Phenological patterns in rain-fed pastures are strongly coupled to regional climate conditions, making these systems particularly susceptible to climate change. In line with this, shifts in annual plant growth cycles have been documented widely in the past decades. These changes in pasture phenology can significantly impact the ecological functions that underpin primary productivity and the delivery of key ecosystem services, with consequences for their management and the implementation of effective policies. Understanding phenology patterns and their driving factors is critical for identifying vulnerable areas, mitigating risks, and managing these agroecological systems to sustain their agricultural ecological functions. We analyzed regional pasture phenology patterns along a climatic gradient, using a space for time approach, from semi-arid to temperate climates. Using high-resolution remote sensing data, we found three distinct typologies of pastures displaying spatial continuity, based on variations in the timings of the growing season and productivity metrics. Machine learning ensemble modelling revealed that climatic variables determined phenology groups, particularly aridity and summer temperatures, alongside soil and topography controlling water availability. Non-linear relationships between phenology groups and their drivers suggest potential critical climate tipping points between pasture ecosystem states, namely, below an Aridity Index of similar to 0.8 and above maximum summer temperatures of similar to 29 degrees C. Our findings highlight the susceptibility of pasture ecosystems to climate change and their potential tipping points. This knowledge can guide about future climate impacts on pasture ecosystem services and inform adaptive management strategies and agricultural policies to enhance resilience.