Spatial statistics to quantify patterns of herd dispersion in a savanna herbivore community


  • A. Stein
  • N. Georgiadis


Understanding the spatial distribution of species is a fundamental issue in ecology, yet quantitative descriptions of animal species’ distributions are rare. In this chapter, we use a spatialstatistics approach to describe the spatial distribution of herds of large herbivores in Laikipia, central Kenya. We used Global Positioning System technology and spatial point pattern analysis (F-, G- and J-functions) to characterise herd distributions of the 9 most abundant species comprising large herbivore communities in African savannas. F-function analysis is based on estimating the probability of a herd occurring within radius r of randomly selected focal points. G-function analysis is similar, but based on randomly selected focal herds. The J-function is derived from the ratio of G- and F-functions. Comparing results from the different functions was instructive about possible causes of spatial patterning at the landscape level. All species displayed consistently aggregated distributions under F- and J-function analyses, partly because wildlife has been displaced by humans and livestock from sections of the study area. By contrast, the G-function provides a description of dispersion under more natural conditions because areas lacking herds are excluded from the analysis. G-function results showed 5 species to display random or nearly random dispersion patterns (zebra, impala, Grant’s gazelle, eland and hartebeest), while the remainder were aggregated (African elephant, giraffe, African buffalo and Thomson’s gazelle). When data for all species were pooled, G-function results revealed an emergent property of this community: wild herbivore herds were arrayed across the landscape in a significantly regular fashion. Two possible causes of this pattern, invoking interspecific complementarity in habitat preference, or disaggregation by prey herds to counter predators, could not be distinguished. Both mechanisms may have been operating in savannas over such long evolutionary time that their effects cannot be separated without experimentation