Excessive grazing by livestock is claimed to displace wild ungulates in the Trans-Himalaya. This study compares the seasonal diets and habitat use of sympatric wild naur Pseudois nayaur and domestic goat Capra hircus, sheep Ovis aries and free-ranging yak Bos grunniens in north Nepal and analyses their overlap both within and across seasons. Alpine meadow and the legumes Oxytropis and Chesneya were critical resources for all animal groups. High overlap occurred cross-seasonally when smallstock (sheep and goats) in summer used the spring and autumn ranges of naur. Relatively high total ungulate biomass (3028 kg km−2) and low recruitment of naur (56 young per 100 adult females in autumn) suggested interspecific competition. The spatio-temporal heterogeneity in composition and phenology of food plants across the steep gradient of altitude, together with rotational grazing, appears to indirectly facilitate coexistence of naur and smallstock. However, owing to high cross-seasonal (inter-seasonal) overlaps, competition is likely to occur between these two groups at high stocking densities. Within seasons, naur overlapped more with free-ranging yak than with smallstock. As their habitat use and diets were most similar in winter, when both fed extensively on the same species of shrubs, naur was most likely to compete with yak during that season.

Assessing the abundance and distribution of mammalian carnivores is vital for understanding their ecology and providing for their long-term conservation. Because of the difficulty of trapping and handling carnivores many studies have relied on abundance indices that may not accurately reflect real abundance and distribution patterns. We developed statistical analyses that detect spatial correlation in visitation data from combined scent station and camera-trap surveys, and we illustrate how to use such data to make inferences about changes in carnivore assemblages. As a case study we compared the carnivore communities of adjacent communal and freehold rangelands in central Namibia. We used an index of overdispersion to test for repeat visits to individual camera-trap scent stations and a bootstrap simulation to test for correlations in visits to camera neighbourhoods. After distilling our presence-absence data to the most defensible spatial scale, we assessed overall carnivore visitation using logistic regression. Our analyses confirmed the expected pattern of a depauparate fauna on the communal rangelands compared to the freehold rangelands. Additionally, the species that were not detected on communal sites were the larger-bodied carnivores. By modelling these rare visits as a Poisson process we illustrate a method of inferring whether or not such patterns are because of local extinction of species or are simply a result of low sample effort. Our Namibian case study indicates that these field methods and analyses can detect meaningful differences in the carnivore communities brought about by anthropogenic influences.