Long-term, livestock-mediated redistribution of nitrogen and phosphorus in an East African savanna

1. The effect of livestock on African rangelands has been a major focus of recent research, but little attention has been paid to the way livestock affects the distribution and availability of soil nutrients. In East African savannas, overnight containment of livestock in thorn-scrub corrals or ‘bomas’ concentrates large quantities of nutrients into small areas, potentially altering the landscape distribution of nitrogen (N) and phosphorus (P) in soils and plants. 2. This study was designed to (i) measure the density, turnover rates and soil nutrient concentrations of abandoned cattle bomas on nutrient-poor rangeland in central Kenya; (ii) determine whether long-term glades dominated by Cynodon plectostachyus are derived from abandoned bomas; and (iii) evaluate the effect of cattle bomas on the landscape-level distribution of N and P. 3. In the study area, glades (> 39 years old) averaged 0·71 ha in size and occurred at a density of 0·71 km − 2 . Abandoned bomas (1–39 years since abandonment) averaged 0·39 ha and occurred at a density of 1·21 km − 2 . During 1961–2000, no glades reverted to bushland vegetation, while 53 bomas were abandoned. 4. All characteristics of soils measured across a boma–glade chronosequence indicated glades were indeed derived from abandoned bomas. 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 significantly enriched relative to surrounding bushland. In contrast, at 40–65 cm depth beneath bomas, glades and bushland, soil N was similar. The texture of surface soils from bomas, glades and bushland was similar, indicating glades were not derived from a unique parent material. 5. Leaves of C. plectostachyus from 12–24-year bomas and long-term glades were enriched in P, calcium (Ca) and N relative to leaves of Cynodon dactylon from nearby bushland sites. In particular, P in boma and glade grass was above recommended levels for growing and lactating livestock, while P content of bushland grass was lower than recommended levels. 6. Cattle management via bomas exerts a greater effect on the distribution of P relative to N within the landscape. For cattle grazing an area of 20–25 km 2 boma − 1 , an estimated 0·24–0·30 g N m − 2 year − 1 is removed from the rangeland and deposited into bomas. Within 1·5 years of boma abandonment, 70% of this N is already lost from the manure and upper soil layer. Permanent N loss does occur via leaching, but the majority is probably volatilized and redistributed in rainfall. N deposition in rainfall (0·43 g N m − 2 year − 1 ) is more than sufficient to offset losses due to cattle grazing and deposition in bomas. In contrast, P deposited in bomas is more tightly retained, creating small P-enriched ‘hotspots’ while causing a permanent loss of the order of 0·021–0·026 g P m − 2 year − 1