Liana effects on biomass dynamics strengthen during secondary forest succession.

Secondary forests are important carbon sinks, but their biomass dynamics vary markedly within and across landscapes. The biotic and abiotic drivers of this variation are still not well understood. We tested the effects of soil resource availability and competition by lianas on the biomass dynamics of young secondary tropical forests in Panama and assessed the extent to which liana effects were mediated by soil resource availability. Over a five-year period, growth, mortality, and recruitment of woody plants of ≥1 cm diameter were monitored in 84 plots in 3-30-year-old secondary forests across the Agua Salud site in central Panama. Biomass dynamics and the effects of lianas and soil resources were examined using (generalized) linear mixed-effect models and a model averaging approach. There was strong spatial and temporal variation in liana biomass within and across the plots. The relative biomass of lianas had a strong negative effect on overall tree growth, growth of understory trees decreased with soil fertility and dry season soil water content, and the effect of lianas on tree mortality varied with soil fertility. Tree recruitment was not associated with any of the predictor variables. Our model indicates that tree biomass growth across our landscape was reduced with 22% due to competition with lianas, and that the effect of lianas increased during succession, from 19% after five years to 32% after 30 years. The projected liana-induced growth reduction after 60 years was 47%, which was consistent with data from a nearby site. Our study shows that the observed liana proliferation across tropical forests may reduce the sequestration and storage of carbon in young secondary forests, with important implications for the carbon balance of tropical forest landscapes and consequently for global climate change. Our study highlights the need to incorporate lianas and soil variables in research on the biomass dynamics of secondary forest across tropical landscapes, and the need for well-replicated longitudinal studies to cover landscape-level variability in the relevant abiotic and biotic components.