Data through the mid 1990s indicate generally negative mass balances for small Arctic glaciers (Dyurgerov and Meier, 1997), parallelling a global tendency. Few of the Arctic records extend before 1960. Overall, these changes are consistent with the warming of recent decades. While mass balances for some glaciers, such as in the montane parts of Scandinavia and Iceland, had been positive due to increased winter precipitation, there are recent indications of mass loss. Based on data from 1979 to 1999, Abdalati and Steffen (2001) show a tendency towards increased melt over the Greenland Ice Sheet. While this trend was interrupted in 1992, apparently as a result of stratospheric dust from the eruption of M. Pinatubo, the upward trend subsequently resumed. The area of the ice sheet experiencing summer melt set a new record in 2002 (Steffen, personal communication). In the terrestrial domain, data through the 1990s indicate warming of permafrost in Alaska and Russia (Pavlov, 1994; Osterkamp and Romanovsky, 1999). Permafrost cooling in eastern Canada through the mid 1990s (Wang and Allard, 1995) appears to have ended, to be replaced by warming (ACIA, 2005). Satellite records from 1966 through 2002 point to the dominance of negative snow cover anomalies over both continents since the late 1980s, largely due to spring and summer deficits (Figure 11.5). The change is more of a step function than a trend, and even within the period of generally less snow cover some months are well above average.
In contrast, based on analyses from 1900 though 2002, annual precipitation over terrestrial regions as averaged over the 55-85° N band (expressed as anomalies with respect to 1951-80 means), has shown a general upward tendency (Figure 11.6). The annual pattern is most strongly allied with changes during winter, summer and autumn. However, the largest changes occurred during the first half of the century.
Zonal means of course mask regional trends. In the Lena Basin, Yang et al. (2002) document that over the period 1935-98, there have been small positive trends in precipitation from October through June, particularly for November, December, March and May. By comparison, precipitation in July and August shows downward trends, strongest in August (15 mm over the study period). In a study for the period 196099, Serreze et al. (2003a) also find winter precipitation increases in the neighboring Yenisey amounting to 14 mm over the 40 years. They also document a small reduction in summer precipitation in this basin but a much more pronounced reduction in aerologically derived summer P —ET (Chapter 6), indicative of strong drying. The analysis of Thompson et al. (2000) for the period 1968-96 (which we will return to later) also indicates high-latitude increases in winter precipitation, which have been most pronounced over Eurasia extending to about 60° E, the central part of the continent, and parts of Alaska. There have also been local negative trends. While these studies have examined different periods and regions, the common thread is a broad increase in winter precipitation and regional summer decreases.
Peterson et al. (2002) find that the aggregate average annual discharge from the six largest Eurasian rivers draining into the Arctic Ocean increased by about 7% over the period 1936-99 (Figure 11.7). While this pattern is not seen for all rivers, aggregate discharge is now about 128 km3 greater than it was when routine measurements began. To put this in perspective, the annual mean discharge from the Yenisey river is about 630 km3 yr—1.
Yang et al. (2002) examined changes in discharge at the mouth of the Lena Basin for the period 1935-99. During the cold season, there have been significant increases (35-90%) in discharge, attended by decreases in river ice thickness. There has also been a hydrologic shift toward more discharge in May, resulting in lower daily maximum discharge in June. Hydrologic changes in summer are less apparent. They suggest that the winter increases in discharge relate in part to the winter warming and increased winter precipitation noted above. These two factors result in higher ground
temperatures (snow cover acting as an insulator, see Sokratov and Barry (2002)) leading to a delay in both the initial and complete freeze-up of a thicker active layer the following autumn. The thicker active layer has greater groundwater storage capacity due to both melt of ground ice and increased winter precipitation input. The increase in groundwater storage implies a greater contribution of subsurface water to river systems, which increases the winter discharge. The hydrographic shift resulting in more May and less June discharge is attributed to climate warming during the snow melt period. Serreze et al. (2003a) document similar, albeit larger winter discharge
increases in the Yenisey as well as as shifts in peak discharge to earlier in the season, which are similarly interpreted.
However, direct human alterations of river hydrology appear to be a significant problem for some basins. Indeed, the subsequent study by McClelland et al. (2004) provides convincing evidence that while both the annual trend in discharge and the shift in peak discharge to earlier in the season represent true climate signals, the winter increases in discharge are largely a result of dam operations, with this effect most pronounced in the Yenisey. They interpret the upward trend in annual discharge as primarily a result of increasing precipitation, with changes in permafrost as well as increases in forest fire frequency possibly playing contributing roles. Regarding the latter, fire scar chronologies in some forests of central Siberia and the Russian far east point to substantial increases in fire frequency (Arbatskaya and Vaganov, 1997; Cushman and Wallin, 2002). With loss or damage of vegetation, evapo-transpiration decreases, so more of the precipitation becomes runoff.
A definitive answer may remain elusive. For example, accounting for the observed increase in annual discharge would require an increase in net precipitation (as averaged over the six major Eurasian rivers for 1936 through 1999) of only 30 mm (from 406 mm to 436 mm). Such a change may well be too small to detect unambiguously with the sparse and error-prone precipitation network (McClelland et al., 2004). The small regional precipitation changes cited earlier must be viewed with the same caveat.
Was this article helpful?