Résumé
The sharing of habitat by wild and domestic animals may result in pathogen transmission, notably via ectoparasite vectors such as ticks. Interfaces between protected and communal lands constitute sharp transitions between areas occupied by host communities that are extremely contrasted in terms of composition, diversity and density. Empirical characterizations of tick communities and of their vertebrate hosts are strongly relevant for understanding the mechanisms leading to disease transmission between wild and domestic animals. In the present study we aimed at depicting the pattern of spatial variation in the density of immature ticks at such an interface located in Zimbabwe. At the end of the 2011 rainy season, we applied a hierarchical repeated protocol to collect ticks. We used the drag-sampling method in the vegetation surrounding water pans used by ungulates in 3 distinct landscape compartments (i.e. national park, mixed compartment and communal lands) characterized by a differential use by wild and domestic hosts. We combined generalized linear mixed models with site occupancy models to (1) assess tick aggregation levels at different spatial scales, (2) identify and disentangle factors which influence the density and probability of tick detection, and (3) compare robust estimations of tick densities among the landscape compartments. Ticks belonging to the Amblyomma and Riphicephalus genuses were found to be the most abundant. At small scale, ticks were more often detected in the afternoon and were more abundant close to water pans for Amblyomma and Riphicephalus genuses. Riphicephalus spp. density was also higher in grassland and bushland vegetation types as compared to woodland vegetation type. At large scale, for the three detected genuses, density was much higher near water pans located in the communal lands as compared to the national park and mixed compartment. Given that host community's diversity is much lower in the communal areas than in the two other landscape compartments, these results are compatible with a dilution effect but not sufficient to demonstrate this effect without additional studies. Up to date, it is the first utilization of these rigorous sampling and statistical modelling methodologies to estimate tick density in African ecosystem simultaneously at large and small scales. (C) 2014 Elsevier B.V. All rights reserved.