Effects of predation and habitat on a desert rodent community in Mahazat as-Sayd Reserve, Saudi Arabia

Abstract

The effects of mammalian predators and habitat on rodent abundance, species composition and diversity were assessed in Mahazat as-Sayd Reserve, in west central Saudi Arabia. Rodent capture rates were related to predator density, predator abundance, season and 22 measures of habitat quantity and quality. Rodent body weights and the proportion of rodents in reproductive condition were compared between three areas of high predator density and a predator-free area to indicate whether rodents were suffering indirect predation effects. To further explore the relationship between habitat and rodent abundance, vegetation was experimentally removed and then replaced at a site, and changes in rodent use were quantified.

Five trapping areas were used. Three (Areas A, B and C) had high densities of mammalian predators (mainly red and Ruppell's foxes). The fourth (Area E) was inside a 400 ha fenced exclosure from which mammalian predators were permanently excluded. The fifth area (D) was severely overgrazed habitat approximately 1 km south east of the Reserve, where mammalian predator density was low. Between November 1996 and October 1997, rodents were live-trapped at four different times of year, at each of 19 sites within the five areas. Habitat data were collected from each site during trapping. Rodent abundance and diversity were compared among the areas with differing densities of mammalian predators. Rodent body weights, and the proportion of the rodent population in reproductive condition, were compared between the areas of high predator densities and the predator-free area. Body weights, indices of relative of bipedalism (IRBs), and habitat use were compared among three rodent species, to investigate reasons for differential predation. Macro-habitat and micro-habitat preferences of three rodent species were determined. In an additional experiment in the predator exclosure, rodents were live-trapped at three treatment and three control plots during four periods; before and after vegetation was removed, before and after vegetation was replaced, and capture rates were compared.

Five rodent species were captured: Meriones crassus Sundevall 1842, Gerbillus nanus Blanford 1875, G. cheesmani Thomas 1919, G. henleyi de Winton 1903 and G. poecilops Yerbury & Thomas 1875. Overall, G. nanus comprised the majority of captured individuals (38.3%, with 122 captures of 80 individuals), followed by M. crassus (34.9%; 141 captures of 74 individuals), G. cheesmani (20.6%; 62 captures of 42 individuals), G. henleyi (4.8 %; 10 captures of 10 individuals), and G. poecilops (1.4 %; three captures of three individuals). M. crassus, G. nanus and G. cheesmani were the main focus of the study because of their higher capture rates. Mammalian predators significantly reduced the abundance of M. crassus and G. cheesmani, but not G. nanus, indicating that they were altering the composition of the rodent community. The Shannon's diversity index of rodent species was not influenced by mammalian predators. Mean weights of M. crassus and G. nanus were lower, and fewer male rodents were in reproductive condition, in the areas of high predator densities than in the predator-free area, indicating that rodents were suffering indirect predation effects. Hence, it is likely that both direct and indirect predation effects contributed to differences in rodent abundance among areas. Of the 23 habitat variables (including season), nine were significant in explaining rodent abundance in a General Linear Model. The significant variables were: trapping session (season), distance from trap to nearest refuge, overall % cover, % silt, % sand, % rock, % tall shrub cover, tall shrub greenness and seeding scores. Each species had different habitat preferences. However, a habitat manipulation experiment failed to show predicted effects on rodent habitat use. Habitat appeared to influence rodent species diversity, indicating that mechanisms of co-existence driven by differential habitat use may be operating in Mahazat as-Sayd Reserve. Implications of my results on the rodent community of Mahazat as-Sayd Reserve are discussed.