Testing a general approach to assess the degree of disturbance in tropical forests

Questions: Is there any theoretical model enabling predictions of the optimal tree size distribution in tropical communities? Can we use such a theoretical framework for quantifying the degree of disturbance? Location: Reserve of Yangambi, northeast region of the Democratic Republic of Congo. Methods: We applied an allometricmodel based on the assumption that a vir- tually undisturbed forest uses all available resources. In this condition, the forest structure (e.g. the tree size distribution) is theoretically predictable fromthe scal- ing of the tree crown with tree height at an individual level. The degree of dis- turbance can be assessed through comparing the slopes of the tree size distribution curves in the observed and predicted conditions. We tested this tool in forest stands subjected to different degrees of disturbance. We inventoried trees >1.3 m in height by measuring the DBH in three plots of 1 ha each, and measured tree height, crownradius and crownlength in a sub-sample of trees. Results: All tree species, independently of the site, shared the same exponents of allometric relationships: tree height vs tree diameter, crown radius vs tree height, crown length vs tree height and consequently crown volume vs tree height, suggesting that similar trajectories of biomass allocation have evolved irrespective of species. The observed tree size distributions appeared to be power laws (excluding the finite size effect) and, as predicted, the slope was steeper in the less disturbed forest (?2.34) compared to the most disturbed (?1.99). The difference in the slope compared to the theoretical fully functional forest (?2.65) represents the metric for assessing the degree of disturbance. Conclusions: We developed a simple tool for operationalizing the concept of ‘disturbance’ in tropical forests. This approach is species-independent, needs minimal theoretical assumptions, the measurement of only a few structural traits and requires a low investment in equipment, time and computer skills. Its simple implementation opens new perspectives for effectively addressing initiatives of forest protection and/or restoration