Skip to main content

Modelling lianas as key drivers of tropical forest responses to climate change

Periodic Reporting for period 3 - TREECLIMBERS (Modelling lianas as key drivers of tropical forest responses to climate change)

Reporting period: 2018-04-01 to 2019-09-30

Tropical forests are essential components of the earth system and play a critical role for land surface feedbacks to climate change. These forests are currently experiencing large-scale structural changes, of which the most apparent may be the increase in liana abundance and biomass. Lianas, as structural parasites of trees, strongly compete with trees for both above and belowground resources. The strong competition from lianas for resources reduces tree growth, reproductive output and recruitment, increases tree mortality and alters the relative allocation to stem and leaf biomass. Lianas therefore have a strong impact on whole-forest respiration, carbon sequestration and residence time, and ongoing liana proliferation has a potential high impact on the future carbon and water cycle of tropical forests.

State-of-the-art global vegetation models have problems to realistically simulate the carbon cycle of tropical forest. Improvement of these models is of major importance to better inform society and political actors on the impact of climate change. Currently, a major source of uncertainty in the global vegetation models is their poor representation of vegetation demographic processes. We are convinced that modeling the ecosystem demography in tropical forests is only possible by accounting for lianas. Nevertheless, no single terrestrial ecosystem model currently includes lianas.

By building the first vegetation model that includes lianas, TREECLIMBERS will generate important insights into the mechanisms by which lianas influence the carbon balance of tropical forest ecosystems. We will make the first integrative study of (1) the contribution of lianas to instantaneous carbon and water fluxes, (2) liana contribution and influence on canopy structure, (3) their role for long term demographic processes, and (4) of their role in forest responses to drought events. To reach this challenging objective, TREECLIMBERS will develop the first liana plant functional type (PFT) based on a unique global meta-analysis and integrate this in the Ecosystem Demography model (ED2), a forerunner of the next generation of vegetation models. By using a model-data fusion framework TREECLIMBERS will for the first time integrate the large amount of available and new collected data on liana ecology. New data collection is focused on important knowledge gaps: (1) characterization of belowground competition for water between lianas and trees using stable water isotopic techniques; and (2) characterizing the extent of the contribution of lianas to the vertical canopy structure, using innovative terrestrial LiDAR 3D forest structure measurements, involving an important methodological effort to identify lianas in LiDAR observations.
During the first half of this project, a first simplified operational model was build, which means that an initial liana plant functional type (PFT) was implemented in the ED2 model. This model is currently still in testing phase and already simulating stable results. There are of course still many improvements to make to this initial model, which will evidently be the focus of the rest of the entire project.

The liana PFT was developed based on a meta-analysis of existing data and collection of targeted new data. In this first period of the project, several extensive field campaigns have been performed in French Guiana and Panama. Data has been collected in view of model parameterization and model validation: Liana inventories have been made at key sites, including quantification of their interaction with trees. Liana leaf traits have been measured, including nutrient compositions and leaf gas exchange measurements. Sap flow sensors have been installed in the footprint of a flux tower to quantify contribution of lianas to whole-forest respiration.
Moreover, specific data collection has been made to gain a mechanistic understanding of important processes: (1) Dual stable water isotopic measurements have been performed and have led to important new insights in the below-ground competition for water between lianas and trees. (2) Terrestrial LIDAR scans (TLS) have been collected at several sites, including a liana removal experimental site in Panama, and have led to methodological advances on the use of TLS for liana extraction. Ongoing research will lead to a better understanding of the vertical canopy structure and the impact of lianas therein.
Several advances have been made beyond the state of the art at the onset of this project:

- Lianas have for the first time been implemented in a dynamic global vegetation model. With ongoing model development continuing as planned, we expect the project will result in a fully operational model including lianas, which will serve as an integration framework for liana data at multiple scales. We foresee it will be possible to make a quantitative assessment of the impact of lianas on the forest carbon and water cycle.

- Important new insights have been gained in belowground competition for water resources from this first period. These results will be implemented in the model. Further model developments will also focus on hydrological processes, and we expect we will be able to characterize the contribution of lianas to water fluxes. Furthermore, we aim to assess the impact of long term drought on the water use of lianas versus trees.

- Important methodological advances have been made for extracting lianas from terrestrial LIDAR scans. We expect to quantify the influence of lianas on the ecosystem canopy structure.

- As an extension of the global meta-analysis, a pan-tropical study on the impact of lianas on tree growth and mortality will be performed based on existing forest inventory data from pan-tropically distributed sites. This will lead to new insights in the importance of liana-tree competition on forest ecosystem dynamics and how (the magnitude of) this competition varies with environmental and forest structural characteristics. This will also serve as an extensive validation dataset for our model predictions.