Abstract Characterisation is used for quality assurance of effluent released into the environment. A different characterisation approach was in this thesis used for more sensitive and ecologically relevant whole effluent assessment. In paper I, organisms from a constructed food chain, including short term tests with bacteria, micro and macro alga, fish embryos and a subchronic test with a copepod (Nitocra spinipes), were exposed to two conventionally and four additionally treated (biofilter, ozone and membrane bio reactor) sewage effluents. Chemical analyses of substances targeted for removal by the additional treatments were also performed. In paper II, a life cycle test with N. spinipes was performed for one conventionally and five additionally treated (active carbon filtration, two different doses of ozone, moving bed biofilm reactor and irradiation with ultraviolet light combined with hydrogen peroxide) effluents and combined with matrix population modelling. The results from the biological tests in paper I could conclude that all additional treatments created effluent with somewhat fewer adverse effects compared to the conventionally treated effluents, however, few effects were found at lower concentrations (<50 % effluent). The chemical analyses showed that ozonation was the most effective treatment for removing the targeted substances. The subchronic test with N. spinipes showed adverse effects at 100% effluent. In paper II, the conventionally treated effluent again showed the most adverse effects, but in this test already at 3 % effluent. The ozonated effluents, and in particular the lower dose ozone, resulted in population level effects most similar to the reference. The individual level endpoints of development over time, however, showed significantly different effects for the ozonated effluents compared to the reference, which did not show an impact in the modelling. In conclusion, combining the biological and chemical approaches in paper I was valuable since they focussed on different aspects of the effluents. Comparing the subchronic test from paper I with the life cycle test from paper II showed that the life cycle test was >30 times more sensitive for a similar effluent tested with the same organism. The population modelling in paper II added value to the characterisation as integrating the endpoints from the life cycle test in the model reduced the uncertainty associated with analysing endpoints individually.