Convergent grazing responses of different‐sized ungulates to low forage quality in a wet savanna

Abstract

We studied how grazing intensity by small and mid-sized ungulate grazers varied with nutritional quality and grass species composition in wet oligotrophic tallgrass savanna of coastal northern Tanzania. Average grazing intensity was low (3-15% by cover), and most grass species were scarcely used by herbivores. Only two grasses, Panicum infestum and Digitaria milanjiana, had nitrogen and phosphorus concentrations that were consistently above the minimum levels (e.g. nitrogen concentrations <7 mg g −1 ) required by the three commonest grazers, Bohor reedbuck, waterbuck and wildebeest. The best predictors of grazing intensity were cover of P. infestum (the most abundant grass, with a mean cover of 15%) and canopy height of ungrazed vegetation. Models did not contain separate predictors for nutritional quality, presumably because quality varied mainly at the grass species level and therefore was fully represented by the variable 'cover of P. infestum'. Given that the three grazers differed greatly in body size and muzzle width (parameters known to influence nutrient requirements and the ability of grazers to feed selectively at the smallest spatial scale), we expected there to be strong resource partitioning that would be detectable in terms of grazing strategies and feeding sites. However, apart from minor differences in canopy height, greenness and diameter of grazed patches (albeit consistent with our expectations), feeding stations of the three grazers were similar and strongly dominated by P. infestum. We conclude that the low quality of herbage in wet oligotrophic savannas restricts foraging choices, which produces a characteristic yet impoverished grazing community that exhibits only limited resource partitioning.

Herding strategies under shifting rainfall conditions: implications for rangeland conservation and pastoralist livelihoods

“…Four caged plots of 1 m 2 each were established within an area of 50 x 100 m of the eight sites, two in each ward named Nainokanoka (highlands), Olbalbal (lowlands) and Ngorongoro and Endulen (midlands), and one in each village (a total of n = 32 exclusions per season). The selection aimed at a proportional sampling of grazing lawns within the open grassland defined by < 5% tree and shrub cover and devoid of large bushes and trees of > 4 m height (Stähli et al, 2015). Grazing lawns are distinct grassland community type, characterized by short-stature and with their persistence and spread promoted by grazing (Hempson et al, 2015).…”

“…All herbaceous biomass was cut to ground level. The clipped materials were collected in paper bags, left to dry until a constant weight was reached (<3 d) (Stähli et al, 2015) and then weighed using a digital balance scale.…”