Modelling the long-term impacts of selective logging on genetic diversity and demographic structure of four tropical tree species in the Amazon forest

In the last three decades, Amazon tropical forests have experienced high rates of deforestation, both by clearing for agriculture and by logging. In this study, we use computer simulations to examine the potential effects of forest logging on genetic diversity and demographic recovery (basal area development) of four neotropical tree species over a time frame reflecting multiple logging events. The study species, Bagassa guianensis Aubl., Hymenaea courbaril L., Manilkara huberi (Ducke) Chevalier, and Symphonia globulifera L.f., are all taxa which are commonly exploited for timber in the Brazilian Amazon. The simulations were parameterized using empirical data from field studies in the Tapajós National Forest, Pará State, Brazil, including genotypes at microsatellite loci, demography, ecology and growth for each species. Eight scenarios, combining two different cutting cycles and two minimum cutting diameters, were examined for each of the four species. The scenarios represent the actual forest practices used in Brazil and French Guiana (cutting diameter 45 and 60cm, and cutting cycle of 30 and 65 years, respectively). Logging scenarios were applied for six cutting cycles, and final genetic and demographic data were compared to baseline data from corresponding control scenarios. At the end of the simulated period the basal area was strongly reduced under all conditions in B. guianensis, H. courbaril, and M. huberi. In only two scenarios was a species able to recover its basal area following logging (S. globulifera with both 45- and 60-cm cutting diameters under a 65-year cutting cycle). In the logging scenarios, all species showed a loss of alleles and genotypes and an increased genetic distance (calculated between each population at the start and the end of the simulations). These effects were higher under the most intensive logging cycles (30 years, 45cm). However, effective number of alleles, expected and observed heterozygosities, and the fixation index were little affected by the logging simulations. Over all, we conclude that, even under very optimistic conditions for growth and recruitment, current logging practices are not sustainable in terms of basal area. Our simulations show that different species respond differently to logging, both demographically and genetically. No single set of logging parameters can be applied to the forest as a whole. Rather, forest management practices must be species-specific, taking into account not only growth parameters but also ecological and reproductive variables, in order to move toward long-term forest sustainability.