Agreement between floristic and soil organic carbon isotope (13C/12C, 14C) indicators of forest invasion of savannas during the last century in Cameroon

In an area of savanna-forest (S-F) mosaic of Cameroon, at Kandara near Bertoua, an enclosed savanna bordered by young semi-deciduous forests was selected for detailed studies of vegetation an soil carbon isotope compositions with a view to estimating the rate of forest advance into savannas. Forest floristic composition and structure were analysed in small plots along two S-F transects and within two large stands. Tree species counts and basal area (BA) measurements gave convergent results defining (1) an edge forest with low BA values that forms an irregular strip parallel to the S-F border, (2) a large colonization zone zone composed of pioneer species (Albizia species, with individuals of very large diameter) and (3) a mature forest composed of abundant Rinorea individuals and large individuals of Triplochiton scleroxylon and Piptadeniastrum africanum. Carbon stable isotopes were determined from organic matter of soil profiles sampled at various depths in savanna, colonization-zone and mature forest. In the deep soil horizons (40-50 cm) of the colonization-zone profiles, delta C-13 values similar to those of the present savanna reveal the past existence of a large-tree savanna. In subsurface horizons (15-20 cm), delta C-13 values intermediate between those of savanna and mature forest prove the encroachment of the forest ecosystem on savanna. Using C-14 measurements, the mean residence time (MRT) of soil organic matter of these last horizons was determined with precision owing to the atmospheric C-14 pulse from nuclear bomb tests prior to 1964. By assuming an exponential age distribution of organic compounds and by taking account of MRT and remaining carbon from the savanna, the colonization zone was found to be 60-80 y old. The age of the colonization zone being the same near the present savanna and near the mature forest, it seems that the forest advance was probably not a linear process but would result from the coalescence of Albizia thickets born in savanna.