Large herbivores and process dynamics in a managed savanna ecosystem

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)




Samuel J. McNaughton


Nitrogen cycling, Savanna, Ungulates, Phosphorus

Subject Categories

Biogeochemistry | Physical and Environmental Geography | Systems Biology | Terrestrial and Aquatic Ecology


Feedbacks between soil nutrients, plant communities, and large mammalian herbivores were studied at the Mpala Ranch and Research Centre in Laikipia, Kenya. The landscape consists of short-grass glades, typically 0.5-1.0 ha in size, dispersed throughout the dominant Acacia bushland vegetation. I examined (1) whether glades are created through the use and abandonment of overnight cattle corrals or 'bomas', (2) the importance of these glades as habitat for native and domestic ungulates, (3) feedback effects of herbivores on aboveground herbaceous productivity (ANPP) and soil nutrient dynamics, and (4) effects of climatic variability on herbivore abundance and soil-grass-grazer interactions.

All soil characteristics measured across a boma-glade chronosequence indicate glades are indeed derived from abandoned bomas. In particular, soil N, P and organic matter quality in the surface (0-15 cm) layer were similar for glades and 30-39 year old bomas, but were enriched relative to surrounding bushland. Soil texture was similar for bomas, glades, and bushland, indicating glades were not derived from a unique parent material. Cynodon leaves from bomas and glades were highly enriched in P, Ca and N relative to Cynodon from nearby bushland sites. Local abundance of impala, zebra and eland was closely tied to the distribution of nutrient-rich glades. Seasonal analyses of impala habitat selection suggested that selection for glade habitat was related both to predation risk and the availability of mineral-rich forage. In particular, P in boma and glade grass was above recommended levels for growing and lactating ruminants, while P content of bushland grass was lower than recommended levels.

Across a soil nutrient gradient, large herbivore consumption rates were linearly related to ANPP. A fertilization experiment and analyses of grass N:P ratios indicate that N and P co-limit productivity on the most nutrient-poor sites. Grazing pressure was consistently high (>60% of ANPP) at all but one site in a dry year (1999), and was greater in nutrient-rich glades (73.0 ± 4.2% of ANPP) compared to nutrient-poor bushland sites (42.7 ± 6.7% of ANPP) in a wet year (2001). Nitrogen budgets constructed for nutrient-rich and nutrient-poor sites showed that large herbivores themselves caused a net N input to the former and a net N loss from the latter. During short (1 month) growing seasons, grazers reduced aboveground productivity regardless of soil nutrient availability. However, over a 5 month growing season, grazers increased ANPP on glades and suppressed ANPP on bushland sites. Grazers increased the size of the inorganic N pool available to plants at onset of the growing season, but also decreased N mineralization rates at all sites early in the growing season. N availability measured via ion-exchange resin bags showed that the net effect of grazers was to increase in N availability at glade sites, but to decrease in N availability at bushland sites. This net effect on N availability mirrored grazer effects on ANPP in the high-rainfall year. Overall, results indicate that grazer effects on N balance and N availability in ecosystems are closely linked to effects on productivity and resilience to drought, and that both soil fertility and climatic variation mediate grazer effects on ecosystem processes.