The Relationship Between Biodiversity and Ecosystem Function in Tropical Ecosystems

The Relationship between Biodiversity and Ecosystem Function in Tropical Ecosystems

Maria del Carmen Ruiz Jaen

Faculty of Science, Department of Biology and the Neotropical Environment Option

McGill University Montreal, Quebec, Canada August 2011

A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Doctor of Philosophy

395 Wellington Street 395, rue Wellington Ottawa ON K1A 0N4 Ottawa ON K1A 0N4 Canada Canada Your file Votre référence ISBN: 978-0-494-78693-2

Our file Notre référence ISBN: 978-0-494-78693-2

NOTICE: AVIS: The author has granted a non- L'auteur a accordé une licence non exclusive exclusive license allowing Library and permettant à la Bibliothèque et Archives Archives Canada to reproduce, Canada de reproduire, publier, archiver, publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public communicate to the public by par télécommunication ou par l'Internet, prêter, telecommunication or on the Internet, distribuer et vendre des thèses partout dans le loan, distrbute and sell theses monde, à des fins commerciales ou autres, sur worldwide, for commercial or non- support microforme, papier, électronique et/ou commercial purposes, in microform, autres formats. paper, electronic and/or any other formats.

The author retains copyright L'auteur conserve la propriété du droit d'auteur ownership and moral rights in this et des droits moraux qui protege cette thèse. Ni thesis. Neither the thesis nor la thèse ni des extraits substantiels de celle-ci substantial extracts from it may be ne doivent être imprimés ou autrement printed or otherwise reproduced reproduits sans son autorisation. without the author's permission.

In compliance with the Canadian Conformément à la loi canadienne sur la Privacy Act some supporting forms protection de la vie privée, quelques may have been removed from this formulaires secondaires ont été enlevés de thesis. cette thèse.

While these forms may be included Bien que ces formulaires aient inclus dans in the document page count, their la pagination, il n'y aura aucun contenu removal does not represent any loss manquant. of content from the thesis.

A mis padres, hermanos y abuelos

Table of Contents List of Tables 6 List of Figures 8 Preface 10 Contributions of Authors 11 Acknowledgements 12 Thesis Abstract 14 Résumé 15 General Introduction 17 Statement of Originality 29 Chapter 1: Tree diversity explains variation in ecosystem function in a Neotropical forest 30 Abstract 31 Introduction 32 Methods 33 Results 37 Discussion 39 Conclusion 42 References 43 Appendix 1: TREE carbon storage is not related to total stem abundance in Fort Sherman (n = 117). 54 Appendix 2: Description of soil physical variables. 55 Appendix 3: Correlations among environmental variables and the ordination axes of a Principal Component Analyses (PCA). 56 Appendix 4. Maps to illustrate the scores for the Principal Component of Neighboring Matrix (PCNM) at broad, medium and fine scales. 57 Appendix 5: Description of the environmental, spatial, and diversity variables. 58

Appendix 6: Size class distribution of the dominant species in Fort Sherman. 60 References for Appendices 61 Linking Statement 1 62 Chapter 2: Can we predict carbon stocks in tropical ecosystems from tree diversity? Comparing species and functional diversity in a plantation and a natural forest 63 Abstract 64 Introduction 65 Methods 67 Results 71 Discussion 72 References 76 Linking Statement 2 92 Chapter 3: Aboveground biomass, forest, structure and functional diversity: Patterns at different scales 93 Abstract 94 Introduction 95 Methods 97 Results 100 Discussion 101 References 105 Appendix 1: Correlation between the scores of the Principal Component Analysis (PCA) and eleven soil nutrients and soil pH at three scales at the 50-ha Dynamic Plot of Barro Colorado Island, Central Panama. 118 Appendix 2: Distribution of two functional traits, maximum height and leaf mass per area, within the 307 species present in the 50-ha Dynamic Plot of Barro Colorado Island, Central Panama. Vertical lines represent the cuts to classify species by low, medium, and high for each trait. 119 Linking Statement 3 120

Chapter 4: Biodiversity-Ecosystem functioning along a precipitation gradient in Panamanian tropical forests 121 Abstract 122 Introduction 123 Methods 124 Results 125 Discussion 126 References 128 Final summary and conclusions 136 References 138

List of Tables Chapter 1 Table 1. Variation partitioning results from the redundancy analyses (RDA) on tree carbon storage of trees and palms in Fort Sherman. Lower case letters represent single fractions of variation: (a) environment (E), (b) space (S), (c) diversity (D), (d) E*S, (e) S*D, (f) E*D, and (g) E*S*D. 50

Chapter 2 Table 1. Mean and coefficient of variation of tree carbon storage, stem density, light availability, species richness, functional diversity (standard deviation of maximum height

(Hmax), leaf mass per area (LMA), and nitrogen fixers (NF)), species dominance (relative basal area (BA) of A. excelsum and T. rosea), and functional dominance (community weighted mean of Hmax, LMA, and NF weighted by the BA) in the mixed-species plantation of Sardinilla and the natural forest of Barro Colorado Island (BCI) in Central Panama. 85 Table 2. Pearson correlations (r) between tree carbon storage controlled by stem density and light availability with species richness, functional diversity, species dominance, and functional dominance in the mixed-species plantation of Sardinilla and the natural forest of Barro Colorado Island (BCI) in Central Panama. 86 Table 3. Description of the two tree species used to explain the variation of carbon storage in the mixed-species plantation, Sardinilla and the natural forest of Barro Colorado Island (BCI) located in Central Panama. 87

Chapter 3 Table 1. Mean and coefficient of variation (in parentheses) of: (a) aboveground biomass (AGB), (b) forest structure (Gini coefficient), (c) species diversity measures (species richness and species evenness), (d) functional diversity measures (functional richness and functional evenness), (e) environment (slope, scores of first and second axis of principal

component analysis (PCA)) at three spatial scales in the 50 ha Dynamic Plot in Barro Colorado Island, Central Panama. 113 Table 2. Adjusted Coefficient of determination (R2) after partitioned the variation of aboveground biomass among species diversity measures (species richness and species evenness), functional diversity measures (functional richness and functional evenness), and environment (slope, scores of first and second axis of principal component analysis (PCA)) at three spatial scales in the 50 ha Dynamic Plot in Barro Colorado Island, Central Panama. Lower case letters represent single fractions of variation: (a) environment (E), (b) Species (S), (c) Function (F), (d) E*S, (e) S*F, (f) E*F, and (g) E*S*F. 114 Table 3. Adjusted coefficient of determination (R2) after partititioned the variation of Gini coefficient among species diversity measures (species richness and species evenness), functional diversity measures (functional richness and functional evenness), and environment (slope, scores of first and second axis of principal component analysis (PCA)) at three spatial scales in the 50 ha Dynamic Plot in Barro Colorado Island, Central Panama. Lower case letters represent single fractions of variation: (a) environment (E), (b) Species (S), (c) Function (F), (d) E*S, (e) S*F, (f) E*F, and (g) E*S*F. 115

Chapter 4 Table 1. Description of the three study sites along the Panama Canal Watershed. 131 Table 2. Mean and coefficient of variation (in parentheses) of: (a) aboveground biomass, (b) species richness (palms, canopy dicots, and total), (c) forest structure (the Gini coefficient, maximum diameter at breast height (DBH), number of trees larger than 70 cm DBH), in three forest along the Panama Canal Watershed. 132

List of Figures Chapter 1 Figure 1. Tree carbon storage in canopy trees distributed across the 124-20 x 20 m2 subplots in Fort Sherman, Panama. In parenthesis are the x and y coordinates used for the spatial analyses. Subplots marked with an X were not included in the statistical analysis. 51 Figure 2. Correlation biplots from redundancy analysis of tree carbon storage in understory trees and palms and canopy trees and palms constrained by the explanatory matrices of: (a) diversity and (b) environment and space (arrows) in Fort Sherman, Panama (see Appendix 5 for details in explanatory variables). Circles indicate the subplots (n=117). Crosses indicate the centroids of the response variables. Angles between tree carbon storage (variables in bold) and arrows of explanatory variables reflect their correlations. If the projection of tree carbon storage from the center of the axis is parallel to an arrow of an explanatory variable, then they are related. Arrow size is positive related to its effect level. 52

Chapter 2 Figure 1. Proportion of tree carbon storage variation explained after controlling for stem density and light availability in the mixed-species plantation of Sardinilla (black bars) and the natural forest of Barro Colorado Island (grey bars) located in Central Panama. Independent matrices were: species richness (SpDiv), functional diversity (FunDiv), species dominance (SpDom), and functional dominance (FunDom). 88 Figure 2. Species ranked by their dominance in the relative basal area in the mixed- species plantation of Sardinilla (a) and the natural forest of Barro Colorado Island (BCI; b), Central Panama. Values are average percent of basal area (BA) by species in each plot. Lines are 95% confidence intervals. Note that only the first 100 species of the natural forests are included, the rest are rare species that contribute very little to total BA in each plot. AE, Anacardium excelsum; TR, Tabebuia rosea 89 Figure 3. Species ranking comparison based on four functional traits of the species shared across the mixed-species plantation (Sardinilla) and the natural forest (Barro Colorado Island, BCI). Traits are: wood density (WD; a), log maximum diameter at breast height

(DBH; b), maximum height (c), and leaf mass per area (LMA; d). Species abbreviations: AG, Astronium graveolens; CO, Cedrela odorata; CG, Colubrina glandulosa; DO, Dipteryx oleifera, GU, Guazuma ulmifolia; HC, Hura crepitans; IP, Inga punctata; LS, Luehea seemannii; OM, Ormosia macrocalyx; TR, Tabebuia rosea; TA, Terminalia amazonia. 90

Chapter 3 Figure 1. Variation of aboveground biomass (Mg ha-1) within the 50- ha Dynamic Plot of Barro Colorado Island, Central Panama at three spatial scales: (a) 0.04 ha, (b) 0.25 ha, and (c) 1-ha. Values of aboveground biomass are log-transformed and increase from dark grey to light grey. Contour lines indicate the topography of the 50-ha plot as a measure of elevation 116 Figure 2. Positive relationship between the Gini coefficient and aboveground biomass at three spatial scales: (a) 0.04 ha, (b) 0.25 ha, and (c) 1 ha at the 50-ha Dynamic Plot of Barro Colorado Island, Central Panama. 117

Chapter 4 Figure 1. Distribution trees based on their diameter at breast height (DBH) in three forests along the Panama Canal Watershed: (a) dry, (b) moist, and (c) wet forest. 133 Figure 2. Left panels show the effect of (a) palm richness, and (b) canopy dicot richness on aboveground biomass in three forests along the Panama Canal Watershed. Each  represents data from dry forest,  from moist forest, and  from wet forest. Dotted lines correspond to linear regression models for the dry forest, the dashed lines for the moist forest, and the solid line for the wet forest. 134 Figure 3. The effect of (a) the Gini coefficient and (b) big trees ( ≥ 70 cm DBH) on aboveground biomass at three forest along the Panama Canal Watershed. Each  represents data from dry forest,  from moist forest, and  from wet forest. Dotted lines correspond to linear regression models for dry forest, the dashed lines for the ones in the moist forest, and the solid line for the ones in the wet forest. 135

Preface

This is a manuscript-based thesis, consisting of a collection of papers of which I am the primary author. Chapter 1 and 2 has been published; Chapter 3 and 4 are in preparation for submission. The manuscripts and associated journals are as follows:

Chapter 1 Ruiz-Jaen, MC, Potvin C. 2010. Tree Diversity Explains Variation in Ecosystem Function in a Neotropical Forest. Biotropica 42: 638-646.

Chapter 2 Ruiz-Jaen MC, Potvin C. 2011. Can we predict carbon stocks in tropical ecosystems from tree diversity? Comparing species and functional diversity in a plantation and a natural forest. New Phytologist 189: 978-987.

Chapter 3 Ruiz-Jaen MC, Wright SJ, Potvin C. In preparation. Aboveground biomass, forest structure and functional diversity: Patterns at different scales.

Chapter 4 Ruiz-Jaen MC, Wright SJ, Potvin C. In preparation. Biodiversity-Ecosystem functioning along a precipitation gradient in Panamanian tropical forests

Contributions of Authors

I am the primary author of all the studies included in this thesis. I formulated the hypothesis, sampling design, collected the data, analyzed the data, and wrote the manuscripts. Catherine Potvin supervised the conceptual framework, sampling design, interpretation of results, and writing of all the manuscripts in this thesis. S. Joseph Wright supervised the data collection, interpretation of results, and writing of Chapter 3 and 4.

Acknowledgements

First of all, I would like to thank Catherine Potvin. She has been an amazing mentor to me. Her guidance, intellectual support, wise comments, pragmatism, and patience were essential for the development of this thesis. I will always be indebted to her. Many thanks to the Potvin Lab, Johanne Pelletier, Ignacia Holmes, and the recent additions, Guillaume Peterson-St. Laurent, Gerardo Vergara-Asenjo, and Paulina Lezama- Nuñez, for all insightful discussions during my time at McGill. I am very thankful to my supervisory committee, Sylvie de Blois, Andrew Gonzalez, and Michel Loreau, who gave constructive comments and helped improve my thesis. I am especially grateful to Joe Wright. Even though, he was a late addition to my supervisory committee, he quickly became a key member for his endless knowledge on tropical forests, his engaging discussions, and his clear vision in research. He also gave me the opportunity to work in Barro Colorado Island (BCI) and provided support to collect the plant functional trait data in BCI and Sardinilla. This thesis would not have been possible with the constant support of Susan Bocti, who reminded me all the deadlines, and helped me keep updated in all the McGill paperwork. I also appreciate the support of Martine Dolmiere, Susan Gabe, Ancil Gittens, Robert Lamarche, Anna McNicoll, Linda Morai, Luisa Sabaz, and Carole Verdone-Smith. I would like to thank my “adoptive” Labs, Lechowicz’s and Leung’s, who offered their expertise during Catherine’s travels. Most especially, I would like to thank Xavier Morin, Juan Posada, and Brian Leung who were always ready to help, to discuss results, and to make my life in Montreal more entertaining. I would like to thank my fellow NEO students: Karina Benessaiah, Cristian Correa, Luis Fernando de Leon, Bruno Guay, Kecia Kerr, Sohinne Mazumbar, Manfred Meiners, Alida Mercado, Sky Oestreicher, Oscar Puebla who share with me the difficulties and advantages of living in two countries and offered different perspectives in Science. Most especially to Carlos Arias and Sergio Estrada, who supported me many times and were always encouraging.

I would like to thank all the friends that I met during this journey. I am very grateful to my friends, Marcelo Castillo, Xavier Gravend-Tirole, Alex de Roode, Jeffrey Barnes, and Phillip Crete for always being supportive and even prepare meals for me during stressful times. Most especially to Marc Dunn and Rosa Ponce who offer me shelter during transitional stages and made my time in Montreal a very special one. I am very grateful to Salomon Aguilar, Javier Ballesteros, Sebastian Bernal, Santiago Bonilla, David Brassfield, Daddiviam Gomez, Jordan Leblanc, Lady Mancilla, Eduardo Medina, Jose Monteza, Jose Angel Quintero, Rolando Perez, Cristina Salvador, Paulino Villarreal for their help in the collection and processing of data during my dissertation. I am most indebted to Andy Jones for having the patience of reading my chapters several times, for discussing results, and for giving moral support during the ups and downs of my PhD. I would like to thank my family for believing in me and making this experience a positive one.

Thesis Abstract Experimental studies, in temperate grasslands, assessing the role of biodiversity and its effects on ecosystem functioning have generally shown that a decline in species number has negative effects on ecosystem functioning. Even though, this relationship has been intensively studied in the last ten years, little is known about it in complex and hyperdiverse tropical ecosystems and where species diversity is not manipulated. My research examines the relationship between biodiversity and ecosystem functioning in natural tropical forests with a special focus on scale. This research centers on field studies. The field studies address the relationship between natural tree biodiversity and aboveground biomass, as the ecosystem function of interest, in forest plots of similar physiognomy but different species composition. Specifically, I explored the following questions: (1) How can the relationship between biodiversity and ecosystem functioning be detected in a naturally varying environment and space?, (2) How can different measures of diversity (species versus function) explain tree carbon stocks?, (3) Can we confound the effect of species diversity on tree carbon storage with that of forest structure?, (4) How does this relationship change with different spatial scales?, and (5) Can we extrapolate the results of biodiversity-ecosystem functioning found in experimental plantations to natural forests? Overall, my thesis has found that environmental factors related to topography, soil physical factors, and nutrients have little effect on aboveground biomass in tropical ecosystems. Species richness alone cannot be used as a predictor for aboveground biomass, however, if reduced to functional types, its explanatory power increases. Functional traits can be useful to unveil the relationship of aboveground biomass and tree diversity, by reducing species to functional types. Forest structure correlates strongly with aboveground biomass independently of scale, but forest structure is interlinked with species functional traits. Finally, we have to be cautious in extrapolating results found in experimental plantations to natural forests.

Résumé Des études expérimentales qui ont été effectuées dans des systèmes expérimentaux herbacés du milieu tempéré afin d’évaluer le rôle de la diversité biologique et ses effets sur le fonctionnement de l’écosystème ont généralement pu montrer que le déclin dans le nombre d’espèces a un effet négatif sur le fonctionnement de l’écosystème. Même si cette relation a été étudiée de façon intensive au cours des dix dernières années, très peu est encore connu à ce propos dans les écosystèmes tropicaux, beaucoup plus complexes et mégadiversifiés. Mes travaux de recherche examinent la relation entre diversité biologique et fonction de l’écosystème dans les forêts naturelles tropicales, en mettant l’emphase sur la question d’échelle. Cette recherche est basée sur une approche de terrain, en opposition avec une approche théorique. L’approche de terrain aborde la relation entre la diversité biologique à l’état naturel et les fonctions d’écosystème dans des parcelles forestières de physionomies similaires mais d’une composition en espèces différente. De façon plus spécifique, j’explore les questions suivantes : (1) Comment la relation entre diversité biologique et fonction d’écosystème peut-elle être détectée dans un environnement et un espace naturel changeant?, (2) Comment différentes mesures de diversité (espèces vs fonction) expliquent-elles les stocks de carbone dans les arbres?, (3) Pouvons-nous confondre l’effet de la diversité en espèces sur l’entreposage du carbone dans les arbres avec l’effet de la structure de la forêt? (4) Comment cette relation change-t-elle avec différentes échelles spatiales? (5) Pouvons-nous extrapoler les résultats obtenus dans des plantations expérimentales aux forêts naturelles dans les tropiques en ce qui a trait à la relation entre biodiversité et fonction d’écosystème? Ainsi, ma thèse montre que pour les écosystèmes tropicaux, les facteurs environnementaux tels que ceux reliés à la topographie, les facteurs physiques et les nutriments des sols ont peu d’effet sur l’entreposage du carbone dans les arbres. De plus, la richesse en espèces ne peut pas à elle prédire de l’entreposage de carbone dans les arbres, cependant lorsque celle-ci est divisée par types fonctionnels, sa puissance explicative augmente. Les traits fonctionnels peuvent donc être utiles pour révéler une relation entre le stockage du carbone et la diversité en arbres en réduisant les espèces en des types fonctionnels. La structure de la forêt est le moteur principal du stockage arboricole du carbone indépendamment de l’échelle, par

contre cette dernière est reliée aux traits fonctionnels des espèces. Finalement, nous avons pu montrer que la prudence est de mise en ce qui concerne toute possible extrapolation de résultats provenant de plantations expérimentales à des forêts naturelles dans les tropiques.

Translated by Johanne Pelletier

General Introduction

Biodiversity and Ecosystem Functioning The role of biodiversity and its effects on ecosystem functioning (BEF) has been studied intensely over the last twenty years (Naeem et al. 2003, Balvanera et al. 2006). In general, these studies have shown that a decline in species numbers has negative effects on ecosystem functioning (Tilman et al. 1996, Naeem et al. 2003, Spehn et al. 2006). For example, productivity declines with the loss of plant species in experimental plots in studies conducted in temperate grasslands of eight European countries (Hector et al. 1999). However, these studies have shown that the relationship between biodiversity and ecosystem function is asymptotic, i.e., only a few species (less than 15) are important to ecosystem functioning (Hector et al. 2001, Balvanera et al. 2005). Results of BEF have been explained mostly by two mechanisms, species complementarity versus sampling or selection effects (Loreau 2000, Loreau and Hector 2001). The complementarity effect is related to niche partitioning, where each species uses resources differently, so that a site with a group of species performs better than sites dominated by a single species. The sampling or selection effect occurs when the probability of including the most productive species increases as the number of species increases. Most BEF studies have used small experimental plots (0.1 m2 to 100 m2) in temperate grasslands or aquatic microcosms (Hopper et al. 2005, Balvanera et al. 2006, Cardinale et al. 2006). In a meta-analysis of 103 studies with a wide range of ecosystem functions, Balvanera et al. (2006) found that, overall, biodiversity has a positive effect on ecosystem function. Cardinale et al. (2007) found that polycultures produced more biomass than monocultures due to niche complementarity and that this relationship becomes stronger over time. In general, these studies have provided evidence that species diversity is linked to ecosystem functioning; however, these results may not apply to all ecosystems or across scales that are relevant for conservation (Lawler et al. 2002, Srivastava and Vellend 2005). Perennial grassland communities, a common model ecosystem used in BEF studies, allow for easy manipulation and rapid measures of productivity. These studies can also be carried out at relatively small spatial scales (< 20 m2), thereby allowing for detailed examinations of BEF in homogeneous environments (Roscher et al. 2005). However, comparatively few studies of BEF have been conducted in forest tree communities.

Forests are composed of long-lived trees (Chambers et al. 1998), which make it difficult to manipulate diversity over short time periods. In addition, the minimum spatial scale over which ecosystem functioning can be effectively measured (e.g., primary productivity) in forested systems is between 0.04 ha to 0.25 ha (Chave et al. 2004), which makes it difficult to control for site heterogeneity. When conducted in natural ecosystems, some BEF studies have shown contradictory effects (Naeem 2002). For example, less diverse forests can be more resistant to perturbation (e.g., hurricanes) than diverse ones (Tanner and Bellingham 2006). Similarly, plant productivity decreased in less diverse mature natural temperate grasslands (Thompson et al. 2005). In order for investigations of BEF in forest ecosystems to be relevant to conservation, it is important to understand how biodiversity affects ecosystem functions at larger scales and across a variety of different ecosystems (Srivastava and Vellend 2005) particularly in ecosystems where biodiversity is high, such as in tropical wet forests.

Biodiversity and Ecosystem Functioning in the Tropics Tropical forest biodiversity is currently threatened by forest degradation (Wright 2005) and this threat is expected to continue into the foreseeable future (Sala et al. 2000). Recently, attention has been given as to how deforestation can be avoided in tropical systems through the exchange of “carbon credits” (Santilli et al. 2005, Baker et al. 2010). If adopted by the UNFCCC (United Nations Framework on Climate Change Convention), this new approach has the potential to encourage forest conservation in the tropics (Venter et al. 2009). Despite enormous interest in tropical forest conservation, only a handful of studies have actually examined the biodiversity-ecosystem function relationship in these forests (Balvanera et al. 2006). Attempts to study BEF in tropical forests have included simulations, experimental plantations, and observations. For example, Balvanera et al. (2005) simulated the effect of timber extraction on carbon storage by modeling the removal of commercially valuable species ≥ 30 cm diameter at breast height (DBH) in a conserved tropical forest. They showed that timber extraction could reduce up to 60% of carbon storage in the forest and change the abundance distribution of species that are important carbon contributors, without affecting species loss. The shortcoming of this

study is that the authors did not include compensation effects (e.g., compensatory growth, species replacement) after logging. In another study, Bunker et al. (2005) simulated 18 extinction scenarios of species loss and their effects on aboveground biomass in the forests of Barro Colorado Island, Central Panama. These simulations included trait-based extinction scenarios based on possible losses due to population traits (abundant/rare), forest management (selective harvest), and environmental changes (carbon dioxide concentrations, forest disturbance, precipitation). The authors reported that the species identity of those species that were lost had a considerable effect on determining aboveground biomass (Bunker et al. 2005). They used the compensatory growth of remaining species to further assess projected changes in aboveground biomass. However, their approach did not consider the probability of replacement of the species that are removed thus ignoringthe natural dynamics of the community, where colonization by other species occurs as a natural process following the removal of an individual. Species replacement depends on available seedlings (Janzen 1970, Connell et al. 1984, Wills and Condit 1999, Wills et al. 2006) and adult trees (Comita et al. 2007) that are close to the tree removal site. Seedling growth potential will also depend on neighborhood composition and density, where the presence of conspecifics can be disadvantageous (Janzen 1970). Plantations offer a useful option for BEF studies in the tropics, because they allow manipulation of species richness. There have been two approaches to this problem: forestry-oriented and species-manipulated plantations. Forestry-oriented plantations manipulate tree diversity in a manner similar to the experimental approach used in early BEF experiments in grasslands, even though the objective of the former is frequently one of increased timber productivity (Erksine et al. 2006). A review of these studies showed that diameter growth rates were higher in mixed-plantations than in monocultures (Piotto 2008). Such studies have been carried out in the lowland tropics of Central America and Borneo with manipulations of species richness varying from 1 to 28 species (Menalled et al. 1998, Scherer-Lorenzen et al. 2005, Potvin et al. 2011). In general, these studies have found that mixed-species plantations have higher biomass and annual carbon sequestration than monocultures (Healy et al. 2008, Potvin et al. 2011).

Nevertheless, plantations do not contain the inherent variability that occurs in natural forests (Scheren-Lorenzen et al. 2005, Vilà et al. 2007). For example, plantations are even- aged stands, while natural forests consist of uneven-aged stands. The latter provides both horizontal and vertical diversity, which enhances niche differences. Moreover, plantations are assembled in a deterministic fashion because species presence and abundance are by definition manipulated. Individuals within natural communities assemblages can depend on the abundance and species identity of neighbors (Wills et al. 2006) and the available species pool in the metacommunity is limited by dispersal (Loreau et al. 2003, Chust et al. 2006). Field observations have been mostly done to compare ecosystem functions under different land uses. These studies have shown that carbon storage (Balvanera et al. 2005, Kirby and Potvin 2007) and nutrient availability (Ewel et al. 1991) were higher in forests at older successional stages than younger stages. For example, Kirby and Potvin (2007) aimed to understand BEF in human landscapes by comparing species composition and carbon storage in plots under different land-use (pastures, agroforestry systems, and managed forests). They found that agroforestry systems stored 30% more carbon than pastures, but 50% less than managed forests. However, agroforestry provided several livelihood benefits; thus, the authors suggested the development of agroforestry systems in pastures as a forest restoration effort, instead of monoculture plantations. They also suggested a species-based management plan, since some species that disproportionally contributed to forest carbon are also preferred for timber extraction. Biodiversity Measures A main difference between BEF studies done in plantations and in natural forests relates to the amount of diversity per se. The Sardinilla experimental plantation in Panama is one the most diverse experimental BEF tree plantations ever created, with up to 18 species per 400-m2 plot (Scherer-Lorenzen et al. 2005). In tropical forest systems, however, biodiversity can be as high as 300 species per ha (Condit et al. 2005). Therefore, this raises the question of how should biodiversity be best measured. Previous studies of biodiversity and ecosystem functioning have yielded different results, depending on how biodiversity was measured (Hooper et al. 2005). In the initial BEF studies, biodiversity was reduced to species richness (i.e., the number of species; Naeem et al. 1994, Tilman et al. 1996, Hector et al. 1999, Loreau 2000), yet this approach is too simplistic, as

biodiversity effects encompass a wider range of traits (e.g., identity, density, evenness, function; Condit et al. 1998, Purvis and Hector 2000, Hooper et al. 2005). For example, community composition can have a major impact on ecosystem functioning (Lovelock and Ewel 2005). Hooper and Vitousek (1997) found that species composition (i.e., species identity) explained nitrogen production and dynamics in Californian grasslands. Similarly, if species richness remains constant, the relative abundance and identity of species is related to total biomass in old fields (Wisley and Potvin 2000). In fact, species identity is often more strongly related to ecosystem functioning than is species richness (Hopper et al. 2005, Roscher et al. 2005). Moreover, functional diversity could be an important and more reliable measure of biodiversity in BEF studies than species diversity (Diaz and Cabildo 2001, Lavorel and Garnier 2002, Petchey and Gaston 2002). Functional diversity requires the partitioning of species into functional groups based on their traits (McGill et al. 2006). For example, Petchey and Gaston (2002) proposed an objective way to determine functional groups of species by using cluster analysis on species functional traits to classify species in functional groups. Several other studies have addressed different ways of measuring functional diversity (Violle et al. 2004, Leps et al. 2006, Petchey and Gaston 2006), but most of these have still used plant functional traits to classify species in a subjective way.

Thesis Overview Building upon findings in the previous section, I expanded on the biodiversity- ecosystem function relationship in natural tropical ecosystems. Specifically, I studied the relationship between tree diversity and carbon storage, as a key ecosystem function in the tropics. In Chapter 1, I hypothesized that environment and space can confound the effect of tree diversity on tree carbon storage. To address this hypothesis, I removed the effect of environmental variables (topography, soil physical characteristics) and space (I used principal coordinates of neighbor matrices, PCNM) to study only the biodiversity- ecosystem function relationship in a wet tropical forest on the Caribbean coast of Panama. I also studied the relative importance of niche complementarity and selection effect on carbon storage. To address niche complementarity, I calculated species richness of four

functional groups with very different carbon storage capacity: understory dicots, understory palms, canopy dicots, and canopy palms. To address the sampling effect, I calculated the relative basal area of the species dominating the total basal area in a plot. I found that species diversity measures explained tree carbon storage to a greater degree than did space and environment. I also found that niche complementarity and selection effects are complementary. In Chapter 2, I was also interested in studying the relative contributions of species versus functional diversity in tropical ecosystems. I thus compared the relative importance of species diversity, functional diversity, species dominance, and functional dominance on tree carbon storage in the species rich-experimental plot in Sardinilla, versus the nearby natural forest of Barro Colorado Island, Panama. For functional diversity and dominance, I used the dispersion and community-weighted means of maximum height, leaf mass per area (LMA), and nitrogen-fixers. This chapter also addressed the possible use of functional traits of species in natural forests to design carbon-rich plantations. Therefore, I compared the consistency of relative and absolute values of key functional traits (wood density, maximum diameter at breast height, maximum height, and LMA) between the experimental plantation and the natural forest. I found that species-related variables were more important to tree carbon storage in the experimental plantation, while functionally related variables were more important in the natural forest. I also found that functional traits varied in terms of their absolute and relative values between sites. In Chapter 3, I determined whether or not the effects of species and functional diversity on aboveground biomass were scale-dependent. In previous chapters, studies were conducted at a scale of 20 m x 20 m to maintain homogeneity in environmental conditions for each plot. To scale up BEF, I studied the aforementioned relationships at 20 m x 20 m, 50 m x 50 m, and 100 m x 100 m scales in the natural forest of Barro Colorado Island, Panama. In this study, I also explored the importance of environmental variables (topography and soil nutrients) and forest structure on aboveground biomass. I found that functional evenness and aboveground biomass decreased in parallel at all spatial scales. Also, I confirmed a previous finding that at the forest level, environment plays a minor role in carbon storage. Finally, forest structure was highly correlated with aboveground biomass. These results were scale-independent.

In Chapter 4, I built upon previous chapters and attempted to generalize the biodiversity-ecosystem function relationship in natural tropical ecosystems. To do that, I studied the effect of tree diversity divided in two functional groups, palms versus dicots, on aboveground biomass in three forests growing under different precipitation regimes. In addition, I explored the effect of forest structure on aboveground biomass in these sites. I confirmed that selection effect rather than complementarity is at play in these ecosystems.

REFERENCES Baker TR, Jones JPG, Thompson ORR, Cuesta RMR, del Castillo D, Aguilar IC, Torres J, Healey JR. 2010. How can ecologists help realize the potential of payments for carbon in tropical forest countries? Journal of Applied Ecology 47(6): 1159-1165. Balvanera P, Kremen C, Martinez-Ramos M. 2005. Applying community structure analysis to ecosystem function: Examples from pollination and carbon storage. Ecological Applications 15(1): 360-375. Balvanera P, Pfisterer AB, Buchmann N, He JS, Nakashizuka T, Raffaelli D, Schmid B. 2006. Quantifying the evidence for biodiversity effects on ecosystem functioning and services. Ecology Letters 9(10): 1146-1156. Bunker DE, DeClerck F, Bradford JC, Colwell RK, Perfecto I, Phillips OL, Sankaran M, Naeem S. 2005. Species loss and aboveground carbon storage in a tropical forest. Science 310: 1029-1031. Cardinale BJ, Wright JP, Cadotte MW, Carroll IT, Hector A, Srivastava DS, Loreau M, Weis JJ. 2007. Impacts of plant diversity on biomass production increase through time because of species complementarity. Proceedings of the National Academy of Sciences of the United States of America 104 (46): 18123-18128. Chambers JQ, Higuchi N, Schimel JP. 1998. Ancient trees in Amazonia. Nature 391(6663): 135-136. Chave J, Condit R, Aguilar S, Hernandez A, Lao S, Perez R. 2004. Error propagation and scaling for tropical forest biomass estimates. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 359(1443): 409-420.

Chust G, Chave J, Condit R, Aguilar S, Lao S, Perez R. 2006. Determinants and spatial modeling of tree beta-diversity in a tropical forest landscape in panama. Journal of Vegetation Science 17(1): 83-92. Comita LS, Aguilar S, Perez R, Lao S, Hubbell SP. 2007. Patterns of woody plant species abundance and diversity in the seedling layer of a tropical forest. Journal of Vegetation Science 18(2): 163-174. Condit R, Ashton P, Bunyavejchewin S, Dattaraja HS, Davies S, Esufali S, Ewango C, Foster R, Gunatilleke I, Gunatilleke CVS, Hall P, Harms KE, Hart T, Hernandez C, Hubbell SP, Itoh A, Kiratiprayoon S, LaFrankie J, de Lao SL, Makana JR, Noor MNS, Kassim AR, Russo S, Sukumar R, Samper C, Suresh HS, Tan S, Thomas S, Valencia R, Vallejo M, Villa G, Zillio T. 2006. The importance of demographic niches to tree diversity. Science 313(5783): 98-101. Condit R, Foster RB, Hubbell SP, Sukumar R, Leigh EG, Manokaran N, Lao S, LaFrankie J, Ashton PS. 1998. Assessing forest diversity on small plots: calibration using species- individual curves from 50-ha plots. In: Dallmeier F, Comiskey JA eds. Forest biodiversity research, monitoring and modeling: conceptual background and old world case studies. Paris, France: UNESCO. Connell JH, Tracey JG, Webb LJ. 1984. Compensatory recruitment, growth, and mortality as factors maintaining rain forest tree diversity. Ecological Monographs 54: 141-164. Diaz S, Cabido M. 2001. Vive la différence: plant functional diversity matters to ecosystem processes. Trends in Ecology & Evolution 16(11): 646-655. Diaz S, Fargione J, Chapin FS, Tilman D. 2006. Biodiversity loss threatens human well- being. Plos Biology 4(8): 1300-1305. Erskine PD, Lamb D, Bristow M. 2006. Tree species diversity and ecosystem function: Can tropical multi-species plantations generate greater productivity? Forest Ecology and Management 233(2-3): 205-210. Ewel JJ, Mazzarino MJ, Berish CW. 1991. Tropical soil fertility changes under monocultures and successional communities of different structure. Ecological Applications 1(3): 289-302.

Healy C, Gotelli NJ, Potvin C. 2008. Partitioning the effects of biodiversity and environmental heterogeneity for productivity and mortality in a tropical tree plantation. Journal of Ecology 96(5): 903-913. Hector A, Joshi J, Lawler SP, Spehn EM, Wilby A. 2001. Conservation implications of the link between biodiversity and ecosystem functioning. Oecologia 129: 624-628. Hector A, Schmid B, Beierkuhnlein C, Caldeira MC, Diemer M, Dimitrakopoulos PG, Finn JA, Freitas H, Giller PS, Good J, Harris R, Högberg P, Huss-Danell K, Joshi J, Jumpponen A, Körner Ch, Leadley PW, Loreau M, Minns A, Mulder CPH, O'Donovan G, Otway SJ, Pereira JS, Prinz A, Read DJ, Scherer-Lorenzen M, Schulze E-D, Siamantziouras A-SD, Spehn EM, Terry AC, Troumbis AY, Woodward FI, Yachi S, Lawton JH. 1999. Plant diversity and productivity experiments in European grasslands. Science 286: 1123-1127. Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge DM, Loreau M, Naeem S, Schmid B, Setala H, Symstad AJ, Vandermeer J, Wardle DA. 2005. Effects of biodiversity on ecosystem functioning: A consensus of current knowledge. Ecological Monographs 75(1): 3-35. Hooper DU, Vitousek PM. 1997. The effects of plant composition and diversity on ecosystem processes. Science 277: 1302-1305. Janzen DH. 1970. Herbivores and the number of tree species in tropical forests. American Naturalist 104: 501-528. Kirby KR, Potvin C. 2007. Variation in carbon storage among trees species: Implications for the management of small-scale carbon sink project. Forest Ecology and Management 246: 208-221. Lavorel S, Garnier E. 2002. Predicting changes in community composition and ecosystem functioning from plant traits: revisiting the Holy Grail. Functional Ecology 16(5): 545- 556 Lawler SP, Armesto JJ, Kareiva P. 2002. How relevant to conservation are studies linking biodiversity and ecosystem functioning? Functional consequences of biodiversity. Princeton: Princeton University Press, 294-313. Leps J, de Bello F, Lavorel S, Berman S. 2006. Quantifying and interpreting functional diversity of natural communities: Practical considerations matter. Preslia 78: 481-501.

Loreau M. 2000. Biodiversity and ecosystem functioning: Recent theoretical advances. Oikos 91: 3-17. Loreau M, Hector A. 2001. Partitioning selection and complementarity in biodiversity experiments. Nature 412(6842): 72-76. Loreau M, Mouquet N, Gonzalez A. 2003. Biodiversity as spatial insurance in heterogeneous landscapes. Proceedings of the National Academy of Sciences of the United States of America 100(22): 12765-12770. Lovelock CE, Ewel JJ. 2005. Links between tree species, symbiotic fungal diversity and ecosystem functioning in simplified tropical ecosystems. New Phytologist 167(1): 219- 228. McGill BJ, Enquist BJ, Weiher E, Westoby M. 2006. Rebuilding community ecology from functional traits. Trends in Ecology & Evolution 21: 178-185. Menalled FD, Kelty MJ, Ewel JJ. 1998. Canopy development in tropical tree plantations: A comparison of species mixtures and monocultures. Forest Ecology and Management 104(1-3): 249-263. Naeem S. 2002. Ecosystem consequences of biodiversity loss: the evolution of a paradigm. Ecology 83: 1537-1552. Naeem S, Thompson LJ, Lawler SP, Lawton JH, Woodfin RM. 1994. Declining biodiversity can alter the performance of ecosystems. Nature 368(6473): 734-737. Naeem S, Wright JP. 2003. Disentangling biodiversity effects on ecosystem functioning: Deriving solutions to a seemingly insurmountable problem. Ecology Letters 6: 567-579. Petchey OL, Gaston KJ. 2006. Functional diversity: Back to basics and looking forward. Ecology Letters 9(6): 741-758. Piotto D. 2008. A meta-analysis comparing tree growth in monocultures and mixed plantations. Forest Ecology and Management 255(3-4): 781-786. Potvin C, Mancilla L, Buchmann N, Monteza J, Moore T, Murphy M, Oelmann Y, Scherer-Lorenzen M, Turner BL, Wilcke W, Zeugin F, Wolf S. 2011. An ecosystem approach to biodiversity effects: carbon pools in a tropical tree plantation. Forest Ecology and Management 261: 1614-1629. Purvis A, Hector A. 2000. Getting the measure of biodiversity. Nature 405(6783): 212- 219.

Roscher C, Temperton VM, Scherer-Lorenzen M, Schmitz M, Schumacher J, Schmid B, Buchmann N, Weisser WW, Schulze ED. 2005. Overyielding in experimental grassland communities - irrespective of species pool or spatial scale. Ecology Letters 8(4): 419- 429. Sala OE, Chapin III FS, Armesto JJ, Berlow E, Bloomfield J, Dirzo R, Huber-Sanwald E, Huenneke LF, Jackson RB, Kinzig A, Leemans R, Lodge DM, Mooney HA, Oesterheld M, Poff NL, Sykes MT, Walker BH, Walker M, Wall DA. 2000. Biodiversity - global biodiversity scenarios for the year 2100. Science 287(5459): 1770-1774. Santilli M, Moutinho P, Schwartzman S, Nepstad D, Curran L, Nobre C. 2005. Tropical deforestation and the Kyoto protocol. Climatic Change 71: 267-276. Scherer-Lorenzen M, Potvin C, Koricheva J, Schmid B, Hector A, Bornik Z, Reynolds G, Schulze ED. 2005. The design of experimental tree plantations for functional biodiversity research. In: M. Scherer-Lorenzen, C. KornerE. D. Schulze eds. Forest diversity and function: Temperate and boreal systems. Heidelberg, Germany: Springer, 347-376. Spehn EM, Hector A, Joshi J, Scherer-Lorenzen M, Schmid B, Bazeley-White E, Beierkuhnlein C, Caldeira MC, Diemer M, Dimitrakopoulos PG, Finn JA, Freitas H, Giller PS, Good J, Harris R, Hogberg P, Huss-Danell K, Jumpponen A, Koricheva J, Leadley PW, Loreau M, Minns A, Mulder CPH, O'Donovan G, Otway SJ, Palmborg C, Pereira JS, Pfisterer AB, Prinz A, Read DJ, Schulze ED, Siamantziouras ASD, Terry AC, Troumbis AY, Woodward FI, Yachi S, Lawton JH. 2005. Ecosystem effects of biodiversity manipulations in european grasslands. Ecological Monographs 75(1): 37- 63. Srivastava DS, Vellend M. 2005. Biodiversity-ecosystem function research: Is it relevant to conservation? Annual Review of Ecology Evolution and Systematics 36: 267-294. Tanner EVJ, Bellingham PJ. 2006. Less diverse forest is more resistant to hurricane disturbance: evidence from montane rain forests in Jamaica. Journal of Ecology 94: 1003-1010. Thompson K, Askew AP, Grime JP, Dunnett NP, Willis AJ. 2005. Biodiversity, ecosystem function and plant traits in mature and immature plant communities. Functional Ecology 19(2): 355-358.

Tilman D, Wedin D, Knops J. 1996. Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature 379: 718. Venter O, Laurance WF, Iwamura T, Wilson KA, Fuller RA, Possingham HP. 2009. Harnessing carbon payments to protect biodiversity. Science 326(5958): 1368-1368. Vila M, Vayreda J, Comas L, Ibanez JJ, Mata T, Obon B. 2007. Species richness and wood production: a positive association in Mediterranean forests. Ecology Letters 10(3): 241-250. Violle C, Navas ML, Vile D, Kazakou E, Fortunel C, Hummel I, Garnier E. 2007. Let the concept of trait be functional! Oikos 116(5): 882-892. Wills C, Condit R. 1999. Similar non-random processes maintain diversity in two tropical rainforests. Proceedings of the Royal Society of London Series B-Biological Sciences 266(1427): 1445-1452. Wills C, Harms KE, Condit R, King D, Thompson J, He F, Muller-Landau HC, Ashton P, Losos L, Comita L, Hubbell SP, LaFrankie J, Bunyavejchewin P, Dattaraja HS, Davies S, Esufali S, Foster R, Gunatilleke N, Gunatilleke S, Hall P, Itoh A, John R, Kiratiprayoon S, Loo de Lao S, Massa M, Nath C, Noor S, Kassim AR, Sukumar R, Suresh HC, Sun I-F, Tan S, , Yamakura T, Zimmerman J. 2006. Nonrandom processes maintain diversity in tropical forests. Science 311(5760): 527-531. Wilsey BJ, Potvin C. 2000. Biodiversity and ecosystem functioning: Importance of species evenness in an old field. Ecology 81(4): 887-892. Wright SJ. 2005. Tropical forests in a changing environment. Trends in Ecology & Evolution 20(10): 554-559.

Statement of Originality

This research is the first attempt to study the relationship between biodiversity and ecosystem function (BEF) in natural forests in the tropics. Specifically, I studied the link between tree diversity and tree carbon storage, as the ecosystem function of interest. First, I partitioned out the effect of environment and space to understand the relationship between species diversity and carbon storage in naturally varying environments (Chapter 1). In Chapter 2, I provided the first attempt to generalize BEF in the tropics by comparing a natural forest and an experimental plantation and also comparing the importance of species versus functional diversity on an ecosystem function. I found that the effects of species diversity found in plantations couldn’t be extrapolated to natural forests. I also found that reducing the number of species to its functional traits increases the predictability of tree carbon storage in natural forests. In Chapter 3, I tested if the previous relationships were scale dependent. Finally, in Chapter 4, I tested BEF in three tropical forests of different species composition and precipitation regimes to generalize my previous findings. This research contributed to new knowledge on four unresolved issues regarding BEF in tropical ecosystems: (1) its relationship to different biodiversity measures (2) its detection in natural communities, (3) its changes through out different forests, and (4) its changes with scale.

CHAPTER I: Tree diversity explains variation in ecosystem function in a Neotropical forest

Maria C. Ruiz-Jaen1 and Catherine Potvin1-2 1 Department of Biology, McGill University, 1205 Dr Penfield, Montréal H3A-1B1, Québec, Canada 2 Smithsonian Tropical Research Institute, Apartado 0843-03092, Balboa, Panama

Status: Ruiz-Jaen, MC, Potvin C. 2010. Tree Diversity Explains Variation in Ecosystem Function in a Neotropical Forest. Biotropica 42: 638-646.

ABSTRACT Many experimental studies show that a decline in species number has a negative effect on ecosystem function, however less is known about this pattern in natural communities. We examined the relative importance of environment, space, and diversity on ecosystem function, specifically tree carbon storage in four plant types (understory/canopy; trees/palms), in a tropical forest in central Panama. The objectives of this study were to detect the relationship between tree diversity and carbon storage given the environmental and spatial variation that occur in natural forests and to determine which species diversity measure is more important to tree carbon storage: richness or dominance. We used redundancy analyses to partition the effect of these sources of variation on tree carbon storage. We showed that together environment, space, and diversity accounted for 43% of tree carbon storage, where diversity (19%) alone is the most important source of variation and explained more variation than space (13%) and environment (1%) together. Therefore, even in natural forests where substantial environment and spatial variation can be found, its still possible to detect the effect of diversity on ecosystem function at scales relevant to conservation. Moreover, both richness and dominance are important to explain the variation on tree carbon storage in natural forests suggesting that these two diversity measures are complementary. Thus, tree diversity is important to predict tree carbon storage in hyperdiverse forests.

Keywords: dominance; environment; Panama; principal coordinates of neighbor matrices (PCNM); redundancy analyses (RDA); space; species richness; tree carbon storage

INTRODUCTION While Species diversity conditions ecosystem functioning in natural communities like grasslands (Loreau and Hector 2001, Tilman et al. 2006), the relationship between diversity and ecosystem function in complex and hyperdiverse ecosystems like tropical forests is unclear (Srivastava and Vellend 2005). This relationship has been explored in the tropics either through simulations (Bunker et al. 2005; Balvanera et al. 2005) or by using native tropical tree plantations (e.g. Erskine et al. 2006, Healy et al. 2008). Simulations suggest that, in tropical forests, aboveground carbon stocks depend on species composition and on the identity and characteristics of the species being lost (Bunker et al. 2005) with a few species contributing disproportionately to carbon storage (Balvanera et al. 2005). Furthermore, tropical mixed species plantations tend to produce higher biomass than monocultures (Erskine et al. 2006, Healy et al. 2008). In natural forests, the effect of diversity on ecosystem function could be masked by the variability of environment and space (Huston and McBride 2002, Vila et al. 2005). Studies have often shown that abiotic factors such as soil and topographic characteristics (hereafter referred to as the environment) play a major role for species distribution (Tuomisto et al. 2003, John et al. 2007, Russo et al. 2008). Moreover, space has been shown to be more important than environment in determining woody plant species distribution (Svenning et al. 2004). Therefore to understand the role of species diversity in natural landscapes, we need to factor out the effects of environment and space (Chesson et al. 2002, Cardinale et al. 2004). In a tree diversity plantation established in Panama, the effect of environment on tree biomass was greater than that of tree diversity (Healy et al. 2008), a result that corroborates earlier findings by Huston (1997) for temperate grasslands. We build on these results to explore the relationship between diversity and ecosystem function in a natural tropical forest of Panama that harbors ~240 tree species. We used aboveground tree carbon storage as our measure of ecosystem function, since trees account for more than 90% of carbon in aboveground biomass (Nascimento and Laurance 2002, Chave et al. 2003, Kirby and Potvin 2007) and 35% to 85% of total ecosystem carbon storage (Kirby and Potvin 2007, Kraenzel et al. 2003). Our analysis focused on four plant groups (understory/canopy; trees/palms), since it has often been suggested that grouping of plants

into functional types would improve the ability to detect a relationship between diversity and ecosystem function, because certain functional traits are related to specific ecosystem function (Scherer-Lorenzen et al. 2007; De Deyn et al. 2008). We used partial redundancy analysis (Borcard et al. 2004; Healy et al. 2008) to answer the following questions: (1) does the environment and spatial variation in plant abundance mask the true relationship between diversity and tree carbon storage in four plant groups? If we can detect a relationship, (2) which diversity measure is more important to explain tree carbon storage, species richness or dominance? Answering these questions contributes to one of ecology’s most pressing challenges: What is the role of species diversity in determining ecosystem function in complex, diverse ecosystems?

METHODS Study site - Our study site is a lowland wet tropical forest located in San Lorenzo National Park on the Caribbean coast of Panama (9°17’N, 79°58’ W), locally known as Fort Sherman. The forest receives a mean annual precipitation of 3,100 mm with a short dry season (January to March; Santiago and Mulkey 2005). The soil is a Saprist (Histosol) with rich organic material lying on sedimentary substrate from Chagres Sandstone parent material (Pyke et al. 2001, Santiago et al. 2005). In comparison to other forests in the Panama Canal Watershed, the soils have high percent of carbon (C) and nitrogen (N) and low pH and bulk density (Santiago et al. 2005). The site includes a 5.95 ha L-shape forest dynamics plot where all stems ≥ 1 cm diameter at breast height (DBH) have been identified to species, measured and mapped in 1999 by the Center for Tropical Forest Science (CTFS; http://ctfs.si.edu/site/sherman/). Most of the plot (4.96 ha) is covered with of ~ 200 year old growth forest, while the rest is covered by secondary forest. To homogenize forest age as much as possible, we selected the old-growth forest portion of the CTFS plot and subdivided it in 124 subplots of 20 x 20 m for sampling purposes. Response matrix: Tree carbon storage - We chose aboveground tree carbon storage as an important measure of ecosystem function in tropical ecosystems, since they store one quarter of terrestrial carbon (Bonan 2008). To estimate the variance amongst subplots’ tree carbon storage, we first calculated the aboveground biomass (AGB) of individual trees using the allometric regression equation for wet tropical forests in Chave et al. (2005).

This equation multiplies the wood specific gravity (WSG) with the wood volume. Where the wood volume is estimated from relationships between volume and diameter at breast height (DBH). For WSG values, we used published (Santiago et al. 2004, Chave et al. 2006) and unpublished (S.J. Wright, unpublished data) material. When the WSG was not known for a species, we used 0.54 g/cm3 (Chave et al. 2003). We used the same allometric regressions models for both trees and palms individuals (Chave et al. 2008). We included the values of all trees (≥ 1 cm DBH) to avoid underestimation of total AGB in subplots (DeWalt and Chave 2004). For individuals with multiple stems, we calculated AGB of each stem and summed them. We then calculate tree carbon content, which is approximated 50 percent of the AGB (Hugues et al. 1999). Since tree carbon storage may simply be a function of stem density within a subplot, previous to the analysis, we determined the effect of total stem abundance and total tree carbon storage and found a non-significant relationship (Appendix 1). For the statistical analysis, we separated tree carbon storage in four plant types akin to functional groups with different capacity to store carbon aboveground (DeWalt and Chave 2004) and with different responses to environment and space (Vormisto et al. 2000, de Castilho et al. 2006). First, we separated palms from trees, since palms are an important component of the forest community in Fort Sherman (13% of total individuals). We further separated understory from canopy trees/palms based on species maximum height at maturity, since forests in Fort Sherman have a high proportion of understory stems (Santiago et al. 2004). This forest has an average canopy height of 36 m (Kurzel et al. 2006), thus, the vertical stratification was done at 10 m, where understory trees were species less than 10 m height at maturity and canopy trees were the remainders of the species. Each species maximum height was based on the CTFS classification of species in the Fort Sherman plot (S.J. Wright, unpublished data). Hereafter, we will use the term TREE carbon storage to refer to the carbon storage aboveground of trees and palms in the understory and canopy layer.

Explanatory Matrices Three sets of explanatory variables: environment, space, and diversity were used as independent variables to explain variation in TREE carbon storage. Each of the three sets

of independent variables served to construct an explanatory matrix, which is described below. Environment - To characterize the environment, we chose five abiotic variables that were measured, at the center, and at 5 meters East and West from the center for each subplot. These values were averaged to obtain one entry per subplot for each abiotic variable. To characterize topography, we classified slopes in a quadrant as: flat (1), gentle (2), and steep slope (3). Slope determined changes in AGB (Chave et al. 2003, de Castilho et al. 2006) and stand structure of tropical forests (Robert and Moravie 2003). We measured soil depth to have an estimate of tree growth belowground. We used a calibrated iron pole of 1.50 m, since most of the root biomass is present at this soil depth (Jackson et al. 1996). To characterize other soil physical properties, we measured the following variables at 0-10 cm depth: bulk density, texture, and color using standard protocols (see Appendix 2 for details in methodology). Soil bulk density is the measure of soil compaction and can decrease root density in the topsoil (Watson and Kelsey 2006). Soil texture can affect soil water retention and carbon storage (Silver et al. 2000). Soil color provides information of the mineral, organic, texture composition, and can be use as a surrogate of soil fertility (Ketterings and Bigham 2000, Fontes and Carvalho 2005). For example, soil color has been correlated with Fe oxides content (Fontes and Carvalho 2005), which is negatively correlated with phosphorus availability in soils (Agbenin 2003). Environmental variables were obtained for 117 of the 124 subplots, since two were around a canopy crane and five others were with large tree falls. These plots were excluded from the analysis (see Fig. 1 for the location of these plots). Space – To remove the methodological artifact of space and detect any spatially structured pattern, we estimated the spatial variation with principal coordinates of neighbor matrices (PCNM) analyses (Borcard et al. 2004). This method generates a set of orthogonal sine waves constructed from a truncated matrix of Euclidean distances among sampling units using the x and y coordinates from the center of each subplot (Fig. 1). Since, our study site was sampled irregularly (See Fig. 1), we filled the empty space by adding points where it was needed to avoid disruption of the sine waves (Borcard and Legendre 2002). PCNM reconstruct spatial patterns from fine to broad scales among the study site. For this study, the PCNM wavelengths ranged from 20 to 340 m. It is important to note that PCNM

scores allow us to detect spatial structure of environment, diversity and TREE carbon storage and to interpret these results; we need to graph these scores against the sampling sites coordinates. We generated the spatial variables (i.e. eigenvectors) by using the spacemakeR package in R (R Development Core Team 2008). Diversity - We considered two components of diversity: richness (Hooper et al. 2005) and dominance (Hillebrand et al. 2008). For richness, we calculated the number of species for each four-plant type, since we expected that the relationship between richness and ecosystem function might be very different between palms and trees due to intrinsic differences in growth pattern, morphology, and density. For dominance, we calculated the relative basal area (BA) of all species in a subplot and identified the species with the highest value, and selected the species that dominated the BA in more than eight subplots. Then, we determined the effect of both relative BA dominance and the identity of the dominant species by using the relative BA of six identified dominant species. Data analyses - We first reduced the covariation among environmental variables and examined their variability at the study site using a principal component analyses (PCA). We then quantified the proportion of the variation in TREE carbon storage explained by environment, space, and diversity. To do so, we selected subsets of space and diversity variables that exerted a significant effect on TREE carbon storage using forward selection with 999 random permutations (Blanchet et al. 2008). Forward selection chooses from a set of explanatory variables a parsimonious subset of variables to model multivariate response data. Then, using partial redundancy analyses (RDA; Borcard et al. 1992; Healy et al. 2008), we partitioned the variation in TREE carbon storage explained by environment, space, and/or diversity. RDA is the canonical extension of a multiple regression that analyzes multivariate response data (Legendre and Legendre 1998). To test for multicollinearity (i.e. correlation) among the explanatory variables in the RDA models, we used the vif.cca function in the vegan package in R (Oksanen et al. 2009). This function calculates the variance inflation factors for each variable in the model. For example, if variables are linear combinations or have an inflation factor of more than 10, they are removed from the estimation. Entries for the response matrix were: (1) C of understory trees, (2) C of canopy trees, (3) C of understory palms, and (4) C of canopy palms. The explanatory matrices

used after the forward selection were: (a) environment expressed as the scores for the first four axis obtained in the principal components analysis, (b) space (13 PCNM scores), and (c) diversity (understory tree richness, understory palm richness, canopy palms richness, canopy tree richness, relative BA of six tree species as dominance). Entries in the explanatory matrices were centered and standardized when necessary. The significant relative contribution of these three sources of variation was tested by a series of seven RDAs: (1) with all three explanatory subsets with no covariables, (2) for each explanatory matrix, using the other two explanatory matrices as covariables (e.g. environment was analyzed using space and diversity as covariables), and (3) with two subsets of explanatory matrices. To estimate the variation explains by each explanatory matrix, we used adjusted R2 to reduce the R2 biased towards sample size and number of predictors (Peres-Neto et al. 2006). Moreover, the significance of each RDA model was determined after 999 permutations using the anova.cca function in the vegan package in R (Oksanen et al. 2009).

RESULTS TREE carbon storage - TREE carbon storage at the subplot level ranged from 26.1 MgC/ha to 284 MgC/ha, most of this carbon was distributed among the canopy trees that on average represented 87% (94.24 ± 47.54 MgC/ha with a coefficient of variation (CV) of 50%; Fig. 1) of carbon storage, followed by understory trees that, on average represented, 11% (10.89 ± 15.03 MgC/ha with a CV of 138%) of carbon storage. Palms were the smallest component of TREE carbon storage. Understory palms carbon storage was on average 0.17 ± 0.12 MgC/ha with a CV of 74%, while canopy palms was 2.10 ±1.25 MgC/ha with a CV of 59%. Environment and Space - The first 4 PCA axes explained 75% of the environmental variation in Fort Sherman (Appendix 3). These principal components separated subplots by clay to sandy clay soils for axis 1 (29% of variance). Axis 2 separated subplots from low to high redness factor (soil color) and from shallow to deep soils (17% of variance). Axis 3 separated subplots from flat to steep slopes (15% of variance), and axis 4 separated subplots from compacted to soil with higher aeration (bulk density; 13% of variance; see Table S1 for values of soil physical variables). For space, subplots coordinates generated

113 orthogonal variables. After forward selection, we retained 13 PCNM variables that were significant to the TREE carbon storage. These PCNM variables were arbitrarily divided in broad (>150 m apart), medium (50 – 150 m apart), and fine scales (<50 m apart) based on their spatial pattern detected in relation to the x and y coordinates along the study site (see Appendix 4 for an example the spatial patterns detected by the PCNM scores at the three scales). Diversity - Total species richness ranged from 30 to 61 species per plot (400 m2), with an average of 52. Richness of understory trees ranged from 7 to 29 species and had an average of 17 species in a subplot and a CV of 25 percent. For the trees with the potential to reach the canopy, their richness varied from 19 to 43 species, with an average of 30 species and a CV of 14.8 percent. Palms were less species rich than trees, but varied greatly among subplots. Understory palms richness ranged from 1 to 7 species, with an average of 3 species and a CV of 41 percent. For canopy palms, we only found three species, with an average of 2.2 for subplot and a CV of 35.2 percent. The six dominant species in terms of their relative BA were all canopy trees with dominance values as high as 60% of total BA. The species with the highest relative BA were: Aspidosperma spruceanum Benth. Ex Mull. Arg. (n=12 subplots, average dominance = 31% of BA, range=20-50%, Apocynaceae, high wood density), Brosimum utile (Kunth) Pittier (n=36 subplots, average dominance = 34% of BA, range=22-60%, Moraceae, medium wood density), Calophyllum longifolium Willd. (n=9 subplots, average dominance = 35% of BA, range=20-60%, Clusiaceae, medium wood density), Manilkara bidentata (A. DC.) A. Chev. (n=11 subplots, average dominance = 32% of BA, range=19- 59%, Sapotaceae, high wood density), Tapirira guianensis Aubl. (n=8 subplots, average dominance = 31% of BA, range=23-45%, Anacardiaceae, low wood density), and Vochysia ferruginea Mart. (n=9 subplots, average dominance = 32% of BA, range=21- 49%, Vochysiaceae, low wood density). Variation partitioning - Together the variables of environment, space, and diversity that we retained explained 44% of the total variation in TREE carbon storage (Table 1, See Appendix 5 for description of explanatory variables). Notwithstanding, diversity alone is the most important source of variation, explaining more TREE carbon storage variation than environment and space together (Table 1). The first canonical axis of the RDA model

with the diversity matrix factoring out environment and space showed that carbon storage in canopy palms and trees increased with canopy palm richness (r=0.68) and when B. utile (r=0.38), C. longifolium (r=0.25), and M. bidentata (r=0.53) dominated BA, but decreased when T. guianensis dominated BA (r=- 0.34; Fig. 2a). For the second RDA axis, carbon storage in understory palms increased in subplots with high richness of understory palms (r=0.73) and trees (r=0.36), but decreased when V. ferruginea dominated BA (r=-0.25). In general, species richness had higher correlation coefficients with TREE carbon storage than the identity of the dominant species and its dominance values based on basal area. The environment-space RDA controlling the effect of diversity explained the variation of carbon storage in understory and canopy palms in the first axis and understory and canopy trees in the second axis (Fig. 2b). Understory palms carbon storage was high in sandy clay soils (r=-0.29) located at steeper slopes (r=-0.22) and varied spatially at medium scales (PCNM 42; r=0.28; PCNM 72; r=0.21; Fig. 2b). Canopy palms carbon storage was high in sites with deep soils with high redness factor (i.e. nutrient poor sites; r=0.30) and flat slopes (r=0.22) and varied at broad scales (PCNM 18; r=-0.50). Understory tree carbon storage increased spatially at medium scales (PCNM 84; r=0.33), but decreased at fine scales (PCNM 106; r=-0.44 and PCNM 113; r=-0.47). Carbon storage in canopy trees decreased spatially at medium (PCNM 29; r=-0.22) to fine (PCNM 106; r=-0.44) scales. In summary, spatial variables best explained the variation in the carbon storage in trees, while environmental factors are the best to explain carbon storage in palms.

DISCUSSION Does the environment and spatial variation in plant abundance mask the true relationship between diversity and tree carbon storage in four plant groups? To answer this question, we partition the variation of TREE carbon storage and show that diversity and space explained the highest proportions. Thus, the answer to our question is that, even in natural forests where substantial environment and spatial variation can be found, it is still possible to detect the effect of diversity on ecosystem function at scales relevant to conservation. Our results support the hypothesis that tree diversity is spatially structured and is driven in part by the environment. For example, environment only affected the carbon storage in

palms (Andersen et al. 2010), where slope (Svenning 1999, Goldsmith and Zahawi 2007) and soil texture (Costa et al. 2009) have been identified as the strong predictors. In Fort Sherman, understory palms dominates along riparian forest, which is characterize for steep slopes and sandy clay soils. However, when we factor out space and diversity, environment is not significant to explain TREE carbon storage. This finding runs contrary to previous studies that suggest that habitat heterogeneity is the main driver of ecosystem function in local communities or across larger scales (Healy et al. 2008; Jiang et al. 2009). A salient feature of the present analysis is the high proportion of variation explained by space, which suggests that ecosystem function is spatially structured (Borcard et al. 2004). This spatial structure could be caused by the interaction of space- environment and space-diversity (Table 1) and these interactions vary with the scale. Carbon storage in trees was affected at fine scales. This can be explained by the light guild of the neighboring trees that could reduce or increase individual tree carbon sequestration through competition or facilitation. For example, competition for light could occur if neighbors have similar light guilds (Uriarte et al. 2004). Another example of fine scale process is the distribution of tree fall gaps creating stands of different ages (Jones et al. 2008). These create differences in light environments within the forest and are important for AGB dynamics (Feeley et al. 2007). The presence of a gap can skew the DBH size class of different species according to their light growth requirement (Wright et al. 2003) and can promote the presence of palms (Svenning 1999). The change in carbon storage of palms at medium and broad scales could be explained by the spatially structured distribution of soil nutrients (Paoli et al. 2008) that can play a major role in determining aboveground biomass. The variation of tree carbon storage at medium to broad scales could be explained by the presence of large trees along the forest. Large trees are usually rare in Neotropical forests yet they account for a large proportion of aboveground biomass (Clark and Clark 1996, Chave et al. 2003). Moreover, seed dispersal would also affect the spatial distribution of species of trees and palms (Svenning 1999, Andersen et al. 2010) at medium to broad scales. Biodiversity-ecosystem functions experiments have likewise highlighted the spatial component of this relationship (Potvin and Dutilleul 2009, Weigelt et al. 2007) suggesting that spatial heterogeneity needs to be incorporated as a key explanatory factor of biodiversity.

Which diversity measure is more important to explain tree carbon storage, species richness or dominance? Our results indicate that species richness increases TREE carbon storage among subplots. We suggest that the increase in carbon storage in understory palms occurring with the increase in understory tree and understory palm richness may be due to an increase in light availability. Most of the understory palms species in Fort Sherman have clonal growth and their number of ramets increases with light availability (Chazdon 1986, Svenning 2000). High canopy palm richness could also increase light availability in the canopy, which might also explain why plots with high palm richness have high carbon storage in canopy trees. It has been shown that the mixture of tree and palm in the canopy layer creates different light environments that may enhance tree growth (Bohlman and O’Brien 2006). Overall, our results support the niche complementarity hypothesis (Loreau and Hector 2001) and suggest that among different resources (e.g. nutrients, light, and water; Wright 2002), light might play a crucial role in carbon sequestration. Several authors had previously suggested that the niche complementary hypothesis should be prevalent in tropical natural systems where a high number of species with weak species-species effects (Paine et al. 2008) coexist in a small area (Tylianakis et al. 2008) and in mature stands (Cardinale et al. 2007; Fargione et al. 2007). Beyond the effect of niche complementarity, the identity of the dominant species also explained subplot TREE carbon storage in Fort Sherman. This result is consistent with previous studies showing that the identity of dominant species, thus species-specific traits of the dominant species, ultimately has a large effect on carbon storage (Balvanera et al. 2005, Kirby and Potvin 2007) and productivity (Healy et al 2008), thus playing a key role in maintaining ecosystem functions (Smith and Knapp 2003). We propose that the effect of the dominant species depends on the distribution of the DBH size classes of the species in the study site and of key functional traits such as wood density (Baker et al. 2004). For example, Tapirira guianensis is a low wood density species with the DBH size classes skewed towards lower sizes, thus when it dominates a subplot, the overall carbon storage is likely to be low (Appendix 6). In contrast, subplots dominated by Brosimum utile are likely to have high carbon storage due to both its skewed size classes distribution towards large trees as well as its medium wood density values (Appendix 6). The occurrence of large trees along the landscape has been previously identified as good

predictors for AGB in tropical ecosystems (Clark and Clark 1996). Thus, our results also support the mass ratio hypothesis that predicts that ecosystem properties are controlled by the dominant species (Grime 1998; Smith and Knapp 2003). This hypothesis should occur in natural landscapes because some species become dominant along the mosaic of environmental patches through habitat specialization (Chesson et al. 2002; Cardinale et al. 2004). Our results suggest that species richness (niche complementarity hypothesis) had a greater explanatory power than dominance (mass ratio hypothesis) on TREE carbon storage (i.e. ecosystem function). The importance of these two components on ecosystem function has been a debate. For example, the properties of the dominant species have been the primarily mechanism to explain ecosystem function (Cardinale et al. 2006), however, as the forest stands get older, complementarity of species richness gain more relevance (Cardinale et al. 2007). We found that in an old-growth natural forest with high species diversity, these two hypotheses are not mutually exclusive.

CONCLUSION While a vast literature examines the relationship between biodiversity and ecosystem function in grasslands and microcosms (Hooper et al. 2005; Balvanera et al. 2006), our analysis presents the first empirical results for the relationship between diversity and ecosystem function in a natural tropical forest. Our analyses allowed us to rank the sources of variation for tree C storage in terms of their importance and we found that diversity and space were the two most important factors. We also highlight different responses of trees and palms to these independent variables. The distinction between these functional groups is often omitted despite the importance of palms in tropical forests (DeWalt and Chave 2004). Thus, future work in diverse natural forests should define different functional groups to determine the role of diversity in the ecosystem function.

ACKNOWLEDGEMENTS MCRJ was supported by IFARHU-SENACYT from the Panamanian Government and CP acknowledges a Discovery Grant from NSERC (Canada). FA Jones, MJ Lechowicz, KR Kirby, J Pelletier, M Peña-Claros and two anonymous reviewers provided helpful

comments. We thank L Mancilla, JA Quintero, D Gomez, J Leblanc, and S Bonilla for their help in the collection and processing of environmental data. We also thank S Lao for assistance with database management of CTFS and X Thibert-Plante for his help in R functions.

REFERENCES Agbenin, JO. 2003. Extractable Iron and Aluminum effects on phosphate sorption in a Savanna Alfisol. Soil Science Society of America Journal 67: 589-595. Andersen KM, Turner BL, Dalling JW. 2010. Soil-based habitat partitioning in understory palms in lower montane tropical forests. Journal of Biogeography 37: 278-292. Baker TR, Phillips OL, Malhi Y, Almeida S, Arroyo L, Di Fiore A, Eerwin T, Killen TJ, Laurance SG, Laurance WF, et al. 2004. Variation in wood density determines spatial patterns in Amazonian forest biomass. Global Change Biology 10: 545-562. Balvanera P, Kremen C, Martinez-Ramos M. 2005. Applying community structure analysis to ecosystem function: Examples from pollination and carbon storage. Ecological Applications 15: 360-375. Balvanera P, Pfisterer AB, Buchmann N, He JS, Nakasizuka T, Raffaeli D, Schmid B. 2006. Quantifying the evidence for biodiversity effects on ecosystem functioning and services. Ecology Letters 9: 1146-1156. Blanchet FG, Legrengre P, Borcard D. 2008. Forward selection of explanatory variables. Ecology 89: 2623-2632. Bohlman S, O’Brien S. 2006. Allometry, adult stature and regeneration requirement of 65 tree species on Barro Colorado Island, Panama. Journal of Tropical Ecology 22: 123- 136 Bonan GB. 2008. Forests and climate change: Forcings, feedbacks, and the climate benefits of forests. Science 320: 1444-1449. Borcard D, Legendre P. 2002. All-scale spatial analyses of ecological data by means of principal coordinates of neighbour matrices. Ecological Modelling 153: 51-68. Borcard D, Legendre P, Avois-Jacquet C, Tuomisto H. 2004. Dissecting the spatial structure of ecological data at multiple scales. Ecology 85: 1826-1832. Borcard D, Legendre P, Drapeau P. 1992. Partialling out the Spatial Component of

Ecological Variation. Ecology 73: 1045-1055. Bunker DE, Declerck F, Bradford JC, Colwell RK, Perfecto I, Phillips OL, Sankaran M, Naeem S. 2005. Species loss and aboveground carbon storage in a tropical forest. Science 310: 1029-1031. Cardinale BJ, Ives AR, Inchausti P. 2004. Effects of species diversity on the primary productivity of ecosystems: extending our spatial and temporal scales of inference. Oikos 104: 437-450. Cardinale BJ, Srivastava DS, Duffy JE, Wright JP, Downing AL, Sankaran M, Jouseau C. 2006. Effects of biodiversity on the functioning of trophic groups and ecosystems. Nature 443: 989-992. Cardinale BJ, Wright JP, Cadotte MW, Carroll IT, Hector A, Srivastava D, Loreau M, Weis JJ. 2007. Impacts of plant diversity on biomass production increase through time because of species complementarity. Proceedings of the National Academy of Science USA 104: 18123-18128 Chave J, Andalo C, Brown S, Cairns MA, Chambers JQ, Eamus D, Folster H, Fromard F, Higuchi N, Kira T, et al. 2005. Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145: 87-99. Chave J, Condit R, Lao S, Caspersen JP, Foster RB, Hubbell SP. 2003. Spatial and temporal variation of biomass in a tropical forest: results from a large census plot in Panama. Journal of Ecology 91: 240-252. Chave J, Condit R, Muller-Landau HC, Thomas SC, Ashton PS, Bunyavejchewin S, Co LL, Dattaraja HS, Davies SJ, Esufali S, et al. 2008. Assessing evidence for a pervasive alteration in tropical tree communities. Plos Biology 6: 455-462. Chave J, Muller-Landau HC, Baker TR, Easdale TA, ter Steege H, Webb CO. 2006. Regional and phylogenetic variation on wood density across 2456 neotropical tree species. Ecological Applications 16: 2356-2367. Chazdon R. 1986. Light variation and carbon gain in rain forest understory palms. Journal of Ecology 74: 995-1012. Chesson P, Pacala SW, Neuhauser C. 2002. Environmental niches and ecosystem functioning. In Kinzing AP, Pacala SW, Tilman D (Eds.). The functional consequences of biodiversity: empirical process and theoretical extensions, pp. 213-245. Princeton

University Press, Princeton, U.S.A. Clark DB, Clark DA. 1996. Abundance, growth, and mortality of very large trees in Neotropical lowland rain forests. Forest Ecology and Management 80: 235-244. Costa, FRC, Guillaumet J-L, Lima AP, Pereira OS. 2009. Gradient within gradients: the mesoscale distribution patterns of palms in a central Amazon forest. Journal of Vegetation Science 20: 69-78. De Castilho CV, Magnusson WE, De Araujo RNO, Luizao RCC, Lima AP, Higuchi N. 2006. Variation in aboveground tree live biomass in a central Amazonian Forest: Effects of soil and topography. Forest Ecology and Management 234: 85-96. De Deyn GB, Cornelissen JHC, Bardgett RD. 2008. Plant functional traits and soil carbon sequestration in contrasting biomes. Ecology Letters 11: 516-531. DeWalt SJ, Chave J. 2004. Structure and biomass of four lowland Neotropical forests. Biotropica 36: 7-19. Erskine PD, Lamb D, Bristow M. 2006. Tree species diversity and ecosystem function: Can tropical multi-species plantations generate greater productivity? Forest Ecology and Management 233: 205-210. Fargione J, Tilman D, Dybzinski R, Lambers JHR, Clark C, Harpole WS, Knops JMH, Reich PB, Loreau M. 2007. From selection to complementarity: shifts in the causes of biodiversity-productivity relationships in a long-term biodiversity experiment. Proceedings of the Royal Society: Biology 274: 871-876. Feeley KJ, Davies SJ, Ashton PS, Bunyavejchewin S, Nur Supardi MN, Kassim AR, Tan S, Chave J. 2007. The role of gap phase processes in the biomass dynamics of tropical forests. Proceedings of the Royal Society: Biology 274: 2857-2864. Fontes MPF, Carvalho IA. 2005. Color attributes and mineralogical characteristics, evaluated by radiometry, of highly weathered tropical soils. Soil Science Society of America Journal 69: 1162-1172. Goldsmith GR, Zahawi RA. 2007. The function of stilt roots in the growth strategy of Socratea exorrhiza (Arecaceae) at two neotropical sites. International Journal of Tropical Biology 55: 787-793. Grime JP. 1998. Benefits of plant diversity to ecosystems: immediate, filter and founder effects. Journal of Ecology 86: 902-910.

Healy C, Gotelli NJ, Potvin C. 2008. Partitioning the effects of biodiversity and environmental heterogeneity for productivity and mortality in a tropical tree plantation. Journal of Ecology 96: 903-913. Hillebrand H, Bennett DM, Cadotte MW. 2008. Consequences of dominance: A review of evenness effects on local and regional ecosystem processes. Ecology 89: 1510-1520. Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge DM, Loreau M, Naeem S, et al. 2005. Effects of biodiversity on ecosystem functioning: A consensus of current knowledge. Ecological Monographs 75: 3-35. Hugues RF, Kauffman JB, Jaramillo VJ. 1999. Biomass, carbon, and nutrient dynamics of secondary forests in a humid tropical region of Mexico. Ecology 80: 1892-1907. Huston MA. 1997. Hidden treatments in ecological experiments: Re-evaluating the ecosystem function of biodiversity. Oecologia 110: 449-460. Huston MA, McBride AC 2002. Evaluating the relative strengths of biotic versus abiotic controls on ecosystem processes. In M. Loreau, S. Naeem, and P. Inchausti (Eds.). Biodiversity and Ecosystem Functioning: Synthesis and Perspectives, pp. 36-46. Oxford University Press, Oxford, UK. Jackson RB, Canadell J, Ehleringer JR, Mooney HA, Sala OE, Schulze ED. 1996. A global analysis of root distributions for terrestrial biomes. Oecologia 108: 389-411. Jiang L, Wan S, Li L. 2009. Species diversity and productivity: why do results of diversity-manipulation experiments differ from natural patterns? Journal of Ecology 97:603-608. John R, Dalling JW, Harms KE, Yavitt JB, Stallard RF, Mirabello M, Hubbell SP, Valencia R, Navarrete H, Vallejo M, Foster RB. 2007. Soil nutrients influence spatial distributions of tropical trees species. Proceedings of the National Academy of Science U.S.A. 104: 864-869. Jones MM, Tuomisto H, Borcard D, Legendre P, Clark DB, Olivas PC. 2008. Explaining variation in tropical plant community composition: influence of environmental and spatial data quality. Oecologia 155: 593-604. Ketterings QM, Bigham JM. 2000. Soil color as an indicator of slash-and-burn fire severity and soil fertility in Sumatra, Indonesia. Soil Science Society of America Journal 64: 1826-1833.

Kirby KR, Potvin C. 2007. Variation in carbon storage among trees species: implications for the management of small-scale carbon sink project. Forest Ecology and Management 246: 208-221. Kraenzel M, Castillo A, Moore T, Potvin C. 2003. Carbon storage of harvest-age teak (Tectona grandis) plantations, Panama. Forest Ecology and Management 173: 213-225. Kurzel BP, Schnitzer SA, Carson WP. 2006. Predicting liana crown location from stem diameter in three Panamanian lowland forests. Biotropica 38: 262-266. Legendre P, Legendre L. 1998. Numerical Ecology. 2nd Edition. Elsevier Science B.V. Amsterdam, The Netherlands. Loreau M, Hector A. 2001. Partitioning selection and complementarity in biodiversity experiments. Nature 412: 72-76. Nascimento HEM, Laurance WF. 2002. Total aboveground biomass in central Amazonian rainforests: a landscape-scale study. Forest Ecology and Management 168: 311-321. Oksanen J, Kindt R, Legendre P, O’Hara B, Simpson GL, Solymos P, Henry M, Stevens H, Wagner H. 2009. Vegan: Community Ecology Package. R package version 1.15-4. R Foundation for Statistical Computing, Vienna, Austria. http://CRAN.R- project.org/package=vegan Paine CET, Harms KE, Schnitzer SA, Carson WP. 2008. Weak competition among tropical tree seedlings: Implications for species coexistence. Biotropica 40: 432-440. Paoli GD, Curran LM, Slik JWF. 2008. Soil nutrients affect spatial patterns of aboveground biomass and emergent tree density in southwestern Borneo. Oecologia 155: 287-299. Peres-Neto PR, Legendre P, Dray S, Borcard D. 2006. Variation partitioning of species data matrices: estimation and comparison of fractions. Ecology 87: 2614-2625. Potvin C, Dutilleul P. 2009. Neighborhood effects and size-asymmetric competition in a tree plantation varying in diversity. Ecology 90: 321-327. Pyke CR, Condit R, Aguilar S, Lao S. 2001. Floristic composition across a climatic gradient in a neotropical lowland forest. Journal of Vegetation Science 12: 553-566. R Development Core Team. 2008. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051- 07-0, URL http://www.R-project.org

Robert A, Moravie MA. 2003. Topographic variation and stand heterogeneity in a wet evergreen forest of India. Journal of Tropical Ecology 19: 697-707. Russo, SE, Brown P, Tan S, Davies SJ. 2008. Interspecific demographic trade-offs and soil-related habitat associations of tree species along resource gradients. Journal of Ecology 96: 192-203. Santiago LS, Goldstein G, Meinzer FC, Fisher JB, Machado K, Woodruff D, Jones T. 2004. Leaf photosythetic traits scale with hydraulic conductivity and wood density in Panamanian forest canopy trees. Oecologia 140: 543-550. Santiago LS, Mulkey SS. 2005. Leaf productivity along a precipitation gradient in lowland Panama: patterns from leaf to ecosystem. Trees 19: 349-356. Santiago LS, Schuur EAG, Silvera K. 2005. Nutrient cycling and plant-soil feedback along a precipitation gradient in lowland Panama. Journal of Tropical Ecology 21: 461-470. Scherer-Lorenzen M, Schulze ED, Don A, Schumacher J, Weller E. 2007. Exploring the functional significance of forest diversity: A new long-term experiment with temperate tree species (BIOTREE). Perspectives Plant Ecology Evolution and Systematics 9: 53- 70. Silver WL, Neff J, McGroddy M, Veldkamp E, Keller M, Cosme R. 2000. Belowground carbon and nutrient storage in a lowland Amazonian forest ecosystem. Ecosystems 3: 193-209. Smith MD, Knapp AK. 2003. Dominant species maintain ecosystem function with nonrandom species loss. Ecology Letters 6: 509-517. Srivastava DS, Vellend M. 2005. Biodiversity-ecosystem function research: Is it relevant to conservation? Annual Review Ecology, Evolution and Systematics 36: 267-294. Svenning JC. 1999. Microhabitat specialization in a species-rich palm community in Amazonian Ecuador. Journal of Ecology 87: 55-65. Svenning JC. 2000. Growth strategies of clonal palms (Arecaceae) in a neotropical rainforest, Yasuni, Ecuador. Australian Journal of Botany 48:167-178 Svenning JC, Kinner DA, Stallard RF, Engelbrecht BMJ, Wright SJ. 2004. Ecological determinism in plant community structure across a tropical forest landscape. Ecology 85: 2526-2538. Tilman D, Reich PB, Knops JMH. 2006. Biodiversity and ecosystem stability in a decade-

long grassland experiment. Nature 441: 629-632. Tuomisto H, Ruokolainen K, Yli-Halla M. 2003. Dispersal, environment, and floristic variation of western Amazonian forests. Science 299: 241-244. Tylianakis JM, Rand TA, Kahmen A, Klein AM, Buchmannn N, Perner J, Tscharntke T. 2008. Resource heterogeneity moderates the biodiversity-function relationship in real world ecosystems. Plos Biology 6: 947-956. Uriarte M, Condit R, Canham CD, Hubbell SP. 2004. A spatially explicit model of sapling growth in a tropical forest: does the identity of neighbours matter? Journal of Ecology 92: 348-360. Vila M, Inchausti P, Vayreda J, Barrantes O, Gracia C, Ibañez JJ, Mata T. 2005. Confounding factors in the observational productivity-diversity relationship in forests. In Scherer-Lorenzen M, Korner C, Schulze E-D. Forest Diversity and Function: Temperate and Boreal Systems, Vol. 177, pp. 65-86. Springer, Berlin, Vormisto J, Phillips OL, Ruokolaine K, Tuomisto H, Vasquez R. 2000. A comparison of fine-scale distribution of four plant groups in an Amazonian rainforest. Ecography 23: 249-359. Watson GW, Kelsey P. 2006. The impact of soil compaction on soil aeration and fine root density of Quercus palustris. Urban Forestry and Urban Greening 4: 69-74 Weigelt A, Schumacher J, Walther T, Bartelheimer M, Steinlein T, Beyschlag W. 2007. Identifying mechanisms of competition in multi-species communities. Journal of Ecology 95: 53-64. Wright SJ. 2002. Plant diversity in tropical forests: a review of mechanisms of species coexistence. Oecologia 130: 1-14. Wright SJ, Muller-Landau HC, Condict R, Hubbell SP. 2003. Gap-dependent recruitment, realized vital rates, and size distributions of tropical trees. Ecology 84: 3174-3185.

TABLES Table 1. Variation partitioning results from the redundancy analyses (RDA) on tree carbon storage of trees and palms in Fort Sherman. Lower case letters represent single fractions of variation: (a) environment (E), (b) space (S), (c) diversity (D), (d) E*S, (e) S*D, (f) E*D, and (g) E*S*D. Sources of variation Fractions included Df Adj. R2 F-ratio P-values

E + S + D [a+b+c+d+e+f+g] 27 0.43 4.20 <0.005

S + D [b+c+d+e+f+g] 23 0.42 4.39 <0.005

E + D [a+c+d+e+f+g] 14 0.30 3.64 <0.005

E + S [a+b+d+e+f+g] 17 0.24 2.46 <0.005

Decomposed variation (single fractions)

D controlled for E and S [c] 10 0.19 4.30 <0.005

S controlled for E and D [b] 13 0.13 2.76 <0.005

E controlled for S and D [a] 4 0.01 1.37 0.3

Unexplained 0.57

Df, degrees of freedom for each RDA model. We reported the adjusted R2 due to the bias associated with R2 values. The F-ratio and p-values were obtained after 999 permutations.

FIGURES Figure 1

Figure 1. Tree carbon storage in canopy trees distributed across the 124-20 x 20 m2 subplots in Fort Sherman, Panama. In parenthesis are the x and y coordinates used for the spatial analyses. Subplots marked with an X were not included in the statistical analysis.

Figure 2

Figure 2. Correlation biplots from redundancy analysis of tree carbon storage in understory trees and palms and canopy trees and palms constrained by the explanatory matrices of: (a) diversity and (b) environment and space (arrows) in Fort Sherman, Panama (see Appendix 5 for details in explanatory variables). Circles indicate the subplots (n=117). Crosses indicate the centroids of the response variables. Angles between tree carbon storage (variables in bold) and arrows of explanatory variables reflect their correlations. If the projection of tree carbon storage from the center of the axis is parallel to an arrow of an explanatory variable, then they are related. Arrow size is positive related to its effect level.

APPENDICES Appendix 1: TREE carbon storage is not related to total stem abundance in Fort Sherman (n = 117).

Appendix 2: Description of Soil Physical Variables Soil physical variables measured at the 20 × 20 m2 subplots in Sherman, Panama (N = 117). Mean values of bulk density (g/cm3), soil depth (cm), sand, silt, clay, and soil color (redness factor) with their coefficient of variation in parenthesis.

Soil physical Mean values Methods variables (CV)

Bulk Density 0.51 (31.5%) Oven-dried soils cores (31.4 cm3) at 105°C for 48 hours (Maynard & Curran 2008)

Soil Depth 105.77 (21.8%) Calibrated iron pole of 150 cm

Soil Texture Used the Buoycous method (Elliott et al. 1999)

Sand 28.28 (36.1%)

Silt 6.01 (40.9%)

Clay 65.71 (15.6%)

Texture type US soil texture chart (Juma 1999)

Clay Soil 94 subplots

Sandy Clay 23 subplots

Color, redness factor 3.45 (34.0%) Redness factor (RF) = ([10 – H] – C/V) (Fontes & Carvalho 2005). Where H, C, and V are the Munsell hue, chroma, and lightness value, respectively (Munsell 2000).

Appendix 3: Correlations among environmental variables and the ordination axes of a Principal Component Analyses (PCA). Correlations found among environmental variables with the first four ordination axes of a Principal Component Analyses (PCA). The PCA scores from these axes were used for the redundancy analyses to explain carbon storage in four functional groups in Fort Sherman, Panama. The percent of variation explained by each axis is in parenthesis.

Variable Axis 1 (30%) Axis 2 (17%) Axis 3 (15%) Axis 4 (13%)

Bulk density –0.37 0.41 0.33 –0.62

Clay 0.85 –0.22 0.10 –0.24

Sand –0.65 -0.03 –0.03 –0.40

Slope –0.08 –0.17 0.90 0.21

Soil color 0.25 0.69 –0.16 –0.02

Soil depth 0.17 0.70 0.21 0.35

Appendix 4. Maps to illustrate the scores for the Principal Component of Neighboring Matrix (PCNM) at broad, medium and fine scales. We arbitrary grouped the PCNM scores in three scales: broad (PCNM 18), medium (PCNM 33), and fine (PCNM 113) scale across the 124-20 x 20 m2 subplots in Fort Sherman, Panama. Filled squares indicate positive values and the open squares negative values with a proportional area equal to their absolute value. For example, we can see that at the broad scale a spatial pattern was only detected at the southwestern side of the study site. In contrasts, at fine scale, spatial pattern is detected at several locations across the study site.

Appendix 5. Description of environmental, spatial, and diversity variables. Description of environmental, spatial, and diversity variables used for the partial redundancy analyses. These entries were retained after forward selection. Abbreviations: BA – Basal Area, PCA – Principal Component Analysis, PCNM - Principal Coordinates of Neighbor Matrices.

Explanatory Variable name Description Range matrix

Environment Soil texture Scores from the PCA axis –3.05/2.20

Soil depth & color Scores from the PCA axis –2.87/4.02

Slope Scores from the PCA axis –1.45/3.15

Bulk density Scores from the PCA axis –2.18/2.19

Space PCNM 5, 18, 26 PCNM scores at broad scales (> 150 m)

PCNM 29, 33, 42, PCNM scores at medium 59, 72, 84 scales (50–150 m)

PCNM 93, 100, 106, PCNM scores at fine scales 113 (< 50 m)

Diversity - Rich. U. Trees Number of tree species with 2.65/5.39 Richness a max. height of < 10 m

Rich. C. Trees Number of tree species with 4.47/6.56 a max. height > 10 m

Rich. U. Palms Number of palm species 1.00/2.65 with a max. height < 10 m

Rich. C. Palms Number of palm species 0/3.00

with a max. height > 10 m

Diversity - A. spruceanum Relative BA of Dominance 0/50% Dominance B. utile Relative BA of Dominance 0/60%

C. longifolium Relative BA of Dominance 0/60%

M. bidentata Relative BA of Dominance 0/59%

T. guianensis Relative BA of Dominance 0/45%

V. ferruginea Relative BA of Dominance 0/49%

Appendix 6: Size class distribution of the dominant species in Fort Sherman, Panama. Logarithmic distribution found for the diameter breast height size classes for the six canopy tree species that dominates the basal area in Fort Sherman, Panama. The coefficient of skewness (g1) summarized the symmetry in lognormal distributions (Bendel et al. 1989).

REFERENCES (FOR APPENDICES) Bendel RB, Higgins SS, Teberg JE, Pyke DA. 1989. Comparison of skewness coefficient, coefficient of variation, and Gini coefficient as inequality measures within populations. Oecologia 78: 394-400. Elliott ET, Heil JW, Kelly EF, Monger HC. 1999. Soil structural and other physical properties. In Robertson GP, Coleman DC, Bledsoe CS, Sollins P (Eds.). Standard soil methods for long-term ecological research, pp. 74-85. Oxford University Press, New York, New York. Fontes MPF, Carvalho IA. 2005. Color attributes and mineralogical characteristics, evaluated by radiometry, of highly weathered tropical soils. Soil Science Society of America Journal 69: 1162-1172. Juma NG. 1999. The pedosphere and its dynamics: A systems approach to soil science. Salman Productions, Inc., Edmonton, CA. Maynard DG, Curran MP. 2008. Bulk Density Measurement in Forest Soils. In Carter MR, Gregorich EG (Eds.). Soil Sampling and Methods of Analysis, pp. 863-869. CRC Press, Boca Raton, Florida. Munsell. 2000. Munsell soil color charts. Macbeth Division of Kollmorgen Corporation, Baltimore, Maryland.

LINKING STATEMENT 1 In Chapter 1, I have detected an effect of tree diversity on tree carbon storage in naturally varying environments. In order to make generalizations of biodiversity- ecosystem function in tropical ecosystems, I compared a natural forest with a nearby experimental plantation in Chapter 2. An important finding of Chapter 1 was the different responses among plant functional groups. Therefore in Chapter 2, I compared the effect of species and functional diversity on tree carbon storage.

CHAPTER 2: Can we predict carbon stocks in tropical ecosystems from tree diversity? Comparing species and functional diversity in a plantation and a natural forest

Maria C. Ruiz-Jaen1 and Catherine Potvin1,2 1 Department of Biology, McGill University, 1205 Dr Penfield, Montréal H3A-1B1, Québec, Canada 2 Smithsonian Tropical Research Institute, Apartado 0843–03092, Balboa, Panama.

Status: Ruiz-Jaen MC, Potvin C. 2011. Can we predict carbon stocks in tropical ecosystems from tree diversity? Comparing species and functional diversity in a plantation and a natural forest. New Phytologist 189: 978-987.

ABSTRACT Linking tree diversity to carbon storage can provide further motivation to conserve tropical forests and to design carbon-enriched plantations. Here, we examine the role of tree diversity and functional traits in determining carbon storage in a mixed-species plantation and in a natural tropical forest in Panama. We used species richness, functional trait diversity, species dominance, and functional trait dominance to predict tree carbon storage across these two forests. Second, we compared the species ranking based on wood density, maximum diameter, maximum height, and leaf mass per area (LMA) between sites to reveal how these values changed between different forests. Increased species richness, a higher proportion of nitrogen fixers, and species with low LMA increase carbon storage in the mixed-species plantation, while a higher proportion of large trees and species with high LMA increased tree carbon storage in the natural forest. Furthermore, we found that tree species greatly vary in their absolute and relative values between study sites. Different results in different forests mean that we cannot easily predict carbon storage capacity in natural forests using data from experimental plantations. Managers should be cautious when applying functional traits measured in natural populations in the design C-enriched plantations.

Key words: dominance, functional trait diversity, functional traits, mixed-species plantations, Panama, species diversity, tree carbon storage, and tropical forests

INTRODUCTION To fully understand the effect of species loss within tropical ecosystems, we need to link measures of biodiversity to ecosystem functions, like carbon sequestration. The relationship between diversity and ecosystem function has been most intensively studied in temperate grasslands, where a decline in species number has negative effects on plant productivity (Tilman et al., 1996; Schwartz et al., 2000; Spehn et al., 2005). Similarly, tropical mixed-species plantations have greater total biomass and higher annual rates of carbon sequestration than monocultures (Erskine et al., 2006; Healy et al., 2008; Piotto, 2008). Moreover, the dominant species has been shown to partly determine carbon storage in tropical plantations and natural forests where depending on its functional characteristics it can increase or decrease carbon storage (Balvanera et al., 2005; Kirby & Potvin, 2007; Ruiz-Jaen & Potvin, 2010). There have been attempts to link tree diversity to carbon storage in tropical ecosystems using experimental plantations (Erskine et al., 2006; Healy et al., 2007), since they can allow for testing the mechanisms responsible for this link, but do not contain the natural variability that occurs in natural forests where species richness can vary by an order of magnitude or more (Scherer-Lorenzen et al., 2005). Thus, we do not know if results found in mixed-species plantations can be extrapolated to predict what is found natural forests. For example, the most diverse tropical plantations have on average eight species in a quarter of a hectare (Scherer-Lorenzen et al., 2005; Erskine et al., 2006; Piotto, 2008), while natural tropical forests have an average of 26 species in a similar area but with an upper limit of up to 300 or more species (Condit et al., 2005). If we can find a similar relationship between tree diversity and carbon storage in natural forests and experimental plantations, we can generalize the importance for conserving biodiversity in tropical forests. In the present study, we examined the relationship between tree carbon storage, species diversity, and functional diversity in a well-studied natural tropical forest in central Panama and compare results to a mixed-species plantation nearby. We are interested in understanding what explains variation in tree carbon storage to identify carbon sinks and help to decrease atmospheric CO2 concentrations (Malhi & Phillips, 2004; Houghton, 2005; Houghton et al., 2009). Tropical forests store more than one quarter of the

terrestrial carbon (Bonan, 2008; Houghton et al., 2009) and trees represent 90% of this (Kirby & Potvin, 2007). This carbon storage can be determined in part by species diversity (Balvanera et al., 2005; Spehn et al., 2005; Vila et al., 2007). Functional traits have also been proposed as an improved way to understand forest dynamics in hyperdiverse tropical forests (Wright et al., 2010), because functional traits considered the redundancy in function of species. This function can group species according to their resource use and life history strategies (Grime 2002). To link functional traits to a specific function, we need to select traits that are related to the ecosystem function of interest, in this case, carbon storage. Therefore, we choose wood density, maximum diameter at breast height (DBH), maximum height, leaf mass per area, and nitrogen fixers as functional traits related to tree carbon storage (Ishida et al., 2008). Wood density partly determines aboveground biomass (Baker et al., 2004) and correlates with growth rates and tree mortality (King et al., 2005; Chave et al., 2009), with high wood density being correlated with slow growth and longer-lived tree species. Maximum diameter is an indirect measure of the maximum aboveground biomass that can be attained by a species (Nelson et al. 1999; Chave et al., 2003) and a measure of light competition among species

(Sheil et al., 2006). Maximum height (Hmax) is also measure of light competition among species (Poorter et al., 2005; Moles et al., 2009). Leaf mass per area (LMA) is leaf biomass invested to produce the light-capturing foliar area and it is negatively correlated to photosynthetic rates (Thomas & Winner, 2002; Rozendaal et al., 2006; Poorter, 2009). The presence of root nodules in some trees to fix atmospheric nitrogen can increase the overall carbon uptake by plants (Cornelissen et al., 2003b; Hedin et al., 2009). With these data available, we asked the following questions. Does species richness, functional trait diversity, species dominance and functional trait dominance similarly explain proportions of tree carbon storage in a mixed-species plantation and in a natural forest? We are also interested in exploring how functional traits vary between sites within and among species. If functional traits can order species similarly in natural forest and mixed-species plantations, we can use the available information on functional traits from different natural forests to design plantations for carbon sequestration (Baroloto et al., 2010). Therefore, our second question was: How functional traits are differently ranked within tropical forests and plantations? Plantations can act as carbon sinks in areas

where deforestation is a threat for preserving native species and could provide an alternative source of timber (Canadell & Raupach, 2008; Paquette & Messier, 2010).

MATERIAL AND METHODS Study sites - This study was conducted in a natural forest on Barro Colorado Island (BCI) and a mixed-species plantation in Sardinilla less than 20 km apart with similar precipitation regimes. We used the same sampling protocols to collect wood density, maximum diameter, maximum height, and leaf mass per area for all the species present at both sites. Sardinilla: We used a high diversity experimental plantation established in a pastureland in Sardinilla, Panama (9˚18’N, 79˚38’W) in 2003. More than 800 trees were planted in 24 plots of 18 m x 18 m with treatments of 6, 9 or 18 species replicated 8 times (Scherer-Lorenzen et al., 2005). These species were randomly drawn from a pool of 28 species found in the nearby natural forest of BCI based on how common they were in midstory to canopy layer (i.e. more than 15 m of maximum height at maturity), their shade tolerance (i.e. light demanding, light intermediate, and shade tolerant), and their timber value (Delagrange et al., 2008). The maximum number of species planted in the experimental plantation was based on the average number of species larger than 10 cm DBH found in 20 x 20 m subplots in BCI (Scherer-Lorenzen et al., 2005). In this plantation, tree height, basal diameter at 10 cm from the ground and diameter at breast height of all individual has been recorded yearly since its establishment. For statistical analysis, we used the data of the census in 2009. The topography along the Sardinilla is almost uniformly flat, with a total elevation difference of 5 m (Potvin et al., 2004). This plantation receives a mean annual precipitation of 2,350 mm with very low rainfall reported during the months of January to March and with an annual mean temperature of 25 °C (Scherer-Lorenzen, 2007). The soil is an Alfisols dominated by clay over a Tertiary limestone and other sedimentary rocks (Potvin et al., 2004). As the natural forests, we used the nearby tropical moist forest of BCI that has a 50-ha permanent dynamic plot (9°09’N, 79°51’ W). This site is a seasonal forest with a mean annual rainfall of 2600 mm with very dry periods during January-April (Leigh, 1999). Soils are weathered kaolinitic Oxisols composed mostly of red light clay (John et al., 2007; Barthold et al., 2008). In the 50-ha permanent plot, every individual > 1 cm dbh

has been mapped and identified to species and has been censused every five years since 1985 (Hubbell & Foster, 1987). For statistical analysis, we used data of the last available census on BCI, 2005. To compare BCI with Sardinilla, we examined equivalent plot and tree sizes. For the plot size, we subdivided the 50-ha forest plot into plots of 20 x 20 m and excluded the ones along streams and with steep slopes (see Harms et al., 2001 for habitat classification). The small plot size helps to control for the effect of habitat heterogeneity, even though it increases carbon storage variation among plots (Chave et al., 2004). Moreover, the choosen plot size reflects the scale of individual tree competition, since neighbor effects are detected in radius of less than 20 m (Hubbell et al., 2001; Wiegand et al., 2007). For tree size, we restricted our analysis to species with a maximum height at maturity of more than 15 m, since the mixed-species plantation is composed only of midstory to canopy trees.

Tree Carbon Storage - We calculated the aboveground biomass (AGB) for individual trees using the following allometric regression for moist forest (Chave et al., 2005): AGB = WD*EXP (-1.499 + 2.148 ln (DBH) + 0.207 (ln (DBH))2 – 0.0281 (ln (DBH))3), where DBH is the diameter at breast height (1.30 m) in cm and WD is the wood density for each species in g cm-3. To determine the WD values in the experimental plantation, we collected wood cores of five individuals per species and followed the standardized protocols in Cornelissen et al. (2003b). S.J. Wright provided the WD values for the natural forest (Wright et al., 2010). For those species with no WD values, we used the average wood density, 0.54 g cm-3 for moist forests in Panama (Chave et al., 2003). For individuals with multiple stems, we calculated AGB of each stem and summed them. The estimation of AGB in the natural forests was done using the biomass function in the CTFS package in R (Hall, 2006). Then, we estimated tree carbon storage per plot using: C (Mg C ha-1) = AGB * 0.46/ plot area (Elias & Potvin, 2003).

Species and functional traits - We examined how species richness, functional diversity, and species dominance determined carbon storage in natural forest and mixed-species plantation. Species richness was estimated using rarefaction curves (Hurlbert, 1971), with

the minimum number of stems in the mixed-species plantation (19 individuals) as the sample size. For functional diversity, we calculated the dispersion for wood density, maximum diameter at breast height (DBH), maximum height (Hmax), leaf mass per area (LMA), and potential for nitrogen fixers (NF) based on the trait value of the species present at each plot. Functional dispersion for each functional trait comes from a species- species distance matrix, where it calculates the species average distance from the centroid of each plot. We weighted species distance by the basal area of the species within each plot (Laliberte & Legendre, 2010). We estimated the functional diversity using the FD package in R using the function dbFD (Laliberté, 2009). Wood density is the oven-dry mass divided by its fresh volume (Cornelissen et al., 2003b) and was calculated from wood cores of five individuals of each species (Wright et al., 2010). In Sardinilla (25 species), we chose the largest individuals of each species within the study site, while for BCI (~300 species); we measured individuals close to the study site. Maximum DBH and Height is the maximum size a species can reach at maturity. To calculate maximum DBH and Hmax, we selected the six largest individuals of each species in Sardinilla (25 species) and in the 50-ha plot of BCI (~300 species) based on previous census in 2009 and in 2005, respectively. Hmax was measured using a telescopic measuring pole for trees smaller than 15 m and a laser rangefinder and a clinometer for trees larger than 15 m (Wright et al. 2010). LMA is the leaf oven-dry weight divided by its fresh area (Cornelissen et al., 2003b) and was calculated from two leaves of five individuals per species in the mixed- species plantation and in the natural forests. Most of the leaves collected were exposed to full sunlight, as suggested by Cornelissen et al. (2003b). However, we collected leaves in the shade when a species did not have individuals that were fully exposed to the sun (for details see Wright et al., 2010). For nitrogen fixers, we assigned a one to species that had the potential to nodulate (i.e. observed having a symbiotic interaction with nitrogen fixing bacteria) based on field observation or existing literature. In the mixed-species plantation, we searched for active root-nodules from five individuals of each species identified as a nitrogen fixer in previous studies (Cornelissen et al., 2003b). In the natural forest, we classified nitrogen fixers based its potential for nodulation according to existing literature (De Faria et al., 1984; Sprent 2005; De Faria et al., 2010).

For species dominance, we calculated the relative basal area for all the species in a plot to account for the contribution to the total basal area of the species dominating the BA in the plot. Previous studies have shown that only few species accounted for more than 90% of total carbon storage (Walker et al., 1999; Balvanera et al., 2005). For functional dominance, we calculated the community weight mean (CWM) for wood density, maximum diameter, Hmax, LMA, and nitrogen fixing. CWM is the mean of each species trait value weighted by the relative basal area of the species at each plot (Lavorel et al., 2008). We estimated the CWM using the FD package in R (Laliberté, 2009).

Statistical Analysis - Before the statistical analysis, we controlled for the variation in stem density and light availability among plots. Stem density has been related to tree carbon storage since stem density is a tradeoff between stem size classes, where a high density of trees is related to smaller average stem size (Clark & Clark 2000, DeWalt & Chave 2004). Light availability in a plot can indicate the presence of a gap or can help to determine stand age. Different light levels can affect species traits (Rijkers et al., 2000) and can enhance tree growth (King et al., 2005). Light availability in the mixed-species plantation was measured taking hemispheric photographs from 1 m from the ground at the center of each plot and used the software, Gap Light Analyzer 2.0 (http://www.ecostudies.org/gla/), to obtain light availability. For the natural forest, we used existing data of light availability (see Ruger et al. 2009) that was calculated as a function of the vertical distribution of canopy density at six height intervals. We regressed tree carbon storage against stem density and light availability and used the model residuals as the dependent variable. To compare the mixed-species plantation in Sardinilla with the natural forest of Barro Colorado Island, we partitioned the variation of the dependent variable among: (1) species richness, (2) functional diversity, (3) species dominance, and (4) functional dominance (Borcard et al., 1992; Legendre, 2008). To estimate the variation explained by each independent matrix and assess model fit, we used adjusted R2, which controls for sample size and number of predictors (Peres-Neto et al., 2006). We performed these analyses using the vegan package in R (Oksanen et al., 2009). We selected independent variables using stepwise procedures to avoid multicollinearity and over-fitting the model (Crawley, 2007). We also used the variance inflator factor to select independent variables

and excluded the ones with values > 10 (Oksanen et al., 2009). After these variable selection procedures, we excluded diversity and dominance of wood density and maximum diameter and retained the relative basal area of two dominant species. To estimate functional trait variability between study sites, we compared species ranking based on four traits (wood density, maximum diameter, maximum height, and LMA) for species shared in Sardinilla and BCI. According to Garnier et al. (2001), no change in species ranking among sites implies that there is no trait plasticity in different environments (e.g. plantation vs natural forests), or that all species responded similarly to the environment. RESULTS Determinants of carbon storage - Tree carbon storage was highly variable at both study sites (Table 1). It ranges from 0.83 to 10.75 Mg C/ha in Sardinilla and from 12.91 to 856.40 Mg C/ha in BCI. Not surprisingly, carbon storage was much higher in BCI because of the young age of Sardinilla. It is interesting to note, that after only 6 years of growth some of the plantation plots had approached the minimum biomass observed in the forest plots. Nitrogen Fixers represent 32% of the species in the Sardinilla (from a total of 25 species), while 14% of the species in Barro Colorado Island (from a total of 157 species

≥ 15 cm dbh). Functional Diversity and the CWM of Hmax, LMA, and N fixers had similar coefficient of variation at the plot level in both sites (Table 1). Species richness, functional diversity, species dominance, and functional dominance together explained 67% of the variation of carbon storage in Sardinilla and 49% for BCI and their contribution to carbon storage variation differs between the two sites. Species richness explained most of the tree carbon storage variation in Sardinilla, but did not contribute significantly to carbon on Barro Colorado Island (Fig. 1). In contrast, Functional dominance explained most of the variation in the natural forest (Fig.

1) with CWM of Hmax as the variable most responsible for this pattern (Table 2). Another important source of variation was the interaction between species dominance and functional dominance; it explained 8% of carbon storage variation in Sardinilla and 12% in BCI. Functional diversity was only important for BCI, which explained 6% of carbon storage variation (Fig. 1; Table 2). Species Dominance explained higher proportion of tree carbon storage variation in

Sardinilla (12%) than in BCI (2%). The two species chosen after forward selection are non-pioneer species and allocated carbon differently (Table 2, 3). Anacardium excelsum has higher investment producing branches than Tabebuia rosea, which invests in vertical growth (Table 3). In BCI, T. rosea is more abundant (0.23% of the stems) than A. excelsum (0.02% of the stems), but achieved smaller DBH sizes (10.5 cm and 105.6 cm average DBH, respectively). In Sardinilla, these two species along with five others contribute to more than half of the basal area (Fig 2a). For BCI, A. excelsum, along with 12 other species, contributed half of the basal area and T. rosea was ranked as the 49th largest contributor to basal area (Fig 2b).

Species trait ranking - None of the traits give a consistent ranking of species between sites.

Wood density and Hmax values were consistently lower in Sardinilla than BCI (Fig. 3a,c). For maximum DBH and LMA trait values increase or decrease depending on the tree species (Fig. 3b,d). There is a large range of LMA values in BCI; by contrast, values were low to medium in Sardinilla (Fig. 3d). Based on species ranking by traits, we can point out species that perform very differently in the mixed-species plantation of Sardinilla and in the natural forest of BCI. For example, Astronium graveolens (AG), Dipteryx oleifera (DO), and Ormosia macrocalyx (OM) are consistently lower in the species ranking from BCI to Sardinilla (Fig. 3a-d). In contrast, Colubrina glandulosa (CG), Guazuma ulmifolia (GU), Inga punctata (IP), and Terminalia amazonia (TA) improved their species ranking from BCI to Sardinilla.

DISCUSSION Sources of Variation for Carbon Storage Tree carbon storage in the mixed-species plantation was mainly explained by species richness and in the natural forests by functional trait diversity. For example, species richness only mattered in Sardinilla, where increasing the number of species also increased carbon storage. This relationship was expected, since trees were planted according to different species functions including shade tolerance and nutrient acquisition ability (Scherer-Lorenzen et al., 2005). In contrast, we found no relationship between species

richness and carbon storage in the BCI forest. One explanation of the lack of relationship is that carbon storage in the natural forest may have reached saturation in species richness, an effect that can be found in high species richness treatments in experimental grasslands (Silver et al., 1996; Wright, 1996; Hooper et al., 2005). In those systems, less than 15 species are needed to reach the highest values of plant productivity (Hector et al., 2001; Balvanera et al., 2005). Saturation between species number and carbon storage can vary among sites and will depend on the niche overlap among species. Previous studies on Barro Colorado Island have found that most species have overlapping ecological niches along an environmental gradient within the same forest type (Wright, 1996; Harms et al., 2001; Cardinale et al., 2007). Functional dominance explained most of the variation of tree carbon storage on BCI, more specifically; we found that the more a canopy tree dominates (i.e. higher CWM of Hmax) a plot, the greater the increase carbon storage. Canopy trees have a higher probability of survival and can grow faster due to asymmetry in the size relative to their neighbors’ (Hubbell et al., 2001; Potvin & Dutilleul, 2009). Plots with low Hmax diversity, where most trees have similar Hmax, had higher tree carbon storage. In contrast, we can expect that sites with more variation in Hmax are associated with younger forests and forest gaps but could also be due to greater liana loads (Schnitzer & Carson, 2001), which reduces carbon storage. The diversity and the proportion of nitrogen fixers enhance carbon storage only in the mixed-species plantation in Sardinilla. Nitrogen content in the soil is lower and less variable in Sardinilla (15.56 mg kg-1 N; Zeugin et al., 2010) than in Barro Colorado Island (25.92 mg kg-1 N; John et al., 2007). In Sardinilla, the previous presence of pasturelands has altered the nitrogen cycle by reducing nitrogen content in the soil after forest cutting (Jordan, 1985; Silver et al., 2005). Thus, an increase in the proportion and diversity of nitrogen fixers in Sardinilla has increased carbon storage through soil improvement in these nutrient poor sites. Nitrogen fixers have been found to be facultative, where nitrogen fixation occurs in disturbed sites like in Sardinilla or in forest gaps (Barron et al. 2011). The high nitrogen soil-content BCI could suppress nitrogen fixation by trees by reducing the relative advantage of symbiotic fixation over direct uptake (Barron et al. 2011; Hedin et al., 2009). Moreover, preliminary data along a forest chronosequence have shown that

younger stands, like Sardinilla, have a higher occurrence of nodulation than old-growth forest (S. Batterman et al. unpublished). In both sites, the effect of LMA diversity on carbon storage was generally lower than the ones found for Hmax and nitrogen fixers. In the experimental plantation, we found that plots dominated by trees with low LMA have higher carbon storage. In general, species that have lower LMA have higher growth rates because of their fast resource acquisition (Reich et al., 2007; Poorter et al., 2009), these are also the species that have higher aboveground biomass in the experimental plantations. Thus, when they dominate the plot, there is more tree carbon storage. We found the opposite relationship in the natural forests, since the old-growth forest canopy is mostly dominated by late successional species with high LMA. Similarly, LMA values have been shown to increase with tree height (Rijkers et al., 2000; Poorter et al., 2009). The dominance of a given species alone is not a strong determinant for tree carbon storage. However, we still found significant effect of a species and this effect depend on the basal area each species occupied. For example, the positive effect of A. excelsum on carbon storage in the natural forest is mostly due to their capacity to reach large DBH even thought it has a low wood density. In natural forests, aboveground biomass in trees has been mostly determined by their diameter size and not by their wood density (Chave et al., 2004).

Species trait ranking There is little interspecific trait consistency among species present at both study sites. Others studies have found that their high intraspecific trait variability providing no clear pattern (Rozendaal et al., 2006; Albert et al., 2010). Changes in interspecific trait ranking have been observed from seedlings grown in greenhouse with plants found in natural environments (Cornelissen et al., 2003a). Species ranking based on single functional traits can be attributed to ontogenetic changes between study sites and to different responses of species to the environment (Garnier et al., 2001; Cornelissen et al., 2003a). For example, there are changes in wood density and LMA at different ontogenetic stages in other tropical forests (Rozendaal et al., 2006; Ollivier et al., 2007), where low wood density at early stages is associated with a higher proportion of sapwood relative to heartwood

(Chave et al., 2009). We would have expected to observe lower values for most of the functional traits in Sardinilla than in BCI due to ontogeny since this plantation has only been established for 6 yr. However, we were surprised to see no clear pattern in the species ranking by all functional traits between study sites. The convergence in the LMA values observed in Sardinilla can be explained by the higher light availability for each individual compared to the forest of Barro Colorado Island (Table 1). Similarly, some species in Sardinilla invested more in growing horizontally (i.e. achieved higher values for maximum DBH than Hmax) than vertically. Investing in tree height is a response of light competition in tropical forests (Poorter et al., 2005; Kooyman & Westoby, 2009). Thus, if light is not a limiting resource, tree species will not invest in height, but growing branches. Species that reached larger sizes in the mixed-species plantation of Sardinilla showed no clear similarities among their functional traits. However, we found these species had low LMA and low wood density in natural forest. These characteristics are associated with pioneer species, which have a tradeoff between high growth and low carbohydrate storage (Poorter & Kitajima, 2007; Chave et al., 2009). In contrast, the species that could be considered good carbon storers (e.g. Dipteryx oleifera and Astronium graveolens), based on their high wood density and Hmax in the natural forest, had low carbon storage in Sardinilla, even though, these species have high growth rates relative to their wood density (i.e. growth rate is negatively correlated to wood density) and are light demanding species (Condit et al., 2006). Wishnie et al. (2007) has found that these two species grow poorly in other mixed-species plantations in Panama. Can we use plant functional traits to design carbon-enriched plantations? We have to be cautious if we want to use information on functional traits from natural forest plots to design experimental high carbon plantation. However, some species have been identified as potential for mixed-species plantations, they can obtain larger sizes, are good for timber (Kirby & Potvin, 2007) and can store carbon for long periods of time (Chambers et al., 1998). Despite the poor grouping of species based on functional traits, we can recommend the use of Colubrina glandulosa, Guazuma ulmifolia, Hura crepitans, Tabebuia rosea, and Terminalia amazonia. Most of these have been reported to growth very well in both diameter and height in mixed-species plantations (Wishnie et al. 2007) and more

specifically T. amazonia and I. punctata have been use to reduce degraded areas dominated by grasslands (Kim et al., 2008; Craven et al. 2009).

ACKNOWLEDGMENTS Authors were supported by the Panamanian Government through the IFARHU-SENACYT (MCRJ) and the Canadian Government with the Discovery Grant from NSERC (CP). The F.H. Levinson Fund supported plant trait measurements for Barro Colorado Island (BCI). National Science and McArthur Foundations supported the 50-ha plot censuses. We thank Andy Jones, Richard Norby, and three anonymous reviewers for insightful comments on previous versions of the manuscript. We want to thank Cristina Salvador, David Brassfield, Sebastian Bernal, Paulino Villarreal, Eduardo Medina, Rolando Perez, Salomon Aguilar, and Javier Ballesteros for collecting traits for BCI data. Jose Monteza, Lady Mancilla, and Jurgis Sapijanskas collected traits and light availability in Sardinilla.

REFERENCES Albert CH, Thuiller W, Yoccoz NG, Soudant A, Boucher F, Saccone P, Lavorel S. 2010. Intraspecific functional variability: extent, structure and sources of variation. Journal of Ecology 98: 604-613. Baker TR, Phillips OL, Malhi Y, Almeida S, Arroyo L, Di Fiore A, Erwin T, Killeen TJ, Laurance SG, Laurance WF, et al. 2004. Variation in wood density determines spatial patterns in Amazonian forest biomass. Global Change Biology 10: 545-562. Balvanera P, Kremen C, Martinez-Ramos M. 2005. Applying community structure analysis to ecosystem function: Examples from pollination and carbon storage. Ecological Applications 15: 360-375. Baroloto C, Marcon E, Morneau F, Pavoine S, Roggy JC. 2010. Integrating functional diversity into tropical forest plantation designs to study ecosystem processes. Annals of Forest Science 67: 303. Barron, AR, Purves DW, Hedin LO. 2011. Facultative nitrogen fixation by canopy legumes in a lowland tropical forest. Oecologia 165: 511-520.

Barthold FK, Stallard RF, Elsenbeer H. 2008. Soil nutrient-landscape relationships in a lowland tropical rainforest in Panama. Forest Ecology and Management 255: 1135- 1148. Bonan GB. 2008. Forests and climate change: Forcings, feedbacks, and the climate benefits of forests. Science 320: 1444-1449. Borcard D, Legendre P, Drapeau P. 1992. Partialling out the Spatial Component of Ecological Variation. Ecology 73: 1045-1055. Canadell JG, Raupach MR. 2008. Managing forests for climate change mitigation. Science 320: 1456-1457. Cardinale BJ, Wright JP, Cadotte MW, Carroll IT, Hector A, Srivastava D, Loreau M, Weis JJ. 2007. Impacts of plant diversity on biomass production increase through time because of species complementarity. Proceedings of the National Academy of Sciences of the United States of America 104: 18123-18128. Chambers JQ, Higuchi N, Schimel JP. 1998. Ancient trees in Amazonia. Nature 391: 135- 136. Chave J, Condit R, Lao S, Caspersen JP, Foster RB, Hubbell SP. 2003. Spatial and temporal variation of biomass in a tropical forest: results from a large census plot in Panama. Journal of Ecology 91: 240-252. Chave J, Condit R, Aguilar S, Hernandez A, Lao S, Perez R. 2004. Error propagation and scaling for tropical forest biomass estimates. Philosophical Transactions of the Royal Society of London: Biological Series 359: 409-420. Chave J, Andalo C, Brown S, Cairns MA, Chambers JQ, Eamus D, Folster H, Fromard F, Higuchi N, Kira T, Lescure JP, Nelson BW, Ogawa H, Puig H, Riera B, Yamakura T. 2005. Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145: 87-99. Chave J, Coomes D, Jansen S, Lewis SL, Swenson NG, Zanne AE. 2009. Towards a worldwide wood economics spectrum. Ecology Letters 12: 351-366. Clark DB, Clark DA. 2000. Landscape-scale variation in forest structure and biomass in a tropical rain forest. Forest Ecology and Management 137: 185-198.

Condit R, Ashton P, Balslev H, Brokaw N, Bunyavejchewin S, Chuyong G, Co L, Dattaraja HS, Davies S, Esufali S, et al. 2005. Tropical tree α-diversity: Results from a worldwide network of large plots. Biologiske Skrifter 55: 565-582. Condit R, Ashton P, Bunyavejchewin S, Dattaraja HS, Davies S, Esufali S, Ewango C, Foster R, Gunatilleke I, Gunatilleke CVS, et al. 2006. The importance of demographic niches to tree diversity. Science 313: 98-101. Cornelissen JHC, Cerabolini B, Castro-Diez P, Villar-Salvador P, Montserrat-Marti G, Puyravaud JP, Maestro M, Werger MJA, Aerts R. 2003a. Functional traits of woody plants: correspondence of species rankings between field adults and laboratory-grown seedlings? Journal of Vegetation Science 14: 311-322. Cornelissen JHC, Lavorel S, Garnier E, Diaz S, Buchmann N, Gurvich DE, Reich PB, ter Steege H, Morgan HD, van der Heijden MGA, et al. 2003b. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian Journal of Botany 51: 335-380. Craven D, Hall J, Verjans JM. 2009. Impacts of herbicide application and mechanical cleanings on growth and mortality of two timber species in Saccharum spontaneum grasslands of the Panama Canal Watershed. Restoration Ecology 17: 751-761. Crawley M. 2007. The R Book. West Sussex, UK: John Wiley & Sons. De Faria, SM, Franco AA, De Jesus RM, Menandro MS, Baitello JB, Mucci ESF, Dobereiner J, Sprent JI. 1984. New nodulating legume trees from South-East Brazil. New Phytologist 98: 317-328. De Faria SM, Diedhiou AG, de Lima HC, Ribeiro RD, Galiana A, Castilho AF, Henriques JC. 2010. Evaluating the nodulation status of leguminous species from the Amazonian forest of Brazil. Journal of Experimental Botany 61: 3119-3127. Delagrange S, Potvin C, Messier C, Coll L. 2008. Linking multiple-level tree traits with biomass accumulation in native tree species used for reforestation in Panama. Trees- Structure and Function 22: 337-349. DeWalt SJ, Chave J. 2004. Structure and biomass of four lowland Neotropical forests. Biotropica 36: 7-19.

Elias M, Potvin C. 2003. Assessing inter-and intra-specific variation in trunk carbon concentration for 32 neotropical tree species. Canadian Journal of Forest Research 33: 1039-1045. Erskine PD, Lamb D, Bristow M. 2006. Tree species diversity and ecosystem function: Can tropical multi-species plantations generate greater productivity? Forest Ecology and Management 233: 205-210. Garnier E, Laurent G, Bellmann A, Debain S, Berthelier P, Ducout B, Roumet C, Navas ML. 2001. Consistency of species ranking based on functional leaf traits. New Phytologist 152: 69-83. Grime JP. 2002. Plant strategies, vegetation processes and ecosystem properties. Chichester, UK: John Wiley and Sons, Inc. Hall P 2006. CTFS: The CTFS Large Plot Forest Dynamics Analyses.In. R package, version 1.00. In: [WWW document], URL http://CRAN.R- project.org/src/contrib/Archive/CTFS/ [accessed on 30 Sept. 2010]. Harms KE, Condit R, Hubbell SP, Foster RB. 2001. Habitat associations of trees and shrubs in a 50-ha neotropical forest plot. Journal of Ecology 89: 947-959. Healy C, Gotelli NJ, Potvin C. 2008. Partitioning the effects of biodiversity and environmental heterogeneity for productivity and mortality in a tropical tree plantation. Journal of Ecology 96: 903-913. Hector A, Joshi J, Lawler SP, Spehn EM, Wilby A. 2001. Conservation implications of the link between biodiversity and ecosystem functioning. Oecologia 129: 624-628. Hedin LO, Brookshire ENJ, Menge DNL, Barron AR. 2009. The Nitrogen Paradox in Tropical Forest Ecosystems. Annual Review of Ecology Evolution and Systematics 40: 613-635. Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge DM, Loreau M, Naeem S, Schmid B, Setala H, Symstad AJ, Vandermeer J, Wardle DA. 2005. Effects of biodiversity on ecosystem functioning: A consensus of current knowledge. Ecological Monographs 75: 3-35. Houghton RA. 2005. Aboveground forest biomass and the global carbon balance. Global Change Biology 11: 945-958.

Houghton RA, Hall F, Goetz SJ. 2009. Importance of biomass in the global carbon cycle. Journal of Geophysical Research-Biogeosciences 114: G00e03. Hurlbert SH. 1971. The nonconcept of species diversity: a critique and alternative parameters. Ecology 52: 577-586 Hubbell SP, Ahumada JA, Condit R, Foster RB. 2001. Local neighborhood effects on long-term survival of individual trees in a neotropical forest. Ecological Research 16: 859-875. Hubbell SP, Foster R. 1987. Large-scale spatial structure of a Neotropical forest. Revista de Biología Tropical 35: 7-22. Ishida A, Nakano T, Yazaki K, Matsuki S, Koike N, Lauenstein DL, Shimizu M, Yamashita N. 2008. Coordination between leaf and stem traits related to leaf carbon gain and hydraulics across 32 drought-tolerant angiosperms. Oecologia 156: 193-202. John R, Dalling JW, Harms KE, Yavitt JB, Stallard RF, Mirabello M, Hubbell SP, Valencia R, Navarrete H, Vallejo M, Foster RB. 2007. Soil nutrients influence spatial distributions of tropical trees species. Proceedings of the National Academy of Sciences of the United States of America 104: 864-869. Jordan CF. 1985. Nutrient cycling in tropical forest ecosystems. Chichester, UK: John Wiley & Sons. Kim TJ, Montagnini F, Dent D. 2008. Rehabilitating abandoned pastures in Panama: Control of the invasive exotic grass, Saccharum spontaneum L., using artificial shade treatments. Journal of Sustainable Forestry 26: 192-203. King DA, Davis SJ, Nur Supardi MN, Tan S. 2005. Tree growth is related to light interception and wood density in two mixed dipterocarp forests of Malaysia. Functional Ecology 19: 445-453. Kirby KR, Potvin C. 2007. Variation in carbon storage among trees species: implications for the management of small-scale carbon sink project. Forest Ecology and Management 246: 208-221. Kooyman RM, Westoby M. 2009. Costs of height gain in rainforest saplings: main-stem scaling, functional traits, and strategy variation across 75 species. Annals of Botany 104: 987-993.

Laliberté E 2009. FD: measuring functional diversity from multiple traits.In. R package version 1.0-9. In [WWW document]. URL http://CRAN.R- project.org/web/packages/FD/index.html [accessed on 30 Sept. 2010]. Laliberté E, Legendre P. 2010. A distance-based framework for measuring functional diversity from multiple traits. Ecology 91: 299-305 Lavorel S, Grigulis K, McIntyre S, Williams NSG, Garden D, Dorrough J, Berman S, Quetier F, Thebault A, Bonis A. 2008. Assessing functional diversity in the field - methodology matters! Functional Ecology 22: 134-147. Legendre P. 2008. Studying beta diversity: ecological variation partitioning by multiple regression and canonical analysis. Journal of Plant Ecology 1: 3-8. Leigh EG. 1999. Tropical forest ecology: a view form Barro Colorado Island. Oxford, UK: Oxford University Press. Malhi Y, Phillips OL. 2004. Tropical forests and global atmospheric change: a synthesis. Philosophical Transactions of The Royal Society of London Series B-Biological Sciences 359: 549-555. Moles A, Warton D, Warman L, Swenson N, Laffan S, Zanne A. 2009. Global patterns in plant height. Journal of Ecology 97: 923-932. Nelson BW, Mesquita R, Pereira JLG, de Souza SGA, Batista GT, Couto LB. 1999. Allometric regressions for improved estimate of secondary forest biomass in the central Amazon. Forest Ecology and Management 117: 149-167. Oksanen J, Kindt R, Legendre P, O'Hara B, Simpson GL, Solymos P, Stevens MHH, Wagner H. 2009. vegan: Community Ecology Package.In: [WWW document] http://CRAN.R-project.org/package=vegan. Ollivier M, Baroloto C, Marcon E. 2007. A trait database for Guianan rain forest trees permits intra- and inter-specific constrasts. Annals of Forest Science 64: 781-786. Paquette A, Messier C. 2010. The role of plantations in managing the world's forests in the Anthropocene. Frontiers in Ecology and the Environment 8: 27-34. Peres-Neto PR, Legendre P, Dray S, Borcard D. 2006. Variation partitioning of species data matrices: estimation and comparison of fractions. Ecology 87: 2614-2625. Piotto D. 2008. A meta-analysis comparing tree growth in monocultures and mixed plantations. Forest Ecology and Management 255: 781-786.

Poorter H, Niinemets U, Poorter L, Wright IJ, Villar R. 2009. Causes and consequences of variation in leaf mass per area (LMA): a meta-analysis. New Phytologist 182: 565-588. Poorter L, Bongers F, Sterck FJ, Woll H. 2005. Beyond the regeneration phase: differentiation of heigth-light trajectories among tropical tree species. Journal of Ecology 93: 256-267. Poorter L, Kitajima K. 2007. Carbohydrate storage and light requirements of tropical moist and dry forest tree species. Ecology 88:1000-1011. Potvin C, Dutilleul P. 2009. Neighborhood effects and size-asymmetric competition in a tree plantation varying in diversity. Ecology 90: 321-327. Potvin C, Whidden E, Moore T. 2004. A case study of carbon pools under three different land-uses in Panama. Climatic Change 67: 291-307. Reich PB, Wright IJ, Lusk CH. 2007. Predicting leaf physiology from simple plant and climate attributes: A global GLOPNET analysis. Ecological Applications 17: 1982- 1988. Rijkers T, Pons TL, Bongers F. 2000. The effect of tree height and light availability on photoshynthetic leaf traits of four neotropical species differing in shade tolerance. Functional Ecology 14: 77-86. Rozendaal DMA, Hurtado VH, Poorter L. 2006. Plasticity in leaf traits of 38 tropical tree species in response to light; relationships with light demand and adult stature. Functional Ecology 20: 207-216. Ruger N, Huth A, Hubbell SP, Condit R. 2009. Response of recruitment to light availability across a tropical lowland rain forest community. Journal of Ecology 97: 1360-1368. Ruiz-Jaen MC, Potvin C. 2010. Tree Diversity Explains Variation in Ecosystem Function in a Neotropical Forest in Panama. Biotropica 42: 638-646. Scherer-Lorenzen M. 2007. Tree species richness affects litter production and decomposition rates in a tropical biodiversity experiment. Oikos 116: 2108-2124. Scherer-Lorenzen M, Potvin C, Koricheva J, Schmid B, Hector A, Bornik Z, Reynolds G, Schulze ED 2005. The Design of Experimental Tree Plantations for Functional Biodiversity Research. In: Scherer-Lorenzen M, Korner C, Schulze ED eds. Forest

diversity and function: temperate and boreal systems. Heidelberg, Germany: Springer, 347-376. Schnitzer A, Carson WP. 2001. Treefall gaps and the maintenance of species diversity in a tropical forest. Ecology 82: 913-919 Schwartz MW, Brigham CA, Hoeksema JD, Lyons KG, Mills MH, van Mantgem PJ. 2000. Linking biodiversity to ecosystem function: implications for conservation ecology. Oecologia 122: 297-305. Sheil D, Salim A, Chave J, Vanclay J, Hawthorne WD. 2006. Illumination-size relationships of 109 coexisting tropical forest tree species. Journal of Ecology 94: 494- 507. Silver WL, Brown S, Lugo AE. 1996. Effects of Changes in Biodiversity on Ecosystem Function in Tropical Forests. Conservation Biology 10: 17-24. Silver WL, Thompson AW, Reich A, Ewel JJ, Firestone MK. 2005. Nitrogen cycling in tropical plantation forests: Potential controls on nitrogen retention. Ecological Applications 15: 1604-1614. Spehn EM, Hector A, Joshi J, Scherer-Lorenzen M, Schmid B, Bazeley-White E, Beierkuhnlein C, Caldeira MC, Diemer M, Dimitrakopoulos PG, et al. 2005. Ecosystem effects of biodiversity manipulations in European grasslands. Ecological Monographs 75: 37-63. Sprent JI. 2005. Nodulated legume trees. In: Werner D, Newton WE, eds. Nitrogen fixation in agriculture, forestry, ecology, and the environment. Dordretch, NL: Springer, 113-141. Thomas SC, Winner WE. 2002. Photosynthetic differences between saplings and adult trees: an integration of field results by meta-analysis. Tree Physiology 22: 117-127. Tilman D, Wedin D, Knops J. 1996. Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature 379: 718-720. Vila M, Vayreda J, Comas L, Ibanez JJ, Mata T, Obon B. 2007. Species richness and wood production: a positive association in Mediterranean forests. Ecology Letters 10: 241-250.

Walker B, Kinzing AP, Langridge J. 1999. Plant attribute diversity, resilience, and ecosystem function: The nature and significance of dominant and minor species. Ecosystems 2: 95-113. Wiegand T, Gunatilleke CVS, Gunatilleke I, Huth A. 2007. How individual species structure diversity in tropical forests. Proceedings of the National Academy of Sciences of the United States of America 104: 19029-19033. Wishnie MH, Dent DH, Mariscal E, Deago J, Cedeño N, Ibarra D, Condit R, Ashton PMS. 2007. Initial performance and reforestation potential of 24 tropical tree species planted across a precipitation gradient in the Republic of Panama. Forest Ecology and Management 243: 39-49. Wright SJ 1996. Plant species diversity and ecosystem functioning in tropical forests. In: Orians GH, Dirzo R, Cushman JH eds. Biodiversity and Ecosystem processes in tropical forests. Berlin, DE: Springer-Verlag, 11-31. Wright SJ, Kitajima K, Kraft NJB, Reich PB, Wright IJ, Bunker DE, Condit R, Dalling JW, Davis SJ, Diaz S, et al. 2010. Functional traits and the growth-mortality tradeoff in tropical trees. Ecology 91: 3664-3674. Zeugin, F, Potvin C, Jansa J, Scherer-Lorenzen M. 2010. Is tree diversity an important driver for phosphorus and nitrogen acquisition of a young tropical plantation? Forest Ecology and Management 260: 1424-1433.

TABLES Table 1. Mean and coefficient of variation of tree carbon storage, stem density, light availability, species richness, functional diversity (standard deviation of maximum height

Carbon storage (Mg C ha-1) 5.44 (0.47) 146.80 (0.79)

Stem density (trees ha-1) 700.88 (0.18) 2371.73 (0.23)

Light availability (%) 0.10 (0.38) 0.01 (0.62)

Species richness 7.18 (0.28) 13.33 (0.14)

Functional diversity

Hmax (m) 0.76 (0.22) 0.66 (0.28)

LMA (g m-2) 0.51 (0.29) 0.58 (0.42)

NF 0.40 (0.47) 0.28 (0.23)

Species dominance1

Anacardium excelsum 0.10 (0.92) 0.48 (0.64)

Tabebuia rosea 0.23 (0.38) 0.04 (2.04)

Functional dominance

CWM Hmax (m) 28.26 (0.13) 28.82 (0.11)

CWM LMA (g m-2) 69.64 (0.07) 72.81 (0.14)

CWM NF 0.29 (0.68) 0.09(1.63)

1 Plots removed where these species were absent. CWM, community weighted mean.

Table 2. Pearson correlations (r) between tree carbon storage controlled by stem density and light availability with species richness, functional diversity, species dominance, and functional dominance in the mixed-species plantation of Sardinilla and the natural forest of Barro Colorado Island (BCI) in Central Panama. Carbon Storage (MgC ha-1) Sardinilla (n=24) BCI (n=725)

Species richness 0.544 ** -0.084 *

Functional diversity

Hmax 0.053 ns -0.315 ***

LMA 0.490 * -0.004 ns

NF 0.343 * -0.001 ns

Species dominance

Anacardium excelsum -0.485 * 0.171 ***

Tabebuia rosea -0.494 * 0.008 ns

Functional dominance

CWM Hmax -0.523 ** 0.628 ***

CWM LMA -0.450 * 0.130 ***

CWM NF 0.397 * 0.098 **

-2 Abbreviations: Hmax, maximum height (m); LMA, leaf mass per area (g m ); NF, nitrogen fixers; CWM, community weight mean weighted by basal area. Significance level: ns, nonsignificant; *, P < 0.05; **, P < 0.01, ***, P < 0.001

Table 3. Description of the two tree species used to explain the variation of carbon storage in the mixed-species plantation, Sardinilla and the natural forest of Barro Colorado Island (BCI) located in Central Panama. Description Anacardium excelsum Tabebuia rosea (Bertol.) A. (Bert. & Balb. ex Kunth) DC. Skeels Sardinilla BCI Sardinilla BCI Family Anacardiaceae Bignoniaceae Wood density (g cm-3) 0.31 (0.24) 0.42 (0.06) 0.39 (0.21) 0.49 (0.14) Maximum Diameter at 171.82 1775.67 149.62 (0.17) 752.20 Breast Height (mm) (0.14) (0.08) (0.13) Maximum Height (m) 10.45 35.11 (0.09) 10.20 (0.04) 40.42 (0.07) (0.02) LMA (g m-2) 82.82 98.26 (0.14) 116.06 (0.09) 101.15 (0.12) (0.22) Economic Use 1 Timber value High timber value and ornamental Traits values represented the average of the six largest individuals and values in parenthesis are the coefficient of variation (Wright et al. 2010). LMA, leaf mass per area. 1Information taken from Delagrange et al. (2008).

FIGURES Figure 1

Figure 1. Proportion of tree carbon storage variation explained after controlling for stem density and light availability in the mixed-species plantation of Sardinilla (black bars) and the natural forest of Barro Colorado Island (grey bars) located in Central Panama. Independent matrices were: species richness (SpDiv), functional diversity (FunDiv), species dominance (SpDom), and functional dominance (FunDom).

Figure 2

Figure 2. Species ranked by their dominance in the relative basal area in the mixed- species plantation of Sardinilla (a) and the natural forest of Barro Colorado Island (BCI; b), Central Panama. Values are average percent of basal area (BA) by species in each plot. Lines are 95% confidence intervals. Note that only the first 100 species of the natural forests are included, the rest are rare species that contribute very little to total BA in each plot. AE, Anacardium excelsum; TR, Tabebuia rosea.

Figure 3

Figure 3. Species ranking comparison based on four functional traits of the species shared across the mixed-species plantation (Sardinilla) and the natural forest (Barro Colorado Island, BCI). Traits are: wood density (WD; a), log maximum diameter at breast height (DBH; b), maximum height (c), and leaf mass per area (LMA; d). Species abbreviations: AG, Astronium graveolens; CO, Cedrela odorata; CG, Colubrina glandulosa; DO, Dipteryx oleifera, GU, Guazuma ulmifolia; HC, Hura crepitans; IP, Inga punctata; LS, Luehea seemannii; OM, Ormosia macrocalyx; TR, Tabebuia rosea; TA, Terminalia amazonia.

LINKING STATEMENT 2 In Chapter 2, I showed that the community weighted mean (CWM) of maximum height weighted by basal area, a measure of functional dominance, explained 63% of the variation on tree carbon storage in the natural forest. I hypothesized that the effect of CWM of maximum height was confounded with the effect of forest structure on tree carbon storage. Therefore in Chapter 3, I related forest structure to aboveground biomass and calculated functional diversity independent of forest structure. In Chapter 3, I also explored if the relationships found at 20 x 20 m scale in the previous chapters changed as the scale increase to 50 m x 50 m then to 100 m x 100 m, in other words, I asked if the BEF relationship was scale dependent.

CHAPTER 3: Aboveground biomass, forest structure and functional diversity: Patterns at different scales

Ruiz-Jaen, Maria C. 1, S. Joseph Wright 2, Catherine Potvin 1,2

1 Department of Biology, McGill University, 1205 Dr Penfield, Montréal H3A-1B1, Québec, Canada 2 Smithsonian Tropical Research Institute, Apartado 0843-03092, Balboa, Panama

Status: Ruiz-Jaen MC, Wright SJ, Potvin C. In preparation. Aboveground biomass, forest structure and functional diversity: Patterns at different scales.

ABSTRACT Over a decade of experimental studies in temperate grasslands has established that reducing species diversity affects ecosystem function. However, this conclusion may not apply to all ecosystems or across spatial scales that are relevant for conservation. For example, little is known about the relationship between biodiversity and ecosystem function in tropical ecosystems. We measured the effect of species diversity, functional diversity, forest structure and environmental heterogeneity on aboveground biomass at different spatial scales (0.04 ha, 0.25 ha, and 1-ha) in a moist tropical forest plot on Barro Colorado Island, Panama. Topography, soil nutrients, and species diversity alone had little effect on aboveground biomass, but functional evenness and aboveground biomass were negatively correlated at all spatial scales. Forest structure, as measured by basal area inequality, and aboveground biomass were strongly positively correlated at each spatial scale examined. Specifically, sites with a high basal area inequality also have high aboveground biomass independent of the spatial scale examined. We conclude that spatial variation of aboveground biomass in the BCI forest is largely explained by the interaction between the size that trees can attain and the identity of the functional group of those species that dominate the stand.

Key words: aboveground biomass, Barro Colorado Island, biodiversity, deciduousness, ecosystem functioning, functional diversity, Gini coefficient, leaf mass per area, maximum height, Panama, tropical forests, species diversity

INTRODUCTION Following the rationale that global environmental change and other anthropogenic activities may negatively effect species diversity (Sala et al. 2000), several studies have tried to understand the effect of species loss on ecosystem function (Balvanera et al. 2006, Cardinale et al. 2006). To date, most studies have reported a positive relationship between ecosystem function, e.g. plant productivity, with increasing species diversity (Vila et al. 2007, Paquette and Messier 2011). However, these relationships can often be asymptotic, with the steepest increase in ecosystem function observed as species richness increases from one to about 15 species (Balvanera et al. 2005). Despite strong evidence supplied by small-scale manipulative experimental studies, these results may not apply to all ecosystems or across spatial scales that are relevant for conservation (Lawler et al. 2002, Srivastava and Vellend 2005). In earlier studies, we have reported that, in wet and semi- wet forests of Panama, species diversity (Ruiz-Jaen and Potvin 2010) and functional diversity (Ruiz-Jaen and Potvin 2011) were positively correlated with aboveground biomass at small spatial scales (0.04 ha). Here we explore how spatial scale affects the biodiversity-ecosystem function by examining correlates of spatial variation in aboveground biomass within the 50-ha forest dynamics plot on Barro Colorado Island, Panama. The extent to which the relationship between biodiversity and ecosystem function remains constant with scale remains to be studied. A first analysis, by Roscher et al. (2005), shows that the relationship between species richness and aboveground productivity estimated at the scales of 0.0012 ha and 0.04 ha, was unchanged. The results however pertain to an experimental system, the Jena diversity experiment, with relatively little environmental variation. In natural environments, as the size of the study plot (i.e. grain size) increases, environmental heterogeneity also increases (Chust et al. 2006). Based on the assumption that species are spatially distributed along environmental gradients (Balvanera and Aguirre 2006), increasing environmental heterogeneity might promote species diversity by providing a range of conditions that allow for species coexistence. Chesson et al. (2002) proposed a conceptual model to tackle the biodiversity-ecosystem function relationship in spatially heterogeneous environments. They argue that, under the “sampling” hypothesis, in any given patch the best species will dominate while, if several

patches are aggregated, several species will coexist resulting in over yielding since no species is superior across all environmental conditions. The result is therefore undistinguishable from the prediction of the “complementarity” hypothesis states that species using resources differently in species rich-communities enhance an ecosystem function. The experimental design of Barro Colorado’s 50 ha plot allowed us to use plot data to scale up measures of biodiversity and of ecosystem function at scales between 0.04, 0.25, and 1 ha while maintaining a large sample sizes at each scale (n=50 ha at the scale of 1 ha). We hypothesized that measures of species diversity will have a greater explanatory power for aboveground biomass at large scale than at small spatial scale. We further hypothesized that the variability in ecosystem function, i.e. above ground biomass, will be larger if estimated at small spatial scale then at large ones since the performance of different dominant species could be different from the mean performance of a group of species (Chesson et al. 2002). Taking into account the results from a companion paper (Ruiz-Jaen and Potvin 2010), we considered the identity of the functional group (e.g. understory/canopy, trees/palms), environmental factors such as soil fertility (Laurance et al. 1999, Paoli et al. 2008), and topography (de Castilho et al. 2006) as additional sources of variation for tree aboveground biomass, a trait related to aboveground productivity itself a key ecosystem function. Earlier studies (Potvin and Dutilleul 2009, Ruiz-Jaen and Potvin 2011) have shown that aboveground biomass is strongly correlated with forest structure. Forest structure can be estimated using stem density, the basal area of different diameter at breast height (dbh) size classes (Chave et al. 2003, DeWalt and Chave 2004), total basal area (Malhi et al. 2006), or the presence of big trees (Clark and Clark 2000). We chose to quantify forest structure by quantifying basal area inequality amongst individuals within a plot using the Gini coefficient. This coefficient includes, in a single statistic, information about the distribution of basal area of individuals within a plot (Wittebolle et al. 2009). The Gini coefficient, developed in economics as a measure of wealth inequality (Castello and Domenech 2002), has been used in ecology as a measure of size hierarchies within populations (Weiner and Solbirg 1984, Bagchi 2007) or as a measure of community evenness (Wittebolle et al. 2009). Here we use the Gini coefficient to estimate tree size

evenness in plots of different sized areas. We examined the extent to which forest structure varies across spatial scales with the hope of further shedding light on the determinants of forest carbon stocks. Forest carbon stocks are increasingly considered an important ecosystem service in the context of climate change mitigation. Recently negotiations, under the UN Framework Convention of Climate Change (UNFCCC), have attracted attention to the possibility of financial transfers to developing countries to reduce historical rates of deforestation (Stickler et al. 2009, Venter et al. 2009).

METHODS Study site. – Our study site is the 50-ha Forest Dynamic Plot (FDP) of Barro Colorado Island (BCI), Panama (9°9’N, 79°51’ W). In the FDP, every individual > 1 cm dbh has been mapped and identified to species and the plot has been recensused every five years since 1985 (Chave et al. 2008). We used the 2005 census in our analyses. The 2005 census recorded 208,387 living tree species representing 299 species > 1 cm dbh . The plot is located in a seasonal moist tropical forest receiving circa 2600 mm of rain yearly with a mean annual temperature of 27°C (Dietrich et al. 1996, Leigh 1999). Most of the FDP is located at a plateau level circa 140 m.a.s.l. within an old-growth forests (Leigh et al. 2004). The soil in the FDP is mostly weathered kaolinitic Oxisol with red light clay (John et al., 2007; Barthold et al., 2008) with an andesitic cap on the plateau. Aboveground Biomass We calculated the aboveground biomass (AGB) for individual trees using the allometric regression for moist forest (Chave et al., 2005): AGB = WD*EXP (-0.099 + 2.148 ln (DBH) + 0.207 (ln (DBH))2 – 0.0281 (ln (DBH)3), where DBH is the diameter at breast height (1.30 m) in cm and WD is the wood density for each species in g cm-3. We determined WD for 259 of the 299 species present in Barro Colorado Island. WD was calculated from wood cores of five individuals of each species collected near the study site (Wright et al., 2010). Assuming a symmetrical trunk, wood cores were broken down into segments and weighted by the area of the annulus that each segment represents. Wood density is represented as the weighted average for each segment (Wright et al. 2010). For those species with no WD values, we used the WD values in Chave et al. (2006) or the average wood density, 0.55 g cm-3, of 259 species present in BCI (Wright et al. 2010). We

included the values of all trees (≥ 1 cm DBH) to avoid underestimation of total AGB in subplots (DeWalt and Chave 2004). For individuals with multiple stems, we calculated AGB of each stem. For palms, we used the allometric regression model for palms in Nascimento and Laurence (2002). The estimation of AGB was done using the biomass function in the CTFS package in R (Hall, 2006). Forest structure To calculate the Gini coefficient, we first calculated the Lorenz Curve (Lorenz 1905) by relating the cumulative number of individuals ordered by their basal area on the abscissa and the cumulative basal area on the ordinate. We then calculated the Gini coefficient for each subplot by using the equation G=A/(A+B), where A is the area that lies between the line of equality and the Lorenz curve over the total area under the line of equality (A + B; area below the Lorenz curve) (Gini 1921, Weiner and Solbirg 1984). The Gini coefficient ranges from zero to one, with Gini equaling zero when all the stems in a plot have the same basal area and it approaches one when a single individual has a disproportionally large basal area within a plot. The Gini coefficient was calculated using the reldist package in R (Handcock and Morris 1999). Species and functional diversity We used different components of diversity based on the number, equitability, and relative abundance of species and species functional groups. The number (i.e. richness) of species or functional group has been the most used measure to link diversity to a given ecosystem function (Hopper et al. 2005, Purvis and Hector 2000), since it is easy to interpret (Purvis and Hector 2000). Species or functional group richness, however, does not include any information of the abundance of the species present, thus giving the same weight to all the species or functional group present. To incorporate a measure of abundance and equitability of species or functional groups, we calculated Simpson’s Index for species or functional groups within a plot. Simpson’s Index is a measure of species or functional group evenness and dominance that represents the likelihood that two individuals choosen at random from a plot are from the same species or functional group (McCune and Grace 2002). All diversity indices were determined from abundance matrices of species or functional groups and were calculated using the vegan package in R (Oksanen et al. 2010).

Species were places within different functional groups based upon three key traits involved in carbon allocation by trees: maximum height, leaf mass per area (Cornelissen et al, 2003), and deciduousness (Bohlman 2010). These three traits are related to plant resource acquisition (Baker et al. 2004, Santiago et al. 2004) and growth-mortality tradeoffs (Wright et al. 2010). Maximum height is the maximum size a species can reach at maturity and it is related to competitive ability for light (King et al. 2006). To calculate height, we selected the six largest individuals in the 50-ha plot of BCI based on the census of 2005 (For details in methodology see Wright et al. 2010, Ruiz-Jaen and Potvin, 2011). Leaf mass per area (LMA) is the leaf oven-dry mass divided by its fresh area (Cornelissen et al., 2003) and was calculated from two leaves of five individuals per species. Most of the leaves collected were exposed to full sunlight as suggested by Cornelissen et al. (2003). However, we collected leaves in the shade when a species did not have individuals that were fully exposed to the sun (for details see Wright et al., 2010, Ruiz- Jaen and Potvin, 2011). Deciduousness of a species is a response to water availability and seasonal drought (Eamus 1999) Poorter and Markesteijn 2008). Deciduous species were classified as those that lose their leaves for periods of days to several weeks. Classification of species by level of deciduousness followed Perez (2008). We determined the functional group of each species by dividing the distribution of values among species for each continuous trait into three parts with equal numbers of species for the dicots (see Appendix 2 for distribution of traits by species). This produced 18 different functional groups, and there was at least one species in each group. We also classified palms as being understory and canopy, resulting in a total of 20 different functional groups. Palms were included as a separate functional group based on a companion study where understory and canopy palms affect aboveground biomass differently (Ruiz-Jaen and Potvin 2010). Environmental heterogeneity Topography and soil properties have been documented at the scale of 20 by 20 m plots within the 50-ha plot. Topography was taken from the CTFS R package (Hall 2006) and is a measure of mean slope determined from a 1-m resolution topographic mean. Soil properties have been collected by John et al. (2007) and included pH in water, and extracted available Al, B, Ca, Cu, Fe, K, Mg, Mn, P and Zn using Mehlich-III extracts.

Nitrogen was estimated colorimetrically on KCL extracts and N mineralization rates were estimated by incubating soil for 28 days (John et al. 2007). Since soil properties are highly correlated, we characterized them by the scores of the first three principal components (Appendix 1). For the 50 x 50 m and 100 x 100 m PCA scores, we averaged values of the 20 x 20 m plots contained within each higher spatial scale before performing their Principal Component Analysis. Data Analysis: We used type II regressions to relate aboveground biomass (AGB) to forest structure (Gini coefficient) using the lmodel2 package in R (Legendre 2008). We used variation partitioning to measure the proportion of AGB and forest structure variation explain by species diversity, functional diversity, and environment. For these analyses, we used the vegan package in R (Oksanen et al. 2010). All the analyses were done at three spatial scales: 0.04, 0.25, and 1 ha in the 50-ha plot of Barro Colorado Island (BCI), Panama. We treated each subplot as an independent measure, since Chave et al. (2004) have found no spatial autocorrelation of AGB at any spatial scale in BCI.

RESULTS Our results showed that two measures of ecosystem function, AGB and Gini coefficient of basal area inequality, are scale independent (Table 1). Aboveground biomass averages approximately 300 Mg ha-1 with a Gini coefficient of 88-90%. Variability in aboveground biomass among plots, however is highly scale dependent (Table 1, Fig. 1). Aboveground biomass ranged from 22-1909 Mgha-1 at the 0.04 ha scale, from 135-659 Mgha-1 at the 0.25 ha scale, and 174-439 Mgha-1 at the 1-ha scale. Aboveground biomass and Gini coefficient were highly correlated at all spatial scales (0.04 ha: R2=0.78, slope 5.75; 0.25 ha: R2=0.78, slope 4.78; 1-ha: R2=0.76, slope 4.14) (Fig. 2). Sites with a high Gini coefficient (i.e high basal area inequality) consistently had higher aboveground biomass than the ones where trees had similar basal areas. The effect of species richness, in contrast with both aboveground biomass and the Gini coefficient, is highly scale dependent with a three-fold increase as scale increases from 0.04 ha to 1 ha (Table 1). Interestingly, richness, whether estimated as species and functional richness, was more variable within plots and across scales than evenness (Table 1). Environmental variables (slope, soil fertility, and nitrogen availability) were also

100

highly variable within the plots at all spatial scales (Table 1). Forest structure measured as the Gini coefficient did not vary greatly among plots at any spatial scale (Table 1). The observed Gini coefficient ranges from 0.73 to 0.98 at the 0.04 ha scale, from 0.83 to 0.94 at the 0.25 ha scale and from 0.85 to 0.93 at the 1-ha scale. Understanding Aboveground Biomass and Forest Structure When comparing the effects of the environment, species diversity, and functional diversity, only functional diversity significantly affected aboveground biomass (Table 2). The effect of functional diversity on AGB was largely explained by the variation in species functional evenness (loadings of -0.78, -0.61, -0.65 at 0.04, 0.25, and 1 ha) and not by functional richness (loadings of -0.16, -0.34, -0.24 at 0.04, 0.25, and 1 ha). The full model explained only 3% of the variation at small spatial scale but explained as much as 38% of the variation in tree carbon storage at the scale of 1 ha (Table 1). Similar to AGB, the full model of environment, species diversity, and functional diversity explained less variation of forest structure at small scale (0.04 ha: R2 adj=0.04) than at large scales (1-ha: R2 adj=0.32) (Table 3). We found that there are higher basal area inequalities in flat areas (slope loadings: 0.04 ha: -0.20; 0.25 ha: -0.54, 1-ha: -0.54), fertile sites (PC1 loadings: 0.04 ha: 0.40; 0.25 ha: 0.62; 1-ha:0.72), sites with low species diversity (species richness loadings: 0.04 ha: -0.32; 0.25 ha: -0.37; 1-ha: -0.24), and low functional evenness (functional evenness loadings: 0.04 ha: -0.87; 0.25 ha: -0.45, 1-ha: - 0.34).The contribution of individual factors to forest structure varied with spatial scales, where functional diversity explained the variation at small to medium scales, but not at larger scales (Table 3).

DISCUSSION Globally, there is a strong positive relationship between aboveground biomass and species diversity largely because tropical ecosystems are both species-rich and carbon-rich (Strassburg et al. 2010). It is not clear however if the relationship persists at smaller spatial scale. Here we examined how the relationship between aboveground biomass and tree diversity changes with spatial scale. We found that environment and species diversity alone had little effect on aboveground biomass and that functional evenness and aboveground biomass were negatively correlated at all spatial scales. Forest structure and

101

aboveground biomass were strongly positively correlated at each spatial scale examined. Therefore, the spatial variation of aboveground biomass in the BCI forest is largely explained by the interaction between the size that trees can attain and the identity of the functional group of those species that dominate the stand. Our results have both theoretical and practical relevance. As pointed out in the introduction, Chesson et al. (2002) hypothesized that in spatially heterogeneous environment, the predictions made by invoking either the “Sampling” or the “Complementarity” effect (sensu Loreau and Hector 2001) would become identical. In accordance with the predictions of Chesson et al (2002), we hypothesized that the variance in aboveground biomass would decrease with scale as the “sampling” effect would be overrun by the “Complementarity” effect and that, consequently, measures of species diversity would have a greater explanatory power at large spatial scales within the BCI forest plot. The patterns observed for aboveground biomass fully support both hypotheses. Roscher et al (2005), in contrast with the present study, found that species- aboveground productivity relationships did not change with spatial scale in experimental grasslands. We see two possible explanations for the difference in results. On the one hand, our results from BCI show that the effect of environment and the interactions between environment and our different measures of diversity change in importance as scale increases. While environment alone was not a significant source of variation for neither aboveground biomass nor forest structure, the interactions of environment with both species and functional diversity were always statistically significant. Soil fertility and topography have been commonly suggested as determinants of aboveground biomass (AGB, Chave et al. 2003, Dewalt and Chave 2004, de Castilho et al. 2006, Paoli et al. 2008, Slik et al. 2010). BCI has relatively fertile soils in comparison to other tropical forests (Powers 2004, John et al. 2007). Slopes can have a negative (Chave et al. 2003), positive (de Castilho et al. 2006) or no (Losos et al. 2004) effect on AGB. For example, Losos et al. (2004) reported that for 12 forest plots along the tropics, topography was not a good predictor of tree density and basal area. Other studies on BCI have found that aboveground biomass (AGB) is higher on slopes (Chave et al. 2003, Muller-Landau et al. unpublished). We suggest that spatial heterogeneity might be higher over the BCI 50 ha plot than in the Jena experiment, since the latter was established in an agricultural field. If

102

environmental heterogeneity is an important determinant of spatial variation then the absence of a scale effect in the Jena experiment might be explained by reduced environmental heterogeneity at all spatial scale. On the other hand, Condit et al. (2000), showed that species distributions in the 50 ha plot of Barro Colorado was highly spatially aggregated. In fact the authors reported similar spatial aggregation in five other tropical forests and concluded that the spatial distribution of trees was not random. Therefore such non-random pattern of species distribution might explain the strong spatial patterns observed in our study as we scaled up from 0.04 ha, 0.25 ha to 1-ha. It has been said that tropical forests species-ecosystem function relationship might reach an asymptote relatively quickly (Wright 1996), which might prevent us from detecting an increase in ecosystem function (e.g. aboveground biomass) with increasing tree species diversity as tree species of different functional groups are added. An alternative way to characterize diversity is to use functional richness. Functional groupings classify species based on their similarities of response to the environment (Petchey and Gaston 2006) and life history traits (Grime 2001), among others. In very diverse ecosystems like the tropics, reducing the number of species to their functional groups could help to understand the link between tree diversity and carbon storage (Silver et al. 2000, Powers and Tiffin 2010). In a previous study, we found that functional dominance, measured as the community-weighted mean of maximum height and Leaf Mass per Area (LMA), increased AGB at small spatial scales on Barro Colorado Island (Ruiz-Jaen and Potvin 2011). Here we found that functional diversity has a greater explanatory power than species diversity for AGB. We also found that species richness increased with scale, a pattern that did not match by neither aboveground biomass nor forest structure. One possible explanation of this pattern is that the accumulation of species as spatial scale increases is largely due to understory species. For example, on Barro Colorado Island, as many other tropical forests, understory species are 80% of all species, comprise 75% of all stems (King et al. 2006), but contribute very little to aboveground biomass (Nascimento and Laurance 2002, Kirby and Potvin 2007, Ruiz-Jaen and Potvin 2010). Finally, we found that aboveground biomass and forest structure, measured by basal area inequality (i.e. a high Gini coefficient), are strongly positively correlated on

103

each spatial scale examined and that forest structure varies in response to the interaction between environmental characteristics, species and functional diversity. Favorable environmental conditions (soil fertility, flat areas), low species evenness and functional evenness increase basal area inequality. These conditions may favor differential growth of tree basal area, which increases aboveground biomass. Going back to the model of Chesson et al. (2002), we propose that certain combination of environmental conditions and species presence will determine the identity of the dominant species and that the species and functional identity of the dominant species will play a crucial role in determining the aboveground biomass of the plot. For example, Baroloto et al. (2011) found that aboveground biomass was mostly explained by variation in basal area in other Neotropical forests at medium scales. According to Potvin and Dutilleul (2009) forest structure in a tropical tree plantation is strongly influenced by light competition among individuals, where the growth of large trees is enhanced by differential size asymmetry of neighboring trees. Similarly, Stoll and Newbury (2005) found that competition for plants belowground causes asymmetric growth. Our analysis of forest structure therefore begins to provide a mechanism to explain the observed relationship between diversity and ecosystem function, in this case aboveground biomass.

Practical implications Tropical forests are underrepresented among studies that consider biodiversity and ecosystem function (Balvanera et al. 2006), despite the fact that they harbor more than half of terrestrial diversity (Grace and Meir 2009) and that they are globally highly threatened (Wright 2005, Malhi 2010). Tropical forests occupy only 12% of the terrestrial surface but hold over 40% of terrestrial carbon (Houghton et al. 2009) and their carbon emissions from deforestation and forest degradation represents 12-15% of global emissions (van der Werf 2009). Currently, there is an increasing incentive for tropical countries to have good estimates of aboveground biomass due to the potential for carbon-payments through reduced emissions from deforestation and forest degradation (REDD) (Baker et al. 2010). A future REDD mechanism could provide funding sources to tropical countries (Harvey et al. 2010). Some have argued that REDD could help preserve biodiversity (Diaz et al. 2009) while other suggested that protecting high carbon landscape might not be sufficient

104

to protect biodiversity (Venter et al. 2009). At the regional level, Stickler et al (2009) examined different scenarios for the Amazon and concluded that the spatial distribution of forest carbon to be protected or restores was vitally important to ensure biodiversity protection and other ecosystem services, such as water management. The results of our study, carried out at small to medium spatial scale confirm that indeed, there might be a tradeoff between species diversity and carbon. For countries who consider biodiversity conservation as an important co-benefit of REDD, care must be taken to identify and protect areas with both high carbon and high species richness.

ACKNOWLEDMENTS Authors were supported by the Panamanian Government through the IFARHU-SENACYT (MCRJ) and the Canadian Government with the Discovery Grant from NSERC (CP). The F.H. Levinson Fund supported plant trait measurements for Barro Colorado Island (BCI). National Science and McArthur Foundations supported the 50-ha plot censuses. Comments by Robin Chazdon and Andy Jones improved the manuscript. We want to thank Cristina Salvador, David Brassfield, Sebastian Bernal, Paulino Villarreal, Eduardo Medina, Rolando Perez, Salomon Aguilar, and Javier Ballesteros for collecting plant functional traits.

REFERENCES Bagchi, S. 2007. Relationship between size hierarchy and density of trees in a tropical dry deciduous forest of western India. Journal of Vegetation Science 18:389-394. Baker, T. R., O. L. Phillips, Y. Malhi, S. Almeida, L. Arroyo, A. Di Fiore, T. Erwin, T. J. Killeen, S. G. Laurance, W. F. Laurance, S. L. Lewis, J. Lloyd, A. Monteagudo, D. A. Neill, S. Patiño, N. C. A. Pitman, J. N. M. Silva, and R. V. Martinez. 2004. Variation in wood density determines spatial patterns in Amazonian forest biomass. Global Change Biology 10:545-562. Baker, T. R., J. P. G. Jones, O. R. R. Thompson, R. M. R. Cuesta, D. del Castillo, I. C. Aguilar, J. Torres, and J. R. Healey. 2010. How can ecologists help realise the potential of payments for carbon in tropical forest countries? Journal of Applied Ecology 47:1159-1165. Balvanera, P., C. Kremen, and M. Martinez-Ramos. 2005. Applying community structure analysis

105

to ecosystem function: Examples from pollination and carbon storage. Ecological Applications 15:360-375. Balvanera, P. and E. Aguirre. 2006. Tree diversity, environmental heterogeneity, and productivity in a Mexican tropical dry forest. Biotropica 38: 479-491. Balvanera, P., A. B. Pfisterer, N. Buchmann, J. S. He, T. Nakashizuka, D. Raffaelli, and B. Schmid. 2006. Quantifying the evidence for biodiversity effects on ecosystem functioning and services. Ecology Letters 9:1146-1156. Baroloto, C., S. Rabaud, Q. Moltos, L. Blancs, C. Fortunel, B. Herault, N. Davila, I. Mesones, M. Rios, E. Valderrama, and P.V.A. Fine 2011. Disentangling stand and environmental correlates of aboveground biomass in Amazonian forests. Global Change Biology 17: 2677-2688. Barthold, F. K., R. F. Stallard, and H. Elsenbeer. 2008. Soil nutrient-landscape relationships in a lowland tropical rainforest in Panama. Forest Ecology and Management 255:1135-1148. Bohlman S. A. 2010. Landscape patterns and environmental controls of deciduousness in forests of central Panama. Global Ecology and Biogeography 19: 376-385. Cardinale, B. J., D. S. Srivastava, J. E. Duffy, J. P. Wright, A. L. Downing, M. Sankaran, and C. Jouseau. 2006. Effects of biodiversity on the functioning of trophic groups and ecosystems. Nature 443:989-992. Castello, A., and R. Domenech. 2002. Human capital inequality and economic growth: Some new evidence. Economic Journal 112:C187-C200. Chave, J., R. Condit, S. Lao, J. P. Caspersen, R. B. Foster, and S. P. Hubbell. 2003. Spatial and temporal variation of biomass in a tropical forest: results from a large census plot in Panama. Journal of Ecology 91:240-252. Chave, J., C. Andalo, S. Brown, M. A. Cairns, J. Q. Chambers, D. Eamus, H. Folster, F. Fromard, N. Higuchi, T. Kira, J. P. Lescure, B. W. Nelson, H. Ogawa, H. Puig, B. Riera, and T. Yamakura. 2005. Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145:87-99. Chave, J., H. C. Muller-Landau, T. R. Baker, T. A. Easdale, H. Ter Steege, and C. O. Webb. 2006. Regional and phylogenetic variation on wood density across 2456 Neotropical tree species. Ecological Applications 16:2356-2367. Chave, J., R. Condit, H. C. Muller-Landau, S. C. Thomas, P. S. Ashton, S. Bunyavejchewin, L. L.

106

Co, H. S. Dattaraja, S. J. Davies, S. Esufali, C. E. N. Ewango, K. J. Feeley, R. B. Foster, N. Gunatilleke, S. Gunatilleke, P. Hall, T. B. Hart, C. Hernandez, S. P. Hubbell, A. Itoh, S. Kiratiprayoon, J. V. LaFrankie, S. L. de Lao, J. R. Makana, M. N. S. Noor, A. R. Kassim, C. Samper, R. Sukumar, H. S. Suresh, S. Tan, J. Thompson, M. D. C. Tongco, R. Valencia, M. Vallejo, G. Villa, T. Yamakura, J. K. Zimmerman, and E. C. Losos. 2008. Assessing evidence for a pervasive alteration in tropical tree communities. Plos Biology 6:455-462. Chesson, P., S. W. Pacala, and C. Neuhauser. 2002. Environmental niches and ecosystem functioning. Pages 213-245 in A. P. Kinzing, S. W. Pacala, and D. Tilman, editors. The functional consequences of biodiversity: empirical process and theoretical extensions. Princeton University Press. Chust, G., J. Chave, R. Condit, S. Aguilar, S. Lao, and R. Perez. 2006. Determinants and spatial modeling of tree beta-diversity in a tropical forest landscape in Panama. Journal of Vegetation Science 17:83-92. Clark, D. B., and D. A. Clark. 2000. Landscape-scale variation in forest structure and biomass in a tropical rain forest. Forest Ecology and Management 137:185-198. Condit, R., P. S. Ashton, P. Baker, S. Bunyavejchewin, S. Gunatilleke, N. Gunatilleke, S.P. Hubbell, R. B. Foster, A. Itoh, J. V. LaFrankie, H. S. Lee, E. Losos, N. Manokaran, ,R. Sukumar, T. Yamakura. 2000. Spatial patterns in the distribution of tropical tree species. Science 288:1414-1418. Cornelissen, J. H. C., S. Lavorel, E. Garnier, S. Diaz, N. Buchmann, D. E. Gurvich, P. B. Reich, H. ter Steege, H. D. Morgan, M. G. A. van der Heijden, J. G. Pausas, and H. Poorter. 2003. A handbook of protocols for standardized and easy measurement of plant functional traits worldwide. Australian Journal of Botany 51:335-380. de Castilho, C. V., W. E. Magnusson, R. N. O. de Araujo, R. C. C. Luizao, A. P. Lima, and N. Higuchi. 2006. Variation in aboveground tree live biomass in a central Amazonian Forest: Effects of soil and topography. Forest Ecology and Management 234:85-96. DeWalt, S. J., and J. Chave. 2004. Structure and biomass of four lowland Neotropical forests. Biotropica 36:7-19. Diaz, S., A. Hector, and D. A. Wardle. 2009. Biodiversity in forest carbon sequestration initiatives: not just a side benefit. Current Opinion in Environmental Sustainability 1:55-

107

60. Dietrich, W. E., D. M. Windsor, and D. Thomas. 1996. Geology, Climate, and Hydrology of Barro Colorado Island. Pages 21-46 in E. G. Leigh, A. S. Rand, and D. M. Windsor, editors. The Ecology of a Tropical Forest. Smithsonian Institution, Washigton, DC. Eamus D. 1999. Ecophysiological traits of deciduous and evergreen woody species in the seasonally dry tropics. Trends in Ecology and Evolution 14: 11-16. Gini, C. 1921. Measurement of Inequality of Incomes. The Economic Journal 31:124-126. Grace, J., and P. Meir. 2009. Tropica rain forests as old-growth forests. Pages 391-410 in C. Wirth, G. Gleixner, and M. Heimann, editors. Old-growth forests. Springer-Verlag, Heidelberg, Germany. Grime, J. P. 2001. Plant strategies, vegetation processes, and ecosystem properties. 2nd edition. Wiley, Chichester, UK. Hall, P. 2006. CTFS: The CTFS Large Plot Forest Dynamics Analyses. R package, version 1.00. Handcock, M. S., and M. Morris. 1999. Relative Distribution Methods in the Social Sciences. Springer, New York. Harms, K. E., R. Condit, S. P. Hubbell, and R. B. Foster. 2001. Habitat associations of trees and shrubs in a 50-ha neotropical forest plot. Journal of Ecology 89:947-959. Harvey, C. A., B. Dickson, and C. Kormos. 2010. Opportunities for achieving biodiversity conservation through REDD. Conservation Letters 3:53-61. Houghton, R. A., F. Hall, and S. J. Goetz. 2009. Importance of biomass in the global carbon cycle. Journal of Geophysical Research-Biogeosciences 114:G00E03. John, R., J. W. Dalling, K. E. Harms, J. B. Yavitt, R. F. Stallard, M. Mirabello, S. P. Hubbell, R. Valencia, H. Navarrete, M. Vallejo, and R. B. Foster. 2007. Soil nutrients influence spatial distributions of tropical trees species. Proceedings of the National Academy of Sciences of the United States of America 104:864-869. King, D., S. J. Wright, and J. H. Connell. 2006. The contribution of interspecific variation in maximum tree height to tropical and temperate diversity. Journal of Tropical Ecology 22:11-24. Kirby, K. R., and C. Potvin. 2007. Variation in carbon storage among trees species: implications for the management of small-scale carbon sink project. Forest Ecology and Management 246:208-221.

108

Laurance, W. F., P. M. Fearnside, S. G. Laurance, P. Delamonica, T. E. Lovejoy, J. M. Rankin-de Merona, J. Q. Chambers, and C. Gascon. 1999. Relationship between soils and Amazon forest biomass: a landscape-scale study. Forest Ecology and Management 118: 127-138. Lawler, S. P., J. J. Armesto, and P. Kareiva. 2002. How relevant to conservation are studies linking Biodiversity and ecosystem functioning? Pages 294-313. Functional Consequences of Biodiversity. Princeton University Press, Princeton. Legendre, P. 2008. lmodel2: Model II Regression. R package version 1.6-3. Leigh, E. G. 1999. Tropical Forest Ecology: A view from Barro Colorado Island. Oxford University Press, New York, USA. Leigh, E. G., S. Lao, R. Condit, S. Hubbell, R. Foster, and R. Perez. 2004. Barro Colorado Island Forest Dynamic Plot, Panama Pages 451-463 in E. Losos and E. G. Leigh, editors. Tropical Forest Diversity and Dynamism: Findings from a Large-Scale Plot Network. The University of Chicago Press, Chicago, USA. Loreau, M., and A. Hector. 2001. Partitioning selection and complementarity in biodiversity experiments. Nature 412:72-76. Lorenz, M. O. 1905. Methods of measuring the concentration of wealth. Journal of the American Statistical Association 9:209-219. Losos, E. C., P. S. Ashton, N. Brokaw, S. Bunyavejchewin, R. Condit, G. Chuyong, L. Co, H. S. Dattaraja, S. Davies, S. Esufali, C. Ewango, R. Foster, N. Gunatilleke, S. Gunatilleke, T. Hart, C. Hernandez, S. Hubbell, A. Itoh, R. John, M. Kanzaki, D. Kenfack, S. Kiratiprayoon, J. LaFrankie, H. S. Lee, I. Liengola, S. Lao, J. R. Makana, N. Manokaran, H. Navarrete, T. Ohkubo, R. Perez, N. Pongpattananurak, C. Samper, K. Sri-ngerbyang, R. Sukumar, I. F. Sun, H. S. Suresh, S. Tan, D. Thomas, J. Thompson, M. Vallejo, G. Villa Munoz, R. Valencia, T. Yamakura, and J. K. Zimmerman. 2004. The structure of tropical forests. Pages 69-78 in E. C. Losos and E. G. Leigh, editors. Tropical Forest Diversity and Dynamics. The University of Chicago Press, Chicago, USA. Malhi, Y. 2010. The carbon balance of tropical forest regions, 1990-2005. Current Opinion in Environmental Sustainability 2:237-244. Malhi, Y., D. Wood, T. R. Baker, J. Wright, O. L. Phillips, T. Cochrane, P. Meir, J. Chave, S. Almeida, L. Arroyo, N. Higuchi, T. J. Killeen, S. G. Laurance, W. F. Laurance, S. L. Lewis, A. Monteagudo, D. A. Neill, P. N. Vargas, N. C. A. Pitman, C. A. Quesada, R.

109

Salomao, J. N. M. Silva, A. T. Lezama, J. Terborgh, R. V. Martinez, and B. Vinceti. 2006. The regional variation of aboveground live biomass in old-growth Amazonian forests. Global Change Biology 12:1107-1138. McCune, B., and J. B. Grace. 2002. Analysis of Ecological Communities. MjM Software Design, Gleneden Beach, Oregon. Nascimento, H. E. M., and W. F. Laurance. 2002. Total aboveground biomass in central Amazonian rainforests: a landscape-scale study. Forest Ecology and Management 168:311-321. Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, R. B. O'Hara, G. L. Simpson, P. Solymos, M. H. H. Stevens, and H. Wagner. 2010. vegan: Community Ecology Package. R package version 1.17-3. http://CRAN.R-project.org/package=vegan. Paoli, G. D., L. M. Curran, and J. W. F. Slik. 2008. Soil nutrients affect spatial patterns of aboveground biomass and emergent tree density in southwestern Borneo. Oecologia 155:287-299. Paquette, A., and C. Messier. 2011. The effect of biodiversity on tree productivity: from temperate to boreal forests. Global Ecology and Biogeography 20:170-180. Perez, R. A. 2008. Arboles de los bosques del Canal de Panama. Smithsonian Tropical Research Institute. Panama, Panama. 466 pp. Petchey, O. L., and K. J. Gaston. 2006. Functional diversity: back to basics and looking forward. Ecology Letters 9:741-758. Poorter, L. and L. Markesteijn. 2008. Seedling traits determine drought tolerance of tropical tree species. Biotropica 40: 321-331. Potvin, C and P. Dutilleul. 2009. Neighborhood effects and size-asymmetric competition in a tree plantation varying in diversity. Ecology 90: 321-327. Powers, J. S. 2004. New perspectives in comparative ecology of Neotropical rain forests: Reflections on the past, present, and future. Biotropica 36:2-6. Powers, J. S., and P. Tiffin. 2010. Plant functional type classifications in tropical dry forests in Costa Rica: leaf habit versus taxonomic approaches. Functional Ecology 24:927-936. Purvis, A., and A. Hector. 2000. Getting the measure of biodiversity. Nature 405:212-219. Roscher, C., V. M. Temperton, M. Scherer-Lorenzen, M. Schmitz, J. Schumacher, B. Schmid, N. Buchmann, W. W. Weisser, and E. D. Schulze. 2005. Overyielding in experimental

110

grassland communities - irrespective of species pool or spatial scale. Ecology Letters 8:419-429. Ruiz-Jaen, M. C., and C. Potvin. 2010. Tree diversity explains variation in ecosystem function in a Neotropical forest in Panama. Biotropica 42:638-646. Ruiz-Jaen, M. C., and C. Potvin. 2011. Can we predict carbon stocks in tropical ecosystems from tree diversity? Comparing species and functional diversity in a plantation and a natural forest. New Phytologist 189:978-987. Sala, O. E., F. Stuart Chapin III, Juan J. Armesto, Eric Berlow, Janine Bloomfield, Rodolfo Dirzo, Elisabeth Huber-Sanwald, Laura F. Huenneke, Robert B. Jackson, Ann Kinzig, Rik Leemans, David M. Lodge, Harold A. Mooney, Martõ «n Oesterheld, N. LeRoy Poff, Martin T. Sykes, Brian H. Walker, M. Walker, and D. H. Wall. 2000. Biodiversity - Global biodiversity scenarios for the year 2100. Science 287:1770-1774. Santiago, L. S., K. Kitajima, S. J. Wright, and S. S. Mulkey. 2004. Coordinated changes in photosynthesis, water relations, and leaf nutritional traits of canopy trees along a precipitation gradient in lowland tropical forest. Oecologia 139: 495-502. Silver, W. L., J. Neff, M. McGroddy, E. Veldkamp, M. Keller, and R. Cosme. 2000. Effects of soil texture on belowground carbon and nutrient storage in a lowland Amazonian forest ecosystem. Ecosystems 3:193-209. Slik, J. W. F., S. I. Aiba, F. Q. Brearley, C. H. Cannon, O. Forshed, K. Kitayama, H. Nagamasu, R. Nilus, J. Payne, G. Paoli, A. D. Poulsen, N. Raes, D. Sheil, K. Sidiyasa, E. Suzuki, and J. van Valkenburg. 2010. Environmental correlates of tree biomass, basal area, wood specific gravity and stem density gradients in Borneo's tropical forests. Global Ecology and Biogeography 19:50-60. Srivastava, D. S., and M. Vellend. 2005. Biodiversity-ecosystem function research: Is it relevant to conservation? Annual Review of Ecology Evolution and Systematics 36:267-294. Stickler, C. M., D. C. Nepstad, M. T. Coe, D. G. McGrath, H. O. Rodrigues, W. S. Walker, B. S. Soares-Filho, and E. A. Davidson. 2009. The potential ecological costs and cobenefits of REDD: a critical review and case study from the Amazon region. Global Change Biology 15: 2803-2824. Strassburg, B. B. N., A. Kelly, A. Balmford, R. G. Davies, H. K. Gibbs, A. Lovett, L. Miles, C. D. L. Orme, J. Price, R. K. Turner, and A. S. L. Rodrigues. 2010. Global congruence of

111

carbon storage and biodiversity in terrestrial ecosystems. Conservation Letters 3:98-105. van der Werf, G. R., D. C. Morton, R. S. DeFries, J. G. J. Olivier, P. S. Kasibhatla, R. B. Jackson, G. J. Collatz, and J. T. Randerson. 2009. CO2 emissions from forest loss. Nature Geoscience 2:737-738. Venter, O., W. F. Laurance, T. Iwamura, K. A. Wilson, R. A. Fuller, and H. P. Possingham. 2009. Harnessing Carbon Payments to Protect Biodiversity. Science 326:1368-1368. Vila, M., J. Vayreda, L. Comas, J. J. Ibanez, T. Mata, and B. Obon. 2007. Species richness and wood production: a positive association in Mediterranean forests. Ecology Letters 10:241- 250. Weiner, J., and O. T. Solbrig. 1984. The meaning and measurement of size hierarchies in plant- populations Oecologia 61:334-336. Wittebolle, L., M. Marzorati, L. Clement, A. Balloi, D. Daffonchio, K. Heylen, P. De Vos, W. Verstraete, and N. Boon. 2009. Initial community evenness favours functionality under selective stress. Nature 458:623-626. Wright, S. J. 1996. Plant Species Diversity and Ecosystem Functioning in Tropical Forests. Wright, S. J. 2005. Tropical forests in a changing environment. Trends in Ecology & Evolution 20:554-559. Wright, S. J., K. Kitajima, N. J. B. Kraft, P. B. Reich, I. J. Wright, D. E. Bunker, R. Condit, J. W. Dalling, S. J. Davies, S. Diaz, B. M. J. Engelbrecht, K. E. Harms, S. P. Hubbell, C. O. Marks, M. C. Ruiz-Jaen, C. M. Salvador, and A. E. Zanne. 2010. Functional traits and the growht-mortality trade-off in tropical trees. Ecology 91:3664-3674.

112

TABLES Table 1. Mean and coefficient of variation (in parentheses) of: (a) aboveground biomass (AGB), (b) forest structure (Gini coefficient), (c) species diversity measures (species richness and species evenness), (d) functional diversity measures (functional richness and functional evenness), (e) environment (slope, scores of first and second axis of principal component analysis (PCA)) at three spatial scales in the 50 ha Dynamic Plot in Barro Colorado Island, Central Panama. Variables/Scale (ha) 0.04 0.25 1.00 a. AGB (Mg ha-1) 299.68 (0.78) 299.68 (0.31) 299.68 (0.18) b. Gini coefficient 0.88 (0.05) 0.90 (0.03) 0.90 (0.02) c. Species Richness 49.41(0.16) 110.51 (0.09) 164.22 (0.05) Evenness 0.93 (0.03) 0.94 (0.02) 0.94 (0.02) d. Function Richness 13.72 (0.11) 17.20 (0.05) 18.54 (0.05) Evenness 0.82 (0.04) 0.84 (0.02) 0.84 (0.02) e. Environment Slope 4.94 (0.75) 4.88 (0.55) 4.90 (0.46) PCA axis soil fertility 0.11 (22.83) 0.19 (13.74) 0.28 (9.54) PCA axis N availability 0.01 (105.68) 0.03 (37.22) 0.07 (17.24)

113

Table 2. Adjusted Coefficient of determination (R2) after partitioned the variation of aboveground biomass among species diversity measures (species richness and species evenness), functional diversity measures (functional richness and functional evenness), and environment (slope, scores of first and second axis of principal component analysis (PCA)) at three spatial scales in the 50 ha Dynamic Plot in Barro Colorado Island, Central Panama. Lower case letters represent single fractions of variation: (a) environment (E), (b) Species (S), (c) Function (F), (d) E*S, (e) S*F, (f) E*F, and (g) E*S*F. Sources of variation Fractions included 0.04 0.25 1.00

Environment + Species + Function [a+b+c+d+e+f+g] 0.028* 0.178* 0.389*

Species + Function [b+c+d+e+f+g] 0.027* 0.139* 0.361*

Environment + Species [a+c+d+e+f+g] 0.012* 0.091* 0.193*

Environment + Function [a+b+d+e+f+g] 0.028* 0.159* 0.263*

Decomposed variation (single fractions)

Species controlled for E and F [c] 0.0001 0.019 0.126

Function controlled for E and S [b] 0.016* 0.087* 0.196*

Environment controlled for S and F [a] 0.0007 0.038 0.028

We reported the adjusted R2 due to the bias associated with R2 values. Significance level: *, P<0.01

114

Table 3. Adjusted coefficient of determination (R2) after partitioned the variation of Gini coefficient among species diversity measures (species richness and species evenness), functional diversity measures (functional richness and functional evenness), and environment (slope, scores of first and second axis of principal component analysis (PCA)) at three spatial scales in the 50 ha Dynamic Plot in Barro Colorado Island, Central Panama. Lower case letters represent single fractions of variation: (a) environment (E), (b) Species (S), (c) Function (F), (d) E*S, (e) S*F, (f) E*F, and (g) E*S*F. Sources of variation Fractions included 0.04 0.25 1.00

Environment + Species + Function [a+b+c+d+e+f+g] 0.037* 0.200* 0.325*

Species + Function [b+c+d+e+f+g] 0.030* 0.109* 0.184*

Environment + Species [a+c+d+e+f+g] 0.014* 0.140* 0.268*

Environment + Function [a+b+d+e+f+g] 0.037* 0.175* 0.244*

Decomposed variation (single fractions)

Species controlled for E and F [c] 0.0001 0.025* 0.081

Function controlled for E and S [b] 0.023* 0.061* 0.057

Environment controlled for S and F [a] 0.007 0.091* 0.140

We reported the adjusted R2 due to the bias associated with R2 values. Significance level: *, P<0.01

115

Figure 1

Figure 1. Variation of aboveground biomass (Mg ha-1) within the 50- ha Dynamic Plot of Barro Colorado Island, Central Panama at three spatial scales: (a) 0.04 ha, (b) 0.25 ha, and (c) 1-ha. Values of aboveground biomass are log-transformed and increase from dark grey (2.10-2.30) to light grey (2.50-2.80). Contour lines indicate the topography of the 50-ha plot as a measure of elevation.

116

Figure 2

Figure 2. Positive relationship between the Gini coefficient and aboveground biomass at three spatial scales: (a) 0.04 ha, (b) 0.25 ha, and (c) 1 ha at the 50-ha Dynamic Plot of Barro Colorado Island, Central Panama.

117

Appendix 1: Correlation between the scores of the Principal Component Analysis (PCA) and eleven soil nutrients and soil pH at three scales in the 50-ha Dynamic Plot of Barro Colorado Island, Central Panama PC1 PC2 PC3 Scale 0.04 0.25 1.00 0.04 0.25 1.00 0.04 0.25 1.00 (ha) Al 0.45 0.32 0.35 -0.44 0.80 0.79 0.77 0.36 0.38 B -0.91 -0.88 -0.90 0.16 -0.17 -0.18 0.13 0.27 0.27 Ca -0.95 -0.96 -0.97 -0.06 -0.11 -0.10 -0.05 -0.03 -0.04 Cu -0.82 -0.83 -0.86 -0.27 0.20 0.17 0.14 0.21 0.28 Fe -0.74 -0.79 -0.81 -0.42 0.42 0.45 0.16 -0.08 -0.07 K -0.95 -0.94 -0.96 -0.01 0.02 0.03 0.02 -0.05 -0.04 Mg -0.89 -0.92 -0.94 -0.06 -0.16 -0.13 -0.03 -0.04 -0.10 Mn -0.68 -0.65 -0.63 -0.31 0.23 0.21 0.25 0.48 0.55 P 0.10 0.23 0.29 0.29 0.60 0.63 0.83 0.51 0.48 Zn -0.91 -0.89 -0.91 0.01 0.12 0.15 -0.06 -0.16 -0.18 N -0.34 -0.03 -0.03 0.75 -0.46 -0.52 0.29 0.73 0.73 N. min -0.72 -0.72 -0.75 -0.13 0.40 0.44 -0.13 -0.41 -0.35 pH -0.64 -0.39 -0.40 0.59 -0.64 -0.70 -0.01 0.26 0.20 Soil nutrients and soil pH at 0.25 ha and 1ha scale were averaged from the values at 0.04 ha before doing the Principal component analysis. .

118

Appendix 2: Distribution of two functional traits, maximum height and leaf mass per area within the 307 species present in the 50-ha Dynamic Plot of Barro Colorado Island, Central Panama. Vertical lines represent the cuts to classify species by low, medium, and high for each trait.

119

LINKING STATEMENT 3

In Chapter 3, I found that species diversity explained little variation in aboveground biomass and that functional evenness decreases aboveground biomass. These findings differ from those reported in Chapter 1. I believe that the difference can be explained by the fact that both chapters were developed in different forests using different biodiversity measures. Therefore, in Chapter 4, I tested the effect of biodiversity on aboveground biomass in three different forest types using the same biodiversity measures.

120

Chapter 4: Biodiversity-Ecosystem functioning along a precipitation gradient in Panamanian tropical forests

Ruiz-Jaen, Maria C. 1, S. Joseph Wright 2, Catherine Potvin 1,2

1 Department of Biology, McGill University, 1205 Dr Penfield, Montréal H3A-1B1, Québec, Canada 2 Smithsonian Tropical Research Institute, Apartado 0843–03092, Balboa, Panama.

Status: Ruiz-Jaen MC, Wright SJ, Potvin C. In preparation. Biodiversity-Ecosystem functioning along a precipitation gradient in Panamanian tropical forests

121

ABSTRACT

To generalize the relationship between species diversity and ecosystem function (BEF) in tropical ecosystems, we addressed BEF along a precipitation gradient (1950 – 3100 mm y- 1) in three high species rich tropical forests in Panama with different forest structure and species composition. In addition, we studied the relative contribution between tree diversity and forest structure on aboveground biomass. Our study emphasized that in species rich tropical ecosystems, forest structure and the identity of the dominant species affect aboveground biomass. In addition, we found that the strength of the relationship between basal area inequality and aboveground biomass increase with forest age. These results were site independent, suggesting a general pattern.

Keywords: aboveground biomass, biodiversity, dry forests, ecosystem functioning, moist forest, palms, Panama, species richness, wet forest

122

INTRODUCTION In the recent years we have been working in Panama to elucidate the relationship between tree diversity and a key function of tropical ecosystem, aboveground biomass. We have studied two different tropical forests, a moist (2000 - 3000 mm annual rainfall) and a wet (> 3000 mm annual rainfall) forest, both distant from less than 40 Km and located in the Atlantic cost of Panama. In the wet forest, palms richness was key to understand the diversity versus carbon storage relationship with an increased of tree carbon storage of canopy palms and eudicots as palm diversity increased (Ruiz-Jaen and Potvin 2010). In the moist forest however, palm diversity did not play a significant role in explaining the patterns in carbon storage nor did total species richness (Ruiz-Jaen and Potvin 2011, Ruiz-Jaen et al. unpublished). In this later forest, structure was the overwhelming source of variation in aboveground biomass (Ruiz-Jaen et al. unpublished). Building one from another, these studies used different metric of diversity as we sought to improve the explanatory power of our models. In the wet tropical forest, where we first worked, species were assigned to one of four functional groups (understory/canopy, palms/eudicots) and the relationship between these functional groups and carbon storage was examined without paying attention to forest structure. In the moist forest, we first used total species richness as our measure of diversity and thereafter rely on functional diversity estimated from trait-based measurements (Ruiz-Jaen et al. unpublished). We hypothesized that in this highly diverse forest, functional diversity might be more powerful than species richness in revealing diversity-ecosystem functions relationships. Together the results of our previous studies in Panama, suggest that the biodiversity-ecosystem functioning relationship is affected by the way species diversity is measured and that we have to consider forest structure to avoid confounding effects. Given that we tested biodiversity-ecosystem functioning using different approaches and in different forests, we now seek to unify our results by testing the hypothesis that above and beyond diversity effect, forest structure plays a central role in determining carbon stocks in a wide range of forests. This analysis will allow us to draw unified conclusion to our efforts in the last few years, to address the important issue of ecosystem functioning in the most diverse terrestrial systems, the tropical forests.

123

METHODS We used three forests dynamic plots of the Center for Forest Science (CTFS) along the precipitation gradient of the Panama Canal Watershed (Condit 1998, Condit et al. 2004). For the dry forest, we used the 4 ha plot in Cocoli located in the Pacific slope. For the moist forest, we used the 50 ha plot in Barro Colorado Island located in the middle of the Panama Canal. For the wet forest, we used the 4.96 ha plot in Fort Sherman located in the Caribbean coast of Panama. These sites have a very different species composition (Pyke et al. 2001, Condit et al. 2004) and forest structure (Condit et al. 2004, Fig. 1). The moist forest has the highest species richness and the dry forest has the lowest (Condit et al. 2004). The wet forest has the highest abundance of stems and the lowest in the dry forest, but the basal area of trees ≥ 30 cm follows the opposite pattern (Condit et al. 2004). Table 1 showed the difference among sites based on their climate, geology, and forest age. Aboveground biomass was calculated using allometric regression equation for each forest type (Chave et al. 2005). The wood density values for each species come from Wright et al. (2010) and Wright et al. unpublished data. For those species with no wood density values, we used the mean wood density for each forest. These calculations included all stems ≥ 1 cm diameter at breast height (DBH). We report total aboveground biomass for all three sites as was done in a previous study on BCI (Ruiz-Jaen and Potvin 2011, Ruiz-Jaen et al., unpublished) instead of dividing it into four functional groups (understory/canopy, palms/dicots) as was done in the study of the wet tropical forest (Ruiz-Jaen and Potvin 2010). For tree diversity, we used palm, and canopy tree richness. We divided species from palms and dicots because in a previous study, we reported that these two functional groups have different effect on tree carbon storage (Ruiz-Jaen and Potvin 2010). Palms and dicots responded differently to the environment and space (Vormisto et al. 2000, de Castilho et al. 2006) and allocated carbon differently (DeWalt and Chave 2004). Moreover, we only used canopy tree richness, since >50% of the species in these forests are understory species, which contributed very little to total aboveground biomass. We used two indicators of forest structure, the Gini coefficient, and the number of big trees. In a previous study of the moist forest, we reported that the Gini coefficient was strongly correlated (R2 ranged from 0.76 to 0.78) with aboveground biomass at all spatial

124

scales (Ruiz-Jaen et al. unpublished). We classified a tree as big when its DBH size was ≥ 70 cm (Lugo and Brown 1992, Clark and Clark 1996). Big trees could be use as an indicator of an old growth forest (Brown et al. 1997), account for a large proportion of the AGB (Clark and Clark 2000), and serve as carbon sinks (Lugo and Brown 1992).

Statistical Analysis - We used linear regressions to relate tree species richness and forest structure to aboveground biomass in the three study sites. We log transformed aboveground biomass to meet normalization requirements. We divided the three forests in subplots of 50 x 50 m (dry forest = 15 plots, moist forest = 200 plots, and wet forest = 20 plots). This size is the minimum plot size suggested by Chave et al (2004) to estimate AGB.

RESULTS Aboveground biomass at the plot level was twice higher in the moist forest than in the dry forest while the wet forests showed an intermediate value some 40% lower than that of the moist forest (Table 2). At the 50 x 50 m scale, aboveground biomass ranged 134.4-215.3 Mg ha-1, 135.5 - 658.7 Mg ha-1, and 125.2-255.1 Mg ha-1; respectively for the dry, moist, and wet forests. Aboveground biomass increases with forest age (Table 1 and 2). Palms were most diverse in the wet forest and decrease with the precipitation gradient (Table 2). At the 50 m x 50 m scale, palm richness ranged from 0 to 6 species in the dry and in moist forest, and ranged from 7 to 11 species in the wet forest. Overall, species richness positively increased with precipitation (Table 1). The moist forest was densest than the other one as shown by the high number of stems (Fig. 1). The Gini coefficient was relatively high in all sites with very little variation among plots of the same forest type. Big trees (≥ 70 cm DBH) were more common in the dry forests (2 to 9 trees), but the largest sizes were in the moist forest (dry: maximum DBH 165 cm; moist: maximum DBH 276.5 cm; wet: maximum DBH 120 cm) (Table 2). Different measures of species richness were regressed with aboveground biomass for each of the forests and the results show that neither palm, nor canopy trees, affected AGB of any of the three study sites (Fig. 2a-b). In contrast, the variables related to forest structure (the number of big trees, and the Gini coefficient) were positively related to AGB. In the

125

dry forest, AGB was mostly explained by the number of big trees (R2 = 0.49, p < 0.01), and by the Gini coefficient (R2 = 0.26, p < 0.05; Fig. 3a-b). In the moist and wet forest, the Gini coefficient (moist: R2 = 0.54, p< 0.001, wet: R2 = 0.59, p < 0.001) was the most important component of forest structure, followed by the number of big trees (moist: R2 = 0.51, p< 0.001; wet: R2 = 0.10, p < 0.5; Fig. 3a-b).

DISCUSSION Our study builds on important experimental work carried out in the last decade to disentangle the relationship between biodiversity and ecosystem functioning (Loreau et al. 2001, Hooper et al. 2005, Tilman et al. 2006). One of the experiments, BioDepth, led a group of investigators to propose that the net effect of biodiversity in fact consisted in two components, the complementarity and the selection effects (Tilman et al. 1999). Complementarity seeks to describe positive interaction among species leading to improve ecosystem functioning when species growing together performed better than the sum of individual species performance (Loreau and Hector 2001). The selection effect conversely can be attributed to the dominant effect of a species in an ecosystem; it has been referred to as the “big species” effect (Tilman et al. 1997). We brought these questions to the native tropical forest, in a series of studies carried out in Panama, moving away from controlled experimental set up. A central result of our work in Panama’s moist forest is that that in spatially heterogeneous environment, as suggested by Chesson et al (2002), the complementarity and the sampling effect cannot be disentangled and that aboveground biomass and forest structure are strongly correlated at small to medium scales (Ruiz-Jaen et al. unpublished). The present study tested the generality of this later finding by working in two additional forests growing in contrasting precipitation regime and forest age. While we did not find biodiversity-ecosystem function relationship among sites, we found a large difference in aboveground biomass (moist>wet>dry) and suggest that this difference can be explained in part by species composition. These sites have large differences in species composition and they varied in levels of deciduousness. For example, the forests only shared 46 species (25% of the species in the dry forest, 20% in the wet forest, and 15% of the moist forest) and precipitation seems to largely explain species identity in the different sites. In the dry forest more than 40% of the species are

126

deciduous and less than 15% of deciduous species in wet forest, (Condit et al. 2000). The level of deciduousness decrease with rainfall and forest age (Bohlman 2010) and deciduousness increases forest productivity and consequently tree carbon storage (Condit et al. 2000). Moreover, some of the “top ten” species for aboveground biomass in the moist forests (based on the maximum volume they can attained) like Ceiba pentandra (Malvaceae), Dipteryx oleifera (Fabaceae), and Hura crepitans (Euphorbiaceae) are absent in the dry forest, and in the wet forest. However, there are some top carbon storers present in the dry forest, Anacardium excelsum (Anacardiaceae) and Cavanilesia platanifolia (Malvaceae), but this species are not abundant (Condit et al. 2004) and their stems does not attained the large size found in the moist forest. The small stem size in the dry forest could be a result of forest age (Condit et al. 2000, see Table 1). In a review of experimental studies addressing the relationship biodiversity- ecosystem function, Cardinale et al. (2006) found that the identity of the dominant species and not species richness affected ecosystem function. The authors followed the rationale that in more diverse communities there is a higher probability to encounter the most productive species (i.e. sampling effect). Following these results, other studies have found that few species have a disproportionate effect on tree carbon storage (Balvanera et al. 2005). For example Kirby and Potvin (2007) reported that 12 tree species, all of them reaching large DBH, accounted for 46% of the biomass in another moist forest of Panama. In the moist forests, we sampled in an area an order of magnitude larger than the dry and wet forests. Increasing the area sampled, also increase the probability of encountering the most productive species, in other words, the sampling effect. Consequently, diversity might play a larger role on above ground biomass based on the identity of species as well as their abundance. In this context species identity, not just diversity becomes a central component of ecosystem functioning. Our results show that aboveground biomass is mostly is highly correlated with size inequality in forests. However, the strength of this relationship seems to increase with forest age (Table 1, Fig 3 a). This suggests that tropical forests continue to increase their aboveground biomass after centuries of forests succession. Zhou et al. (2006) found that as a stand matures, the soil carbon also increases in old-growth forests in China. Therefore,

127

at the level of the tropical forest, old growth stands stored high aboveground biomass and maintains high tree richness

REFERENCES Balvanera P, Kremen C, Martinez-Ramos M. 2005. Applying community structure analysis to ecosystem function: Examples from pollination and carbon storage. Ecological Applications 15(1): 360-375. Bohlman SA. 2010. Landscape patterns and environmental controls of deciduousness in forests of central Panama. Global Ecology and Biogeography 19(3): 376-385. Cardinale BJ, Srivastava DS, Duffy JE, Wright JP, Downing AL, Sankaran M, Jouseau C. 2006. Effects of biodiversity on the functioning of trophic groups and ecosystems. Nature 443(7114): 989-992. Chave J, Andalo C, Brown S, Cairns MA, Chambers JQ, Eamus D, Folster H, Fromard F, Higuchi N, Kira T, Lescure JP, Nelson BW, Ogawa H, Puig H, Riera B, Yamakura T. 2005. Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145(1): 87-99. Clark DB, Clark DA. 1996. Abundance, growth and mortality of very large trees in neotropical lowland rain forest. Forest Ecology and Management 80(1-3): 235-244. Condit R. 1998. Ecological implications of changes in drought patterns: Shifts in forest composition in Panama. Climatic Change 39(2-3): 413-427. Condit R, Aguilar S, Hernandez A, Perez R, Lao S, Angehr G, Hubbell SP, Foster RB. 2004. Tropical forest dynamics across a rainfall gradient and the impact of an El Nino dry season. Journal of Tropical Ecology 20: 51-72. Condit R, Watts K, Bohlman SA, Perez R, Foster RB, Hubbell SP. 2000. Quantifying the deciduousness of tropical forest canopies under varying climates. Journal of Vegetation Science 11(5): 649-658. de Castilho CV, Magnusson WE, de Araujo RNO, Luizao RCC, Lima AP, Higuchi N. 2006. Variation in aboveground tree live biomass in a central Amazonian Forest: Effects of soil and topography. Forest Ecology and Management 234(1-3): 85-96. DeWalt SJ, Chave J. 2004. Structure and biomass of four lowland Neotropical forests. Biotropica 36(1): 7-19.

128

Elias M, Potvin C. 2003. Assessing inter- and intra-specific variation in a trunk carbon concentration for 32 neotropical tree species Canadian Journal of Forest Research 33: 1039-1045. Guariguata MR, Chazdon RL, Denslow JS, Dupuy JM, Anderson L. 1997. Structure and floristics of secondary and old-growth forest stands in lowland Costa Rica. Plant Ecology 132(1): 107-120. Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge DM, Loreau M, Naeem S, Schmid B, Setala H, Symstad AJ, Vandermeer J, Wardle DA. 2005. Effects of biodiversity on ecosystem functioning: A consensus of current knowledge. Ecological Monographs 75(1): 3-35. Kirby KR, Potvin C. 2007. Variation in carbon storage among trees species: implications for the management of small-scale carbon sink project. Forest Ecology and Management 246: 208-221. Loreau M, Hector A. 2001. Partitioning selection and complementarity in biodiversity experiments. Nature 412(6842): 72. Loreau M, Naeem S, Inchausti P, Bengtsson J, Grime JP, Hector A, Hooper DU, Huston MA, Raffaelli D, Schmid B, Tilman D, Wardle DA. 2001. Biodiversity and ecosystem functioning: Current knowledge and future challenges. Science 294(5543): 804-808. Losos EC, Ashton PS, Brokaw N, Bunyavejchewin S, Condit R, Chuyong G, Co L, Dattaraja HS, Davies S, Esufali S, Ewango C, et al. 2004. The structure of tropical forests. In: Losos EC, Leigh EG eds. Tropical forest diversity and dynamics. Chicago, USA: The University of Chicago Press, 69-78. Dolores R. Piperno. 1990. Fitolitos, Arqueología y Cambios Prehistóricos de la Vegetación en un Lote de Cincuenta Hectáreas de la Isla de Barro Colorado. In: A. Stanley Rand;Donald M. Windsor;Jr. Leigh, Egbert Giles (Ed.), Ecología de un bosque tropical: ciclos estacionales y cambios a largo plazo. Balboa, Republic of Panama: Smithsonian Tropical Research Institute, 153-156. Pyke CR, Condit R, Aguilar S, Lao S. 2001. Floristic composition across a climatic gradient in a neotropical lowland forest. Journal of Vegetation Science 12(4): 553-566. Ruiz-Jaen MC, Potvin C. 2010. Tree diversity explains variation in ecosystem function in a Neotropical forest in Panama. Biotropica 42: 638-646.

129

Ruiz-Jaen MC, Potvin C. 2011. Can we predict carbon stocks in tropical ecosystems from tree diversity? Comparing species and functional diversity in a plantation and a natural forest. New Phytologist 189: 978-987. Svenning JC. 1999. Recruitment of tall arborescent palms in the Yasuni National Park, Amazonian Ecuador: are large treefall gaps important? Journal of Tropical Ecology 15: 355-366. Tilman D. 1999. The ecological consequences of changes in biodiversity: A search for general principles. Ecology 80(5): 1455-1474. Tilman D, Lehman CL, Thomson KT. 1997. Plant diversity and ecosystem productivity: Theoretical considerations. Proceedings of the National Academy of Sciences of the United States of America 94(5): 1857-1861. Tilman D, Reich PB, Knops JMH. 2006. Biodiversity and ecosystem stability in a decade-long grassland experiment. Nature 441(7093): 629-632. Vandermeer J, Stout J, Miller G. 1974. Growth rates of Welfia georgii, Socratea durissima, and Iriartea gigantea under varios conditions in a natural rainforest in Costa Rica. Principles 18: 148-154. Vormisto J, Phillips OL, Ruokolainen K, Tuomisto H, Vásquez R. 2000. A comparison of fine- scale distribution patterns of four plant groups in an Amazonian rainforest. Ecography 23: 349-359. Wright SJ, Kitajima K, Kraft NJB, Reich PB, Wright IJ, Bunker DE, Condit R, Dalling JW, Davies SJ, Diaz S, et al. 2010. Functional traits and the growht-mortality trade-off in tropical trees. Ecology 91: 3664-3674. Zhou, G., S. Liu, Z. Li, D. Zhang, X. Tang, C. Zhou, J. Yan, and J. Mo. 2006. Old-growth forests can accumulate carbon in soils. Science 314: 1417.

130

TABLES Table 1. Description of the three study sites along the Panama Canal Watershed. Variables Dry Forest Moist Forest Wet Forest Mean annual precipitation (mm yr-1) 1950 2600 3100 Duration of the dry season (days) 1 129 118 106 Geology 2 Miocene and Miocene Basalt Chagres Pre-Tertiary and the Caimito sandstone Basalt Formation of parent tuffaceous material sandstones Forest Age (years) ~100 > 1,5003 ~200 1 Condit et al. (2000) 2 Pyke et al. (2001) 3 Piperno (1990)

131

Table 2. Mean and coefficient of variation (in parentheses) of: (a) aboveground biomass (AGB), (b) species richness (palms and canopy trees), (c) forest structure (the Gini coefficient, maximum diameter at breast height (DBH), number of trees larger than 70 cm DBH), in three forest along the Panama Canal Watershed. Variables Dry Forest Moist Forest Wet Forest a. AGB (Mg ha-1) 165.64 (0.13) 299.67 (0.32) 203.19 (0.16) b. Species Richness Palm 3.2 (0.43) 2.44 (0.43) 9.64 (0.11) Canopy Trees 30.53 (0.14) 29.38 (0.12) 54.45 (0.12) Total 69.2 (0.12) 110.93 (0.09) 105.14 (0.11) c. Forest structure Gini coefficient 0.93 (0.01) 0.89 (0.03) 0.87 (0.02) Maximum DBH (cm) 133.75 (0.17) 115.28 (0.39) 90.37 (0.18) Number of big trees 5.6 (0.35) 2.60 (0.67) 2.23 (0.62)

132

FIGURES Figure 1 a)

Figure 1. Distribution tree diameter classes at breast height (DBH) in three forests along the Panama Canal Watershed: (a) dry, (b) moist, and (c) wet forest. Data in the moist forest is a subsample of 23000 individuals to be comparable with the other forests types.

133

Figure 2 (a)

(b)

Figure 2. Left panels show the effect of (a) palm richness and (b) canopy trees richness on aboveground biomass in three forests along the Panama Canal Watershed. Symbols are: () Dry forest,  moist forest, and , wet forest. Dotted, dashed and solid lines correspond to linear regression models for the dry, moist, and wet forests, respectively.

134

Figure 3 a)

Figure 3. The effect of (a) the Gini coefficient and (b) the number of big trees ( ≥ 70 cm DBH) on above ground biomass for three forests in the Panama Canal Watershed. Each  represents data from dry forest,  from moist forest, and  from wet forest. Dotted lines correspond to linear regression models for dry forest, the dashed lines for the ones in the moist forest, and the solid line for the ones in the wet forest.

135

Final Summary and Conclusions

Lessons from the tropical forests in Panama My dissertation was inspired by studies in temperate grasslands that tried to link biodiversity to ecosystem function to understand the effect of species loss (Hector et al. 1999, Loreau et al. 2001, Tilman et al. 2006). In general, these studies have found that losing species can decrease an ecosystem function and explained these results with two main mechanisms: niche complementarity (species richness) and selection effect (species dominance). The importance of these two mechanisms for ecosystem function has been a debate. For example, the properties of the dominant species have been the primarily mechanism to explain ecosystem function (Cardinale et al. 2006), however, as the forest stands get older, complementarity of species richness gain more relevance (Cardinale et al. 2007). Since I have always being interested in tropical forests, I focused my research in the relationship between biodiversity and ecosystem function (BEF) in these ecosystems. To explore BEF, I look at the relationship between tree species richness and functional richness (niche complementarity) and species and functional evenness (selection effect) and aboveground biomass in three natural ecosystems and in a mixed-species experimental plantation. Since, the studies were developed in naturally varying environment and with long-lived trees, we considered for the effect of environmental variation and forest structure.

After five years of studying BEF in tropical ecosystems, I found that: (1) Environmental variables related to topography, soil nutrients and physical characteristics do not affect aboveground biomass significantly in my field sites. (2) The way biodiversity is measured affected BEF relationship considerably. (3) We found that functional richness varied very little among plots and most of the functional groups were present, suggesting that these sites might be functionally saturated, thus adding species or functional groups do not increase aboveground biomass. Moreover, the species diversity of tropical forests is mostly allocated to understory species that contributed very little to aboveground biomass.

136

(4) Both the identity of species and species composition can affect aboveground biomass indirectly, suggesting that the sampling effect may be at play in tropical forests. (5) Functional evenness decreases aboveground biomass from small to medium scales. (6) Forest structure is highly correlated to aboveground biomass at small to medium scales and at different forests types.

My results also have implications for forest management and for assessing carbon stocks in tropical ecosystems. Specifically, I found that to design carbon-rich plantations, forest managers should be cautious when applying functional trait measured in natural populations.

Future research Existing anthropogenic (e.g., land-use change, Nascimento and Laurance 2004, Wright 2005, Chave et al. 2008) and climatic (e.g. increase in droughts, Condit 1998, Phillips et al. 2002) pressures are likely to modify tropical ecosystems (Phillips et al. 2004, Engelbrecht et al. 2007, Gardner et al. 2009, Slik et al. 2010). Our results have found that the identity of the dominant species and the species composition affect tree carbon storage, therefore, I suggest two major lines of research for understanding biodiversity-ecosystem function in tropical ecosystems: (1) cross-site comparison, and (2) incorporating forest dynamics. Cross-site comparison Our last chapter attempted to generalize BEF in tropical ecosystems by studying forests of different precipitation regimes. I found a consistent BEF pattern among sites, but with different explanatory power. Studies with cross comparisons along the tropics have shown that at a continental scales forests have different carbon allocation (Malhi 2010), species composition (Ashton et al. 2004), and forest structure (Losos et al. 2004). For example, Feldpaush et al. (2011) found that the relationship between height and diameters of trees was highly variable across the tropics. Therefore, these changes in tree allometry could affect the strength of the relationship between tree diversity and aboveground biomass.

Incorporating forest dynamics

137

Stability of carbon sequestration is going to play a central role in recent initiatives to pay for carbon credits by reducing greenhouse gas emission from deforestation and forest degradation (REDD, Baker et al. 2010). Stability of an ecosystem function has been poorly studied in tropical ecosystems. In an early paper, Tanner and Bellinghan (2006) found that areas where hurricanes are frequent, forest productivity was higher in less diverse sites. Forest stability would also be influence by climate change. For example, Feeley et al. (2007) reported a decelerated growth in two different tropical forests due to increase in dry days. Slow tree growth and high tree mortality are related to ocurrance of strong drought episodes (Malhi et al. 2009)) that are associated with El Niño events (Chave et al. 2008). Increasing drougth incidence could also increase liana occurrence, since liana density have been shown to increase in sites with dry season lenght (DeWalt et al. 2010). Given the proliferation of lianas in the last decade (Phillips et al. 2002, Wright et al. 2004, Ingwell et al. 2010), several studies have raised the importance of lianas in tropical ecosystems, especially in their effect on carbon sequestration. Therefore, I can test the relative contribution of functional groups (trees versus lianas) on aboveground biomass stability. These two research lines could be address using the network of the Center for Tropical Forest Science (CTFS)-Smithsonian Institution Global Earth Observatories (SIGEO). CTFS-SIGEO has > 20 forest dynamic plots along tropical ecosystems and some of these plots have been censused at least three times (Wills et al. 2006, Chave et al. 2008). In all this plots all stems ≥ 1 cm in diameter at breast height have been mapped, measured its DBH, and identified to species. In addition, Stephan Schnitzer and colleagues monitored for the first time lianas density, distribution, and diversity in one CTFS-SIGEO plot with the aim to extend this protocols to the rest of the network. Therefore, I would like to take advantage of this network to conduct BEF-forest structure and carbon dynamics analyses across forest sites.

REFERENCES Ashton P, Brokaw N, Bunyavejchewin S, Condit R, Chuyong G, Co L, Dattaraja HS, Davies S, Esufali S, Ewango C, Foster R, et al. 2004. Floristic and vegetation of the forest dynamics plots. In: Losos E, Leigh EG eds. Tropical forest diversity and

138

dynamics. Chicago, USA: The University of Chicago Press, 90-102. Baker TR, Jones JPG, Thompson ORR, Cuesta RMR, del Castillo D, Aguilar IC, Torres J, Healey JR. 2010. How can ecologists help realise the potential of payments for carbon in tropical forest countries? Journal of Applied Ecology 47(6): 1159-1165. Chave J, Condit R, Muller-Landau HC, Thomas SC, Ashton PS, Bunyavejchewin S, Co LL, Dattaraja HS, Davies SJ, Esufali S, Ewango CEN, Feeley KJ, Foster RB, Gunatilleke N, Gunatilleke S, Hall P, Hart TB, Hernandez C, Hubbell SP, Itoh A, Kiratiprayoon S, LaFrankie JV, de Lao SL, Makana JR, Noor MNS, Kassim AR, Samper C, Sukumar R, Suresh HS, Tan S, Thompson J, Tongco MDC, Valencia R, Vallejo M, Villa G, Yamakura T, Zimmerman JK, Losos EC. 2008. Assessing evidence for a pervasive alteration in tropical tree communities. Plos Biology 6(3): 455-462. Condit R. 1998. Ecological implications of changes in drought patterns: Shifts in forest composition in Panama. Climatic Change 39(2-3): 413-427. DeWalt SJ, Schnitzer SA, Chave J, Bongers F, Burnham RJ, Cai ZQ, Chuyong G, Clark DB, Ewango CEN, Gerwing JJ, Gortaire E, Hart T, Ibarra-Manriquez G, Ickes K, Kenfack D, Macia MJ, Makana JR, Martinez-Ramos M, Mascaro J, Moses S, Muller- Landau HC, Parren MPE, Parthasarathy N, Perez-Salicrup DR, Putz FE, Romero-Saltos H, Thomas D. 2010. Annual Rainfall and Seasonality Predict Pan-tropical Patterns of Liana Density and Basal Area. Biotropica 42(3): 309-317. Engelbrecht BMJ, Comita LS, Condit R, Kursar TA, Tyree MT, Turner BL, Hubbell SP. 2007. Drought sensitivity shapes species distribution patterns in tropical forests. Nature 447(7140): 80-U82. Feeley KJ, Wright SJ, Nur Supardi MN, Kassim AR, Davies SJ. 2007. Decelerating growth in tropical forest trees. Ecology Letters 10: 461-469. Feldpausch TR, Banin L, Phillips OL, Baker TR, Lewus SL, Quesada CA, Affum-Baffoe K, Arets EJMMM, Berry NJ, Bird M, Brondizio ES, de Camargo P, Chave J, Djagbletey G, Domingues TF, Drescher M, Fearnside PM, Franca MB, Fyllas NM, Lopez-Gonzalez G, Hladik A, Higuchi N, Hunter MO, Iida Y, Salim KA, Kassim AR, Keller M, Kemp J, King DA, Lovett JC, Marimon BS, Marimon-Junior BH, Lenza E, Marshall AR, DJ Metcalfe, Mitchard ETA, Moran EF, Nelson BW, Nilus R, Nogueira EM, Palace M, Patiño S, Peh KSH, Raventos MT, Reitsma JM, Saiz G, Schrodt F,

139

Sonke B, Taedoumg HE, Tan S, White L, Woll H, Lloyd J. 2011. Height-diameter allometry of tropical forest trees. Biogeosciences 8: 1081-1106. Gardner TA, Barlow J, Chazdon R, Ewers RM, Harvey CA, Peres CA, Sodhi NS. 2009. Prospects for tropical forest biodiversity in a human-modified world. Ecology Letters 12(6): 561-582. Hector A, Schmid B, Beierkuhnlein C, Caldeira MC, Diemer M, Dimitrakopoulos PG, Finn JA, Freitas H, Giller PS, Good J, Harris R, Hogberg P, Huss-Danell K, Joshi J, Jumpponen A, Korner C, Leadley PW, Loreau M, Minns A, Mulder CPH, O'Donovan G, Otway S, Pereira JS, Prinz A, Read DJ, Scherer-Lorenzen M, Schulze ED, Siamantziouras ASD, Spehn EM, Terry AC, Troumbis AY, Woodward FI, Yachi S, Lawton JH. 1999. Plant diversity and productivity experiments in European grasslands. Science 286: 1123-1127. Ingwell LL, Wright SJ, Becklund KK, Hubbell SP, Schnitzer SA. 2010. The impact of lianas on 10 years of tree growth and mortality on Barro Colorado Island, Panama. Journal of Ecology 98(4): 879-887. Loreau M, Naeem S, Inchausti P, Bengtsson J, Grime JP, Hector A, Hooper DU, Huston MA, Raffaelli D, Schmid B, Tilman D, Wardle DA. 2001. Biodiversity and ecosystem functioning: Current knowledge and future challenges. Science 294(5543): 804-808. Losos EC, Ashton PS, Brokaw N, Bunyavejchewin S, Condit R, Chuyong G, Co L, Dattaraja HS, Davies S, Esufali S, et al. 2004. The structure of tropical forests. In: Losos EC, Leigh EG eds. Tropical forest diversity and dynamics. Chicago, USA: The University of Chicago Press, 69-78. Malhi Y. 2010. The carbon balance of tropical forest regions, 1990-2005. Current Opinion in Environmental Sustainability 2(4): 237-244. Malhi Y, Aragao L, Galbraith D, Huntingford C, Fisher R, Zelazowski P, Sitch S, McSweeney C, Meir P. 2009. Exploring the likelihood and mechanism of a climate- change-induced dieback of the Amazon rainforest. Proceedings of the National Academy of Sciences of the United States of America 106(49): 20610-20615. Nascimento HEM, Laurance WF. 2004. Biomass dynamics in Amazonian forest fragments. Ecological Applications 14: S127-S138. Phillips OL, Baker TR, Arroyo L, Higuchi N, Killeen TJ, Laurance WF, Lewis SL, Lloyd

140

J, Malhi Y, Monteagudo A, Neill DA, et al.. 2004. Pattern and process in Amazon tree turnover, 1976-2001. Philosophical Transactions of The Royal Society Of London Series B-Biological Sciences 359(1443): 381-407. Phillips OL, Martinez RV, Arroyo L, Baker TR, Killeen T, Lewis SL, Malhi Y, Mendoza AM, Neill D, Vargas PN, Alexiades M, Ceron C, Di Fiore A, Erwin T, Jardim A, Palacios W, Saldias M, Vinceti B. 2002. Increasing dominance of large lianas in Amazonian forests. Nature 418(6899): 770-774. Slik JWF, Aiba SI, Brearley FQ, Cannon CH, Forshed O, Kitayama K, Nagamasu H, Nilus R, Payne J, Paoli G, Poulsen AD, Raes N, Sheil D, Sidiyasa K, Suzuki E, van Valkenburg J. 2010. Environmental correlates of tree biomass, basal area, wood specific gravity and stem density gradients in Borneo's tropical forests. Global Ecology and Biogeography 19(1): 50-60. Tanner EVJ, Bellingham PJ. 2006. Less diverse forest is more resistant to hurricane disturbance: evidence from montane rain forests in Jamaica. Journal of Ecology 94: 1003-1010. Tilman D, Reich PB, Knops JMH. 2006. Biodiversity and ecosystem stability in a decade- long grassland experiment. Nature 441(7093): 629-632. Wills C, Harms KE, Condit R, King D, Thompson J, He FL, Muller-Landau HC, Ashton P, Losos E, Comita L, Hubbell SP, et al. 2006. Nonrandom processes maintain diversity in tropical forests. Science 311(5760): 527-531. Wright SJ. 2005. Tropical forests in a changing environment. Trends in Ecology & Evolution 20(10): 554-559. Wright SJ, Calderon O, Hernandez A, Paton S. 2004. Are lianas increasing in importance in tropical forests? A 17-year record from Panama. Ecology 85(2): 484-489.

141