SCOPE 42 - Biogeochemistry of Major World Rivers

15.1 INTRODUCTION

The previous chapters in this book record results and reflect ideas of many authors who have studied different aspects of the world river system. A number of chapters deal with regional views of the flux of dissolved and particulate carbon, mineral nutrients and solid detritus. Other chapters discuss more general topics in terms of satellite imagery, molecular nature of humic compounds, clay mineral distribution in major world rivers, and particle formation in riverine and estuarine environments. This summary chapter attempts to integrate the observations made in the earlier contributions and tries to present an overview on the interaction of biogeochemical cycles in major world rivers.

From the beginning of mankind rivers have received special attention. Cultures which depend mainly on rivers as a water resource¾Hydraulic Societies in the sense of Wittfogel (1977)¾are particularly vulnerable to regional changes of rainfall pattern if the area of catchment is small. Moreover, minor shifts of climatic zones will have dramatic consequences, if the catchment is located at the boundary of such zones. Excellent examples in this respect are the Euphrates for the Babylonians, the Changjiang and Huanghe for the Chinese and the Nile for the Egyptians. We shall select Egypt as a case study to illustrate the dependence of human societies on major rivers and climatic change.

15.2 PALEOHYDROLOGY OF THE NILE

The Nile's discharge varies on a continuous spectrum of time scales in response to regional climatic change as well as to even minor shifts in the hemisphere circulation pattern of the atmosphere (Degens and Spitzy 1983). The river is fed to almost 80% by summer rains falling in the Ethiopian mountains, a tiny catchment area just 4° of longitude times 4° of latitude. The overall situation is reminiscent of the Colorado River which receives a large part of its water from a very limited area in Arizona, 2000 m above sea level. Comparison of these two regions makes it clear to what extent relatively limited mountainous terrains, positioned in precipitation-strategic regimes, can control the water supply of distant cities such as Cairo or Los Angeles.

Water resources of Egypt are fed by two spatially and seasonally distinct phenomena; Mediterranean winter rain at the northern margin of the Sahara; and tropical summer rain in eastern Africa. Both events are associated with two zones of pronounced meteorological gradients, that is to the north, the polar front, and at equatorial latitudes the intertropical convergence zone (ITCZ).

The spatial distribution of yearly mean precipitation across Africa is rather homogeneous zonally, but shows a strong latitudinal gradient as shown in Figure 15.1. In winter, the polar front moves south and brings rain to the northern margin of the Sahara, while the ITCZ moves south leaving precipitation in the Ethiopian mountains at a minimum. During summer, the situation is reversed: the polar front retires northwards and no rain falls along the Mediterranean coast of Africa. Similarly, the ITCZ moves northwards across the Equator to the Ethiopian mountains, where its heavy rainfalls feed the tributaries of the Blue Nile. It is here that the Nile's flood is born.

From this modern analog it becomes obvious that the water supply for Egypt must have been entirely different in the past, when the polar front and the ITCZ moved around different mean latitudinal positions. How the mean positions of the polar front and ITCZ have changed over the past 20 000 years, i.e. since about the time of the Weichselian glacial maximum, is depicted in Figure 15.2. In a simplified manner, the past 20 000 years may be characterized by three distinct patterns for the general circulation of the atmosphere with the concurrent mean latitudinal positions of the polar front and the ITCZ. These broad patterns characterize: (i) the period around the glacial maximum 18 000 years ago; (ii) the Holocene warm phase between 10 000 and 4000 years BP (before present); and (iii) today's situation, prevailing for the past 3000 years. In short, the cold and dry period that is the time slot 20 000 to 12 000 years BP (Figure 15.2a), appeared to be characterized by a southward shift of the polar front as well as the tropic front, and Egypt got more rain but less Nile water. With the outgoing Glacial and the onset of the Holocene (Figure 15.2b), the tropic front moved further north, reaching the main catchment of the Nile. On the other hand, the polar front remained at a more southern position as compared to today, because the glaciers of the northern hemisphere had not yet fully retreated. The general warming that followed the last Glacial reached its climax between 6000 and 4000 years BP . Mean global temperatures were probably 1 to 1.5 °C higher than at present, and today's arid zones were significantly reduced from the Equator to the Mediterranean and a changed humidity regime was established (Figure 15.2c). During the past few thousand years the polar front moved north and left the Sahara without water.

Figure 15.1 Yearly precipitation in Africa

All these scenarios cannot be more than schematic pictures of the real events. In a series of papers (Flohn 1980; Flohn and Nicholson 1980) it was said that spring and autumn play an important role for the water cycle in the Sahara. One reason is that the tropical easterly jet stream¾driven by the Tibetan Plateau¾is only effective during summer. Associated with the wind regime in its delta is the sinking of air masses. This sinking suppresses the northward movement of rain-intensive tropical weather systems, and provides the answer to the question 'Why is the Sahara so dry?' (Flohn 1966). During spring and autumn, this mechanism is less dominant, so that tropical disturbances may easily move northward and bring rain and sand storms as far as the northern margin of the Sahara. They are more familiarly known under the term Sahara Depressions, which are triggered off by low level elements of the ITCZ, that is the easterly waves. Should conditions be favorable for an intensified triggering of Sahara Depressions, a more humid Sahara might develop without having to move the ITCZ. Such conditions were established a few thousand years ago, when rivers and lakes existed in the western Egyptian and Libyan deserts, where precipitation is now almost nil. Truly a Fata Morgana seen across the sands of time.

Figure 15.2 Relationships of the polar front and ITCZ in Africa over the past 20 000 years

In conclusion, polar and tropic fronts control the climatic events in Africa in space and time. They are critical elements of the general circulation of the atmosphere and the driving force of other climate events, the monsoons. Monsoonal influence is felt across practically all land masses. In a word, the tropical belt cooks the weather for the globe, and small changes injected into that system may alter the motion of air and sea, and with it induce climatic change and alter the water runoff of rivers. What has been said for long-term changes of riverine discharges, equally applies to short-term changes as they are manifested in the records of flood from various stations along the Nile, ever since the Nile's runoff has been monitored (Riehl and Meitin 1979; Riehl et al. 1979). The Nilometer readings on the island of Roda in Cairo, which start in AD 622, the year in which the Mohammedan calendar begins, show characteristic features of 'fat' and 'lean' periods, each one lasting a few decades. All of this suggests that rivers do not represent a steady-state system but are¾as far as runoff is concerned¾a true reflection of global climatic change. In consequence, climatic changes as envisioned for the coming decades, will find their expression in physical and chemical changes of rivers.

To illustrate how rivers respond to challenges in climatic change, we will briefly examine river runoff into the Black Sea over the past 20 000 years (Degens and Kempe 1982). Today's total suspended load in the Black Sea is in the order of 2 x 10⁸ t/year. Since that load will produce annually a blanket of 0.3 mm thick, observed changes in rates of sedimentation recorded in the sediments deposited simply reflect changes in river runoff and the surface characteristics in terms of vegetation, soil and climate. In the case of the Black Sea, we have observed 20-fold changes in sediment transport during an interval of only 20 000 years (Figure 15.3). All this suggests that we have to be rather cautious when using one-year river records as a steady-state figure and draw from them conclusions on global transport. Rivers have short memories, and their discharge may change from season to season, from year to year, and throughout the millenia in response to climatic change.

Figure 15.3 Changes in sediment transport in the Black Sea in the last 20 000 years

15.3 PREVIOUS STUDIES

The old saying by Plinius that Tales sunt aquae, qualis terra per quam fluunt (waters take their nature from the strata through which they flow) carries indeed a profound meaning. A river flowing through crystalline rocks will certainly carry a different chemical fingerprint than a river fed from a limestone source. In addition: climate, elevation, length, and type of vegetation will profoundly affect kind and amount of water, salt, organic matter and minerals eventually discharged to the sea.

At the beginning of research on major world rivers, work was executed in a matter of random approach. It was more or less left to chance what rivers happened to be looked at and what ingredients were measured. From such an inhomogeneous source of references, archives and data banks were prepared which then served as a base to assess the discharge characteristics of rivers in terms of water, volume, chemistry and mineralogy. In this context, the pioneering work of Frank Wigglesworth Clarke, entitled The Data of Geochemistry and published in 1916, should be mentioned. This book holds a special chapter on lakes and rivers. A more recent account on rivers is the highly recommended treatise by Livingstone (1963) which still is the best general compendium on riverine geochemistry. Concerning the mineralogical flux associated with rivers, an up-to-date survey has been given by Milliman and Meade (1983) from where older references can be obtained too.

With the growing awareness of human impact on the quality of rivers and streams, emphasis has shifted from the point of just monitoring for this or that parameter, towards a more holistic treatment of a river as an ecosystem, such a study is the work on the Amazon by Richey et al. (this volume, Chapter 3). In turn, research does now include the study of dissolved and particulate organic matter, mineral nutrients, heavy metals and other pollutants which indeed contribute significantly to the 'well-being' of a river. A substantial body of information has meanwhile been gathered (e .g. Meybeck 1979; Bolin et al. 1979; Kempe 1979; Schlesinger and Melack 1981).

In the present volume an attempt has been made to review in a systematic fashion and on a global scale work that has been done over the past decade on the biogeochemistry of riverine systems. This involved not only research on dissolved and particulate carbon compounds, but encompassed work on water runoff, mineralogy of the detritus, nutrients, major ions, trace elements, a number of pollutants, planktonic organisms, etc., in close to 50% of all waters running from land to sea.

15.4 INVENTORY

Global flux rates for minerals, salt and water discharge have been summarized in the form of graphs by Milliman and Meade (1983) and Degens (1989). Such figures certainly help to obtain a general overview on the quantity of material that becomes mobilized via riverine erosion and moreover where most of the action happens to be. It was for that purpose that these authors prepared a number of graphics. Here we refrain from presenting such graphs because closer inspection reveals that the measurements on which they were based are at best semi-quantitative.

This is principally so, because individual researchers often employed different sampling devices, relied on surface samples rather than depth-integrated samples, used non-standardized analytical techniques and so forth. In addition, sampling stations were not always representative of the total output of a given river, and last but not least, rivers fluctuate by a wide margin in water volume and overall biogeochemistry both seasonally and interannually. Inasmuch as critical background information is frequently lacking in the published record, intercomparison based on literature data is at best tentative. For example, the Zambezi has been quoted as discharging 225 km³/year traditionally (Ambroggi 1980; Lerman 1981). However, recent evaluation of water discharge at Kariba Dam for the sampling period 1962 to 1983 yielded an average of 46 km³/year only (Borchert and Kempe 1985). Water runoff at this station stems from about half the tributary area of the Zambezi implying that total Zambezi discharge is in the order of 80 km³/year. Or take the figures for the detrital load reported for the Rivers Nile and Indus which amount to 125 x 10⁶ t/year and 400 x 10⁶ t/year, respectively. By contrast, recent data show almost no detrital discharge for the Nile River (~2 10⁶ t/year) and for the Indus River a much reduced mass (100 x 10⁶ t/year) (Arain 1985; Kempe 1989; Subramanian and Ittekkot, this volume, Chapter 7). For the salt load, discrepancies are equally pronounced and the Rivers Elbe and Rhine are listed as examples. Namely, salts due to present or former mining activities have increased by a factor of 30 to 50 since last century, and the amount of dissolved nitrate and phosphorus, largely derived from agricultural, industrial and household activities, has been substantially raised during this interval (Kempe et al., this volume, Chapter 8). One of the major difficulties in assessing annual transport and discharge of an individual river are episodic events. Take, for example, the River Paraná. It has experienced a century flood in response to the 1982/1983 El Niño Event with skyrocketing discharge values for the year 1983 (Depetris and Paolini, this volume, Chapter 5).

We have taken data pertaining to global riverine fluxes from the preceding chapters (Table 15.1). In view of the reservations expressed by us, we hope that the numbers will not be taken as absolute figures, since they reflect only the present state of the art. Revisions will certainly have to be made once the monitoring of major world rivers has been accomplished across a much larger time frame. It will only be then that modelling of individual rivers as well as general conclusions on river systems at large become meaningful. The transport of total organic carbon is estimated to 0.33 x 10¹⁵ g/year. This is higher than the estimate of 0.19 of Kempe (1989) and close to the first estimate of 0.28 obtained from the SCOPE/UNEP data set in 1985 (Kempe et al. in Degens et al. 1985). It is, however, lower than some of the other estimates which appeared in recent publications. The value of total DIC transport of 0.41 x 10¹⁵ g C/year is very close to the value of 0.44 estimated by Kempe (1979) on the basis of the Livingstone (1963) data.

Table 15.1 Drainage area, total discharge volume and carbon fluxes of prominent world rivers: a summary

Note: Total transports were obtained using the following formula: Continental Discharge x (Total listed Transport/Total listed Discharge): Continent runoff volumes have been taken from Baumgartner and Reihel (1975). This allows to estimate the carbon transport from non-listed rivers In proportion to the average carbon transport of a certain continent from listed rivers.

15.5 DISCUSSION

Streams can be examined from outer space, by aircraft, from a bridge or a vessel, and on a river's ground. Each vantage point is unique and provides information complementary to the observations from the other points. Thus, an integrated approach is essential to define physically the riverine environment and to reveal its biogeochemical working.

River plumes can create stratified environments. Therefore, the structure of river plumes is very important for the cycling of nutrients in estuaries and resultant primary productivity. Szekielda and McGinnis (this volume, Chapter 1) could demonstrate that plankton patches monitored in different plumes of the Amazon, the Changjiang and the Rhone showed a direct relationship between patch size and its lifespan. It is evident that patch sizes in the range of kilometers may be of statistical significance for plankton development or eutrophication in general.

Substantial progress has been made in remote sensing of water substances over rivers, estuaries and coastal waters (Fischer et al., this volume, Chapter 2). For a successful retrieval of substances from multispectral radiance measurements in these types of waters, the interpretation method is important. From eigenvalues and factor analysis, applied to measured and simulated mutlispectral radiances, these authors found that suspended matter, phytoplankton and 'Gelbstoff' (humic acids-fulvic acids) are detectable under fair measuring conditions. The interpretation method used to detect and discriminate between these three substances, is an inverse modelling technique, based on a radiative transfer model for aqueous environment plus the atmosphere. Fischer et al. (this volume, Chapter 2) present aircraft and satellite data showing the usefulness of this procedure for the evaluation of a number of parameters present at or near the surface of natural water bodies. In addition to such complex remote sensing techniques, a simple method¾an aerial photograph¾can be extremely informative when it comes to the mapping of suspended matter distribution in rivers and estuaries (Fischer et al., this volume, Chapter 2).

South America (Depetris and Paolini, this volume, Chapter 5) is a continent of unusual properties: 94.4% of its 11 100 km³/year of discharge is directed to its eastern side. Also, its specific discharge (19.71/s/km²) is more than twice as high as for any other continent. This is due to the existence of a large landmass in humid tropical latitudes. Furthermore, it contains the largest river system on Earth, the Amazon, largest with regard to tributary area (6.3 x 10⁶ km²), discharge (5500 km³/year) and, almost, length (6577 km), thus second only to the Nile. In specific discharge (28.0 l/s/km²) it is surpassed worldwide only by its northern neighbor: the Orinoco (32.7l/s/km², 1130 km³/year). The Paraná/Paraguay/Uruguay system serves a vast discharge basin also, about half as large as that of the Amazon, but since it is situated in less humid zones a much smaller total (470 km³/year) and specific discharge (5.31/s/km²) is reached.

Not much had been known on the biogeochemistry of these three large South American rivers ten years ago. Since then the Amazon has been explored through the international and multidisciplinary CAMREX Project (Richey et al., this volume, Chapter 3) and Paraná and Orinoco were investigated by Argentine and Venezuelan SCOPE Projects (Depetris and Paolini, this volume, Chapter 5). Information on the Uruguay, the São Francisco and the Magdalena was also gathered.

The world's largest stream, the Amazon, has been examined in a manner unsurpassed by any other river system. Interdisciplinary research groups from many nations have studied in-depth the dynamics of this tropical river for more than a decade. Chapter 3 by Richey et al. (this volume) reviews major accomplishments and highlights research on the biogeochemical cycling of elements in main stem and tributaries of the Amazon system. Special attention is given to carbon, whose behavior in large areas of the Andean and tributary basins, the floodplains and the main stem sheds considerable light on the interplay and dynamics of biogeochemical elements operative in waters, soils and sediments. Inasmuch as collection of data was done at different time scales, that is from hours, weeks to years, and at varying space intervals, that is from kilometers to hundreds of kilometers, short- and long- term processes affecting bioactive elements throughout the drainage system could be revealed. Such a data set is pertinent to assess, for instance, human impact on rain forests with respect to the amount of carbon dioxide and methane released to the atmosphere. Inasmuch as considerable quantities of organic matter are carried to the sea (Richey et al., this volume, Chapter 3), it is from this data set that one can now realistically appraise the release potential of certain greenhouse gases from tropical soils in the aftermath of a massive clearing of rain forests by 'ax and fire'. Figures for global CO₂ discharge due to deforestation seem to be at the lower side of recent estimates, i.e. c. 1 Gt C/year. Aside from elucidating riverine biogeochemical processes operative at several different time and space scales, the Amazon Model of Richey and his associates carries signals essential for evaluating climatic change as a result of anthropogenic inputs of greenhouse gases into our atmosphere.

Generalized, the Orinoco is fed on its right bank by black water and from its left bank by white water rivers. Black water rivers are low in pH, conductivity and suspended matter but high in DOC (possibly composed largely of fulvic acids) and white water rivers are high in pH, dissolved solids, alkalinity and suspended matter but low in DOC. The highest average monthly discharge (August/September) is a factor of 6.6 higher than the lowest monthly discharge (February) and the concentrations of most ions follow a dilution model.

The Paraná proper is formed by the confluence of two very different rivers, the upper Paraná which has a low conductivity and pH but high DOC and is high in dissolved fructose, and the Paraguay which has a high conductivity and high TSS. An Andean tributary to the Paraguay, the Bermejo, has an average TSS concentration of 4500 mg/l and is therefore the main source of sediment to the river. Paraná TSS loads can fluctuate from 40 to 100 x 10⁶ t/year .

On the lower courses of both rivers extensive floodplains have developed. These are possibly the sources for the high DOC concentrations observed at high discharges. This is also suggested by relatively high concentrations of arabinose and non-protein dissolved amino acids indicative of bacterial reworking of organic matter. DOC average concentration is 4.8 mg/l in the Orinoco. In the Paraná the El Niño/Southern Oscillation event of 1982 caused a 'century flood', inundating the entire floodplain. During this event the DOC concentration increased from 6.1 mg/l to 10.2 mg/l and the export more than doubled from 2.8 to 7.5 x 10⁶ t/year. In contrast, POC exports decreased from 1.6 to 0.9 x 10⁶ t/year, suggesting that erosion is small once the floodplain is inundated and leaching becomes the dominant process of organic carbon mobilization.

TSS concentration is, in contrast to DOC, highest during the early over- bank in both rivers. Even though the POC content of TSS fans off exponentially with increasing TSS concentrations, POC concentrations (in mg/l) peak also during early overbank. Compared to DOC, POC transport is, however, second in both rivers. Total transports of DOC, POC and DIC amount to 4.5, 2.1 and 1.7 x .10⁶ t/year and to 5.9, 1.3 and 3.0 x 10⁶ t/year, respectively for the Orinoco and Paraná.

The Arctic territory belongs essentially to three nations: Soviet Union, Canada, and United States. In all three of them, more than half of the fresh water reserves are carried or contained in the Arctic rivers and lakes. Many of them drain permafrost regions and are frozen until late May. For the first time, long-term records from Siberian and North American rivers were drawn together (Telang et al., this volume, Chapter 4). The delicate biogeochemical nature in terms of mineral nutrients, low salinity, restricted biota, and high seasonality in water discharge make them readily vulnerable to deleterious effects from human activities, e.g. acid rain, hydrocarbon exploration, large dams or electric power stations. Intercomparison is made not only among the Arctic rivers but their distinctive biogeochemical features in relation to rivers from the tropical belt are elaborated on. Respiration and reworking of Quaternary peat deposits account for much of the carbon flux in these northern streams.

Large areas of Africa are without peripheral discharge. Existing rivers fall into two categories: (i) rivers of the tropical rain forest, and (ii) rivers of the tropical and subtropical savannah (Martins and Probst, this volume, Chapter 6). The Zaïre, belonging to the first group, has the second highest discharge of all world rivers (1300 km³/year) and a pronounced specific discharge (11.8 l/s/km²). Its very low ratio of 1.7 between highest and lowest monthly discharge is related to seasonal latitudinal shifts of the intertropical convergence zone (ITCZ). In contrast, 'savannah' rivers represented by the Niger, Nile, Zambezi, Orange and Senegal exhibit a much lower absolute discharge in relation to the large size of basins they serve. Specific discharge values range from 0.3 (Orange) to 4.5 l/s/km² (Zambezi); seasonal discharge ratio can be as high as 13 (Niger). The extremely low ratio for the Nile (2.6) is due to water regulation imposed by the Aswan Dam. Furthermore, intensive usage of water for irrigation in Egypt raises appreciably the Nile's salinity, possibly causing massive precipitation of CaCO₃ (Kempe 1983). Increase of salt in groundwaters and soils (Kempe 1989) appears to proceed so fast that Egypt may loose most of its agricultural area much earlier than the calculated lifespan of the Aswan Dam (> 300 years), i.e. within the next decades. A number of savannah rivers display a decadal frequency in water runoff which coincides with the sunspot cycle (Faure and Gac 1981; Borchert and Kempe 1985).

The flow of all major African rivers, except the Zaïre, is interrupted by dams with resultant effects on river dynamics. For instance, the chemistry of the Orange is nearly constant throughout the seasons (Hart 1985). In the case of the Nile, the bulk of suspended matter amounting to about 100 x 10⁶ t/year accumulates behind the Aswan Dam. Rivers of the tropical rain forest are high in dissolved organic carbon (DOC) and for the Zaïre a mean value of 8.5 mg/ml is reported. Savannah rivers are low in DOC (e.g. Orange: 2.3 mg/l). The amount of particulate organic carbon (POC) on the other hand is more related to riverine hydraulics. The Niger is highest in POC content (3.7 mg/l) due to its extended floodplains. In general, savannah floodplain rivers carry the bulk of the detritus during the early overbank, that is the concentration peak precedes the discharge peak. In contrast, the Zaïre carries highest concentrations at times of low discharge. Savannah rivers exhibit an inverse relationship between POC content and suspended matter (TSS) concentration, which nevertheless results in an increase in absolute POC concentrations (highly significant positive linear correlation). Again, the Zaïre shows a different pattern: the POC content of suspended matter seems to increase with TSS concentration; large-scale planktonic contributions are the possible reason for this pattern, while in the savannah rivers the bulk of the POC probably represents eroded topsoil.

The Asian rivers, especially those draining the Himalayas discussed by Subramanian and Ittekkot (this volume, Chapter 7), account for more than 40% of the total annual sediment discharge from the land to the sea. Most of this discharge takes place within a short period of time between July and September of each year. Three other factors make these rivers of special significance not only from an academic but also from an environmental point of view: (i) the existing steep gradient between the sources of these rivers and the river mouths; (ii) deforestation in the drainage areas; and (iii) their direct connection to the deep sea via canyons and a network of submarine channels. The first two factors ensure the transport of large quantities of carbon and minerals by these rivers to the adjacent seas. In contrast to the, other rivers, the materials transported by the Asian rivers are mostly carried to offshore areas and are deposited in the huge submarine fans extending several thousand kilometers from the river mouths.

Europe is a continent with numerous, relatively small drainage basins. Its largest two rivers, the Danube and the Wolga, do not rank among the top ten world rivers in terms of discharge, basin size, length nor specific discharge. However, if it comes to nutrient load or specific discharge of salts, then Europe becomes the leading continent (Kempe et al., this volume, Chapter 8). Due to its temperate and humid climate and due to its large percentage of carbonate rocks, it is also the continent with the fastest chemical weathering and with the highest dissolved to particulate transport ratio.

In comparison to other continents, the chemistry of Europe's rivers is known better and for a much longer span of time. However, scientific evaluation of these records is still at the beginning and large-scale comparative studies are missing. Furthermore, it is not sufficient-as is the case for the streams of South America-to study just the largest five or ten rivers in order to obtain discharge figures for more than 50% of the water runoff. Discharge data of 112 rivers, which represent a cumulated drainage area of 6.915 x 10⁶ km² or 65.9% of the continental area of Europe, sum up to 1544 km³ /year. This figure is 55.1% of Europe's theoretical continental discharge (Baumgartner and Reichel1975).

One important aspect of studying European rivers has to do with the effects anthropogenic activities impose on riverine biogeochemistry. For instance, the transport of sodium chloride in the Rhine has increased by a factor of 30 to 50 since the beginning of industrialization (Zobrist and Stumm 1981). The cumulated loads in Rhine, Weser and Elbe for total phosphorus (50 x 10³ t/year) and nitrogen (624 X 10³ t/year) match the P and N load the 50 times larger Amazon carries to the sea.

The case study of the Rhine provided insight into the internal workings of a large 'industrial' river. One of the expressions of its well-being is its carbon dioxide pressure (PCO₂) (Kempe 1982). Where photosynthesis prevails, pCO₂ tends to be low. In eutrophic lakes, such as Lake Constance, the source of the Rhine, it can be below the pressure of ambient air during the summer, making the lake a temporary sink for atmospheric CO₂. If respiration prevails, then the pCO₂ rises and the seasonality changes from the limnic to the riverine mode, i.e. it is highest when temperature is high and the concentration of organic matter is high such as is the case during low water periods in summer or spring. In the Rhine long-term means have been calculated exceeding 6000 ppmv pCO₂, i.e. pressures 20 times those of ambient air. The river is a local source of CO₂ to the atmosphere. As can be shown, the pCO₂ rises from Lake Constance downstream with the further input of labile organic matter to the Rhine. At such high pCO₂ all of the oxygen should have been consumed to fuel the ongoing respiration. This, however, is not the case. Instead nitrate seems to serve as a second oxygen source and denitrification becomes an important process in summer preventing total anaerobia in the Rhine. Investigation of the long-term chemical records of the Rhine, which date back to 1963, show that pCO₂ and BOD (biological oxygen demand) have risen until 1972. Since then, the input of labile organics to the river has decreased thanks to the construction of sewage plants. However, such a reverse in trend cannot be found in the case of the nutrient inputs, which are still rising or, at best, stable at high levels in European rivers. The consequences for the coastal seas have become evident in the last years; in 1988, a catastrophic bloom of a flagellate wiped out the fish population throughout much of the Kattegat and Skagerrak.

Riverine DOC originates from photosynthesis in the water column (autochthonous) and from the leaching of plants and soils (allochthonous) .The autochthonous fraction is predominantly aliphatic, rich in nitrogen and composed largely of low molecular weight material, that may be readily decomposed and recycled to CO₂, ammonia, nitrate, phosphate, etc. The allochthonous fraction is rich in aromatics, low in nitrogen, of high molecular weight and refractory. Whereas in undisturbed ecosystems (e.g. the Amazon or Zaïre) or in rivers with high suspension loads suppressing photosynthesis (e.g. the Changjiang or Huanghe) the allochthonous fraction comprises the bulk of the DOC, human impacts that improve the conditions for photosynthesis (such as reservoir building or nutrient loading due to domestic and agricultural effluents) will lead to an increased contribution of autochthonous DOC (Spitzy and Leenheer, this volume, Chapter 9).

On a global scale one-third of a gigatonne of DOC reaches the river mouths annually. For temperate and high latitude rivers conservative mixing with sea water takes place. In the tropics a certain degree of abiotic decomposition is indicated. Further studies are warranted in order to substantiate existence and magnitude of this potentially important process for global trace gas budgets.

If all riverine DOC would reach the open sea and given our current understanding of the global oceanic DOC pool, its oxidation kinetics and radiocarbon age, then a long-term steady state (a few thousand years) between river input and deep sea oxidation is established.

Environmentally, DOC is significant as a transport medium for metals and pesticides/insecticides, which readily yield complexes with DOC. Thus, instead of being precipitated or sedimented after sorption on to suspended particles, such complexes may be transported within the main stem of a river and reach coastal and open ocean waters. Alternatively, they may be deposited on floodplains from where they become episodically mobilized. In this context it is of note that riverine DOC serves as a primary nutrient or micronutrient for a number of organisms. In turn, deleterious impacts on organisms due to pollutants incorporated into DOC are feasible.

Ittekkot and Laane (this volume, Chapter 10) give an overall treatment of the data on particulate organic carbon (POC) in various classes of suspended matter concentrations encountered in world rivers. The major trend is a decrease in organic carbon content with logarithmically increasing concentrations of suspended matter. This results from the dilution of organic matter by mineral matter, and by a reduction in the autochthonous inputs during high turbidity conditions. The estimated annual POC transport in rivers is c. 230 x 10¹² g C. Breakdown of fluxes according to various classes of suspended matter concentrations shows that the maximum transport of POC occurs in the class of TSS concentrations between 500 and 1500 mg/l. These concentrations are usually encountered in the rivers of South Asia, especially those draining the Himalayas (Subramanian and Ittekkot, this volume, Chapter 7).

Chemical analyses of river suspended matter for individual organic constituents such as carbohydrates and amino acids have permitted estimates of the fluxes of labile (metabolizable) and residual (non-metabolizable) organic carbon. About 35% of POC transport in rivers occurs in the labile and the rest in the residual fractions. These results suggest that a certain fraction of riverine organic matter is available for heterotrophic utilization in the estuaries and coastal areas. However, compared to the estimated annual primary production in estuaries and coastal seas this fraction represents no more than 1%. The residual fraction of the riverine particulate organic matter together with the refractory organic matter derived from primary production in the estuaries and coastal seas may account for a significant constituent of organic matter being trapped in the modern marine environment.

Stable isotopes are an extremely powerful tool to understand the dynamics or biogeochemistry or river systems. It is of note, however, that isotopic species of elements, in particular of hydrogen, carbon, nitrogen, oxygen and sulphur, are of a variegated nature. For instance, hydrogen and oxygen isotopes are extremely useful to decipher the sources of water masses, because fractionation of hydrogen and oxygen isotopes is controlled by the meteorological cycle. In consequence, a water molecule carries an isotopic fingerprint as to whether it came from rain or snow, from low or high elevation, from areas distant or near to the ocean, from high or low latitudes, from this or that climatic zone, etc. Nitrogen isotope fractionation on the other hand is more seen in connection with the activity of organisms, that is nitrification, denitrification, conversion to or from ammonia, nitrate, nitrite and nitrogenous organic matter, as well as anthropogenic effects (fertilization). Sulphur isotope fractionation is of special relevance in redox reactions on floodplains and touches upon organic and inorganic mechanisms operating in fresh waters. The impact of acid rain, for example on Arctic rivers, can be traced by means of sulfur isotopes. The behavior of stable carbon isotopes in riverine systems has received special attention (Mook and Tan, this volume, Chapter 11). In aqueous environments, the element carbon can principally be partitioned into carbon dioxide, bicarbonate, carbonic acid, carbonate (e.g. CaCO₃), and organic matter. The ¹³C/¹²C ratio in any one of these molecules is determined by a great number of biogeochemical processes or physical mechanisms which makes it often difficult to relate a given isotope signal determined in a compound to a well-defined process. Fortunately, however, systematic monitoring of rivers in space and time (e.g. Maackenzie River, Amazon, European rivers) provides trends which unequivocally can be assigned to large contributions of biogenic (soil) CO₂ in case dissolved carbonates are enriched in ¹²C, while a predominance of rock weathering would be registered in a lower ¹²C content (Mook and Tan, this volume, Chapter 11).With respect to riverine organic matter, both the dissolved and particulate fractions, discrimination can be made between contributions from C₃ and C₄ plants, land vegetation versus aquatic life, and in the case of estuarine environments between marine and fresh water contributions. Mook and Tan have carefully weighted the possible C-fractionations that proceed between different carbon-containing molecules¾inorganic as well as organic¾and shown us how to utilize sensibly carbon isotope techniques for the elucidation of the path biogeochemical elements take in the fresh water ecosystem.

In a German folksong Das Wandern ist des Mülers Lust the second and third strophes deal with riverine transport. In the second verse it says: Vom Wasser haben wir's gelernt, vom Wasser, and in the third verse it goes: Die Steine ach so schwer sie sind, die Steine ach so schwer sie sind, die Steine. Sie hab'n nicht Ruh bei Tag und Nacht, sind stets auf Wanderschaft bedacht, die Steine. Along the same line goes the 'Song of the Moldau' (in: Bertold Brecht's Schweyk during World War II): Am Grunde der Moldau wandern die Steine. . .(at the Moldau's ground the rocks wander. . .). So it appears that not only water and its dissolved and particulate load is carried to the sea, but pebbles too.

According to Milliman and Meade (1983), of the approximately 15 x 10⁹ tonnes of sediment load carried annually to the ocean via rivers, roughly 85% is minerals, and the remainder is organic matter. The bulk, that is 13.5 x 10⁹ t/year, reaches the sea as suspended matter, and just 10% is transported as bottom load. Since many of the particles come temporarily to rest on banks, floodplains and river beds, it may take an episodic event to resuspend that load. As a consequence of these circumstances, discharge figures for water, salt and minerals vary from year to year, from river to river, and from region to region. As a result all published runoff data for this or that constituent are beset by many uncertainties¾locally, regionally and globally¾so that recent estimates are at best 'orientation figures'.

Irion (this volume, Chapter 12) assumes that the observed pattern in clay mineral distribution found in rivers is largely determined by the petrology of the drainage area. Only in the tropics the intense weathering of surface rocks produces newly formed clay minerals which then determine the mineralogical composition of river suspensions. Konta (1985, 1988) in contrast suggests that climate is the dominant factor which determines the clay mineral distribution in rivers.

The grain-size distribution is extremely variable but with regard to major streams, these predominantly transport particles of clay/silt- and sand-size in suspension or as bedload, respectively. The alteration of the transported minerals and rock fragments (crushing and dissolution of less stable components) is mainly restricted to the upper reaches, whereas in the lower reaches of rivers changes in mineral composition are largely a function of mixing with other rivers of different provenance.

Mineral associations encountered both in bedload and in suspension vary widely and may consist of salt minerals (in very restricted areas), carbonates, feldspars, heavy minerals, quartz and the whole sequence of clay minerals. Iron sulfides (e.g. pyrite framboids) are present in case river banks become flooded, thus resuspending reducing sediments accumulated there. Furthermore, amorphous particles of volcanic origin and biogenic debris (shell carbonates, siliceous tests, phosphatic remains) contribute to material transport.

Specific parameters that might be sensitive to climatic or source changes have been described by Irion (this volume, Chapter 12) and Konta (1985, 1988). Both authors agree that the mineral composition is hardly a result of a single factor. Konta points out that high ratios of the sum of clay minerals to the sum of the chemically less stable minerals, notably feldspars, amphiboles, calcite and dolomite, typically occur in tropical regions. Subtropical or mild humid regimes display medium to low ratios; the lowest ratios were found in rivers draining either deserts or polar regions. Since this is not only a function of a decreasing density of vegetation but also of an increasing relief gradient, other factors may leave their impact on these compositional changes.

Processes affecting particulate matter in the estuaries are discussed in detail by Eisma and Cadée (this volume, Chapter 13). The role of estuaries in terms of importers or exporters of organic matter is still being debated. It appears that depending upon the type and nature of estuaries they may either import organic matter from the sea or export it to the sea.

Particulate matter in rivers and estuaries is mostly in the form of flocs of different sizes and shapes, and consisting of both organic and inorganic particles. Conventional techniques of size analysis do allow a realistic estimation of floc sizes because they tend to cover only broken-up flocs. New techniques on floc studies show flocs with sizes ranging from c. 100 µm to several millimetres. Observations of the macroflocs in nature suggest that salt flocculation is not of much importance in floc formation except when colloidal particles are involved. In a number of estuaries a decrease in macrofloc size occurred at the, fresh water-sea water interface, and this coincides with a peak in dissolved carbohydrates. Borate present in sea water is suggested to play a role in the mobilization of carbohydrates present in the organic matter in flocs. Carbohydrate peaks may also result from the dissociation of fulvic acids that hold the particles together. These flocs have an important role in the sedimentation of fine-grained particles because of their high settling rates.

Esser and Kohlmaier (this volume, Chapter 14) pay special emphasis to the anthropogenic input of a variety of biogeochemical elements to rivers, not-ably carbon, nitrogen, sulfur and phosphorus. They identify five categories contributing to the release of such elements:

• Fossil fuel emission
• Food and feed production
• Fertilizer production
• Detergent production
• Tropical deforestation

They try to assign numbers to those various input categories. This work clearly shows that the anthropogenic impact upon rivers is principally felt by the aquatic biota. Additions of nitrates and phosphates stimulate photosynthesis, while on the other hand infiltration of pesticides and metals deteriorates environmental quality. This again makes it apparent that an isolated study of a single element will not contribute significantly to the understanding of the riverine system as well as of the global environment in general. Instead an integrated approach involving all biogeochemical elements critical to life seems to be mandatory.

15.6 CONCLUSIONS AND OUTLOOK

In Table 15.1 the best available estimates for riverine discharge are summarized. The Asian rivers are the chief contributors of dissolved and particulate matter to world oceans. The rivers of South America follow far behind. Compared to these two regions, which totally deliver about 75% of all globally eroded (dissolved plus particulate) material that reaches the sea, the rest of the world has less impact on total global riverine discharge. On the other hand, it is Europe, North America and the rivers of the Soviet Union, which contribute the bulk of agricultural, industrial and household waste to estuaries and oceans. It is here that immediate action is to be taken to restore environmental quality because to paraphrase Pindar Water is the best of all things .

What should be done next? The answer is:

1. To continue monitoring of major world rivers for another decade and for the same essential parameters in order to understand better the interactions of biogeochemical elements in river systems across a much larger time frame.
2. To make a long-term investigation of lakes, man-made dams and groundwaters for the same biogeochemical elements in order to elucidate and assess the integrated operation principles in fresh water systems.
3. To extend biogeochemical studies to estuaries and the open sea along the same lines as in fresh waters in order eventually to close and bring to light the biogeochemical cycle of elements at work in aquatic ecosystems.

REFERENCES

Ambroggi, R. P. (1980) Water. Sci. Amer. 243(3), 90-104.

Arain, R. (1985) Carbon and mineral transport of Indus River 1982-1983. 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, pp. 487-94.

Baumgartner, A. and Reichel, E. (1975) The World Water Balance. R. Oldenbourg, München, Wien, 179 pp.

Bolin, B., Degens, E. T., Kempe, S. and Ketner, P. (Eds) (1979) The Global Carbon Cycle. SCOPE Report 13, John Wiley & Sons, Chichester, New York, Brisbane, Toronto, 491 pp.

Borchert, G. and Kempe, S. (1985) A Zambezi aqueduct. 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, pp. 443-57.

Clarke, F. W. (1916) The Data of Geochemistry. Govn. Print. Off., Washington, US Geol. Surv. Bull. 616, 821 pp.

Degens, E. T. (1989) Perspectives on Biogeochemistry. Springer Verlag, Berlin, Heidelberg, New York, 423 pp.

Degens, E. T. and Kempe, S. (1982) Riverine carbon-an outlook. In: Degens, E. T . (Ed.) Transport of Carbon and Minerals in Major World Rivers, Pt. 1. Mitt. Geol.- Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 52, pp. 757-64.

Degens, E. T. and Spitzy, A. (1983) Paleohydrology of the Nile. In: Degens, E. T., Kempe, S. and Soliman, H. (Eds) Transport of Carbon and Minerals in Major World Rivers, Pt. 2. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 55, pp. 1-20.

Degens, E. T. (Ed.) (1982) Transport of Carbon and Minerals in Major World Rivers, Pt. 1. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 52, 766 pp.

Degens, E. T., Kempe, S. and Soliman, S. (Eds) (1983) Transport of Carbon and Minerals in Major World Rivers, Pt. 2. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 55, 535 pp.

Degens, E. T., Kempe, S. and Herrera, R. (Eds) (1985) Transport of Carbon and Mineals in Major World Rivers, Pt. 3. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 58, 645 pp.

Degens, E. T., Kempe, S. and Gan Weibin (Eds) (1987) Transport of Carbon and Minerals in Major World Rivers, Pt. 4. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 64, 512 pp.

Degens, E. T., Kempe, S. and Naidu, A. S. (Eds) (1988) Transport of Carbon and Minerals in Major World Rivers, Lakes and Estuaries, Pt. 5. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd.66, 422 pp.

Depetris, P. and Paolini, J. (1990) Biogeochemical aspects of South American rivers: the Paraná and the Orinoco. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 105-25.

Eisma, D. and Cadée, G. C. (1990) Particulate matter processes in estuaries. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 283-96.

Esser, G. and Kohlmaier, G. (1990) Modelling terrestrial sources of nitrogen, phosphorus, sulphur and organic carbon to rivers. In: Degens, E. T., Kempe, S. and Richey J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 297-322.

Faure, H. and Gac, J. Y. (1981) Will the Sahelian drought end in 1985? Nature 291, 475-8.

Fischer, J., Doerffer, R. and Grassl, H. (1990) Remote sensing of water substances in rivers, estuarine and coastal waters. In: Degens, E. T., Kempe, S. and Richey, J . (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 25-55.

Flohn, H. (1966) Warum ist die Sahara trocken? Z. Met. 17, 316-20.

Flohn, H. (1980) Possible climatic consequences of a man-made global warming. Bericht RR-80-30, IIASA, Laxenburg.

Flohn, H. and Nicholson, S. E. (1980) Climatic fluctuations in the arid belt of the 'old world' since the last glacial maximum: possible causes and future implications. In: Zinderen, E. M. van et al. (Eds) Paleontology of Africa and the Surrounding Islands, Balkema, Rotterdam, pp. 3-21.

Hart, R. C. (1985) Aspects of the hydrogeochemistry of the upper Orange River. 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, pp. 435-42.

Irion, G. (1990) Minerals in rivers. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 265-81.

Ittekkot, V. and Laane, R. W. P. M. (1990) Fate of riverine particulate organic matter. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 233-43.

Kempe, S. (1979) Carbon in the freshwater cycle. In: Bolin, B., Degens, E. T., Kempe, S. and Ketner, P. (Eds) The Global Carbon Cycle. SCOPE Report 13, John Wiley & Sons, Chichester, New York, Brisbane, Toronto, pp. 317-42.

Kempe, S. (1982) Long-term records of CO₂ pressure fluctuations in fresh waters. In: Degens, E. T. (Ed.) Transport of Carbon and Minerals in Major World Rivers, Pt. 1. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 52, pp. 91-332.

Kempe, S. (1983) Impact of Aswan High Dam on water chemistry of the Nile. In: Degens, E. T., Kempe, S. and Soliman, H. (Eds) Transport of Carbon and Minerals in Major World Rivers, Pt. 2. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, SCOPE/ UNEP Sonderbd. 55, pp. 401-23.

Kempe, S. (1989) Reservoir case study: Lake Nasser. In: Meybeck, M., Chapman, D. and Helmer, R. (Eds) Global Freshwater Quality, A First Assessment, Ch. 2, pp. 243-52, GEMS, WHO, UNEP, Blackwell, Oxford.

Kempe, S., Pet tine, M. and Cauwet, G. (1990) Biogeochemistry of European rivers. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 169-211.

Konta, J. (1985) Mineralogy and chemical maturity of suspended matter in major rivers sampled under the SCOPE/UNEP project. 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, pp. 569-92.

Konta, J. (1988) Minerals in rivers. In: Degens, E. T., Kempe, S. and Naidu, S. (Eds) Transport of Carbon in Major World Rivers, Lakes and Estuaries, Pt. 5. Mitt. Geol.- Paläont. Inst. Univ. Hamburg, SCOPE/UNEP Sonderbd. 66, pp. 341-65.

Lerman, A. (1981) Controls on river composition and the mass balance of river systems. In: Martin, J. M., Burton, J. D. and Eisma, D. (Eds)River Inputs to Ocean Systems. Proc. Workshop Rome, 26-30 March, 1979, UNEP, IOC, SCOR, pp. 1-4.

Livingstone, D. A. (1963) Chemical Composition of Rivers and Lakes. US Geol. Surv. Prof. Pap. 440-G, 64 pp.

Martins, O. and Probst, J .-L. (1990) Biogeochemistry of major African rivers: carbon and mineral transport. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 127-55.

Meybeck, M. (1979) Concentrations des eaux fluviales en éléments majeurs et apports en solution aux océans. Rev. de Géol. Dynamique et de Geogr. Phys,, 21(3), 215-46.

Milliman, J. D. and Meade, R. H. (1983) World wide delivery of river sediment to the oceans. J. Geol. 91, 1-21.

Mook, W. G. and Tan, F. C. (1990) Stable carbon isotopes in rivers and estuaries. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 245-64.

Richey, J., Victoria, R. L., Salati, E. and Forsberg, B. R. (1990) The biogeochemistry of a major river system: the Amazon case study. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 57-74.

Riehl, H. and Meitin, J. (1979) Discharge of the Nile River: a barometer of short-period climatic variation. Science 206, 1178- 79.

Riehl, H., EI-Bakry, M. and Meitin, J. (1979) Nile River discharge. Monthly Weather Rev. 107, 1546-53.

Schlesinger, W. H. and Melack, J. M. (1981) Transport of organic carbon in the world's rivers. Tellus 33, 172-87.

Spitzy, A. and Leenheer, J. (1990) Dissolved organic carbon in rivers. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 213-32.

Subramanian, V. and Ittekkot, V. (1990) Carbon transport by the Himalayan rivers. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 157-68.

Szekielda, K.-H. and McGinnis, D. (1990) Spatial variability of river plumes and eutrophication. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 1-24.

Telang, S. A., Pocklington, R., Naidu, A. S., Romankevich, E. A., Gitelson, I. I. and Gladyshev, M. I. (1990) Carbon and mineral transport in major North American, Russian arctic, and Siberian rivers: the St Lawrence, the Mackenzie, the Yukon, the arctic Alaskan rivers, the arctic basin rivers in the Soviet Union, and the Yenisei. In: Degens, E. T., Kempe, S. and Richey, J. (Eds) Biogeochemistry of Major World Rivers. SCOPE Report 42, John Wiley & Sons, Chichester, pp. 75-104.

Wittfogel, K. A. (1977) Die Orientalische Despotie. Ullstein Veri., Frankfurt, 625 pp.

Zobrist, J. and Stumm, W. (1981) Chemical dynamics of the Rhine catchment area in Switzerland, extrapolation to the 'pristine' Rhine river input to the ocean. In: Martin, J. M., Burton, J. D. and Eisma, D. (Eds) River Inputs to Ocean Systems, UNEP, IOC, SCOR, 52-64.