OOS 5-3 - Remotely sensing the influence of lianas on tropical forest structure and function

Background/Question/Methods
Tropical forests are major reservoirs of carbon and have been acting as an important carbon sink for the past decades. However, these forests are currently undergoing large-scale structural changes that can be attributed primarily to anthropogenic disturbances. One such change is the increasing liana (woody climbing plants) abundance, especially in neotropical forests. Lianas compete strongly with trees thereby reducing tree growth and increasing tree mortality. Here, we quantify the contribution of increasing liana abundance to the accelerated carbon sink decline in Barro Colorado Island (BCI), a structurally intact tropical forest in Panama. We collected high resolution terrestrial laser scanning (TLS) data from a 4-ha area on BCI between January and March 2019. We manually segmented all trees ≥ 40 cm DBH from this area to study the impact of lianas on individual tree height, crown area and volume. We then derived allometric models for tree height, crown area and volume according to their liana infestation level (low, medium and high). Finally, we simulated the liana impact on forest carbon accumulation by using these TLS-derived allometric models in ED2, a global dynamic vegetation model, which currently uses pantropical allometric model that do not take lianas into account as its default.

Results/Conclusions
We found that trees that host many lianas (≥ 50% of its crown) were 9% shorter with smaller crown area of 43% and could store 16% less carbon than liana-free trees. Scaling this to stand-level for BCI resulted in 9% (mean ranging between 6% to 11.9%) liana-induced reduction in carbon stocks when compared to the current pantropical allometric model estimates. When simulating this TLS-derived impact of lianas on carbon accumulation using ED2, we found a 69% (4.8 vs. 1.5 Mg/ha/yr) reduction in the net aboveground biomass accumulation for the BCI site compared to the current pantropical allometries. And a 30% observed increase in liana density on BCI in a decade could result in further 49% (1.5 vs. 0.74 Mg/ha/yr) decrease in the carbon sink capacity of forest as opposed to a mere 13% reduction estimated by current pantropical allometries (4.8 vs. 4.2 Mg/ha/yr). The unique laser scanning data of this study allowed us to show for the first time that quantifying tree-level biomass reduction through liana-induced height, crown area, and woody biomass allometric change is essential for more accurate estimates of tropical forest carbon stocks and fluxes.