with the most abundant rainfall (see also von Blanckenburg et al. (2004)).

3.3

Topographic parameters

Aalto et al. (2006) showed a strong correlation between topographic parameters like Slope or average altitude and the sediment Yield (i.e. the denudation rate) in a set of 47 Bolivian catchments. In our study, the weak impact of the average slope of catchments (table 4) is easily explained by the similarity of these values, in that our catchments have very similar relieves. To the contrary, Aalto et al. (2006) studied embedded catchments, which are more representative of along-stream variations in erosion rates. The high erosion rates and slopes of the sub-catchments located upstream can be explained 7

by the progression of incision towards the interior of the mountain (Carretier et al., 2006; Vassallo et al., 2007; Vanacker et al., 2007b) and by lithological differences (Aalto et al., 2006). Rainfall does not have any correlation to the denudation rates studied by Aalto et al. (2006). Thus, according to these authors, denudation rates in the sub-catchments are mainly controlled by the response to uplift. Over longer time scales, Safran et al. (2005) reached the same conclusion. Safran et al. (2005) calculated average Holocene denudation rates for the Bolivian basins from the concentration of cosmonucleides in the sands of the rivers. They showed that there are very weak correlations between these denudation rates and the climate and lithology, concluding that the denudation rates are controlled by transient adjustments of the catchments to the Altiplano’s uplift. Our results for the large Bolivian basins bordering the range confirm this interpretation.

3.4

Climatic-Topographic coupled model

To estimate denudation rate in other Andean catchment, we determined simple models that calculate denudation rate based on several of our parameters. We carried out a statistical multiple regression study to provide both the best linear and no-linear regressions using Xlstat software (http://www.xlstat.com/fr/). Several forms of no-linear regression were studied. The one presented on fig. 4 has been chosen because of its simplicity, its efficiency and because its feature is similar to those of Aalto et al. (2006). The entire set of variables used in table 4 were studied and we limited at three the number of variables used to model the denudation rate. The software computed all the possible combination between variables according to the regression form asked. Regressions presenting the best correlations are shown in fig. 4. The denudation rate is efficiently estimated using an single variable equation with CVp . However, the most reliable models to estimate Er are combining Cvp, PLI and the Slope parameters. Thus, taking into account both secondary parameters are improving our results. It demonstrates that even if their correlation with the denudation rate are not significant, parameters linked to topography (Slope) and lithology (PLI) are playing an indirect control on erosion and should be taken into account.

4

CONCLUSION

This study presents for the first time, the suspended sediment yield and associated current denudation rates of the Eastern side of the central Andes. Denudation rates have been computed over eight large catchments and averaged over time spans of 2-43 years. Denudation rates range between 0.25 8

and 1.20 mm.yr−1 and a North to South gradient is detectable. Analysis of correlations between denudation rates and climate, topographic and lithologic parameters suggests that denudation rate is mainly controlled by climate and especially its variability. The slope and lithology determined to be important parameters for sediment yield control of Bolivian embedded catchments (Aalto et al., 2006) are only secondary parameters of denudation control along the Andes range. At large spatial scale, erosion is favourized in catchments submitted to arid and variable climate and disadvantaged in catchment submitted to high-rainy and constant climate. This non-intuitive result suggests that the climate plays a complexe role on erosion control, notably because it is influencing other variables like the vegetation cover. Thus, the important effect of climate variability on erosion and sediment transport suggested previously by Dadson et al. (2004) appears in our studied field case enhance by the potential shielding effect of the vegetation cover directly controlled by the climate regime. Finally, this study shows it is primordial to take into account the variability of the climate in order to deal with denudation rate control. Acknowledgements : We kindly thank S. Carretier for his suggestions that improved the first version of this manuscript.

9

References Aalto, R., Dunne, T., Guyot, J., 2006. Geomorphic Controls on Andean Denudation Rates. The Journal of Geology 114, 85–99. Armijos, E., LARAQUE, A., BARBA3, S., BOURREL, L., CERON, C., LAGANE, C., MAGAT, P., MOQUET, J., POMBOSA, R., SONDAG, F., VERA, A., GUYOT, J., ????a. . JSAES. Armijos, E., LARAQUE, A., BARBA3, S., BOURREL, L., CERON, C., LAGANE, C., MAGAT, P., MOQUET, J., POMBOSA, R., SONDAG, F., VERA, A., GUYOT, J., ????b. Yields of suspended sediment and dissolved solids from the Andean basins of Ecuador. Hydrological Sciences Journal. Barnes, J., Ehlers, T., N.McQuarrie, OSullivan, P., Pelletier, J., 2006. Eocene to recent variations in erosion across the central Andean fold-thrust belt, nothern Bolivia: Implications for plateau evolution. Earth Planet. Sci. Lett. 248, 118–133. Barnes, J., Pelletier, J., 2006. Latitudinal variation of denudation in the evolutino of the Bolivian Andes. Am. J. of Science 506, 1–31. Buytaert, W., Celleri, R., Willems, P., Bivre, B. D., Wyseure, G., 2006. Spatial and temporal rainfall variability in mountain areas: A case study from the south Ecuadorian Andes. Journal of Hydrology ?, ? Callede, J., Kosuth, P., Guyot, J.-L., Guimaraes, V., 2000. Discharge determination by acoustic doppler current profilers (ADCP) : a moving bottom error correction method and its application on the river Amazon at Obidos. Hydrological Sciences Journal 45(6), 911–924. Carretier, S., Niviere, B., Giamboni, M., , Winter, T., 2006. Do river profiles record along-stream variations of low uplift rate? J. Geophys. Res. 111, F02024. Cochonneau, G., Sondag, F., Guyot, J., Geraldo, B., Filizola, N., Fraizy, P., Laraque, A., Magat, P., Martinez, J., Noriega, L., Oliveira, E., Ordonez, J., Pombosa, R., Seyler, F., Sidgwick, J., Vauchel, P., 2006. L’Observatoire de Recherche en Environnement, ORE HYBAM sur les grands fleuves amazoniens. In: Climate Variability and Change-Hydrological impacts. Vol. 308. pp. A232–A232. Collins, B., T.Dunne, 1989. Gravel transport, gravel harvesting and channelbed degradation in rivers draining the southern Olumpic Mountains, Washington, USA. Environ. Geol. Water. Sci. 13, 213–224. Cooper, M., Addison, F., Alvarez, R., Coral, M., Graham, R., Hayward, A., Howe, S., Martinez, J., Naar, J., Penas, R., Pulham, A., Taborda, A., 1995. Basin development and tectonic history of the Llanos Basin, Eastern Cordillera, and Middle Magdalena Valley, Colombia. AAPG Bulletin 79(10), 1421–1443. Dadson, S. J., Hovius, N., Chen, H., Dade, W., Hsieh, M.-L., Willet, S., Hu, J.C., Horn, M.-J., Chen, M.-C., Stark, C., Lague, D., Lin, J.-C., 2003. Links between erosion, runoff, variability and seismicity in the Taiwan orogen. Nature 426, 648–651. 10

Dadson, S. J., Hovius, N., Chen, H., Dade, W., Lin, J.-C., Hsu, M.-L., Lin, C.-W., Horn, M.-J., Chen, T.-C., Milliman, J., Stark, C., 2004. Earthquaketrigered increase in sediment delivery from an active mountain belt. Geology 32, 733–736. Ege, H., Sobel, E., Scheuber, E., Jacobshagen, V., 2007. Exhumation history of the southern Altiplano plateau (southern Bolivia) constrained by apatite fission track thermochronology. Tectonics 26, TC1004. Espinoza, J., Fraizy, P., Guyot, J., Ordonez, J., Pombosa, R., Ronchail, J., 2006. La variabilit´e des d´ebits du Rio Amazonas au P´erou. In: Climate Variability and Change-Hydrological impacts. IAHS, pp. (308)424–429. Espinoza, J., Ronchail, J., Guyot, J., Cochonneau, G., Filizola, N., Lavado, W., Noriega, L., Deolivera, E., Pombosa, R., Romero, H., Vauchel, P., accepted. Spatio-temporal rainfall in the amazon basin countries (Brazil, Peru, Boliva, Colombia and Ecuador). Int. J. of Climatology. Far´ıas, M., Charrier, Carretier, S., Martinod, J., Fock, A., Campbell, D., Caceres, J., Comte, D., 2008. Late Miocene high and rapid surface uplift and its erosional response in the Andes of Central Chile (33-35S). Tectonics 27, TC1005. Farr, T., Rosen, P., Caro, E., Crippen, R., Duren, R., Hensley, S., Kobrick, M., Paller, M., Rodriguez, E., Roth, L., Seal, D., Shaffer, S., Shimada, J., Umland, J., Werner, M., Oskin, M., Burbank, D., Alsdorf, D., 2007. The Shuttle Radar Topography Mission. Geophysics 45, RG2004. Filizola, N., 2003. Transfert s´edimentaire actuel par les fleuves amazoniens. Ph.D. thesis, Universit P. Sabatier, Toulouse, France. Filizola, N., Guyot, J.-L., 2004. The use of doppler technology for suspended sediment discharge determinations in the river amazon. Hydrological Sciences Journal 49(1), 143–153. Garzione, C., Molnar, P., Libarkin, J., Macfadden, B., 2006. Rapid late Miocene rise of the Bolivian Altiplano: Evidence for removal of mantle lithosphere. Earth Planet. Sci. Lett. 241, 543–556. Gillis, R., Horton, B., Grove, M., 2006. Thermochronology, geochronology, and upper crustal structure of the Cordillera Real: Implications for Cenozoic exhumation of the central Andean plateau. Tectonics 25, TC6007. Goulding, D., 2003. The Smithsonian Atlas of the Amazon. Smithsonian BooksWashington and London, USA and UK . Gregory-Wodzicki, K., 2000. Uplift history of the central and northern andes: A review. Geol. Soc. Am. Bull. 112, 10911105. Guyot, J., Filizola, N., Quintanilla, J., Cortez, J., 1996. Dissolved solids and suspended sediment yields in the Rio Madeira basin, from the Bolivian Andes to the Amazon. In: D.E. Walling and B.W. Webb (Ed.), Erosion and Sediment yield : Global and Regional Perspectives. Vol. 236. pp. 55–63. Guyot, J.-L., 1993. Hydrog´eochimie des fleuves de l’Amazonie bolivienne. Ph.D. thesis, Collection Etudes et Th`eses, ORSTOM, Paris. Guyot, J. L., Cardenas, E. A., Bonnet, M., Bourrel, L., Calmant, S., Cochonneau, G., Crave, A., Oliveira, E. D., Villar, J. E., Villar, R. E., Fraizy, 11

N. F. P., Labat, D., Lagane, C., Laraque, A., Casimiro, W. L., Martinez, J., Moquet, J., Moreira-Turcq, P., Flores, L. N., Loza, R. P., Ronchail, J., Seyler, F., Seyler, P., Sondag, F., Vauchel, P., 2011. Hydrology and sediment transport in the Amazon basin, from the Andes to the Ocean. Status and future of the world’s large rivers. Guyot, J. L., Filizola, N., Laraque, A., 2005. R´egime et bilan du flux s´edimentaire de l‘Amazone ´a Obidos (Br´esil), de 1995 a´ 2003. In: D.E. Walling and A.J. Horowitz (Ed.), Sediment budgets, IAHS Symposium, Foz de Iguazu, April 2005. Vol. 291. pp. 347–356. Hicks, D. M., Hill, J., Shankar, U., 1996. Variation of suspended sediment yields around new zealand: the relative importance of rainfall and geology. in erosion and sediment yield: global and regional perspectives. Int. Assoc. Hydrol. Sci. 236, 149–156. Hovius, N., Stark, C. P., Tutton, M. A., Abbott, L. D., 1998. Controls on sediment supply by large rivers. In Shanley, K. W., and McCabe, P. J., eds. In Relative role of eustasy, climate and tectonics in continental rocks. Soc. Econ. Paleontol. Mineral. Spec. Publ., 3–16. Johnson, A., 1976. The climate of peru, bolivia and ecuador. World survey of climatology 12, 147–218. Laraque, A., C´eron, C., Armijos, E., Pombosa, R., Magat, P., Guyot, J., 2004. Sediments yields and erosion rates in the Napo River Basin: an Ecuadorian Andean Amazon tributary. In: V. Golosov and V. Belyaev and D.E. Walling (Ed.), Sediment Transfer through the Fluvial System. Vol. 288. pp. 220–225. Laraque, A., Ronchail, J., Cochonneau, G., Pombosa, R., Guyot, J.-L., 2007. Heterogeneous distribution of rainfall and discharge regimes in the Ecuadorian Amazon Basin. Journal of Hydrometeorology 8(6), 1364–1381. Lavado, W., Labat, D., Guyot, J., Vrugt, J., Glvez, J. O., submitted. A mesoscale and distributed conceptual monthly water balance models for large watersheds in the Peruvian Amazon-Andes basin. Hydrological Processes. Hydrological Sciences Journal. Louvat, P., Allegre, C., 1997. Present denudation rates on the island of Runion determined by river geochemistry: Basalt weathering and mass budget between chemical and mechanical erosions. Geochim. Cosmochim. Acta 61, 3645–3669. Malbrunot, A., 2006. Sedimentolog´ıa de la cuenca alta del r´ıo Pilcomayo. S´ıntesis de los conocimientos hasta la fecha, datos disponibles y elaboraci´on de los par´ametros principales aguas arriba de Misi´on La Paz (Argentina). In: IRD - SENAMHI. La Paz, Bolivia, p. 147. Meade, R., Dunne, T., Richey, J., Santos, U., Salati, E., 1985. Storage and Remobilizatin o suspended sediment in the lower Amazon River of Brazil. Science 228, 488–490. Moquet, J., Crave, A., Viers, J., Seyler, P., Armijos, E., Bourrel, L., Chavarri, E., Lagane, C., Laraque, A., Lavado, W., Pombosa, R., Noriega, L., Vera, A., Guyot, J., 2011. Chemical weathering and atmospheric/soil co2 uptake in the andean and foreland amazon basins. Chemical Geology 287, 1–26. 12

Pepin, E., Carretier, S., Guyot, J., Escobar, F., 2010. Specific suspended sediment yields of the Andean rivers of Chile and their relationship to climate, slope and vegetation. Hydrological Sciences Journal 55 (7), 1190–1205. Pinet, P., Souriau, M., 1988. Continental erosion and large-scale relief. Tectonics. Probst, J., 1990. G´eochimie et hydrologie de l ´erosion continentale: m´ecanismes, bilan global actuel et fluctuations au cours des 500 derniers million d’ann´ees. Ph.D. thesis, Institut de G´eologie, Strasbourg. Pulwarty, R., Barry, R., Hurst, C., Sellinger, K., Mogollon, L., 1998. Precipitation in the Venezuelan Andes in the context of regional climate. Meteorology and Atmospheric Physics 67, 217–237. Reiners, P., Ehlers, T., Mitchell, S., Montgomery, D., 2003. Coupled spatial variations in precipitation and long-term erosion rates across the Washington Cascades. Nature 426, 645–647. Restrepo, J., Kjerfve, B., Hermelin, M., Restrepo, J., 2006. Factors controlling sediment yield in a major South American drainage basin: the Magdalena River, Colombia. Journal of Hydrology 316, 213–232. Riquelme, R., H´erail, G., Martinod, J., Charrier, R., Darrozes, J., 2007. Late Cenozoic geomorphologic signal of Andean forearc deformation and tilting associated with the uplift and climate changes of the Southern Atacama Desert (26S-28S) . Geomorphology 86, 283–306. Roche, M. A., Aliaga, A., Campos, J., Pea, J., Cortes, J., Rocha, N., 1990. Htrognit des prcipitations sur la cordire des Andes boliviennes,. In: H. lang and A. Musy, IAHS Publ. 193 (Ed.), In Situ Produced Cosmogenic Nuclides and Quantification of Geological Processes. Hydrology in Mountainous Regions, pp. 381–388. Roche, M. A., Fernandez, C., 1988. Water resources. salinity and salt yields of the rivers of the bolivian amazon. Journal of Hydrology 101, 305–331. Roddaz, M., Viers, J., Brusset, S., Baby, P., H´erail, G., 2005. Sediment provenances and drainage evolution of the neogene amazonian foreland basin. Earth Planet. Sci. Lett. 239, 57–78. Ronchail, J., Bourrel, L., Cochonneau, G., Vauchel, P., Phillips, L., Castro, A., Guyot, J.-L., de Oliveira, E., 2005. Climate and inundation in the mamor basin (south-western amazon bolivia). Journal of Hydrology 302, 223–238. Safran, E. B., Bierman, P., Aalto, R., Dunne, T., Whipple, K., Caffee, M., 2005. Erosion rates driven by channel network incision in the bolivian andes. Earth Surf. Proc. Land. 30, 1007–1024. Safran, E. B., Blythe, A., Dunne, T., 2006. Spatially variable exhumation rates in orogenic belts: An andean example. Journal of Geology 114, 665–681. Schildgen, T. F., Hodges, K., Whipple, K., Reiners, P., Pringle, M., 2007. Uplift of the western margin of the Andean plateau revealed from canyon incision history, southern Peru. Geology 35, 523–526. Thouret, J.-C., Worner, G., Gunnell, Y., Singer, B., Zhang, X., Souriot, T., 2007. Geochronologic and stratigraphic constraints on canyon incision and Miocene uplift of the Central Andes in Peru. Earth Planet. Sci. Lett. 263, 13

151–166. Vanacker, V., Molina, A., Govers, G., Poesen, J., Deckers, J., 2007a. Spatial variation of suspended sediment concentrations in a tropical Andean river system: The Paute River, southern Ecuador. geomorphology 87, 53–67. Vanacker, V., von Blanckenburg, F., Govers, G., Molina, A., Poesen, J., Deckers, J., Kubik, P., 2007b. Restoring dense vegetation can slow mountain erosion to near natural benchmark levels. Geology 35, 303–306. Vassallo, R., Ritz, J. F., Braucher, R., Jolivet, M., Carretier, S., Larroque, C., Chauvet, A., Sue, C., Todbileg, M., Bourles, D., Arzhannikova, A., Arzhannikov, S., 2007. Transpressional tectonics and stream terraces of the GobiAltay, Mongolia. Tectonics 26(5), TC5013. Verstraeten, G., Poesen, J., 2001. Factors controlling sediment yield from small intensively cultivated catchments in a temperate humid climate. Geomorphology 40(1-2), 123–144. von Blanckenburg, F., Hewawasam, T., Kubik, P., 2004. Cosmogenic nuclide evidence for low weathering and denudation in the wet, tropical highlands of Sri Lanka. J. Geophys. Res. 109, F03008. Whipple, K. X., Tucker, G. E., 1999. Dynamics of the stream-power incision model: implication for heigth limits of mountain ranges, landscape response timescales, and research needs. J. Geophys. Res. 104, 17,661–17,674.

14

Stations Code

Latitude Longitude Altitude (deg.) (deg.) (m)

Rio

Area (km2)

Basins Avg. Max Avg. Vegetation Ande slope altitude altitude cover PLI (%) (%) (deg ) (m) (m)

FRS

Napo at Francisco de Orellana station Coca at San Sebastiàn station

-0.473 -0.343

-76.982 -77.007

225

17 640

86

14

5 855

1 265

52

4.2

BOR

Marañon at Borja station

-4.470

-77.548

140

114 270

100

19

6 160

2140

35

5.0

CHA

Huallaga at Chazuta station

-6,570

-76.119

155

70 270

100

18

4 970

1725

44

3.9

ATA

Ucayali at Atalaya station

-10.731

-73.747

185

191 020

89

17

6 305

2920

24

3.8

RUR

Beni at Rurrenabaque station

-14.441

-67.535

180

69 980

100

21

6 405

2165

33

4.3

ABA

Grande at Abapo station

-18.913

-63.194

420

59 480

100

16

5 135

2535

11

5.0

SAA

Parapeti at San Antonio station

-20.012

-63.194

605

7 430

100

14

3 530

2 650

18

4.0

Pilcomayo at Villamontes

-21.257

-63.502

380

87 350

100

14

5 915

3 260

3

3.6

VIL

Table 1 Location and various topographic and geomorphologic parameters of the catchments studied. ”Andes” is the percentage of the catchment area located above 400 m of altitude. PLI is defined in table 3. See figure 1. Rainfall (P) Avg. annual Runoff code (mm/an) CvP FRS BOR CHA ATA RUR ABA SAA VIL

3315 1310 1640 1010 1310 750 800 840

0.29 0.43 0.43 0.78 0.66 0.93 1.00 1.17

Water discharge (Qe)

P data period (yr-yr) 1963-1998 1963-2004 1963-2004 1963-2004 1963-2003 1963-2003 1968-2006 1968-2005

Qe Runoff_Qe (m3/s) (mm/an) CvQe 1680 4976 3028 4800 2034 310 89 292

2988 1373 1359 792 917 164 378 105

0.25 0.00 0.00 0.00 1.37

Suspended sediment data

Qe data period 2000-2010 1986-2010 2003-2010 1967-2010 1976-1990 1976-1985 1973-2006

Sediment flux Qs (Kt/an) 11300 136623 70256 200000 161828 178240 18164 175690

Qs data period

Basin denudation Yield rate Er (t/km2/yr) (mm/an)

2000-2010 2004-2010 2004-2010 2004-2006 1969-2010 1976-1990 1976-1983 1977-2005

640 1200 1000 1050 2310 3000 2440 2010

Table 2 Results per catchment coming from the rainfall, water discharge and suspended sediment databases. The density used to calculate the denudation rate is 2.5 t/m3.

igneous

Assigned PLI metamorphique

1

plutonic volcanic

1 2 2 4 4 10 40

sedimentary

mafic siliciclastic carbonate lutite evaporite

PLI ratio

Repartition (%) FRS

BOR

CHA

ATA

RUR

11 13 17 0 38 0 21 0

14 9 12 1 23 6 34 0

18 12 2 0 46 9 13 0.1

9 12 17 0 43 5 14 0

0 2 0 0 90 2 6 0

4.2

5.1

3.9

3.9

4.3

ABA

VIL

SAA

0 0 3 0 79 0 18 0

9 0 5 0 83 3 0 0

0 0 0 0 99 0 1 0

5.0

3.63

4.06

Table 3 Probst lithologic index PLI which represents the relative rate of erosion observed for each rock type Probst (1990) has been assigned to each major lithology defined by Moquet et al. (2011) according to previous work (Guyot, 1993; Aalto et al., 2006)

15

0.25 0.48 0.40 0.42 0.92 1.20 0.98 0.80

Slope Avg. Altitude PLI Veg. cover Runoff CVp Qs Er

Area 0.43 0.71 -0.07 -0.15 -0.34 0.04 0.75 -0.30

Slope

Avg. Altitude

PLI

Veg. Cover

Runoff

Cvp

Qs

0.21 0.36 0.11 -0.36 -0.19 0.52 0.14

-0.14 -0.68 -0.58 0.57 0.92 0.23

0.10 -0.04 -0.29 0.12 0.25

0.83 -0.97 -0.55 -0.73

-0.78 -0.60 -0.68

0.41 0.73

0.30

Table 4 Matrix of Pearson´s correlation. A matrix value of higher than 0.63 is necessary to indicate a correlation with 95% of significance.

16

CAL VENEZUELA

80°0'W

75°0'S

70°0'W

65°0'W

60°0'W

Colombia

0°0'S

Ecuador FRSNapo

es

Ama

Peru

BOR

Brazil

Solimo

zon

as

Marañon

5°0'S

CHA ali

Ucay

ga

alla Hu

ATA

15°0'S

20°0'S

Altitude 0m 500 m 2000 m 4000 m 6500 m 0 275

550

RUR

ré Mamo

Be

ni

10°0'S

Gra nde

1100 km

Bolivia

ABA SAA VIL

Chile

Fig. 1. Definition of the basins studied. See table 1 for the names, slopes and surfaces. Catchment outlets consist of hydrological stations (red square) where water discharge and suspended sediment concentrations are measured.

17

1,2

A

B

2

R = 0,8365

2000

R2 = 0,9958

0,4

1000 0

0,8

CvP

Avg. Rainfall (mm/yr.)

3000

0

0

1000

2000

Avg. discharge (mm/yr.)

3000

0

0,5

1,0

CvQ

Fig. 2. A. Correlation between average inter-annual rainfall and water discharges normalized by drainage area. B Correlation between the standard deviation of average monthly rainfall and water discharges.

18

1,5

300 mm

sep

feb

aug sep

100 mm

sep

feb

feb

aug

aug

100 mm

sep

feb

aug

100 mm

sep

feb

Mean runoff (mm/yr) 3000

100 mm

aug

Mean denudation rate Er (mm/yr.)

sep

aug

1 0,9 0,8

1500

feb

100 mm

0,4 0.25

sep

feb

aug

1000 750

station

sep

feb

aug

Fig. 3. Current denudation rates calculated from average inter-annual suspended sediment discharges (in gray scale). Drops illustrate mean average runoff calculated from rainfall time series. Diagrams indicate annual precipitation regimes and highlight the increase of climate variability southern of the studied area.

19

Denudation rate Er (mm/yr)

1.2

R 2 = 0.55

0.8

0.4

0

0

0.4

0.90 CVp

Denudation rate Er (mm/yr)

1.2

0.8

1.2

0.8

1.2

0.77

R 2 = 0.78

0.8

0.4

0

0

0.4

Denudation rate Er (mm/yr)

-1.53+0.96 CVp+0.29 PLI+0.017 Slope

R 2 = 0.85

1.2

0.8

0.4

0

0

0.4

0.8 1.7

0.05 CVp

1.24

. PLI

1.2

.Slope

Fig. 4. A. No-linear regression between the denudation rate and the Climate variability parameter CVp that is the main factor of control. B. Combining CVp with the lithologic and Slope secondary factors improves the estimation of the denudation rate using both linear regression feature (B) and no-linear regression feature 20 (C).