Sulfur dynamics in the F1 catchment at Lake Gårdsjön, about 40 km NE of Gothenburg (Western Sweden; 58o04'N, 12o03'E), have been investigated for four years (1990-1993). The F1 catchment is 3.6 ha and has discharge through a small creek to the north. Average rainfall and temperature in the catchment is 1198 mm and 5oC respectively. Soil cover is very thin, on average about 45 cm, but the lower parts of the catchment can have thicker soil cover. Around the creek and especially near the outlet in the Lake Gårdsjön, small peat areas can be found. The total cover of peat areas are about 10% of the total catchment area. Samples have been taken from bulk deposition, throughfall, groundwater (two wells; #5 and #9) and runoff. d34S mean value in the throughfall deposition is, +7.34 (weighted mean) and the groundwater wells in the catchment have d34S mean values of +8.02 , ±0.74; 95% C.F. interval; n=30 (well #5) and +6.94 , ±0.12; 95% C.F. interval; n=28 (well #9) respectively. The runoff d34S mean value is +7.51 (weighted mean; n=45). Two different patterns have been observed for the d34S values in the two groundwater wells, seasonal variation (well #5) and stable values (well #9). Runoff data showed seasonal variation, which over the year is observed as higher d34S values during summer and lower values during winter. The groundwater well with seasonal variation (well #5) and runoff showed a correlation of (r2=) 0.57 between d34S values and 0.75 between sulfate concentrations. Low discharge during summer can result in both extremely low d34S values (lowest record was -1.02 ) and high d34S values (up to +11 ).
The only source for the sulfur is the deposition, but no correlation for the d34S values can be observed between the deposition and runoff. Consequently, short term d34S variations in the deposition is not reflected in runoff, while in the long run the d34S values in runoff reflects the deposition. The runoff sulfate concentration show some anticorrelation with the d34S value (r2=0.22). Dissimilatory sulfate reduction is the only process that can at the same time shift d34S values and the sulfate concentration. Since dissimilatory sulfate reduction occur in the saturated zone, the sulfate in the unsaturated zone is the source sulfur. It is here suggested that this source sulfur is what is measured in the stable groundwater well (well #9). The groundwater well #9 is situated in the upper part of the catchment. It has been shown that the sulfate's isotope composition in groundwater and runoff can be used to identify the isotope composition of the B-horizon sulfate's isotope composition (Mörth & Torssander, 1995). The well #9 receives water from surroundings with no peat areas. Therefore, the d34S value in well #9 mirrors the B-horizon d34S value. Subsequently the incoming water to the peat areas bear this signature. Almost all water in the F1 catchment (not less than 80%, observed during snow melt) in runoff comes from groundwater (Rodhe, 1987).
Mass balance studies in F1 have shown that the sulfate budget seems to be balanced over a 10-year period (Hultberg & Grennfelt, 1992). However the error is large, about ±20%. Therefore the isotope shift within the catchment may be used to calculate the net retention of sulfur by dissimilatory sulfate reduction. In order to do that a one box steady-state model has been developed. The model calculates the amount net retained by using the d34S shift and a given fractionation factor. The amount sulfur net retained is about 3 % using a fractionation factor (a) of 1.020.
Hultberg, H. & Grennfelt P., Environ. Pollut. 75, 215-222 (1992).
Mörth, C-M. & Torssander, P., Water, Air and Soil Pollut. 79, 261-278 (1995).
Rodhe, A., The origin of streamwater traced by oxygen-18, Uppsala University, Dept. of Physical Geography, Report Series A, No. 41 (1987).