The present study examines the contribution of combined sewer overflows (CSO) to loads and concentrations of trace contaminants in receiving surface water. A simple method to assess the ratio of CSO to wastewater treatment plant (WWTP) effluents was applied to the urban River Spree in Berlin, Germany. The assessment indicated that annual loads are dominated by CSO for substances with removal in WWTP above w95%. Moreover, it showed that substances with high removal in WWTP can lead to concentration peaks in the river during CSO events. The calculated results could be verified based on eight years of monitoring data from the River Spree, collected between 2000 and 2007. Substances that are well removed in WWTP such as NTA (nitrilotriacetic acid) were found to occur in significantly increased concentration during CSO, while the concentration of substances that are poorly removable in WWTP such as EDTA (ethylenediaminetetraacetic acid) decreased in CSO-influenced samples due to dilution effects. The overall results indicate the potential importance of the CSO pathway of well-removable sewage-based trace contaminants to rivers. In particular, high concentrations during CSO events may be relevant for aquatic organisms. Given the results, it is suggested to include well-removable, sewage-based trace contaminants, a substance group often neglected in the past, in future studies on urban rivers in case of combined sewer systems. The presented methodology is suggested for a first assessment, since it is based solely on urban drainage data, which is available in most cities.

The effect of combined sewer overflow (CSO) control measures should be validated during operation based on monitoring of CSO activity and subsequent comparison with (legal) requirements. However, most CSO monitoring programs have been started only recently and therefore no long-term data is available for reliable efficiency control. A method is proposed that focuses on rainfall data for evaluating the effectiveness of CSO control measures. It is applicable if a sufficient time-series of rainfall data and a limited set of data on CSO discharges are available. The method is demonstrated for four catchments of the Berlin combined sewer system. The analysis of the 2000-2007 data shows the effect of CSO control measures, such as activation of in-pipe storage capacities within the Berlin system. The catchment, where measures are fully implemented shows less than 40 % of the CSO activity of those catchments, where measures have not yet or not yet completely been realised.

A mixed surface and sub-surface flow riparian zone in Brittany (France), which is mainly fed by water from drainage ditches, was monitored for nitrate retention over three years from 2005 to 2007. Results show high time-averaged nitrate retention of >90 % for subsurface and ~70 % for surface passage. However, no retention could be detected during major rain events, which reduced the overall (flow-averaged) retention to ~40 %. Based on the findings, higher nitrate retention can be reached by increasing (i) the water residence time in buffer systems, (ii) the fraction of subsurface passage or (iii) denitrification rates in the system. (i) is only feasible if (active) buffer volume is enlarged, which may be difficult in practice. In the case of Brittany an enlargement can also be reached by extending buffer systems into existing drainage ditches. (ii) is of particular importance in areas with low soil permeability. In such areas, addition of gravel or sand beds can be considered. Regarding (iii), denitrification turns maximal under anaerobic conditions if sufficient carbon sources are available. In straw- and bark-filled column experiments we found high nitrate retention rates of >99 % and ~40 %, respectively, during a comparably low residence time of ~5 hours. As a result, the addition of external carbon sources to buffer systems is suggested. Currently, several pilot sites are constructed in the Ic watershed in Brittany attempting to take into account points (i) to (iii). For the following four buffer types, monitoring will start in February 2010: (a) two short drainage ditches, filled with carbon sources, (b) one drainage ditch and (c) one riparian wetland, each filled with a gravel filter, and optional upstream addition of carbon sources.

Stormwater impact guidelines for dissolved oxygen (DO) were applied to the Berlin River Spree, which (a) receives the effluents of more than 100 combined sewer discharge points and (b) is subject to significant anthropogenic background pollution. Discrimination of DO depressions, which are the direct result of combined sewer overflows (CSO) from DO depressions which are not related to CSO was achieved by combining stormwater impact guidelines with the analysis of data for: (i) rain events before critical DO depressions, (ii) water temperature (T) and conductivity as indicators for CSO impact in the river and (iii) T and DO before critical DO depressions to assess the effect of background pollution. Results indicate that the River Spree is in a critical state regarding DO for two main reasons: (a) upstream of the stretch with CSO discharge points because of background pollution and (b) downstream of the stretch because of CSO. Highly critical situations with DO < 2 mg L-1 only occurred under CSO influence. Nevertheless, the analysis underlines the importance of measures to reduce both CSO and background pollution in urban rivers.

We quantified the areal hypolimnetic mineralization rate (AHM; total areal hypolimnetic oxygen depletion including the formation of reduced substances) in two formerly eutrophic lakes based on 20 yr of water-column data collected during oligotrophication. The upward diffusion of reduced substances originating from the decomposition of organic matter in the sediment was determined from pore-water profiles and related to the time of deposition. More than 80% of AHM was due to degradation of organic matter in the water column (including sediment surface) and diffusion of reduced substances from sediment layers younger than 10 yr. Sediments older than 10 yr, including the eutrophic past, accounted for , 15% of AHM. This ‘‘old’’ contribution corresponds to a 20–43% fraction of the total sediment-based AHM. The contribution from old sediment layers to AHM is expected to be even lower in lakes with deeper hypolimnia (. 12 m). In summary, oxygen consumption in stratified hypolimnia is controlled mainly by the present lake productivity. As a result, technical lake management measures, such as oxygenation, artificial mixing, or sediment dredging, cannot efficiently decrease the flux of

The Aquisafe project aims at mitigation of diffuse pollution from agricultural sources to protect surface water resources. The first project phase (2007-2009) focused on the review of available information and preliminary tests regarding (i) most relevant contaminants, (ii) system-analytical tools to assess sources and pathways of diffuse agricultural pollution, (iii) the potential of mitigation zones, such as wetlands or riparian buffers, to reduce diffuse agricultural pollution of surface waters and (iv) experimental setups to simulate mitigation zones under controlled conditions. The present report deals with (i), providing information on trace substances, which enter surface water predominantly via diffuse sources in rural or semi-rural environments. In particular, it provides a priority list of relevant substances to aid planning of monitoring programs at waterworks, which abstract surface water from rural watersheds, for which information on substance use is sparse. As this ranking is limited to substances for which broad data sets are available from literature, it is compared to actual screening programs in predominantly rural catchments in Brittany (France) and Indiana (USA). The literature review identified pesticides as the dominant known diffuse contaminant group in rural and semi-rural settings (section 2.1). This is confirmed for the agriculturally dominated Ic Catchment in France and Upper White River Watershed in the USA, where pesticides were found to dominate the diffuse source compounds (section 3). Seven agricultural pesticides were detected in the Ic Catchment with AMPA and atrazine being the most common compounds, detected in 54 % and 41 % of all the samples, respectively. In the White River Basin 26 of the 38 detected compounds were pesticides making them the largest group of chemicals detected. Based on literature values on pesticide detection in surface waters in Germany, France and the USA, a priority list was established in section 2.2 of this report (see Table on page vi). Only seven substances were among the 20 most relevant pesticides, both in the USA and in Europe. Accordingly, US and European substances are distinguished in the priority list. Most frequently detected substances were atrazine, metolachlor and simazine for the USA, AMPA (metabolite of glyphosate), diuron and atrazine for France and diuron, atrazine and isoproturon for Germany. The importance of atrazine in Europe is interesting, since it was already banned at the time of the monitoring, indicating the high persistency of atrazine in groundwater. In some cases in Germany, concentrations in surface waters were found to follow typical seasonal application patterns, indicating illegal use (pers. Comm.. M. Bach). Although the list of substances in the USA and in Europe differ, there is an agreement to the fact that many of the pesticides applied in agriculture find their way into surface waters. The concentrations found are often beyond 0.1 µg/L. For the EU this level already corresponds to the drinking water limit. Thus, if surface water is used for drinking water production pesticides seem to be of high relevance. In finished drinking water, frequently-used Isoproturon and Bentazon were most frequently detected in Germany and France. The importance for drinking water production is emphasized by frequent detections above 0.1 µg/L in finished drinking water in nine waterworks in the US. Regarding drinking water regulation, the thresholds in the USA are substance-specific and generally more than one magnitude higher than 0.1 µg/L. As a result threshold exceedance was mainly found for Atrazine. In terms of treatability in water works, the priority list includes the efficiency of classical treatment (flocculation, filtration, ozonation) and of powdered activated carbon (PAC), which is often added in emergency situations. Particularly problematic are triazines (such as atrazine), phenoxy-type substances (such as 2,4-D and Mecoprop) and Anilides/Anilines (such as Metolachlor and Acetochlor). The pesticides found in the screenings are in good agreement with the priority list of most problematic pesticides for the US and Europe. AMPA and atrazine, the substances detected most frequently in the Ic catchment, as well as 2,4-D and dichlorprop, which were found in high concentrations > 0.1 µg/L in one sample are all included in the Europe top 20 of the priority list. Other substances on the list may not have been found because they were not measured, because of relatively high analytical detection limits of the screening or simply because they are not used in the basin, dominated by corn and wheat cultures. In the White River Basin, atrazine, acetochlor and simazine were detected at concentrations exceeding early warning levels utilized by several states in the United States, indicating their high relevance concerning drinking water production. They are also included in the US top 20 of the priority list. The priority list is a reliable basis for potentially problematic pesticides. It can thus be used as a starting point for monitoring programs in rural catchments, where no specific information on pesticide use are available. If looking for pesticides in surface water, it is important to take times of application of regarded pesticides into consideration, as shown by strong fluctuations in atrazine concentrations in the source water of a waterworks in Indiana (Figure 12 of this report). The screening results indicate that also other contaminants than pesticides may play a role in rural catchments. In the screening in the semi-rural catchments in Indiana, twelve of the detected 38 substances were not pesticides, but belonged to other groups, such as domestic use products, manufacturing additives or gasoline hydrocarbons. Of these twelve substances, seven were only found in one of the two catchments, showing a strong catchment-specific relationship. The findings indicate that other substances than pesticides may be of local importance, though in the case study all 12 substances were at least 50-fold below human health benchmarks (if defined). We conclude that the pesticide priority list given below is a good starting point for diffuse pollution screening even though it may possibly not be sufficient if major local influences, such as factories, large roads with stormwater discharges, CSO or specific local pesticide uses are present.