LAUSR

Quantifying biomass and carbon stocks in tropical forests is essential for the implementation of climate change mitigation strategies, including the mechanism for reducing emissions from deforestation and forest degradation. Scientists have reached agreement on the use of a general allometric biomass equation combined with height-diameter allometry adjusted to local non-destructive data. However, tree allometry varies between species and functional strategies on the one hand, and between forest types and environmental conditions on the other. The aim of this PhD thesis was to study the variations in tree allometry among coexisting tree species and between sites and forest types, in order to improve estimations of biomass and carbon stocks in Central Africa. At the species scale, variations in height-diameter and crown dimensions-diameter allometries were reported among the 45 coexisting tree species in a representative site of semi-deciduous Celtis forest in northern Congo. The relationships between architectural traits (total height and crown dimensions) derived from species-specific allometries and functional traits highlighted a continuum of species between large-statured canopy species and small-statured understory species. The large-statured species tended to be light-demanding, wind-dispersed, deciduous and major contributors to forest biomass, while the small-statured understory species tended to be shade-tolerant, animal-dispersed, evergreen and most abundant in terms of stem density. At the site scale, we showed the variations in height-diameter allometry between forest types and, for the first time in Central Africa, variations in crown-diameter allometry between forest types. These trends were also confirmed for species common to different forest types, suggesting an environmental control in tree allometry. From the site-specific allometries, architectural attributes (total height and crown dimensions), in addition to the more classic structural attributes (basal area) and composition attributes (wood density), were derived and related to spatial variations in biomass and carbon stocks. This PhD thesis provided new data needed for the future biomass maps that will support forest carbon monitoring in Central Africa.