Skip to main content

Data from: Long-term persistence of wildlife populations in a pastoral area

Cite this dataset

Kiffner, Christian et al. (2021). Data from: Long-term persistence of wildlife populations in a pastoral area [Dataset]. Dryad.


Facilitating coexistence between people and wildlife is a major conservation challenge in East Africa. Some conservation models aim to balance the needs of people and wildlife, but the effectiveness of these models is rarely assessed. Using a case-study approach, we assessed the ecological performance of a pastoral area in northern Tanzania (Manyara Ranch) and established a long-term wildlife population monitoring programme (carried out intermittently from 2003-2008 and regularly from 2011-2019) embedded in a distance sampling framework. By comparing density estimates of the road-transect based long-term monitoring to estimates derived from systematically distributed transects, we found that the bias associated with non-random placement of transects was non-significant. Overall, cattle, and sheep and goat reached the greatest densities and several wildlife species occurred at densities similar (zebra, wildebeest, waterbuck, Kirk’s dik-dik) or possibly even greater (giraffe, eland, lesser kudu, Grant’s gazelle, Thomson’s gazelle) than in adjacent national parks in the same ecosystem. Generalized linear mixed models suggested that most wildlife species (8 out of 14) reached greatest densities during the dry season, that wildlife population densities either remained constant or increased over the 17-year period, that herbivorous livestock species remained constant, while domestic dog population decreased over time. Cross-species correlations did not provide evidence for interference competition between grazing or mixed livestock species and wildlife species but indicate possible negative relationships between domestic dog and warthog populations. Overall, wildlife and livestock populations in Manyara Ranch appear to coexist over the 17-year span. Most likely, this is facilitated by existing connectivity to adjacent protected areas, effective anti-poaching efforts, spatio-temporal grazing restrictions, favourable environmental conditions of the ranch and spatial heterogeneity of surface water and habitats. This long-term case study illustrates the potential of rangelands to simultaneously support wildlife conservation and human livelihood goals if livestock grazing is restricted in space, time and numbers.


  • line transects driven along routes; same routes were used each season
  • conventional distance sampling was used to estimate season-specific densities
  • detection functions were estimaetd based on pooled observations

Usage notes

  • Densities_MR: contains seasonal density estimates of livestock and wildlife secies from 2003-2019; from 2003-2008 livestock species were not consistently recorded (therefore do not treat as zero but as NA)
  • Dspecies is the point estimate of the species; Lspecies is the lower 95% confidence interval of this estimate, and Uspecies the upper 95% confidence interval 
  • Densities_TME: contains density estimates and associated 95% confidence intervals (notation as above) of 12 wildlife species in four study areas of the Tarangire-Manyara ecosystem (MR: Manyara Ranch; CA: Mto wa Mbu Game Controlled Area; LMNP: Lake Manyara National Park; TNP: Tarangire National Park). Estimates and confidence intervals are based on point estimates from seasonal surveys carried out between 2011-2019.