Three different methods for fi ltration characterization in Membrane Bioreactor (MBR) systems were compared. These were the Delft Filtration Characterization Method (DFCm), the Berlin Filtration Method (BFM) and an ex situ side-stream fi ltration test cell for the determination of the critical fl ux. The ex situ fi ltration test cell and the DFCm fi lter activated sludge from a tank, while the BFM works in situ with a test cell directly submerged into the biological tank at similar operational conditions to a typical MBR plant. The mixed liquor of four different MBR units was characterised several times with the three fi ltration methods. The three tested methods seemed to agree in the classifi cation of the tested mixed liquors in terms of fi lterability except for one of the tested activated sludges. Additionally, three critical fl ux protocols were studied using the BFM fi ltration test cell. The fi rst consisted in the classical fl ux-step method, the second included relaxation between fi ltration steps and in the third protocol, 2 min fi ltration at a fi xed fl ux were performed before every fi ltration step. The last protocol was selected as the most representative of full scale MBR operation and the most interesting one for giving valuable information about the irreversibility of the fouling.

A continuous monitoring, using UV-VIS spectrometers, was carried out in Berlin from 2010 to 2012. It combined (i) continuous measurements of the quality and flow rates of combined sewer overflows (CSO) at one main CSO outlet downstream of the overflow structure and (ii) continuous measurements of water quality parameters at five sites within the urban stretch of the receiving River Spree. Locally, the collection of data aims at (i) characterizing CSO emissions, (ii) assessing the local dynamics and intensity of CSO impacts on the river and (iii) calibrating sewer and river water quality models being part of a planning tool for future CSO management in Berlin (Riechel et al., 2011). UV-VIS spectrometers are in-situ probes, which measure absorbance spectra ranging from UV to visual wavelengths. Concentrations, such as chemical oxygen demand (COD), are calculated from these spectra. Due to the varying composition of waste and river water a local calibration is required to enhance the measurement quality. According to Gamerith et al. (2011), manufacturer global calibration can lead to systematic error up to 50% for COD measurements.