Can off-river water and shade provision reduce cattle intrusion into drinking water catchment riparian zones?

Cattle defecation into rivers and overgrazing of riparian zones, are major concerns for drinking water catchment managers. Behaviour modification has been proposed instead of fencing for managing impacts, but reported success varies. Our study aimed to resolve whether provision of off-stream water and shade on real working farms could reduce the likelihood of cattle entering watercourses feeding Sydney, Australiaʹs primary water supply, Lake Burragorang (34° S, 150° E). Cattle herds (1.4 and 11 Animal Units ha−1) at two sites were fitted with Global Positioning System (GPS) collars (n = 12). Cattle movements were tracked following installation of industry-recommended off-stream water and shade (twelve 2 week duration control + treatment experiments). Some statistically significant differences in movement (Mann–Whitney U Prob. < 0.0001) were observed, as judged by comparisons of riparian, water trough and shade NEAR distances, and riparian zone visit number, duration and frequency. But effect magnitudes were small, inconsistent between different experiments, and insufficient to justify widespread water and shade provision. These findings contrast with the marked reductions in riparian impacts reported for rangeland pastured (>≈1 km2) cattle, but were not inconsistent with smaller scale grazing studies. Statistically significant correlations (Spearman R) were, however, observed (Goulburn_1, Robertson_2 experiments respectively) between the movement of cattle within the same herd (0.94, 0.85), cattle in adjacent fields (0.7, 0.64), and heat stress related factors (temperature, light, humidity, wind) (0.1–0.5) indicating GPS tracking was sound and other factors more strongly influenced animal location. We hypothesize that our off-stream water and shade did not markedly influence cattle movement because our paddocks were relatively small (1.5 and 20 ha) compared to rangeland pastures. The studyʹs main limitation was that GPS error prevented differentiation of riparian zone interaction from full stream contact. We recommend in future using direct video to overcome this, and differential quantification of these impacts.