Search

Characterization of groundwater-surface water interactions remains challenging,especially in anthropogenically affected and geologically complex river systems. In thisstudy, groundwater inflow into the Ammer River was quantified by optimizingexfiltration rates in an implicit numerical model (FINIFLUX) to measured data ofseveral environmental tracers (e.g., sulfate (SO42−), radon (222Rn)) and organicconservative pollutants like carbamazepine and sulfamethoxazole emitted via awastewater treatment plant (WWTP). Modelling results of the selected tracers werecompared. The investigated Ammer River is strongly karstified and anthropogenically modified (millcanals, weirs, numerous artificial inputs) but representative for many small river systems inEurope. 222Rn activities in the river ranged from 500 to 3300 Bq m−3 and SO42−concentrations from 224 to 291 mg L−1. Concentrations of carbamazepine (90-130 ng L−1)and sulfamethoxazole (55-100 ng L−1) in the river water result from effluents of aWWTP. All tracers indicate major local groundwater inflows over the 6 km riverreach. Additional high peaks mainly observed in 222Rn activities indicate inputsalong geological fault lines, whereby increasing concentrations of carbamazepineand sulfamethoxazole may be attributed to inputs of groundwater contaminated byinfiltration of wastewater from a nearby WWTP. 222Rn-based calculations showed thebest optimization performance (R2=0.90) and result in a reasonable cumulativegroundwater inflow of 0.274 m3 s−1 (28.5% of total discharge) for the investigated reach.Poor optimizations for the other tracers may be due to the diurnal pattern of ionconcentrations induced by the WWTP and measurement uncertainties and the smallconcentration differences between river water and groundwater for carbamazepine andsulfamethoxazole. In our case study, 222Rn seems to be the most suitable environmental tracer despiteuncertainties in degassing and determination of the groundwater endmember concentration.However, only a multi tracer approach can account for the complexity of the investigatedriver system and may thus be indispensable for future investigations aiming to understandgroundwater-surface water interactions. [ABSTRACT FROM AUTHOR]