Ecosystems occasionally respond abruptly to small changes in external conditions. Such abrupt responses can represent critical transitions that occur at tipping points where the ecosystem shifts to an alternative regime. Despite the often profound consequences of tipping events, our ability to predict them is still limited. To meet this challenge, generic early-warning signals for critical transitions have recently been proposed. These signals are generic, because they may in principle work irrespectively of the specific mechanism responsible for the tipping, making their potential field of application very broad. However, they have so-far been mostly analyzed in simple models that neglect the high spatial and structural complexity that characterizes most ecosystems. The proposed project aims to help bridging this gap between complex reality and simple models by 1) analyzing how critical transitions arise in spatially and structurally complex ecological networks, and 2) investigating how this kind of critical transitions might be detected by generic early-warning signals. The anticipated results will be of more than academic interest, as the complex webs of interactions in nature are under increasing pressure and there is an urgent need to understand how we may estimate the risk of systemic collapses.