influences of rainfall, seasonality, age and sex on ungulate abundance
Fluctuations in rainfall play a fundamental role in herbivore population dynamics in savannas, and should be evaluated and accounted for before the effects of other factors (such as predation) can be reliably revealed. All seven ungulate species showed either humped, sigmoidal or exponentially increasing relations with the accumulated late wet-, late dry-, or wet- and dry-season rainfall components. The aggregate and quarter-size to adult age classes of zebra were also most abundant at intermediate values of accumulated monthly rainfall. Excluding the newborn class, abundance was best correlated with accumulated monthly rainfall only for zebra; late wet-season rainfall for topi, giraffe, kongoni, impala and warthog; and annual rainfall for waterbuck. The relationship between aggregated abundance and rainfall characteristically reflected the relationship between the abundance of the adult age class and rainfall. This shows that demographic monitoring is necessary to reveal influences of variability in rainfall on immature ungulates. In general, the abundances of older animals were influenced more by past rainfall accumulated over longer periods than were the abundances of newborn calves.
The 1993, 1997, 1999–2000 droughts and the 1997–98 floods had devastating impacts on ungulate populations, decimating individuals directly through starvation, and indirectly by weakening them and hence amplifying their vulnerability to predation, diseases and parasites. Droughts and floods are often associated with outbreaks of disease, including epizootics of viral diseases in the Mara-Serengeti (Packer et al. 1999). As there was no veterinary monitoring, it is hard to assess the role of diseases; however, no mass die-offs or unusual numbers of carcasses were reported by the field team during the monitoring period. Nevertheless, extreme climatic events increase the susceptibility of animals to disease outbreaks, and probably did so during the monitoring period. As an example, mass mortality of impalas attributed to anthrax occurred in the Serengeti (S. Cleveland, pers. comm.), whereas massive livestock losses, human fatalities and high incidences of livestock diseases were reported elsewhere in northern Kenya and southern Somalia (Little, Mahmoud & Coppock 2001) and in Ngorongoro and Loliondo pastoral areas of Tanzania (Galvin et al. 2001) during the 1997 drought and the El Niño floods of 1997–98.
The distributions of mean densities across months suggested that some species engaged in small-scale seasonal movements, probably between the pastoral ranches and the reserve (cf. Stelfox et al. 1986). Feeding specializations probably determine the species involved in these movements, whereas rainfall seasonality determines the direction of movements. In the late wet season, surface water and high-quality food are widely available, but predation risk is elevated in the tall grass of low quality that predominates in the reserve at this time. This forces resident herbivores specializing on young and mature grasses such as topi, zebra and Thomson's gazelle (Murray & Brown 1993) to disperse to the pastoral ranches, where heavy livestock grazing keeps grass in a short, tender and nutritious state and maintains high visibility. In the dry season, when surface water and food availability are reduced, heavy livestock grazing in the pastoral ranches forces wildlife to disperse to the reserve, where dry-season fires and migratory wildebeest and zebra have removed the tall, rank grass and improved visibility. Thus heavy livestock grazing in the pastoral ranches facilitates some grazers during the late wet season, but competition with livestock in the dry season for food and surface water forces wildlife into the reserve, where they are facilitated by migratory wildebeest and zebra, which also absorb most of the predation pressure. Competition with migratory ungulates and livestock for food and surface water is expected to increase with decreasing rainfall and increasing density of livestock and migratory ungulates. Kongoni, a ruminant grazer specializing on tall grass (Murray & Brown 1993), were reduced in numbers in the dry season when migratory herbivores were present, but increased in numbers in the late wet season in the reserve. The number of animals involved in these movements probably increased after the mass die-offs of buffalo, Synceros caffer (Sparrman), in 1993/94 (Kenya Wildlife Service, unpublished reports), which greatly reduced the ability of residential ungulates to keep the grass short in the late wet season and hence reduce predation risk. Stelfox et al. (1986) described these seasonal movements in detail.
Past studies have shown the population growth of greater kudu, Tragelaphus strepsiceros, in Kruger Park, South Africa, to be dependent on wet-season rainfall (Owen-Smith 1990), and annual population changes in a number of African ungulates to be dependent on variation in dry-season rainfall (Mduma et al. 1999; Ogutu & Owen-Smith 2003; Owen-Smith & Ogutu 2003). Our results reinforce and extend these findings by revealing, for the first time, that accumulated monthly rainfall determines the abundance of newborn calves and that accumulated late wet-season rainfall primarily determines the abundance of older animals among the seven ungulate species we considered. Our results also demonstrate the necessity of demographic monitoring on fine spatial and temporal frames in order to reveal the full spectrum of effects of rainfall fluctuations on the abundance of African savanna ungulates.
Reduced rainfall during the dry season, which is projected to be exacerbated in the future on continental Africa (Hulme et al. 2001), will probably reduce ungulate populations. Recent analyses have highlighted species’ range shifts in both animal and plant taxa in response to climate change during the past century (Parmesan & Yohe 2003; Root et al. 2003). Such changes can decouple the contemporary relationships among ungulate species and create novel combinations of species (Hughes 2000). Future forecasts for continental Africa suggest substantial species losses and major reshuffling in community composition, due to the composite effects of climate change and land transformations (Thuiller et al. 2006). Repercussions caused by the indirect effects of rearrangement of mammal communities and changes in the patterns of interspecific interactions (Stenseth et al. 2002) may fundamentally transform the character of current species interactions and essential ecosystem processes in unforeseen ways (Walther et al. 2002).
Anthropogenic land-use and land-cover changes that reduce or preclude the potential for species to spread to future climatically suitable habitats may severely threaten plant and animal populations worldwide (Sala, Chapin & Armesto 2000). This may happen through extensive habitat alterations with the potential to generate huge gaps between suitable patches (Brown & Lomolino 1998; Tokeshi 1999), thus hindering free mobility of species to track climate change. The inverse relationship between range size and extinction probability (Gaston 1994) suggests that range contractions will probably amplify the risk of local extinctions of species, including ungulates (Thomas, Cameron & Green 2004; Thuiller et al. 2005).
Our results also have important management implications for African ungulates.
Reduced rainfall and rising temperatures in the Mara-Serengeti during 1989–2003 (Ogutu et al. 2007) diminished vegetation production and availability of surface water, accelerating habitat desiccation and intensifying human–wildlife conflicts. Land-use transformations, such as those occurring in the buffer zones of the Mara-Serengeti (Serneels, Said & Lambin 2001), accentuate effects of climatic change on local floral and faunal assemblages. Intensification of land use will probably amplify effects of climatic extremes on wildlife and their habitat, requiring coordinated management of the Mara-Serengeti and its agro-pastoral buffer zones. Oscillations in regional rainfall in equatorial East Africa (Rothe & Virji 1976; Ogallo 1984; Nicholson & Entekhabi 1986), and in the Mara-Serengeti in particular (Ogutu et al. 2007), suggest that, as a recurrent and persistent feature of this region, droughts should be incorporated into management plans and policies aimed at minimizing drought mortalities. Policies that promote mobility and flexible access of wildlife to resources, such as maintaining open dispersal and migratory routes, can mitigate impacts of recurrent droughts.
In conclusion, zebra were most responsive to monthly rainfall, suggesting that rainfall affected their movements. In contrast, for resident ungulates, changes in abundance were most strongly correlated with cumulative past rainfall, suggesting that rainfall affected their survival and reproduction. Herbivore responses to rainfall were apparently independent of the degree of breeding seasonality and synchrony. Hence topi that breed seasonally and synchronously followed a pattern very similar to impala that breed year-round. Similarly, warthog that breed synchronously followed a pattern very similar to giraffe, kongoni and waterbuck, all of which breed asynchronously. Rainfall affected food availability and quality more strongly than the availability of surface water for residents. Thus, with the exception of newborns, rainfall responses of the highly water-dependent waterbuck and warthog were similar to those of the less water-dependent kongoni, giraffe and impala. Furthermore, the similarity in rainfall responses of the pure grazers (kongoni, waterbuck, warthog and topi) to those of giraffe (a pure browser) and impala (a mixed feeder) suggests that the influence of rainfall on abundance was not substantially modified by the dietary guild of herbivores. Finally, excluding newborns, immature zebra, giraffe, kongoni, waterbuck and warthog showed responses to rainfall similar to those of fully grown animals. Only immature impala and topi showed contrasting responses to rainfall compared with fully grown animals.