Bushclump-grass interactions in a south-east African savanna : processes and responses to bush control.

Abstract

The objective of this study was to investigate woody-grass interactions and the initial response of vegetation to bush control in the mesic Eastern Cape bushclump savannas. The occurrence of multi-species bushclumps, rather than single-trees, presented an interesting variation to an otherwise well-studied interaction. The effect of bushclumps on their local environment was characterized. Since all woody-grass interactions involve competition for irradiance, nutrients and moisture, a factorial experiment was designed to discriminate these individual and interactive effects. Mechanical and chemical bush control measures were investigated in a formal, replicated experiment. The herbaceous, woody and soil responses to bush control treatments, for the first two seasons, are reported. Bushclumps had a moderating effect on their microclimate when compared with the open grassland. Lower maximum and higher minimum temperatures, and higher humidity were the result of an 80-90% reduction in the irradiance regime. Soils beneath bushclumps were more fertile than grassland soils. The importance of bushclumps on sandier soils was discussed. Bushclumps were characterized by a sparse shade-tolerant herbaceous layer which contributed little to grazing capacity. An aspect effect increased grass production in the grassland on the south-facing side of bushclumps. Initial results suggest that the lateral spread of woody roots could be as far as 25 m. The factorial experiment tested the individual and interactive effects of irradiance (normal sunlight, 40% and 80% shade), nutrients (normal nutrient level, low and high nutrient addition levels) and moisture (low, normal and additional moisture levels) on the herbaceous layer. The interaction of 80% shade and high nutrients had a detrimental effect on herbaceous production. Deep shade did not affect herbaceous production, but Themeda triandra showed etiolated growth, aerial tillering, an increase in the number of leaves, and an increase in the proportion of stem under deep shade. The root mass of the herbaceous layer also decreased. This suggested that below-ground biomass production was impaired at the expense of maintaining aboveground biomass. The addition of nutrients significantly increased herbaceous production and resulted in a change in sward composition. Moisture was not an important factor in this experiment. Mechanical clearing in the bush control experiment resulted in a significant increase in herbaceous production. Panicum maximum colonized the ex-bushclump zone and contributed significantly to the increased production. Oversowing with Chloris gayana significantly increased grass yields. The two contrasting seasons revealed the importance of rainfall in affecting herbaceous production. The second season was characterized by lower soil fertility and a decline in grass quality. This was attributed to high grass production in the above-average rainfall season. A four-fold increase in woody stem density after two seasons demonstrated the coppicing ability of the woody layer once mechanically cleared. Most of the coppice occurred within the first season. Exceptional coppice growth characterized the second season. Acacia karroo recruitment was mainly from seed. Woody plants showed their susceptibility to chemical poisoning by dropping their leaves within the first season. Many of these individuals succumbed during the second season. Mortality was greatest in woody plants with a smaller basal circumference. Owing to the difficulty of accessing all woody stems in a bushclump, mortality in bushclumps was lower than that in the open grassland. Grass production in the bushclump and its periphery were significantly increased in both seasons. This was attributed to the increased productivity of mainly Panicum maximum which took advantage of the increased irradiance regime. Both the mechanical and chemical treatments displayed significantly greater grass production in the open grassland zone. This demonstrated the extent to which the woody layer had competitively dominated the herbaceous layer.