Search for: All records

Abstract Lianas are key components of tropical forests, particularly at sites with more severe dry seasons. In contrast, trees are more abundant and speciose in wetter areas. The seasonal growth advantage (SGA) hypothesis postulates that such contrasting distributions are produced by higher liana growth relative to trees during seasonal droughts. The SGA has been investigated for larger size classes (e.g., ≥5 cm diameter at 1.3 m, dbh), but rarely for seedlings. Using eight annual censuses of >12,000 seedlings of 483 tree and liana species conducted at eight 1‐ha plots spanning a strong rainfall gradient in central Panama, we evaluated whether liana seedlings had higher growth and/or survival rates than tree seedlings at sites with stronger droughts. We also tested whether an extreme El Niño drought during the study period had a more negative effect on tree compared to liana seedlings. The absolute density of liana seedlings was similar across the rainfall gradient, ranging from 0.32 individuals/m²(0.20–0.49, 95% credible interval [CI]) at the driest end of the gradient and 0.27 individuals/m²(0.13–0.51 95% CI) at the wettest end of the gradient. The relative density of liana seedlings compared to tree seedlings was higher at sites with stronger dry seasons (0.27, 0.21–0.33, 95% CI), compared to wetter sites (0.12, 0.04–0.20 95% CI), due to lower tree seedling densities at drier sites. However, liana seedlings did not grow or survive better than tree seedlings in drier sites compared to wetter sites. Tree seedlings were more negatively impacted in terms of mortality by the extreme El Niño drought than liana seedlings, with an increase in annual mortality rate of 0.013 (0.003–0.025 95% CI) compared to lianas of −0.009 (−0.028 to 0.008 95% CI), but not growth. Our results indicate that lianas do not have a SGA over trees at the seedling stage. Instead, higher survival of liana versus tree seedlings during severe droughts or differences in liana versus tree fecundity or germination across the rainfall gradient likely explain why liana seedlings have higher relative densities at drier sites. 

Lianas are a quintessential tropical plant growth-form; they are speciose and abundant in tropical forests worldwide. Lianas compete intensely with trees, reducing nearly all aspects of tree performance. However, the negative effects of lianas on trees have never been combined and quantified for multiple tropical forests. Here, we present the first comprehensive standardized quantification of the effect of lianas on trees across tropical forests worldwide. We used data from 50 liana removal experiments and quantified the effect size of lianas on tree growth, biomass accretion, reproduction, mortality, leaf water potential, sap flow velocity, and leaf area index (LAI) across different forest types. Using a three-level mixed-effect meta-analysis, we found unequivocal evidence that lianas significantly reduce tree growth and biomass accretion in ecological, logging, and silvicultural studies. Lianas also significantly reduce tree reproduction, recruitment, and physiological performance. The relative detrimental effect of lianas on trees does not increase in drier forests, where lianas tend to be more abundant. Our results highlight the substantial liana-induced reduction in tree performance and biomass accumulation, and they provide quantitative data on the effects of lianas on trees that are essential for large-scale plant demographic and ecosystem models that predict forest change and carbon dynamics. 

Data on individual tree crowns from remote sensing have the potential to advance forest ecology by providing information about forest composition and structure with a continuous spatial coverage over large spatial extents. Classifying individual trees to their taxonomic species over large regions from remote sensing data is challenging. Methods to classify individual species are often accurate for common species, but perform poorly for less common species and when applied to new sites. We ran a data science competition to help identify effective methods for the task of classification of individual crowns to species identity. The competition included data from three sites to assess each methods’ ability to generalize patterns across two sites simultaneously and apply methods to an untrained site. Three different metrics were used to assess and compare model performance. Six teams participated, representing four countries and nine individuals. The highest performing method from a previous competition in 2017 was applied and used as a baseline to understand advancements and changes in successful methods. The best species classification method was based on a two-stage fully connected neural network that significantly outperformed the baseline random forest and gradient boosting ensemble methods. All methods generalized well by showing relatively strong performance on the trained sites (accuracy = 0.46–0.55, macro F1 = 0.09–0.32, cross entropy loss = 2.4–9.2), but generally failed to transfer effectively to the untrained site (accuracy = 0.07–0.32, macro F1 = 0.02–0.18, cross entropy loss = 2.8–16.3). Classification performance was influenced by the number of samples with species labels available for training, with most methods predicting common species at the training sites well (maximum F1 score of 0.86) relative to the uncommon species where none were predicted. Classification errors were most common between species in the same genus and different species that occur in the same habitat. Most methods performed better than the baseline in detecting if a species was not in the training data by predicting an untrained mixed-species class, especially in the untrained site. This work has highlighted that data science competitions can encourage advancement of methods, particularly by bringing in new people from outside the focal discipline, and by providing an open dataset and evaluation criteria from which participants can learn.