Claude J. Allegre

Global silicate weathering and CO2 consumption rates deduced from the chemistry of large rivers

Abstract The main problem associated with the study of silicate weathering using river dissolved load is that the main control of solute chemistry is lithology and that all rivers are influenced by carbonate and evaporite weathering. In this paper, newly compiled data on the 60 largest rivers of the world are used to calculate the contribution of main lithologies, rain and atmosphere to river dissolved loads. Technically, an inverse method is used to solve a model containing of a series of mass budget equations relating river concentrations to chemical weathering products and atmospheric inputs. New estimates of global silicate weathering fluxes and associated CO2 consumption fluxes are given. The role of basalt weathering on oceanic islands and volcanic arcs is emphasized. For each large river, an attempt is made to calculate chemical weathering rates of silicates per unit area. Only relative chemical weathering rates can be calculated. The relationships between the chemical weathering rates of silicates and the possible controlling parameters are explored. A combined effect of runoff-temperature and physical denudation seems to explain the variability of modern silicate chemical weathering rates. The results of this study highlight the coupling between the physical and the chemical processes of silicate weathering. Only an active physical denudation of continental rocks seems to be able to maintain high chemical weathering rates and significant CO2 consumption rates.

Major and trace elements of river-borne material: The Congo Basin

The Congo river Basin is the second largest drainage basin in the world, after the Amazon. The materials carried by its main rivers provide the opportunity to study the products of denudation of a large fraction of the upper continental crust of the African continent. This paper presents the chemical composition of the different phases carried in the Congo rivers and is followed by a companion paper, devoted to the modelling of major and trace elements. n nThe Congo river between Bangui and Brazzaville as well as its main tributaries, including a few organic-rich rivers, also called Black rivers, were sampled during the 1989 high water stage. The three main phases (suspended load, dissolved load, and bedload) were analysed for twenty-five major and trace elements. n nConcentrations normalized to the upper continental crust show that in each river, suspended sediments and dissolved load are chemical complements for the most soluble elements (Ca, Na, Sr, K, Ba, Rb, and U). While these elements are enriched in the dissolved loads, they are considerably depleted in the corresponding suspended sediments. This is consistent with their high mobility during weathering. Another type of complementarity is observed for Zr and Hf between suspended sediments and bedload, related to the differential velocity of suspended sediments and zircons which are concentrated in bedloads. n nCompared to other rivers, absolute dissolved concentrations of Ca, Na, Sr, K, Ba, Rb, and U are remarkably low. n nSurprisingly, high dissolved concentrations are found in the Congo waters for other trace elements (e.g., REEs), especially in the Black rivers. On a world scale, these concentrations are among the highest measured in rivers and are shown to be pH dependent for a number of dissolved trace elements. The dissolved loads are systematically normalized to the suspended loads for each river, in order to remove the variations of the element abundances owing to source rock variations. Normalized diagrams for REEs are presented and extended to the other elements. They strongly support the argument that the apparent higher solubility of trace elements in the Congo waters is due to the presence in the dissolved load of a colloidal phase (as a result of 0.2 μm filtration). An important result is that these colloids are strongly depleted in Fe and Al with respect to the other elements. n nFinally, the comparison of the dissolved, suspended, and sandy transport fluxes of each element in the Congo Basin rivers shows that, although the proportions of, for example, the REEs in the dissolved loads of the majority of the Congo Basin rivers is close to 10% of the total transport flux, up to 80% of the REEs are transported by the so-called “dissolved”load in the Black rivers.

Geochemistry of large river suspended sediments: Silicate weathering or recycling tracer?

This study focuses on the major and trace element composition of suspended sediments transported by the world’s largest rivers. Its main purpose is to answer the following question: is the degree of weathering of modern river-borne particles consistent with the estimated river dissolved loads derived from silicate weathering? n nIn agreement with the well known mobility of elements during weathering of continental rocks, we confirm that river sediments are systematically depleted in Na, K, Ba with respect to the Upper Continental Crust. For each of these mobile elements, a systematics of weathering indexes of river-borne solids is attempted. A global consistency is found between all these indexes. Important variations in weathering intensities exist. A clear dependence of weathering intensities with climate is observed for the rivers draining mostly lowlands. However, no global correlation exists between weathering intensities and climatic or relief parameters because the trend observed for lowlands is obscured by rivers draining orogenic zones. An inverse correlation between weathering intensities and suspended sediment concentrations is observed showing that the regions having the highest rates of physical denudation produce the least weathered sediments. Finally, chemical and physical weathering are compared through the use of a simple steady state model. We show that the weathering intensities of large river suspended sediments can only be reconciled with the (silicate-derived) dissolved load of rivers, by admitting that most of the continental rocks submitted to weathering in large river basins have already suffered previous weathering cycles. A simple graphical method is proposed for calculating the proportion of sedimentary recycling in large river basins. Finally, even if orogenic zones produce weakly weathered sediments, we emphasize the fact that silicate chemical weathering rates (and hence CO2 consumption rates by silicate weathering) are greatly enhanced in mountains simply because the sediment yields in orogenic drainage basins are higher. Hence, the parameters that control chemical weathering rates would be those that control physical denudation rates.

Geochemical applications of quantitative ion-microprobe analysis

A method for quantitative analysis of silicates with an ion-microprobe has been developed by suppressing the intensities of interfering molecular ion species based on the difference in the kinetic energy distributions between molecular and single-atom ions. Molecular ions drop their intensities rapidly with increasing energy and are virtually eliminated at the energy level 150eV higher than the energy acquired by acceleration. n nOutlines of geochemical applications of the technique are given for: n1. n(a) analysis of isotopic composition of Pb in galena n n2. n(b) analysis of isotopic composition of Mg in anorthite and hibonite in inclusions in the Leoville meteorite n n3. n(c) quantitative analysis of major and trace elements in Ca-rich pyroxene, garnet and plagioclase n n4. n(d) determination of REE pattern in hornblende; and n n5. n(e) in-depth profiling of 18O16O ratios to determine oxygen self-diffusion in silicates in hydrothermal conditions. n n n n nAlthough the potential capabilities of an ion-microprobe are not well-realized at the present stage, due in part to the lack of well-documented standards for trace elements and a relatively large beam spot size, it is concluded, considering the technical development foreseeable in the near future, that the ion-microprobe should prove extremely useful in geochemistry.

Trace elements behavior in granite genesis: A case study The calc-alkaline plutonic association from the Querigut complex (Pyrénées, France)

The different granitoids of the zoned Querigut complex (Hercynian Pyrenees) are associated with a series of basic to intermediate rocks ranging from hornblende-bearing peridotites to quartz-diorites. The whole complex appears as a calc-alkaline plutonic suite typical of orogenic zones. The distribution of lanthanides and other trace elements amongst coexisting minerals indicate they are essentially held by accessory phases, particularly in granitoids. This restricts the use of those elements in the calculation of petrogenetic models for acidic plutonic rocks. Magmatic differentiation, mainly by hornblende + plagioclase fractionation, can produce the basic series. This differentiation cannot directly produce the different granitoids, which require a preponderant contribution of crustal melts. The sequence of different granitoids can be explained either by an heterogeneity in the source region, or by magmatic differentiation. The most plausible interpretation of the whole complex calls for the emplacement of a mantle-derived magma into a wet, anatectic continental crust, with interactions between basic rocks and the soproduced acidic melts.

Chemical composition of the Earth and the volatility control on planetary genetics

Abstract Using trace element ratios with a common reference to a refractory element, we have shown that carbonaceous chondrites define a straight line in every diagram including the semi-volatile and volatile elements with the relative position of CI, CM, CO and CV always following the same order. We show that bulk Earth values estimated only by terrestrial consideration, using Mg/Al for the refractories or K/U, Rb/Sr for the volatiles, plot on the carbonaceous chondrite line but not within the group of ordinary chondrites. The position on the carbonaceous chondrite line varies according to volatility. Highly refractory elements are close to CI, moderate refractories close to CM and volatiles away from CV. Such systematics permit the calculation of the bulk composition of the Earth for every element. Those observations are in agreement with a condensation temperature of the Earth ranging from 11 to 1200°K.

Isotopic and Chemical Effects Produced in a Continuously Differentiating Convecting Earth Mantle

Trace element variations have established the present concept of chemical heterogeneity of the Earth’s mantle. A continuous range of variations is observed for the mantle source of basalts, from mantle depleted in incompatible trace elements to mantle enriched in such elements, in particular light rare earth elements. This heterogeneity is supported by the study of orogenic lherzolites and ultramafic nodules from volcanoes and kimberlite pipes. Radiogenic isotopes of Sr, Pb and Nd confirm this heterogeneity and show that it results from ancient chemical fractionations. This trace element and radiogenic isotope heterogeneity is identifiable in the Precambrian mantle. Pb isotopes are the clearest tracers of this past heterogeneity. Correlations are observed between variations in incompatible element ratios, radiogenic isotope ratios and radiogenic isotope and trace element ratios. These correlations define large domains (source of alkali basalts, source of mid-oceanic ridge basalts) and more restricted domains (e.g. the Canary Islands). Thus, they show that the mantle is heterogeneous on very large scales (ocean scales) as well as in more limited areas. Radiogenic data on ultramafic nodules also show that these mantle materials have isotopic heterogeneities at the mineral scale. These data also allow an estimation of the time scale of creation of these heterogeneities: they are of the order of some 109 years (= 1 Ga). The mechanisms involved in the chemical evolution of the mantle are discussed. Heterogeneity is created by chemical fractionation during petrogenetic processes; essentially oceanic lithosphere formation, storage into the continental crust and also recycling of sediments during lithospheric plate subduction. This heterogeneity tends to be erased in the mantle by convection and diffusion. The former efficiently mixes the mantle at large scales. The latter is very inefficient in solid mantle conditions, but can homogenize the radiogenic isotopes during partial melting. All these processes have been continuously active through geological time. A mathematical model is proposed which describes the chemical evolution of a continuously differentiating convecting mantle. The correlations, and in particular mantle isochrons, appear as artefacts without time significance. Evolution of both trace element ratios and isotopic compositions can be described simultaneously if isotopic homogenization is easier than chemical mixing. Variations of lead isotopes tend to indicate an ancient (initial) heterogeneity of the mantle which can be possibly attributed to loss of lead from mantle to core. The rate of chemical fractionation of the mantle cannot have been constant with time but was faster during the Archaean than at present.

Stochastic melting of the marble cake mantle: evidence from local study of the East Pacific Rise at 12°50′N

Abstract Isotopes (Nd, Sr and Pb) and trace elements (REE, Ba, Sr, Rb) have been measured on a set of basaltic glasses from a restricted area (40 × 10 km) at 12°50′N on the East Pacific Rise. The huge variation of incompatible element concentrations (factor 70 for Ba concentrations), and the variable degrees of correlation between element concentrations cannot be explained by usual models of melting and fractional crystallization. A rough correlation between the Ce/Yb ratio and the isotopic ratios favors a “source effect” for the genesis of the glasses. We have developed a model including both partial melting process acting on a heterogeneous mantle source with two components (peridotites and pyroxenites; “marble cake mantle” of Allegre and Turcotte) and fractional crystallization. The purpose of this model is not to obtain values of the four parameters involved (degree of melting in the peridotites, in the pyroxenites, proportion of pyroxenites involved in the melting, degree of fractional crystallization) for each analyzed glass, but to model the whole set of glasses by stochastic genesis and sampling of liquids. We have used the stochastic procedure for the four controlled parameters, currently generating 10, 000 “samples”. Our preferred model for this portion of the East Pacific Ridge is obtained with a degree of melting in the peridotites and in the pyroxenites varying uniformly from 6 to 20%, and from 6 to 50% respectively. The degree of mixing between liquids issued from the two sources varies from 0 to 100%, and the degree of fractional crystallization remains small, without noticeable effect on the concentrations, varying from 0 to 6%.

A global geochemical mass budget applied to the Congo basin rivers: Erosion rates and continental crust composition

Rivers carry the products of continental denudation either in a dissolved form (chemical erosion) or in a solid form (physical erosion). We focus in this paper on the relationship between physical erosion and chemical erosion. We establish the mass budget of the Congo Basin Rivers using chemical complementarities between river suspended sediments, sandy bedload, and dissolved load of the Congo Basin rivers reported in a previous paper (Dupre et al., 1995). A series of equations are presented, assuming that the physical and chemical erosion processes are in a steady state during one year. The total mass of river-borne material (dissolved and particulate) transported in the river over a given period of time should balance the mass of upper continental crust eroded during this time. n nWe show that the local continental crust on each drainage basin can be estimated and solve our steady-state weathering model using an inversion procedure. The very good agreement between modelled and measured values of the river-suspended sediment concentrations validates the steady-state hypothesis in this wet tropical area. Consequently, in this area, the sediment yield provide a good estimate of the rates of mechanical denudation. This result also validates the calculation of the chemical and isotopic composition of the local continental upper crust using the bulk river load. Erosion rates for the silicate upper crust and thus independent of the lithological variability (silicates, evaporites, and carbonates) of the drainage basins are calculated. Mechanical erosion rates and chemical erosion rates for the Congo Basin at Brazzaville are 8 t/km2/y and 5 t/km2/y. The corresponding consumption of atmospheric CO2 by weathering process is estimated to 51 × 103 mol/km2/an. These weathering and consumption rates are low in spite of the severity of the weathering conditions, of the high soil temperature, and of the intensity of precipitations. These conclusions indicate the limiting influence the dynamic equilibrium of soils for silicate weathering. Finally, by estimating the local continental crust chemical composition before the onset of erosion processes, especially for the most soluble elements, we can test the model of Taylor and McLennan.

Systematic use of trace elements in igneous processes

Trace element concentration data can be used in a systematic way for the study of igneous processes by means of constructing models of such processes which satisfactorily account for the observations. We propose to treat the problem as an inverse problem. The concept of trace element paths (TEP) is introduced as a representation of the solution to the direct problem. The inverse problem consists of estimating, by a resolution of the equations, the various parameters of a model so as to provide a best fit to observed TEP. A detailed account of the theory is given in the case of equilibrium fractional crystallization. The estimated parameters are then those figuring in the Rayleigh distillation law, namely, 1) the initial concentrations of trace elements in the parental magma, 2) the bulk partition coefficients of the elements, and 3) the degree of crystallization corresponding to each sample of the magmatic suite analyzed.A slightly generalized maximum likelihood method is used to solve the linearized equation by a stable, iterative algorithm. Information theory is then shown to yield an account of the distribution and flow of information during the process of solving the inverse problem. The concept of Data Importances is generalized, and its use in optimizing the study justified. The technique is successfully applied to a synthetic data set, and then illustrated on a data set from Terceira (Azores). The results are used to refine the conclusions reached in part I, and permit a more detailed discussion of the model.