SCOPE 42 - Biogeochemistry of Major World Rivers

11.1 INTRODUCTION

The stable carbon isotopic compositions, ¹³C/¹²C or d¹³C, of total dissolved inorganic carbon (DIC), particulate organic (POC) and particulate inorganic (PIC) carbon occurring in natural waters should be closely related. This is true for autochthonous carbonate and organic matter. However, suspended matter and sediment may also originate from other sources and may be transported to the sampling location over long distances. In fact, estuarine suspended matter may find its origin in the river drainage basin as well as in the sea and be transported upstream by tidal effects.

In order to be able to distinguish between the different carbon sources we have to know the d¹³C values of the various contributors and understand the isotopic fractionation processes in the water. Therefore, it is imperative to understand the carbon isotopic composition of the total dissolved inorganic carbon (CO₂ + HCO₃ + CO₃^2-) or rather the d¹³C values of the separate compounds, because isotopic fractionation relates the d¹³C of organic matter and carbonate with that of the compound (CO₂ or HCO₃^- or CO₃^2-) from which it is derived. Knowing d¹³C of the DIC is not sufficient, because at one specific d¹³C value of DIC the d¹³C value of the CO₂ or HCO₃^- fraction depends on the relative amounts of CO₂, HCO₃^- and CO₃^2- present in the solution. These ratios are determined by pH, temperature and salinity. These values thus have to be known. Unfortunately, the literature does not always provide sufficient information to unravel the relations.

In this chapter we shall present a brief review of carbon isotopes in nature as far as it is relevant here to river studies. Figure 11.1 is a schematic diagram for the major carbon pathways and the resulting isotope fractionation effects.

Figure 11.1 Schematic diagram showing the sources of dissolved inorganic carbon (DIC), suspended particulate organic carbon (POC) and particulate inorganic carbon (PIC) in rivers and estuaries. Full lines with arrows indicate chemical and isotopic formation processes, dashed lines refer merely to transport and mixing. The numbers are representative average d¹³C values (in per thousand), of which the realistic ranges and deviations are discussed in the text. The inorganic carbon isotope fractionation values at moderate temperatures are taken from Mook et al. (1974) and Mook (1986)

The carbon isotope results are reported with respect to the VPDB standard (Vienna PDB, defined by d¹³C  of NBS19 = + 1.95 per thousand).

11.1.1 ATMOSPHERIC CO₂

In 1980 the average d¹³C of atmospheric CO₂ over the oceans was -7.5 per thousand, varying with latitude and season (Mook et al. 1983; Keeling et al. 1984). Representative equations describing the d¹³C  variation with time for the southern hemisphere and northern hemisphere at mid-latitude are respectively:

and

In continental air d¹³C averages around -8 to -9 per thousand but varies largely as a function of factors such as the uptake and release of biospheric CO₂ and the combustion of fossil fuel. These short-term variations in d¹³C are closely related to the CO₂ concentration by (valid for 1980 at northern mid-latitude):

              
where -7.6 per thousand and 338 ppm are the d¹³C and CO₂ concentration, respectively, and -26 per thousand refers to the biospheric component. A fossil-fuel-derived component may be as negative as -29 per thousand. The variation amounts to approximately -0.06 per thousand per ppm of CO₂ increase.

11.1.2 LAND PLANTS

During photosynthesis carbon becomes depleted in ¹³C. Plants using the Calvin or C₃ photosynthesis pathway have a mean d¹³C of about -26 per thousand (Park and Epstein 1961). Grasses and other plants using the Hatch-Slack or C₄ pathway have a mean d¹³C of about -14 per thousand. A third group of plants using the CAM (crassulacean acid metabolic) pathway have d¹³C values which range from -9 to -29 per thousand with no well-defined peak in the distribution of values (review by Deines 1980).

Although the C₃ type plants are the most common in species as well as in quantity of biomass (by far), effects of C₄ plants can be observed in coastal waters. The fact is that tidal flats can be protected by certain types of grasses which have C₄ d¹³C values. For instance, along the Dutch North Sea coast often d¹³C values of sedimentary organic matter are found as high as -15 per thousand or even -12 per thousand, especially in the coarse fraction of the sediment (Laane et al. 1990).

11.2 DISSOLVED INORGANIC CARBON (DIC)

11.2.1 SEA WATER

In sea water d¹³C values of total dissolved inorganic carbon (DIC = CO₂ + H₂CO₃ + HCO₃^- + CO₃^2-) vary regionally, locally, and with time and depth. The d¹³C range is from 0 to +2 per thousand. In general, deep waters below the mixed layer have d¹³C values of +0.5 ±0.5 per thousand, while in surface water they tend to be higher, generally correlated with the occurrence of primary production which removes isotopically light carbon from the water.

11.2.2 RIVERS AND GROUNDWATER

The carbon isotopic composition of DIC in fresh water is controlled by the carbon sources and sinks and results from isotope fractionation both between solid, dissolved and gaseous phases and between oxidation states (Degens 1969). The major sources of carbon contributing to DIC in natural waters are CO₂ derived from the decay of organic matter in continental soils and from the dissolution of carbonate, while in general the contribution of atmospheric CO₂ is negligibly small. Processes involved in the removal of DIC include photosynthesis, carbonate precipitation and air-water exchange.

In most rivers the water is closely linked to the groundwater in the drainage basin. The pattern of the d¹³C of DIC related to the dissolved carbon chemistry has been discussed by Mook (1980). The normal type of groundwater and river water is the calcium bicarbonate type of which d¹³C is determined by the reaction:

From soil limestone of marine origin and biogenic soil CO₂ d¹³C of the dissolved carbonate is expected to be about -12 per thousand. Additional biogenic CO₂ at pH values below 7.5 may reduce the d¹³C of DIC to more negative values ( -14 ± 2 per thousand). At very low limestone content of the soils in the drainage basin and the presence of a large amount of vegetation, the d¹³C value may become as negative as -26 per thousand at low pH. In the absence of vegetation, atmospheric CO₂ may contribute to DIC by rock weathering, resulting in d¹³C values as high as -7 per thousand at low CO₂ concentrations.

Hitchon and Krouse (1972) determined the d¹³C of DIC in 68 river waters of the Mackenzie River drainage basin. A histogram of variations in the d¹³C of DIC is shown in Figure 11.2. Dissolution of soil carbonate minerals by soil organic CO₂ derived from land vegetation can explain most of the d¹³C values observed. This means that, regarding the DIC component of the water, the river contains predominantly groundwater with d¹³C values of the HCO₃^- fraction of -11 ± 1 per thousand with varying amounts of additional biogenic CO₂. The low d¹³C values of DIC should therefore combine with lower pH values. The slightly higher observed values of -9 ± 1 per thousand at the high d¹³C end can be caused by the occurrence of rock weathering (carbonate d¹³C= -2 ± 2 per thousand, Figure 11.2) and by atmospheric CO₂ of -8 ± 1 per thousand.

Figure 11.2 Histograms of d¹³C values of DIC in rivers from North and South America and the Netherlands

Longinelli and Edmond (1983) determined the carbon isotopic composition of DIC in the Amazon basin. Extremely negative d¹³C values of -28 per thousand in the mainstem may be attributed to a relatively large contribution of biogenic (soil) CO₂ with low d¹³C, while values around -11 per thousand to -14 per thousand are normal for groundwater rivers (Figure 11.2).

A series of large and small rivers in northwestern Europe appear to have d¹³C values in a narrow range between -10 and -12 per thousand (Figures 11.2 and 11.3), due to the circumstances that climate is temperate while the drainage basins are the sedimentary type with modest vegetation (Mook 1970). The Dutch rivers also showed a seasonal variability in d¹³C. The higher d¹³C values of the HCO₃^- fraction in summer are attributed to photosynthetic activity and isotopic exchange with atmospheric CO₂. In the fall, at lower temperatures and heavy rainfall groundwater discharge causes d¹³C to decrease to normal values of around -12 per thousand.   

Figure 11.3 Seasonal variation of d¹³C of the dissolved bicarbonate fraction of large (Rhine, Meuse, Ijssel) and small rivers (Vecht, Drentse Aa) in the Netherlands

11.2.3 ESTUARIES

In estuaries the d¹³C of DIC is determined by the mixing ratio of fresh and sea water. In cases where the estuary has one predominant river input, DIC as well as ¹³C behave conservatively, resulting in a linear relation for DIC with chlorinity and a curved relation for d¹³C. The latter occurs in general where the DIC concentrations of the river and sea water differ. An example has been reported for the western Scheldt Estuary in the Netherlands (Mook 1970) (Figure 11.4). Conservative mixing requires the d¹³C of DIC to obey the ¹³C mass balance relation:

where Cl, Cl_f ( 0 per thousand) and Cl_m ( 19.4 per thousand) refer to the chlorinities of the actual sample, and the fresh water and sea water components, respectively, while C_T is equal to the DIC concentration and the subscripts f and m refer to the river water and sea water, respectively.

11.2.4 LAKES

A discussion of lakes is beyond the scope of this review. However, some basic processes may occur in estuarine or river floodplain lakes as well. For shallow lakes the original d¹³C of DIC in the river water or groundwater source may change to less negative values in time and during prolonged contact of surface water with the atmosphere. This is due to primary production removing isotopically light carbon. This effect depends on the ratio between primary productivity and the mineralization of the organic matter because only this results in a net removal of carbon. In general a continuous removal of carbon combined with isotope fractionation is described by the Rayleigh equation:

where the zero subscripts refer to the original conditions and e ( = a -1) is the isotope fractionation during carbon removal. For instance, in the case of primary production with a fractionation e = -23 per thousand, the transition of 10% of the DIC into organic matter would change d¹³C of the DIC by +2.4 per thousand.

A second cause of d¹³C changes is chemical and isotopic exchange between DIC and atmospheric CO₂, in combination with increasing pH.

Figure 11.4 d¹³C of DlC obtained during several cruises in the western Scheldt Estuary (Netherlands). The dashed line is calculated according to Equation (11.4), using d¹³C (DIC) values of the North Sea and River Scheldt as indicated. d¹⁸O of the water represents the mixing ratio of fresh water and sea water, rather than the chlorinity

Thisprocess is quantitatively determined by the residence time of the water in the lake (Broecker and Walton 1959; Mook 1970). The highest d¹³C values ( +1 ± 1 per thousand) are a result of isotopic equilibrium between the DIC fractions and atmospheric CO₂ at a pH value of 8.5 to 9.5.

11.3 PARTICULATE INORGANIC CARBON (PIC)

The d¹³C values of carbonates ( = particulate inorganic carbon) are controlled by several factors: (1) the d¹³C of the DIC; (2) the equilibrium fractionation factor between carbonate, and for instance, the HCO₃^- fraction and the variation in this factor with temperature (Mook 1986); (3) biological or kinetic effects possibly leading to non-equilibrium fractionation during carbonate formation; (4) possible diagenetic processes altering the original d¹³C value of the carbonate.

The scope of this review allows us to present only briefly the ranges of d¹³C values, sufficient to understand the observed variations in rivers and estuaries. Figure 11.1 shows these ranges for primary marine carbonates (d¹³C = -1 to +2 per thousand) and various fresh water carbonates. In the latter the d¹³C of PIC is largely determined by the carbon isotopic composition of DIC. Again we have to stress that fractionation factors apply to single compounds, for instance the reaction CO₃^2- ® CaCO₃ or HCO₃^- ® CaCO₃, rather than to the isotopic difference between carbonate and DIC as a whole.

Figure 11.1 also contains d¹³C data on carbonate rock, because weathering brings carbonates into the river as well. In fact, in most cases a considerable part of the PIC content of river sediments and suspended matter appears to originate from rock weathering. An example is shown for samples collected from the Dutch rivers Rhine and Meuse. It appears that, in general, suspended matter can better be characterized if two parameters are considered, rather than one. 

Figure 11.5 Relation between d¹³C of organic carbon (POC) and the carbonate fraction (PIC) of suspended matter in two rivers. The coarse and fine fractions are distinguished as > 63 µm and < 63 µm

Figure 11.6 Relation between d¹⁸O of water (indicating fresh water and sea water mixing) and d¹³C of dissolved HCO₃^- in the western Scheldt Estuary (full line) and the former Wadden Sea-Zuiderzee Estuary (finely dashed line because not actually observed). The d¹³C and d¹⁸O values of carbonate shell collected in situ appear to represent isotopic equilibrium with H₂O (for ¹⁸O) and HCO₃^- (for ¹³C). The double line refers to the isotopic composition of purely marine carbonate which is temperature dependent (Mook 1971)

Figure 11.5 shows the relation between d¹³C values of the carbonate (PIC) and the organic fraction (POC). The PIC data for the Rhine are probably more influenced by secondary soil carbonates. d¹³Cvalues from estuarine shell samples which doubtless are the best representations of autochthonous carbonates are shown in Figure 11.6 (Mook 1971). The close correlation with the d¹³C values of the HCO₃^- fraction representing the DIC is obvious.

11.4 PARTICULATE ORGANIC CARBON (POC)

Information on the sources of organic carbon is essential for our understanding of carbon cycles in fresh water systems. Organic carbon in rivers is derived from two major sources: (1) autochthonous organic matter such as river phytoplankton and (2) allochthonous (terrestrial) material. Because of the isotopic difference between fresh water plankton and terrestrial organic carbon, it is worthwhile attempting to use stable carbon isotopes to ascertain the relative proportion of allochthonous versus autochthonous organic carbon in river systems. Although carbon isotopes have been used extensively to study the sources of organic carbon in nearshore environments (Tan and Strain 1979, 1983; Sackett and Thompson 1963), few attempts have been made to use this method to study the organic carbon sources in rivers. Again, as with the carbonates, a close correlation is to be expected between the carbon isotopic compositions of DIC and autochthonous POC.

11.4.1 MARINE PLANKTON

The carbon isotopic composition of marine plankton ranges between -20 and -30 per thousand (Sackett et al. 1965), and depends on the isotope fractionation between the phytoplankton and the various fractions of DIC. The magnitude of this fractionation is related to temperature, species and growth rates (Sackett et al. 1965; Wong and Sackett 1978; Fontugne and Duplessy 1981). d¹³C values of marine phytoplankton from tropical and temperate environments range around -20 per thousand while those from Antarctic regions are about -27 per thousand. The isotopic composition of near surface particulate organic carbon is generally similar to that of plankton. Deep water POC, however, is about 2 to 3 per thousand and more negative than surface material. This is attributed to the selective removal of ¹³C-enriched biochemical components from the surface material by biological processes, or to significant contribution of an isotopically light Antarctic component (Eadie and Jeffrey 1973; Eadie et al. 1978).

As we mentioned before, d¹³C values as high as -11 per thousand can be found in coastal sediments, originating from relatively coarse detritus from grasses with C₄ metabolism grown on tidal flats.

11.4.2 FRESH WATER PLANKTON

In principle, similar fractionations control the d¹³C of POC produced in fresh water. Since the d¹³C of fresh-water-dissolved CO₂ (d¹³C -20 to - 26 per thousand) is 12 to 18 per thousand more negative than marine-dissolved CO₂ (d¹³C -8 per thousand), fresh water plankton shows d¹³C values of -32 (high HCO₃^- water at moderate temperatures) to -44 per thousand (high CO₂ water at low temperatures). For example, plankton from a subalpine fresh water lake has been reported as -44 to -47 per thousand (Rau 1978). Mook (1970) found plankton d¹³C of -31 per thousand in a Dutch lake combined with a d¹³Cof dissolved HCO₃^- of -8 per thousand where in a later stage of the water circulation d¹³C of plankton was -23 per thousand at d¹³C (HCO₃^-) = 0 per thousand. This results in an isotope fractionation between dissolved HCO₃^- and plankton of -23 per thousand at moderate temperature (equivalent to a fractionation between dissolved CO₂ and plankton of -13 per thousand) equal to that observed under estuarine and marine conditions (Tan and Strain 1983).

Some examples will serve to illustrate observations on d¹³C in riverine POC. As will be shown part of the collected material appears to be of autochthonous (i.e. produced in the river) organic matter, part originates from, for instance, erosion of continental peat with a d¹³C of -27 ± 1 per thousand.

11.4.3 RIVERS

The sources of particulate organic carbon in the Amazon River and its tributaries were studied by Cai et al. (1988). Figures 11.7 and 11.8 show the locations and d¹³C values for the middle and lower reaches of the Amazon main channel and its tributaries. The d¹³C values are observed for the middle and lower reaches of the Amazon main channel, showing a range of -27.9 to -30.1 per thousand, averaging at -28.9 per thousand. The d¹³C values of POC for the tributaries are generally more negative than those from the main channel. The carbon isotopic composition of terrestrial plant organic carbon in the Amazon drainage basin has been determined by Medina and Minchin (1980) and Hedges et al. (1986). The first authors reported a range of d¹³C values of leaves collected at different levels in two forest communities located in the Negro River basin of -28.7 to -35.2 per thousand. A composite sample of leaf litter in the podsol forest has a d¹³C value of -29.4 per thousand and is similar to leaves in the upper canopy. The litter value is identical to the POC measured in the Negro River (-29.3 per thousand) and similar to the mean d¹³C value observed for the Amazon main channel. Hedges et al. (1986) reported a range of d¹³C values for 15 common species of individual tree leaves as -28.7 to -31.3 per thousand.

The second possible source of organic carbon is the in situ production of aquatic plankton which occurs principally in floodplain lakes (varzea) and in the mouth bays of the tributaries (Rai and Hill 1984). Using the isotopic composition of DIC reported by Longinelli and Edmond (1983) and the isotopic fractionation between riverine plankton and dissolved HCO₃^-, we expect a range of -33 to -39 per thousand for Amazon fluvial plankton. The dominant source of POC in the Amazon River system is therefore terrestrial, with the highest d¹³Cvalues for the POC in the upper reaches of the main channel due to peat and soil erosion with possibly some contribution of isotopically heavy C₄ grasses (Junk 1973) and lower values downstream due to a contribution of in situ plankton.

Figure 11.7 Sample locations and d¹³C values of suspended POC in the Peruvian Amazon   

From the St Lawrence River at Quebec City (Canada) d¹³Cshows a seasonal variation (Tan 1987) (Figure 11.9), with the relatively negative values around -27 per thousand occurring in April-June and less negative values of -24 per thousand in August-October. This variation was repeatedly observed during three consecutive sampling years. The most negative d¹³Cvalues are associated with the high spring discharge, while the least negative values are observed during the low fall discharge. This effect is to be attributed to the more dominant contribution of terrestrial organic carbon such as, for instance, peat with d¹³C= -27± 1 per thousand during periods of high discharge, while the d¹³Cvalues of -24 per thousand are more likely to originate from other terrestrial plant material.

Figure 11.8 Sample location and d¹³C values of suspended POC in the Brazilian Amazon

Figure 11.9 Monthly variations in d¹³C of  POC in the St Lawrence River during 1981 to 1984

Suspended matter samples, collected on several days from the River Rhine at the Dutch/German border, show values of -26.8 per thousand for the coarse and -27.0 per thousand for the fine fraction. Similarly the River Meuse at the Dutch-Belgian border gave -28.1 and -28.7 per thousand for coarse and fine fractions, respectively. Peat erosion is certainly the cause of d¹³Cvalues of -29 per thousand in the Ems Estuary, northwest Germany (Salomons and Mook 1977; Eisma et al. 1985).

11.4.4 ESTUARIES

 Organic matter produced in estuaries will behave similarly to POC produced in marine and fresh water, with an isotope fractionation of about -23 per thousand relative to the dissolved bicarbonate fraction. However, because the contribution of autochthonous material is generally small, we will focus our attention briefly on the application of carbon isotopes for determining the ratio between marine POC ( d¹³C= -21 ± 1 per thousand) and terrestrial POC (d¹³C = -27 ± 1 per thousand) (Fry and Sherr 1984).

The isotopic study of Eisma et al. (1985) in the Gironde Estuary (France) illustrates the different sources of organic matter (Figure 11.10). The d¹³C clearly shows the mixing of fluviatile POC with marine POC, indicating the upstream transport of marine organic matter and the associated marine sediment and suspended matter. Under these conditions the relative contribution of fluviatile and marine POC to a sediment sample can be determined with an accuracy in the order of 10%. Figure 11.11 shows the d¹³C values of POC for the middle and lower reaches of the Amazon River and Estuary ( Cai et al. 1988) ; the d¹³C changes from the main channel (-29.3 per thousand) to the marine planktonic value (-18.0 per thousand).

Figure 11.10 d¹³Cvariation in POC along the Gironde Estuary, France. Station numbers proceed from riverine to marine conditions. Pyrolysis mass spectrometry (PyMS) reveals the chemical constituents of the organic matter, where high numbers indicate fluvial material, low numbers marine organic matter (Eisma et al. 1985)

Figure 11.11 Plot of d¹³Cof POC as a function of distance from the river mouth for the Brazilian Amazon Estuary

11.5 CONCLUDING REMARKS

In order to understand the d¹³C values of POC in rivers and estuaries, it is imperative to know the d¹³C of DIC in the water. Furthermore, it is necessary to determine the d¹³C of single compounds ( dissolved CO₂, HCO₃^- or CO₃^2-), because isotope fractionation operates during the transition between two single compounds. For instance, carbon isotope fractionation between organic primary production and dissolved CO₂ or HCO₃^- has a specific-only temperature dependent-value, whereas fractionation between POC and DIC ( = CO₂ + HCO₃- + CO₃^2-) depends on the ratio between the DIC fractions which in turn depend on pH and temperature.

River water DIC generally originates from groundwater with d¹³C values of the HCO₃^- fraction of -12 ± 1 per thousand. Sea water DIC has d¹³C values of + 1 ± l per thousand. Consequently in estuaries intermediate d¹³C values are observed depending on the mixing ratio of river and sea water , generally represented by the chlorinity. Particulate organic carbon derived from primary production is subjected to carbon isotope fractionation of about -20 to -25 per thousand at high and low temperatures, respectively, relative to dissolved bicarbonate (average -23 per thousand). The resulting d¹³C values of fresh water and marine POC are thus -35 ± 3 per thousand and - 22 ± 2 per thousand, respectively, at moderate temperatures. Conversely, rivers often contain terrestrial carbon, derived from eroded soils and land plants. In many cases this is even the dominant source of fluviatile organic matter.

It should be emphasized that in general only elaborate research allows a full understanding of the behaviour of POC in sediment and suspended matter. The attention has to be focused on:

1. d¹³C of DIC and the temperature, pH and chlorinity of the water.
2. d¹³C of fluviatile POC, depending on season.
3. d¹³C of fluviatile PIC (the carbonate fraction).
4. Grain size differentiation.

 Other techniques additionally characterize the following:

5. The carbonate by X-ray analysis (calcite/dolomite differentiation).
6. The organic matter by pyrolysis (gas chromatography), mass spectrometry (PyGMS or PyMS) or other techniques applied as carbon source  indicators.
7. Further useful tracers may be found in stable isotopes of nitrogen ¹⁵N/¹⁴N), sulphur (³⁴S/³²S) and hydrogen (²H/¹H).
8. A last extremely valuable contribution would be the analysis of ¹⁴C by applying accelerator mass spectrometry techniques to the generally small samples. This method can nicely differentiate between relatively old POC from peat erosion and river primary production. In this case it is also necessary to determine ¹⁴C in the DIC.

REFERENCES

Broecker, W. S. and Walton, A. (1959) The geochemistry of ¹⁴C in fresh water systems. Geochim. Cosmochim. Acta 16, 15-38.

Cai, D. L. , Tan, F. C. and Edmond, J. M. (1988) Sources and transport of particulate organic carbon in the Amazon River and Estuary. Estuar. Coast. Shelf Sci. 26, 1-14.

Degens, E. T. (1969) Biogeochemistry of stable carbon isotopes. In: Eglinton, G. and Murphy, M. T. J. (Eds) Organic Geochemistry, Methods and Results, Springer Verlag, Heidelberg, pp. 304-29.

Deines, P. (1980) The isotopic composition of reduced organic carbon. In: Fritz, P. and Fontes, J. C. (Eds) Handbook of Environmental Isotope Geochemistry, Elsevier, Amsterdam, Oxford, New York, PP. 329-406.

Eadie, B. J. and Jeffrey, L. M. (1973) d¹³C analyses of oceanic particulate matter. Mar. Chem., 1, 199-209.

Eadie, B. J., Jeffrey, L. M. and Sackett, W. M. (1978) Some observations on the stable carbon isotope composition of dissolved and particulate organic carbon in the marine environment. Geochim. Cosmochim. Acta 42, 1265-9.

Eisma, D., Bernard, P., Boon, J. J., van Grieken, R., Kalf, J. and Mook, W. G. (1985) Loss of particulate organic matter in estuaries as exemplified by the Ems and Gironde Estuaries. In: Degens, E. T., Kempe, S. and Herrera, R. (Eds) Transport of Carbon and Minerals in Major World Rivers, Pt. 3. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 58, 397-412.

Fontugne, M. R. and Duplessy, J. C. (1981) Organic carbon isotopes fractionation by marine plankton in the temperate range -1 to 31°C. Oceanologica Acta 4, 85-90.

Fry, B. and Sherr, E. B. (1984) ¹³C measurements as indicators of carbon flow in marine and fresh-water ecosystems. Contrib. Mar. Sci. 27, 13-47.

Hedges, J. I., Clark, W. A., Quay, P. D., Richey, J. E., Devol, A. H. and Santos, U. de M. (1986) Compositions and fluxes of particulate organic material in the Amazon River. Limnol. Oceanogr. 31(4),717-38.

Hitchon, B. and Krouse, H. R. (1972) Hydrogeochemistry of the surface waters of the Mackenzie River drainage basin, Canada-III, stable isotopes of oxygen, carbon and sulphur. Geochim. Cosmochim. Acta 36, 1337-57.

Junk, W. J. (1973) Investigations on the ecology and production biology of the floating meadows (Paspalo-Echinochloetum) on the middle Amazon, Part 2. The aquatic fauna in the root zone of floating vegetation. Amazôniana 4, 9-102.

Keeling, C. D., Carter, A. F. and Mook, W. G. (1984) Seasonal, latitudinal and secular variations in the abundance and isotopic ratios of atmospheric carbon dioxide. 2. Results from oceanographic cruises in the tropical Pacific Ocean. J. Geophys. Res. 89, 4615-28.

Laane, R. W. P. M., Turkstra, E. and Mook, W. G. (1990) Stable carbon isotope composition of pelagic and benthic organic matter in the North Sea and adjacent estuaries. In: Ittekkot, V., Kempe, S., Michaelis, W. and Spitzy, A. (Eds) Facets of Modern Biogeochemistry, Festschrift for the 60th anniversary of E. T. Degens, Springer Verlag, Berlin, Heidelberg, pp. 214-24.

Longinelli, A. and Edmond, J. M. (1983) Isotope geochemistry of the Amazon basin: a reconnaissance. J. Geophys. Res. 88, 3703-17.

Medina, E. and Minchin, P. (1980) Stratification of d¹³C values of leaves in Amazonian rain forests. Oecologia 45, 377-8.

Mook, W. G. (1970) Stable carbon and oxygen isotopes of natural waters in the Netherlands. In: Proceedings  IAEA Conference on Isotopes in Hydrology, Vienna, pp.163-90.

Mook, W. G. (1971) Palaeotemperatures and chlorinities from stable carbon and oxygen isotopes in shell carbonate. Palaeogeogr. Palaeoclim. Palaeooecol. 9, 245- 63.

Mook, W. G. (1980) Carbon-14 in hydrogeological studies. In: Fritz, P. and Fontes, J. Ch. (Eds) Handbook of Environmental Isotope Geochemistry, Elsevier, pp. 49-74.

Mook, W. G. (1986) ¹³C in atmospheric CO₂. Neth. J. Sea Res. 20, 211-23.

Mook, W. G., Bommerson, J. C. and Staverman, W. H. (1974) Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide. Earth Planet. Sci. Lett. 22, 169-76.

Mook, W. G., Koopmans, M., Carter, A. F. and Keeling, C. D. (1983) Seasonal, latitudinal and secular variations in the abundance and isotopic ratios of atmospheric carbon dioxide. J. Geophys. Res. 88, 10,915-10,933.

Park, R. and Epstein, S. (1961) Metabolic fractionation of ¹³C and ¹²C in plants. Plant Physiol. 36, 133-8.

Rai, H. and Hill, G. (1984) Primary production in the Amazonian aquatic ecosystem. In: Sioli, H. (Ed.) The Amazon-Limnology and Landscape Ecology of a Mighty Tropical River and its Basin, Junk, The Hague, pp. 312-35.

Rau, G. (1978) Carbon 13 depletion in a subalpine lake: carbon flow implications. Science, 20, 901-2.

Sackett, W. M. and Thompson, R. R. (1963) Isotopical organic carbon composition of recent continental derived clastic sediments of the eastern Gulf of Mexico. Amer. Ass. Petrol. Geol. Bull. 47, 525-38.

Sackett, W. M., Eckelman, W. R., Bender, M. L. and Be, A. W. H. (1965) Temperature dependence of carbon isotope composition in marine plankton and sediments. Science 148, 235- 7.

Salomons, W. and Mook, W. G. (1977) Trace metal concentrations in estuarine sediments: mobilization, mixing or precipitation. Neth. J. Sea Res. 11, 199-209.

Tan, F. C. and Strain, P. M. (1979) Organic carbon isotope ratios in recent sediments in the St Lawrence Estuary and the Gulf of St Lawence. Estuar. Coast. Mar. Sci. 8, 213-25.

Tan, F. C. and Strain, P. M. (1983) Sources, sinks and distribution of organic carbon in the St Lawrence Estuary, Canada. Geochim. Cosmochim. Acta 47, 125-32.

Tan, F. C. (1987) Discharge and carbon isotope composition of particulate organic carbon from the St Lawrence River, Canada. In: Degens, E. T., Kempe, S. and Gan Weibin (Eds) Transport of Carbon and Minerals in Major World Rivers. Mitt. Geol-Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 64, 301-10.

Wong, W. W. and Sackett, W. M. (1978) Fractionation of stable carbon isotopes by marine phytoplankton. Geochim. Cosmochim. Acta 42, 1809-15.