Elsevier

Geochimica et Cosmochimica Acta

Volume 69, Issue 4, 15 February 2005, Pages 839-850
Geochimica et Cosmochimica Acta

Geochemical processes in the Onyx River, Wright Valley, Antarctica: Major ions, nutrients, trace metals

https://doi.org/10.1016/j.gca.2004.08.001Get rights and content

Abstract

We present data on major ions, nutrients and trace metals in an Antarctic stream. The Onyx River is located in Wright Valley (77–32 S; 161–34 E), one of a group of ancient river and glacier-carved landforms that comprise the McMurdo Dry Valleys of Antarctica. The river is more than 30 km long and is the largest of the glacial meltwater streams that characterize this relatively ice-free region near the Ross Sea. The complete absence of rainfall in the region and the usually small contributions of glacially derived tributaries to the main channel make this a comparatively simple system for geochemical investigation. Moreover, the lack of human impacts, past or present, provides an increasingly rare window onto a pristine aquatic system.

For all major ions and silica, we observe increasing concentrations with distance from Lake Brownworth down to the recording weir near Lake Vanda. Chemical weathering rates are unexpectedly high and may be related to the rapid dissolution of ancient carbonate deposits and to the severe physical weathering associated with the harsh Antarctic winter. Of the nutrients, nitrate and dissolved reactive phosphate appear to have quite different sources. Nitrate is enriched in waters near the Lower Wright Glacier and may ultimately be derived from stratospheric sources; while phosphate is likely to be the product of chemical weathering of valley rocks and soils. We confirm the work of earlier investigations regarding the importance of the Boulder Pavement as a nutrient sink.

Dissolved Mn, Fe, Ni, Cu, and Cd are present at nanomolar levels and, in all cases, the concentrations of these metals are lower than in average world river water. We hypothesize that metal uptake and exchange with particulate phases along the course of the river may serve as a buffer for the dissolved load. Concurrent study of these three solute classes points out significant differences in the mechanisms and sites of their removal from the Onyx River.

Introduction

Insight into chemical weathering processes, including the comparative behavior of major solute classes along the length of river channels, can often be obtained in those rare natural settings which offer a somewhat restricted list of physical, chemical and biologic variables. Even the rates of chemical weathering and, consequently, of carbon dioxide drawdown, can be more easily, though roughly, estimated in those streams which are of smaller scale and which are not significantly complicated by rainfall or by the usual extraneous inputs, whether natural or human. The present study of the Onyx River was undertaken with a view toward capitalizing on these possible advantages in scale and complexity to compare, through simultaneous measurement, the behavior of major ions, nutrients and trace metals in a remote and relatively simple stream and to examine chemical weathering rates and carbon dioxide uptake in a cold-climate system. We note that this is the first study to report on dissolved and particulate trace metals over the entire length of an Antarctic river and the first to establish preliminary weathering rates for the Onyx River itself. Although the exotic setting of this river, and, indeed, other streams of the McMurdo Dry Valleys, expands the range of flowing water environments generally considered by ecologists, the evolving principles of solute acquisition and transformation apply.

A small number of ice-free desert oases occur along the coastline of an otherwise ice-burdened Antarctic Continent. The largest of these, the McMurdo Dry Valleys, cuts inland from the Ross Sea—in an east-west direction—toward the vast ice cap of the Polar Plateau. Extreme polar deserts, where precipitation falls only as dry snow, the valleys are characterized by a mean annual temperature of −20°C and by fierce desiccating winds that act as agents of extreme physical weathering and material transport. The lithology of the area has been described in some detail by Vocke and Hanson (1981), and Claridge and Campbell (1977) and Keys and Williams (1981) have discussed soil development in the region. Briefly, basement rocks consist of pre-Ordovician schists, hornfels and marbles; a younger set of lower Paleozoic granites and granite gneiss intrusions; overlying Beacon sandstones intruded by Ferrer Dolerites; and thin Quaternary dimictics and McMurdo volcanics. Soils are low in organic content but contain salts of either marine origin or derived from chemical weathering. Jones and Faure (1978), Green et al. (1988) and Lyons et al. (1998) have examined weathering and transport of ions by streams and have discussed the evolution of closed-basin lakes in Wright and Taylor Valleys.

The Onyx River is the longest of the many glacial meltwater streams that flow through the McMurdo Dry Valleys during the warmer (six to eight) weeks of the austral summer. Over 30 km in length, the Onyx flows westward from the Lower Wright Glacier into Lake Vanda, a highly stratified, closed-basin lake (Fig. 1). Along its course, it moves through bedrock channels, alluvial flats, moraine complexes, and a boulder pavement. Unlike more familiar rivers throughout the world, the Onyx is not influenced by rainfall and, during cooler years, even runoff from tributary streams is minor. Many of the factors (Allan, 1995) that contribute to the complexities of stream composition and compositional variations over space and time elsewhere are either absent or greatly attenuated in this system. The river, therefore, presents a rare opportunity to examine solute acquisition and removal processes along a channel which is largely unperturbed by ancillary inputs.

While it has been dubbed (humorously) the Mississippi and the Amazon of the Antarctic Continent, the Onyx is quite small by temperate zone standards, having a depth of approximately one half meter in a channel that is, depending on location, roughly 10 m wide. In the late 1960s, the New Zealand Ministry of Works established a permanent weir site near Lake Vanda, and since then instantaneous and annual river discharges have been recorded. Discharges have been highly variable from year to year, but in the past two decades there has been a notable increase in flow that has resulted in a 10-m rise in the level of Lake Vanda since 1973 (Howard-Williams et al., 1997). Total water inputs to Lake Vanda since measurements began in the late 1960s average between 2.5 and 5.5 × 109 L/yr (Howard-Williams et al., 1997).

In addition to hydrological investigations, there have been a number of studies that provide a background for the present work. Shaw and Healy (1980) examined the morphology of the river and reported on a characteristic hierarchy of channels that ranges from a sandy, narrow, inner channel, some 10 m wide, to a series of broad outlying terraces typically associated with higher flows. Mosley (1988) investigated bedload transport and sediment yield in the river and estimated a specific sediment yield of 5.9 t km− 1 yr−1. It was noted that this yield was some two orders of magnitude less than Arctic and Alpine proglacial rivers.

As part of their effort to determine the chemical evolution of Lake Vanda, Green and Canfield (1984) studied the major ion geochemistry of the river and observed that the Onyx is relatively rich in sodium and bicarbonate, in contrast to the lake itself, which is dominated by calcium and chloride. Their study showed that the Onyx River alone could not have produced a brine having the qualitative geochemical features of the lower saline waters of Lake Vanda. In a related study, Canfield and Green (1985) reported on the nutrient chemistry of the Onyx River and determined nitrogen and phosphorus loading, N/P ratios, and variations in nutrient concentrations along the river course. They concluded that glacial ice was an important source of nitrogen and that phosphorus was derived from the weathering of valley soils.

Howard-Williams et al. (1997) presented detailed information on nutrient sinks within the Onyx River system. They observed that Lake Brownworth, near the Lower Wright Glacier (Fig. 1) has extensive cyanobacterial mats on the lake bed and appears to serve as a significant nutrient filter for both nitrogen and phosphorus. An even more important region for nutrient extraction is the “Boulder Pavement,” located some four kilometers upstream from Lake Vanda. This 1.5-km stretch of river is characterized by large flat rocks that have been colonized by algae and cyanobacteria. Howard-Williams et al. (1997) showed that in this region nutrients that otherwise would have entered Lake Vanda are stripped from the river. The “Boulder Pavement” thus serves as an efficient nutrient sink and contributes to the lake’s ultra-oligotrophic condition (Howard-Williams et al., 1997).

Studies of trace metals in the Onyx River, and, indeed, in Dry Valley streams generally, are limited. Green et al. (1986) reported values for Mn, Fe, Cu, and Cd in several samples collected at the recording weir near the input to Lake Vanda, and they used these data to approximate metal residence times in the lake. Their trace metal investigation, however, was restricted to a single sampling site and made no claims about longitudinal changes along the river. Masuda et al. (1982) presented data on 14 transition series and rare earth metals in one unfiltered river sample and suggested atmospheric pathways for metal entry into the river. More systematic work on metals in Antarctic streams has not been reported.

Chemical weathering is linked to climate through the carbon cycle. Both carbonate and silicate weathering result in the uptake of atmospheric carbon dioxide and in the production of bicarbonate ions as part of the dissolved stream load. These processes can be represented in simplified form by the two equations below (Berner and Berner, 1994): for Carbonates:CO2+H2O+CaCO3=Ca2+HCO3 for Silicates:2CO2+3H2O+CaSiO3=HCO3+H4SiO4

Classically, it has been argued (see, for example, White and Blum, 1995) that higher temperatures result in greater atmospheric moisture content, increased precipitation, higher runoff rates, enhanced rates of reaction, and, ultimately, increased rates of carbon dioxide uptake. On this view, chemical weathering serves as a negative feedback in the climate system, since warmer temperatures naturally promote carbon dioxide drawdown and, hence, a return to cooler regimes. This temperature-dependence, “thermostat” model, which would predict very low rates of chemical weathering for Antarctic streams, has been challenged by a number of recent studies. On a grand scale, for example, Raymo and Ruddiman (1992) have argued that orogeny, as in the case of the Himalayas, can override the temperature factor by exposing new rock surfaces to stream action, thereby augmenting weathering rates and carbon dioxide removal from the atmosphere. The “ Raymo Hypothesis” attempts to account for global cooling over the past 40 million years in terms of this mountain building effect.

Recently, Edmond and Huh (1997) computed basin-scale consumption rates of carbon dioxide from terrain in hot and cold climates and observed that the total dissolved solids flux and the carbon dioxide uptake flux for the Siberian Craton are comparable to those observed in the tropics. In explaining their cold region data, they emphasized the importance of such physical processes as frost shattering, which is responsible for exposing fresh surfaces to chemical attack. Edmond and Huh (1997) employed a number of measures to compute carbon dioxide fluxes. Among these were the weathering rates of silica, the total number of equivalents of cations in the dissolved load, and the bicarbonate flux. The last of these—obtained from the bicarbonate alkalinity—is a direct measure of fixed carbon dioxide, provided that it is not compromised by the presence of unanalyzed organic acids.

On a much smaller scale, Lyons et al. (1997) and Nezat et al. (2001) have shown, perhaps surprisingly, that the Taylor Valley (Antarctica) streams in their study had weathering rates (as determined by silica and bicarbonate fluxes) that were comparable to rates in far warmer climate regimes. For example, the values for streams in the Lake Fryxell and Lake Hoare Basins were well in excess of world average weathering rates (64 × 103 mol HCO3 km−2 yr−1) and higher than rates reported for the Mekong and Amazon Rivers. These data suggested to Nezat et al. (2001) that elevated temperature and precipitation, both of which are absent from the McMurdo Dry Valleys, are likely not determinative of denudation rate. No estimate of carbon dioxide uptake was made in their study, but we will calculate this from their data for discussion purposes.

Section snippets

Field

Sampling was conducted to obtain information on: (a) chemical changes along the length of the river; (b) changes across the Boulder Pavement; and (c) compositional variations with time and discharge at the recording weir. The longitudinal study (a) was carried out on January 15, 1996. Eleven samples were collected at the sites shown in Figure 1. These include waters entering the Onyx system from the north (site 1), direct runoff from the Lower Wright Glacier (site 2), and nine samples along the

Major ion profiles and estimated chemical weathering rates

The chemical composition of the Onyx River is given at 11 sites along its course in Table 2. We took the outflow of Lake Brownworth as the origin of flow (0.0 km) and indicate downstream distances as positive and upstream distances as negative. Thus, the first sample listed in the table was collected 3.5 km upstream of Lake Brownworth. The third sample was collected at 0.35 km downstream of this lake.

For this profile, samples were collected in the evening over a 2-h period (aided by helicopter

Conclusions

With the exception of the small amounts of soluble material associated with glacial ice, solutes in the Onyx River are ultimately derived from the weathering of valley rocks and soils. Paleolake carbonate deposits, aluminosilicate minerals associated with the basement complex wind-blown marine salts, manganese- and iron-rich varnishes and calcite crusts are all likely solute sources. This study has shown that, despite the many variables associated with channel width, stream velocity and mineral

Acknowledgments

We wish to thank Bonnie Jo Bratina and Bradley Stevenson for assistance and companionship in the field. We are grateful to Clive Howard-Williams for helping us to plan the study of the Onyx River and for continuing discussions. We thank Betty Marak for typing and reviewing the manuscript. Also, we wish to gratefully acknowledge the many thoughtful critical comments and suggestions presented by the AE and by several of our reviewers. This research was funded through NSF award OPP93-19044, and we

References (39)

  • K.K. Turekian

    The fate of metals in the ocean

    Geochim. Cosmochim. Acta

    (1977)
  • K.A. Welch et al.

    Determination of major element chemistry in terrestrial waters from Antarctica by ion chromatography

    J. Chromatogr. A.

    (1996)
  • A.F. White et al.

    Effects of climate on chemical weathering in watersheds

    Geochim. Cosmochim. Acta

    (1995)
  • J.D. Allan

    Stream EcologyStructure and Function of Running Waters

    (1995)
  • R.A. Berner et al.

    Global EnvironmentWater, Air and Geochemical Cycles

    (1994)
  • D.E. Canfield et al.

    The cycling of nutrients in a closed-basin Antarctic lakeLake Vanda

    Biogeochemistry

    (1985)
  • G.G.C. Claridge et al.

    Origin of nitrate deposits

    Nature

    (1968)
  • G.G.C. Claridge et al.

    The salts in Antarctic soils and their distribution and relationship to soil processes

    Soil Sci.

    (1977)
  • J.M. Edmond et al.

    Chem. weathering yields from basement and orogenic terrains in hot and cold climates

  • Cited by (43)

    • Temporal and diurnal analysis of trace elements in the Cryospheric water at remote Laohugou basin in northeast Tibetan Plateau

      2017, Chemosphere
      Citation Excerpt :

      Glacier meltwater is usually low solute concentrations than that of other rural sites as its snow-ice based formation process. For example, previous study at the Wright Valley of the Antarctic showed that, the element concentrations (e.g. Mn, Fe and Ni) are lower than those in world's stream waters (Green et al., 2005). The Tibetan Plateau, usually called “the third pole” (Yao et al., 2012), is one of the most remote and isolated regions in the world.

    View all citing articles on Scopus

    Associate editor: K. Falkner

    View full text