『Abstract
　Rivers from the upper Rio Madeira basin (Bolivia) have been studied with uranium-series isotopes in order to constrain the timescales of weathering and sediment transfer from the Andes through the Amazon tropical plain. Uranium (U), thorium (Th) and radium (Ra) isotopes (²³⁸U-²³⁴U-²³⁰Th-²²⁶Ra and ²³²Th) have been analyzed in the suspended load (＞0.2μm) of rivers. Increasing ²³⁰Th excesses relative to ²³⁸U in suspended particles from the Andes to the tropical plain is interpreted as an increasing duration of weathering during sediment transport and storage in the foreland basin. Model calculations for (²³⁰Th/²³⁸U) and (²²⁶Ra/²³⁰Th) activity ratios in suspended particles using a continuous weathering model indicates that: (i) the timescale for production, storage and transport of sediments in the Andean Cordillera is only a few kyr, (ii) the storage time of suspended sediments in the foreland basin is 5±1 kyr and (iii) the average transfer time of suspended sediments from the Andes to the confluence of Rio Madeira with the Amazon River is 17±3 kyr. An implication of these short timescales is that the bedrock eroded must have lost part of its uranium during one or several past erosion cycles. This demonstrates the recycling of sediments through several erosion cycles before transfer to the oceans. The calculation of long-term (＞1 kyr), steady-state erosion rates indicates that they are much lower than present-day rates. This increase in denudation rates must be recent and could be explained by an increase in precipitation 〜4 ka ago, as suggested by palaeoclimatic evidences and the draining of transient sedimentary basins encountered on the Altiplano and easily eroded. This suggests that climatic variability rather than tectonics alone produces high erosion rates.

Keywords: radioactive disequilibrium; uranium-series; erosion; Amazon; Andes; sediment transport』

1. Introduction
2. The upper Madeira drainage basin
3. Analytical procedures
4. Results
5. Timescale of weathering during transport
6. Assessment of the steady-state nature of erosion
7. Conclusions
Acknowledgments
References

Fig. 2. Map of the studied area showing the mean residence times of sediments in each region (from south to north: Andes, Andes + foreland basin, upper Madeira basin, whole Madeira basin; see text for details on the calculations). 〔Dosseto,A., Bourdon,B., Gaillardet,J., Maurice-Bourgoin,L. and Allegre（最初のeの頭に｀）,C.J.(2006): Weathering and transport of sediments in the Bolivian Andes: Time constraints from uranium-series isotopes. Earth and Planetary Science Letters, 248, 759-771.〕から〕

Table 4. Physical erosion rates as measured by sediment gauging (present-day) and predicted by a steady-state erosion model using the ²³⁰Th-²³⁸ disequilibrium

River name	Erosion rate (mm/kyr)
	Present-day	Steady-state
Coroico (Guanay)	690	99 ± 3
Tipuani	100	-
Challana	37	-
Mapiri	410	53 ± 1
Alto Beni	1900	1060 ± 30
Beni (Rurrenabaque)	1100	41 ± 2
Madre de Dios	200	-
Beni (Riberalta)	480	58 ± 2
Orthon	16	23 ± 1
Mamore（eの頭に´）	42	67 ± 2
Madeira (Porto Velho)	115	17 ± 1
Erosion rates are calculated using mean annual discharges, drainage areas from Table 1 and suspended sediment concentrations. For present-day rates, sediment concentrations are taken from Table 1 and represent multi-year average sediment gauging [8] and [28]. For steady-state rates, sediment concentrations are calculated using the 230Th-238　disequilibrium in the dissolved and particulate loads (Table 2) and assuming steady-state erosion (see text for details). Errors are given at the 2σ level.