Mediterranean mountain landscapes feature pastoralism, and the maintenance of sheep and goat grazing is the main challenge for their sustainability. To deliver pastoralism ecosystem services, particularly in reducing fire risk amidst global change, it is essential to align the herds' size and impact with the current foraging status of the ecosystems in real time. This approach allows for more efficient herding, thereby maintaining the resilience of these landscapes. Understanding complex grazing processes at a landscape scale can now be aided by the various sensors and trackers available. The development of geolocation technologies makes it possible to combine information at different scales on time to adapt pastoralism more efficiently in the context of rural abandonment. In this paper, we analyze artificial intelligence-based IoT data from the daily routes of three flocks of goats and sheep over a year in the mountains of central Portugal. A principal components analysis of the temporal geolocalization of the herds discriminates the spatial scale of the territorial selection attributed to the herders and the animals, analytically confirming what has been described empirically by several authors. The results confirmed the relevance of these operational solutions for understanding complex landscape processes and efficiently supporting sustainable management. Finally, our result shows that these innovative tools are reliable for monitoring animal ecosystem services in Mediterranean mountain landscapes in support of policy decisions on their remuneration.