The chemical response of alpine catchments to snowmelt is quite important for aquatic biota. Our knowledge of the well-known ionic pulse, whereby the majority of the ionic content of the snowpack is released with the earliest fraction of meltwater [ Bales, 1992; Berg, 1992; Peters and Leavesley, 1993] is being integrated with a fundamental understanding of how the chemical composition of snowmelt runoff changes as it passes from the snowpack through a basin with a heterogeneous distribution of rock, soil, and vegetation. The snowpack itself is not a static medium, but undergoes post-depositional chemical and biological changes [ Davies et al., 1991].
In a lysimeter study at 2950-m Mammoth Mountain in the Sierra Nevada of California, the inverse relation between meltwater discharge from the base of the snowpack and discharge electrical conductivity was interpreted as the combination of a concentrated signal from regions in the pack less subject to leaching and a relatively dilute signal from near the snow surface where the snow was actively melting [ Bales et al., 1993]. It was also shown that compared to continuous conductivity measurements, nonsystematic grab sampling of snowpack meltwater can be misleading because of multiple ionic pulses over the ablation season and strong diurnal fluctuations in chemical concentrations. Twice-daily sampling, together with continuous conductivity measurements has been shown to be a quite satisfactory characterization of an alpine stream [ Gurnell et al., 1994]. Mathematical modeling of solute release at a point has taken a similar approach to that used in soil and groundwater [ Bales, 1991].
Additions to the top of the snowpack later in the melt season can produce multiple ionic pulses [ Jenkins et al., 1993], as can melt-freeze action [ Davis, 1991]. Evidence for dry deposition to the wet snowpack surface was noted in one season-long snow-pit and lysimeter study in the 3300 m Glacier Lakes (GLEES) study area in the Snowy Range of Wyoming [ Bales et al., 1990]. Snow-pit and snowmelt sampling in Emerald Lake basin in California have consistently failed to show contributions from dry deposition in that 2800-m high alpine catchment [ Williams and Melack, 1991 a; Williams and Melack, 1991 b].
Biological influences on snow chemical composition include: i) deposition to the snowpack by plants and animals [ Jones, 1991], ii) algal and bacterial activity in the snowpack itself [ Hoham, 1991; Duval, 1993; Brooks et al., 1993], and iii) bacterial and other activity in soil and litter beneath the snowpack. These influences can be particularly important for nitrogen species. Concentrations of in snow gas at GLEES were intermediate between those in the soil and atmosphere [ Sommerfeld, 1993]. concentrations in snow gas were several times those in the air above the pack, and at times elevated above soil concentrations [ Sommerfeld, 1991].
Reports of chemical analyses of alpine snow continue to show that it is a dilute medium compared to rain or to snow from lower-elevation sites. The central-Alaska snowpack was found to be very clean, in spite of the Arctic-wide air pollution known as Arctic haze, suggesting a lower than world-average removal rate of aged air contaminants in the Arctic [ Shaw et al., 1993]. Arctic samples from Alaska, however, showed evidence of a higher flux than did samples from the interior of Alaska [ Jaffe and Zukowski, 1993].
One alpine area that is close to urban pollution is the Chinese Tien Shan. is especially elevated [ Wake et al., 1991]; but as it is largely associated with , aeolean dust may be a major source [ Williams et al., 1992; Tonnessen et al., 1991]. In central and southern California, concentrations up to 7.4 were observed in fresh snow, and 60-95% of the total was calculated to have been derived from S(IV) oxidation [ Gunz and Hoffmann, 1990]. Another report from central and southern California showed a wide variation in snow and rime chemical compositions [ Berg et al., 1991]. Rime generally has higher chemical concentrations than does unrimed snow [ Duncan, 1992].
The acidifying influence of snowpack nitrogen has been a question in many snow-dominated catchments. In 11 Adirondack watersheds, it was found that most watersheds released nitrate ion during snowmelt, in addition to the release of nitrate from the snowpack [ Schaefer and Driscoll, 1993]. Studies in the Emerald Lake basin suggest that oxidation of snowpack in soils was a source of the observed pulse in stream waters [ Williams et al., 1991]. In an analysis of data from Eastern Brook Lake basin, on the east side of the Sierra Nevada, it was suggested that the average snowpack anion release rate is a good indicator of the acidification potential of the snowpack [ Chen et al., 1991]. The three-year average rate for Eastern Brook, 1.4 , was only one-tenth that for the Woods Lake watershed in the Adirondacks of New York. Several problems with this approach of using an average rate for both Eastern Brook and for the Sierra Nevada have been noted [ Bales et al., 1993]. For example, focusing on annual averages does not address impacts that may arise from due to high, short lived concentrations at the onset of spring snowmelt.
Combinations of chemical species in snow have been used as tracers for hydrologic flowpaths during snowmelt runoff. Much of the snowmelt runoff from even sparsely vegetated alpine basins with little soil appears to infiltrate soils and then reach streams as saturated return flow [ Williams, 1992]. Because the infiltration capacity of soils is limited, alpine lakes and streams are particularly susceptible to acidic pulses from changes in either snowmelt rates or chemical loadings.