Northern latitude chemical weathering rates:: Clues from the Mackenzie River Basin, Canada

The main scope of this study is to investigate parameters controlling chemical weathering rates for a large river system submitted to subarctic climate. More than 110 river water samples from the Mackenzie River system (northern Canada) have been sampled and analyzed for major and trace elements and Sr isotopic ratios in the dissolved phase. The three main morphological units are reflected in water chemistry. Rivers from the Canadian Shield are very dilute, dominated by silicate weathering (Millot et al., 2002), whereas the rivers of the Rocky and Mackenzie Mountains as well as the rivers of the sedimentary Interior Platform are dominated by carbonate weathering and are SO4 rich. Compared to the rivers of the Mackenzie and Rocky Mountains, the rivers of the interior plains are organic, silica, and Na rich and constitute the dominant input term to the Mackenzie River mainstream. Rivers of the Canadian Shield area do not significantly contribute to the Mackenzie River system. Using elemental ratios and Sr isotopic ratios, a mathematical inversion procedure is presented that distinguishes between solutes derived from silicate weathering and solutes derived from carbonate weathering. Carbonate weathering rates are mostly controlled by runoff, which is higher in the mountainous part of the Mackenzie basin. These rates are comparable to the carbonate weathering rates of warmer areas of the world. It is possible that part of the carbonate weathering is controlled by sulfide oxidative weathering, but its extent remains difficult to assess. Conversely to what was stated by Edmond and Huh (1997), overall silicate weathering rates in the Mackenzie basin are low, ranging from 0.13 to 4.3 tons/km(2)/yr (Na + K + Ca + Mg), and confirm the negative action of temperature on silicate weathering rates for river basins in cold climates. In contrast to what has been observed in other large river systems such as the Amazon and Ganges Rivers, silicate weathering rates appear 3 to 4 times more elevated in the plains than in the mountainous headwaters. This contradicts the "Raymo hypothesis" (Raymo and Ruddiman, 1992). Isotopic characterization of suspended material clearly shows that the higher weathering rates reported for the plains are not due to the weathering of fine sediments produced in the mountains (e.g., by glaciers) and deposited in the plains. Rather, the relatively high chemical denudation rates in the plains are attributed to lithology (uncompacted shales), high mechanical denudation, and the abundance of soil organic matter derived from incomplete degradation and promoting crystal lattice degradation by element complexation. The three- to fourfold factor of chemical weathering enhancement between the plains and mountains is similar to the fourfold factor of enhancement found by Moulton et al. (2000) between unvegetated and vegetated watershed. This study confirms the negative action of temperature on silicate weathering for cold climate but shows that additional factors, such as organic matter, associated with northern watersheds are able to counteract the effect of temperature. This acceleration by a factor of 4 in the plains is equivalent to a 6degreesC increase in temperature. Copyright (C) 2003 Elsevier Science Ltd.