Climate Change 2001:
Working Group II: Impacts, Adaptation and Vulnerability
Other reports in this collection

16.2.6.4. Sensitivity of Arctic Ocean to River Flow

Increased future melting from the Greenland ice sheet and Arctic glaciers and ice caps has potential to increase significantly freshwater runoff to the northern circumpolar seas on time scales as short as decades. Combined with changes in Arctic runoff, there will be an effect on Arctic sea ice. Freshwater input to the Arctic Ocean is important for growth, duration, and melt of sea ice, particularly on the large, shallow continental shelf areas that constitute 30% of the area of the Arctic basin. Spreading over the ocean of lower density freshwater from summer runoff, together with sinking of denser brine rejected during autumn and winter sea-ice production, maintain characteristic salinity and temperature layering. This leads to the Arctic halocline—a cold, salt-stratified layer that separates the upper mixed layer from underlying warm saline waters. The strength and position of surface and deep currents in the slope water south of Newfoundland are thought to vary as a coupled system in relation to the dipole in atmospheric sea-level pressure known as the NAO (Keigwin and Pickart, 1999).

Overall, runoff to the Arctic Ocean is approximately 3-4 x 103 km3 yr-1 (Prowse and Flegg, 2000)—twice that produced by precipitation minus evaporation for the Arctic Ocean or the influx of low-salinity ocean water (Barry and Serreze, 2000). Export of freshwater (as low-salinity water and sea ice) from the Arctic Ocean into the North Atlantic couples northern latitudes to the world thermohaline circulation.

Almost 70% of the total runoff to the Arctic Ocean is provided by four rivers: the Ob, Yenisey, and Lena in Russia and the Mackenzie in Canada. The three Russian rivers have flow measurements dating back to the 1930s, whereas the Mackenzie record begins only in 1972. Although there is no significant long-term trend in flow for these rivers, there has been a slight increase in runoff beginning in the late-1970s for other rivers in European Russia and western Siberia, with some dependence on season (Georgievskii et al., 1996; Georgievskii, 1998; Grabs et al., 2000; Shiklomanov et al., 2000). Shiklomanov et al. (2000) found that historical temperature shifts of -0.5 to +1.0°C had little effect on river inflows to the Arctic Ocean, which remained within 3-5% of their long-term average.

For a range of future climate scenarios, results from hydrological routines embedded in climate models and from independent hydrological models that are driven by climate model output indicate that discharge from the major Arctic rivers will increase significantly (Miller and Russell, 1992; Shiklomanov, 1994; van Blarcum et al., 1995; Shiklomanov, 1997; Hagemann and Dümenil, 1998; Lewis et al., 2000). Mackenzie basin studies, however, have projected some reduction in streamflow as a result of expected climate warming (MBIS, 1997). Summarizing the general results, Shiklomanov et al. (2000) note that with an atmospheric CO2 doubling, total annual inflow to the Arctic Ocean will increase by 10-20%, with a 1.5- to 2.0-fold increase during winter, although most of the flow still will occur in summer. Such predictions must be treated cautiously, however, because most climate models have been found to overestimate Arctic precipitation and simulate soil moisture and evaporation poorly (Robock et al., 1998; Walsh et al., 1998). Uncertainty in hydrological modeling of high-latitude river systems must be resolved before the risk of changes to the Arctic Ocean freshwater budget and thermohaline circulation can be quantified accurately (Carmack, 2000; Prowse and Flegg, 2000).



Other reports in this collection