This paper intends to investigate which are the major factors controlling river sediment fluxes to the oceans, and whether these factors can be used for prediction and modeling of river sediment yields (FTSS). For this purpose, the empirical relationships between FTSS and a large number of hydroclimatic, biological, and geomorphological parameters are investigated at different scales.
For a set of 60 major world river basins, the environmental characteristics were extracted from various global data sets, and the river sediment fluxes were taken from the literature. The selected river basins are representative for the major climate types on Earth (Ludwig et al., 1996), and FTSS vary almost within three orders of magnitude (4 - 2514 t km-2 yr-1). At this level, FTSS can be best correlated by forming the products of hydroclimatic and geomorphological factors, that is runoff intensity, basin slope, an index characterizing rock hardness, and an index characterizing rainfall variability over the year. When extrapolated to the total continental area on the basis of the corresponding data sets, the empirical models result in a spatial distribution of river sediment yields that is in good agreement with field data (Ludwig and Probst, 1998), confirming thus the applicability of our models for global scale modeling.
On more regional scales, however, this is not necessarily the case. We compare therefore our results with the results of a similar study that has been done on a much smaller scale (Maneux, 1998). In this study, the river sediment fluxes in 44 catchments of the French Adour, Dordogne, andGaronne river systems have been measured by weekly to daily sampling during three consecutive years. The environmental characteristics of the watersheds were determined by integrating regional an/or european databases in a Geographical Information System (GIS). Although climatic variability between the catchments is naturally less important compared to our global scale study, FTSS still vary within 1-2 orders of magnitude (3 - 150 t km-2 yr-1). Also at this level, FTSS can be correlated with the product hydroclimatic and geomorphological factors, but several outliers occur in the regressions, which seem to be related to more erodible soil types and/or lithologies within the watersheds (such as for the Tarn river).
These factors are still difficult to quantify in modeling studies. Going from smaller to larger scales, their effect on mechanical erosion rates and river sediment yields may average out, and hydroclimatic and morphological factors may become the dominant controls, for which variability increases. Nevertheless, also for global scale modeling, the role of lithology is one of the major unknowns (Ludwig and Probst, 1998), and better parameterizations of this factor may also improve modeling at larger scales.
Ludwig, W, Probst, J-L & Kempe, S, Global Biogeochemical Cycles, 10, 23-41, (1996).
Ludwig, W & Probst, J-L, American Journal of Science, 296, 265-295, (1998).
Maneux, E, PhD thesis, University Bordeaux I, 252 p, (1998).
The role of climate in silicate weathering (in particular temperature) is still highly controversial. A comparison of river chemistries for the Mackenzie and Amazon river basins offers the possibility of comparing silicate weathering rates under cold (arctic and subarctic basin of the Mackenzie river) and hot climate (Amazon river basin).
We will present the first Sr isotopic systematics for the dissolved and solid products of erosion transported in the Mackenzie River basin. 87Sr/86Sr ratio range from 0.7102 to 0.7343 in the sedimentary part of the basin, while the Canadian Shield displays higher values close to 0.7424. These data as well as major and trace element ratios confirm the dominant control of lithology for river dissolved load. Based on mixing diagrams, we estimate the contribution from the various lithologies (marine limestones, dolomites, evaporites and aluminosilicates) present in the river basin.
The fraction of each element coming from silicate weathering and CO2 consumption rates for the Mackenzie river and its main tributaries, can then be calculated. We find that 160, 24, 14 and 28*109 mol of CO2/year, are consumed by silicate weathering in the Mackenzie, Liard, Peace and Athabasca rivers respectively. These values are three to four times higher than those given by Edmond et al., (1996) for silicate weathering. We have also calculated the Total Dissolved Solids (TDS) load produced by silicate weathering for the main tributaries of the Mackenzie river basin, in order to compare with the values for the Amazon river basin (Gaillardet et al., 1997). The silicate TDS values for the andean tributaries of the Amazon river basin are six to nine times higher than those of the Mackenzie river basin. This result reemphasizes the idea that silicate weathering rates are highly sensitive to temperature on a global scale.
Finally, the role of organic matter in weathering processes can be addressed in the Mackenzie river basin, which gives examples of both low dissolved organic carbon (DOC) rivers (highland rivers) and DOC rich rivers (black rivers of the lowlands). We find a positive correlation between sodium or alcalinity derived from silicate weathering and DOC concentrations, suggesting that under the subarctic conditions of the Mackenzie basin, dissolved organic matter may be a key parameter controlling the weathering processes and possibly the silicate weathering rates.
Edmond J.M., Palmer M.R., Measures C.I. & Huh Y., Geochim. Cosmochim. Acta, 60, 2949-2976, (1996).
Gaillardet J., Dupré B., Allègre C.J. & Négrel P., Chem. Geol., 142, 141-173, (1997).
Knowledge of the relative proportions of silicate and carbonate rock which contribute to the dissolved load in rivers and the relationship of river water 87Sr/86Sr ratios to those of the carbonate and silicate bedrock are critical for understanding the impact of tectonics on long-term climate change and the significance of oceanic 87Sr/86Sr as a proxy for global weathering rates. Recent work on the headwaters of a Himalayan river (Blum et al., 1998) has concluded that small fractions of carbonate dispersed in predominantly silicate gneisses contribute disproportionately to the riverine dissolved load. Our results on a number of tributaries draining similar High Himalayan Crystalline Series gneisses in the Headwaters of the Ganga do not show the elevated Ca/Sr ratios ascribed to the dispersed carbonate. However carbonate rocks in the catchment exhibit a wide range of Ca/Sr molar ratios (0.002 to 0.04) and some impure carbonate rocks have 87Sr/86Sr ratios in excess of 1.0. Leaching experiments on silicate rocks, carbonate rocks and synthetic mixtures show that care has to be taken to remove exchangeable Sr and in the leaching of carbonate from predominantly silicate rocks. The results show that two components contribute to the high 87Sr/86Sr in the river waters. One component comes from impure carbonate rocks which have elevated 87Sr/86Sr ratios due to their early Proterozoic age, coupled with recent metamorphic exchange between silicate and carbonate minerals. The other is the high 87Sr/86Sr silicate rocks. Calculation of the fractions of silicate and carbonate contributing to the dissolved load is not straightforward. We show that it is essential to combine information on bed rock composition as well as the dissolved load, suspended load and bed load to calculate the relative fluxes.
Blum JD, Gazis CA, Jacobson, AD & Chamberlain C Page, Geology, 26, 411-414
Continental erosion and river transport of organic and inorganic materials are an important sink of atmopsheric/soil CO2, particularly silicate rock weathering. Several factors are controlling these biogeochemical processes at a global scale but the most important factors are the climate (precipitation, air temperature and runoff) as shown by Berner (1991) and White and Blum (1995) and the relief (elevation, slope) as shown by Raymo et al. (1988).
If we look at the present-day rock weathering, the fluxes of atmospheric/soil CO2 consumed by silicate and carbonate rock weathering is mainly controlled by the runoff intensity. This has been shown for small monolithologic watersheds (Amiotte Suchet and Probst, 1993 and 1995) as well as for large river basins (Amiotte Suchet, 1995 and Boeglin and Probst, 1998). Moreover, the flux of dissolved silica, mainly released by silicate rock weathering and transported by rivers into the oceans is also proportional to the runoff (Probst, 1992). But for a given runoff intensity, the dissolved silica river flux is higher in tropical and equatorial regions (higher air temperature) than in temperate and subarctic regions (lower air temperature). Nevertheless this air temperature effect is less important than the runoff control at this large river basin scale.
An important question is whether atmospheric CO2 variation and climate change are due to enhanced weathering or whether these variations have changed weathering rate and CO2 uptake. It has been proposed for example that increased rates of chemical weathering and the related drawdown of atmospheric CO2 on the continents may have at least partly contributed to the low CO2 concentrations during the glacial periods (Gibbs and Kump, 1994; Munhoven and François, 1996). Variations in continental erosion could thus be one of the driving forces for the glacial/interglacial climate cycles during Quaternary times for example. Nevertheless it could be shown recently (Ludwig and Probst, in press and Ludwig et al. in press) that LGM climatic conditions reconstructed on the basis of GCM outputs were drier than today, i.e the runoff was 9% lower during LGM than today. Consequently the LGM CO2 uptake by silicate rock weathering were 10% lower than today.
These results show that among the climatic factors, the runoff plays a major role in controlling silicate weathering rate and CO2 uptake. These results show also that at the time scale of glacial/interglacial periods, continental weathering is probably not the driving force of atmospheric CO2 content and climate change but it represents a response to climate change, particularly runoff variation, and consequently a possible feedback mechanism with regards to increase or decrease of atmospheric CO2 content.
Variations in the isotopic composition of osmium in seawater, recorded by marine sediments, are a proxy of the relative variations of continent, mantle and meteorites contributions. The osmium isotopic ratios of these three potential sources to the ocean enable to estimate the fraction of seawater osmium coming from the continent to about 80% in modern times. Rivers should then provide the major flux of osmium to the oceans and play a major role in defining the marine 187Os/186Os ratio. Variations in this source are most probably an important cause of variation of the composition of seawater osmium through time and as such the osmium composition in seawater is potentially a good record of continental weathering. In order to understand the significance of the temporal changes in the seawater, it is necessary to understand the contemporary relations between the continental weathering processes and osmium composition and flux in rivers leading to the mean continental flux to the ocean. Due to the analytical difficulty to measure its very low concentration in natural waters, very little is known, by now, about osmium in rivers. Here we present the first extensive results of isotopic ratios and concentration of osmium in river water on a collection of samples from major world rivers (Amazon, Congo, Orinoco, Yangtze, Lena, Ganges, St-Lawrence, Mackenzie, Red river, HuangHe, Yukon, Narmada) analysed using the method we have already applied to seawater. The isotopic composition and concentration measurements of the sampled major rivers vary between 8.8 and 19.6 (187Os/186Os ratio) and between 3.5 and 52.3 pg/kg respectively (although the whole variation scale of the composition is larger if we consider some small rivers). We will discuss these results in term of flux to the ocean and try to decipher the endmembers of which weathering products determine the composition of river osmium.
Continental crust is on average more radiogenic in its Os isotopic composition than present-day seawater and thus constitutes the radiogenic source of Os in the oceans. The most radiogenic runoff observed thus far is from old cratonic shields surrounding the North Atlantic (Peucker-Ehrenbrink and Ravizza, 1996; Peucker-Ehrenbrink and Blum, 1998), leading to slightly elevated Os isotope ratios in Atlantic seawater compared to other ocean basins (Burton et al., 1996). The source lithology for mobile radiogenic Os, however, remains elusive. Prime candidates are Precambrian crystalline rocks and black shales. If radiogenic Os were primarily released through weathering of silicate continental rocks, changes in the Os isotopic composition of seawater would record changes in the intensity of silicate weathering - one of the major controls of atmospheric carbon dioxide on geologic time scales. It is tempting to interpret the observed secular trend in the marine Os isotope record during the Cenozoic toward more radiogenic values in such a way as this is consistent with the observed long-term cooling of the Earth. If, however, weathering of black shales controls the delivery of radiogenic Os to seawater, the observed trend in the marine Os isotope record is closely related to release rather then sequestration of carbon dioxide through organic matter oxidation during the Cenozoic. We have studied the mobility of Os during black shale weathering by comparing time-equivalent unweathered (drill core) and weathered (surface) late Ordovician (Caradocian) black shales from the Taconic Foreland Basin of Quebec, Canada. Fresh samples are characterized by Os concentrations of 0.5 - 2.6 ng/g and 187Os/188Os of 2.49 - 9.43. They are thus significantly more radiogenic than average continental crust (187Os/188Os ~ 1.3 - 1.9) and contain one to two orders of magnitude more Os. Weathered surface samples are much less radiogenic than their fresh equivalents (187Os/188Os of 0.66-0.84), indicating preferential release of radiogenic Os, and are significantly depleted in Os (Os concentrations of ~0.2 ng/g). Our data set indicates substantial (64-92%) loss of radiogenic Os during weathering of black shales. In contrast, the leachable (mobile) Os fraction in crystalline rocks is typically less than 5% (Peucker-Ehrenbrink and Blum, 1998). Weathering of one mass unit black shale therefore releases an equivalent amount of Os as weathering of several thousand mass units felsic crystalline rocks. This order of magnitude estimate highlights the importance of black shale weathering for the marine Os isotope record. It also raises doubts as to whether the marine Os isotope record can be used as a faithful recorder for changes in the intensity of the globally integrated silicate weathering intensity over time.
Peucker-Ehrenbrink B & Blum, JD, Geochim. Cosmochim. Acta, 62, in press, (1998).
Peucker-Ehrenbrink B & Ravizza, G, Geology, 24, 327-330, (1996).
Burton KW, et al, J. Conf. Abs, 1, 92, (1996).
The mean osmium isotope composition of sea water (187Os/188Os = 1.04 reflects the balance between the input of osmium released during the chemical weathering of continental rocks (187Os/188Os = 1.3) and osmium derived from the dissolution of cosmic dust and/or hydrothermal alteration (187Os/188Os = 0.12). Changes in the magnitude or composition of these fluxes over time will be reflected as changes in the osmium isotope composition of seawater. While not well-constrained, the oceanic residence time of osmium is less than 40 Kyr. Thus the seawater Os isotope curve for the Pleistocene (as recorded in marine sediments) can provide information on changes in chemical weathering over these timescales. While the Pleistocene has seen large changes in ice cover, mean global temperature and continental aridity, changes in mean global relief and exposed lithology have been minimal or can be well-constrained. This means that the potential exists to isolate changes in chemical weathering due to climatic influences.
The Pleistocene marine osmium isotope record should help to establish the links or lack thereof between climatic variables and chemical weathering rates. Minima in the 187Os/188Os of seawater occur late in oxygen isotope stages 2 and 6 (indicating reduced weathering rates of the continents) (Oxburgh, 1998). There does not appear to be a direct correllation between mean global temperature and weathering rate; rather it seems that chemical weathering rates are reduced during times when it is both extremely cold and extremely dry. However, the published data are not of sufficient resolution to identify precisely the timing of the shifts in seawater composition and thus also their phasing relative to climatic changes that may drive the changes in weathering rate. I here present a high resolution osmium isotope record from rapidly accumulating (30 cm/Kyr) sediments from the Cariaco Basin that span the past 20 Kyr. These data allow the timing and nature of the change in composition since the last glacial maximum to be well defined and thus improve constraints on the role of climate in driving changes in chemical weathering rate.
Oxburgh, R., Earth and Planet. Sci. Lett., 159, 183-191
The erosional mass flux into the oceans can be constrained by the isotopes of lead, because they trace the provenance of the material eroded. Furthermore, the isotopic composition of Pb that is mobilised from rocks and soils is controlled by the degree and duration of weathering in the source area (Erel et al. 1994). Hence the isotopic composition of natural Pb that is released into seawater is a function of both provenance and the maturity of weathering in the source area.
The areas surrounding the Labrador Sea are amongst the lowest in 206Pb /204Pb and the highest in 208Pb/206Pb world-wide. In contrast, pre-anthropogenic seawater Pb in Northwest Atlantic deepwater is the highest in 206Pb/204Pb world-wide (von Blanckenburg et al. 1996). We have leached Pb from Archean gneisses and from the detrital fraction of Baffin Bay and Labrador Sea core top sediments. Fresh rocks leached for two days in an aqueous NaCl solution release about 0.1% of the rock's total Pb. 206Pb/204Pb in the leacheate increases from 11.5 to 13.0, and from 15.2 to 18.8. In a 0.05 M HCl leach, 206Pb/204Pb increases from 11.5 to 14.8 and from 15.2 to 21.1. In the sediments, 0.1 to 0.2% Pb was released in the NaCl solution, with 206Pb/204Pb increasing from 16.9 to 17.2, and to 17.8 in 0.05 M HCl. In one rock 208Pb/204Pb of 40 increased to 190. In all cases 208Pb/206Pb is lower in the leacheates than in the total rock.
Labrador Sea core top sediments show a pronounced release of slightly more radiogenic Pb, despite going through the entire erosion-sedimentation cycle. This is in line with rapid glacial erosion and ice rafting being the predominant mechanism of sediment transport in this area, and explains the contradiction between the non-radiogenic source area and the radiogenic seawater Pb. Also, time series analyses of Pb in dated NW Atlantic ferromanganese crusts has revealed a marked increase in paleo-seawater 206Pb/204Pb and a decrease in 208Pb/206Pb since the Pliocene (Burton et al. 1997, O'Nions et al. 1998). In pre-glacial times, or in other areas, where slow weathering is leading to more mature erosion products, the released Pb is expected to be less radiogenic. Hence Pb isotopes in detrital marine sediments have potential to serve as a proxy for paleo-weathering conditions in the source area.
Burton KW, Ling HF & O'Nions RK, Nature, 386, 382-385, (1997).
Erel Y, Harlavan Y, Blum JD, Geochim. Cosmochim. Acta, 58, 5299-5306, (1997).
O'Nions RK, Frank M, von Blanckenburg F, Ling HF, Earth Planet. Sci. Lett, 155, 15-28, (1998).
von Blanckenburg F, O'Nions RK & Hein JR, Geochim. Cosmochim. Acta, 60, 4957-4963, (1996).
The combined information about the Oligo/Miocene floral change in Europe, the stratigraphies from the foreland basins surrounding the Swiss Alps, and the structural evolution of the Alpine orogen allow an evaluation of the controls of climate on the large-scale Alpine tectonic evolution. At the Oligocene-Miocene boundary, the floral assemblage changes from leaves that are characteristic for a deciduous forest (e.g. cypresses, pecans) towards an abundance of legumes and pines that indicate the presence of a notophyllus woodland. This floral change implies a significant decrease of average precipitation rates. The Oligo/Miocene aridization of climate appears to have resulted in a decrease of Alpine erosion rates. Indeed, volumetric data from the North and South Alpine Foreland Basins (NAFB, SAFB) indicate a ca. 50% decrease of average sediment flux to the basins from 10-20 km2/My (NAFB) and 5-10 km2/My (SAFB) prior to 20 Ma to 5-10 km2/My (NAFB) and 2-5 km2/My (SAFB) after 20 Ma. Also at ca. 20 Ma, rerouting of the Alpine paleorivers from an across-strike to the present-day along-strike orientation was initiated, indicating surface uplift. This implies the inability of erosion to keep pace with tectonic uplift. Most importantly, the decrease of average erosion rates at that time coincides with cessation of slip movement along the Insubric Line (Insubric phase of deformation) which prior to ca. 20 Ma represented the southernmost extent of the Alpine orogen. At 20 Ma, the tip of the orogenic wedge shifted by >50 km towards distal sites (Lombardic phase of deformation). Deactivation of the Insubric Line as the limiting shear zone and expansion of the orogenic tip towards distal sites suggests that the Swiss Alps entered a phase of orogenic growth. According to coupled erosion-mechanical models of orogens this phase of orogenic growth could have been caused by the reduction of average erosion rates at ca. 20 Ma. Since paleobotanic data suggest a shift towards drier climate at that time, we speculate that climate was the primary control on initiation of the phase of orogenic growth.
In order to evaluate mutual influence of climate and uplift during the Oligocene and Miocene evolution of the Eastern Alps, we aimed to reconstruct the uplift history and climatic changes since Oligocene times. The palaeoclimatic analyses of latest Eocene to Late Pannonian palyno- and megafloras from the Eastern Alps and surrounding sedimentary basins yield relatively homogeneous climatic conditions in space and time. However, the data from the northern Alpine area show distinct cooling and a decrease of summer precipitation at the Eocene/Oligocene boundary. In addition to the global cooling signal at that time, the isolation of the Paratethys Sea caused a significant weakening of the oceanic influence in that region. Around the Early/Middle Miocene boundary a cooling trend for winter temperatures, but stagnant summer temperatures are recorded for the Molasse basin; thus, the seasonality increased. The synchronous regression of the Paratethys in the Molasse Basin and of the North Sea again indicates that the decrease of oceanic influence was the main force responsible for the palaeoclimate development. In the Pannonian area, a second significant cooling event (decreasing summer and winter temperatures) in combination with a decrease in summer precipitation took place during the Middle and/or Late Miocene. Here the increasing tectonical uplift of the easternmost Alps together with the final isolation of the Paratethys from the oceans seem to control the mechanisms of climatic changes. Reconstruction of the geomorphologic evolution reveals that the western part of the Eastern Alps already was mountainous since middle Oligocene times; peak elevations certainly exceeded 2000 m. At the same time, the eastern part of the region remained mainly hilly. This part did not attain a mountainous relief before the Late Miocene. From the palaeoclimatic data we infer that the morphology of the Eastern Alps did not substantially influence the climate pattern before Late Miocene times, when mean annual temperatures and precipitation differentiated north, east, and south of the Eastern Alps. The data show that the land-sea-distribution is of prime importance for homogeneity and seasonality and thus represents a ruling factor in the Tertiary climatic development of the Eastern Alpine region. Geomorphologic influence only became an important factor in the very late history of the mountain range.
Against a general trend of global cooling beginning in the Eocene, the precipitous recent drop in ocean temperature and the growth of northern hemisphere glaciation began between 4(?) and 3 Ma ago. The reason why northern hemisphere glaciers suddenly grew and stabilised is not known. Suggestions have included oceanographic circulation changes (e.g. closure of Panama Isthmus), and tectonic uplift of large plateaux such as Tibet or western United States, which would have changed atmospheric circulation and affected its chemistry. This paper proposes that there is a more plausible explanation: the rise of the Coastal Mountains of northwestern North America from latitude 46N to central Alaska during the last 7 Ma. This tectonically-controlled coastal mountain system would have induced significant changes to atmospheric circulation and affected heat transport to the arctic, and tipped the scales in favour of continental glaciation. Fission track, palaeoclimatic, geological and geomorphological evidence, mainly in British Columbia, implies that the bedrock of the present Coast Mountains rose 1.5-3.5 km in the past 7 Ma. Detailed studies north of Vancouver (51N), near Mt Waddington (lat. 53N), Bella Coola-Prince Rupert (55N), southeast Alaska, and near Mt Logan (60N) all imply major mountain uplift and deep valley incision at rates of 0.5-1.5 km/Ma in the past several million years, resulting in the current rugged topography. The present Coastal mountain system varies in elevation but forms an important weather barrier with peaks between 2-4 km, with the St. Elias Range exceeding 5 km. Prior to the rise of these mountains, when topography was much less pronounced, warmer moist Pacific air would have flowed well into the interior of the continent, providing a temperate moderating influence: a climatic pattern consistent with Miocene flora of the continental interior and intermontane regions of western Canada. The rise of the mountains beginning in late Miocene time would have exerted a strong influence on this circulation pattern, channelling Pacific air northwest more parallel with the coast of North America, thus isolating the continental interior and allowing cold winter air masses to become established: an important prerequisite for the growth and stability of North American continental ice sheets. The changes in northern hemisphere atmospheric circulation affected the balance of heat transport to northern high latitudes. This longitudinal change in heat transport is a logical consequence of the rise of a continuous system of mountains along the coast of western North America, and may provide a more direct and most plausible explanation for the onset of stable northern hemisphere glaciation in the Pliocene.
There is an ongoing controversy whether widespread Late Cenozoic uplift has caused climate change or whether Cenozoic climate change has increased erosion rates in such a way as to create the illusion of uplift. Detrital erosion rates have increased in both high and low latitudes by about a factor of four since the mid-Cenozoic. The increase in detrital erosion rates could be due to: 1) increased elevation or relief, 2) a change from evenly distributed to seasonal rainfall, 3) changes in vegetative cover of land, and 4) the sea-level falls associated with increasingly extensive glaciation. Detrital erosion rates increase with increasing average elevation of drainage basins, as well as with basin and local relief. This has often been the basis for arguing that uplift has occurred. A change from rainfall evenly distributed throughout the year to strongly seasonal rainfall (monsoons) could also cause a major increase in erosion rates. This may have occurred as a result of increasing continentality associated with the closure of the Tethys and its loss as a source of water vapor. The evolution and spread of water-conserving C4 plants may have increased aridity and promoted increased erosion rates. The lowering of sea-level as glaciation has spread may have caused offloading of sediment from the continental shelves into the deep sea. Discrimination among these possibilities is difficult using geologic data, but may be possible by analysis of coupled climate-vegetation-erosion models.
Index of EUG 10 Volume
Further EUG 10 Information
Index of the Journal of Conference Abstracts
Cambridge Publications Home Page
Last Updated on Wednesday, March 17, 1999.
© 1997 Cambridge Publications