This thesis aims to unpack the complexity of trajectories of land use change in semiarid Mediterranean agro-ecosystems - illustrated using findings from the Torrealvilla catchment in south-eastern Spain. The research looks at multiple dimensions of land use change and addresses the past, present, and future . It comprises of three key stages. Stage 1 quantifies land use change between 1956 and 2008, assesses the impacts on landscape fragmentation, and identifies important drivers of the detected changes. Stage 2 examines the linkage between farmer heterogeneity and the diversity of agricultural land uses in the studied landscape by developing a typology of farmers. It also investigates how farmers respond to future changes that potentially affect their farming and describes potential future trajectories of agriculture in the area and the implications for agro-ecosystem services provision. Stage 3 models the future trajectories of land use change for the studied landscape using two different approaches: Cellular Automata (CA) and Agent-Based scenario analysis. The research shows that over the past five decades the catchment has undergone significant changes with pronounced effects on landscape configuration. The changes involved three major trajectories (forest expansion, abandonment of rain-fed farming, and intensification) and were attributable to various interacting biophysical and socioeconomic drivers. The study differentiates six types of farmers according to the characteristics of their households and of their farm management. Different types of farmers are expected to pursue different land use directions with important impacts on future provision of agro-ecosystem services. Results from the final stage highlight a diverging future . "'To illustrate, while CA projection suggests scope for further expansion of irrigated farming, the agent-based scenario analysis indicates the opposite. Overall the thesis confirms that land use changes in Mediterranean agro-ecosystems are indeed complex entailing multiple, contrasting, and perhaps competing trajectories and are the manifestation of the interacting effects of various types of drivers. The thesis highlights conceptual and practical challenges in undertaking an integrated evaluation of land use change complexity. Future policy making should consider the various factors that are jointly shaping agricultural land use trajectories.