Phenomics

Repeatability and resolution robustness of GPS-derived movement traits in Merino ewes across irrigated valley and summer rangeland environments

---

Abstract

GPS collars can provide phenotypes for genetic evaluation, but it is unclear which movement traits are repeatable, how much variation arises from collars, and how sensitive estimates are to logging interval. We evaluated GPS-derived movement traits calculated from ~8 consecutive 2-week trial periods in Merino ewes grazing irrigated valley pasture and summer rangeland. Merino ewes from the Rafter 7 flock (University of Nevada, Reno) wore i-gotU-120B collars logging at 5-min intervals. After cleaning fixes and removing extreme travel days ( >40 km), we derived three daily movement traits we hypothesized to be affected if a ewe is stressed: total distance travelled (km; sum of step lengths between consecutive fixes), radius of gyration (km; the root-mean-square distance of fixes from the day's centroid, describing how widely a ewe ranged around its core location), and minimum convex hull area (km²; the area of the smallest polygon enclosing all fixes for the day, a simple proxy for daily home-range size). Daily locations were used to classify ewe-days as valley or summer. The analysed dataset comprised 97 ewes and 6,906 ewe-days, with collars contributing data for an average of 10.4 days per deployment (SD 2.3). We fitted linear mixed models by environment and trait, including day (date) as a fixed effect and ewe (and, where relevant, collar) as random effects to estimate repeatability, and a bivariate valley-summer model for daily distance. To assess logging-interval effects, we recomputed distance after thinning 5-min fixes to 10, 15, 30, and 60 min and compared repeatability and correlations of ewe predicted effects (PE) with the 5-min baseline. Daily distance showed low-to-moderate repeatability overall (R = 0.19, SE = 0.01) and higher R in summer (0.24, SE = 0.01) than in the valley (0.16, SE = 0.02). Ewe effects for distance were positively correlated between valley and summer (r = 0.44, SE = 0.07). In summer, distance and radius had similar repeatabilities (both ≈0.24), whereas hull area was much less repeatable (R ≈ 0.04); in the valley, radius and hull showed little between-ewe signal. Ewe-level PEs for distance were strongly correlated with those for radius (r = 0.78) and hull (r = 0.84), indicating that these traits largely capture a shared movement dimension. Downsampling from 5 to 10-30 min increased repeatability for distance (R ≈ 0.27-0.28), and ewe PEs remained highly correlated with the 5-min baseline (r_PE ≥ 0.96 up to 30 min; 0.92 at 60 min). Taken together, our results suggest that daily distance is a promising, moderately repeatable movement phenotype for Merino ewes, with collar effects that are present but modest and quantifiable. Modest reductions in fix rate (to ≤30-min intervals) can substantially reduce logging burden while largely preserving ewe rankings for downstream genetic and resilience analyses.

Keywords: 2026

How to Cite:

Hess, A., Bergeron, J., Huber, S. & Cartagena, F., (2026) “Repeatability and resolution robustness of GPS-derived movement traits in Merino ewes across irrigated valley and summer rangeland environments”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2295677. doi: https://doi.org/10.31274/wcgalp.24358

Rights: 1