Factors Affecting Livestock Guarding Dogs’ Proximity to Their Herd and Association with Perceived Effectiveness

Abstract: Livestock guarding dogs (LGDs) are utilized worldwide as a non-lethal strategy to alleviate human-wildlife conflict. However, while studies show their effectiveness, resulting in reductions of livestock loss, there is limited research into the factors that influence individual dog variation. One factor which may be important is the proximity of the dog to the herd it is guarding, as well as subtle aspects of its behavior. This small-scale study investigates individual differences in proximity and behavior and the potential influencing environmental factors. It compares these to the dog’s perceived effectiveness, as rated by the herder who regularly works with them. We attached GPS collars to 8 LGDs and a ‘lead’ goat from their respective herd to track proximity during grazing visits to the Namibian bush. For 4 dogs, the data spanned 2 years, so we examined consistency of proximity, both within, and between, years. We also examined associations between temperature, time of day and vegetation density and the LGD’s proximity to the herd. In addition, behavioral observations were conducted for 3 of the dogs to compare to the data derived from the collars. Individual dogs varied in their proximity to the herd and showed consistency within (F(7,4180)=36.8, P < 0.001) and between the 2 years (r=0.99. P = 0.012). Proximity was significantly associated with time of day (F(1,4180)=19.3, P < 0.001), but not with daily temperature (F(1,4180) = 0.003, P = 0.96). Dogs tended to be closer to the herd at the start and end of the day. Whilst vegetation density did not have a universal effect (F(2,4180) = 0.99, P = 0.393), some dogs showed distinct patterns of change in proximity with differences in vegetation density (F(14,4180) = 3.60, P < 0.01). This suggests some dogs were modifying their behavior in response to relative predation risk. The herder tended to rate those dogs, which were recorded to be closest, as more effective. Behavioral observations of distance correlated well with the data gathered from GPS collars but, in most cases, the dog was observed to be closer to other herd members than to the collared lead goat. We conclude that measuring dog proximity to the herd is a useful way to monitor behavior, but that monitoring multiple sheep or goats, and also conducting behavioral observations is integral to understanding ways to maximize performance.