Contrasting elephants and humans as agents of disturbance in Miombo woodlands

Abstract

The Miombo woodlands are “disturbance driven savannas”, where people, elephants and fire prevent canopy closure and maintain tree-grass coexistence. Since the end of the 18th century, the woodlands have undergone dramatic changes to their disturbance regimes, through increases in human and decreases in elephant populations. This has resulted in a high concentration of human-driven disturbances in some areas of the woodlands and low disturbances in other areas. Consequently, in the areas affected by humans, there are high levels of biomass loss, while other areas experience woody plant encroachment. The ecological impacts of humans and elephants are comparable, but there are important differences. Although both remove woody biomass, they might impact vegetation composition and structure differently. The aim of this study was to investigate whether people can perform a similar functional role to elephants, with regards to biomass removal and potential changes to savanna structure and composition. Aboveground woody biomass estimates were obtained from Synthetic Aperture Radar backscatter and used to (1) compare biomass change at different levels of human (2010 human population density) and elephant influence (2009 Niassa elephant density); and (2) identify how the intensity of biomass loss differs for humans and historical elephant densities (1.4 elephants/km2) under different rainfall and initial biomass conditions. Biomass change was calculated between 2007 and 2010 and compared across equal areas inside and outside of Niassa National Reserve. For human influence, a regional analysis was also done. Published estimates of past elephant densities without poaching for Niassa National Reserve were used as the historical elephant density. Biomass loss increases with population density, for both people and elephants. However, rainfall responses show opposite patterns: biomass loss decreases with rainfall within Niassa National Reserve (under elephant influence) but increases outside of protected areas (under human influence). Under more mesic conditions, humans tend to remove more biomass than elephants. This suggests that woodlands with low rainfall and low human influence have more trees and higher biomass than if elephants were still present at their historical densities, and that some harvesting by people may be beneficial. People and elephants both remove woody biomass; however, they do not necessarily select the same stems. Therefore, species and size class preference of humans and elephants were compared, to identify potential differences in their impacts on vegetation structure and composition. I found that humans prefer a wider range of species than elephants, so might filter the canopy layer more strongly. Elephants’ preference for slightly larger stems suggests that areas utilised by elephants might result in a more left-skewed size-class distribution than humans in areas without a charcoal industry. The high regeneration ability of many species, the large overlap in stems used by people and iii elephants and the relative flexibility in what stems can be used for a specific purpose, suggests that many of these differences in impacts on composition and structure could be mitigated. While elephants and people clearly have different impacts on woody vegetation, people are a totally novel ecological and evolutionary force due to some similarities in how both use woody biomass. The type of human utilisation will determine the stems used and I suggest that since almost all aboveground woody biomass can be utilised for charcoal, the human impact on the woodlands could be reduced by decentralising the charcoal industry. Encouraging small scale production in low rainfall, low human impact areas, could alleviate the pressure on the hotspots of biomass loss. This could provide rural populations with an additional source of income, and the woodlands with a disturbance necessary in maintaining, to some extent, the ecosystem structure and functioning. However, care needs to be taken as increasing accessibility in remote areas could result in rampant deforestation. Results from this thesis suggest that forest management principals developed in other ecosystems, which haven’t had mega-fauna for thousands of years, need to be tailored to African ecosystems. African forest managers therefore need to develop policies that consider the past and present disturbance regimes of these ecosystems.