Structure comparée du peuplement avien dans trois sites de forêt primaire en Guyane

Abstract

The bird communities of three rain forest areas of southern French Guiana, two of them “virgin” and the third hunted, are described. The sampling procedure involved 408 one hour, 250 m long, transect counts randomly distributed over a 70 km²surface area, and seven 3-days periods of mistnetting along similar 250 m lines. The 361 resident species identified certainly did not represent the whole bird community, as the species/area curves would not reach an asymptote without a 2-fold increase of the sampling effort. The overall calculated density ranged from 1,226 to 1,615 birds/km². Mistnet lines gave a mean hourly catch in the low undergrowth similar to the results of the transect counts, but the proportions of the species were different. The total biomass (fresh weight) reaches at least 70 to 150 kg/km². The three more important guilds were, by number of individuals, the insectivores (54-65 %), the omnivores (16-20 %), and the frugivores (13-15 %) ; if these guilds were ranked by weight, the results was reversed (15-25 %, 24 %, and 38-50 %, respectively), because of the large size of many frugivores. About 1 % of the species were common, 10 % regular with moderate densities, the rest being occasional to accidental species together making up half the total of the species present. 79 to 85 % of the species appeared in less than 5 % of the sample counts, although each count included on average 20-25 birds of 10-13 species. Consequently, the diversity index was high. Species densities were lognormaly distributed and taxonomic diversity was high (72 genera per 100 species). The differences in species richness, turnover rate, relative abundance, population structure, density and total biomass were greater between sites than within site plots, and higher than expected from habitat distribution — which means that vegetation alone could not account for the differences observed. Frugivores had a bimodal vertical distribution 74-82 % of their biomass being found at ground level and in the canopy ; insectivores were the dominant guild in the intermediate layers. More than half the consuming biomass was found in the canopy, but the maximum number of species per volume unit of vegetation was found in the low undergrowth. Most of the species were randomly dispersed along the transect lines and widely spaced : for one third of those recorded, successive contacts were more than 23 km apart, on average, and for only 2 % of the species the mean spacing was less than 1 km. Most species were habitat or food specialists, and 135 were recorded in mixed feeding flocks. Hunting pressure significantly reduced species richness, diversity, density and biomass. Not only game birds, but also raptors and insectivores were affected, probably through complex interactions and the disappearance of top predators. Changes in community structure included a greater abundance of the more common species, and the scarcity if not the disappearance of rare species. The 3-fold decrease of the biomass of frugivores might have a detrimental effect on the seed dispersal of some trees, and therefore lead to a longterm modification of the forest structure. Bird biomass, but not density, is significantly higher in tropical than in temperate forests. The higher species richness of the former is therefore due to a tighter species packing, with a lower mean density of individuals. If the size of a self-supporting, isolated population is of 1,000 breeding pairs, i.e. about 5,000 individuals, then one third of these primary forest species would need a reserve of more than one million hectares to survive according to the estimated densities, and without allowing for the distributional gaps and the seasonal movements of unknown extent which also occur.