Search for a command to run...
• Using Global Positioning System (GPS) tracking 22 cattle herds at 5-minute intervals from August 2011 to August 2015, we collected 16993 daily grazing orbits and used them to quantify pastoralist mobility in East Africa. • We modeled each grazing orbit using a double Gompertz nonlinear regression model, revealing outward, resource, and return phases. • Using Partitioning Around Medoids (PAM) clustering, we identified four groups that characterize distinct resource acquisition strategies among pastoralists. • The defined versus exploratory grazing trips reveal adaptive strategies, linking herd mobility patterns to vegetated landscapes. • Pastoralists balance predictable grazing units with exploratory trips to maximize forage uptake while minimizing energy expenditure. GPS collar data collected from cattle herds provide detailed movement metrics that reveal how pastoralists in southern Ethiopia adapt to fluctuating resource availability. While daily grazing orbits are increasingly well documented, their relationship to grazing units remains poorly understood. Grazing orbits represent herd movement trajectories, while grazing units denote distinct vegetation patches where grazing occurs. This study explores how orbit typologies derived from GPS collar data correspond to grazing units, operationally defined as vegetation clusters distinguished by abundance, transition dynamics, and terrain accessibility. We tracked 22 cattle herds at 5-minute intervals from August 2011 to August 2015, generating 16,993 full-day grazing orbits. The study objectives are threefold: (a) classify daily grazing orbit types into groups reflective of resource use strategies using GPS collar data and nonlinear modeling; (b) assess how herd size, sites, individuals, and seasonality influence grazing orbit usage; (c) reclassify grazing units from published datasets based on orbit typologies to evaluate correspondence with vegetated landscapes across wet and dry seasons. We identified four distinct orbit types, each linked to grazing units and representing resource acquisition strategies. These findings highlight the potential of GPS-based orbit typologies to refine grazing unit classifications and improve understanding of pastoralist adaptation. We developed an operational framework that requires minimal data input and has high potential for transfer to satellite imagery, such as Landsat-8 and -9 and Sentinel-2. Future work should expand the routine by accessing the full data archive and exploring aggregated spectral-temporal metrics, rather than relying on a single-date satellite imagery.