Final Activity Report Summary - RESES (Modelling the dynamics of resilience in coupled social-ecological systems) The resilience of social-ecological systems and their capacity to adapt to global or local changes are valuable system properties in face of global change. They determine the degree to which a system can flexibly react and cope with unforeseen events or gradual change without losing major functions, such as the delivery of ecosystem goods and services. It is therefore one of the principal aims of adaptive management to enhance the resilience of a given system. However, the dynamics of resilience, factors and mechanisms that determine resilience in a complex environmental management situation and approaches to manage for resilience are insufficiently known. This project developed agent based and equation based models of dynamic interactions between actors and an exploited resource system, namely water, fish or vegetation, to investigate the resilience of those coupled systems to short term disturbance, such as water scarcity or rainfall variability, and long term changes, such as increased variability of water flows. An agent-based model of water use in a semiarid river basin was utilised to test the impact of different governance structures and diversification of water use on the performance of the social-ecological system in face of high variability in water flows. A comparison of a centralised versus a fragmented governance system revealed that, under the given assumptions, the resilience of the system was higher under a centralised regime while the fragmented, local system could perform better in case actors were given the opportunity to diversify their water use. In the latter case the actors benefited from different dynamic responses of the exploited resources to variability in water availability, i.e. the buffer capacity of a fish population versus the direct reaction of the agricultural sector to short term water scarcity. In addition, a model of a rangeland system was used to test rising variance in a state variable as an indicator of the systems approach to a threshold and thus loss of resilience. It revealed that variance actually increased when the system was close to the threshold; however, the signal was not unambiguous and might be difficult to detect in a real system. Overall, the project revealed that a multitude of interacting factors could affect the resilience of a system on different scales. Resilience of a coupled social-ecological system was to a large extent determined by the behavioural responses of individual and collective actors to environmental and social change, as well as by their capacity to adapt institutions and resource use to changing conditions. Nevertheless, research on institutional dynamics and approaches for modelling change in collective behaviour and institutions was only at its beginning by the time of the project completion.