Ecology of a Grazing Ecosystem: The Serengeti

Abstract

Primary productivity and herbivory were studied in the Serengeti National Park, Tanzania, and Masai Mara Game Reserve, Kenya, during the annual cycle of 1974—1975, and wet—dry season transitions in 1976—1979. Basic state variables measured were aboveground plant biomass inside permanent and temporary fences, and outside fences. Productivity was calculated as the sum of positive plant biomass increments. Control productivity (cPn) was calculated from biomass dynamics inside permanent fences. Temporary fences were moved in concert with grazing by the region's abundant ungulates to estimate actual aboveground primary productivity (aPn). Primary productivity was highly stochastic with productive periods poorly synchronized even among nearby sites. Short—term productivities could be extremely high, exceeding 30 g°m—2°d—1. Grazing animals adjusted their densities in relation to grassland productivity. The average proportion of annual aPn that was consumed by herbivores was 0.66, with a minimum of 0.15 and a maximum of 0.94. Green forage was available everywhere late in the wet season in May but was available only at high rainfall sites in the northwest late in the dry season in November. By the end of the dry season, the residual plant biomass outside fences averaged only 8% of cPn. Nomadic grazers moved seasonally in response to grassland productivity. The growing season ranged from 76 d in low rainfall areas to virtually continuous in high rainfall areas. Annual cPn was linearly related to rainfall and averaged 357 g°m—2°yr—1 over the year and 1.89 g°m—2°d—1 during the growing season. Actual aPn was substantially greater than cPn at most sites, averaging 664 g°m—2°yr—1. Growing season aPn averaged 3.78 g°m—2°d—1. Grazing stimulated net primary productivity at most locations, with the maximum stimulation at intermediate grazing intensities. Stimulation was dependent upon soil moisture status at the time of grazing. Rain had a diminishing effect on primary productivity as the wet season progressed and plant biomass accumulated. Part of the stimulation of grassland productivity by grazing was due to maintenance of the vegetation in an immature, rapidly growing state similar to that at the beginning of the rainy season. Since grazers overrode rainfall—determined productivity patterns, aPn was more closely related to grazing intensity than to ranfall. Grazing was heavier on grasslands that were intrinsically more productive. Rate of energy flow per unit of plant biomass was much higher in grazed vegetation. Grazers ate green leaves almost exclusively during the wet season, but species composition of the diets of different grazers differed markedly. Diets of nomadic grazers were very different in the wet and dry seasons. Vegetation dried out rapidly at the onset of the dry season and dry plant tissues made up a substantial proportion of ungulate dry season diets. However, green forage commonly was more abundant in diets than in the vegetation. Grazing increased both forage quality and its rate of production. Zebras supplemented a high—bulk diet by eating the seeds of awnless grasses. The foraging patterns of different grazers were differentiated by several vegetation properties, including productivity, structure, and species composition, in a manner suggesting resource partitioning. The relationship between the stability of vegetation functional properties and community species diversity was positive in five of seven tests. Greater species diversity was associated with greater biomass stability through the seasons, greater resistance to grazing by a single species of ungulate in both the wet and dry seasons, and greater resilience after grazing. Species diversity was not associated with greater resistance to grazing by several ungulate species or to plant species extinction. Specific properties of trophic web members were identified that produced greater functional stability in more diverse communities. Fire does not appear to have important effects upon the functional properties of the grasslands except for a weak stimulation of productivity in the wet season immediately following dry season burning. Fire did have an important effect upon structural properties of the vegetation that would tend to regulate ungulate feeding. The ecology of neither the plants nor the animals in the Serengeti ecosystem can be understood in isolation; many traits of both suggest coevolution among trophic web members. The functional dynamics of the trophic web suggest that the acceleration of energy and nutrient flow rates due to intense herbivory has resulted in the development of an entire consumer food web due to additive fluxes rather than mere quasi—parasitic fluxes from plants to animals.