In the city of Berlin combined sewer overflows (CSO) can lead to severe depressions in dissolved oxygen (DO) of receiving urban rivers and hence to acute stress for the local fish fauna. To quantify CSO impacts and optimize sewer management strategies, a model-based planning instrument has been developed. It couples the urban drainage model InfoWorks CS which simulates hydraulics and pollutant transport in the sewer with the river water quality model QSim which simulates hydraulics, mass transport and various biogeochemical processes in the receiving water body. To identify simulated CSO impacts, concentration-durationfrequency-thresholds for DO are applied to river model results via an impact assessment tool. Two kinds of impacts are distinguished: i) suboptimal conditions and ii) critical conditions for which acute fish kills are possible. In the case of Berlin, suboptimal conditions are observed on up to 92 days per year, predominantly during periods of low discharge and high temperatures whereas critical conditions only occur after CSO. For model calibration and validation, continuous measurements in both river and sewer are used. First simulations show good accordance between simulated and measured DO concentration in the river with Nash-Sutcliffe efficiencies between 0.70 and 0.79 for an eight-month time period at three different river monitoring points. However, to assure satisfactory model performance for adverse DO conditions in particular, impact assessment results for measured and simulated data are compared. Regarding suboptimal DO conditions simulated and measured data show good agreement. Nevertheless model representation for critical conditions is poor for some river sections and requires further improvement for CSO conditions. The results underline the importance of combining different validation approaches when dealing with complex systems.

In the city of Berlin regular combined sewer overflows (CSO) lead to acute stress of aquatic organisms in the receiving River Spree and its side channels. Of most concern are oxygen depressions, following the inflow of degradable organic matter via ~180 CSO outlets, along a river stretch of 16 km. For the assessment of the severity of these oxygen depressions, an existing impact-based approach suggested by Lammersen (1997) was combined with information on the local fish fauna. Application of this locally adapted assessment method to seven years of oxygen measurements at a CSO hotspot in the river yielded an annual average of 14 periods with suboptimal conditions for which adverse effects on the fish fauna are expected and 20 periods with critical conditions for which acute fish kills are possible. Further investigation on rain and sewer management data proved that such critical conditions only occurred as a direct result of CSO events, whereas suboptimal conditions are also possible at dry weather and may last up to 32 days (Riechel et al. 2010).

The centre of Berlin, Germany, is drained by a combined sewer system. The receiving waters Havel and Spree are characterized by low flow velocities and an increased risk of eutrophication. High demands towards a reduction of the emission loads of combined sewer overflows (CSOs) down to 20 % of the mean annual runoff load of TSS, COD and BOD5 are formulated by the Berlin Water Authority. Therefore a pollution control plan will be carried out until the year 2020 that will lead to a storage enlargement of the combined sewer system by 100 %. To assess if these efforts will lead to the expected water quality of the receiving water regarding the objectives of the European Water Framework Directive, a method will be developed to evaluate in advance the achievable improvement. Starting from the actual status of the water body this model based method should allow for an estimation, if the good status will be achieved after the realization of the measures of storage upgrading in the sewer system. The study currently concentrates on the integrated water quality modelling of the high dynamic processes in the sewer system and the receiving water. The paper focuses on the simulation of oxygen concentration in the receiving water.

The objective of the studies performed in the scope of the Integrated Sewage Management (ISM) project on combined sewer overflows in Berlin, Germany was to develop methods that would make it possible to assess wastewater management measures performed under the city’s water management permit as well as more sophisticated strategies (e.g., global real time control) through the application of water body-related criteria. For this purpose, a preliminary study was first performed to characterize the underlying water body-specific processes and hydraulic, physical, chemical and ecological parameters relevant to the status of Berlin’s surface waters (LESZINSKI et al., 2007a). The second step involved the development of a method for water quality-oriented assessment of wastewater management measures (LESZINSKI ET AL., 2007b). In addition to the already recognized thresholds for dissolved oxygen concentration during continuous, long-term water load conditions, particular focus was placed on formulating requirements for oxygen demand under peak load conditions. Ammonia toxicity due to sewage input, another important stress factor for aquatic ecosystems, was also analyzed and threshold values for both chronic and acute peak ammonia loads were defined. The results of the third phase of this research are described in this report. Two numerical simulation models (for urban drainage networks and surface waters) were combined and the feasibility of the developed method was evaluated based on the case of a combined sewer overflow event documented by the surface water monitoring. The simulations were performed using InfoWorksTM CS hydrological/hydrodynamic urban drainage network modeling software (ISM model) and the GERRIS/HYDRAX/Qsim unsteady ecosystem modeling system. The latter model was developed by the Federal Institute of Hydrology in Koblenz and is used by the Senate Department of Health, Environment and Consumer Protection (SenGesUmV). The present report describes the theoretical principles of the utilized models, the base of data available for analysis of the selected event, and the assumptions made in cases of missing input data for hydraulic modeling as well as for the water quality simulations. The one-dimensional hydraulic modeling results for the branched surface water system of the reach Berlin-Charlottenburg demonstrated that the hydraulic conditions can be simulated with satisfactory accuracy using the current data. In the case of water temperature, it was also possible to achieve a high degree of agreement between the measured and computed values in spite of the lack of highresolution temporal input data from the tributaries (Landwehr Canal, Panke River, BerlinSpandau Ship Canal). However, this was not the case for dissolved oxygen concentration, the main parameter used for evaluation of combined water treatment. The DOC simulations computed using input data based on a monthly sampling interval did not show satisfactory agreement with the online measurements in the water system. Dry-weather biological processes, which were associated with high-level, short-term oxygen enrichment or consumption, could not be depicted in the simulations. After completion of the water quality simulations, the effect of variation of individual input parameters was assessed. This analysis showed that no significant improvement of agreement with the measured values could be achieved by adjusting the assumptions for individual parameters (chlorophyll-a and BSB5). In the case of ammonia, the second most important parameter, the available sampling data from the tributaries in the investigated water system were collected only once a month, if at all. Therefore, it cannot be expected that the temporal distribution of this parameter was correctly reflected by the model. The number of validation measurements taken within the water system was also insufficient. Summarizing the results of the study of the linked urban drainage/surface water quality model, which was tested for the first time, it can be concluded that InfoWorks CS and GERRIS/HYRDRAX/Qsim provide problem-oriented simulation tools for reaching the objective of ISM study of assessing various scenarios for reduction of impacts from combined sewer overflows. By contrast, the available data are deficient and do not allow to adjust and calibrate the models to meet the specific needs of this task, particularly in light of the fact that short-term effects of combined sewer overflows are to be analyzed.

The development of the integrated control of sewage network and wastewater treatment plant has progressed during the last decade. Nevertheless, an operational implementation of the concepts for huge, complex systems has hardly been realised. That was an obvious reason to initiate the project "Integrated Sewage Management (ISM)". The ISM project aimed at the development of strategies for an integrated management of the Berlin sewage system consisting of sewer networks (both, combined and separate system), pump stations, pressure mains and wwtp. For these purposes a numerical model of the collection system has been built up. Those catchments have been chosen that have a significant quantity of wastewater and are connected to at least one of the three main wastewater treatment plants of Berlin (Ruhleben, Waßmannsdorf and Schönerlinde). To enable an evaluation of total emissions it was necessary to incorporate not only catchment area and collection system but also the wwtp into the model. Furthermore, the Berlin specific transport of wastewater through pressure mains had to be considered. Both, advective pollutant transport and the limiting pressure situation had to be taken into account. An integrated model of collection system, pressure mains and wwtp has been set up for the catchment of wwtp Ruhleben for the study of a global control concept. Those processes that were of particular importance for the control concepts or had a significant influence on the criteria (derived from the objectives) had to be simulated adequately. Hence, for the Berlin model the main attention was paid to an accurate reproduction of in-pipe storage activation and the transport of wastewater through the pressure pipes. A sufficient set of data was available to model the system structure. For process parameter estimation the necessary information was taken from the operational SCADA system. Some gaps in the data could be closed by additional measurement campaigns (Bln VII, 2001; Bln X, 2002; Heiligensee, 2003). For modelling the collection system the dynamic flow routing model InfoWorks CS of Wallingford Software Limited has been chosen due to its user-friendliness (window navigation, GIS) and comprehensiveness (pollutant load calculation, long-time simulation, spatial rainfall distribution, rtc module). A suitable approach to the simulation of the Berlin pressure mains was found to be based on EPANET 2 of the U.S. Environmental Protection Agency. The software SIMBA® 5 of ifak System GmbH has been used to simulate the dynamic treatment processes. For the activated sludge conversion part the Activated Sludge Model No. 1 (ASM 1) has been used. The three models have been coupled in sequence on the basis of simple input and output files. Further on, in the framework of three sub studies the ISM model has been applied to operational questions. The applicability of the ISM model for the assessment of the impact of the NPA control on the wwtp was tested. NPA stands for “new pump automatic (Neue Pumpen Automatik)” and signifies a control concept that is implemented in the framework of the LISA project (BWB). The inflow to wwtp Schönerlinde has been simulated for one rain event and the NPA control of the pump stations could be simulated well on the basis of the InfoWorks rtc module. Furthermore, the ISM model has been applied to evaluate a concept for a level dependant real-time control (Pegelgesteuerte Förderstromregelung) of sewage pump stations. The idea of the concept was to build an easy function that allowed continuously varying the pumpage and implicitly managing available inline storage capacities within the trunk sewers. The objective was to smooth the delivery towards the treatment plant to avoid peak loads. The evaluation showed that it is possible to manage available inline storage volume by applying the control function. But only if there is an adequate retention volume of around 60.0 m³/ha Aimp or more a significant improvement of the flow characteristic towards the wwtp is possible. Consequently, in Berlin only two catchments have the potential for the introduced control concept (Charlottenburg III und Ruhleben). Finally, the effects and the benefit from global pump station control in comparison to local control have been studied on the basis of the integrated model. The assessment of the Berlin drainage system that was carried out before arrived at the conclusion that there is a high potential for the control of the total system. The positive rating can partly be ascribed to the high storage volume that can be activated within the trunk sewers and the high number of pump stations that are used as actuators. However, this potential is already used by locally controlling the pump stations and storing sewage in the collectors. The potential of a global control of sewage pump stations arises from the non-uniform distribution of rainfall and the non-uniform distribution of storage volumes over the system. Those conditions usually lead to a non-uniform utilisation of storage capacities and further on to sewer overflows that cannot be balanced by local control. A look on the simulated total emissions showed that concerning discharged quantities the load from the wwtp is highly dominant, since most of the time (under dry weather conditions) wwtp effluents are the only impact on the receiving water. Furthermore, the global control concept only works during rain situation and does not have an influence on dry weather effluents. Consequently, the influence of global control on yearly total emissions is marginal. Nevertheless, it could be shown that global control can avoid peak load situations at the inflow to the wwtp and consequently reduce peak loads in the effluent. The control concepts had a significant influence on the emissions from combined sewer overflows. The reduction of sewer overflows plays a prominent role since they present a highly dynamic impact on the water body. The simulations showed that on average during periods of cso 2.5 t COD/h enter the receiving water. Compared to that load the continuous impact from the wwtp effluent was only 0.4 t COD/h. Moreover, due to the high fraction of biodegradable organic substrate the impact from combined sewer overflows is of special relevance. In contrary to the refractory COD from wwtp effluents, 60 % of the COD from combined sewer overflows are biodegradable leading to extreme oxygen depletion within the receiving water. It could be shown that under current conditions at the wwtp (rain weather capacity of wwtp Ruhleben = 6700 l/s) a local control (= local automation) of the pump stations has an adverse effect on the performance of the sewage system. In contrary to an optimum coordination of the pump stations local control leads to an overloading of the wwtp and an increase of emissions from combined sewer overflows by 9 % (volume), 15 % (COD) and 20 % (TKN). Due to that reason the current operation provides for manual interventions in case of rain events to coordinate the delivery of the pump stations. This necessity will persist under the LISA automation. Assuming a future upgrade of wwtp Ruhleben and an increase in rain weather capacity up to 7650 l/s, global pump station control will result in cso emissions that are 19 % (volume), 20 % (COD) and 25 % (TKN) below that under local control (= local automation). The major deliverable of the ISM project is the model for the Berlin collection system (18 combined and 29 separate sewer systems that are connected to the three main wastewater treatment plants Ruhleben, Waßmannsdorf and Schönerlinde). The further application and maintenance of the sewer model will take place at BWB, department NA-G. The scope of studies that will be supported by the model covers operational planning as well as general, conceptual and investment planning (storage optimisation, problem of parasite water). Concerning the implementation of the global control concept that has been developed in the framework of the ISM project first tests shall be carried out in 2006 and 2007. Therefore, the follow-up project EVA (Entscheidungshilfesystem zur Verbundsteuerung von Abwasserpumpwerken / Decision support system for global control of sewage pump stations) was planned at KWB to enable support and a further cooperation between KWB and BWB. The algorithm has to be adapted to the operational and technical boundary conditions and a detailed practical planning in terms of control engineering has to be carried out. The main prerequisite for an implementation of the introduced control concept is the technical ability of the pump stations to increase delivery beyond the value of 2 * Qd,16. Simultaneously, an authorisation is necessary to introduce a flexible regulation of the pump station’s rain weather delivery off the value of 2 * Qd,16 as demanded nowadays by the Berlin water authority. If the necessary data is available (usually given by the existing scada system of BWB) and if the used pumps can be controlled according to the above-stated technical requirements, thestudied control concept can be implemented without any further constructional investment.