Regeneration in natural and logged tropical rain forest : modelling seed dispersal and regeneration

Abstract

Regeneration and disturbance are thought to play key roles in the maintenance of the high tree species diversity in tropical rain forests. Nevertheless, the earliest stages in the regeneration of tropical rain forest trees, from seed production to established seedlings, have received little attention in population dynamical studies and in simulation models that are used to evaluate forest management in logged areas. The aim of the research described in this thesis was to gather information on the earliest stages of the regeneration of tropical rain forest trees, and to use this knowledge in a simulation model to study the long-term effects of natural and anthropogenic disturbance on the composition of the forest. To this end, the seed production, seed dispersal, seed fate and spatial distribution of seedlings were investigated in natural and logged tropical rain forest in central Guyana. Most species in central Guyana have large seeds. Species with large seeds appeared to be ‘seed limited’. They produced few, large seeds that were dispersed over short distances only and the majority of the seeds will fall within a few metres from the crown. The seeds of large-seeded species germinated very well in the dark understorey, but did not tolerate the high insolation in open logging gaps. As canopy openness increased their germination success decreased due to desiccation. The short seed dispersal distances largely seemed to determine the spatial distribution patterns of their seedlings, which were clumped around adult trees. Their distribution was not consistently related to light availability or disturbance history. Species with small-seeds are rare in Guyana and their seed production and dispersal could not be studied, but generally it is assumed that seeds of these species are widely dispersed. These species are more likely to be ‘site limited’. Their seeds germinated better in gaps with intermediate canopy openness than in the understorey or very large gaps. Their seedlings were clumped some distance away from adult trees in sites with intermediate light availability and in logging gaps, reflecting their germination requirements. Tree growth and yield models that have been developed for tropical rain forests rarely include detailed descriptions of regeneration, which may lead to an underestimation of the effect of disturbance. Therefore, a model that explicitly includes regeneration processes was developed, within the SYMFOR modelling framework, based on the collected information and data from a partner project. Test simulations of undisturbed forest with a length of 300 years showed that the forest composition did not remain stable in this model. However, simulations of a 25-year logging cycle suggested that the forest can rapidly transform into a secondary forest dominated by pioneer species with small seeds and light wood and with low yields per logging event. Despite the seemingly correct response to logging it was decided that the model should not be used in this form to analyse the long-term effects of disturbance on the functional group composition. Probably the accuracy of the functions in the model could be improved if more data were available, for more species and spanning longer periods.