Linking foraging behaviour of free-ranging, lactating beef cows with diet quality and weight gain in semi-arid rangeland

Monitoring free-ranging livestock foraging behaviour and health with on-animal sensors has emerged as a potential means to enhance adaptive management for ranching operations. We evaluated the use of GPS collars collecting animal locations at 5-min intervals and estimating activity (grazing, walking, or stationary) via a 3-axis accelerometer to quantify foraging behaviour of lactating Bos taurus beef cows with calves on a ∼7600 ha working ranch in a sagebrush grassland ecotone in northeast Wyoming. We used this sensor data to quantify five metrics of foraging behaviour at a daily time step including (1) mean velocity while grazing (VG), (2) mean grazing bout duration (GBD), (3) mean turn angle while grazing (TAG), (4) total daily grazing time (TTG) and (5) total daily travel distance (TD), and related these metrics to measurements of cattle diet quality, weight gain, and stock density. While foraging behaviour metrics varied among individual cows, we found that daily estimates of VG, GBD, TAG and TD were all significantly related to stock density and variation in remotely-sensed estimates of dietary crude protein content. Over longer time periods of weeks to months, cattle diet quality and weight gain were significantly related to VG (positive correlation; R² = 0.42 – 0.87), and GBD (negative correlation; R² = 0.58 – 0.78). These findings indicate that ranchers have the ability to influence diet quality and animal performance via (1) the rotation of herds among pastures of varying composition and quality, and (2) changes in herd size relative to pasture size (i.e., animal density). Furthermore, our results indicate there is substantial potential utility for near-real-time monitoring of foraging behaviour as an indicator of animal performance via the combination of GPS tracking, accelerometer data, and a method to wirelessly transmit data to the internet. However, operationalizing such a system will likely depend on continuing improvements in sensor durability and data management efficiency, and reductions in sensor and data transmission costs.