Over the last decades there has been a substantial increase in the research on critical transitions, and more recently on their potential early warning signals. Most of this work focusses on modelling the transitions and in this way trying to detect various ways of predicting them. To validate the early warning signals predicted by models, this PhD project will focus on experimentally testing whether the early warning signals found in models occur in real life. To do this, we will focus on three different systems. First we will use a simple bistable chemical system to test various noise regimes and look at flickering. Second, we will use photobioreactors to create single and multiple species systems. Here we will also look at various noise regimes and furthermore explore the effect of species interactions. Last we will perform experiments in mesocosms placed in a ditch to explore if we can detect early warning systems in a complex environment.