The role of fire and elephants in shaping a Central African forest-savanna mosaic

This thesis aimed to investigate the ecology of a forest-savanna transition in Central Africa, with a specific focus on how it was affected by fire and elephants. It had three objectives, each of which was addressed through a research article. The data presented in these articles were collected at forty forest-savanna transition sites in the forest-savanna mosaic of Lopé National Park, Gabon. The objective of the first article, entitled "Grass species flammability, not biomass, drives changes in fire behaviour at tropical forest-savanna transitions", was to empirically describe the fire suppression threshold and the interactive abiotic and biotic controls on it. Results from this work showed fires to become less intense and less hot as they approached the forest-savanna transition. These changes in fire behaviour were induced by a change in grass species composition, not by changes in grass biomass. Approaching the forest-savanna transition, the grass community shifted from one dominated by a fine, dry, highly flammable species, to one dominated by a thick-culmed, slower drying species. This shift was accompanied by a shift to higher levels of tree canopy cover. This article showed how fire was mediated by grass species composition, which was mediated by tree canopy cover, which was in turn mediated by fire. It thus elucidated on the mechanism that allows forest and savanna to co-exist in tropical mosaicked landscapes. The objective of the second article, entitled "Distinct transition tree community suggests possible antiquity of forest-savanna mosaics", was to assess whether a distinct transition tree community could be identified at the forest-savanna transition, and to investigate its ecology. Results from this work showed that a distinct woody community existed at the forest-savanna transition. This community was intermediately fire resistant, with a suite of functional traits suited to survive at the interface of forest and savanna, rather than in the interior of either. The species in this transition community were not forest gap species forming a community representing one successional stage, but rather showed signs of being a stable community that had experienced multiple recruitment events. The presence of such a unique woody community at the transition suggested that this mosaic, and the fire regime that maintained it, are ancient. The objective of the third article, entitled "The role of forest elephants in shaping tropical forest-savanna co-existence", was to quantify forest elephant use of the forest-savanna transition, assess the drivers of this use, and determine the consequences for transition dynamics. Results from this work showed that the forest-savanna transition was an important foraging site for forest elephants. Forest elephant foraging here was strongly seasonal and driven by fruit availability. Unlike savanna elephants, forest elephants foraged relatively non-destructively, and had little direct effect on woody vegetation structure. However, their frequent use of the transition created substantial paths, which acted as fire breaks during savanna burning. Inadvertently, forest elephants thus protected the fire-sensitive forest interior from burning. This thesis concluded that fire is the main determinant of forest-savanna transition dynamics and vegetation distribution in mosaics, and likely has been for millennia. Although forest elephants did not have the ability to create and maintain an open canopy in the same way as fire, their foraging at the forest edge reinforced the fire-mediated feedbacks that allow forest and savanna to co-exist in this landscape.