Factors influencing thespatial distribution of dry forests in a forest-savanna mosaic in West Africa

Hennenberg Klaus ¹,Goetze Dethardt ¹, Fischer Frauke ², Porembski Stefan ¹

¹ Institute ofBiodiversity Research, Department of Botany, University of Rostock;

² Department ofAnimal Ecology & Tropical Biology, University of Wuerzburg

The forest-savannamosaic of the Comoé National Park region (CNP, Ivory Coast, transitionzone between Guineo-Congolian rain forest and Sudanian woodland) includes dryand wet forest islands as well as gallery forests. A digital elevation model ofthe investigated area shows that the region is characterised by a shallowrelief with slope inclinations normally less than 3%. The more elevated sitesshow shallower soils with coarser and more skeletal soil material than theslopes and the valleys.

Dry forest islandsmostly are located at the more elevated sites in the relief, which can be shownby satellite images and GPS mapping. On the valley slopes, the dry forestsadjoin to tree dominated savanna types which are linked with grass savanna oroccasionally wet forest types in the temporally flooded valley bottoms.

Thefollowing question arises: Why are dry forests restricted to the elevated sitesof the hills with less favourable soil conditions, and why does savannadominate on the slopes?

To answer thisquestion, eight transects of a length of 300m each and a width of 10 (-50)mwere installed in the undisturbed area of the CNP. The transects are located ateight independent forest islands along the ecological gradient between forestand savanna. This gradient is characterised by a higher daily maximum of airtemperature in the savanna and of relative humidity in the forest. Unexpectedlysuch clear edge effects are not apparent for the mean values. Treeindividuals of all size classes have been mapped and will be controlled regularly.Especially Anogeissus leiocarpus is of crucial importance atthe forest border. Composition and density of vegetation and site parametersare being analysed. Investigations of biomass (grasses, herbs and litter) showthat grass biomass is clearly correlated to fire events. A modelling approachhas been developed in order to test the relevance of different factorsinfluencing the position of dry forests in the investigated area.

Our study isembedded in the project BIOTA (Biodiversity Transect Analysis),funded by the German Federal Ministry of Education and Research (BMBF).