DIFFERENTIATING ATMOSPHERIC AND MINERAL SOURCES OF SULFUR DURING SNOWMELT USING 34S AND 35S OF SULFATE AND 18O OF WATER

During the 2000 snowmelt at Sleepers River, Vermont, we used ³⁴S to quantify atmospheric and mineral weathering sources of sulfate to streamwater. We compared these data for consistency with independent measures of the atmospheric S component using ³⁵S, and old/new water fractions using ¹⁸O.

Weathering of sulfide minerals in the phyllite underlying Sleepers River watershed yields sulfate with a distinctively high d³⁴S value near +8.0 permil. The d³⁴S of the snowpack was +5.6 permil, consistent with the relatively constant regional value for precipitation. Using these two values as end members, the mineral and atmospheric inputs of sulfate to stream water during snowmelt were quantified using a simple two-component mixing model. At the most intensively sampled site, a 13-ha forested headwater catchment, 11 samples were taken during all phases of snowmelt from late March through April. The maximum atmospheric sulfate contribution reached 50% during an early rain-on-snow event on 28 March and again during a late diurnal peak. For most of the melt period, the atmospheric sulfate contribution varied between 10 and 40%, with higher percentages at higher flows.

Atmospheric sulfate contains ³⁵S, a radioactive isotope with a half-life of 87 days. The presence of ³⁵S in streamwater is diagnostic of an atmospheric source, although allowance must be made for ³⁵S decay in the snowpack. No ³⁵S was detected in a low-flow sample during early melt, when the ³⁴S data indicated 100% mineral sulfate. In contrast, the highest ³⁵S activity of 2.6 mBq/L occurred during the 28 March event, compared to 2.2 mBq/L from a snowpack sample. The excess activity in the stream probably reflects aging of the snowpack, loss of snowpack sulfate during mid-winter melt events, and new ³⁵S introduced in rain.

Oxygen-18 was used to calculate old and new water percentages based on two-component mixing of a constant groundwater signal and a variable meltwater input. The atmospheric sulfate percentage closely tracked the new water percentage (e.g. 40% new water with 50% atmospheric sulfate on 28 March), until an isotopically heavy rain-on-snow event on 4 April precluded further ¹⁸O separations.

Collectively, these results show that large amounts of recently deposited atmospheric sulfur enter the stream with the flush of new water during snowmelt.