Patterns of tree-fall and branch-fall in a tropical rain forest in French Guiana

1. Tree-falls and branch-falls of all trees with a diameter at breast height (d.b.h.) of over 10 cm were studied in 12 ha of primary rain forest at Nouragues, French Guiana. Consequences in terms of the number of damaged trees and number and sizes of resulting canopy gaps were investigated. Spatial distribution of tree-falls and spatial relation with pre-existing canopy gaps were analysed. Forest turnover rates were calculated on different bases. 2. Of the 6314 trees initially present in the plot 1.5% had fallen and 1.3% lost at least one major branch after one year. In total, 4.9% of the standing basal area had been felled or damaged. 3. The major cause of tree-fall and tree damage is falling trees and limbs: 75% of the fallen and 83% of the damaged trees were felled or damaged by others ('domino-wise' felled or damaged trees). Of the remaining ('initiating') fallen trees, 38% were uprooted and 62% were snapped-off above the ground. Similar percentages were observed for domino-wise felled trees. 4. The spatial distribution of the 38 trees which initiated events was random. Preexisting canopy gaps did not increase the chances of surrounding trees to initiate a tree- or branch-fall. However, initiating tree- and branch-falls were found clustered at around 50 m from large gaps, while few gaps occurred at around 25 m from old gaps. 5. Some 35% of the fallen trees did not create canopy gaps, and 42% of the damaged trees were not involved in the creation of a canopy gap. Initiating tree-fall that resulted in a canopy gap had no larger d.b.h. than initiating tree-fall not resulting in a canopy gap. In contrast, gap-creating branch-fall trees had a larger d.b.h. than those that did not create gaps. 6. Gap size increased with the number and basal area of the trees involved. Gap size measured following the Brokaw (1982a) definition was not very relevant in quantifying the direct consequences of gap creation, and a modified gap definition is proposed. 7. Both the forest turnover rate based on the number of fallen trees and the turnover rate based on the area of canopy gaps underestimate the actual disturbance regime of the forest. To quantify the disturbance regime of a forest, the basal area of the annually fallen and damaged trees deserves more attention.