Field-based measurement and characterization of transport and turn-over of wastewater contaminants in streams

Abstract

The present work investigated the spatio-temporal variability of the input and fate of common wastewater contaminants (e.g. pharmaceuticals, flame retardants, personal-care products, etc.) in three contrasting rivers in the area of Tübingen, southern Germany. Field sampling particularly focused on assessing reactive transport and key parameters affecting contaminant removal in streams, as well as their association to particles during high-discharge events. Among the three investigated catchments, River Steinlach was the main investigation site with a catchment characterized by mixed land use. River Ammer exhibits a mainly agricultural and slightly more urbanized catchment and River Goldersbach drains a primarily forested catchment devoid from treated wastewater discharge or heavy traffic. Reactive transport was investigated during two sampling campaigns using novel approaches of the Lagrangian sampling scheme in a segment of River Steinlach downstream of a wastewater treatment plant (WWTP). For optimal results, environmental conditions favorable for natural attenuation were chosen – i.e. baseflow conditions, sunny and dry weather. One sampling approach, the so-called “tracer-based sampling”, coupled a conservative-tracer test with water sampling at four monitoring sites during a day- and a night-time experiment. The tracer breakthrough curves were deconvoluted to calculate transfer functions, mathematical operators describing the travel time distribution of a water parcel flowing between the measuring stations. The transfer functions served to disentangle conservative transport processes affecting observed concentrations from degradation of compounds investigated and fit first-order decay constants for photo-dependent and -independent elimination. The other approach, the so-called “24 h sampling”, monitored mass fluxes along a 24 h-water parcel at two sampling sites. A mass balance of studied compounds was calculated during a winter campaign to serve as a comparison to a previous identical study conducted during the summer season. The length of the water parcel was chosen to minimize biases by inaccuracies from timing and dispersion effect and to integrate complete diurnal patterns in contaminant input and removal. Since concentrations of suspended sediments are typically low during baseflow conditions in the investigated rivers, particle-associated transport was assessed by sampling of turbid water during flood events in Rivers Steinlach and Ammer. Linear correlations between total contaminant and particle concentrations were tested as it had previously been done for polycyclic aromatic hydrocarbons (PAHs). Finally, contaminant fluxes were monitored during a monthly sampling in River Steinlach, Ammer and Goldersbach and yearly loads were estimated and compared. Contaminants observed in River Steinlach exhibited conservative (e.g. carbamazepine) to highly reactive behaviors (e.g. oxcarbazepine was completely removed during the tracer-based sampling). The comparison between day- and night-time, and summer and winter conditions, highlighted the occurrence of photo-dependent and -independent elimination processes as well as their seasonal dependence. In winter, reactivity was reduced, with an absence of photo-dependent elimination and limited or no photo-independent elimination of most analyzed compounds. Contaminant removal was higher in the tracer-based study compared to the summer 24 h sampling or previous field studies. Enhanced photo-dependent and independent elimination observed during the tracer-based sampling were linked to local parameters such as the high proportion of treated wastewater, water temperature, dissolved organic carbon and nitrate concentrations, as well as low flow conditions with a larger flow fraction through the hyporheic zone. During day-time, photo-dependent degradation was highly efficient for metoprolol, bisoprolol or venlafaxine in the tracer-based study, but its impact on contaminant removal was comparatively similar or lower than photo-independent processes when averaged over 24 hours. A linear correlation between total contaminant and particle concentration was only found for six of the fourteen contaminants analyzed during the flood sampled in River Ammer. The method is based on the assumption that the dissolved concentration of contaminants and their average concentration on the particles does not vary over time. But the simultaneous input of dissolved contaminants during combined-sewer overflows parallel to the mobilization of particles may have led to such apparent correlations between total concentrations of contaminants and particle concentration. Without analyzing directly compounds onto the sediments, particle-associated transport could not be confirmed, nor a constant loading of the contaminants on suspended particles. If any, less than 51% for carbamazepine, diclofenac, naproxen, TCPP and triclosan and less than 67% for lidocaine was associated to particles for suspended sediment concentration around 300 mgL-1 during the Ammer event sampled. Fluxes calculated during the sampling campaigns identified the WWTP as the main source of wastewater contaminants in River Steinlach during dry-weather conditions. Yearly load estimations were challenged by contaminant input variability from the WWTP, the contaminant degradability and the erratic activation of additional contaminant sources by precipitation. While loads of individual contaminants in the pristine part of the Goldersbach catchment were negligible, they reached up to 10 kg a-1 (for HHCB-lactone and TCPP) in the Steinlach and Ammer catchments downstream of WWTPs. In comparison, PAHs, that are mainly released into rivers sorbed to particles during precipitation events, exhibited loads of 4 and 6 kg a-1 in River Steinlach and Ammer, respectively. In conclusion, contaminants released continuously by WWTPs can undergo no to complete natural attenuation in receiving rivers. But attenuation processes, particularly photo-dependent elimination, are strongly governed by prevailing environmental conditions and follow their variability in time. This variability is yet to be fully assessed, further than day and night or season alternation, to capture a more representative picture of compound reactivity according to environmental parameters met in single streams. Also, higher fluxes of contaminants are released during episodic precipitation events and sorption to particles in turbid waters during floods may occur, but remains to be demonstrated.