Search

Understanding herbivore selection and utilization of vegetation types is fundamental to conservation of multispecies communities. We tested three hypotheses for how ungulate species select their habitats and how this changes with season: first, resources are distributed as a mosaic of patches so that ungulates are also distributed patchily; this distribution reflects habitat selection, which changes with season, the different ungulates behaving differently. Second, resources become scarcer in the dry season relative to those in the wet season. If interspecific competition prevails, then all species should show a contraction of habitats chosen. Third, if predation is limiting, competition will be minimal, and hence, habitat selection by herbivores will not differ between seasons. We used frequencies of occurrence in four common vegetation types in western Serengeti National Park to determine selection coefficients and utilization patterns and Chi-square analysis to test the hypotheses. The results showed that selection changes differently in each species, agreeing with the first hypothesis. Herbivores did not all become more selective, as predicted by the competition hypothesis, nor did selection remain the same across seasons, as predicted by the predation hypothesis. These results can be useful in constructing habitat suitability maps for ungulate species with special conservation needs.

Accurate detection of individual animals and estimation of ungulate population density might be a function of vegetation cover, animal size, observation radius or season. We assessed the effect of these factors on estimates of detection probability and density using five ungulate species in Western Serengeti National Park, Tanzania. Estimates were derived from information collected using ground surveys involving line transects targeting three resident species (impala, topi and buffalo) and two migrants (wildebeest and zebra) and analysed using DISTANCE, MANOVA, t-test and Pearson correlation. Results showed that ground surveys that take observation radii of 100 m would appreciably estimate at least 80% of the available ungulates. Beyond 100 m radii, surveys would leave approximately 43% of individuals undetected, the reason being a substantial influence of animal size, vegetation cover and observation radius on the detection. Animal size and observation radius have interactive effects. On their own seasonal differences, they do not have any effect but in interaction with animal size have significant effects especially on the migrant species. As reliable estimates of detection and density are required for making reasonable inferences, we urge that surveys using DISTANCE approach should consider incorporating both ground and aerial survey methods and ensure adequate sample replication. Resume La detection individuelle precise d'animaux et l'estimation de la densite de population d'ongules peuvent dependre du couvert vegetal, de la taille des animaux, du rayon d'observation et de la saison. Nous avons evalue l'effet de ces facteurs sur la probabilite des estimations de detection et sur la densite en utilisant cinq especes d'ongules dans l'ouest du Parc National de Serengeti, en Tanzanie. Nous avons tire les estimations des informations recoltees par des reconnaissances au sol impliquant des transects en ligne visant trois especes residentes (impala, topi et buffle) et deux especes migratrices (gnou et zebre), analysees avec DISTANCE, ANOVA, test-t et coefficient de correlation lineaire. Les resultats ont montre que les reconnaissances au sol qui optent pour un rayon d'observation de 100 m estimaient assez correctement au moins 80% des ongules disponibles. Au-dela de 100 m, les reconnaissances ne pourraient detecter quelque 43% des animaux, la raison etant une influence substantielle de la taille des animaux, de la couverture vegetale et du rayon d'observation sur la detection. La taille des animaux et le rayon d'observation ont des effets interactifs. En soi, les differences saisonnieres n'ont aucun effet mais, en interaction avec la taille des animaux, elles ont des effets significatifs specialement sur les especes migratrices. Puisqu'il faut des estimations fiables de detection et de densite pour faire des deductions raisonnables, nous insistons pour que les reconnaissances utilisant une approche DISTANCE envisagent d'integrer des methodes de reconnaissance aerienne et au sol et s'assurent de pouvoir repliquer correctement l'echantillonnage.