Recent climate change research reveals a novel and significant trend: weather patterns at mid-latitudes, such as in temperate western Europe, are getting more persistent. With respect to rainfall, this means longer droughts, but also longer periods with excessive rain. No comprehensive study has hitherto investigated the ecological consequences of such regime shifts. Can ecosystems adapt, or will the alternation between drought stress and soil water saturation exhaust them? Will this select for communities with novel trait combinations and more volatile species dynamics? And will these novel systems still be robust in the face of further changes in the environment? This study explores the potential impact of the ongoing shift in the frequency of dry/wet cycles at multiple, connected levels of biological organization. It does so in a new, large-scale set-up at UAntwerp built in the framework of the developing European infrastructure for ecosystem research 'AnaEE'. The design simulates changes in rainfall and associated temperature changes in the open air, using a gradient with eight precipitation regimes so that non-linearity and tipping points can be discerned with great precision. The project scope ranges from plants to soil biota such as bacteria and fungi, and from metabolism and genetic regulation assessed with bioinformatics to ecosystem processes. This multi-scale approach explicitly acknowledges the interwoven nature of ecosystems, with knowledge of molecular and cellular changes being instrumental to mechanistically explain the whole-system-scale effects on productivity, greenhouse gas fluxes and biodiversity dynamics. Different experiments are planned each year: (i) year 1 features a gradient in alternating dry/wet cycles, from 1 to 60 days, across a full growing season; (ii) year 2 focuses on legacy effects and the importance of changes of soil communities; (iii) year 3 matches precipitation regimes to corresponding temperature regimes to study the impact of drought-associated warming (an important natural feedback that can greatly increase plant stress). A series of connected, hypothesis-driven measurements is carried out, which will be integrated using structural equation modelling (path analysis) and ecosystem modelling. The project team has successfully collaborated in the past, and the complementary expertise brought together here should yield both significantly increased understanding of key processes as well as new avenues to climate change impact mitigation.