| dc.description.abstracteng | Vascular epiphytes are non-parasitic plants that germinate and grow on trees without contact to the soil. Their arboreal life style implies a strong dependence on forest structure and dynamics. Tree architectures change constantly during ontogeny, and large and old branches of the inner crowns are generally more suitable for epiphyte colonization and survival than small branches of the outer crowns. In addition, microclimatic conditions within canopies, such as light, temperature or humidity, are directly influenced by forest structure. While the influence of such gradients within trees and forests on the vertical distribution of epiphyte species is undisputed, our understanding of the relationship between epiphyte distribution and functional traits is limited. Moreover, a causal relationship between the dynamics of trees and forests and the dynamics of epiphyte assemblages is obvious, but our quantitative knowledge on this topic is strikingly scarce.
In this thesis, I provide a detailed analysis of how forest structure and dynamics influence the structure and dynamics of epiphyte assemblages and their functional traits via both field studies (chapters 2 and 3) and modelling studies (chapters 4 and 5).
In chapter 2, I analyzed vertical gradients of ten leaf traits based on leaf samples of > 1100 individuals belonging to 83 epiphyte species in a Panamanian lowland forest. I found that community mean trait values of many leaf traits were significantly correlated with height above ground. Trait-height correlations were particularly strong for specific leaf area (SLA), leaf thickness, leaf chlorophyll concentration and carbon isotope ratio. Both linear and non-linear trends were observed, and while the leaf thickness, for instance, increased linearly, the SLA decreased non-linearly with height. Furthermore, intraspecific trait variability was pronounced and accounted for one-third of total observed trait variance. Intraspecific trait adjustments along the vertical gradient were common and seventy per cent of all species showed significant trait–height relationships. In addition, intraspecific trait variability was positively correlated with the vertical range occupied by species; however, this correlation was rather weak. I also observed significant trait differences between major taxonomic groups (orchids, ferns, aroids, bromeliads) that were linked to their vertical distributions. Orchids, for instance, had on average the thickest leaves and lowest SLA values, while in ferns the leaf dry matter content was almost twofold higher than in the other taxonomic groups. My study represents the most comprehensive study on vertical trait gradients of vascular epiphytes to date and demonstrates that leaf trait syndromes and intraspecific trait variability play important roles in explaining the vertical zonation of vascular epiphyte species and taxonomic groups.
In chapter 3, I addressed the role of forest dynamics on community structure and mortality patterns of epiphyte assemblages by exploring the forest floor as source of information. To this end, I surveyed fallen branches and epiphytes in 96 transects in rainforests in Brazil and Panama. I found that trends in epiphyte abundance, richness and composition over branch diameter on the forest floor reflected trends in the forest canopy. This finding suggests that forest floor surveys provide useful demographic information, particularly on epiphytes occurring on the thinnest branches which are least accessible with the most common techniques to access the forest canopy. Furthermore, the density of epiphytes on the forest floor was high, and I estimated mortality rates of at least 4% per year at the community level, and of ~13% per year when considering epiphytes on branches < 10 cm in diameter. The results of this study highlight the importance of tree and forest dynamics for the demography of vascular epiphytes.
In chapter 4, I developed a dynamic forest stand model in which trees are represented by their three-dimensional (3D) structure. In this model, tree species were characterized by a set of leaf traits under consideration of trade-offs and correlations among traits. Applying the principles of the pipe model theory, these leaf trait trade-offs were scaled to whole-tree growth. This approach reproduced fundamental life history variation between different functional tree groups with regard to their growth, survival, and light demand. For instance, species with high SLA values had high initial growth rates, but lower maximum heights and shorter life spans, i.e. characteristics associated with pioneer species. Tree growth patterns in my model were largely consistent with observations and support the notion that the growth-survival trade-off across tropical tree species is, at least partly, determined by leaf traits. Furthermore, I coupled the trait-based tree model with a forest stand model which simulates key demographic processes and integrates between-tree competition. This stand model successfully reproduced a number of important ecological patterns. A dynamic equilibrium state was reached after ~ 100 years, and in this equilibrium twelve important forest attributes (e.g. above-ground biomass, basal area, stem number, net-primary production or leaf area index) were within typical ranges of Neotropical lowland forests. Moreover, complex patterns like the vertical leaf area density or the diameter-height relationship closely matched observations, indicating that a structurally realistic forest can be simulated with my model. To my knowledge, the presented modelling approach allowing detailed 3D long-term simulations of forest dynamics is unique and paves the way for further model-based analyses of forest dynamics, or model-based studies of canopy-dwelling organisms requiring a detailed representation of forest structures and their dynamics.
In chapter 5, I developed the first mechanistic model for epiphytes which explicitly simulates population dynamics while being coupled with a structurally-realistic forest model. This epiphyte model is three-dimensional, spatially-explicit, and trait- and individual-based. After the model was validated by comparing model results with field data, I used simulation experiments to assess how differences in natural forest dynamics, logging strategies, and the size of forest patches influenced the long-term dynamics of epiphyte assemblages. Tree turnover rates in natural tropical rainforest typically vary between 1% and 3% per year, and such variations had a marked impact on epiphyte assemblages, i.e. forests with low tree turnover rates had considerably lower extinction rates and higher epiphyte abundances. It has been observed that even in mature forests with low tree turnover rates, epiphyte assemblages show no sign of saturation, and my simulations demonstrated that the saturation level was clearly influenced by forest dynamics. Furthermore, logging had the hypothesized negative effect on epiphyte diversity and abundance. Strikingly, a slight reduction in size of logged trees from 45 to 40 cm in diameter at breast height had a catastrophic effect on epiphyte assemblages and resulted in nearly complete extinction. In contrast, epiphyte extinction rates decreased with increasing forest patch sizes. The coupled epiphyte-forest model presented in this study provided valuable insights on how forests stand parameters influence epiphyte assemblages and has the potential to address pending question in the field of epiphyte ecology and conservation in future studies.
In summary, the findings of my thesis represent a major advance towards a better understanding of the relationship between forest structure and dynamics and (trait) structure and dynamics of epiphyte communities. My thesis constitutes the most comprehensive study on the community trait structure of vascular epiphytes to date and introduced complex mechanistic models to the field of epiphyte ecology. The modelling approaches open new avenues for future studies of spatial and temporal dynamics of vascular epiphyte assemblages while integrating epiphyte research in a more theoretical context. | de |