The ecology of lianas and trees in tropical forest canopies

Tropical forests occupy 12% of the terrestrial surface area, contain ca. 25% of all terrestrial carbon, and recycle tens of petagrams (1015 gram) carbon in photosynthesis and respiration processes annually. These forests are hyper-diverse in tree and liana species, which drive carbon stocks and dynamics and create a large variation in structure, microhabitats and food items for other plants and animals. Over the past decades, several reports suggest that lianas are rapidly increasing in abundance and biomass relative to trees, with potential negative impacts for tree species reproduction, growth, and survival; and for ecosystem-level processes such as carbon accumulation. Despite these major consequences, the mechanisms underlying the increase of lianas in tropical forest are not understood. It is speculated that such dramatic, rapid changes result from an increase in light owing to an increase in natural and/or anthropogenic disturbances and to increasing droughts due to global warming. We defined the general question: “What determines the increase in abundance and dominance of lianas, relative to trees, in tropical forests?” With the work done in this thesis, we compared the form and functionality of lianas and trees, and their response to different resource conditions and seasonality, to provide information and elucidate some of the processes behind an increase in lianas relative to trees in some tropical forests. In this study, we focus on canopy lianas and trees because they are important for the structure and dynamics of the forest. The canopy is where most high-light interception and also high-water losses occur, affecting overall plant performance (i.e., gas exchange, photosynthesis, transpiration). Branches in the canopy are key for carbon gain and bottlenecks for water transport. Large lianas function fundamentally different than small saplings. We make use of two 50-m tall canopy cranes in two forests in the Republic of Panama, which provide a unique opportunity to study the environmental impacts of large canopy lianas and trees. These two forests have contrasting environmental conditions and are located at the extreme sides of the precipitation gradient extending across the Isthmus of Panama. The drier forest, located in Parque Natural Metropolitano near Panama City and the Pacific coast, has a precipitation of ca. 1860 mm.yr-1 and strong seasonality. The wetter forests, located in Bosque Protector San Lorenzo, has a precipitation of ca. 3200 mm.yr-1 and a weaker dry season. Both canopy cranes give access to ca. 0.91 ha of forest. For simplicity, Parque Natural Metropolitano (PNM) will hereafter be referred to as the drier forest and Bosque Protector San Lorenzo (BPSL) as the wetter forest. At each forest site, we selected a set of eight liana and eight tree species. These species differed between forests given the strong effect of precipitation and soil fertility on plant species distribution in the well spatially structured forests of Panama. Using the canopy cranes, we observed the differences in vegetative phenology between lianas and trees as a measure of plant performance under similar resource conditions and seasonality (i.e., light and water). We further measured plant functional traits at the leaf, stem, and individual level to assess differences in the form and functionality between lianas and trees (chapter three and four). Finally, we integrated observations of vegetative phenology and plant functional traits in a process-based plant growth model to quantify how differences in functional traits and plant structure drive the productivity patterns of lianas and trees. With the model, we also assess the effect of environment on the capacity of lianas and trees for carbon uptake, as a measure of productivity (chapter five). One commonly accepted idea behind the functionality of lianas is that lianas benefit from better access to soil moisture during dry conditions and thus take advantage of high-light availability due to low cloud cover. We tested whether this idea applied to both forests at the extreme opposites of the precipitation gradient in the Panama Canal Watershed (chapter two). For this, we explored the differences in performance between canopy lianas and trees and their response to seasonality by assessing their ability to add leaf area for 17 months, including one wet and two dry seasons. We followed leaf area production because it is a component of growth and is strongly linked to cambial dormancy and reductions in growth. We predicted that lianas have better access to available soil moisture than trees, particularly during seasonal drought. We further predicted that these differences in available soil moisture between lianas and trees are associated with a more rapid relative increase of leaf area on lianas than on trees during peak dry-season light conditions. Surprisingly, both lianas and trees converged in their ability to add leaf area over the entire study period for the drier forest. During the dry season, both lianas and trees were water limited as indicated by an observed decrease in predawn leaf water potentials (a proxy for available soil moisture) and leaf areas. Decreases in leaf areas are commonly associated with control of water stress and as a mechanism to maintain hydraulic integrity in plants. Contrasting with the drier forest, lianas in the wetter forest were indeed favored by a higher capacity to access available soil moisture while trees showed water stress, particularly at the end of the dry season. Observations from chapter two suggest that in forests with low precipitation and strong dry season, both lianas and trees suffer from reduced available soil moisture in dry periods. With increasing precipitation, a higher capacity to access available soil moisture favor lianas over trees since lianas can benefit more from dry season high-light conditions. The contrasting patterns in the performance of lianas and trees observed in chapter two suggested differences in functionality and their response to resource availability and seasonality. Therefore, we assessed the intrinsic differences between canopy lianas and trees in chapter three and four by exploring leaf and stem level functional traits and their relation with plant structure. Via the analysis of functional traits, we assessed differences in resource acquisition strategies. Specifically, in chapter three, we explored branch development, branch structure, and the costs of leaf display for lianas and trees to assess the efficiency of leaf display in the forest canopy. We observed that lianas are more effective than trees in exploring the forest canopy in the drier forest. This efficiency was indicated by the development of larger leaf areas per unit branch cross-sectional area and more slender stems of lianas. This efficiency is partially lost in the wetter, more light-limited forest. The main driver of the differences in leaf display between lianas and trees in the canopy of the drier forest was the construction cost of the leaf tissue, as indicated by differences in leaf mass per unit area (LMA). In the drier forest, lianas had a lower LMA than trees while no differences were observed in the wetter forest. We argued that for the drier forest, the allometry that favors exploring and intercepting light in the forest canopy contributes to the success of lianas observed for forests with relatively dry environments, whereas this advantage is partially lost in more shaded, wetter environments. In the fourth chapter, we focused on the differences of functional traits representative of the leaf economics spectrum (LES), wood economics spectrum (WES), and plant hydraulics, essential for overall plant economics and growth. We predicted lianas to be representative of species with a high resource acquisition strategy and trees to have a more conservative resource acquisition strategy. Among our main observations, lianas in the drier forest showed a more acquisitive strategy indicated by lower leaf construction costs, shorter leaf longevities, and relatively higher foliar nutrient content than trees. In the wetter forest, we observed a high non-systematic variation between species, and no differences between lianas and trees, suggesting a convergence in their functionality. Results from chapter three and four indicated that there are functional differences between lianas and trees. However, they do not quantify the implications of those differences for productivity and growth. To effectively understand the effects of differences in traits and structure between lianas and trees for carbon uptake (GPP – gross primary productivity) in response to light and water, we integrated the observed functional traits and branch structure into a process-based plant growth model in chapter five. We hypothesized that lianas have higher carbon uptake rates than trees, driven by their more resource acquisitive strategy and more efficient leaf display over stem support than trees; particularly so, during high-light seasonal drought. The model provided estimates for carbon uptake (GPP – gross primary productivity) in lianas and trees from both the drier and wetter forest. These estimates of GPP were standardized per unit leaf area and unit branch cross-sectional areas for comparison and to assess efficiency in GPP at the organ level. For the drier forest, we observed that GPP per unit leaf area was lower in lianas than in trees, suggesting that liana leaves are less efficient than trees in carbon uptake. However, a higher GPP per unit branch-crossectional area of the liana form suggested that lianas are overbuilt at the stem level. We found that the Huber value (HV, the ratio of branch-crossectional area to leaf area) and stomatal control drove these differences in productivity between lianas and trees. Lianas in the drier forests had lower HV than trees, that is, relatively larger leaf areas per unit branch cross-sectional area. Low HV pose hydraulic constraints in plants, given the higher rates of transpiration. It is strongly suggested that lianas compensate for higher rates of transpiration via higher conductivities of the xylem. For lianas in the drier forest, we observed that xylem conductivity did not completely compensate for the high transpiration costs of the leaves most likely because the water transported through the xylem was distributed over a large surface leaf area, and thus each individual leaf was hydraulically constrained. Interestingly, we observed that lianas maintained higher plant water status at midday, which suggest a stronger stomatal control, a common mechanism to maintain hydraulic integrity in plants but that limits carbon uptake; which explains the simulated low carbon uptake per unit leaf area from this study. Although liana leaves in the drier forest were less efficient in terms of carbon uptake than tree leaves, the relatively larger leaf areas per unit branch cross-sectional area indicated by their lower HV explain why lianas show higher GPP per branch cross-sectional area than trees, as the latter maintains a less optimal structure (HV) and functionality. The patterns observed for both GPP per leaf and branch cross-sectional area in lianas and trees from the wetter forest contrasted with the simulated patterns observed for the drier forest. Lianas and trees in the wetter forest showed very similar patterns of carbon uptake, mostly driven by their high similarity in form and functionality (chapter three and four) at the branch and leaf organ level. In this thesis, we showed that there may be a gradient in form and functionality between lianas and trees. The differences between lianas and trees may be strongly driven by resource availability. In the drier forest, the high acquisitive strategy of the liana form is constrained due to hydraulic stress in the dry season. However, with increasing available soil moisture in the wet season, lianas may use their acquisitive strategy and perform better than trees. Over longer periods, this acquisitive strategy may explain why lianas show relatively higher growth patterns than trees in forests with lower precipitation and strong seasonality. In contrast, with increasing precipitation, lianas and trees converge in their form and functionality, most likely driven by changes in resource availability. We show that physiological and morphological traits of lianas and trees depend very much on resource availability, such as light, water, and potentially, soil nutrient supply. Lianas benefit from high-light conditions during drought in relatively wet forests. In dry forests, lianas and trees are under the effect of water stress as indicated by lower leaf water potentials and a decrease in crown development. However, lianas may benefit from the high availability of light and soil nutrients in drier forests, where lianas adopt a more acquisitive strategy that allows them to rapidly colonize the canopy and display their leaves more effectively than trees. Thus, lianas may achieve higher carbon uptake by their exposed branches. In wetter forests, lower light levels constrain lianas in taking a more acquisitive strategy, probably in combination with the lower levels of soil nutrient supply. There, lianas are very similar to trees in form and functionality. This thesis indicates that the physiological, morphological, and growth responses of exposed canopy branches of lianas versus trees strongly depend on both climate and soil conditions and questions the generality of the commonly expected positive effects of drought on lianas and associated implications for the carbon and water cycle of entire forests.