Forest Responses to Climate Change Along an Andes-To-Amazon Elevational Gradient

Home , Abarema jupunba, Alsophila erinacea, Andes to Amazon, Buddleja montana, Calyptranthes densiflora, Cinchona pubescens

FOREST RESPONSES TO CLIMATE CHANGE ALONG AN ANDES-TO-AMAZON ELEVATIONAL GRADIENT

WILLIAM FARFAN RIOS

A Dissertation Submitted to the Graduate Faculty of the

WAKE FOREST UNIVERSITY GRADUATE SCHOOL OF ARTS AND SCIENCES

in Partial Fulfillment of the Requirements

for the Degree of

DOCTOR OF PHILOSOPHY

Biology

August 2019

Winston-Salem, North Carolina

Approved By:

Miles R. Silman, Ph.D., Advisor

Nigel C. Pitman, Ph.D., Chair

T. Michael Anderson, Ph.D.

Kathleen A. Kron, Ph.D.

William K. Smith, Ph.D.

Clifford W. Zeyl, Ph.D. DEDICATION

To my beloved parents Julio Farfan and Visitacion I. Rios, and my brothers Maikol and

Henry Farfan

ACKNOWLEDGEMENTS

First and foremost, I would like to sincerely thank my advisor, Miles R. Silman, for his continuous support of my research, for his encouragement, knowledge, patience and for being a great mentor and friend over the past several years. I deeply appreciate his time, enthusiasm, and encouragement in pushing me beyond the limits my own mind to develop my critical thinking. If I had to do it all over again, undoubtedly, I would do it with Miles. Likewise, I thank my committee members, Dr. Kathleen A. Kron, Dr.

William K. Smith, Dr. Michael T. Anderson, Dr. Clifford W. Zeyl, and Dr. Nigel C.

Pitman for their time, helpful ideas and comments that enriched my research.

I have been immensely blessed by a wonderful and incredibly family, their love with unwavering support were essential to finish graduate school and perform my research. I owe so much gratitude to my beloved parents Julio Farfan, Visitacion I. Rios and my brothers Maikol and Henry Farfan.

Collaboration is important to do science and I was extremely fortunate to work closely with great scientists that enriched my research on so many levels. My sincere gratitude to professors Kenneth Feeley, Yadvinder Malhi, Sassan Saatchi, Oliver Phillips,

Norma Salinas, Patrick Meir, Greg Asner, Timothy Baker, Paul Baker, Sherilyn Fritz,

James Pease, Brian Enquist and Alexander Shenkin.

Thanks to my old and new Silman lab friends for their support, comments on the different drafts of the dissertation chapters and helpful ideas to improve my research,

iii thanks to Josh Rapp, Danny Lough, Karina Garcia, Rachel Hillyer, Noah Yavit, Cassie

Freund, Ellen Quinlan, Max Messinger, Stephanie Bilodeau, John Gorelick, Jared Beaver and Jorge Caballero. Likewise, thanks for the support to my Deacon friends Jenny

Howard, Emily Tompkins, Nicholas Huffeldt, Yoyi Fernandez, Scott Cory, Deusdedith

Rugemalila, Robbie Baldwin, Felipe Estela, and Mary J. Carmichael. Thanks to the

Causita’s group members Jorge Ortiz, Ciro Astete, Armando Alfaro, Jorge Curo, Juan

Gibaja, Victor Chama, Juan Costa, Vicky Huaman, Javier Silva, Johnny Farfan, Jean

Latorre and Saul Zuñiga for their support and encouragement to finish graduate school and complete my research. Thanks to Alex Nina for his thoughtful assistance translating the abstract in Quechua. Finally, thanks to my friends Leydi, Gaile, Gerald, Janell,

Richard and Craig for their support during my graduate school life.

I express my deep gratitude to Robin Foster, Abel Monteagudo, Paul Mass,

Ronald Liesner, Jun Wen, Kenneth Wurdack, Nancy Hensold, Henk van der Werff,

Fabian A. Michelangeli, Lucia Kawasaki, Percy Nuñez, Karina Garcia, Felipe Sinca and

Charlote M. Taylor for the help in the botanical identifications. Likewise, my recognition and thanks to the directors and curators of the different herbaria, including the Herbario

Vargas (CUZ), Herbario del Museo de Historia Natural- Universidad Nacional Mayor de

San Marcos (USM), Herbario Nacional de Bolivia (LPB), al Field Museum (F) and the

Missouri Botanical Garden Herbarium (MO), Herbario Selva Central - Oxapampa

(HOXA), Herbario Augusto Weberbauer (MOL) for access and permission to work with the botanical collections. Thanks to GBIF, JSTOR, TROPICOS, TNRS and to all the institutions that made available the botanical data essential for this research.

This fieldwork effort was not one-man labor, it was only possible thanks to an extraordinary group of people with high levels of enthusiasm and hard work. First, I thank Alex Nina, Jhonatan Sallo, Karina Garcia, Luis Imunda, Natividad Raurau, Judith

Huaman and Alberto Gibaja for their crucial help in getting everything done coordinating the different aspects of the fieldwork and plants identification process. Lastly but no less important, I thank the Peruvian students that shared with me the extreme fieldwork conditions, leaving sweat, blood, and tears in our field campaigns. Without them, I would not have been able to complete this research. My eternal gratitude to Adam J. Ccahuana,

Albino Quispe, Alex Nina, Alexander Quispe, Alex A. Caceres, Andrea Palomino,

Angela Rozas, Bryan G. Valencia, Carlos A. Salas, Catherine Bravo, Claudio Lipa,

Chaska Chavez, Cristian E. Solis, Cristian Alvarez Galvez, Cristhian Alvarez C., Darcy

F. Galiano, David Lopez, Dayana Suni, Dino Tapia, Edith R. Clemente, Erickson G.

Urquiaga, Felipe Sinca, Félix F. Farfan, Flor M. Perez, Flor M. Zamora, Fredy Guizado,

Ghylmar J. Tinoco, Giuliana M. Palomino, Gladys Castillo, Guido Vilcahuaman, Guisela

J. Zans, Irving C. Costas, Israel Cuba, Janet Mamani, Javier E. Silva, , Jesus Castaneda,

Jesus M. Banon, , Jhonatan Sallo, Jhoel Delgado, Jhon G. Quispe, Jhon S. Cansaya,

Jimmy R. Chambi, Jimy S. Mesicano, Joel Mendoza, Jonyer H. Zapata, Jose A.

Quintano, Jose L. Mancilla, Jose Sanchez, Juan A. Gibaja, Juan J. Calvo, Judit Huaman,

Julina Pelaes, , Karina Garcia, Karina Luna, Katherine Quispe, Katia Quispe, Kilmenia

Luna, Leonit Tueros, Leydi V. Auccacusi, Leyni Caballero, Lucely Vilca, Luis Imunda,

Luis Mancilla, Luz M. Cabrera, Manuel J. Marca, Marco A. Maldonado, Maria A.

Cuentas, Maria I. Manta, Mario V. Sanchez, Marlene Mamani, Miguel Salas, Milenka X.

Montoya, Mireya N. Raurau, Nohemi Lizarraga, Omar Tacusi, Omayra V. Colque, Paul

E. Santos, Paul Iturbe, Paul Ricardo, Percy O. Chambi, Reymerth Peña, Richard Tito,

Rosa M. Castro, Rosalbina Butron, Rudi S. Cruz, Ruly Pillco, Shirley J. Serrano, Tatiana

E. Boza, Vicky Huaman, Walter Huaraca, Washinton Luis, Wilson Huaman, Yan C.

Nina, Yaneth Quispe, Yeferson Davalos, Yessica Halanocca, Yezen Sanchez, Ylenia

Moron, Yovana Yllanes, and Yorka Gutierrez.

This study is a product of the Andes Biodiversity and Ecosystem Research Group

(ABERG; http://www.andesconservation.org/) with contributions for lowland plot data from John Terborgh and affiliated networks RAINFOR (http://www.rainfor.org/), GEM

(http://gem.tropicalforests.ox.ac.uk/), and the ForestPlots.net data management utility for permanent plots. SERFOR, SERNANP, and personnel of Manu National Park - Peru provided assistance with logistics and permission to work in the buffer zone and protected areas in Peru. Pantiacolla Tours and the Amazon Conservation Association provided logistical support. Funding came from Wake Forest University through the

Vecellio Grant, Elton C. Cocke Travel Award, Alumni Student Travel Award, and

Richter Scholarship Award. Funding also came from the Gordon and Betty Moore

Foundation’s Andes to Amazon initiative and the US National Science Foundation (NSF)

DEB 0743666, NSF Frontiers in Earth Systems Dynamics (FESD) 1338694, and NSF

Long-Term Research in Environmental Biology (LTREB) 1754647. The research was also supported by the National Aeronautics and Space Administration (NASA) Terrestrial

Ecology Program grant # NNH08ZDA001N-TE/ 08-TE08-0037. Support for RAINFOR and ForestPlots.net plot monitoring in Peru has come from a European Research Council

(ERC) Advanced Grant (T‐FORCES, “Tropical Forests in the Changing Earth System”,

291585), Natural Environment Research Council grants (including NE/F005806/),

NE/D005590/1, and NE/N012542/1), and the Gordon and Betty Moore Foundation.

vii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS....………..…………..………………………………...….iii

LIST OF TABLES...…………………………………………………………………...xi

LIST OF FIGURES.….……………………………………………………...……….xiii

ABSTRACT….……………………….……………………………………………...xvii

CHAPTER I

Introduction……..…..…………………………………………………………..1

CHAPTER II

Landscape-scale wood density variation across an Andes-to-Amazon

elevational gradient

ABSTRACT………………….………………………………….………..26

INTRODUCTION……………………………….…………………...... 28

METHODS…….…...………………………………………………...... 32

RESULTS….………...…………………………………………………...38

DISCUSSION………….…………………………………....…………....42

LITERATURE CITED.…………….……………………….……………55

SUPPORTING INFORMATION.…………………..………….…..…….84

CHAPTER III

Movement or mortality? Pervasive but slow upslope tree migration in the

Amazon to the Andes revealed through 38 years of forest monitoring

ABSTRACT…….……………………………………………….………129

viii

INTRODUCTION…………………………………………………...... 131

METHODS…………………………………………………………...... 135

RESULTS……………………………………………………………….143

DISCUSSION….……………………………………………………...... 148

LITERATURE CITED.……….……………………………..………….158

SUPPORTING INFORMATION …..…………………….…………….180

CHAPTER IV

Long-term stand and carbon dynamics along the Amazon-to-Andes elevation

gradient

ABSTRACT…………….……………………………………….………211

INTRODUCTION…………………………………………………...... 213

METHODS…………………………………………………………...... 217

RESULTS……………………………………………………………….224

DISCUSSION……………….…………………………………………..230

LITERATURE CITED……….…………………………………………238

SUPPORTING INFORMATION ………………………...... 263

CHAPTER V

An annotated checklist of trees and their relatives in tropical montane forests of

southeast Peru: the importance of continued botanical collecting

ABSTRACT………………..……………………………….…..….……277

INTRODUCTION…………………………………………………...... 279

METHODS………………………………………………………...... 281

RESULTS…………………………………………………………….....283

DISCUSSION…………………….…………………………………..…283

LITERATURE CITED………………………………...………………..288

SUPPORTING INFORMATION ……………………………………....353

CURRICULUM VITAE…………………………………………………………..…357

LIST OF TABLES

Table II - 1. Sites description and mean wood density values across the Andes-to-

Amazon elevational gradient…………………………………………………..69

Appendix II – Table S1. Species list and basic wood specific gravity values along an

Andean-to-Amazon elevational gradient.………………………..…………….86

Appendix II – Table S2. Statistical moments of mean wood density distribution on

species and stem levels across the elevational gradient.…….……...………..119

Appendix III - Table S1. Description of the 40 permanent forests plots along the

Andes-to-Amazon elevational gradient…………...…………………………181

Appendix III - Table S2. The estimated plot-level thermophilization rates………....184

Appendix III - Table S3. The estimated genus-level thermal migration rates for genera

weighted by number of individuals………………..………………………....188

Appendix III - Table S4. The estimated genus-level thermal migration rates for genera

weighted by basal area…………………………………………….………....192

Appendix III - Table S5. The estimated species-level thermal migration rates weighted

by number of individuals…………………………..………………………....196

Appendix III - Table S6. The estimated species-level thermal migration rates weighted

by basal area………………………………….……………………………....200

Appendix IV - Table S1. Description of the 40 permanent forests plots along the

Andes-to-Amazon elevational gradient……………………………………....264

Appendix IV - Table S2. Parameters values to estimate height in meters with diameters

expressed in centimeters…………..………………………………………….268

Table V - 1. List of trees and arborescent species…….....………………………...…292

xii

LIST OF FIGURES

Figure II - 1. Location of the permanent forest plots along the elevational

gradient…………………………………………………………………...…....76

Figure II - 2. Distribution of species slopes of the linear regression and intraspecific

variation of wood density………………………………………………..…….77

Figure II - 3. Plot-level mean wood density variation across 41 permanent forest plots

along the Andes-to-Amazon elevational gradient………………………..……78

Figure II - 4. Mean species wood density weighted by basal area…………………….79

Figure II - 5. Wood density distribution for species and stems……………………….80

Figure II - 6. Plot-level wood density variation across forests types …………...... 81

Figure II - 7. Mean wood density variation for stems across DBH classes………...…82

Figure II - 8. Mean wood density and landslide stability……………………………...83

Appendix II - Figure S1. Effect of drying temperature wood specific gravity…..…..122

Appendix II - Figure S2. Overall wood density distribution along the Andes-to-

Amazon……………………………………………………………………….123

Appendix II - Figure S3. Variance partitioning of wood density……………….…...124

Appendix II - Figure S4. Sampled wood density along an Andes-to-Amazon

elevational gradient…………………………………………………….…….125

Appendix II - Figure S5. Plot-level mean wood density variation…………………..126

Appendix II - Figure S6. Mean plot-level wood density for genus and family

basis………………………………………………………………………..…127

xiii

Appendix II - Figure S7. Wood density as a function of stem diameter……………..128

Figure III - 1. The estimated community temperature index (CTI) for the 38 permanent

forest plots……………………………………………………………………175

Figure III - 2. Estimated plot-level thermophilization rates for the Andean and

Amazonian forests plots……………………………………………………...176

Figure III - 3. Estimated changes in plot-level thermophilization rates due mortality,

recruitment and stem growth…………………………………………………177

Figure III - 4. Estimated genus-level thermal migration rates for 78 genera along the

Andes-to-Amazon elevational gradient………………………………………178

Figure III - 5. Estimated species-level thermal migration rates for 79 tree species along

the Andes-to-Amazon elevational gradient…………………………………..179

Appendix III - Figure S1. Species richness, number of individuals and basal area along

the Manu-Tambopata elevational transect…………………...………………204

Appendix III - Figure S2. Estimate elevational rages for 1896 arboreal species

including trees, tree ferns, palms and lianas along the gradient…………...…205

Appendix III - Figure S3. Mean community temperature index calculated based in

number of individuals and basal area…………..…………………...………..206

Appendix III - Figure S4. Estimated community temperature index including landslides

along the gradient…………………………...……………………………..…207

Appendix III - Figure S5. Linear least-square regression lines fitted to the estimated

CTI and elevation…………………………………………….………………208

Appendix III - Figure S6. Distribution of the thermophilization rates for the

Amazonian and Andean forests plots………………………………………...209

xiv

Appendix III - Figure S7. Thermophilization rates and the relationship with tree

mortality tree recruitment and growth…………………….………………….210

Figure IV - 1. Mean annualized vital rates along the Andes-to-Amazon elevational

gradient for mortality, recruitment, turnover and net stem change………..…258

Figure IV - 2. Above-ground carbon density along the Manu-Tambopata elevational

transect………………………………………………………………………..259

Figure IV - 3. Mean aboveground carbon density (ACD) along the Andes-to-Amazon

elevational gradient for mortality, recruitment and ACD net change………..260

Figure IV - 4. Trends in vital rates for 38 years of forests plots monitoring for stand

mortality, recruitment, turnover, and net stem change……………………….261

Figure IV - 5. Trends in aboveground carbon density (ACD) across 38 years of forests

plots monitoring for mortality, productivity, and net ACD change………….262

Appendix IV - Figure S1. Monthly precipitation (mm) recorded at the Rocotal

meteorological station at 2010 m of elevation ……..………………………..269

Appendix IV - Figure S2. Number of individuals along the Manu-Tambopata

elevational transect ……...…………………………………………………...270

Appendix IV - Figure S3. Trends in vital rates for 38 years of forests plots monitoring

for tree mortality, recruitment, turnover and net stem change………….……271

Appendix IV - Figure S4. Trends in aboveground carbon density for mortality,

productivity and net carbon change along the elevational gradient……….…272

Appendix IV - Figure S5. Relationship between annual net change in aboveground

carbon density and their annual change in stem number………………….….273

Figure V - 1. Panoramic view of the study site………..………….…………...……..352

Appendix V - Figure S1. Illustrated plates of tree species….……………..………..354

Appendix V - Figure S2. Illustrated plates of tree species….…………..…………..355

Appendix V - Figure S3. Illustrated plates of tree species….………..……………..356

xvi

ABSTRACT

A pressing question in current biology is how ecosystems in general, and species in particular, are responding to climate change in time and space. This dissertation responds to this question using an elevational gradient as a tool to understand the influence of temperature in biodiversity and ecosystem function. This approach has a long history in ecology and biogeography because these “natural laboratories” provide insights into how species and ecosystems respond to climate change. Particularly, I look at Amazonian and Andean forest response to climate gradients and changes in those gradients through time, understanding patterns and processes of tree species-specific functional traits (wood density), shifts in species composition towards taxa that have warmer mean ranges (thermophilization process) and changes in tree demography (rates of tree mortality and recruitment) in temporal- and spatial-scales. These studies are based on comprehensive long-term inventory forest data that includes 41 permanent plots along a 3500 m elevational gradient in Eastern Peru spanning 38 years.

Results of these studies showed: (1) Mean wood density declined from the

Amazonian forest to middle elevations, with an abrupt change corresponding to the elevation where clouds form, with an increase again to the montane treeline, and that interspecific variation in wood density along the elevational gradient was explained mainly by species turnover rather than elevation. (2) Amazonian-Andean tree communities showed slow thermophilization rates compared to the few results from previous studies across the tropics and the changes are mainly explained by increases in

xvii tree mortality over time rather than any range extensions or true migration. In addition, the Amazonian tree community’s response to warming was weak to nonexistent, in contrast to Andean forests that showed slow but non-zero responses to warming. (3) Tree demographic rates along the gradient showed that Andean forests can be as dynamic as

Amazonian forests, with equivalent dynamism up to ~2500 m of elevation. The net change in stem density and aboveground carbon was not related to elevation and the big losses were around the cloud base. Trends throughout time showed that tree dynamism and net carbon storage did not show an increase over time as previous studies showed.

However, the Amazonian-Andean forests are still acting as a long-term carbon sink, though a tendency was observed for long-term net carbon accumulation to decline. (4)

Plant inventory along the gradient showed that Andean flora is poorly known, particularly between 1000 - 2500 m of elevation, with 40 % of the species remaining as morphospecies (only identified to a genus or family level), unidentifiable by experts.

Together these studies showed that tree mortality has driven the shifts in species composition and carbon accumulation along the gradient, suggesting that forests will endure large scale compositional and ecosystem changes due to climate regimes with increased temperatures and more frequent drought events.

xviii

RESUMEN

Una pregunta apremiante en la biología actual es cómo los ecosistemas en general, y las especies en particular, están respondiendo al cambio climático en el tiempo y espacio. Esta tesis doctoral responde a esa pregunta utilizando un gradiente de elevación como herramienta para comprender la influencia de la temperatura en la biodiversidad y función del ecosistema. Este enfoque tiene una larga historia en la ecología y biogeografía porque estos "laboratorios naturales" proporcionan información crucial sobre cómo las especies y los ecosistemas están respondiendo al cambio climático. Este estudio en particular está enfocado en la respuesta de los bosques amazónicos y andinos a las gradientes climáticas y a los cambios en esas gradientes a través del tiempo, entendiendo los patrones y procesos de los rasgos funcionales de especies arbóreas especificas (densidad de la madera), los cambios en la composición de las especies hacia taxa que tengan rangos de elevación medios más cálidos (proceso de termofilización) y cambios en la demografía arbórea (tasas de mortalidad y reclutamiento arbóreo) en escalas temporales y espaciales. Estos estudios se basan en un inventario forestal comprehensivo a largo plazo, que incluye 41 parcelas permanentes a lo largo de un gradiente de elevación de 3500 m en el este peruano, abarcando 38 años de monitoreo.

Los resultados de estos estudios mostraron: (1) La densidad de madera promedio disminuyó desde los bosques amazónicos hasta elevaciones medias, con un cambio brusco que corresponde a elevaciones donde se forman las nubes, y un aumento hacia la línea de árboles montanos. La variación interespecífica en la densidad de madera a lo largo del gradiente se explica principalmente por el recambio de especies en lugar de la xix elevación. (2) Las comunidades arbóreas amazónicas y andinas mostraron tasas lentas de termofilización (migración) en comparación con resultados de estudios previos, estos cambios se explican principalmente por el aumento de la mortalidad arbórea a lo largo del tiempo en lugar de cualquier extensión de rango o migración verdadera. Además, las respuestas de las comunidades arbóreas amazónicas al calentamiento fueron débiles o inexistentes, en contraste con los bosques andinos que mostraron respuestas lentas, pero no nulas al calentamiento. (3) Las tasas demográficas arbóreas a lo largo del gradiente mostraron que los bosques andinos pueden ser tan dinámicos como los bosques amazónicos, con un dinamismo similar hasta ~2500 m de elevación. El cambio neto en el número de tallos y el carbono sobre el suelo no se relacionó con la elevación y las grandes pérdidas se produjeron alrededor de la base de nubes. Las tendencias a lo largo del tiempo mostraron que el dinamismo de los árboles y el almacenamiento neto de carbono no mostraron un aumento con el tiempo, como mostraron estudios anteriores. Sin embargo, los bosques amazónicos y andinos todavía están actuando como un sumidero de carbono a largo plazo, aunque se observó una tendencia de disminución en la acumulación neta de carbono a largo plazo. (4) El inventario de plantas a lo largo del gradiente mostró que la flora andina es poco conocida, particularmente entre 1000 - 2500 m de elevación, con el 40% de las especies arbóreas manteniéndose como morfoespecies

(solo identificadas a nivel de género o familia), no identificables por expertos. En conjunto, estos estudios muestran que la mortalidad arbórea está impulsando los cambios en la composición de especies y la acumulación de carbono a lo largo del gradiente de elevación, lo que sugiere que los bosques soportarán cambios en la composición de

xx especies y ecosistemas a gran escala debido a los regímenes climáticos con temperaturas más altas y eventos de sequía más frecuentes.

xxi

PISIYACHIYNIN

Mosoq watakunapin, huk ancha hatun tapakuy qallarin pachamamanchismanta, kausaykunamantaima, imaynatan paykuna kutipakunku, pisi paraman nishu ruphayman ima. Kay hatun yachaywan chayraqmin kutichisunchis chay tapukuyta, chay raykun huk hatunkaray orqopin umanchasunchis, imaynatan kausaykuna, pachamanchis ima, ruphaywan kausanku. Kay umanchaymin ñaupaq watakunamantaraqmin kallarin, tapukunku, imaynatan mallkikuna chhalanku pampamanta wichaykama. Yunka mallkikunawanmi, rupa rupa mallkikunawanmi ima munayku yachayta, imaynan paikuna wiñanku hoq niraq pachakunapi, imaynatan chhalankaku raphinkunata, tullunkunata ima.

Hamuq watakunapi, sach’a sach’a kaskallanchu kankaku, icha chhalankakuchu.

Mallkikunawan llank’ayka, ñawpaq watakunamantaraqmin qallarikun chay raykun

Manuq sach’a sach’anpin 41 suyukuna ruwakunku, chaypin tapukuykunatan kutichisunchis.

Imatataq tariyku? (1) Yunkapampapi kaq mallkikunaq tullun chuchulla, chaymantan k’aspikuna ñapuytan kallarinku phuyu sach’a sach’a chayaspan mallkikunaq tullunka kallarin chuchuyta, hawan orcoman chayaspan, tulluqa sinchi chuchuma tukupun. (2) Manu sach’a sach’apin, nishutan mallkikuna wañunku chay raykun, yunkapampapi kaq mallkikuna rupa rupa mallkikuna ima, alliliamanta thamanku orkota.

(3) Rupa rupa mallkikuna wiñanku wachanku ima yunkapampamallkikuna hina.

Mallkikunan, uraymanta hawankaman qasqallanmi wachakunku poqokunku ima, phuyupi kaq mallkikunallan pisita wachakunku. Manuq sach’a sach’aka carbonuta hap’ishallanku, ichaqa hatun pachakunapin chay carbonuta hap’iskanku pisiyapun. (4)

xxii

Ashqa mallkikunaq sutinmi manan yachakunchu, yakachan kuskan. Ni suti churaq amautakunapas yachankuchu. Tukunapaq niyku; mallkikuna wañuspanmi kallarichinku wiñachiyta mosoq mallkikunata, carbono ima chhalaykun yunkamanta ruparupaman.

Hamuq watakunapaqmi umullikun, sichi chhalaykuna manu sach’a sach’api, rupay, ch’aki pacha ima aswantaraq yanapanka.

xxiii

CHAPTER I

INTRODUCTION

One of the fundamental quests in plant ecology is to understand environmental influences on plant demography (Harper 1967) with a related and central theme being understanding species distributions across environmental gradients (von Humboldt 1838,

Whittaker 1967, Gentry 1988). Over the last four decades the understanding of ecological processes affecting plant demography and species distributions in different biomes has grown, and in the case of tropical forests this has been primarily based on long-term studies in lowland ecosystems (Hubbell and Foster R.B. 1992, Phillips and Gentry 1994,

Condit 1995). However, tropical montane ecosystems remain poorly studied (Budd et al.

2004). This is even though the montane forests of the tropical Andean-Amazonian elevational gradient are considered a natural laboratory for studying biotic and environmental controls on plant distribution and ecosystem function (Malhi et al. 2010,

Silman 2014). The Andes-Amazon system is also a model system for investigating ecological and evolutionary processes in plant communities and is imperative for evaluating the responses of tropical forests under the ongoing anthropogenic global climate change (Foster 2001). This area of research is expanding the frontiers of the understanding of longstanding questions related to diversity, species distribution and ecosystem function (Silman 2014).

Here I review the basic underpinnings of this research, including global change effects on tropical forests (i.e. temperature and drought) as well as the basic forest responses to these climate changes. Furthermore, this dissertation describes the changes in tree demography (rates of mortality and recruitment) from lowland-to-upland forests and their responses to climate change, showing empirically the changes in species distributional ranges (species migration) in response to warming, and how the ecosystem function is affected (carbon dynamics). A unique and continuous Andes-to-Amazon gradient spanning 3500 m in elevation in Eastern Peru is used to show evidence of forest responses to climate change and leverage a comprehensive long-term inventory forest data across 41 permanent plots spanning 38 years of field-based monitoring.

Tropical forest under global warming

Over the last 50 years, anthropogenic climate change has had a significant impact on physical and biological properties on Earth systems (Rosenzweig et al. 2008, Malhi et al. 2014). The increase in emissions of long-lived greenhouse gases coupled with the radiative forcing of the climate system is raising global temperatures (Forster et al. 2007) at a rate and magnitude unprecedented for millennia (Mann and Jones 2003). It has been predicted that global mean surface air temperature is likely to exceed 4 °C by the end of this century (Stocker et al. 2013) creating novel climates for current species (Williams et al. 2007). One of the central ongoing research questions in modern ecology is understanding the effects of anthropogenic climate change on Earth’s biomes, and within that specifically, it’s most diverse biome—the tropical forest (Malhi et al. 2008).

Markedly, human activity has dramatically increased CO2 emissions since the late

1700s, with fossil fuel combustion, deforestation, and cement production the primary sources. Deforestation is the second largest anthropogenic source of carbon dioxide to the atmosphere after fossil fuel combustion (van der Werf et al. 2009), contributing to 12 -

25% of global anthropogenic greenhouse gas emissions (van der Werf et al. 2009, Le

Quéré et al. 2015) with a continuous growth trend of CO2 emissions in the last decade (Le

Quéré et al. 2015). Global forests cover ~30% of the land surface and stores around 45% of the terrestrial carbon (Bonan 2008). At the global scale, both land and ocean ecosystems showed an increase in carbon uptake in the past 50 years (Ballantyne et al.

2012) acting as important carbon sinks. Tropical forests account for ~44% of global forest cover, storing ~55% of total forest carbon and play a central role in atmospheric carbon sequestration as they account for ~70% of global terrestrial net primary productivity (FAO 2011, Pan et al. 2011).

Rapid atmospheric warming is already pushing species out of their fundamental niches and even threatening their extinction. Tropical regions have experienced warming at an average of 0.26 °C per decade since the mid-1970s (Malhi and Wright 2004), and

2015 was the hottest year for Amazonia in the last century (Jiménez-Muñoz et al. 2016).

It is suggested that temperature increase could be the most pervasive climate impact on tropical ecosystems because tropical species are adapted to limited geographic and seasonal variation in temperature (Janzen 1967, Wright et al. 2009, Corlett 2011). For example, results of a simple model of temperature effects on species ranges predicts that a warming of 5 °C—expected over the next 100 years in tropical forests (Carter et al.

2007, Urrutia and Vuille 2009)—would place nearly 100% of Amazonian trees outside of their currently measured climatic niches (Feeley and Silman 2010).

One of the ways in which tree species are expected to respond to rapid warning is through shifting their distributions (e.g. Davis 1969, Williams et al. 2007) both poleward and towards higher elevations (Parmesan et al. 1999, Colwell et al. 2008, Chen et al.

2009, Feeley et al. 2011, 2013, Morueta-Holme et al. 2015). Individuals of long-lived organisms such as canopy trees will have to acclimate to long-term warming climates or die (Corlett 2011, Feeley et al. 2012). For example, the life-span of Amazonian canopy trees routinely exceed 300 years and can last over 1400 years (Chambers et al. 1998,

Fichtler et al. 2003), and many individuals alive now will still be so in 2100 if they can acclimate as they did to climate changes they have already experienced.

Range shifts and changes in ecosystem properties such as carbon storage are both underpinned by demographic changes, and recent studies have shown that tropical forests are responding to global environmental changes by increasing or decreasing tree population sizes over time. For instance, Amazonian forests are becoming more dynamic in terms of both mortality and recruitment rates (Phillips et al. 2004, Rice et al. 2004,

Susan et al. 2009) and are hypothesized to have been acting as a long-term carbon sink since the 1970s in response to environmental conditions (Lewis et al. 2004, Valencia et al. 2009, Brienen et al. 2015). However, it is unknown whether Andean forest dynamics and long-term carbon accumulation are also changing in response to global change. In

Amazonian forests, the gain in stems over the loss is accelerating tree dynamics over time (Lewis et al. 2004, Phillips et al. 2004) and potentially increasing the carbon stocks in tropical forests. Though changes in tree demography (trees dying, recruiting, and

4 growing) are crucial to understanding tree dynamics and carbon balance in tropical forests, these basic but fundamental trends for the Andes remain unknown, much less their underlying causes.

Tree growth responses to global warming are fundamental to understanding forest ecosystem ecology and predicting future ecosystem services. In the tropics, tree growth is positively correlated with temperature (Kitayama and Aiba 2002, Rapp et al. 2012, Clark et al. 2015). However, while some pantropical studies show a net-carbon sink in tropical forests due to increased growth rates and a net shift in the balance of mortality and recruitment (Lewis et al. 2004, Phillips et al. 2004), other studies report a decrease in growth rates in lowland forest (Feeley et al. 2007, Clark et al. 2010) that could be explained by the negative correlation between tree growth and minimum daily temperature. These may be caused by elevated respiration costs when air temperature increases, particularly variations in the mean annual night-time temperature (Feeley et al.

2007, Clark et al. 2010). Both cases indicate that carbon balance is sensitive to temperature across a range of tropical tree species. A central goal of this dissertation is to examine the processes of growth, recruitment, and mortality on species ranges and ecosystem carbon dynamics.

The pace of droughts in tropical forest

Drought events play a major role in shaping tropical forests and have large effects on tree mortality and growth (Phillips et al. 2010, Bonal et al. 2016). Mortality of tropical trees has been widely observed in natural droughts that are often associated with multi-

5 year climatic cycles explained by El Niño-Southern Oscillation (ENSO) events

(Williamson et al. 2000, Condit et al. 2004, Slik 2004). However, the 2015-2016 ENSO drought event was the strongest since 1979 and is attributed to sea surface warming in the

Central Equatorial Pacific (Jiménez-Muñoz et al. 2016), and is an event that potentially will increase tree mortality in the current decade. In the tropics, some abnormal drought events are associated with an increase in sea surface temperatures (SSTs) in the tropical

North Atlantic Ocean and are thought to be 1-in-100 year events (Marengo et al. 2008,

2011, Malhi et al. 2008). In the last decade, two major warming Atlantic SST-associate drought events were identified in 2005 and 2010, greatly affecting the Amazon basin

(Phillips et al. 2009, Lewis et al. 2011, Marengo et al. 2011), resulting in peaks of fire activity, dried lakes, the low levels in many rivers in southwestern Amazon.

These droughts increased tree mortality in lowlands forests with major changes in forest composition and structure (e.g. Phillips et al. 2010, Esquivel-Muelbert et al. 2018).

One of the short-term responses of forests to drought is high tree mortality, with evidence of lagged mortality in one or more years changing forests from carbon sinks into sources in some tropical regions (Phillips et al. 2009, Corlett 2016). The droughts had particularly strong effects on large trees (stems ˃ 40 cm in diameter) (Phillips et al. 2010). Hydraulic failure and carbon starvation have been proposed as mechanisms inducing mortality in large vs. smaller trees, with hydraulic failure being the most supported hypothesis (Choat et al. 2012, Rowland et al. 2015). Understanding of forest ecosystem responses to drought in lowland Amazonian systems is increasing (Phillips et al. 2010, Gatti et al.

2014), with advances in tying drought effects to functional traits (Chao et al. 2008, Kraft et al. 2010). However, drought effects in montane regions such as the tropical Andes

6 remain unstudied, and it is unclear whether simple extrapolations from lowland systems are sufficient for understanding upland changes in the moisture regime because of cloud immersion and steep temperature gradients (Pounds et al. 1999, Halladay et al. 2012,

Rapp and Silman 2012).

While predicting tree mortality in response to drought is difficult, it has been suggested that wood density—an important functional trait in plant life history—could be used as a proxy to predict mortality risk in tropical lowland forest (Chao et al. 2008, Kraft et al. 2010). This wood density-mortality risk relationship can be explained by the positive relationship of high wood density and thick-walled vessels to protect vessels from implosion when water shortages create strongly negative xylem potentials (Hacke et al. 2001). This dissertation explores the patterns of wood density variations across the

Andes-Amazon gradient with a focus on explaining patterns of mortality and drought susceptibility.

Research objectives

This dissertation presents a study of how Andean-Amazonian forests respond to ongoing climate change that uses long-term empirical field data along an elevational gradient. For the purposes of this dissertation, I installed a network of 24 1-ha permanent plots spanning from 400 to 3725 m of elevation with multiple censuses since 2003, as a part of the Andes Biodiversity and Ecosystem Research Group (ABERG). Additionally, since science is also often a collaborative effort, I included data from17 lowland plots

(below 500 m) from enthusiastic collaborators of RAINFOR and GEM plot networks.

Together, the studies in this dissertation provide insights into the responses of Andean and Amazonian forests to global change based on comprehensive long-term forest plots’ data including over 41,000 stems and ~1,900 tree species across 41 permanent forest plots, along a 3500 m elevational gradient that spans the past 38 years in Eastern Peru (38 years below 500 m, 14 years above 500 m). All stems ≥10 cm of diameter at breast height in each permanent forest plot are marked, identified to species level and measured multiple times for up to four decades.

Chapter II describes the patterns of wood density along the elevational gradient.

Wood density is a functional trait that integrates many aspects of plant form and function and is central in forest carbon estimations. Using field-collected wood samples to calculate species-level wood density and combined with global databases and permanent plot data, I test the extent to which elevation, species composition, phylogenetic affinity, and forest structure determine variation in wood density along the elevational gradient.

Chapter III and IV examine tree demography looking at the mortality (dead trees) and recruitment rates (new trees ≥10 cm of diameter at breast height into the plot) changes in response to climate change along the elevational gradient over 38 years.

Chapter III examines shifts in species distribution due to climate change looking at both temporal and spatial changes in tree demography. I tested the thermophilization hypothesis (shifts in community composition due to the replacement of species with cooler tolerances with those with warm tolerances at any point) and the species thermal migration hypothesis (changes in the distributional range of specific plant species) in response to climate change. Chapter IV describes the regional- and temporal-scale patterns of tree mortality, recruitment and stem-change rates along the elevational

8 gradient, and how those demographic rates influence aboveground carbon dynamics. I integrated the temporal and spatial forest plot data with wood density data (from Chapter

II) and ground measured stem length into an allometric equation to estimate aboveground carbon density accumulation and change for the past 38 years. Chapter V ( published as

Farfan-Rios et al. 2015) describes the depth of the floristic work that underpins all of these ecological efforts and presents an annotated checklist for the Andean arborescent life forms including trees, arborescent ferns, palms, and lianas for the elevational gradient used in this study. It presents species elevational ranges (including range expansions), the distribution of endemic species, and new records for the Peruvian flora, along 3000 m of elevation change across 20 km of geographic distance.

In their sum, the results presented in this dissertation represents a novel contribution to understanding the fate of Amazonian and Andean forests in the face of ongoing and future climate change, as well as providing basic knowledge of the structural and compositional dynamics of Andean forests.

Literature cited

Ballantyne, A. P., C. B. Alden, J. B. Miller, P. P. Tans, and J. W. C. White. 2012.

Increase in observed net carbon dioxide uptake by land and oceans during the past

50 years. Nature 488:70–72.

Berry, Z. C., N. C. Emery, S. G. Gotsch, and G. R. Goldsmith. 2019. Foliar water uptake:

Processes, pathways, and integration into plant water budgets. Plant, Cell &

Environment 42:410–423.

Bonal, D., B. Burban, C. Stahl, F. Wagner, and B. Hérault. 2016. The response of tropical

rainforests to drought—lessons from recent research and future prospects. Annals of

Forest Science 73:27–44.

Bonan, G. B. 2008. Forests and climate change: forcings, feedbacks, and the climate

benefits of forests. Science (New York, N.Y.) 320:1444–9.

Bony, S., B. Stevens, D. M. W. Frierson, C. Jakob, M. Kageyama, R. Pincus, T. G.

Shepherd, S. C. Sherwood, A. P. Siebesma, A. H. Sobel, M. Watanabe, and M. J.

Webb. 2015. Clouds, circulation and climate sensitivity. Nature Geoscience 8:261–

268.

Brienen, R. J. W., O. L. Phillips, T. R. Feldpausch, E. Gloor, T. R. Baker, J. Lloyd, G.

Lopez-Gonzalez, A. Monteagudo-Mendoza, Y. Malhi, S. L. Lewis, R. Vásquez

Martinez, M. Alexiades, E. Álvarez Dávila, P. Alvarez-Loayza, A. Andrade, L. E.

O. C. Aragão, A. Araujo-Murakami, E. J. M. M. Arets, L. Arroyo, G. A. Aymard C.,

O. S. Bánki, C. Baraloto, J. Barroso, D. Bonal, R. G. A. Boot, J. L. C. Camargo, C.

V. Castilho, V. Chama, K. J. Chao, J. Chave, J. A. Comiskey, F. Cornejo Valverde,

L. da Costa, E. A. de Oliveira, A. Di Fiore, T. L. Erwin, S. Fauset, M. Forsthofer, D.

R. Galbraith, E. S. Grahame, N. Groot, B. Hérault, N. Higuchi, E. N. Honorio

Coronado, H. Keeling, T. J. Killeen, W. F. Laurance, S. Laurance, J. Licona, W. E.

Magnussen, B. S. Marimon, B. H. Marimon-Junior, C. Mendoza, D. A. Neill, E. M.

Nogueira, P. Núñez, N. C. Pallqui Camacho, A. Parada, G. Pardo-Molina, J.

Peacock, M. Peña-Claros, G. C. Pickavance, N. C. A. Pitman, L. Poorter, A. Prieto,

C. A. Quesada, F. Ramírez, H. Ramírez-Angulo, Z. Restrepo, A. Roopsind, A.

Rudas, R. P. Salomão, M. Schwarz, N. Silva, J. E. Silva-Espejo, M. Silveira, J.

Stropp, J. Talbot, H. ter Steege, J. Teran-Aguilar, J. Terborgh, R. Thomas-Caesar,

M. Toledo, M. Torello-Raventos, R. K. Umetsu, G. M. F. van der Heijden, P. van

der Hout, I. C. Guimarães Vieira, S. A. Vieira, E. Vilanova, V. A. Vos, and R. J.

Zagt. 2015. Long-term decline of the Amazon carbon sink. Nature 519:344–348.

Budd, P., I. May, L. Miles, and J. Sayer. 2004. Cloud Forest Agenda. United Nations

Environment - Programme-World Conservation Monitoring Centre, Cambridge,

UK.

Bush, M. B. 2002. Distributional change and conservation on the Andean flank: A

palaeoecological perspective. Global Ecology and Biogeography 11:463–473.

Bush, M. B., M. R. Silman, and D. H. Urrego. 2004. 48,000 years of climate and forest

change in a biodiversity hot spot. Science (New York, N.Y.) 303:827–9.

Carter, T. R., R. N. Jones, X. Lu, S. Bhadwal, C. Conde, L. O. Mearns, B. C. O’Neill, M.

D. A. Rounsevell, and M. B. Zurek. 2007. New Assessment Methods and the

Characterisation of Future Conditions. Climate Change 2007: Impacts, Adaptation

and Vulnerability. Contribution of Working Group II to the Fourth Assessment

Report of the Intergovernmental Panel on Climate Change. Page (M. Parry, O.

Canziani, J. Palutikof, P. Van Der Linden, and C. Hanson, Eds.). Cambridge

University Press, Cambridge, UK.

Chambers, J. Q., N. Higuchi, and J. P. Schimel. 1998. Ancient trees in Amazonia. Nature

391:135–136.

Chao, K. J., O. L. Phillips, E. Gloor, A. Monteagudo, A. Torres-Lezama, and R. V

Martinez. 2008. Growth and wood density predict tree mortality in Amazon forests.

Journal of Ecology 96:281–292.

Chen, I.-C., H.-J. Shiu, S. Benedick, J. D. Holloway, V. K. Chey, H. S. Barlow, J. K.

Hill, and C. D. Thomas. 2009. Elevation increases in moth assemblages over 42

years on a tropical mountain. Proceedings of the National Academy of Sciences of

the United States of America 106:1479–1483.

Choat, B., S. Jansen, T. J. Brodribb, H. Cochard, S. Delzon, R. Bhaskar, S. J. Bucci, T. S.

Feild, S. M. Gleason, U. G. Hacke, A. L. Jacobsen, F. Lens, H. Maherali, J.

Martínez-Vilalta, S. Mayr, M. Mencuccini, P. J. Mitchell, A. Nardini, J. Pittermann,

R. B. Pratt, J. S. Sperry, M. Westoby, I. J. Wright, and A. E. Zanne. 2012. Global

convergence in the vulnerability of forests to drought. Nature 491:752–5.

Clark, D. B., D. A. Clark, and S. F. Oberbauer. 2010. Annual wood production in a

tropical rain forest in NE Costa Rica linked to climatic variation but not to

increasing CO2. Global Change Biology 16:747–759.

Clark, D. B., J. Hurtado, and S. S. Saatchi. 2015. Tropical Rain Forest Structure, Tree

Growth and Dynamics along a 2700-m Elevational Transect in Costa Rica. PloS one

10:e0122905.

Colwell, R. K., G. Brehm, C. L. Cardelús, A. C. Gilman, and J. T. Longino. 2008. Global

warming, elevational range shifts, and lowland biotic attrition in the wet tropics.

Science (New York, N.Y.) 322:258–61.

Condit, R. 1995. Research in large, long-term tropical forest plots. Trends in Ecology &

Evolution 10:18–22.

Condit, R., S. Aguilar, A. Hernandez, R. Perez, S. Lao, G. Angehr, S. P. Hubbell, and R.

B. Foster. 2004. Tropical forest dynamics across a rainfall gradient and the impact of

an El Nino dry season. Journal of Tropical Ecology 20:51–72.

Corlett, R. T. 2011. Impacts of warming on tropical lowland rainforests. Trends in

Ecology and Evolution 26:606–613.

Corlett, R. T. 2016. The Impacts of Droughts in Tropical Forests. Trends in plant science.

Cramer, W., A. Bondeau, S. Schaphoff, W. Lucht, B. Smith, and S. Sitch. 2004. Tropical

forests and the global carbon cycle: impacts of atmospheric carbon dioxide, climate

change and rate of deforestation. Philosophical Transactions of the Royal Society of

London B: Biological Sciences 359:331–343.

Davis, M. B. 1969. Climatic changes in southern Connecticut recorded by pollen

deposition at Rogers Lake. Ecology 50:409–422.

Esquivel-Muelbert, A., T. R. Baker, K. G. Dexter, S. L. Lewis, R. J. W. Brienen, T. R.

Feldpausch, J. Lloyd, A. Monteagudo-Mendoza, L. Arroyo, E. Álvarez-Dávila, N.

Higuchi, B. S. Marimon, B. H. Marimon-Junior, M. Silveira, E. Vilanova, E. Gloor,

Y. Malhi, J. Chave, J. Barlow, D. Bonal, N. Davila Cardozo, T. Erwin, S. Fauset, B.

Hérault, S. Laurance, L. Poorter, L. Qie, C. Stahl, M. J. P. Sullivan, H. ter Steege, V.

A. Vos, P. A. Zuidema, E. Almeida, E. Almeida de Oliveira, A. Andrade, S. A.

Vieira, L. Aragão, A. Araujo-Murakami, E. Arets, G. A. Aymard C, P. B. Camargo,

J. G. Barroso, F. Bongers, R. Boot, J. L. Camargo, W. Castro, V. Chama Moscoso,

J. Comiskey, F. Cornejo Valverde, A. C. Lola da Costa, J. del Aguila Pasquel, T. Di

Fiore, L. Fernanda Duque, F. Elias, J. Engel, G. Flores Llampazo, D. Galbraith, R.

Herrera Fernández, E. Honorio Coronado, W. Hubau, E. Jimenez-Rojas, A. J. N.

Lima, R. K. Umetsu, W. Laurance, G. Lopez-Gonzalez, T. Lovejoy, O. Aurelio

Melo Cruz, P. S. Morandi, D. Neill, P. Núñez Vargas, N. C. Pallqui, A. Parada

Gutierrez, G. Pardo, J. Peacock, M. Peña-Claros, M. C. Peñuela-Mora, P. Petronelli,

G. C. Pickavance, N. Pitman, A. Prieto, C. Quesada, H. Ramírez-Angulo, M. Réjou-

Méchain, Z. Restrepo Correa, A. Roopsind, A. Rudas, R. Salomão, N. Silva, J. Silva

Espejo, J. Singh, J. Stropp, J. Terborgh, R. Thomas, M. Toledo, A. Torres-Lezama,

L. Valenzuela Gamarra, P. J. van de Meer, G. van der Heijden, P. van der Hout, R.

Vasquez Martinez, C. Vela, I. C. G. Vieira, and O. L. Phillips. 2018. Compositional

response of Amazon forests to climate change. Global Change Biology.

FAO. 2011. State of the World’s Forests 2011. Rome.

Farfan-Rios, W., K. Garcia-cabrera, N. Salinas, M. N. Raurau-quisiyupanqui, and M. R.

Silman. 2015. Lista anotada de árboles y afines en los bosques montanos del sureste

peruano : la importancia de seguir recolectando. Revista Peruana de Biología

22:145–174.

Feeley, K. J., J. Hurtado, S. Saatchi, M. R. Silman, and D. B. Clark. 2013. Compositional

shifts in Costa Rican forests due to climate-driven species migrations. Global change

biology 19:3472–80.

Feeley, K. J., S. Joseph Wright, M. N. Nur Supardi, A. R. Kassim, and S. J. Davies. 2007.

Decelerating growth in tropical forest trees. Ecology Letters 10:461–469.

Feeley, K. J., E. M. Rehm, and B. Machovina. 2012. perspective: The responses of

tropical forest species to global climate change: acclimate, adapt, migrate, or go

extinct? Frontiers of Biogeography 4.

Feeley, K. J., and M. R. Silman. 2010. Biotic attrition from tropical forests correcting for

truncated temperature niches. Global Change Biology 16:1830–1836.

Feeley, K. J., M. R. Silman, M. B. Bush, W. Farfan, K. G. Cabrera, Y. Malhi, P. Meir, N.

S. Revilla, M. N. R. Quisiyupanqui, and S. Saatchi. 2011. Upslope migration of

Andean trees. Journal of Biogeography 38:783–791.

Fichtler, E., D. A. Clark, and M. Worbes. 2003. Age and Long-term Growth of Trees in

an Old-growth Tropical Rain Forest, Based on Analyses of Tree Rings and 14C1.

BIOTROPICA 35:306.

Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D. W. Fahey, J. Haywood,

J. Lean, D. C. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz, and R.

Van Dorland. 2007. Changes in Atmospheric Constituents and in Radiative Forcing.

Chapter 2.

Foster, P. 2001. The potential negative impacts of global climate change on tropical

montane cloud forests. Earth-Science Reviews 55:73–106.

Gatti, L. V, M. Gloor, J. B. Miller, C. E. Doughty, Y. Malhi, L. G. Domingues, L. S.

Basso, A. Martinewski, C. S. C. Correia, V. F. Borges, S. Freitas, R. Braz, L. O.

Anderson, H. Rocha, J. Grace, O. L. Phillips, and J. Lloyd. 2014. Drought sensitivity

of Amazonian carbon balance revealed by atmospheric measurements. Nature

506:76–80.

Gentry, A. H. 1988. Changes in Plant Community Diversity and Floristic Composition on

Environmental and Geographical Gradients. Annals of the Missouri Botanical

Garden 75:1–34.

Goldsmith, G. R., N. J. Matzke, and T. E. Dawson. 2013. The incidence and implications

of clouds for cloud forest plant water relations. Ecology Letters 16:307–314.

Hacke, U. G., J. S. Sperry, W. T. Pockman, S. D. Davis, and K. A. McCulloh. 2001.

Trends in wood density and structure are linked to prevention of xylem implosion by

negative pressure. Oecologia 126:457–461.

Halladay, K., Y. Malhi, and M. New. 2012. Cloud frequency climatology at the

Andes/Amazon transition: 2. Trends and variability. Journal of Geophysical

Research-Atmospheres 117:D23103, doi:10.1029/2012JD017789.

Harper, J. L. 1967. A Darwinian Approach to Plant Ecology. Source Journal of Applied

Ecology 4:267–290.

Hubbell, S. P., and R. Foster R.B. 1992. Short-Term Dynamics of a Neotropical Forest.

BioScience 42:822–828. von Humboldt, A. 1838. Notice de Deux Tentatives d’Ascension du Chimborazo (A.

Pihan de la Forest, Paris).

Janzen, D. H. 1967. Why Mountain Passes are Higher in the Tropics. The American

Naturalist 101:233–249.

Jiménez-Muñoz, J. C., C. Mattar, J. Barichivich, A. Santamaría-Artigas, K. Takahashi, Y.

Malhi, J. A. Sobrino, and G. van der Schrier. 2016. Record-breaking warming and

extreme drought in the Amazon rainforest during the course of El Niño 2015–2016.

Scientific Reports 6:33130.

Kitayama, K., and S.-I. Aiba. 2002. Ecosystem structure and productivity of tropical rain

forests along altitudinal gradients with contrasting soil phosphorus pools on Mount

Kinabalu, Borneo. Journal of Ecology 90:37–51.

Kraft, N. J. B., M. R. Metz, R. S. Condit, and J. Chave. 2010. The relationship between

wood density and mortality in a global tropical forest data set. The New phytologist

188:1124–36.

Lewis, S. L., P. M. Brando, O. L. Phillips, G. M. F. van der Heijden, and D. Nepstad.

2011. The 2010 Amazon Drought. Science 331:554.

Lewis, S. L., O. L. Phillips, T. R. Baker, J. Lloyd, Y. Malhi, S. Almeida, N. Higuchi, W.

F. Laurance, D. A. Neill, J. N. M. Silva, J. Terborgh, A. Torres Lezama, R. Vásquez

Martinez, S. Brown, J. Chave, C. Kuebler, P. Núñez Vargas, and B. Vinceti. 2004.

Concerted changes in tropical forest structure and dynamics: evidence from 50

South American long-term plots. Philosophical Transactions of the Royal Society of

London. Series B: Biological Sciences 359:421–436.

Malhi, Y., T. A. Gardner, G. R. Goldsmith, M. R. Silman, and P. Zelazowski. 2014.

Tropical Forests in the Anthropocene. Annual Review of Environment and

Resources 39:125–159.

Malhi, Y., J. T. Roberts, R. A. Betts, T. J. Killeen, W. Li, and C. A. Nobre. 2008. Climate

change, deforestation, and the fate of the Amazon. Science (New York, N.Y.)

319:169–72.

Malhi, Y., M. Silman, N. Salinas, M. Bush, P. Meir, and S. Saatchi. 2010. Introduction:

Elevation gradients in the tropics: laboratories for ecosystem ecology and global

change research. Global Change Biology 16:3171–3175.

Malhi, Y., and J. Wright. 2004. Spatial patterns and recent trends in the climate of

tropical rainforest regions. Philosophical transactions of the Royal Society of

London. Series B, Biological sciences 359:311–29.

Mann, M. E., and P. D. Jones. 2003. Global surface temperatures over the past two

millennia. Geophysical Research Letters 30.

Marengo, J. A., C. A. Nobre, J. Tomasella, M. F. Cardoso, and M. D. Oyama. 2008.

Hydro-climatic and ecological behaviour of the drought of Amazonia in 2005.

Philosophical Transactions of the Royal Society B-Biological Sciences 363:1773–

1778.

Marengo, J. A., J. Tomasella, L. M. Alves, W. R. Soares, and D. A. Rodriguez. 2011. The

drought of 2010 in the context of historical droughts in the Amazon region.

Geophysical Research Letters 38:n/a-n/a.

Morueta-Holme, N., K. Engemann, P. Sandoval-Acuña, J. D. Jonas, R. M. Segnitz, and

J.-C. Svenning. 2015. Strong upslope shifts in Chimborazo’s vegetation over two

centuries since Humboldt. Proceedings of the National Academy of Sciences of the

United States of America.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. da Fonseca, and J. Kent. 2000.

Biodiversity hotspots for conservation priorities. Nature 403:5.

Pan, Y., R. A. Birdsey, J. Fang, R. Houghton, P. E. Kauppi, W. A. Kurz, O. L. Phillips,

A. Shvidenko, S. L. Lewis, J. G. Canadell, P. Ciais, R. B. Jackson, S. W. Pacala, A.

D. McGuire, S. Piao, A. Rautiainen, S. Sitch, and D. Hayes. 2011. A large and

persistent carbon sink in the world’s forests. Science (New York, N.Y.) 333:988–93.

Parmesan, C., N. Ryrholm, C. Stefanescu, J. K. Hill, C. D. Thomas, H. Descimon#, B.

Huntley, L. Kaila, J. Kullberg, T. Tammaru, W. J. Tennent, J. A. Thomas, and M.

Warren. 1999. Poleward shifts in geographical ranges of butterfly species associated

with regional warming. Nature 399:579–583.

Phillips, O. L., L. E. O. C. Aragão, S. L. Lewis, J. B. Fisher, J. Lloyd, G. López-

González, Y. Malhi, A. Monteagudo, J. Peacock, C. A. Quesada, G. van der

Heijden, S. Almeida, I. Amaral, L. Arroyo, G. Aymard, T. R. Baker, O. Bánki, L.

Blanc, D. Bonal, P. Brando, J. Chave, Á. C. A. de Oliveira, N. D. Cardozo, C. I.

Czimczik, T. R. Feldpausch, M. A. Freitas, E. Gloor, N. Higuchi, E. Jiménez, G.

Lloyd, P. Meir, C. Mendoza, A. Morel, D. A. Neill, D. Nepstad, S. Patiño, M. C.

Peñuela, A. Prieto, F. Ramírez, M. Schwarz, J. Silva, M. Silveira, A. S. Thomas, H.

ter Steege, J. Stropp, R. Vásquez, P. Zelazowski, E. A. Dávila, S. Andelman, A.

Andrade, K.-J. Chao, T. Erwin, A. Di Fiore, E. H. C., H. Keeling, T. J. Killeen, W.

F. Laurance, A. P. Cruz, N. C. A. Pitman, P. N. Vargas, H. Ramírez-Angulo, A.

Rudas, R. Salamão, N. Silva, J. Terborgh, and A. Torres-Lezama. 2009. Drought

Sensitivity of the Amazon Rainforest. Science 323:1344–1347.

Phillips, O. L., T. R. Baker, L. Arroyo, N. Higuchi, T. J. Killeen, W. F. Laurance, S. L.

Lewis, J. Lloyd, Y. Malhi, A. Monteagudo, D. A. Neill, P. N. Vargas, J. N. M. Silva,

J. Terborgh, R. V Martinez, M. Alexiades, S. Almeida, S. Brown, J. Chave, J. A.

Comiskey, C. I. Czimczik, A. Di Fiore, T. Erwin, C. Kuebler, S. G. Laurance, H. E.

M. Nascimento, J. Olivier, W. Palacios, S. Patino, N. C. A. Pitman, C. A. Quesada,

M. Salidas, A. T. Lezama, and B. Vinceti. 2004. Pattern and process in Amazon tree

turnover, 1976-2001. Philosophical Transactions of the Royal Society of London

Series B-Biological Sciences 359:381–407.

Phillips, O. L., and A. H. Gentry. 1994. Increasing Turnover through Time in Tropical

Forests. Science 263:954–958.

Phillips, O. L., G. van der Heijden, S. L. Lewis, G. Lopez-Gonzalez, L. Aragao, J. Lloyd,

Y. Malhi, A. Monteagudo, S. Almeida, E. A. Davila, I. Amaral, S. Andelman, A.

Andrade, L. Arroyo, G. Aymard, T. R. Baker, L. Blanc, D. Bonal, A. C. A. de

Oliveira, K. J. Chao, N. D. Cardozo, L. da Costa, T. R. Feldpausch, J. B. Fisher, N.

M. Fyllas, M. A. Freitas, D. Galbraith, E. Gloor, N. Higuchi, E. Honorio, E.

Jimenez, H. Keeling, T. J. Killeen, J. C. Lovett, P. Meir, C. Mendoza, A. Morel, P.

N. Vargas, S. Patino, K. S. H. Peh, A. P. Cruz, A. Prieto, C. A. Quesada, F.

Ramirez, H. Ramirez, A. Rudas, R. Salamao, M. Schwarz, J. Silva, M. Silveira, J.

W. F. Slik, B. Sonke, A. S. Thomas, J. Stropp, J. R. D. Taplin, R. Vasquez, and E.

Vilanova. 2010. Drought-mortality relationships for tropical forests. New

Phytologist 187:631–646.

Pounds, J., M. Fogden, and J. Campbell. 1999. Biological response to climate change on

a tropical mountain. Nature 398:611–615.

Le Quéré, C., R. Moriarty, R. Andrew, G. Peters, P. Ciais, P. Friedlingstein, S. Jones, S.

Sitch, P. Tans, A. Arneth, T. Boden, L. Bopp, Y. Bozec, J. Canadell, L. Chini, F.

Chevallier, C. Cosca, I. Harris, M. Hoppema, R. Houghton, J. House, A. Jain, T.

Johannessen, E. Kato, R. Keeling, V. Kitidis, K. Klein Goldewijk, C. Koven, C.

Landa, P. Landschützer, A. Lenton, I. Lima, G. Marland, J. Mathis, N. Metzl, Y.

Nojiri, A. Olsen, T. Ono, S. Peng, W. Peters, B. Pfeil, B. Poulter, M. Raupach, P.

Regnier, C. Rödenbeck, S. Saito, J. Salisbury, U. Schuster, J. Schwinger, R.

Séférian, J. Segschneider, T. Steinhoff, B. Stocker, A. Sutton, T. Takahashi, B.

Tilbrook, G. Van Der Werf, N. Viovy, Y. Wang, R. Wanninkhof, A. Wiltshire, and

N. Zeng. 2015. Global carbon budget 2014.

Rapp, J. M., and M. R. Silman. 2012. Diurnal, seasonal, and altitudinal trends in

microclimate across a tropical montane cloud forest. Climate Research 55:17–32.

Rapp, J. M., and M. R. Silman. 2014. Epiphyte response to drought and experimental

warming in an Andean cloud forest. F1000Research 3:7.

Rapp, J. M., M. R. Silman, J. S. Clark, C. A. J. Girardin, D. Galiano, and R. Tito. 2012.

Intra- and interspecific tree growth across a long altitudinal gradient in the Peruvian

Andes. Ecology 93:2061–2072.

Rice, A. H., E. H. Pyle, S. R. Saleska, L. Hutyra, M. Palace, M. Keller, P. B. de

Camargo, K. Portilho, D. F. Marques, and S. C. Wofsy. 2004. Carbon balance and

vegetation dynamics in an old-growth Amazonian forest. Ecological Applications

14:S55–S71.

Rosenzweig, C., D. Karoly, M. Vicarelli, P. Neofotis, Q. Wu, G. Casassa, A. Menzel, T.

L. Root, N. Estrella, B. Seguin, P. Tryjanowski, C. Liu, S. Rawlins, and A. Imeson.

2008. Attributing physical and biological impacts to anthropogenic climate change.

Nature 453:353–7.

Rowland, L., A. C. L. da Costa, D. R. Galbraith, R. S. Oliveira, O. J. Binks, A. A. R.

Oliveira, A. M. Pullen, C. E. Doughty, D. B. Metcalfe, S. S. Vasconcelos, L. V

Ferreira, Y. Malhi, J. Grace, M. Mencuccini, and P. Meir. 2015. Death from drought

in tropical forests is triggered by hydraulics not carbon starvation. Nature 528:119–

122.

Silman, M. R. 2014. Functional megadiversity. Proceedings of the National Academy of

Sciences of the United States of America 111:5763–4.

Slik, J. W. F. 2004. El Niño droughts and their effects on tree species composition and

diversity in tropical rain forests. Oecologia 141:114–120.

Stocker, T. F., D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y.

Xia, V. Bex, and P. M. Midgley, editors. 2013. IPCC, 2013: Climate Change 2013:

The Physical Science Basis. Contribution of Working Group I to the Fifth

Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge

University Press, Cambridge, United Kingdom and New York, NY, USA.

Susan, G. W. L., W. F. Laurance, E. M. N. Henrique, A. Andrade, P. M. Fearnside, R. G.

R. Expedito, and R. Condit. 2009. Long-term variation in Amazon forest dynamics.

Journal of Vegetation Science 20:323–333.

Urrutia, R., and M. Vuille. 2009. Climate change projections for the tropical Andes using

a regional climate model: Temperature and precipitation simulations for the end of

the 21st century. Journal of Geophysical Research 114:D02108.

Valencia, R., R. Condit, H. C. Muller-Landau, C. Hernandez, and H. Navarrete. 2009.

Dissecting biomass dynamics in a large Amazonian forest plot. Journal of Tropical

Ecology 25:473–482. 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.

Whittaker, R. H. 1967. Gradient analysis of vegetation. Biological Reviews 42:207–264.

Williams, J. W., S. T. Jackson, and J. E. Kutzbach. 2007. Projected distributions of novel

and disappearing climates by 2100 AD. Proceedings of the National Academy of

Sciences 104:5738–5742.

Williamson, G. B., W. F. Laurance, A. A. Oliveira, P. Delamonica, C. Gascon, T. E.

Lovejoy, and L. Pohl. 2000. Amazonian tree mortality during the 1997 El Nino

drought. Conservation Biology 14:1538–1542.

Wright, S. J., H. C. Muller-Landau, and J. Schipper. 2009. The future of tropical species

on a warmer planet. Conservation Biology 23:1418–1426.

Wurdack, K. J., and W. Farfan-Rios. 2017. Incadendron : a new genus of Euphorbiaceae

tribe Hippomaneae from the sub-Andean cordilleras of Ecuador and Peru.

PhytoKeys 85:69–86.

CHAPTER II

LANDSCAPE-SCALE WOOD DENSITY VARIATION ACROSS AN ANDES-TO-

AMAZON ELEVATIONAL GRADIENT

William Farfan-Rios1,2, Sassan Saatchi3, Imma Oliveras4, Yadvinder Malhi4,

Chelsea M. Robinson5, Oliver L. Phillips6, Alex Nina-Quispe7, Juan A. Gibaja8, Israel

Cuba8, Karina Garcia-Cabrera1,2, Norma Salinas-Revilla7, John Terborgh9, Nigel

Pitman10, Rodolfo Vasquez11, Abel Monteagudo Mendoza11, Terry Erwin12, Percy Nunez

Vargas2, Fernando Cornejo13, Miles R. Silman1,14

Affiliations

1 Department of Biology, Wake Forest University, 1834 Wake Forest Rd, Winston

Salem, NC 27106, USA

2 Herbario Vargaz (CUZ), Escuela Profesional de Biología, Universidad Nacional de San

Antonio Abad del Cusco, Cusco, Peru

3 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109,

USA

4 Environmental Change Institute, School of Geography and the Environment, Oxford

University, South Parks Road, Oxford, OX1 3QY, UK

5 Department of Geography, University of California, Los Angeles, CA 90095, USA

6 School of Geography, University of Leeds, LS2 9JT, UK

7 Pontificia Universidad Católica del Perú, Av. Universitaria 1801, Lima, Perú

8 Universidad Nacional de San Antonio Abad del Cusco, Cusco, Peru

9 Center for Tropical Conservation, Nicholas School of the Environment, Duke Univ.,

Durham, USA

10 Science and Education, The Field Museum, 1400 S. Lake Shore Drive, Chicago, IL,

60605-2496, USA

11 Jardín Botánico de Missouri, Oxapampa, Pasco, Perú

12 Smithsonian Inst., Washington DC, USA

13 Andes to Amazon Biodiversity Program, Madre de Dios, Perú

14 Center for Energy, Environment and Sustainability, Wake Forest University, Winston-

Salem, NC 27106, USA

Correspondence to: William Farfan-Rios ([email protected], [email protected])

Abstract

Understanding how functional traits are related to species diversity and ecosystem properties is a central goal of ecology. Basic wood specific gravity (hereafter wood density) is a trait that integrates many aspects of plant form and function and is highly variable among species. While previous studies of wood density across elevational gradients have reported declines with increasing elevation, whether this is really the case in the tropics remains unclear. Here, we use one of the longest and most speciose elevational gradients in the world, extending from the Andean treeline to the Amazon basin, to test the extent to which elevation, species composition, phylogenetic affinity, and forest structure determine variation in wood density. Using field-collected wood samples and global databases we compiled 1276 taxa that were imputed to every stem across 41 (47.5 ha) mature forest plots arrayed across a 3,500m vertical gradient.

While elevation had no consistent effect on intraspecific variation in wood density, species turnover explained most of the interspecific variation across the gradient.

Mean wood density, either weighted by abundance, basal area or species were highly variable but tended to decline from low to middle elevations and increase again from mid-elevations to the treeline. As a result of this non-linearity, forests at the Andean treeline were more wood-dense than their lowland Amazon counterparts. We observed an abrupt transition in wood density at the base of persistent cloud formation, where the lowest wood density values were found. The decline of wood density is attributed to significant shift in the life form with abundance of ferns in middle elevation and the lack of stability of the landscape with higher probability of landslides and disturbances demanding for softer wood with higher elasticity. Together, both gradual compositional

26 changes and sharp local changes in the importance of non-dicot life-forms such as arborescent ferns and palms define patterns of forest-level carbon density, with implications for ecosystem properties across the Andes-to-Amazon elevational gradient.

Keywords: Wood density, elevation gradient, Andes, Amazon, species composition, functional trait

Introduction

Understanding how functional traits are related to species diversity and ecosystem properties is a central goal of ecology and is also important for biodiversity conservation, understanding forest nutrient cycles and ecosystem services, and ecosystem responses to global change (Asner et al. 2016, Neyret et al. 2016, Fyllas et al. 2017). Climate and landscape gradients are efficient laboratories for investigating the environmental controls on the ecosystem structure, function, and diversity (von Humboldt 1838, Whittaker 1967,

Malhi et al. 2010). Among natural gradients, tropical montane forests in general, and the

Amazon-Andes region in particular, are among the most diverse and complex forests globally (Silman 2014).

To address questions about changes in diversity and function of the ecosystem in the Amazon-Andes region requires the quantification of the current pattern and distribution of plant diversity and functional traits along the elevational gradient. In this study, we focus on the basic wood specific gravity (hereafter wood density) as the preferred functional trait of the ecosystem that can capture the influence of the environmental variables such as temperature, radiation, moisture, and forest architecture and mechanical characteristics impacted by variation in soil, wind speed, and disturbance along the steep slopes and elevation gradient (Sperry et al. 2008, Chave et al. 2009).

Wood density can be used as an indicator of how trees store nutrients, carbohydrates and secondary chemical compounds (Kozlowski 1992) in what is referred to as the “wood economics spectrum” (Chave et al. 2009). Understanding variations in wood density can give insights into tree life history strategies, growth rate, and the controls of climate and

28 disturbance (Putz et al. 1983, Lawton 1984, van Gelder et al. 2006, Swenson and Enquist

2007, Poorter et al. 2008, Chave et al. 2009, Adler et al. 2014).

Wood density variability and forest function

Wood density is considered a variable characterizing between species, within species, and even within individual ecological functions. Inter- and intraspecific variation in wood density is closely related to variation in diameter growth rate (Putz et al. 1983,

Muller-Landau 2004, King et al. 2005), hydraulic properties of the plant (Zanne et al.

2010), and other wood properties such as porosity, resistance and the number of vessel cells (Chave et al. 2009, Fortunel et al. 2014). At the community level, variation in wood density is often related to forest successional stage and life history trade-offs between light demanding and shade tolerant species. Fast growing light-demanding species typically have lower wood density values than shade-tolerant species (Chave et al. 2009), though the variability in the relationship between the rate of growth and wood density is high, with some fast-growing species have the highest wood density values recorded in tropical forests (e.g. Tabebuia spp.). However, the importance of within-species variation in wood density and within-species responses to environmental gradients as opposed to among-species to understand the complexity of this functional trait and its implication in ecosystem functioning remains poorly understood. Even less understood are the differences in wood density among life forms such as true trees vs. species without secondary xylem such as palms and tree ferns, and the effects that varying abundances of these life forms have on ecosystem-level attributes such as carbon storage.

The importance of within-species variation in wood density and within- and among-species responses to environmental changes remains poorly understood in most tropical ecosystems with complex structure and high species diversity. Furthermore, wood density is considered a key functional trait to follow and understand the growth– mortality trade-off (Wright et al. 2010, Kraft et al. 2010), forest dynamics and carbon cycling (Putz et al. 1983, Zanne et al. 2010). Wood density responses to environmental gradients are similarly poorly constrained, even though this is a central component of modeling carbon cycle responses to droughts and climate warming (Saatchi et al. 2011).

An accurate understanding of species- and stand-level variation of wood density is also central to reducing the uncertainties in carbon estimation when scaling from small sample plots to a landscape (Fearnside 1997, Malhi et al. 2006, Saatchi et al. 2011, Asner et al.

2014b, Phillips et al. 2019). However, it is still unclear how different ways of measuring landscape-level wood density (species level, community-level weighted by individuals or basal area) affect carbon storage calculations.

Wood density variation in tropical environmental gradients

Existing studies of tropical forests across geographic and environmental gradients suggest a great variation in wood density within and among tree communities

(Williamson 1984, Baker et al. 2004, Muller-Landau 2004, Chave et al. 2006, Fortunel et al. 2014). For example, across the Amazon basin, wood density is higher in central and eastern Amazon than northwestern Amazonia, both at species-level (Muller-Landau

2004, Chave et al. 2006) and stand-level per stem basis (Baker et al. 2004). This pattern can be explained by the disproportionate abundance and diversity of taxa with high wood

30 density values associated with poor soils found in central and eastern Amazonia (Baker et al. 2004, Muller-Landau 2004, ter Steege et al. 2006). However, there is still uncertainty with regards to whether these patterns are caused by changes in wood density within all species in a community, or just species that combine high local density with broad distributional ranges in Amazonia, because tree communities are comprised of few abundant species (i.e. oligarchs or hyperdominants) and many rare species (Pitman et al.

2013, ter Steege et al. 2013).

Though wood density is a key functional trait that links plant diversity with ecosystem function in tropical forests, our knowledge of this trait along environmental gradients in the Neotropics is incomplete, measured only at the species-level, and only on the lower range (0 - 2500 m) of an elevation gradient extending up to 4000 m (Chave et al. 2006). This leads to important questions: (1) Given that leaf canopy traits are highly variable along elevational gradients (Asner et al. 2016, Neyret et al. 2016), what is the pattern of wood density variation? and (2) how is wood density related to elevation per se, as indicated by within-species variation with respect to environmental changes, versus turnover in species, and even deeper phylogenetic conservatism?

Given the basic and applied importance of wood density as a functional trait and the paucity of knowledge concerning its variation along elevational gradients, here, we investigate the variation of wood density as a functional trait in an elevational gradient spanning ~3500 m from the Andean treeline to the Amazon basin on the eastern slope of the Peruvian Andes. To our knowledge, this is the first study in the Neotropics that assesses changes in wood density on an extensive elevational gradient using field- collected wood samples and plot-based sampling approaches including three arborescent

31 life forms (trees, palms and tree ferns). We ask (1) what is the pattern of intra- and interspecific variation in wood density across the Andes-to-Amazon gradient? (2) What is the effect of elevation on community wood density variation and distribution, and how does this pattern differ when communities are defined by species composition and stem abundance? and (3) what is the relationship between wood density and stem size across the elevation gradient?

Methods

Study site and climate

The study was performed on the eastern slope of the Peruvian Andes along an elevational gradient extending from the treeline at 3700 m to the Amazon basin at 190 m in the Manu Biosphere Reserve (11.8564° S, 71.7214° W) and Tambopata National

Reserve (12.9206° S, 69.2819° W). Mean Annual Temperature decreases linearly along the gradient with increasing elevation at a lapse rate of 5.2 o C/km, ranging from ~ 26.6 o

C at the lowest elevations to ~ 6.4 o C at treeline (Rapp and Silman 2012, Malhi et al.

2016). Mean annual precipitation varies across the gradient from 2448 to 5500 mm yr-1, with significant inter-annual variability throughout (Rapp and Silman 2012, Malhi et al.

2016). There is a distinct seasonality in rainfall, with highest rainfall in January and

February and the lowest in June and July. Winds vary little throughout the year, with the dominant pattern being upslope winds during the day and downslope winds at night.

Mean wind speeds were higher at lower elevations particularly during the austral spring

(Rapp and Silman 2012). The study area has high cloud frequency in contrast to many

32 other areas of the eastern slope of the Andes, with clouds present in all seasons. Along the elevational gradient, the cloud base zone is estimated between 1500 - 2000 m, with a highest mean annual cloud frequency between 2000 - 3500 m (Halladay et al. 2012).

Wood density calculation

We focused sampling on the dominant montane forest species because they are poorly or not represented in global databases. In our study area we have registered 908 arborescent species above 1000 m of elevation and our field wood samples comprise 34% of those species. We collected wood cores from 892 individuals representing 311 species of the dominant arborescent life forms—including trees, tree ferns (hereafter ferns) and palms—along the elevational gradient during 2009-2015. We stratified sampling of wood cores across the elevational gradient to ensure coverage of a broad range of taxa and to collect at least one individual for every species at each elevation. Core samples were collected in 51 sites ranging from 346 to 3650 m of elevation (Fig. 1). An increment borer was used to extract wood core samples for trees and palms ≥ 10 cm diameter at breast height (DBH). The DBH of the sampled individuals ranged from 10 to 85 cm and core samples were extracted between 1 to 1.3 m above the ground. For trees and palms, the wood cores were taken from the heartwood to the bark to capture the density variation. For the arborescent ferns, sliced samples were taken from the trunk-like rhizomes in six different sections and the average density value of the individual was used. Core samples were taken from individuals of targeted species outside of the permanent plots across the gradient (see below “inventory plot data” section) to avoid effects on plants that are part of long-term studies.

All values here are reported as a wood basic specific gravity, which is defined as oven-dry mass divided by its green volume (Fearnside 1997, Chave et al. 2006,

Williamson and Wiemann 2010) and henceforth called wood density in the text for simplicity. Wood density was calculated using the water displacement method with all samples oven-dried to constant mass and weighted to the nearest 0.001 g (Chave et al.

2006). Values of wood density were first calculated at oven-dry temperature at ~80 o C.

Because the possible presence of bond water in the wood samples (Williamson and

Wiemann 2010), we used a sub-sample (n = 145) to calculate wood density at 105 o C oven-dry. We developed a correction equation (105 o C WD = - 0.0113 + 0.9969 x 75 o C

WD; SI Appendix, Fig. S1) that was applied to calibrate the wood density values to the rest of the wood samples. We observed no significant difference among the wood density values at 105 o C and 80 o C (Mann-Whitney-Wilcoxon test, n = 145, p = 0.31). The overall mean difference between wood density values at 105 o C and 80 o C was 2.4 % ±

0.38 (95 % CI).

Inventory plot data

Plot data were collected from 41 (47.5 ha) permanent mature forest plots across an elevation gradient ranging from 190 to 3625 m elevation, extending from lowlands through the montane forest up to the Andean treeline. A network of 24 1-ha permanents plots were established and monitored by the Andes Biodiversity and Ecosystem Research

Group – ABERG (http://www.andesconservation.org/) ranging from 387 to 3625 m elevation. Additionally, 17 (23.5 ha) permanents plots were established by various investigators in lowland forest and are now monitored by the Amazon Forest Inventory

Network – RAINFOR (http://www.rainfor.org/) and Global Ecosystems Monitoring

Network – GEM (http://gem.tropicalforests.ox.ac.uk/) (Fig. 1). The RAINFOR plot data were extracted from the ForestPlots.net database (Lopez‐Gonzalez et al. 2009, Lopez-

Gonzalez et al. 2011). The permanent forest plots contain 31,330 stems greater than 10 cm DBH and encompass 1,950 species (of which 35 % are morphospecies). Overall, the registered species in the transect belonged to 408 genera and 111 families (sensu APG

IV) which represents 42.2% of the Peruvian tree genera (Pennington et al. 2004). To avoid variation in the data caused by possible long-term directional changes in forest structure and species composition, we used the most recent censuses (2013-14) for all the plots to estimate mean community wood density.

Botanical identification

All botanical vouchers taken with the wood core collections were identified, and then compared and standardized with the permanent forest plots vouchers that were deposited in the Peruvian and USA herbaria (CUZ, HUT, MOL, USM and DAV, MO, F,

WFU respectively). Additionally, local flora and plant checklists were used as references

(Cano et al. 1995, Pennington et al. 2004, Farfan-Rios et al. 2015, Vasquez M. and Rojas

G. 2016) and plant identifications were also confirmed by taxonomic experts. The APG

IV classification (Chase et al. 2016) was followed for the taxonomy names and

Taxonomic Name Resolution Service (TRNS) online application was used for scientific plant names standardization (Boyle et al. 2013).

Data analysis

We analyzed wood density interspecific variation against elevation using each individual of a given species sampled in the field. A restricted maximum likelihood

(REML) analysis was used to test the inter-specific variance of wood density across phylogenetic levels along the gradient (Messier et al. 2010). Variance partitioning analysis was done using the lme and varcomp functions in R where a generalized linear model was fitted to the variance across four scales nested levels: Species, genus, family, and plot. Variance partitioning allowed us to test the role of phylogeny and plot-to-plot variability including elevation. To test the effect of elevation on intra-specific variation in wood density, we used a subset of the field collected samples. We used 46 species with ≥ five individuals and that were present at least in two research sites along the gradient. We then calculated the slopes of the linear regression models for each of the selected species to observe the distribution of slopes and assess the positive, negative, or non-relationship with elevation.

To analyze wood density variation across the elevational gradient at plot-level (n

= 41, 47.5 hectares), we calculated an average species wood density value derived from the wood samples collected in the field (311 species, 892 individuals) and those values were assigned to each stem of a given species in the plot network across the transect. For stems with no measured density values from the transect we incorporated wood density values from the Global Wood Density Data Base (Zanne et al. 2009) and ForestPlots.net network (Lopez‐Gonzalez et al. 2009, Lopez-Gonzalez et al. 2011). Overall, we compiled

1,276 forest taxa from field-collected samples and published resources (SI Appendix,

Table S1). When density values were not available from the combined datasets of field-

36 published resources at the species-level, the mean values at the genus or family level was used. This was the case for the unidentified individuals to a species level (morphospecies) that accounted for 13 % of the total individuals. For the unknown taxa (0.8 % of all taxa), the local plot-level mean value was used. We then calculated the mean wood density of each plot in two ways. First, we calculated the average wood density across all species present in each plot (species mean WD) and then we calculated the mean wood density by weighting each species by its number of stems (stem weighted WD). In addition, species mean WD was also weighed by basal area. We ran the analysis for all arborescent life forms (i.e. trees, ferns and palms) and for trees only, in all the cases we excluded the lianas from the analysis. The outcome of this analysis indicates the influence of the arboreal life forms, the number and size of stems and the species composition turnover on plot-level wood density variation along the elevation gradient.

To allow biogeographical comparations of wood density along the elevational gradient, these plots were divided into five different forests types (Young 1992,

Pennington et al. 2004): Lowland (500 m ≤; including terra firme, floodplain and bamboo dominated), sub montane (500 - 1500 m), lower montane (1500 - 2500 m), upper montane (2500 - 3400 m) and treeline (≥ 3400 m). Finally, wood density variation was calculated across diameter classes for comparison of forest structure across forest types.

We used ordinary least squares linear regression to explore the intra-and inter-specific relationships between wood density and elevation and the smoothing function of generalized additive model (GAM) to fit response curves and to test the relationship of wood density and elevation if a non-linear relationship was observed.

Results

Across the entire elevational gradient, species mean wood density (WD) for all arborescent life forms was 0.578 g cm-3 ± 0.004 (95 % CI). Focusing on single life forms, dicot tree species mean wood density was 0.587 g cm-3 ± 0.004 (95 % CI), palm species mean wood density was 0.410 g cm-3 ± 0.026 (95 % CI), and arborescent fern species mean wood density was 0.351 g cm-3 ± 0.003 (95 % CI). The maximum wood density value was 1.120 g cm-3 for Machaerium acutifolium (Fabaceae) in the sub montane forest, and the minimum value was 0.111 g cm-3 for Erythrina ulei (Fabaceae) in lowland forest. Stem weighted plot-level mean WD for all life forms was 0.547 g cm-3 ± 0.002 (95

% CI), and the mean for only trees, palms and ferns was 0.584 g cm-3 ± 0.001 (95 % CI),

0.347 g cm-3 ± 0.004 (95 % CI) and 0.349 g cm-3 ± 0.008 (95 % CI) respectively. Across all elevations, the overall distribution of species and stem weighed WD for all arborescent life forms and for trees alone was symmetric and normal, but with a slight positive skewness and kurtosis for species WD and negatively skewed for stem weighed

WD (SI Appendix, Fig. S2).

Inter- and intra-specific variation of wood density along elevation

Variance partitioning showed that evolutionary relatedness explained most of the variance in wood density for both species sampled in the field (69.4 % of the variation) and the plot-level including the 41 forests plots (99.7 % of the variation) across the gradient (SI

Appendix, Fig. S3). The differences between families accounted for the largest proportion of the total variation for the field sampled species (28.5 %) and at the plot-level the largest variation was between genera (47.1 %; SI Appendix, Fig. S3).

With regards to the relationship of intra-specific variation in wood density with elevation, 83% of the sampled species (n = 46, ≥ 5 individuals) showed no relationship with elevation and only eight species showed a significant response (Fig. 2a). We found that the modal slope from the regressions was essentially zero (n = 46, 푥̅ = 0.0003 ± 0.0001

SE), with a slight bias towards positive regression slopes, as compared to negative slopes, with only eight slopes significantly different from zero (Fig. 2a). However, we found large variation between tree species in both the sign and strength of the relationship. For example, wood density in Clethra cuneata shows a highly significant decrease with

2 increasing elevation (n = 45, F1,43 = 18.44, adj. R = 0.28, p < 0.0001; Fig. 2b), while

2 Morella pubescens (n = 20, F1,18 = 2.90, adj. R = 0.09, p = 0.11; Fig. 2c) and Weinmannia

2 bangii (n = 26, F1,24 = 0.0004, adj. R = -0.04, p = 0.98; Fig. 2d) have no relationship with

2 elevation. The wood density of Alnus acuminata (n = 17, F1,15 = 7.75, adj. R = 0.30, p =

0.013; Fig. 2e) and Weinmannia fagaroides both increase significantly with increasing

2 elevation (n = 28, F1,26 = 10.55, adj. R = 0.26, p = 0.003; Fig. 2f).

Plot-level wood density variation along the elevational gradient

Across the Andes-to-Amazon gradient, plot-to-plot mean wood density showed a non-linear relationship with elevation (Fig. 3, Table 1, SI Appendix, Fig. S4). Mean WD for all arborescent species decreased slightly from 190 m to 1500 m, remained constant until 2500 m and subsequently increased linearly up to the treeline (Fig. 3a). This trend changes when considering only tree species, with wood density decreasing from 190 m to

1500 m and linearly increasing above cloud base up the treeline; palms and ferns species mean wood density did not show a relationship with elevation (Fig. 3b; SI Appendix, Fig

S5). Stem weighted WD for all arborescent life forms remains constant until 2000 m, declines abruptly between 2250 m and 2500 m and then increases with elevation (Fig. 3c).

This trend changes for trees only, with stem weighted WD slightly declining until the cloud base and then increasing up to the treeline (Fig. 3d). Palm and fern stem weighted WD were not related to elevation (Fig. 3d; SI Appendix, Fig. S5). The influence of life forms on plot-level mean wood density variability was marked at middle elevations by arborescent ferns and by palms in lowlands sites (Fig. 3, Table 1). When mean species level wood density was weighted by basal area, weighted species mean WD for all arborescent life forms showed lower values and the weighted WD for trees only showed a stronger non-linear relationship of wood density with elevation (Fig. 4a, b), with the lowest values recorded in the sub-montane forest (Fig. 4b). The non-linear relationship of wood density and elevation remained constant for both genus and family level. Mean species- level wood densities obtained using genus- and family-level identification were highly correlated with those using species-level data (p < 0.0001, r = 0.94; p < 0.0001, r = 0.89 respectively; SI Appendix, Fig. S6).

We observed that plot-to-plot wood density distributions and their statistical moments varied across elevation for both species and stem weighted WD (Fig. 5). The species WD skewness did not show a clear pattern from low to high elevations. However, the skew was more pronounced at the middle and low elevations for stem weighted WD, indicating a clear influence of the abundance of low wood density taxa (Fig. 5; SI

Appendix, Table S2). Mean wood density differs significantly with forest types at species- and stem-level, but the difference for stem-weighted WD for all arborescent life forms was not significant (Kruskal-Wallis, n = 7; species WD for all arborescent life forms, x2 = 16.19,

40 p = 0.013; for trees, x2 = 15.35, p = 0.018; stem-weighted WD for all arborescent life forms, x2 = 11.62, p = 0.071; for trees, x2 = 15.67, p = 0.016; Fig. 6). For all taxa, species mean

WD across forest types was lower at low elevations, reaching a minimum below cloud base in sub montane forest and increasing through treeline for all life forms and only trees (Fig.

6a, b). Stem weighted WD remains constant towards lower montane forest and increases only in the upper and treeline forest. When growth form was restricted to just trees, the pattern shifts, with a distinct drop in wood density in sub montane forest, and wood density exceeding its lowland values only in upper-montane and treeline forests (Fig. 6c, d).

Wood density and forest structure

Although the relationship between wood density and DBH class varies greatly among forest types along the gradient, we observed a general tendency where mean wood density decreases with DBH across forest types, and that trend was significant in the sub

2 montane (n = 5, F1,3 = 22.46, adj. R = 0.84, p = 0.018) and lower montane (n = 5, F1,3 =

12.69, adj. R2 = 0.74, p = 0.037) forests (Fig. 7a). Only in the bamboo dominated forest did wood density increase across DBH classes (Fig. 7a). The relationship of mean wood density and DBH classes follow different patterns across the elevation and between the forest’s plots with marked variability for big trees over 50 cm DBH in lowlands (floodplain and terra firme forests) and lower montane forests plots (Fig. 7b). An independent dataset for field-sampled individuals (≥ 10 cm DBH) outside of the permanent plots showed a very weak trend of wood density also decreasing with increasing stem size, though 99.5% of the

2 variation in wood density was not explained by stem size (n = 631, F1,629 = 4.25, adj. R =

0.005, p = 0.039; SI Appendix, Fig. S7).

Discussion

Mean wood density substantially changes from lowlands to montane environments across the Andes-to-Amazon elevational gradient. Plot-level mean wood density showed a clear non-linear relationship with elevation and that pattern is explained by changes in species composition rather than elevation. Even though we observed a non- linear relationship between mean wood density and elevation, mean wood density increases with increasing elevation. We observed that most of the wood density variation was a result of among-species differences rather than within species variability across the elevational gradient. Moreover, the abundance and distribution of arborescent ferns and palms had a remarkable effect on mean wood density values along the gradient, decreasing wood density by 21% in montane forest dominated by ferns and by 16 % in lowland forests dominated by palms. To better understand community-level functional traits and their ecological importance, species lists from plots by themselves are not enough and it needs to be combined with the number of individuals and stem size (basal area).

Species composition drives the non-linear relationship of increasing wood density with elevation

The present study provides a new framework to understand how wood density varies at species- and stem-level across forest types and along a broad elevational gradient. Contrasting with the current results, the sole previous study suggested that wood density significantly decreased with increasing elevation (Chave et al. 2006). In light of the current results, the discrepancy could be because we evaluated the variation of wood

42 density across the entire Andes-to-Amazon gradient that is expanded from 190 to 3650 m of elevation; this gradient harbors a continuous forest with marked changes in species composition and forest structure. Although species mean wood density declines with increasing elevation up to ~1500 m, it increases above that up to the treeline (Fig. 3, 6).

The non-linear relationship between wood density and elevation (Fig. 3) is the result of species composition turnover and environment filtering of life history based on wood density or traits associated with wood density rather than a physiological response

(intraspecific variation) to the elevation gradient, which is reflected in the decrease of the distributional variance with increasing elevation at species- and stem-level (SI Appendix,

Table S2). The variation in mean species WD resides predominately at genus- and family-level, and that variation principally occurs between rather than within genera and families (SI Appendix, Fig. S3), indicating that wood density is highly conserved phylogenetically (e.g. Chave et al. 2006; Swenson & Enquist 2007). The same phylogenetic pattern has previously been found for leaf mass per area measured from forest canopies across an Andean elevational gradient (Neyret et al. 2016) as well as a large suite of leaf functional traits (Asner et al. 2014c), demonstrating that a wide range of plant functional traits are evolutionarily conserved. While we know this is true for individual traits, understanding the correlated suites of traits—the covariances among them—would give information about the major axes of variation or syndromes of functional traits, if they do in fact exist (Díaz et al. 2016, Asner et al. 2016). This is important in understanding the effects of environmental filtering and lineage sorting in shaping functional traits patterns across environmental gradients, and raises questions about the relative influences of historical (e.g. Andean uplift) and ecological forces in

43 shaping functional traits (e.g. wood density) variation in tropical forests (Chave et al.

2006).

The gradual decrease in mean wood density from lowland to sub montane forest and then an increase up to the treeline (Fig. 3, 6; Table 1; SI Appendix, Fig. S4) can also be explained by the trade-off between low tree growth and high wood density (Poorter et al. 2008, Díaz et al. 2016), which is supported by the observed decline in average tree growth with increasing elevation in eastern Andean forests (Rapp et al. 2012). The marked changes in species composition across the Andes-to-Amazon elevational gradient

(Gentry 1988) cause the decline of tree growth with increasing elevation (Rapp et al.

2012) and is also responsible for wood density increase with increasing elevation observed in this study. However, environmental factors (e.g. temperature) can also contribute to the negative correlation between wood density and tree growth. Low wood density species grow fast because they require less resources to construct wood at a given volume of stem with at lower cost, and also because they can store and move more water through the large vessels cells, parenchyma and fiber lumen (Poorter et al. 2008, Zanne et al. 2010). This highlight the importance of wood density in understanding resource investment in plant growth, and its central role in the growth-survival trade-off (Poorter et al. 2008, Díaz et al. 2016) along environmental gradients.

Wood properties could also explain the increase of wood density with elevation

While wood density is taken as a comprehensive functional trait, wood has many functions and properties and only some of them are correlated to density. For instance, non-lumen tissue such as vessel walls, fibers, and parenchyma only explain 15 % of the

44 variation in wood density, while vessel lumen fraction is unrelated to wood density

(Zanne et al. 2010), and what selective forces lead vessel and fiber trait variation remain unclear. Colder environments are potentially dominated by taxa that contain small vessels and tracheids that probably evolved before climate occupancy (Tyree and Zimmermann

2002, Zanne et al. 2014). Forests at treeline are exposed to temperatures below 0 oC and can reach ≤ -5 oC in the austral dry season (June) exposing plants to freezing conditions

(Rapp and Silman 2012). Thus, it is expected that these taxa will contain numerous, but short vessels with narrow diameters (Wheeler et al. 2007) and thick-walled fibers and vessels (Chave et al. 2009) explaining the high wood density values at higher elevations.

In addition, wood density has shown evolutionary correlation with other plant traits. For instance, wood density decreases with increasing leaf size but was found generally unrelated to other functional traits such as seed size, fruit size and plant height (Wright et al. 2007). However, there are mixed findings of wood density and leaf mass per area

(LMA) relationship, showing either a positive relationship (Ishida et al. 2008) or none at all (Wright et al. 2007). Along a tropical elevational gradient, LMA increases linearly with increasing elevation (Asner et al. 2016) suggesting a positive relationship of wood density and LMA in our Manu-Tambopata elevational transect, though this remains untested. The complex relationships between wood and leaf function remain unclear but are important to understanding the leaf-wood construction cost in plant growth spectrum between conservative and acquisitive species.

Dominant taxa and life forms influence wood density variation across the gradient

The role of dominant taxa and arborescent life forms markedly impact the observed patterns of plot stem weighted WD variation across the gradient (Fig. 3, 5, 6).

Trends of species and stem weighted WD with elevation for all arborescent life forms was highly nonlinear (Deviance explained = 57.5 %, for species and 36.4 % for stem weighted WD; Fig. 3a, c, S5), and this relationship is even stronger for stem weighted

WD when palms and ferns are excluded (Deviance explained = 39.7 %; Fig. 3d; SI

Appendix, Fig. S5). The non-linear but positive relationship of wood density and elevation can be explained by the increase in the dominance of heavy woody species at higher elevations (sensu Slik et al. 2010). For example, the highest mean wood density value was registered in our highest elevation forest plot (mean wood density 0.653 g cm-

3; APK-01, 3625m) in which tree communities are dominated by hard wooded species

[e.g. Gynoxys nitida (wood density 0.620 g cm-3), Miconia alpina (0.740 g cm-3),

Polylepis sericea (0.732 g cm-3)].

Forests at the treeline are typically dominated by taxa with higher wood density values than their lowlands counterparts, with low values at middle elevations (figs. 3, 5,

6). For instance, the ten most dominant species at our highest plot in the treeline forest

[e.g. Miconia setulosa (0.658 g cm-3), Weinmannia fagaroides (0.571 g cm-3), Clethra cuneata (0.526 g cm-3), Gynoxys nitida (0.620 g cm-3) and Miconia alpina (0.740 g cm-3)] account for 17 % of the total species but hold 70 % of the total stems. This indicates that the highest values of stem weighted WD at higher elevations (figs. 3 c, d), are driven by a few dominant hardwood species. Contrasting with the montane pattern, in our

Amazonian floodplain forest the ten most dominant species [e.g. Iriartea deltoidea (0.265

46 g cm-3), Pseudolmedia laevis (0.618 g cm-3), Quararibea wittii (0.532 g cm-3),

Astrocaryum chonta (0.508 g cm-3) and Otoba parvifolia (0.426 g cm-3)] even though accounting for only 2 % of the total species, hold 38 % of the total stems. If we exclude these dominant species, mean wood density values of the floodplain forests increase from

0.522 to 0.557 g cm-3, although that difference is not significant (Mann-Whitney-

Wilcoxon test, n = 6, p = 0.132). This suggests that the abundance of these few species could explain the lower mean stem weighted WD found in floodplain forests and in general in lowland forest. The plot-level variability in mean species and stem weighted

WD in the Amazonian forest (Fig. 3, 6) can also be associated to the difference in the geomorphology between the Holocene (floodplain) and Pleistocene (terra firme) sediments (Phillips et al. 2019).

While general trends in mean wood density values were clear across the elevation gradient, there is substantial variability within any elevation and much of this variation can be accounted for by the abundances of different arborescent life forms. By far, dicot trees were the most dominant group along the gradient, however, arborescent ferns (i.e.

Cyatheaceae and Dicksoniaceae) and palms (i.e. Arecaceae) had profound effects on mean plot wood density variation in lower montane and lowland forests (Figs. 3, 5; SI

Appendix, Fig. S5). For example, the abrupt decline of mean stem-weighted WD in the sub montane and lower montane forests (Figs. 3, 5, 6) in our Manu-Tambopata elevational transect is explained by the high abundance of ferns (mean wood density 0.35 g cm-3) at middle elevations. Excluding the dominance of arborescent ferns in the montane forest, mean stem-weighted WD significantly increases from 0.542 to 0.585 g cm-3 (Mann-Whitney-Wilcoxon test, n = 13, p = 0.04) and that difference is clearly

47 observed for example in TRU-05 and TRU-06 plots, where the high abundance of ferns encompass 52 % and 48 % of the total stems at each plot respectively, with a difference of 24 % (for TRU-05) and 20 % (for TRU-06) in stem weighted WD (Table 1). The same fern abundance patterns at plot-level were found at middle elevations in the Costa Rica elevational transect (Lieberman et al. 1996) indicating that the high abundance of this functional group has important effects on forest structure that translates to the carbon cycle in Neotropical montane forests.

Life forms had also a large influence in community wood density variation within lowland forests. Lowland forests are dominated by palms (Arecaceae) with a remarkable hyperdominance of Iriatea deltoidea (Pitman et al. 2001, ter Steege et al. 2013).

However, along the elevational transect, Iriatea deltoidea was not only dominant in lowland forests, but also in sub-montane forests up to ~1300 m, in particular in the SAI-

02 (1250 m) plot. These new results show an expansion in the hyperdominance of Iriatea deltoidea to higher elevations in the Andean forests. The dominance of Arecaceae was observed to be strongly associated with wood density variation in lowland forests (Fig. 3,

6). Thus, if palms are excluded, mean stem-weighted WD is significantly higher in the

Amazonian forests, increasing from 0.538 to 0.581 g cm-3 (Mann-Whitney-Wilcoxon test, n = 19, p < 0.001). This highlight the importance of including life forms (i.e. ferns and palms) in wood density calculations at local- and regional-scales to understand the causes of its variation and its effects in forest structure and function. Overall, although these patterns of dominant taxa and the abundance of arborescence life forms are observed in the Manu-Tambopata elevational transect, the floristic composition and the distribution of neotropical tree communities (Gentry 1988, Pitman et al. 2001, Arellano et al. 2014,

Gomes et al. 2018) suggests that the regional-scale patterns in wood density variation across an elevational transect derived from the current study will prove to be spatially consistent across neotropical forests.

Lowest wood density values in sub montane forest

We observed that arborescent life forms have profound effects in the wood density variation in lower montane to lowland forests across the elevation gradient.

However, if non-tree life forms are excluded, plot-to-plot wood density variation in true trees (those with secondary xylem) still follows the non-linear relationship along the gradient and does not fully explain why wood density is lower around the cloud base

(Fig. 3, 6; SI Appendix, Fig. S5). Forest dynamics in tropical mountains are highly influenced by natural disturbances with significant effects on the forest structure, diversity, and function (Crausbay and Martin 2016). Landslides and tree gaps may be the primary driving forces to vegetation turnover and changes in tree species composition in

Andean tropical mountains. Accordingly, a plausible explanation for the consistent trends of lower values of mean wood density at middle elevations around the cloud base may be due to the effects of landslide occurrence. It has been shown that high landslide probability occurs around 1500 m of elevation occurrence (Clark et al. 2015) where the cloud base —below the cloud immersion zone— is estimated to be in our study area

(Halladay et al. 2012). These results could explain the lowest mean species WD values found around the cloud base (Fig. 8). The size, intensity, and recurrence of these landslides around the cloud base may lead to a high tree species turnover, facilitating the establishment of fast-growing species with low wood density [e.g. Urera caracasana

(mean wood density = 0.180 g cm-3), Heliocarpus americanus (0.215 g cm-3), and

Jacaratia digitata (0.223 g cm-3)] and resulting in a highly dynamic and heterogeneous forests with high abundances of low wood density species (e.g. Tachigali setifera, 0.367 g cm -3 in SPD-02 plot; Fig. 5). Even large size trees have significantly lower wood density values in the sub and lower montane forests than their lowlands and uplands counterparts (Fig. 7). The suggested relationship of the lowest mean community wood density around the cloud base with the high landslide may be important to understand how tree communities are assembled and what determines their abundances and distribution ranges.

Wood density relationship with forest dynamics and ecosystem function

The documented association of wood density with the vital demographic rates

(e.g. tree mortality) is reflected in the low wood density-high mortality trade-off (Muller-

Landau 2004, Kraft et al. 2010). Along the gradient, high tree mortality was reported around the cloud base (Farfan Rios 2011) in which the lowest tree species- and tree stem- level wood density values were found in this study. The high probability of landslide recurrence around the cloud base could lead to the tree community reshuffle with the dominance of low wood density species that are more probable to die, according to the hypothesis that wood density can predict tree mortality in tropical forests (Chao et al.

2008). If low wood density and high conductance are advantageous in the cloud base zone, then one might expect these communities to be especially vulnerable to drought.

This is consistent with the hypothesis that trees with high wood density are more capable to resist drought-induced embolism than trees with low wood density (Hacke et al. 2001,

Zanne et al. 2010). The relationship between high wood density and low mortality risk can be explained by the positive relationship between high wood density and thick-walled vessels that protects xylem from implosion when water shortages create strongly negative water potentials (Hacke et al. 2001), although we have no physiological data to confirm this. However, we can hypothesize that arboreal species at the treeline are better able to resist drought-induced embolism than mid-elevation and lowlands species.

Wood density has been suggested to play a key role for understanding ecosystem properties such as carbon cycling (Zanne et al. 2010) and is considered one of the six functional traits that bridges tree diversity with ecosystem function (Díaz et al. 2016). In order to improve aboveground carbon estimations in tropical forests, allometric equations now include wood density values to reduce uncertainties because it is critical to capture spatial patterns of carbon dynamics at local- and regional-scales (Baker et al. 2004, Malhi et al. 2006, Phillips et al. 2019). To reduce uncertainties in aboveground forest biomass calculation, collecting wood density data of 10 % of individuals can provide a good estimate of plot-level wood density variation for carbon estimation from local- to regional-scales (Chelsea et al. in prep.). Here, we provide a wood density database to reduce uncertainties in carbon calculations in particular for Andean montane forests, where have been observed that biomass decline with elevation (Malhi et al. 2016).

Current aboveground biomass estimates across the Amazonian forests includes palm communities in their calculations (Malhi et al. 2006, Asner et al. 2014a). However, even though tree ferns abundance exceeds the 50 % of stems per hectare in some elevations, they are not included in global tropical forests carbon estimates (Saatchi et al. 2011)

51 excluding their contributions to tropical forests carbon stocks. This reinforces the importance of including all arborescent life forms in global forest biomass calculations.

This study provides insights about how wood density variation among functional groups, stem size and habitats could improve carbon dynamic calculations in tropical forests. The sizes of trees (i.e. stem diameter) are important in biomass calculations, in particular, large rainforest trees, that accounts for 2 % of the stems, but store up to 40 % of aboveground biomass per hectare (Clark and Clark 1996, ABERG PlotData 2019). In this study, when wood density was weighted by basal area, the values were lower than the unweighted wood density, suggesting the dominance of species with low wood density in large sizes trees (Fig. 4). We also observed that large size trees across the gradient tend to have lower mean wood density values that small size classes in all forests types except for the bamboo dominated forest (Fig. 7; SI Appendix, Fig. S7). Large trees with the lowest wood density values were coincidentally found at the sub montane forest, where the lowest community wood density was reported here. This negative correlation between wood density and stem diameter was also found in Thai (Sungpalee et al. 2009) and Panamanian tropical forests (Chave et al. 2004), however, the causality of this trend has yet to be resolved.

Data compilations of basic wood specific gravity for a wider range of taxa is still required for a better understanding of species- and stand-level wood density variation across environmental gradients. In our dataset, we still need to identify 35 % of the taxa at species-level. However, as more than 65 % are identified species, it gives us insights of wood density variation across the Andes-to-Amazon forests, and contribute to understanding the effects of wood density (species composition) on ecosystem function

52 in particular when projecting future patterns of carbon dynamics based on projected climate changes.

Acknowledgments

This paper is a product of the Andes Biodiversity and Ecosystem Research Group

(ABERG; http://www.andesconservation.org/) with contributions for lowland plots data from John Terborgh, Percy Nunez and affiliated networks RAINFOR, GEM, and the

ForestPlots.net data management utility for permanent plots. Data included in this study is the result of an extraordinary effort by a large team in Peru specially from the

Universidad Nacional de San Antonio Abad de Cusco. Special thanks go to Luis Imunda and Erickson Urquiaga for their assistance in the field sampling campaigns. SERNANP and personnel of Manu National Park - Peru provided assistance with logistics and permission to work in the protected area. Pantiacolla Tours and the Amazon

Conservation Association provided logistical support. Funding came from the Gordon and Betty Moore Foundation’s Andes to Amazon initiative and the US National Science

Foundation (NSF) DEB 0743666, NSF Frontiers in Earth Systems Dynamics (FESD)

1338694, and NSF Long-Term Research in Environmental Biology (LTREB) 1754647.

The research was also supported by the National Aeronautics and Space Administration

(NASA) Terrestrial Ecology Program grant # NNH08ZDA001N-TE/ 08-TE08-0037.

Support for RAINFOR and ForestPlots.net plot monitoring in Peru has come from a

European Research Council (ERC) Advanced Grant (T‐FORCES, “Tropical Forests in the Changing Earth System”, 291585), Natural Environment Research Council grants

(including NE/F005806/), NE/D005590/1, and NE/N012542/1), and the Gordon and

Betty Moore Foundation.

Literature cited

ABERG PlotData. 2019. Permanent tree plot network.

http://www.andesconservation.org/.

Adler, P. B., R. Salguero-Gómez, A. Compagnoni, J. S. Hsu, J. Ray-Mukherjee, C.

Mbeau-Ache, and M. Franco. 2014. Functional traits explain variation in plant life

history strategies. Proceedings of the National Academy of Sciences 111:740–745.

Arellano, G., V. Cala, and M. J. Macía. 2014. Niche breadth of oligarchic species in

Amazonian and Andean rain forests. Journal of Vegetation Science 25:1355–1366.

Asner, G. P., C. B. Anderson, R. E. Martin, D. E. Knapp, R. Tupayachi, F. Sinca, and Y.

Malhi. 2014a. Landscape-scale changes in forest structure and functional traits along

an Andes-to-Amazon elevation gradient. Biogeosciences 11:843–856.

Asner, G. P., D. E. Knapp, R. E. Martin, R. Tupayachi, C. B. Anderson, J. Mascaro, F.

Sinca, K. D. Chadwick, M. Higgins, W. Farfan, W. Llactayo, and M. R. Silman.

2014b. Targeted carbon conservation at national scales with high-resolution

monitoring. Proceedings of the National Academy of Sciences 111:E5016–E5022.

Asner, G. P., R. E. Martin, C. B. Anderson, K. Kryston, N. Vaughn, D. E. Knapp, L. P.

Bentley, A. Shenkin, N. Salinas, F. Sinca, R. Tupayachi, K. Quispe Huaypar, M.

Montoya Pillco, F. D. Ccori Álvarez, S. Díaz, B. Enquist, and Y. Malhi. 2016, June.

Scale dependence of canopy trait distributions along a tropical forest elevation

gradient. New Phytologist.

Asner, G. P., R. E. Martin, R. Tupayachi, C. B. Anderson, F. Sinca, L. Carranza-Jiménez,

and P. Martinez. 2014c. Amazonian functional diversity from forest canopy

chemical assembly. Proceedings of the National Academy of Sciences of the United

States of America 111:5604–9.

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. Patino, 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.

Boyle, B., N. Hopkins, Z. Lu, J. A. Raygoza Garay, D. Mozzherin, T. Rees, N. Matasci,

M. L. Narro, W. H. Piel, S. J. McKay, S. Lowry, C. Freeland, R. K. Peet, and B. J.

Enquist. 2013. The taxonomic name resolution service: an online tool for automated

standardization of plant names. BMC bioinformatics 14:16.

Cano, A., K. R. Young, B. Leon, and R. B. Foster. 1995. Composition and diversity of

flowering plants in the upper montane forest of Manu National Park, Southern Peru.

Pages 271–280 in S. P. Churchill, H. Balslev, E. Forero, and J. L. Luteyn, editors.

Biodiversity and Conservation of Neotropical Montane Forests: Proceedings of the

Neotropical Montane Forest. New York Botanical Garden Pr Dept, United states of

America.

Chao, K. J., O. L. Phillips, E. Gloor, A. Monteagudo, A. Torres-Lezama, and R. V

Martinez. 2008. Growth and wood density predict tree mortality in Amazon forests.

Journal of Ecology 96:281–292.

Chase, M. W., M. J. M. Christenhusz, M. F. Fay, J. W. Byng, W. S. Judd, D. E. Soltis, D.

J. Mabberley, A. N. Sennikov, P. S. Soltis, P. F. Stevens, B. Briggs, S. Brockington,

A. Chautems, J. C. Clark, J. Conran, E. Haston, M. Möller, M. Moore, R. Olmstead,

M. Perret, L. Skog, J. Smith, D. Tank, M. Vorontsova, and A. Weber. 2016. An

update of the Angiosperm Phylogeny Group classification for the orders and

families of flowering plants: APG IV. Botanical Journal of the Linnean Society

181:1–20.

Chave, J., R. Condit, S. Aguilar, A. Hernandez, S. Lao, and R. Perez. 2004. Error

propagation and scaling for tropical forest biomass estimates. Philosophical

Transactions of the Royal Society B: Biological Sciences 359:409–420.

Chave, J., D. Coomes, S. Jansen, S. L. Lewis, N. G. Swenson, and A. E. Zanne. 2009.

Towards a worldwide wood economics spectrum. Ecology Letters 12:351–366.

Chave, J., H. C. Muller-Landau, T. R. Baker, T. A. Easdale, H. Ter Steege, and C. O.

Webb. 2006. Regional and phylogenetic variation of wood density across 2456

neotropical tree species. Ecological Applications 16:2356–2367.

Clark, D. B., and D. A. Clark. 1996. Abundance, growth and mortality of very large trees

in neotropical lowland rain forest. Forest Ecology and Management 80:235–244.

Clark, K. E., A. J. West, R. G. Hilton, G. P. Asner, C. A. Quesada, M. R. Silman, S. S.

Saatchi, W. Farfan-Rios, R. E. Martin, A. B. Horwath, K. Halladay, M. New, and Y.

Malhi. 2015. Storm-triggered landslides in the Peruvian Andes and implications for

topography, carbon cycles, and biodiversity. Earth Surface Dynamics Discussions

3:631–688.

Crausbay, S. D., and P. H. Martin. 2016. Natural disturbance, vegetation patterns and

ecological dynamics in tropical montane forests. Journal of Tropical Ecology

32:384–403.

Díaz, S., J. Kattge, J. H. C. Cornelissen, I. J. Wright, S. Lavorel, S. Dray, B. Reu, M.

Kleyer, C. Wirth, I. Colin Prentice, E. Garnier, G. Bönisch, M. Westoby, H. Poorter,

P. B. Reich, A. T. Moles, J. Dickie, A. N. Gillison, A. E. Zanne, J. Chave, S. Joseph

Wright, S. N. Sheremet’ev, H. Jactel, C. Baraloto, B. Cerabolini, S. Pierce, B.

Shipley, D. Kirkup, F. Casanoves, J. S. Joswig, A. Günther, V. Falczuk, N. Rüger,

M. D. Mahecha, and L. D. Gorné. 2016. The global spectrum of plant form and

function. Nature 529:167–171.

Farfan-Rios, W., K. Garcia-cabrera, N. Salinas, M. N. Raurau-quisiyupanqui, and M. R.

Silman. 2015. Lista anotada de árboles y afines en los bosques montanos del sureste

peruano : la importancia de seguir recolectando. Revista Peruana de Biología

22:145–174.

Farfan Rios, W. 2011. Changes in forest dynamics along a 2.5 km elevation gradient on

the southeastern flank of the Peruvian Andes. Dissertation, Wake Forest University,

Winston Salem, North Carolina, USA.

Fearnside, P. M. 1997. Wood density for estimating forest biomass in Brazilian

Amazonia. Forest Ecology and Management 90:59–87.

Fortunel, C., J. Ruelle, J. Beauchêne, P. V. A. Fine, and C. Baraloto. 2014. Wood specific

gravity and anatomy of branches and roots in 113 Amazonian rainforest tree species

across environmental gradients. The New phytologist 202:79–94.

Fyllas, N. M., L. P. Bentley, A. Shenkin, G. P. Asner, O. K. Atkin, S. Díaz, B. J. Enquist,

W. Farfan-Rios, E. Gloor, R. Guerrieri, W. H. Huasco, Y. Ishida, R. E. Martin, P.

Meir, O. Phillips, N. Salinas, M. Silman, L. K. Weerasinghe, J. Zaragoza-Castells,

and Y. Malhi. 2017. Solar radiation and functional traits explain the decline of forest

primary productivity along a tropical elevation gradient. Ecology Letters 20:730–

740. van Gelder, H. A., L. Poorter, and F. J. Sterck. 2006. Wood mechanics, allometry, and

life-history variation in a tropical rain forest tree community. The New phytologist

171:367–78.

Gentry, A. H. 1988. Changes in Plant Community Diversity and Floristic Composition on

Environmental and Geographical Gradients. Annals of the Missouri Botanical

Garden 75:1–34.

Gomes, V. H. F., S. D. Ijff, N. Raes, I. L. Amaral, R. P. Salomão, L. D. S. Coelho, F. D.

D. A. Matos, C. V. Castilho, D. D. A. L. Filho, D. C. López, J. E. Guevara, W. E.

Magnusson, O. L. Phillips, F. Wittmann, M. D. J. V. Carim, M. P. Martins, M. V.

Irume, D. Sabatier, J.-F. Molino, O. S. Bánki, J. R. D. S. Guimarães, N. C. A.

Pitman, M. T. F. Piedade, A. M. Mendoza, B. G. Luize, E. M. Venticinque, E. M.

M. D. L. Novo, P. N. Vargas, T. S. F. Silva, A. G. Manzatto, J. Terborgh, N. F. C.

Reis, J. C. Montero, K. R. Casula, B. S. Marimon, B.-H. Marimon, E. N. H.

Coronado, T. R. Feldpausch, A. Duque, C. E. Zartman, N. C. Arboleda, T. J.

Killeen, B. Mostacedo, R. Vasquez, J. Schöngart, R. L. Assis, M. B. Medeiros, M.

F. Simon, A. Andrade, W. F. Laurance, J. L. Camargo, L. O. Demarchi, S. G. W.

Laurance, E. D. S. Farias, H. E. M. Nascimento, J. D. C. Revilla, A. Quaresma, F. R.

C. Costa, I. C. G. Vieira, B. B. L. Cintra, H. Castellanos, R. Brienen, P. R.

Stevenson, Y. Feitosa, J. F. Duivenvoorden, G. A. C. Aymard, H. F. Mogollón, N.

Targhetta, J. A. Comiskey, A. Vicentini, A. Lopes, G. Damasco, N. Dávila, R.

García-Villacorta, C. Levis, J. Schietti, P. Souza, T. Emilio, A. Alonso, D. Neill, F.

Dallmeier, L. V. Ferreira, A. Araujo-Murakami, D. Praia, D. D. Do Amaral, F. A.

Carvalho, F. C. De Souza, K. Feeley, L. Arroyo, M. P. Pansonato, R. Gribel, B.

Villa, J. C. Licona, P. V. A. Fine, C. Cerón, C. Baraloto, E. M. Jimenez, J. Stropp, J.

Engel, M. Silveira, M. C. P. Mora, P. Petronelli, P. Maas, R. Thomas-Caesar, T. W.

Henkel, D. Daly, M. R. Paredes, T. R. Baker, A. Fuentes, C. A. Peres, J. Chave, J. L.

M. Pena, K. G. Dexter, M. R. Silman, P. M. Jørgensen, T. Pennington, A. Di Fiore,

F. C. Valverde, J. F. Phillips, G. Rivas-Torres, P. Von Hildebrand, T. R. Van Andel,

A. R. Ruschel, A. Prieto, A. Rudas, B. Hoffman, C. I. A. Vela, E. M. Barbosa, E. L.

Zent, G. P. G. Gonzales, H. P. D. Doza, I. P. D. A. Miranda, J.-L. Guillaumet, L. F.

M. Pinto, L. C. D. M. Bonates, N. Silva, R. Z. Gómez, S. Zent, T. Gonzales, V. A.

Vos, Y. Malhi, A. A. Oliveira, A. Cano, B. W. Albuquerque, C. Vriesendorp, D. F.

Correa, E. V. Torre, G. Van Der Heijden, H. Ramirez-Angulo, J. F. Ramos, K. R.

Young, M. Rocha, M. T. Nascimento, M. N. U. Medina, M. Tirado, O. Wang, R.

Sierra, A. Torres-Lezama, C. Mendoza, C. Ferreira, C. Baider, D. Villarroel, H.

Balslev, I. Mesones, L. E. U. Giraldo, L. F. Casas, M. A. A. Reategui, R. Linares-

Palomino, R. Zagt, S. Cárdenas, W. Farfan-Rios, A. F. Sampaio, D. Pauletto, E. H.

V. Sandoval, F. R. Arevalo, I. Huamantupa-Chuquimaco, K. Garcia-Cabrera, L.

Hernandez, L. V. Gamarra, M. N. Alexiades, S. Pansini, W. P. Cuenca, W. Milliken,

J. Ricardo, G. Lopez-Gonzalez, E. Pos, and H. Ter Steege. 2018. Species

Distribution Modelling: Contrasting presence-only models with plot abundance data.

Scientific Reports 8.

Hacke, U. G., J. S. Sperry, W. T. Pockman, S. D. Davis, and K. A. McCulloh. 2001.

Trends in wood density and structure are linked to prevention of xylem implosion by

negative pressure. Oecologia 126:457–461.

Halladay, K., Y. Malhi, and M. New. 2012. Cloud frequency climatology at the

Andes/Amazon transition: 1. Seasonal and diurnal cycles. J. Geophys. Res.

117:D23102, doi:10.1029/2012JD017770. von Humboldt, A. 1838. Notice de Deux Tentatives d’Ascension du Chimborazo (A.

Pihan de la Forest, Paris).

Ishida, A., T. Nakano, K. Yazaki, S. Matsuki, N. Koike, D. L. Lauenstein, M. Shimizu,

and N. Yamashita. 2008. Coordination between leaf and stem traits related to leaf

carbon gain and hydraulics across 32 drought-tolerant angiosperms. Oecologia

156:193–202.

King, D. A., S. J. Davies, M. N. N. Supardi, and S. Tan. 2005. Tree growth is related to

light interception and wood density in two mixed dipterocarp forests of Malaysia.

Functional Ecology 19:445–453.

Kozlowski, T. T. 1992. Carbohydrate sources and sinks in woody plants. The Botanical

Review 58:107–222.

Kraft, N. J. B., M. R. Metz, R. S. Condit, and J. Chave. 2010. The relationship between

wood density and mortality in a global tropical forest data set. The New phytologist

188:1124–36.

Lawton, R. O. 1984. Ecological Constraints on Wood Density in a Tropical Montane

Rain Forest. American journal of botany 71:261–267.

Lieberman, D., M. Lieberman, R. Peralta, and G. S. Hartshorn. 1996. Tropical forest

structure and composition on a large-scale altitudinal gradient in Costa Rica. Journal

of Ecology 84:137–152.

Lopez-Gonzalez, G., S. L. Lewis, M. Burkitt, and O. L. Phillips. 2011. ForestPlots.net: a

web application and research tool to manage and analyse tropical forest plot data.

Journal of Vegetation Science 22:610–613.

Lopez‐Gonzalez, G., S. L. Lewis, M. Burkitt, T. R. Baker, and O. L. Phillips. 2009.

ForestPlots.net Database. Available at: www.forestplots.net. Last accessed July

2017.

Malhi, Y., C. A. J. Girardin, G. R. Goldsmith, C. E. Doughty, N. Salinas, D. B. Metcalfe,

W. Huaraca Huasco, J. E. Silva-Espejo, J. del Aguilla-Pasquell, F. Farfán

Amézquita, L. E. O. C. Aragão, R. Guerrieri, F. Y. Ishida, N. H. A. Bahar, W.

Farfan-Rios, O. L. Phillips, P. Meir, and M. Silman. 2016. The variation of

productivity and its allocation along a tropical elevation gradient: a whole carbon

budget perspective. New Phytologist.

Malhi, Y., M. Silman, N. Salinas, M. Bush, P. Meir, and S. Saatchi. 2010. Introduction:

Elevation gradients in the tropics: laboratories for ecosystem ecology and global

change research. Global Change Biology 16:3171–3175.

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. 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.

Messier, J., B. J. McGill, and M. J. Lechowicz. 2010. How do traits vary across

ecological scales? A case for trait-based ecology. Ecology Letters 13:838–848.

Muller-Landau, H. C. 2004. Interspecific and Inter-site Variation in Wood Specific

Gravity of Tropical Trees. Biotropica 36:20–32.

Neyret, M., L. P. Bentley, I. Oliveras, B. S. Marimon, B. H. Marimon-Junior, E. Almeida

de Oliveira, F. Barbosa Passos, R. Castro Ccoscco, J. dos Santos, S. Matias Reis, P.

S. Morandi, G. Rayme Paucar, A. Robles Cáceres, Y. Valdez Tejeira, Y. Yllanes

Choque, N. Salinas, A. Shenkin, G. P. Asner, S. Díaz, B. J. Enquist, and Y. Malhi.

2016. Examining variation in the leaf mass per area of dominant species across two

contrasting tropical gradients in light of community assembly. Ecology and

Evolution 6:5674–5689.

Pennington, T. D., C. Reynel, and A. Daza. 2004. Illustrated guide to the Trees of Peru.

Page (T. D. Pennington, C. Reynel, and A. Daza, Eds.). David Hunt, Sherborne, UK.

Phillips, O. L., M. J. P. Sullivan, T. R. Baker, A. Monteagudo Mendoza, P. N. Vargas,

and R. Vásquez. 2019. Species Matter: Wood Density Influences Tropical Forest

Biomass at Multiple Scales. Surveys in Geophysics:1–23.

Pitman, N. C. A., M. R. Silman, and J. W. Terborgh. 2013. Oligarchies in Amazonian

tree communities: A ten-year review. Ecography 36:114–123.

Pitman, N. C. A., J. W. Terborgh, M. R. Silman, P. Nunez, D. A. Neill, C. E. Ceron, W.

A. Palacios, and M. Aulestia. 2001. Dominance and distribution of tree species in

upper Amazonian terra firme forests. Ecology 82:2101–2117.

Poorter, L., S. J. Wright, H. Paz, D. D. Ackerly, R. Condit, G. Ibarra-Manríquez, K. E.

Harms, J. C. Licona, M. Martínez-Ramos, S. J. Mazer, H. C. Muller-Landau, M.

Peña-Claros, C. O. Webb, and I. J. Wright. 2008. Are functional traits good

predictors of demographic rates? evidence from five neotropical forests. Ecology

89:1908–1920.

Putz, F. E., P. D. Coley, K. Lu, A. Montalvo, and A. Aiello. 1983. Uprooting and

Snapping of Trees - Structural Determinants and Ecological Consequences.

Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere

13:1011–1020.

Rapp, J. M., and M. R. Silman. 2012. Diurnal, seasonal, and altitudinal trends in

microclimate across a tropical montane cloud forest. Climate Research 55:17–32.

Rapp, J. M., M. R. Silman, J. S. Clark, C. A. J. Girardin, D. Galiano, and R. Tito. 2012.

Intra- and interspecific tree growth across a long altitudinal gradient in the Peruvian

Andes. Ecology 93:2061–2072.

Saatchi, S. S., N. L. Harris, S. Brown, M. Lefsky, E. T. A. Mitchard, W. Salas, B. R.

Zutta, W. Buermann, S. L. Lewis, S. Hagen, S. Petrova, L. White, M. Silman, and

A. Morel. 2011. Benchmark map of forest carbon stocks in tropical regions across

three continents. Proceedings of the National Academy of Sciences of the United

States of America 108:9899–904.

Silman, M. R. 2014. Functional megadiversity. Proceedings of the National Academy of

Sciences of the United States of America 111:5763–4.

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. L. C. H. 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.

Sperry, J. S., F. C. Meinzer, and K. A. McCulloh. 2008. Safety and efficiency conflicts in

hydraulic architecture: scaling from tissues to trees. Plant, cell & environment

31:632–45. ter Steege, H., N. C. A. Pitman, O. L. Phillips, J. Chave, D. Sabatier, A. Duque, J. F.

Molino, M. F. Prevost, R. Spichiger, H. Castellanos, P. von Hildebrand, and R.

Vasquez. 2006. Continental-scale patterns of canopy tree composition and function

across Amazonia. Nature 443:444–447. ter Steege, H., N. C. A. Pitman, D. Sabatier, C. Baraloto, R. P. Salomão, J. E. Guevara,

O. L. Phillips, C. V Castilho, W. E. Magnusson, J.-F. Molino, A. Monteagudo, P.

Núñez Vargas, J. C. Montero, T. R. Feldpausch, E. N. H. Coronado, T. J. Killeen, B.

Mostacedo, R. Vasquez, R. L. Assis, J. Terborgh, F. Wittmann, A. Andrade, W. F.

Laurance, S. G. W. Laurance, B. S. Marimon, B.-H. Marimon, I. C. Guimarães

Vieira, I. L. Amaral, R. Brienen, H. Castellanos, D. Cárdenas López, J. F.

Duivenvoorden, H. F. Mogollón, F. D. de A. Matos, N. Dávila, R. García-Villacorta,

P. R. Stevenson Diaz, F. Costa, T. Emilio, C. Levis, J. Schietti, P. Souza, A. Alonso,

F. Dallmeier, A. J. D. Montoya, M. T. Fernandez Piedade, A. Araujo-Murakami, L.

Arroyo, R. Gribel, P. V. A. Fine, C. A. Peres, M. Toledo, G. A. Aymard C, T. R.

Baker, C. Cerón, J. Engel, T. W. Henkel, P. Maas, P. Petronelli, J. Stropp, C. E.

Zartman, D. Daly, D. Neill, M. Silveira, M. R. Paredes, J. Chave, D. de A. Lima

Filho, P. M. Jørgensen, A. Fuentes, J. Schöngart, F. Cornejo Valverde, A. Di Fiore,

E. M. Jimenez, M. C. Peñuela Mora, J. F. Phillips, G. Rivas, T. R. van Andel, P. von

Hildebrand, B. Hoffman, E. L. Zent, Y. Malhi, A. Prieto, A. Rudas, A. R. Ruschell,

N. Silva, V. Vos, S. Zent, A. A. Oliveira, A. C. Schutz, T. Gonzales, M. Trindade

Nascimento, H. Ramirez-Angulo, R. Sierra, M. Tirado, M. N. Umaña Medina, G.

van der Heijden, C. I. A. Vela, E. Vilanova Torre, C. Vriesendorp, O. Wang, K. R.

Young, C. Baider, H. Balslev, C. Ferreira, I. Mesones, A. Torres-Lezama, L. E.

Urrego Giraldo, R. Zagt, M. N. Alexiades, L. Hernandez, I. Huamantupa-

Chuquimaco, W. Milliken, W. Palacios Cuenca, D. Pauletto, E. Valderrama

Sandoval, L. Valenzuela Gamarra, K. G. Dexter, K. Feeley, G. Lopez-Gonzalez, and

M. R. Silman. 2013. Hyperdominance in the Amazonian tree flora. Science (New

York, N.Y.) 342:1243092.

Sungpalee, W., A. Itoh, M. Kanzaki, K. Sri-ngernyuang, H. Noguchi, T. Mizuno, S.

Teejuntuk, M. Hara, K. Chai-udom, T. Ohkubo, P. Sahunalu, P. Dhanmmanonda, S.

Nanami, T. Yamakura, and A. Sorn-ngai. 2009. Intra- and interspecific variation in

wood density and fine-scale spatial distribution of stand-level wood density in a

northern Thai tropical montane forest. Journal of Tropical Ecology 25:359–370.

Swenson, N. G., and B. J. Enquist. 2007. Ecological and evolutionary determinants of a

key plant functional trait: wood density and its community-wide variation across

latitude and elevation. American journal of botany 94:451–9.

Tyree, M. T., and M. H. Zimmermann. 2002. Hydraulic Architecture of Whole Plants and

Plant Performance. Pages 175–214 Springer. Springer, Berlin, Heidelberg.

Vasquez M., R., and R. D. P. Rojas G. 2016. Clave para identificar grupos de familias de

Gymnospermae y Angiospermae del Perú. Jardin Botanico de Missouri.

Wheeler, E. A., P. Baas, and S. Rodgers. 2007, January 1. Variations in dicot wood

anatomy: A global analysis based on the insidewood database. Brill.

Whittaker, R. H. 1967. Gradient analysis of vegetation. Biological Reviews 42:207–264.

Williamson, G. B. 1984. Gradients in Wood Specific Gravity of Trees. Bulletin of the

Torrey Botanical Club 111:51–55.

Williamson, G. B., and M. C. Wiemann. 2010. Measuring wood specific

gravity...correctly. American Journal of Botany 97:519–524.

Wright, I. J., D. D. Ackerly, F. Bongers, K. E. Harms, G. Ibarra-Manriquez, M. Martinez-

Ramos, S. J. Mazer, H. C. Muller-Landau, H. Paz, N. C. A. Pitman, L. Poorter, M.

R. Silman, C. F. Vriesendorp, C. O. Webb, M. Westoby, and S. J. Wright. 2007.

Relationships among ecologically important dimensions of plant trait variation in

seven neotropical forests. Annals of Botany 99:1003–1015.

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. Díaz, 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 growth–mortality trade-off in tropical trees. Ecology

91:3664–3674.

Young, K. R. 1992. Biogeography of the montane forest zone of the eastern slopes of

Peru. Memorias del Museo de Historia Natural U.N.M.S.M. 21:119–154.

Zanne, A. E., G. Lopez-Gonzalez, D. A. Coomes, J. Ilic, S. Jansen, S. L. Lewis, R. B.

Miller, N. G. Swenson, M. C. Wiemann, and J. Chave. 2009. Data from: Towards a

worldwide wood economics spectrum. Dryad Data Repository.

Zanne, A. E., D. C. Tank, W. K. Cornwell, J. M. Eastman, S. A. Smith, R. G. FitzJohn,

D. J. McGlinn, B. C. O’Meara, A. T. Moles, P. B. Reich, D. L. Royer, D. E. Soltis,

P. F. Stevens, M. Westoby, I. J. Wright, L. Aarssen, R. I. Bertin, A. Calaminus, R.

Govaerts, F. Hemmings, M. R. Leishman, J. Oleksyn, P. S. Soltis, N. G. Swenson,

L. Warman, and J. M. Beaulieu. 2014. Three keys to the radiation of angiosperms

into freezing environments. Nature 506:89–92.

Zanne, A. E., M. Westoby, D. S. Falster, D. D. Ackerly, S. R. Loarie, S. E. J. Arnold, and

D. A. Coomes. 2010. Angiosperm wood structure: Global patterns in vessel anatomy

and their relation to wood density and potential conductivity. American journal of

botany 97:207–15.

Table II - 1. Sites description and mean wood density values for 41 (47.5 hectares) forest plots across the Andes-to-Amazon

elevational gradient.

Mean wood basic specific gravity (g cm-3) Basal area (m2 ha-1) Species level Stem level Basal area level Elevation Size Plot Forest type All All All (m) (ha) life Tree All life Tree life Tree life Tree

forms Trees Palms ferns forms Trees Palms ferns forms Trees Palms ferns forms Trees Palms ferns

APK-01 3625 1 Treeline 21.2 21.2 0.654 0.654 0.653 0.653 0.655 0.655

ACJ-01 3537 1 Treeline 37.3 37.3 0.635 0.635 0.596 0.596 0.592 0.592 69

TRU-01 3450 1 Treeline 28.6 28.4 0.2 0.589 0.603 0.348 0.600 0.604 0.348 0.588 0.589 0.348

TRU-02 3250 1 Upper montane 31.2 30.2 1.0 0.567 0.591 0.357 0.600 0.611 0.370 0.598 0.605 0.373

TRU-03 3000 1 Upper montane 20.5 20.0 0.0 0.4 0.587 0.612 0.351 0.348 0.650 0.658 0.351 0.348 0.649 0.655 0.351 0.348

WAY-01 3000 1 Upper montane 34.3 34.3 0.0 0.589 0.602 0.348 0.628 0.629 0.348 0.633 0.633 0.348

ESP-01 2890 1 Upper montane 28.2 27.7 0.5 0.573 0.594 0.348 0.614 0.628 0.348 0.620 0.624 0.348

TRU-04 2750 1 Upper montane 34.7 29.9 4.7 0.562 0.594 0.353 0.561 0.618 0.350 0.594 0.633 0.349

TRU-05 2500 1 Upper montane 43.5 24.6 18.8 0.558 0.601 0.351 0.463 0.586 0.350 0.477 0.574 0.350

TRU-06 2250 1 Lower montane 37.0 24.4 12.6 0.518 0.559 0.350 0.456 0.556 0.348 0.469 0.531 0.348

TRU-07 2000 1 Lower montane 21.0 17.0 4.0 0.561 0.584 0.353 0.564 0.623 0.359 0.581 0.633 0.359

TRU-08 1800 1 Lower montane 29.6 25.4 0.1 4.2 0.562 0.577 0.437 0.354 0.547 0.599 0.437 0.358 0.565 0.599 0.437 0.358

forms Trees Palms ferns forms Trees Palms ferns forms Trees Palms ferns forms Trees Palms ferns

SPD-01 1750 1 Lower montane 36.5 33.5 3.1 0.560 0.571 0.343 0.521 0.569 0.316 0.553 0.575 0.313

CAL-01 1500 1 Sub montane 30.8 30.7 0.0 0.0 0.518 0.521 0.393 0.369 0.517 0.517 0.393 0.377 0.513 0.513 0.393 0.376

SAI-01 1500 1 Sub montane 41.2 40.4 0.7 0.2 0.552 0.559 0.363 0.362 0.553 0.559 0.414 0.353 0.541 0.544 0.412 0.351

SPD-02 1500 1 Sub montane 30.4 29.4 0.1 0.9 0.538 0.546 0.390 0.347 0.499 0.513 0.388 0.332 0.492 0.497 0.388 0.333

CAL-02 1250 1 Sub montane 37.0 36.9 0.0 0.534 0.536 0.230 0.545 0.545 0.230 0.531 0.531 0.230

70 SAI-02 1250 1 Sub montane 42.2 36.3 5.7 0.2 0.553 0.563 0.351 0.348 0.497 0.555 0.310 0.348 0.496 0.526 0.308 0.348

TON-02 1000 1 Sub montane 31.3 31.1 0.1 0.1 0.549 0.558 0.294 0.362 0.529 0.534 0.278 0.364 0.484 0.486 0.257 0.367

PAN-03 850 1 Sub montane 24.2 24.2 0.0 0.616 0.620 0.348 0.622 0.622 0.348 0.614 0.614 0.348

TON-01 800 1 Sub montane 26.5 26.3 0.0 0.2 0.594 0.601 0.388 0.362 0.586 0.589 0.388 0.369 0.590 0.591 0.388 0.370

PAN-02 595 1 Sub montane 27.7 27.7 0.0 0.606 0.608 0.265 0.577 0.577 0.265 0.548 0.548 0.265

PAN-01 425 1 Lowland (TF) 27.3 25.8 1.4 0.564 0.570 0.274 0.551 0.577 0.271 0.550 0.565 0.271

ALM-01 400 2 Lowland (TF) 31.2 26.9 4.3 0.591 0.599 0.419 0.542 0.580 0.347 0.536 0.569 0.327

MNU-08 400 2 Lowland (FP) 41.8 33.9 7.9 0.594 0.599 0.413 0.489 0.550 0.364 0.491 0.529 0.328

BAB-01 387 1 Lowland (BB) 29.3 26.9 2.4 0.0 0.574 0.582 0.332 0.348 0.521 0.563 0.335 0.348 0.505 0.524 0.289 0.348

MNU-04 358 2 Lowland (TF) 28.2 24.3 3.9 0.587 0.593 0.392 0.541 0.580 0.365 0.530 0.559 0.350

forms Trees Palms ferns forms Trees Palms ferns forms Trees Palms ferns forms Trees Palms ferns

MNU-05 347 2.25 Lowland (TF) 30.7 28.6 2.1 0.578 0.582 0.373 0.536 0.551 0.430 0.559 0.572 0.391

MNU-06 345 2.25 Lowland (TF) 32.3 27.0 5.3 0.571 0.578 0.388 0.517 0.559 0.383 0.510 0.542 0.350

MNU-03 312 2 Lowland (TF) 30.3 26.1 4.2 0.582 0.588 0.427 0.518 0.552 0.340 0.507 0.538 0.315

TAM-07 225 1 Lowland (TF) 25.5 25.2 0.4 0.617 0.619 0.539 0.602 0.602 0.588 0.586 0.585 0.609

TAM-05 220 1 Lowland (TF) 27.6 27.1 0.5 0.620 0.623 0.512 0.600 0.606 0.436 0.585 0.589 0.406

71 TAM-08 220 1 Lowland (TF) 22.2 20.3 2.0 0.603 0.606 0.512 0.583 0.614 0.360 0.599 0.626 0.328

TAM-01 205 1 Lowland (TF) 28.8 22.0 6.8 0.609 0.616 0.418 0.504 0.601 0.292 0.519 0.595 0.275

TAM-02 201 1 Lowland (TF) 28.1 22.0 6.1 0.593 0.599 0.434 0.520 0.604 0.315 0.530 0.595 0.293

TAM-06 200 1 Lowland (FP) 37.6 29.4 8.2 0.581 0.591 0.418 0.484 0.574 0.292 0.493 0.550 0.290

TAM-09 197 1 Lowland (TF) 23.3 20.3 3.0 0.610 0.614 0.455 0.567 0.619 0.298 0.571 0.614 0.283

CUZ-01 190 1 Lowland (FP) 27.2 25.5 1.7 0.558 0.567 0.403 0.502 0.518 0.358 0.550 0.565 0.333

CUZ-02 190 1 Lowland (FP) 27.9 24.0 3.9 0.571 0.579 0.405 0.529 0.585 0.315 0.534 0.571 0.305

CUZ-03 190 1 Lowland (FP) 27.7 24.5 3.2 0.588 0.595 0.432 0.551 0.593 0.327 0.566 0.598 0.314

CUZ-04 190 1 Lowland (FP) 28.1 24.4 3.6 0.585 0.593 0.405 0.576 0.607 0.409 0.572 0.602 0.372

Figure legends

Figure II - 1. Location of the wood collection sites (red) and the permanent forest plots

(green) across the Andes-to-Amazon elevational gradient.

Figure II - 2. (a) empirical distribution of within-species slopes of the linear regression between wood density and elevation (n = 46), vertical bars represent the slopes significantly different from zero (8-positives and 1-negative). Intraspecific variation in wood density with respect to elevation for (b) Clethra cuneata (n = 45), (c) Morella pubescens (n = 20), (d) Weinmannia bangii (n = 26), (e) Alnus acuminata (n = 17), and

(f) Weinmannia fagaroides (n = 28). Grey circles represent sampled individuals across elevation, black circles represent the mean wood density among species at each sampled site. Black solid line represents the linear regression fit with 95% confidence limits. Error bars depict bootstrapped 95% confidence intervals.

Figure II - 3. Plot-level mean wood density variation across 41 permanent forest plots along the Andes-to-Amazon elevational gradient for (a) species mean wood density for all arborescent life forms and (b) for trees, palms and ferns species. (c) Stem weighted mean wood density for all arborescent life forms and (d) for trees, palms and ferns. Error bars depict bootstrapped 95% confidence intervals. Solid lines are generalized additive models (GAM) fit using smoothing function with 95% confidence limits. Vertical dashed lines represent the elevation of cloud base.

Figure II - 4. (a) species mean wood density for all arborescent life forms and trees. (b) species mean wood density for all arborescent life forms and trees weighted by basal area. Open triangles represent all arborescent life forms and grey circles represent trees.

Solid and dashed lines are generalized additive models (GAM) fit using smoothing function. Vertical dashed lines represent the position of cloud base along the gradient.

Legend correspond to the same life forms for panels (a) and (b).

Figure II - 5. Wood density distribution for species (blue lines) and stems (red lines) for all plots (n = 41) across the Andes-to-Amazon elevational gradient. Data include all arborescent life forms. Vertical dashed lines indicate corresponding means.

Figure II - 6. Plot-level wood density variation across forests types including: Lowland

(Ltf = Lowland terra firme, Lfp = Lowland floodplain, Lbb = Lowland bamboo dominated forest; < 500 m), sub montane (SM: 500 - 1500 m), lower montane (LM: 1500

- 2500 m), upper montane (UP: 2500 - 3400 m) and treeline (TL: > 3400 m) for (a) species mean WD including all arborescent life forms and (b) tree species. (c) stem weighted WD for all arborescent life forms and (d) tree stems. Box plots show 25% quartile, median and 75% quartile of the distribution (horizontal lines). Forest types are defined based on Young (1992) and Pennington et al. (2004).

Figure II - 7. (a) Mean wood density variation for individual stems across DBH classes.

Regression lines were computed using the mean averaged wood density across diameter classes for each forest type. Stars represent significant relationship. (b) coefficient of variation (CV) of wood density across diameter classes, CV was calculated using the mean averaged wood density across diameter classes for each forest type. Color lines correspond to the same forest types for figures (a) and (b).

Figure II - 8. Mean wood density and landslide stability (LS) as a function of elevation along the Andes-to-Amazon transect. LS was calculated as the inverted scale of landslide

73 probability (1-landslide probability (%yr-1)) taken from Clark et al. (2015). (a) Species mean wood density includes all arborescent life forms and (b) trees only. Vertical dashed lines represent the cloud base across the gradient. For panels (a) and (b) open circles represent mean species wood density and grey triangles landslide stability.

Appendix II – Figure S1. Effect of drying temperature on basic wood specific gravity values as shown by the relationship between wood density values dried at ~80 o C and

105 o C oven temperature (n = 145). The equation y = - 0.0113 + 0.9969(x) was used to calibrate all wood density values obtained at ~80 o C oven temperature.

Appendix II – Figure S2. Overall wood density distribution along the Andes-to-Amazon elevational gradient for (a) species including all arborescent life forms and (b) tree species. Mean wood density for all stems (c) including all arborescent life forms and (d) tree stems. Solid red vertical lines indicate the means and dashed black lines indicate the medians.

Appendix II – Figure S3. Variance partitioning of wood density across phylogenetic levels and environment (plot-level) for (a) field core-sampled and (b) plot level along the

Andes-to-Amazon elevational gradient.

Appendix II – Figure S4. Sampled wood density along an Andes-to-Amazon elevational gradient. Grey open circles represent stem core samples across elevation (n = 892), black solid circles represent the mean wood density among samples at each sampled site (n =

51). Red solid line is the generalized additive model (GAM) fit using smoothing function with 95% confidence limits. Error bars depict bootstrapped 95% confidence intervals.

Appendix II – Figure S5. Plot-level mean wood density variation for trees, palms and tree ferns along the Andes-to-Amazon elevational gradient. Upper panels represent species mean wood density and lower panels represent the stem-weighted wood density. Dashed vertical lines indicate the cloud base across the gradient. Percentages indicate the contribution of each life form.

Appendix II – Figure S6. Mean plot-level wood density for (a) genus and (b) family basis in function of species wood density. Solid lines represent best-fit linear regressions, dashed lines represent the 1:1 relationship. Color ramp corresponds to plot elevations from the highest (blue) to the lowest (red) forest plots. Legend correspond to the same forest plot colors for panels (a) and (b).

Appendix II – Figure S7. Wood density as a function of stem diameter (≥ 10 cm DBH) for sampled species across the Andes-to-Amazon elevational gradient (n = 631). Red solid line represents the linear regression fit with 95% confidence limits. Error bars depict bootstrapped 95% confidence intervals.

FIGURE II - 1

(a) (b)

(e) (f)

FIGURE II - 2

(a)

(b)

(c)

(d)

FIGURE II - 3

(a) (b)

Frequency

------Wood density g cm-3------

FIGURE II - 54

(a) (b)

FIGURE II - 6

(a) (b)

FIGURE II - 7

Supporting information

Title:

Landscape-scale wood density variation across an Andes-to-Amazon elevational gradient

Authors:

William Farfan-Rios1,2, Sassan Saatchi3, Imma Oliveras4, Yadvinder Malhi4,

Chelsea M. Robinson5, Alex Nina-Quispe6, Juan A. Gibaja7, Israel Cuba7, Karina Garcia-

Cabrera7, Norma Salinas-Revilla6, Oliver L. Phillips8, John Terborgh9, Nigel Pitman10,

Rodolfo Vasquez11, Abel Monteagudo Mendoza11, Terry Erwin12, Percy Nuñez Vargas7,

Fernando Cornejo13, Miles R. Silman1,14

FIGURE II - 8

Results

Across all elevations, the overall distribution of species mean wood density for all arborescent life forms and for trees alone was symmetric and normal with a slight positive skewness and kurtosis (All life forms: Skewness = 0.07, kurtosis = 2.96; trees:

Skewness = 0.09, kurtosis = 3.1). Likewise, the distribution of stem weighted mean wood density for all arborescent life forms and for trees alone were symmetric and normal, but was negatively skewed with positive kurtosis for both, all arborescent (skewness = -0.18, kurtosis = 2.70) and trees stems (skewness = -0.13, kurtosis = 3.53) (SI Appendix, Fig.

S2).

Plot-to-plot wood density distributions and their statistical moments varied across elevation

(fig. 5). Species WD distributions shifted from negative skew in the lowlands to positive skew in the sub montane forest and shifting again to negative skew in the montane to treeline forests plots, with this pattern being similar for all arborescent life forms and trees only (Fig. 5; SI Appendix, Table S2). The montane and lowlands plots showed a negative kurtosis, but we observed a positive kurtosis only in the sub montane forests plots (Fig. 5;

SI Appendix, Table S2). For stem WD distributions we observed a markedly increase in the right skew (right tail increasing) in plots with high abundance of few species with low wood density values in the montane (e. g. TRU-05, TRU-06) and lowland forests plots (e.g.

TAM-01, MNU-08; SI Appendix, Table S2). Kurtosis for stem WD distributions for all arborescent life forms and only trees varies along the gradient, with the predominance of negative kurtosis in the lowland forests (Fig. 5; SI Appendix, Table S2).

Appendix II - Table S1. Species list and basic wood specific gravity values along an

Andean-to-Amazonian elevational gradient for 1,276 forest taxa compiled from field- collected samples and published resources.

Family Species Basic wood specific gravity (g cm-3) Achariaceae Lindackeria paludosa 0.56 Achariaceae Mayna parvifolia 0.56 Actinidiaceae Saurauia glabra 0.552 Actinidiaceae Saurauia peruviana 0.552 Adoxaceae Viburnum ayavacense 0.556 Adoxaceae Viburnum hallii 0.556 Adoxaceae Viburnum reticulatum 0.556 Alzateaceae Alzatea verticillata 0.735 Anacardiaceae Astronium graveolens 0.847 Anacardiaceae Astronium lecointei 0.79 Anacardiaceae Spondias mombin 0.428 Anacardiaceae Spondias venosa 0.385 Anacardiaceae Tapirira guianensis 0.466 Anacardiaceae Tapirira obtusa 0.451 Anacardiaceae Tapirira peckoltiana 0.375 Anacardiaceae Tapirira retusa 0.375 Anacardiaceae Thyrsodium spruceanum 0.54 Annonaceae Annona ambotay 0.499 Annonaceae Annona edulis 0.535 Annonaceae Annona excellens 0.535 Annonaceae Annona foetida 0.572 Annonaceae Annona fosteri 0.499 Annonaceae Annona montana 0.499 Annonaceae Annona papilionella 0.499 Annonaceae Annona williamsii 0.535 Annonaceae Duguetia flagellaris 0.747 Annonaceae Duguetia lucida 0.747 Annonaceae Duguetia odorata 0.747 Annonaceae Duguetia quitarensis 0.8 Annonaceae Duguetia spixiana 0.58 Annonaceae Guatteria acutissima 0.333 Annonaceae Guatteria alutacea 0.576

Family Species Basic wood specific gravity (g cm-3) Annonaceae Guatteria boliviana 0.576 Annonaceae Guatteria brevicuspis 0.576 Annonaceae Guatteria buchtienii 0.576 Annonaceae Guatteria citriodora 0.576 Annonaceae Guatteria dielsiana 0.482 Annonaceae Guatteria duodecima 0.431 Annonaceae Guatteria elata 0.576 Annonaceae Guatteria glauca 0.495 Annonaceae Guatteria guentheri 0.576 Annonaceae Guatteria hyposericea 0.576 Annonaceae Guatteria oblongifolia 0.557 Annonaceae Guatteria pteropus 0.576 Annonaceae Guatteria recurvisepala 0.576 Annonaceae Guatteria schomburgkiana 0.6 Annonaceae Guatteria scytophylla 0.515 Annonaceae Guatteria terminalis 0.619 Annonaceae Guatteria tomentosa 0.431 Annonaceae Guatteria ucayalina 0.504 Annonaceae Klarobelia candida 0.589 Annonaceae Klarobelia napoensis 0.589 Annonaceae Malmea diclina 0.39 Annonaceae Malmea dielsiana 0.39 Annonaceae Malmea lucida 0.39 Annonaceae Onychopetalum krukoffii 0.637 Annonaceae Oxandra acuminata 0.665 Annonaceae Oxandra espintana 0.63 Annonaceae Oxandra mediocris 0.737 Annonaceae Oxandra polyantha 0.737 Annonaceae Oxandra riedeliana 0.77 Annonaceae Oxandra xylopioides 0.77 Annonaceae Porcelia nitidifolia 0.4 Annonaceae Pseudomalmea diclina 0.39 Annonaceae Pseudomalmea diclina 0.808 Annonaceae Rollinia andicola 0.37 Annonaceae Rollinia centrantha 0.338 Annonaceae Rollinia cuspidata 0.37 Annonaceae Rollinia edulis 0.338 Annonaceae Rollinia fosteri 0.338 Annonaceae Rollinia pittieri 0.27 Annonaceae Rollinia williamsii 0.37

Family Species Basic wood specific gravity (g cm-3) Annonaceae Ruizodendron ovale 0.589 Annonaceae Trigynaea duckei 0.589 Annonaceae Trigynaea ecuadorensis 0.589 Annonaceae Unonopsis floribunda 0.42 Annonaceae Unonopsis guatterioides 0.463 Annonaceae Unonopsis matthewsii 0.534 Annonaceae Unonopsis peruviana 0.534 Annonaceae Unonopsis spectabilis 0.463 Annonaceae Unonopsis veneficiorum 0.436 Annonaceae Xylopia benthamii 0.6 Annonaceae Xylopia calophylla 0.595 Annonaceae Xylopia ligustrifolia 0.6 Apocynaceae Aspidosperma excelsum 0.792 Apocynaceae Aspidosperma macrocarpon 0.683 Apocynaceae Aspidosperma marcgravianum 0.733 Apocynaceae Aspidosperma megaphyllum 0.673 Apocynaceae Aspidosperma nitidum 0.764 Apocynaceae Aspidosperma parvifolium 0.735 Apocynaceae Aspidosperma rigidum 0.46 Apocynaceae Aspidosperma spruceanum 0.753 Apocynaceae Aspidosperma tambopatense 0.741 Apocynaceae Aspidosperma vargasii 0.741 Apocynaceae Geissospermum reticulatum 0.782 Apocynaceae Himatanthus articulatus 0.536 Apocynaceae Himatanthus sucuuba 0.462 Apocynaceae Lacmellea arborescens 0.513 Apocynaceae Lacmellea peruviana 0.512 Apocynaceae Macoubea guianensis 0.414 Apocynaceae Rauvolfia leptophylla 0.617 Apocynaceae Rauvolfia praecox 0.46 Apocynaceae Rauvolfia sprucei 0.464 Apocynaceae Tabernaemontana cymosa 0.47 Apocynaceae Tabernaemontana psychotriifolia 0.462 Apocynaceae Tabernaemontana sananho 0.462 Aquifoliaceae Ilex aggregata 0.642 Aquifoliaceae Ilex biserrulata 0.642 Aquifoliaceae Ilex gabrielleana 0.677 Aquifoliaceae Ilex guayusa 0.642 Aquifoliaceae Ilex karstenii 0.65 Aquifoliaceae Ilex laurina 0.642

Family Species Basic wood specific gravity (g cm-3) Aquifoliaceae Ilex microdonta 0.731 Aquifoliaceae Ilex nayana 0.551 Aquifoliaceae Ilex nervosa 0.642 Aquifoliaceae Ilex sessiliflora 0.625 Aquifoliaceae Ilex trichoclada 0.637 Aquifoliaceae Ilex villosula 0.667 Araliaceae Dendropanax arboreus 0.42 Araliaceae Dendropanax cuneatus 0.514 Araliaceae Dendropanax umbellatus 0.407 Araliaceae Dendropanax weberbaueri 0.457 Araliaceae Dendropanax williamsii 0.457 Araliaceae Oreopanax capitatus 0.557 Araliaceae Oreopanax kuntzei 0.476 Araliaceae Oreopanax microflorous 0.428 Araliaceae Oreopanax ruizii 0.704 Araliaceae Schefflera allocotantha 0.519 Araliaceae Schefflera inambarica 0.545 Araliaceae Schefflera morototoni 0.455 Araliaceae Schefflera patula 0.545 Araliaceae Schefflera sprucei 0.455 Arecaceae Astrocaryum chonta 0.508 Arecaceae Astrocaryum gratum 0.508 Arecaceae Astrocaryum macrocalyx 0.508 Arecaceae Astrocaryum murumuru 0.508 Arecaceae Attalea butyracea 0.326 Arecaceae Attalea cephalotes 0.326 Arecaceae Attalea maripa 0.326 Arecaceae Attalea phalerata 0.326 Arecaceae Bactris dahlgreniana 0.393 Arecaceae Bactris gasipaes 0.393 Arecaceae Bactris setulosa 0.393 Arecaceae Chelyocarpus ulei 0.393 Arecaceae Dictyocaryum lamarckianum 0.437 Arecaceae Euterpe precatoria 0.388 Arecaceae Geonoma brongniartii 0.393 Arecaceae Geonoma undata 0.351 Arecaceae Iriartea deltoidea 0.265 Arecaceae Jessenia bataua 0.393 Arecaceae Oenocarpus bataua 0.682 Arecaceae Oenocarpus mapora 0.713

Family Species Basic wood specific gravity (g cm-3) Arecaceae Phytelephas macrocarpa 0.393 Arecaceae Scheelea butyracea 0.393 Arecaceae Scheelea cephalotes 0.393 Arecaceae Socratea exorrhiza 0.23 Arecaceae Socratea salazarii 0.23 Asteraceae Barnadesia caryophylla 0.6 Asteraceae Gynoxys nitida 0.62 Asteraceae Nordenstamia repanda 0.604 Asteraceae Vernonanthura patens 0.54 Betulaceae Alnus acuminata 0.429 Bignoniaceae Jacaranda copaia 0.351 Bignoniaceae Jacaranda glabra 0.395 Bignoniaceae Jacaranda obtusifolia 0.483 Bignoniaceae Jacaranda spectabilis 0.394 Bignoniaceae Tabebuia serratifolia 0.924 Bixaceae Bixa arborea 0.37 Bixaceae Bixa platycarpa 0.346 Boraginaceae Cordia alliodora 0.493 Boraginaceae Cordia hebeclada 0.525 Boraginaceae Cordia lomatoloba 0.41 Boraginaceae Cordia mexiana 0.525 Boraginaceae Cordia nodosa 0.39 Boraginaceae Cordia panamensis 0.39 Boraginaceae Cordia ripicola 0.525 Boraginaceae Cordia scabrifolia 0.474 Boraginaceae Cordia tetrandra 0.345 Boraginaceae Cordia toqueve 0.516 Boraginaceae Cordia trachyphylla 0.527 Boraginaceae Cordia ucayaliensis 0.41 Brunelliaceae Brunellia boliviana 0.343 Brunelliaceae Brunellia brunnea 0.316 Brunelliaceae Brunellia cuzcoensis 0.316 Brunelliaceae Brunellia dulcis 0.316 Brunelliaceae Brunellia inermis 0.343 Brunelliaceae Brunellia littlei 0.316 Brunelliaceae Brunellia stenoptera 0.446 Brunelliaceae Brunellia weberbaueri 0.288 Burseraceae Crepidospermum goudotianum 0.579 Burseraceae Dacryodes peruviana 0.53 Burseraceae Protium altsonii 0.684

Family Species Basic wood specific gravity (g cm-3) Burseraceae Protium amazonicum 0.408 Burseraceae Protium apiculatum 0.556 Burseraceae Protium aracouchini 0.49 Burseraceae Protium decandrum 0.512 Burseraceae Protium glabrescens 0.532 Burseraceae Protium hebetatum 0.567 Burseraceae Protium insigne 0.556 Burseraceae Protium montanum 0.53 Burseraceae Protium neglectum 0.556 Burseraceae Protium opacum 0.57 Burseraceae Protium pallidum 0.556 Burseraceae Protium plagiocarpium 0.551 Burseraceae Protium puncticulatum 0.556 Burseraceae Protium rhynchophyllum 0.556 Burseraceae Protium robustum 0.556 Burseraceae Protium sagotianum 0.558 Burseraceae Protium spruceanum 0.556 Burseraceae Protium tenuifolium 0.57 Burseraceae Tetragastris altissima 0.708 Burseraceae Tetragastris panamensis 0.717 Burseraceae Trattinnickia aspera 0.424 Burseraceae Trattinnickia boliviana 0.469 Burseraceae Trattinnickia burserifolia 0.46 Burseraceae Trattinnickia glaziovii 0.521 Burseraceae Trattinnickia peruviana 0.56 Burseraceae Trattinnickia rhoifolia 0.451 Buxaceae Styloceras brokawii 0.638 Calophyllaceae Calophyllum angulare 0.643 Calophyllaceae Calophyllum brasiliense 0.583 Calophyllaceae Caraipa densifolia 0.641 Calophyllaceae Caraipa myrcioides 0.655 Calophyllaceae Marila laxiflora 0.371 Campanulaceae Siphocampylus vatkeanus 0.654 Cannabaceae Celtis schippii 0.616 Cannabaceae Trema micrantha 0.425 Capparaceae Capparis amplissima 0.663 Capparaceae Capparis flexuosa 0.684 Capparaceae Capparis macrophylla 0.684 Capparaceae Capparis sola 0.684 Cardiopteridaceae Citronella incarum 0.635

Family Species Basic wood specific gravity (g cm-3) Cardiopteridaceae Dendrobangia boliviana 0.53 Caricaceae Jacaratia digitata 0.223 Caryocaraceae Anthodiscus klugii 0.684 Caryocaraceae Anthodiscus peruanus 0.69 Caryocaraceae Caryocar amygdaliforme 0.688 Caryocaraceae Caryocar glabrum 0.676 Caryocaraceae Caryocar pallidum 0.84 Celastraceae Cheiloclinium cognatum 0.716 Celastraceae Haydenia urbaniana 0.719 Celastraceae Maytenus ebenifolia 0.718 Celastraceae Maytenus macrocarpa 0.718 Celastraceae Maytenus magnifolia 0.718 Celastraceae Salacia gigantea 0.702 Celastraceae Salacia macrantha 0.702 Celastraceae Salacia micrantha 0.702 Celastraceae Salacia opacifolia 0.702 Celastraceae Tontelea attenuata 0.702 Chloranthaceae Hedyosmum anisodorum 0.437 Chloranthaceae Hedyosmum cuatrecazanum 0.275 Chloranthaceae Hedyosmum goudotianum 0.435 Chloranthaceae Hedyosmum maximum 0.531 Chloranthaceae Hedyosmum peruvianum 0.41 Chloranthaceae Hedyosmum racemosum 0.469 Chloranthaceae Hedyosmum scabrum 0.535 Chloranthaceae Hedyosmum translucidum 0.437 Chrysobalanaceae Couepia bernardii 0.797 Chrysobalanaceae Couepia latifolia 0.797 Chrysobalanaceae Couepia racemosa 0.797 Chrysobalanaceae Hirtella excelsa 0.798 Chrysobalanaceae Hirtella lightioides 0.798 Chrysobalanaceae Hirtella macrophylla 0.74 Chrysobalanaceae Hirtella racemosa 0.782 Chrysobalanaceae Hirtella triandra 0.683 Chrysobalanaceae Licania apetala 0.763 Chrysobalanaceae Licania brittoniana 0.674 Chrysobalanaceae Licania canescens 0.88 Chrysobalanaceae Licania caudata 0.825 Chrysobalanaceae Licania cf. impressa 0.819 Chrysobalanaceae Licania harlingii 0.825 Chrysobalanaceae Licania heteromorpha 0.816

Family Species Basic wood specific gravity (g cm-3) Chrysobalanaceae Licania hispida 0.831 Chrysobalanaceae Licania hypoleuca 0.909 Chrysobalanaceae Licania kunthiana 0.88 Chrysobalanaceae Licania macrocarpa 0.825 Chrysobalanaceae Licania micrantha 0.836 Chrysobalanaceae Licania octandra 0.825 Chrysobalanaceae Licania paraensis 0.825 Chrysobalanaceae Parinari klugii 0.702 Chrysobalanaceae Parinari occidentalis 0.702 Chrysobalanaceae Parinari parilis 0.573 Clethraceae Clethra castaneifolia 0.589 Clethraceae Clethra cuneata 0.526 Clethraceae Clethra ferruginea 0.652 Clethraceae Clethra obovata 0.599 Clethraceae Clethra revoluta 0.511 Clethraceae Clethra scabra 0.688 Clusiaceae Chrysochlamys ulei 0.43 Clusiaceae Clusia alata 0.649 Clusiaceae Clusia ducuoides 0.806 Clusiaceae Clusia elliptica 0.76 Clusiaceae Clusia flavida 0.76 Clusiaceae Clusia pavonii 0.654 Clusiaceae Clusia sp1(1048WFR) 0.76 Clusiaceae Clusia sphaerocarpa 0.658 Clusiaceae Clusia thurifera 0.69 Clusiaceae Clusia trochiformis 0.634 Clusiaceae Garcinia macrophylla 0.67 Clusiaceae Garcinia madruno 0.71 Clusiaceae Garcinia ovalifolia 0.68 Clusiaceae Rheedia acuminata 0.669 Clusiaceae Rheedia brasiliensis 0.669 Clusiaceae Rheedia macrophylla 0.669 Clusiaceae Symphonia globulifera 0.608 Clusiaceae Tovomita laurina 0.695 Clusiaceae Tovomita weddelliana 0.512 Combretaceae Buchenavia fanshawei 0.745 Combretaceae Buchenavia grandis 0.755 Combretaceae Buchenavia oxycarpa 0.713 Combretaceae Buchenavia punctata 0.745 Combretaceae Terminalia amazonia 0.674

Family Species Basic wood specific gravity (g cm-3) Combretaceae Terminalia oblonga 0.694 Cunoniaceae Weinmannia auriculata 0.604 Cunoniaceae Weinmannia balbisiana 0.61 Cunoniaceae Weinmannia bangii 0.612 Cunoniaceae Weinmannia cochensis 0.583 Cunoniaceae Weinmannia crassifolia 0.637 Cunoniaceae Weinmannia fagaroides 0.571 Cunoniaceae Weinmannia lechleriana 0.566 Cunoniaceae Weinmannia mariquitae 0.649 Cunoniaceae Weinmannia multijuga 0.636 Cunoniaceae Weinmannia ovata 0.653 Cunoniaceae Weinmannia pinnata 0.61 Cunoniaceae Weinmannia reticulata 0.667 Cyatheaceae Alsophila cuspidata 0.348 Cyatheaceae Alsophila erinacea 0.348 Cyatheaceae Cyathea andina 0.348 Cyatheaceae Cyathea caracasana 0.391 Cyatheaceae Cyathea carolihenrici 0.348 Cyatheaceae Cyathea catacampta 0.348 Cyatheaceae Cyathea cystolepis 0.348 Cyatheaceae Cyathea delgadii 0.348 Cyatheaceae Cyathea divergens 0.348 Cyatheaceae Cyathea herzogii 0.348 Cyatheaceae Cyathea lechleri 0.347 Cyatheaceae Cyathea multisegmenta 0.348 Cyatheaceae Cyathea pallescens 0.348 Cyatheaceae Cyathea ruiziana 0.348 Cyatheaceae Cyathea squamipes 0.348 Dichapetalaceae Tapura acreana 0.62 Dichapetalaceae Tapura juruana 0.641 Dichapetalaceae Tapura peruviana 0.648 Dicksoniaceae Dicksonia sellowiana 0.348 Dipentodontaceae Perrottetia sessiliflora 0.484 Ebenaceae Diospyros artanthifolia 0.599 Ebenaceae Diospyros capreifolia 0.637 Ebenaceae Diospyros manu 0.599 Ebenaceae Diospyros subrotata 0.599 Elaeocarpaceae Sloanea brevipes 0.807 Elaeocarpaceae Sloanea durissima 0.777 Elaeocarpaceae Sloanea eichleri 0.75

Family Species Basic wood specific gravity (g cm-3) Elaeocarpaceae Sloanea fragrans 0.47 Elaeocarpaceae Sloanea garckeana 0.767 Elaeocarpaceae Sloanea guianensis 0.625 Elaeocarpaceae Sloanea latifolia 0.777 Elaeocarpaceae Sloanea laurifolia 0.816 Elaeocarpaceae Sloanea laxiflora 0.767 Elaeocarpaceae Sloanea macrophylla 0.777 Elaeocarpaceae Sloanea meianthera 0.767 Elaeocarpaceae Sloanea obtusifolia 0.777 Elaeocarpaceae Sloanea picapica 0.777 Elaeocarpaceae Sloanea ptariana 0.777 Elaeocarpaceae Sloanea pubescens 0.767 Elaeocarpaceae Sloanea robusta 0.861 Elaeocarpaceae Sloanea rufa 0.807 Elaeocarpaceae Sloanea sinemariensis 0.777 Elaeocarpaceae Sloanea stipitata 0.777 Elaeocarpaceae Sloanea terniflora 0.777 Elaeocarpaceae Sloanea tuerckheimii 0.807 Elaeocarpaceae Vallea stipularis 0.746 Ericaceae Bejaria aestuans 0.791 Ericaceae Cavendishia bracteata 0.722 Erythroxylaceae Erythroxylum deciduum 0.71 Erythroxylaceae Erythroxylum patens 0.71 Erythroxylaceae Erythroxylum squamatum 0.71 Escalloniaceae Escallonia myrtilloides 0.67 Escalloniaceae Escallonia paniculata 0.747 Euphorbiaceae Acalypha stenoloba 0.527 Euphorbiaceae Alchornea acutifolia 0.541 Euphorbiaceae Alchornea anamariae 0.453 Euphorbiaceae Alchornea brittonii 0.35 Euphorbiaceae Alchornea glandulosa 0.373 Euphorbiaceae Alchornea grandiflora 0.483 Euphorbiaceae Alchornea grandis 0.46 Euphorbiaceae Alchornea hilariana 0.532 Euphorbiaceae Alchornea iricurana 0.381 Euphorbiaceae Alchornea latifolia 0.483 Euphorbiaceae Alchornea pearcei 0.436 Euphorbiaceae Alchornea triplinervia 0.651 Euphorbiaceae Aparisthmium cordatum 0.39 Euphorbiaceae Caryodendron orinocense 0.65

Family Species Basic wood specific gravity (g cm-3) Euphorbiaceae Conceveiba guianensis 0.543 Euphorbiaceae Conceveiba rhytidocarpa 0.413 Euphorbiaceae Croton matourensis 0.45 Euphorbiaceae Croton tessmannii 0.39 Euphorbiaceae Glycydendron amazonicum 0.681 Euphorbiaceae Hevea brasiliensis 0.492 Euphorbiaceae Hevea guianensis 0.486 Euphorbiaceae Hura crepitans 0.367 Euphorbiaceae Mabea maynensis 0.608 Euphorbiaceae Mabea nitida 0.647 Euphorbiaceae Mabea occidentalis 0.608 Euphorbiaceae Mabea speciosa 0.635 Euphorbiaceae Pausandra trianae 0.59 Euphorbiaceae Sagotia racemosa 0.58 Euphorbiaceae Sapium aereum 0.423 Euphorbiaceae Sapium eglandulosum 0.423 Euphorbiaceae Sapium glandulosum 0.421 Euphorbiaceae Sapium ixiamasense 0.423 Euphorbiaceae Sapium laurifolium 0.529 Euphorbiaceae Sapium marmieri 0.41 Euphorbiaceae Senefeldera inclinata 0.96 Euphorbiaceae Tetrorchidium rubrivenium 0.353 Fabaceae Abarema floribunda 0.512 Fabaceae Abarema jupunba 0.585 Fabaceae Acacia loretensis 0.714 Fabaceae Amburana cearensis 0.514 Fabaceae Andira inermis 0.644 Fabaceae Andira legalis 0.656 Fabaceae Andira multistipula 0.752 Fabaceae Andira surinamensis 0.705 Fabaceae Apuleia leiocarpa 0.788 Fabaceae Bauhinia glabra 0.635 Fabaceae Bauhinia tarapotensis 0.6 Fabaceae Cassia grandis 0.783 Fabaceae Cassia sylvestris 0.783 Fabaceae Cedrelinga cateniformis 0.476 Fabaceae Centrolobium paraense 0.711 Fabaceae Copaifera reticulata 0.608 Fabaceae Cyathostegia mathewsii 0.57 Fabaceae Dialium guianense 0.878

Family Species Basic wood specific gravity (g cm-3) Fabaceae Diplotropis martiusii 0.633 Fabaceae Diplotropis purpurea 0.779 Fabaceae Dipteryx alata 0.937 Fabaceae Dipteryx micrantha 0.871 Fabaceae Dipteryx odorata 0.923 Fabaceae Dussia tessmannii 0.553 Fabaceae Enterolobium cyclocarpum 0.39 Fabaceae Enterolobium schomburgkii 0.701 Fabaceae Erythrina ulei 0.111 Fabaceae Hymenaea courbaril 0.782 Fabaceae Hymenaea oblongifolia 0.737 Fabaceae Hymenaea parvifolia 0.877 Fabaceae Inga acreana 0.579 Fabaceae Inga acrocephala 0.513 Fabaceae Inga acuminata 0.579 Fabaceae Inga alba 0.586 Fabaceae Inga aria 0.606 Fabaceae Inga auristellae 0.579 Fabaceae Inga barbata 0.587 Fabaceae Inga bourgonii 0.545 Fabaceae Inga bracteosa 0.579 Fabaceae Inga capitata 0.592 Fabaceae Inga cayennensis 0.53 Fabaceae Inga chartacea 0.512 Fabaceae Inga cinnamomea 0.528 Fabaceae Inga cordatoalata 0.587 Fabaceae Inga coriacea 0.579 Fabaceae Inga coruscans 0.72 Fabaceae Inga densiflora 0.579 Fabaceae Inga edulis 0.587 Fabaceae Inga gracilifolia 0.587 Fabaceae Inga heterophylla 0.438 Fabaceae Inga jenmani 0.579 Fabaceae Inga jenmanii 0.579 Fabaceae Inga killipiana 0.575 Fabaceae Inga lallensis 0.579 Fabaceae Inga laurina 0.62 Fabaceae Inga leiocalycina 0.442 Fabaceae Inga macrophylla 0.587 Fabaceae Inga marginata 0.545

Family Species Basic wood specific gravity (g cm-3) Fabaceae Inga megalobotrys 0.587 Fabaceae Inga multinervis 0.587 Fabaceae Inga nobilis 0.56 Fabaceae Inga pavoniana 0.579 Fabaceae Inga pezizifera 0.621 Fabaceae Inga punctata 0.56 Fabaceae Inga quaternata 0.579 Fabaceae Inga ruiziana 0.565 Fabaceae Inga sapindoides 0.579 Fabaceae Inga semialata 0.6 Fabaceae Inga spectabilis 0.579 Fabaceae Inga splendens 0.6 Fabaceae Inga stenoptera 0.579 Fabaceae Inga stipularis 0.53 Fabaceae Inga striata 0.579 Fabaceae Inga tenuistipula 0.579 Fabaceae Inga thibaudiana 0.658 Fabaceae Inga tomentosa 0.579 Fabaceae Inga vismiifolia 0.599 Fabaceae Lecointea amazonica 0.895 Fabaceae Lecointea peruviana 0.893 Fabaceae Lonchocarpus seorsus 0.754 Fabaceae Lonchocarpus spiciflorus 0.757 Fabaceae Lonchocarpus trifolius 0.757 Fabaceae Machaerium acutifolium 1.12 Fabaceae Machaerium pilosum 0.79 Fabaceae Myroxylon balsamum 0.779 Fabaceae Ormosia amazonica 0.601 Fabaceae Ormosia bopiensis 0.601 Fabaceae Ormosia coarctata 0.601 Fabaceae Ormosia panamensis 0.601 Fabaceae Parkia nitida 0.383 Fabaceae Parkia velutina 0.435 Fabaceae Peltogyne floribunda 0.787 Fabaceae Phyllocarpus riedelii 0.713 Fabaceae Piptadenia adiantoides 0.819 Fabaceae Piptadenia boliviana 0.78 Fabaceae Piptadenia communis 0.819 Fabaceae Piptadenia cuzcoensis 0.78 Fabaceae Piptadenia pteroclada 0.78

Family Species Basic wood specific gravity (g cm-3) Fabaceae Pithecellobium corymbosum 0.512 Fabaceae Pithecellobium latifolium 0.36 Fabaceae Platymiscium pinnatum 0.745 Fabaceae Platymiscium stipulare 0.748 Fabaceae Platymiscium ulei 0.75 Fabaceae Platypodium elegans 0.806 Fabaceae Platypodium viride 0.75 Fabaceae Pseudopiptadenia suaveolens 0.68 Fabaceae Pterocarpus amazonicus 0.491 Fabaceae Pterocarpus amazonum 0.491 Fabaceae Pterocarpus rohrii 0.439 Fabaceae Pterocarpus ulei 0.491 Fabaceae Schizolobium parahyba 0.358 Fabaceae Sclerolobium bracteosum 0.528 Fabaceae Senegalia tenuifolia 0.67 Fabaceae Senna silvestris 0.604 Fabaceae Stryphnodendron guianense 0.568 Fabaceae Stryphnodendron pulcherrimum 0.475 Fabaceae Stryphnodendron purpureum 0.55 Fabaceae Swartzia amplifolia 0.842 Fabaceae Swartzia arborescens 0.835 Fabaceae Swartzia dipetala 0.842 Fabaceae Swartzia leptopetala 0.842 Fabaceae Swartzia myrtifolia 0.9 Fabaceae Tachigali bracteosa 0.579 Fabaceae Tachigali cenepensis 0.579 Fabaceae Tachigali macbridei 0.528 Fabaceae Tachigali poeppigiana 0.579 Fabaceae Tachigali polyphylla 0.637 Fabaceae Tachigali rusbyi 0.579 Fabaceae Tachigali setifera 0.367 Fabaceae Tachigali vasquezii 0.579 Fabaceae Vatairea fusca 0.657 Fabaceae Zygia latifolia 0.75 Gentianaceae Macrocarpaea maguirei 0.589 Humiriaceae Humiriastrum colombianum 0.375 Humiriaceae Sacoglottis mattogrossensis 0.77 Humiriaceae Vantanea guianensis 0.822 Hypericaceae Vismia gracilis 0.398 Hypericaceae Vismia mandurr 0.702

Family Species Basic wood specific gravity (g cm-3) Hypericaceae Vismia tomentosa 0.314 Icacinaceae Calatola costaricensis 0.565 Juglandaceae Juglans australis 0.518 Lacistemataceae Lacistema aggregatum 0.506 Lacistemataceae Lacistema nena 0.506 Lacistemataceae Lozania mutisiana 0.506 Lamiaceae Aegiphila filipes 0.645 Lamiaceae Aegiphila integrifolia 0.86 Lamiaceae Aegiphila racemosa 0.645 Lamiaceae Aegiphila saltensis 0.404 Lamiaceae Vitex cymosa 0.563 Lamiaceae Vitex pseudolea 0.556 Lamiaceae Vitex triflora 0.559 Lauraceae Aiouea grandifolia 0.37 Lauraceae Aniba canelilla 0.952 Lauraceae Aniba firmula 0.626 Lauraceae Aniba guianensis 0.52 Lauraceae Aniba panurensis 0.61 Lauraceae Aniba puchury-minor 0.533 Lauraceae Aniba robusta 0.626 Lauraceae Aniba taubertiana 0.593 Lauraceae Aniba terminalis 0.54 Lauraceae Beilschmiedia latifolia 0.548 Lauraceae Beilschmiedia tovarensis 0.52 Lauraceae Caryodaphnopsis fosteri 0.559 Lauraceae Cinnamomum triplinerve 0.51 Lauraceae Cryptocarya aschersoniana 0.57 Lauraceae Endlicheria bracteata 0.501 Lauraceae Endlicheria directonervia 0.56 Lauraceae Endlicheria dysodantha 0.501 Lauraceae Endlicheria formosa 0.504 Lauraceae Endlicheria griseo-sericea 0.56 Lauraceae Endlicheria klugii 0.487 Lauraceae Endlicheria krukovii 0.455 Lauraceae Endlicheria macrophylla 0.365 Lauraceae Endlicheria paniculata 0.682 Lauraceae Endlicheria ruforamula 0.501 Lauraceae Endlicheria williamsii 0.47 Lauraceae Licaria aurea 0.787 Lauraceae Licaria canella 0.787

100

Family Species Basic wood specific gravity (g cm-3) Lauraceae Licaria cannella 0.939 Lauraceae Licaria pucheri 0.786 Lauraceae Licaria triandra 0.631 Lauraceae Mezilaurus campaucola 0.715 Lauraceae Mezilaurus subcordata 0.748 Lauraceae Nectandra bicolor 0.53 Lauraceae Nectandra cissiflora 0.504 Lauraceae Nectandra cuneatocordata 0.504 Lauraceae Nectandra cuspidata 0.518 Lauraceae Nectandra globosa 0.39 Lauraceae Nectandra lineata 0.431 Lauraceae Nectandra longifolia 0.414 Lauraceae Nectandra membranacea 0.504 Lauraceae Nectandra pulverulenta 0.504 Lauraceae Nectandra reticulata 0.342 Lauraceae Nectandra turbacensis 0.545 Lauraceae Nectandra viburnoides 0.504 Lauraceae Nectandra yarinensis 0.504 Lauraceae Ocotea aciphylla 0.584 Lauraceae Ocotea bofo 0.514 Lauraceae Ocotea camphoromoea 0.514 Lauraceae Ocotea cernua 0.454 Lauraceae Ocotea cuprea 0.539 Lauraceae Ocotea glabriflora 0.71 Lauraceae Ocotea guianensis 0.53 Lauraceae Ocotea insularis 0.559 Lauraceae Ocotea javitensis 0.505 Lauraceae Ocotea leucoxylon 0.462 Lauraceae Ocotea longifolia 0.428 Lauraceae Ocotea oblonga 0.533 Lauraceae Ocotea obovata 0.539 Lauraceae Ocotea puberula 0.455 Lauraceae Ocotea rubrinervis 0.514 Lauraceae Ocotea tessmannii 0.514 Lauraceae Persea areolatocostae 0.515 Lauraceae Persea brevipes 0.515 Lauraceae Persea caerulea 0.437 Lauraceae Persea corymbosa 0.457 Lauraceae Persea ferruginea 0.563 Lauraceae Persea mutisii 0.541

101

Family Species Basic wood specific gravity (g cm-3) Lauraceae Persea nudigemma 0.515 Lauraceae Persea peruviana 0.444 Lauraceae Pleurothyrium cuneifolium 0.473 Lauraceae Pleurothyrium intermedium 0.47 Lauraceae Pleurothyrium krukovii 0.47 Lauraceae Pleurothyrium parviflorum 0.47 Lauraceae Pleurothyrium poeppigii 0.471 Lauraceae Pleurothyrium trianae 0.472 Lauraceae Pleurothyrium vasquezii 0.47 Lauraceae Pleurothyrium williamsii 0.47 Lauraceae Rhodostemonodaphne grandis 0.39 Lauraceae Rhodostemonodaphne kunthiana 0.479 Lecythidaceae Bertholletia excelsa 0.624 Lecythidaceae Cariniana decandra 0.58 Lecythidaceae Cariniana estrellensis 0.565 Lecythidaceae Couratari guianensis 0.507 Lecythidaceae Couratari macrosperma 0.67 Lecythidaceae Eschweilera albiflora 0.86 Lecythidaceae Eschweilera baguensis 0.855 Lecythidaceae Eschweilera coriacea 0.852 Lecythidaceae Eschweilera gigantea 0.78 Lecythidaceae Eschweilera juruensis 0.96 Lecythidaceae Eschweilera klugii 0.855 Lecythidaceae Gustavia augusta 0.65 Lecythidaceae Gustavia hexapetala 0.716 Lepidobotryaceae Ruptiliocarpon caracolito 0.638 Linaceae Hebepetalum humiriifolia 0.871 Linaceae Hebepetalum humiriifolium 0.871 Linaceae Roucheria columbiana 0.77 Linaceae Roucheria punctata 0.825 Loganiaceae Strychnos tarapotensis 0.54 Loranthaceae Gaiadendron punctatum 0.586 Lythraceae Lafoensia punicifolia 0.705 Lythraceae Physocalymma scaberrimum 0.705 Magnoliaceae Magnolia amazonica 0.608 Magnoliaceae Magnolia boliviana 0.618 Magnoliaceae Magnolia gilbertoi 0.628 Malpighiaceae Bunchosia argentea 0.65 Malpighiaceae Byrsonima arthropoda 0.62 Malpighiaceae Byrsonima poeppigiana 0.643

102

Family Species Basic wood specific gravity (g cm-3) Malvaceae Apeiba aspera 0.137 Malvaceae Apeiba glabra 0.32 Malvaceae Apeiba membranacea 0.255 Malvaceae Apeiba tibourbou 0.2 Malvaceae Cavanillesia umbellata 0.165 Malvaceae Ceiba insignis 0.285 Malvaceae Ceiba pentandra 0.335 Malvaceae Ceiba samauma 0.488 Malvaceae Chorisia insignis 0.292 Malvaceae Chorisia integrifolia 0.283 Malvaceae Eriotheca globosa 0.41 Malvaceae Guazuma crinita 0.44 Malvaceae Guazuma ulmifolia 0.533 Malvaceae Heliocarpus americanus 0.215 Malvaceae Huberodendron swietenioides 0.564 Malvaceae Luehea cymulosa 0.51 Malvaceae Lueheopsis duckeana 0.616 Malvaceae Matisia bicolor 0.516 Malvaceae Matisia cordata 0.49 Malvaceae Matisia malacocalyx 0.535 Malvaceae Matisia ochrocalyx 0.57 Malvaceae Matisia rhombifolia 0.542 Malvaceae Ochroma pyramidale 0.148 Malvaceae Pachira insignis 0.384 Malvaceae Pachira trinitensis 0.384 Malvaceae Pseudobombax septenatum 0.248 Malvaceae Pterygota amazonica 0.56 Malvaceae Quararibea guianensis 0.56 Malvaceae Quararibea ochrocalyx 0.488 Malvaceae Quararibea putumayensis 0.488 Malvaceae Quararibea wittii 0.532 Malvaceae Spirotheca rosea 0.33 Malvaceae Sterculia apetala 0.415 Malvaceae Sterculia peruviana 0.485 Malvaceae Sterculia rebeccae 0.49 Malvaceae Sterculia simplex 0.49 Malvaceae Sterculia tessmannii 0.49 Malvaceae Theobroma cacao 0.42 Malvaceae Theobroma speciosum 0.63 Malvaceae Theobroma subincanum 0.47

103

Family Species Basic wood specific gravity (g cm-3) Melastomataceae Axinaea glandulosa 0.606 Melastomataceae Axinaea pennellii 0.606 Melastomataceae Bellucia aequiloba 0.59 Melastomataceae Bellucia grossularioides 0.598 Melastomataceae Bellucia pentamera 0.615 Melastomataceae Graffenrieda cucullata 0.64 Melastomataceae Meriania cuzcoana 0.49 Melastomataceae Miconia affinis 0.624 Melastomataceae Miconia alpina 0.74 Melastomataceae Miconia aristata 0.633 Melastomataceae Miconia astroplocama 0.545 Melastomataceae Miconia aulocalyx 0.66 Melastomataceae Miconia aurea 0.654 Melastomataceae Miconia axinaeoides 0.545 Melastomataceae Miconia barbeyana 0.531 Melastomataceae Miconia brachyanthera 0.579 Melastomataceae Miconia brevistylis 0.541 Melastomataceae Miconia bullata 0.638 Melastomataceae Miconia calophylla 0.468 Melastomataceae Miconia calvescens 0.42 Melastomataceae Miconia centrodesma 0.551 Melastomataceae Miconia cookii 0.676 Melastomataceae Miconia crassipes 0.548 Melastomataceae Miconia crassistigma 0.667 Melastomataceae Miconia cretacea 0.593 Melastomataceae Miconia dolichorrhyncha 0.624 Melastomataceae Miconia elongata 0.627 Melastomataceae Miconia impetiolaris 0.624 Melastomataceae Miconia lamprophylla 0.548 Melastomataceae Miconia longifolia 0.75 Melastomataceae Miconia madisonii 0.591 Melastomataceae Miconia modica 0.419 Melastomataceae Miconia peruviana 0.558 Melastomataceae Miconia prasina 0.705 Melastomataceae Miconia punctata 0.624 Melastomataceae Miconia pyrifolia 0.624 Melastomataceae Miconia setulosa 0.658 Melastomataceae Miconia spennerostachya 0.556 Melastomataceae Miconia splendens 0.66 Melastomataceae Miconia stelligera 0.545

104

Family Species Basic wood specific gravity (g cm-3) Melastomataceae Miconia terborghii 0.593 Melastomataceae Miconia terera 0.553 Melastomataceae Miconia ternatifolia 0.62 Melastomataceae Miconia tetragona 0.629 Melastomataceae Miconia theizans 0.562 Melastomataceae Miconia tomentosa 0.71 Melastomataceae Miconia trinervia 0.62 Melastomataceae Miconia triplinervis 0.624 Melastomataceae Mouriri apiranga 0.84 Melastomataceae Mouriri grandiflora 0.9 Melastomataceae Mouriri nervosa 0.84 Melastomataceae Mouriri nigra 0.895 Melastomataceae Mouriri peruviana 0.84 Melastomataceae Tibouchina dimorphophylla 0.635 Meliaceae Cabralea canjerana 0.478 Meliaceae Cedrela fissilis 0.437 Meliaceae Cedrela odorata 0.427 Meliaceae Guarea glabra 0.503 Meliaceae Guarea gomma 0.65 Meliaceae Guarea guidonia 0.565 Meliaceae Guarea kunthiana 0.645 Meliaceae Guarea macrophylla 0.528 Meliaceae Guarea pterorhachis 0.77 Meliaceae Guarea silvatica 0.605 Meliaceae Ruagea glabra 0.47 Meliaceae Ruagea subviridiflora 0.529 Meliaceae Trichilia adolfi 0.657 Meliaceae Trichilia areolata 0.657 Meliaceae Trichilia elegans 0.657 Meliaceae Trichilia euneura 0.657 Meliaceae Trichilia havanensis 0.622 Meliaceae Trichilia maynasiana 0.69 Meliaceae Trichilia micrantha 0.613 Meliaceae Trichilia pachypoda 0.657 Meliaceae Trichilia pallida 0.68 Meliaceae Trichilia pleeana 0.675 Meliaceae Trichilia poeppigii 0.657 Meliaceae Trichilia quadrijuga 0.548 Meliaceae Trichilia rubra 0.585 Meliaceae Trichilia schomburgkii 0.687

105

Family Species Basic wood specific gravity (g cm-3) Meliaceae Trichilia septentrionalis 0.654 Meliaceae Trichilia solitudinis 0.657 Monimiaceae Mollinedia caudata 0.612 Monimiaceae Mollinedia killipii 0.61 Monimiaceae Mollinedia lanceolata 0.63 Monimiaceae Mollinedia ovata 0.591 Monimiaceae Mollinedia repanda 0.61 Monimiaceae Mollinedia simulans 0.61 Monimiaceae Mollinedia tomentosa 0.61 Monimiaceae Siparuna decipiens 0.612 Moraceae Batocarpus amazonicus 0.533 Moraceae Batocarpus costaricensis 0.533 Moraceae Brosimum alicastrum 0.6 Moraceae Brosimum guianense 0.844 Moraceae Brosimum lactescens 0.656 Moraceae Brosimum parinarioides 0.631 Moraceae Brosimum rubescens 0.825 Moraceae Brosimum utile 0.507 Moraceae Castilla ulei 0.568 Moraceae Clarisia biflora 0.497 Moraceae Clarisia racemosa 0.565 Moraceae Ficus americana 0.542 Moraceae Ficus apollinaris 0.493 Moraceae Ficus boliviana 0.493 Moraceae Ficus casapiensis 0.394 Moraceae Ficus cervantesiana 0.482 Moraceae Ficus citrifolia 0.4 Moraceae Ficus coerulescens 0.401 Moraceae Ficus crocata 0.474 Moraceae Ficus cuatrecasana 0.401 Moraceae Ficus donnell-smithii 0.493 Moraceae Ficus ecuadorensis 0.482 Moraceae Ficus erythrosticta 0.394 Moraceae Ficus gomelleira 0.396 Moraceae Ficus insipida 0.346 Moraceae Ficus juruensis 0.394 Moraceae Ficus killipii 0.394 Moraceae Ficus macbridei 0.533 Moraceae Ficus mathewsii 0.385 Moraceae Ficus maxima 0.556

106

Family Species Basic wood specific gravity (g cm-3) Moraceae Ficus obtusifolia 0.482 Moraceae Ficus pallida 0.493 Moraceae Ficus perforata 0.394 Moraceae Ficus pertusa 0.42 Moraceae Ficus schippii 0.493 Moraceae Ficus schultesii 0.394 Moraceae Ficus sphenophylla 0.493 Moraceae Ficus tonduzii 0.527 Moraceae Ficus tovarensis 0.482 Moraceae Ficus trapezicola 0.482 Moraceae Ficus trigona 0.394 Moraceae Ficus ypsilophlebia 0.394 Moraceae Helicostylis elegans 0.65 Moraceae Helicostylis tomentosa 0.627 Moraceae Maclura tinctoria 0.733 Moraceae Maquira calophylla 0.62 Moraceae Maquira guianensis 0.766 Moraceae Morus insignis 0.559 Moraceae Naucleopsis concinna 0.67 Moraceae Naucleopsis krukovii 0.651 Moraceae Naucleopsis pseudonaga 0.651 Moraceae Naucleopsis ternstroemiiflora 0.612 Moraceae Naucleopsis ulei 0.67 Moraceae Olmedia aspera 0.579 Moraceae Perebea angustifolia 0.517 Moraceae Perebea guianensis 0.563 Moraceae Perebea tessmannii 0.501 Moraceae Perebea xanthochyma 0.56 Moraceae Poulsenia armata 0.503 Moraceae Pseudolmedia laevigata 0.582 Moraceae Pseudolmedia laevis 0.618 Moraceae Pseudolmedia macrophylla 0.66 Moraceae Pseudolmedia murure 0.67 Moraceae Pseudolmedia rigida 0.669 Moraceae Sorocea briquetii 0.649 Moraceae Sorocea hirtella 0.578 Moraceae Sorocea pileata 0.578 Moraceae Sorocea steinbachii 0.578 Moraceae Trophis caucana 0.47 Moraceae Trophis scandens 0.574

107

Family Species Basic wood specific gravity (g cm-3) Myricaceae Morella pubescens 0.596 Myristicaceae Iryanthera juruensis 0.633 Myristicaceae Iryanthera laevis 0.615 Myristicaceae Iryanthera olacoides 0.598 Myristicaceae Iryanthera tessmannii 0.598 Myristicaceae Otoba parvifolia 0.426 Myristicaceae Virola calophylla 0.395 Myristicaceae Virola decorticans 0.478 Myristicaceae Virola duckei 0.382 Myristicaceae Virola elongata 0.523 Myristicaceae Virola flexuosa 0.51 Myristicaceae Virola loretensis 0.478 Myristicaceae Virola minutiflora 0.441 Myristicaceae Virola mollissima 0.478 Myristicaceae Virola multinervia 0.615 Myristicaceae Virola pavonis 0.587 Myristicaceae Virola peruviana 0.478 Myristicaceae Virola sebifera 0.403 Myristicaceae Virola surinamensis 0.418 Myrtaceae Calyptranthes densiflora 0.86 Myrtaceae Calyptranthes longifolia 0.86 Myrtaceae Calyptranthes paniculata 0.86 Myrtaceae Eugenia acrensis 0.737 Myrtaceae Eugenia biflora 0.8 Myrtaceae Eugenia feijoi 0.789 Myrtaceae Eugenia florida 0.855 Myrtaceae Eugenia myrobalana 0.737 Myrtaceae Eugenia punicifolia 0.737 Myrtaceae Eugenia riparia 0.737 Myrtaceae Eugenia uniflora 0.737 Myrtaceae Myrcia aliena 0.738 Myrtaceae Myrcia atrorufa 0.733 Myrtaceae Myrcia egensis 0.706 Myrtaceae Myrcia fallax 0.713 Myrtaceae Myrcia floribunda 0.806 Myrtaceae Myrcia guianensis 0.806 Myrtaceae Myrcia magnifolia 0.809 Myrtaceae Myrcia mollis 0.708 Myrtaceae Myrcia paivae 0.774 Myrtaceae Myrcia rostrata 0.754

108

Family Species Basic wood specific gravity (g cm-3) Myrtaceae Myrcia splendens 0.604 Myrtaceae Myrciaria amazonica 0.658 Myrtaceae Psidium acutangulum 0.74 Myrtaceae Siphoneugena densiflora 0.755 Nyctaginaceae Guapira noxia 0.492 Nyctaginaceae Neea altissima 0.64 Nyctaginaceae Neea chlorantha 0.64 Nyctaginaceae Neea dimorphophylla 0.664 Nyctaginaceae Neea divaricata 0.444 Nyctaginaceae Neea floribunda 0.62 Nyctaginaceae Neea laxa 0.679 Nyctaginaceae Neea macrophylla 0.64 Nyctaginaceae Neea oppositifolia 0.893 Nyctaginaceae Neea ovalifolia 0.77 Nyctaginaceae Neea parviflora 0.64 Nyctaginaceae Neea spruceana 0.64 Nyctaginaceae Neea verticillata 0.64 Nyctaginaceae Neea virens 0.519 Ochnaceae Lacunaria jenmanii 0.803 Ochnaceae Ouratea castaneifolia 0.77 Ochnaceae Ouratea iquitosensis 0.774 Ochnaceae Quiina amazonica 0.92 Ochnaceae Quiina florida 0.728 Ochnaceae Quiina nitens 0.862 Ochnaceae Quiina paraensis 0.862 Ochnaceae Quiina peruviana 0.862 Olacaceae Heisteria acuminata 0.696 Olacaceae Heisteria duckei 0.696 Olacaceae Heisteria nitida 0.696 Olacaceae Heisteria ovata 0.54 Olacaceae Heisteria spruceana 0.73 Olacaceae Minquartia guianensis 0.787 Opiliaceae Agonandra peruviana 0.645 Opiliaceae Agonandra silvatica 0.831 Pentaphylacaceae Freziera angulosa 0.599 Pentaphylacaceae Freziera dudleyi 0.599 Pentaphylacaceae Freziera karsteniana 0.585 Pentaphylacaceae Freziera lanata 0.613 Pentaphylacaceae Ternstroemia brachypoda 0.696 Pentaphylacaceae Ternstroemia globiflora 0.69

109

Family Species Basic wood specific gravity (g cm-3) Peraceae Pera benensis 0.666 Peraceae Pera tomentosa 0.666 Phyllanthaceae Hieronyma alchorneoides 0.631 Phyllanthaceae Hieronyma andina 0.507 Phyllanthaceae Hieronyma duquei 0.7 Phyllanthaceae Hieronyma fendleri 0.629 Phyllanthaceae Hieronyma laxiflora 0.55 Phyllanthaceae Hieronyma macrocarpa 0.603 Phyllanthaceae Hieronyma oblonga 0.603 Phyllanthaceae Margaritaria nobilis 0.595 Phyllanthaceae Phyllanthus attenuatus 0.587 Phyllanthaceae Phyllanthus juglandifolius 0.587 Phyllanthaceae Richeria grandis 0.678 Phytolaccaceae Gallesia integrifolia 0.48 Picramniaceae Picramnia juniniana 0.594 Picramniaceae Picramnia latifolia 0.532 Piperaceae Piper coruscans 0.33 Piperaceae Piper obliquum 0.33 Piperaceae Piper reticulatum 0.33 Podocarpaceae Podocarpus oleifolius 0.551 Podocarpaceae Prumnopitys harmsiana 0.57 Podocarpaceae Retrophyllum rospigliosii 0.56 Polygalaceae Monnina connectisepala 0.582 Polygonaceae Coccoloba densifrons 0.58 Polygonaceae Coccoloba lehmannii 0.58 Polygonaceae Coccoloba lepidota 0.675 Polygonaceae Coccoloba mollis 0.52 Polygonaceae Coccoloba peruviana 0.675 Polygonaceae Coccoloba warmingii 0.675 Polygonaceae Coccoloba williamsii 0.675 Polygonaceae Ruprechtia tangarana 0.635 Polygonaceae Triplaris americana 0.49 Polygonaceae Triplaris peruviana 0.515 Polygonaceae Triplaris poeppigiana 0.49 Polygonaceae Triplaris setosa 0.524 Polygonaceae Triplaris weigeltiana 0.486 Primulaceae Cybianthus comperuvianus 0.731 Primulaceae Geissanthus ambigua 0.535 Primulaceae Myrsine andina 0.796 Primulaceae Myrsine coriacea 0.702

110

Family Species Basic wood specific gravity (g cm-3) Primulaceae Myrsine dependens 0.71 Primulaceae Myrsine manglilla 0.685 Primulaceae Myrsine pellucida 0.653 Primulaceae Myrsine youngii 0.76 Primulaceae Stylogyne ambigua 0.43 Primulaceae Stylogyne cauliflora 0.535 Proteaceae Panopsis pearcei 0.482 Proteaceae Roupala monosperma 0.8 Proteaceae Roupala montana 0.8 Puntranjavaceae Drypetes amazonica 0.67 Putranjivaceae Drypetes amazonica 0.67 Putranjivaceae Drypetes gentryi 0.712 Rhamnaceae Rhamnidium elaeocarpum 0.92 Rhamnaceae Ziziphus cinnamomum 0.835 Rosaceae Hesperomeles ferruginea 0.766 Rosaceae Polylepis pauta 0.694 Rosaceae Polylepis sericea 0.732 Rosaceae Prunus debilis 0.856 Rosaceae Prunus detrita 0.7 Rosaceae Prunus herthae 0.634 Rosaceae Prunus huantensis 0.75 Rosaceae Prunus integrifolia 0.657 Rosaceae Prunus pleiantha 0.645 Rosaceae Prunus stipulata 0.645 Rosaceae Prunus vana 0.856 Rubiaceae Alibertia bertierifolia 0.683 Rubiaceae Alibertia latifolia 0.683 Rubiaceae Alseis blackiana 0.536 Rubiaceae Alseis microcarpa 0.679 Rubiaceae Amaioua corymbosa 0.625 Rubiaceae Amaioua guianensis 0.625 Rubiaceae Bathysa australis 0.64 Rubiaceae Bathysa peruviana 0.64 Rubiaceae Calycophyllum acreanum 0.8 Rubiaceae Calycophyllum megistocaulum 0.72 Rubiaceae Calycophyllum spruceanum 0.605 Rubiaceae Capirona decorticans 0.593 Rubiaceae Chimarrhis glabriflora 0.716 Rubiaceae Chimarrhis hookeri 0.71 Rubiaceae Chomelia apodantha 0.557

111

Family Species Basic wood specific gravity (g cm-3) Rubiaceae Chomelia tenuiflora 0.557 Rubiaceae Cinchona calisaya 0.551 Rubiaceae Cinchona macrocalyx 0.545 Rubiaceae Cinchona micrantha 0.545 Rubiaceae Cinchona pubescens 0.54 Rubiaceae Condaminea corymbosa 0.619 Rubiaceae Coussarea brevicaulis 0.679 Rubiaceae Coussarea ecuadorensis 0.681 Rubiaceae Coussarea hirticalyx 0.645 Rubiaceae Coussarea klugii 0.677 Rubiaceae Coussarea paniculata 0.679 Rubiaceae Dialypetalanthus fuscescens 0.628 Rubiaceae Dioicodendron dioicum 0.614 Rubiaceae Elaeagia mariae 0.521 Rubiaceae Elaeagia myriantha 0.508 Rubiaceae Elaeagia utilis 0.418 Rubiaceae Faramea bangii 0.618 Rubiaceae Faramea candelabrum 0.649 Rubiaceae Faramea cf. verticillata 0.61 Rubiaceae Faramea glandulosa 0.584 Rubiaceae Faramea maynensis 0.584 Rubiaceae Faramea occidentalis 0.584 Rubiaceae Faramea tamberlikiana 0.618 Rubiaceae Faramea torquata 0.618 Rubiaceae Ferdinandusa chlorantha 0.723 Rubiaceae Ferdinandusa guainiae 0.723 Rubiaceae Genipa americana 0.64 Rubiaceae Guettarda crispiflora 0.42 Rubiaceae Guettarda tournefortiopsis 0.42 Rubiaceae Hillia parasitica 0.599 Rubiaceae Isertia laevis 0.52 Rubiaceae Ixora peruviana 0.628 Rubiaceae Joosia umbellifera 0.465 Rubiaceae Ladenbergia oblongifolia 0.615 Rubiaceae Macrocnemum roseum 0.45 Rubiaceae Palicourea amethystina 0.505 Rubiaceae Palicourea guianensis 0.54 Rubiaceae Palicourea lineata 0.505 Rubiaceae Palicourea stipularis 0.47 Rubiaceae Palicourea sulphurea 0.505

112

Family Species Basic wood specific gravity (g cm-3) Rubiaceae Palicourea tinctoria 0.505 Rubiaceae Pentagonia parvifolia 0.628 Rubiaceae Posoqueria coriacea 0.584 Rubiaceae Posoqueria latifolia 0.584 Rubiaceae Psychotria allenii 0.516 Rubiaceae Psychotria conephoroides 0.516 Rubiaceae Psychotria pichisensis 0.516 Rubiaceae Randia armata 0.69 Rubiaceae Rudgea cornifolia 0.57 Rubiaceae Rudgea verticillata 0.57 Rubiaceae Rustia rubra 0.628 Rubiaceae Schizocalyx obovatus 0.628 Rubiaceae Schizocalyx peruvianus 0.628 Rubiaceae Simira catappifolia 0.678 Rubiaceae Simira rubescens 0.8 Rubiaceae Warszewiczia coccinea 0.557 Rubiaceae Warszewiczia cordata 0.58 Rutaceae Galipea trifoliata 0.784 Rutaceae Metrodorea flavida 0.784 Rutaceae Zanthoxylum acreanum 0.586 Rutaceae Zanthoxylum caribaeum 0.76 Rutaceae Zanthoxylum ekmanii 0.586 Rutaceae Zanthoxylum rhoifolium 0.505 Rutaceae Zanthoxylum riedelianum 0.48 Rutaceae Zanthoxylum setulosum 0.385 Rutaceae Zanthoxylum sprucei 0.586 Rutaceae Zanthoxylum weberbaueri 0.586 Sabiaceae Meliosma boliviensis 0.794 Sabiaceae Meliosma frondosa 0.706 Sabiaceae Meliosma glabrata 0.502 Sabiaceae Meliosma herbertii 0.42 Sabiaceae Meliosma pumila 0.67 Sabiaceae Meliosma vasquezii 0.67 Salicaceae Banara guianensis 0.61 Salicaceae Casearia arborea 0.441 Salicaceae Casearia corymbosa 0.622 Salicaceae Casearia decandra 0.647 Salicaceae Casearia gossypiosperma 0.68 Salicaceae Casearia javitensis 0.753 Salicaceae Casearia megacarpa 0.656

113

Family Species Basic wood specific gravity (g cm-3) Salicaceae Casearia singularis 0.68 Salicaceae Casearia sylvestris 0.673 Salicaceae Casearia tachirensis 0.482 Salicaceae Casearia ulmifolia 0.596 Salicaceae Casearia zahlbruckneri 0.623 Salicaceae Hasseltia floribunda 0.58 Salicaceae Laetia corymbulosa 0.59 Salicaceae Laetia procera 0.648 Salicaceae Laetia suaveolens 0.64 Salicaceae Lunania parviflora 0.635 Salicaceae Tetrathylacium macrophyllum 0.62 Salicaceae Xylosma intermedia 0.77 Sapindaceae Allophylus divaricatus 0.559 Sapindaceae Allophylus floribundus 0.39 Sapindaceae Allophylus glabratus 0.491 Sapindaceae Allophylus loretensis 0.491 Sapindaceae Allophylus pilosus 0.491 Sapindaceae Allophylus punctatus 0.531 Sapindaceae Allophylus scrobiculatus 0.53 Sapindaceae Cupania rubiginosa 0.732 Sapindaceae Cupania scrobiculata 0.628 Sapindaceae Matayba arborescens 0.697 Sapindaceae Matayba guianensis 0.729 Sapindaceae Matayba macrocarpa 0.771 Sapindaceae Matayba purgans 0.771 Sapindaceae Matayba scrobiculata 0.729 Sapindaceae Pseudima frutescens 0.8 Sapindaceae Sapindus saponaria 0.743 Sapindaceae Talisia cerasina 0.827 Sapindaceae Talisia cupularis 0.827 Sapindaceae Talisia hexaphylla 0.827 Sapindaceae Talisia mollis 0.827 Sapindaceae Toulicia reticulata 0.608 Sapotaceae Chrysophyllum acreanum 0.671 Sapotaceae Chrysophyllum amazonicum 0.826 Sapotaceae Chrysophyllum argenteum 0.784 Sapotaceae Chrysophyllum lucentifolium 0.787 Sapotaceae Chrysophyllum manaosense 0.75 Sapotaceae Chrysophyllum ovale 0.787 Sapotaceae Chrysophyllum pomiferum 0.766

114

Family Species Basic wood specific gravity (g cm-3) Sapotaceae Chrysophyllum sanguinolentum 0.671 Sapotaceae Chrysophyllum venezuelanense 0.787 Sapotaceae Diploon cuspidatum 0.85 Sapotaceae Ecclinusa guianensis 0.627 Sapotaceae Ecclinusa lanceolata 0.581 Sapotaceae Ecclinusa lancifolia 0.655 Sapotaceae Ecclinusa ramiflora 0.645 Sapotaceae Elaeoluma nuda 0.784 Sapotaceae Manilkara bidentata 0.87 Sapotaceae Manilkara inundata 0.8 Sapotaceae Manilkara surinamensis 0.868 Sapotaceae Micropholis egensis 0.6 Sapotaceae Micropholis guyanensis 0.657 Sapotaceae Micropholis melinoniana 0.53 Sapotaceae Micropholis venulosa 0.67 Sapotaceae Pouteria bangii 0.758 Sapotaceae Pouteria bilocularis 0.708 Sapotaceae Pouteria boliviana 0.758 Sapotaceae Pouteria caimito 0.784 Sapotaceae Pouteria cladantha 0.942 Sapotaceae Pouteria coriacea 0.895 Sapotaceae Pouteria cuspidata 0.9 Sapotaceae Pouteria durlandii 0.69 Sapotaceae Pouteria ephedrantha 0.758 Sapotaceae Pouteria eugeniifolia 1.113 Sapotaceae Pouteria franciscana 0.6 Sapotaceae Pouteria guianensis 0.93 Sapotaceae Pouteria juruana 0.758 Sapotaceae Pouteria krukovii 0.758 Sapotaceae Pouteria lucentifolia 0.758 Sapotaceae Pouteria macrophylla 0.737 Sapotaceae Pouteria pariry 0.758 Sapotaceae Pouteria plicata 0.728 Sapotaceae Pouteria procera 0.758 Sapotaceae Pouteria rata 0.758 Sapotaceae Pouteria reticulata 0.794 Sapotaceae Pouteria sagotiana 0.865 Sapotaceae Pouteria simulans 0.758 Sapotaceae Pouteria tarapotensis 0.758 Sapotaceae Pouteria torta 0.692

115

Family Species Basic wood specific gravity (g cm-3) Sapotaceae Pouteria trilocularis 0.672 Sapotaceae Sarcaulus brasiliensis 0.615 Scrophulariaceae Buddleja montana 0.498 Simaroubaceae Simaba paraensis 0.41 Simaroubaceae Simaba polyphylla 0.41 Simaroubaceae Simarouba amara 0.383 Siparunaceae Siparuna aspera 0.648 Siparunaceae Siparuna bifida 0.662 Siparunaceae Siparuna crassiflora 0.662 Siparunaceae Siparuna cuspidata 0.655 Siparunaceae Siparuna decipiens 0.641 Siparunaceae Siparuna subinodora 0.662 Siparunaceae Siparuna tomentosa 0.648 Solanaceae Cestrum conglomeratum 0.505 Solanaceae Cestrum megalophyllum 0.457 Solanaceae Markea ulei 0.482 Solanaceae Saracha punctata 0.563 Solanaceae Sessea dependens 0.659 Solanaceae Solanum aphyodendron 0.447 Solanaceae Solanum endopogon 0.28 Solanaceae Solanum maturecalvans 0.447 Solanaceae Solanum nutans 0.447 Staphyleaceae Turpinia occidentalis 0.547 Strelitziaceae Phenakospermum guyannense 0.594 Styracaceae Styrax foveolaria 0.527 Styracaceae Styrax pentlandianus 0.527 Symplocaceae Symplocos arechea 0.422 Symplocaceae Symplocos baehnii 0.62 Symplocaceae Symplocos fuliginosa 0.403 Symplocaceae Symplocos mezii 0.59 Symplocaceae Symplocos psiloclada 0.601 Symplocaceae Symplocos quitensis 0.518 Symplocaceae Symplocos reflexa 0.579 Symplocaceae Symplocos spruceana 0.59 Tapisciaceae Huertea glandulosa 0.429 Theaceae Gordonia fruticosa 0.551 Theaceae Gordonia pubescens 0.518 Ulmaceae Ampelocera edentula 0.7 Ulmaceae Ampelocera ruizii 0.648 Ulmaceae Ampelocera verrucosa 0.726

116

Family Species Basic wood specific gravity (g cm-3) Urticaceae Cecropia angustifolia 0.505 Urticaceae Cecropia engleriana 0.49 Urticaceae Cecropia ficifolia 0.266 Urticaceae Cecropia latiloba 0.33 Urticaceae Cecropia membranacea 0.33 Urticaceae Cecropia polystachya 0.312 Urticaceae Cecropia sciadophylla 0.374 Urticaceae Cecropia tessmannii 0.312 Urticaceae Coussapoa ovalifolia 0.462 Urticaceae Coussapoa trinervia 0.462 Urticaceae Coussapoa villosa 0.616 Urticaceae Myriocarpa stipitata 0.269 Urticaceae Pourouma bicolor 0.331 Urticaceae Pourouma cecropiifolia 0.356 Urticaceae Pourouma cucura 0.394 Urticaceae Pourouma cuspidata 0.394 Urticaceae Pourouma guianensis 0.38 Urticaceae Pourouma herrerensis 0.379 Urticaceae Pourouma minor 0.427 Urticaceae Pourouma mollis 0.328 Urticaceae Pourouma palmata 0.394 Urticaceae Pourouma substrigosa 0.394 Urticaceae Pourouma tomentosa 0.395 Urticaceae Urera baccifera 0.18 Urticaceae Urera capitata 0.413 Urticaceae Urera caracasana 0.18 Urticaceae Urera lianoides 0.18 Urticaceae Urera simplex 0.18 Violaceae Gloeospermum sphaerocarpum 0.665 Violaceae Leonia crassa 0.65 Violaceae Leonia glycycarpa 0.6 Violaceae Leonia racemosa 0.633 Violaceae Rinorea apiculata 0.668 Violaceae Rinorea flavescens 0.63 Violaceae Rinorea guianensis 0.78 Violaceae Rinorea lindeniana 0.68 Violaceae Rinorea pubiflora 0.75 Violaceae Rinorea viridifolia 0.515 Violaceae Rinoreocarpus ulei 0.665 Vochysiaceae Qualea paraensis 0.672

117

Family Species Basic wood specific gravity (g cm-3) Vochysiaceae Vochysia biloba 0.476 Vochysiaceae Vochysia kosnipatae 0.476 Vochysiaceae Vochysia leguiana 0.476 Vochysiaceae Vochysia majuscula 0.459

118

Appendix II -Table S2. Statistical moments of mean wood density (g cm-3) distribution on species and stem levels for 41 permanents

plots across the Andes-to-Amazon elevational gradient.

Species level Stem level

Elevation All arborescent life forms Trees All arborescent life forms Trees Plot (m) mean sd skewness kurtosis mean sd skewness kurtosis mean sd skewness kurtosis mean sd skewness kurtosis

APK-01 3625 0.654 0.084 -0.415 -0.984 0.654 0.084 -0.415 -0.984 0.653 0.067 0.085 -1.355 0.653 0.067 0.085 -1.355

ACJ-01 3537 0.635 0.071 0.360 -0.250 0.635 0.071 0.360 -0.250 0.596 0.065 0.451 -0.862 0.596 0.065 0.451 -0.862

TRU-01 3450 0.589 0.115 -0.798 0.065 0.603 0.101 -0.835 0.756 0.600 0.069 -1.114 2.390 0.604 0.061 -0.676 1.195

119 TRU-02 3250 0.567 0.113 -0.480 -0.267 0.591 0.091 -0.340 0.507 0.600 0.103 -0.876 0.529 0.611 0.091 -0.878 1.211

TRU-03 3000 0.587 0.122 -0.618 -0.061 0.612 0.098 -0.529 1.002 0.650 0.088 -0.596 2.926 0.658 0.074 0.368 1.683

WAY-01 3000 0.589 0.107 -0.515 -0.022 0.602 0.093 -0.286 -0.015 0.628 0.066 -1.046 3.475 0.629 0.065 -0.944 3.229

ESP-01 2890 0.573 0.125 -0.424 -0.253 0.594 0.108 -0.444 0.621 0.614 0.089 -1.591 2.815 0.628 0.066 -1.238 3.756

TRU-04 2750 0.562 0.117 -0.576 -0.328 0.594 0.090 -0.607 1.382 0.561 0.130 -0.566 -0.969 0.618 0.078 -0.546 0.386

TRU-05 2500 0.558 0.146 0.220 -0.166 0.601 0.122 0.385 0.952 0.463 0.136 0.728 -0.772 0.586 0.098 0.056 0.195

TRU-06 2250 0.518 0.136 0.014 -0.787 0.559 0.120 -0.348 0.336 0.456 0.137 0.807 -0.768 0.556 0.124 -0.228 -0.699

TRU-07 2000 0.561 0.123 -0.401 -0.379 0.584 0.106 -0.492 0.619 0.564 0.139 -0.330 -1.128 0.623 0.096 -0.496 0.148

TRU-08 1800 0.562 0.123 0.105 -0.305 0.577 0.114 0.169 -0.040 0.547 0.143 0.189 -1.057 0.599 0.116 0.244 -1.014

SPD-01 1750 0.560 0.129 0.553 0.856 0.571 0.122 0.704 1.180 0.521 0.144 0.086 -0.398 0.569 0.115 0.255 0.775

CAL-01 1500 0.518 0.118 0.160 1.152 0.521 0.118 0.131 1.251 0.517 0.118 -0.359 0.854 0.517 0.118 -0.375 0.895

Species level Stem level

Elevation All arborescent life forms Trees All arborescent life forms Trees Plot (m) mean sd skewness kurtosis mean sd skewness kurtosis mean sd skewness kurtosis mean sd skewness kurtosis

SAI-01 1500 0.552 0.119 0.335 0.950 0.559 0.115 0.410 1.160 0.553 0.119 0.448 0.111 0.559 0.117 0.472 0.143

SPD-02 1500 0.538 0.127 0.276 1.167 0.546 0.123 0.268 1.515 0.499 0.124 0.535 0.649 0.513 0.119 0.538 0.949

CAL-02 1250 0.534 0.140 0.415 1.792 0.536 0.138 0.461 1.889 0.545 0.111 -0.285 1.676 0.545 0.110 -0.271 1.686

SAI-02 1250 0.553 0.121 0.050 0.784 0.563 0.115 0.115 1.211 0.497 0.147 0.020 -0.217 0.555 0.112 0.354 1.289

TON-02 1000 0.549 0.132 0.464 0.258 0.558 0.126 0.606 0.283 0.529 0.118 0.514 0.523 0.534 0.115 0.605 0.563

PAN-03 850 0.616 0.140 0.057 -0.620 0.620 0.138 0.082 -0.610 0.622 0.150 0.310 -0.085 0.622 0.150 0.316 -0.075

TON-01 800 0.594 0.135 0.358 -0.037 0.601 0.131 0.402 0.051 0.586 0.134 0.591 0.472 0.589 0.132 0.621 0.543

120

PAN-02 595 0.606 0.148 0.371 0.005 0.608 0.146 0.429 -0.020 0.577 0.171 0.763 -0.053 0.577 0.171 0.770 -0.053

PAN-01 425 0.564 0.143 0.265 -0.226 0.570 0.138 0.381 -0.271 0.551 0.182 0.625 0.041 0.577 0.168 0.849 0.150

ALM-01 400 0.591 0.153 -0.349 -0.220 0.599 0.148 -0.355 -0.105 0.542 0.164 -0.232 -0.690 0.580 0.143 -0.329 -0.106

MNU-08 400 0.594 0.139 -0.114 -0.294 0.599 0.136 -0.116 -0.261 0.489 0.147 0.061 -0.438 0.550 0.120 0.274 0.079

BAB-01 387 0.574 0.151 -0.285 0.007 0.582 0.146 -0.280 0.181 0.521 0.168 -0.062 -0.650 0.563 0.149 -0.127 -0.168

MNU-04 358 0.587 0.145 -0.303 -0.290 0.593 0.140 -0.283 -0.253 0.541 0.150 -0.348 -0.545 0.580 0.128 -0.479 0.197

MNU-05 347 0.578 0.134 -0.237 -0.354 0.582 0.132 -0.232 -0.316 0.536 0.131 0.176 -0.230 0.551 0.128 0.208 -0.390

MNU-06 345 0.571 0.135 -0.045 -0.348 0.578 0.132 -0.042 -0.287 0.517 0.140 -0.086 -0.482 0.559 0.121 0.004 -0.355

MNU-03 312 0.582 0.144 -0.153 -0.274 0.588 0.142 -0.155 -0.232 0.518 0.149 -0.251 -0.579 0.552 0.131 -0.332 0.073

TAM-07 225 0.617 0.145 -0.281 -0.492 0.619 0.142 -0.205 -0.605 0.602 0.146 0.024 -0.893 0.602 0.145 0.069 -0.930

Species level Stem level

Elevation All arborescent life forms Trees All arborescent life forms Trees Plot (m) mean sd skewness kurtosis mean sd skewness kurtosis mean sd skewness kurtosis mean sd skewness kurtosis

TAM-05 220 0.620 0.148 0.079 -0.319 0.623 0.145 0.145 -0.356 0.600 0.149 0.127 -0.604 0.606 0.144 0.227 -0.676

TAM-08 220 0.603 0.142 -0.054 -0.822 0.606 0.139 0.009 -0.889 0.583 0.151 -0.372 -0.287 0.614 0.123 -0.002 -0.261

TAM-01 205 0.609 0.147 -0.152 -0.271 0.616 0.140 0.026 -0.474 0.504 0.187 0.008 -1.180 0.601 0.131 -0.111 -0.039

TAM-02 201 0.593 0.142 -0.110 -0.367 0.599 0.137 -0.010 -0.424 0.520 0.174 -0.147 -1.109 0.604 0.123 -0.346 0.112

TAM-06 200 0.581 0.143 -0.051 -0.171 0.591 0.138 -0.023 -0.102 0.484 0.175 -0.012 -1.072 0.574 0.125 -0.150 0.275

TAM-09 197 0.610 0.144 -0.088 -0.205 0.614 0.139 0.036 -0.315 0.567 0.167 -0.427 -0.530 0.619 0.122 -0.119 0.032

CUZ-01 190 0.558 0.164 -0.291 -0.007 0.567 0.160 -0.337 0.227 0.502 0.169 -0.378 -0.472 0.518 0.168 -0.594 -0.125

121

CUZ-02 190 0.571 0.143 -0.208 -0.162 0.579 0.137 -0.151 -0.076 0.529 0.149 -0.414 -0.542 0.585 0.105 -0.245 0.987

CUZ-03 190 0.588 0.151 -0.238 -0.243 0.595 0.147 -0.256 -0.081 0.551 0.150 -0.485 -0.249 0.593 0.118 -0.505 1.287

CUZ-04 190 0.585 0.137 -0.076 0.264 0.593 0.130 0.054 0.313 0.576 0.136 -0.027 -0.021 0.607 0.118 0.119 0.424

APPENDIX II – FIGURE S1

122

APPENDIX II – FIGURE S2

(a) (b)

123

APPENDIX II – FIGURE S3

(a) (b) 0.3 % 0.0 % 18.9 % 33.2 % 11.6 %

28.5 %

47.1 %

26.9 %

14.0 % 19.4 %

124

APPENDIX II – FIGURE S4

125

APPENDIX I APPENDIX

126

–

FIGURE

APPENDIX II – FIGURE S6

(a)

(b)

127

APPENDIX II – FIGURE S7

128

CHAPTER III

MOVEMENT OR MORTALITY? PERVASIVE BUT SLOW UPSLOPE TREE

MIGRATION IN THE AMAZON TO THE ANDES REVEALED THROUGH 38

YEARS OF FOREST MONITORING

Abstract

Climate change is thought to be causing species distributional shifts along environmental gradients leading to novel species assemblages, but evidence for those shifts in the Andes is scarce and is almost nonexistent for the Amazon. Based on permanent forest plots with data of 41,352 stems and 1,896 tree species during 38 years of monitoring, we assess the community-level shift in species composition towards taxa that have warmer mean ranges (community thermophilization) and species-level distributional shifts (thermal migration) along a 3500 m elevational gradient in Eastern

Peru. Results showed that the Amazonian tree community’s response to climate change is weak to nonexistent, with a mean migration rate of -0.00021 °C yr-1 based on individuals and +0.00024 °C yr-1 based on basal area. For Andean trees, migration was slow but present, at least on average, with +0.0062 °C yr-1 to +0.0044 °C yr-1 in terms of individuals and basal area basis respectively. However, “migration” or

“thermophilization” results are mainly explained by increases in tree mortality at warmer parts of their ranges rather than “true” shifts of entire ranges upslope. The observed

129 migration was also variable and heterogeneous across both elevation and time for thermophilization, with rates in single plots ranging from -0.0207 to +0.0257 °C yr-1 across years, again with much of the thermal migration being explained by mortality of taxa at the warm end of their ranges in response to episodic drought. Species responses varied across space and time as well, with rates ranging from -0.0319 to +0.0355 °C yr-1, demonstrating the idiosyncratic response of each taxon to warming and drought events.

The observed lag in tree migration is out of equilibrium with current climate change increasing the “climatic debt” or migration that must happen for species to remain in equilibrium with climate in tropical forests.

Keywords: Climate change, community assembly, thermal niches, range shifts, tropical biodiversity, species migration, global warming

130

Introduction

Changes in species ranges in response to climate is one of the central processes in population biology. Paleoecological studies have shown vast plant species migrations across environmental gradients, and wholescale reassortment of communities due to climate changes (Davis 1969, Jackson and Overpeck 2000, Williams and Jackson 2007).

A key question to understand plant responses to ongoing global warming is how to reconcile the millennial-scale results that we see in the paleoecological record—the pattern of migration—with current population-level processes observed on annual to decadal time scales. For instance, processes such as changes in the distributional range of specific plant species (species migration), or the shift in community composition due to the replacement of species with cooler tolerances with those with warm tolerances at any point, a process called thermophilization (De Frenne et al. 2013). At all-time scales plant species ranges change through differential recruitment and mortality rates along environmental gradients or in space, and the process of species migration and community thermophilization result from the integration of these population demography processes across the species that form the local community, and understanding ongoing changes and linking them to longer-term population trajectories requires a demographic approach, first advocated by Harper in

1967.

Atmospheric warming is currently pushing species out of their fundamental niche

(Lenoir et al. 2008). In response to such rapid changes, plant species are predicted to shift their distributions to cooler environments to remain within their thermal niches [“species migration” including range expansion, contraction, and shift (Parmesan et al. 1999, Feeley

2012, Svenning and Sandel 2013)]. The current velocity of climate change is causing rapid

131 shifts in species distributions and community composition due to changes in the abundances of populations (Loarie et al. 2009, De Frenne et al. 2013). For instance, studies on plants migrations in temperate an tropical regions have shown that current plant assemblages are shifting their ranges upslope in response to global warming (Gottfried et al. 2012, Morueta-Holme et al. 2015, Freeman et al. 2018). However, a smaller set has looked at plants species migrations exclusively for tree communities in temperate

(Monleon and Lintz 2015, Savage and Vellend 2015) and tropical (Feeley et al. 2013,

Fadrique et al. 2018) regions, though the time-scales covered and inventory sizes provide limited understanding of global forest responses to global warming.

Tropical regions have warmed at an average of 0.26 °C per decade since the mid-

1970s (Malhi and Wright 2004). And, in the eastern slopes of the Andes specifically, temperature has increased approximately 0.10 - 0.11 °C/decade since 1939 (Vuille and

Bradley 2000) and is predicted to increase 2 - 7 °C this century (Urrutia and Vuille 2009), making the pace of climate change higher than any time in the last 50,000 years (Bush et al. 2004) and presenting unprecedented challenges to western Amazonian and Andean plant species. The speed and magnitude of global warming raise critical questions about how Amazonian and Andean forests will respond to climate change. For instance, there has been a substantial advance in our understanding about drivers of forest dynamics and species compositional change in the Amazonian regions (Phillips and Gentry 1994, ter

Steege et al. 2006, Gomes et al. 2018) and, to some extent, the responses of lowland forests to drought events (Esquivel-Muelbert et al. 2018), but there are few modern data on the effects of temperature changes on these systems. Second, we know little about the drivers of forest dynamics and species composition or even basic ecology for Andean montane

132 forests, however, that is where most of the temperature change studies have been conducted

(Feeley et al. 2011, Duque et al. 2015, Fadrique et al. 2018). Third, studies of Amazonian and Andean forests have drawn an artificial line at 500 m (ter Steege et al. 2006) to separate the two forest types, severing what is currently and has been shown to be in the paleo- record to be a continuum of both species distributions and migrating populations.

Evidence of range shifts to date

The prediction of modern shifts in tropical plant distributions to cooler environments in response to climate change (Foster 2001, Bush 2002, Colwell et al. 2008) is supported by an increasing number of studies documenting community tree migrations in Andean and Mesoamerican montane ecosystems (Feeley et al. 2013, Duque et al. 2015,

Fadrique et al. 2018). However, these results are based mainly on vegetation surveys of forests over ~500 m, whose conclusions could be strongly influenced by the exclusion of populations from the lower elevational range limit of the Andean tree species that can reach the Amazonian regions or, by the exclusion of Amazonian tree species populations that can reach the Andean regions. For example, 39 % of the ~700 species of trees >10 cm DBH known from the lowlands of SE Peru have upper ranges that exceed 500 m (data from this study). Furthermore, tree gamma diversity in the Western Amazon is highest in the ranges straddling 300 - 1200 m, tree community alpha diversity stays approximately constant from the lowlands to ~1700 m, and “Amazonian” tree species routinely have populations that extend fluidly from the lowlands into the Andes, with several lowland dominants extending up to 1500 m in elevation (Gentry 1988, Silman 2007). Because of this, evaluation of

133 tropical forest responses to climate change and improvements in predictions of future responses requires the integration of long-term studies of Amazonian-Andean tree communities across temperature gradients, with the lack of studies linking the Amazonian and Andean forests representing a major gap in the understanding of the mechanisms underlying forest responses to climate change.

Attribution of the cause of distributional change in response to climate change is also difficult. For example, the observed widespread Andean “plant migration” has been interpreted in terms of temperature change (Fadrique et al. 2018), even though we know from the Amazon and also from the cloud base studies (Pounds et al. 1999) that moisture regimes can have a profound effects on forest dynamics, particularly through accelerated tree mortality (Farfan Rios 2011, Esquivel-Muelbert et al. 2018). If these mortality events were to be concentrated in only part of a range across a temperature gradient it would indicate a thermal migration, even though the primary cause was drought, and its linkage to temperature is through differential strengths of drought response with respect to temperature. Furthermore, given the predicted interaction of temperature and drought responses, climate change effects on tropical forests are predicted to happen unevenly in time, as mortality is a fast response, but recruitment, particularly into canopy tree sizes, can take decades to centuries. This is particularly true during drought events where it is possible to observe pulsed tree mortality episodes. All of this makes long-term data central to understanding the tempo of forest responses to climate change.

This study seeks to understand population and community changes with respect to temperature changes—tree migration and community thermophilization—in the Amazon-

Andes. We used one of the world’s largest elevational transect located in the eastern slope

134 of the Peruvian Andes, spanning from the lowland western Amazonian forests (190 m) to the Andean treeline (3700 m). The elevational transect has a comprehensive and permanent forest monitoring network spanning the last 38 years and is probably the best representation of mature Amazonian-Andean forests, providing a synoptic view of tropical forest responses to climate change. Here we ask questions about the pace, causes, and underlying demographic basis of tropical forest responses to temperature changes. Specifically: (1) what is the temporal pattern of tree community thermophilization, and how does this vary across the gradient? (2) As thermophilization is at its core a demographic (birth-death- growth) process, what are the contributions of these elements to observed thermophilization? And (3) migration is a population process, but in the paleoecological record, and in all modern studies to date, it has been examined at higher taxonomic levels

(genus, family). Do we observe individual species migration, and how does it compare with observations taken at the genus level? We focus particularly on episodic events in the climate record and ask whether droughts and other events can lead to thermophilization and apparent thermal migration.

Methods

Study site

The study was performed on the eastern slope of the Peruvian Andes along an elevational gradient extending from the Andean treeline at 3700 m to the Amazon basin at 190 m. The elevational gradient lies in unbroken mature forest ranging from 3700 m to 300 m in the Manu Biosphere Reserve (11.8564° S, 71.7214° W) and extending to 190

135 m in the Tambopata National Reserve (12.9206° S, 69.2819° W), hereafter we will refer this gradient as the Manu-Tambopata elevational transect. Most of the montane plots

(elevations 3700m to 425m) are within 60 km of each other, while the lowland plots extend across the lowlands at the base of the Andes with a maximum inter-plot distance of 250 km. Mean Annual Temperature decreases linearly with increasing elevation along the gradient at a lapse rate of 5.5 o C km-1, with mean annual temperatures ranging from ~

26.6 o C at the lowest elevations to ~ 6.4 o C at treeline (Rapp and Silman 2012, Malhi et al. 2016). Mean annual precipitation varies across the gradient from 2448 to 5500 mm yr-

1, with significant inter-annual variability throughout (Rapp and Silman 2012, Malhi et al.

2016). Additionally, in the study area it has been observed that temperature increases have averaged approximately 0.03 - 0.05 o C y-1 since the late 1950s (Feeley et al. 2011).

Inventory forests plots: Ground data

Plot data were collected from 40 permanent inventory forest plots totaling 46.5 ha across a continuous elevation gradient, extending from lowland forest (≤ 500 m), through sub montane (500-1600 m), lower montane (1600-2500 m), upper montane (2500-3400 m) and treeline (≥ 3400 m) forests types (Young 1992, Pennington et al. 2004). Twenty three 1-ha permanent plots were established and are maintained by the Andes

Biodiversity and Ecosystem Research Group – ABERG

(http://www.andesconservation.org/) and range from 400 to 3625 m of elevation and in age from 2003 to 2017. Additionally, 17 (23.5 ha) permanents plots were established by various investigators, particularly John Terborgh, Percy Nunez, and Alwyn Gentry; and

136 are currently monitored by the Amazon Forest Inventory Network – RAINFOR

(http://www.rainfor.org/), the Global Ecosystem Monitoring Network – GEM

(http://gem.tropicalforests.ox.ac.uk/) and the Cocha Cashu Biological Station – CCBS

(https://cochacashu.sandiegozooglobal.org/) in the lowland forest, ranging from 190 - 405 m of elevation and 1979 to 2014 in age. The forest plots were established and remeasured multiple times following standardized protocols over 40 years (Phillips et al. 2016). In our plot network, 38 of the 40 permanent plots were censused at least three times between

1979 and 2017 (38 years). Excluding the TON-01 and PAN-01 plots (plots with one census), the number of multiple plot censuses varies from 3 to 13 times over the 38 years

[average number of censuses = 6.16 (95% CI = 5.3-7.0), median number of censuses =

5). The oldest plot was established in 1979 in the Tambopata terra firme rain forest and has the highest number of censuses (n = 13) (Table S1). Overall, the permanent plots contain 41,352 stems greater than 10 cm diameter at breast height (dbh) and encompass

1,896 arborescent species and morphospecies including trees, tree ferns, palms, and lianas > 10 cm dbh (hereafter, trees) involving more than 174,000 tree measurements.

Plant identifications

All botanical collections taken from the permanent plots from the ABERG and

RAINFOR networks were identified in situ and in different herbaria, then were compared and standardized. The vouchers were deposited in the Peruvian and USA herbaria (CUZ,

HOXA, HUT, MOL, USM, and DAV, MO, F, WFU respectively). Additionally, local flora and plant checklists were used as references (Cano et al. 1995, Pennington et al.

137

2004, Farfan-Rios et al. 2015, Vasquez M. and Rojas G. 2016) and plant identifications were also confirmed by taxonomic experts. We then combined and standardized the species names from all the permanent plots (n = 40). The combined species list was submitted to the Taxonomic Name Resolution Service (TRNS version 4.0, http://tnrs.iplantcollaborative.org/) online application to standardize and validate the species names (Boyle et al. 2013). All taxa with homogenized morphospecies [e.g. sp1(5984WFR)] and invalid names (e.g. “indet”) were assigned as an “undetermined”.

We followed the APG IV plant classification for the valid species names (Chase et al.

2016). All TNRS “accepted” species names with an overall TNRS-score below 0.9 were manually reviewed and the names were confirmed on The Plant List

(http://www.theplantlist.org/) and Tropicos (http://www.tropicos.org) database.

However, if the specific epithet was not confirmed, we used the valid genus names as a unique species identifier. Species with an unassigned accepted TNRS name (e.g.

“invalid”, “illegitimate” or “no opinion”) were also manually reviewed and the species names were corrected following the previous criteria. Unidentified taxa at the genus-level were excluded from the analysis.

General trends in plot-to-plot diversity and structure along the gradient

Along the Manu-Tambopata elevational transect, we observed that lowland species richness (~180 spp. ha-1) is maintained to ~1700 m of elevation and decreases abruptly above this elevation up to the treeline (SI Appendix, Fig. S1a). The number of species ranges from a high of 179 species in our plot at 405 m of elevation (ALM-01), to

138 a low of 17 species in our highest elevation plot (APU-01) at treeline. The number of individuals per plot greater than 10 cm diameter at breast height (DBH) ranged from 447

(TON-02) in the sub montane forest to 1262 (TRU-08) in the lower montane forest and overall, the number of individuals on a plot-to-plot basis showed a non-linear relationship with elevation with high stem density between the lower and upper montane forest

(Farfan 2011; SI Appendix, Fig. S1b). The observed high individual density at the lower and upper montane forests is explained by the high abundance of tree ferns (Farfan Rios

2011). Basal area did not show a relationship with elevation, indicating that the high stem-abundance in Andean montane forests compensates for the larger maximum tree size in Amazonian forests (SI Appendix, Fig. S1c).

Species thermal distributions and the estimated thermal optimum

We used established protocols to estimate species thermal distributions for all arboreal species occurring in the 40 permanent forests plots across the elevational gradient (Feeley et al. 2011). For all species occurring in the plots, we downloaded all available georeferenced herbarium records from the Andean-Amazonian countries (i.e.

Bolivia, Colombia, Ecuador, and Peru) through the Global Biodiversity Information

Facility data portal (GBIF: http://www.gbif.org). Plant records that were missing coordinates, records that were tagged by the GBIF as a having coordinate errors or that had evident georeferenced errors (e.g. falling in large bodies of water), and duplicate records were discarded. The mean annual temperatures (MAT) of all specimens were calculated at the collection locations by extracting the temperature values from the

139

WorldClim extrapolated climate map at a spatial resolution of 30 arc seconds (Hijmans et al. 2005). We estimated the thermal optimum as the mean MAT (o C) for all the collections locations for each species represented by ≥ 10 herbarium collection records.

For species with < 10 available records, the thermal optimum was estimated as the average collection temperature calculated from all available records of congeneric individuals collected from the tropical Andean-Amazonian region (Feeley et al. 2011).

Community temperature index (CTI) and thermophilization rates (TR)

In this study, we integrated the western Amazonian forests permanent plots below

500 m to include the lower limit populations of Andean tree species (down to 190 m) along their elevational range. If we restrict the analysis of “Andean” taxa (by truncating the analysis at 500 m), we exclude the lower limit populations of 176 Andean species, and therefore their entire geographical distributions along the elevational with the reciprocal being true, excluding the upper range of 98 of our “Amazonian” species along the gradient (SI Appendix, Fig. S2). We calculated the community temperature index

(CTI) using two different approaches. First, we used individual-weighted approach because changes in demography are based on individuals dying (mortality) and living

(recruitment). Second, basal area-weighted approach, because tree size is ecologically important, for instance, the death of a large tree will affect more the ecosystem function than the death of a small tree. In addition, large trees are potentially more closely tied to their climate optimum than smaller trees.

140

CTI was calculated for each forest plot (n = 40 plots) as the average thermal optimum of the arboreal species weighted by the number of individuals and weighted by the relative total basal area at breast height (summed cross-sectional area at 1.3 m above ground) in the respective plot (Feeley et al. 2011). In the CTI individual-weighted approach we exclude individuals with multiple-stems below breast height and only consider the principal stem because changes in CTI integrates the effects of individual tree demography (changes in mortality, recruitment and growth rates). On the other hand, the CTI basal area-weighted approach was calculated combining the basal area of individuals including their multiple stems to give a total basal area for each individual tree. The annual changes in CTI values at the individual- and basal area-weighted approach were then calculated for each forest plot with more than three remeasurements

(n = 38 plots) over all possible consecutive census intervals (n = 234 census intervals).

We observed that trees that died in landslide events caused significant changes in CTI over time for the affected plots (e.g. SPD-01; Fig. S3). Landslides are inherent in the forest dynamics on montane ecosystems and kill a large proportion of trees in short period time. Accordingly, we excluded the trees that died in landslides (outliers) since their started time that the plot was established. The new calculated CTI values were then used to estimate the thermophilization rates (TR) and test the thermophilization hypothesis.

Thermophilization rate (°C) was calculated as the net change in CTI values for

-1 each plot overall consecutive censuses and the overall annualized TRplot (°C yr ) was calculated for each plot as the slope of the linear least-square regression between CTI and the census year, at the individual- and basal area-weighted approach. We also calculated

141

TRplot for each plot only for the tree basal area growth (TRgro), tree basal area recruitment

(TRrec), and tree basal area mortality (TRmor). The TRgro of a plot is the difference in the plot’s CTI as calculated using the initial vs. final basal areas of just the trees surviving through the census period. The TRrec of a plot is the difference in its CTI as calculated using the final basal areas of all stems recorded at the end of the census interval vs. its

CTI as calculated using the final basal areas of just the stems that survived through the census interval. Finally, TRmor will be the difference in its CTI as calculated using the initial basal areas of all stems recorded in the first census vs. its CTI as calculated using the initial basal areas of just the stems that survived through the census interval (Feeley et al. 2013). To test the contributions of TRmor, TRrec and TRgro in changes of TRplot we used generalized linear models (GLMs) using the glm function in R.

Thermal migration rates (TMR) at genus- and species-level

Changes in TRplot is the result of changes in the relative abundance of different taxa along the elevational gradient. Because of this, we also calculated the thermal migration rates (TMR) of dominant taxa at the genus- and species-level to test the changes in the mean temperatures between censuses along the gradient. TMRgenus and

TMRspecies were calculated for all tree genera and species that occurred in at least two of the 38 forest plots and had more than 50 individuals at the initial census along the gradient. For each genus and species that met the criterion, TMR was calculated as the slope of the linear least-square regression between the average of the mean temperatures and the census year weighted by the number of individuals and by the relative basal area

142 within genera and species across plots. We used a binomial probability test to determine whether the number of plots, genera, and species with positive TRplot, TMRgenus, and

TMRspecies values differ significantly from the null expectations of no change in TR and

TMR over time.

Results

Thermophilization of tree communities along the Amazonian-Andean gradient

The average community temperature index (CTI) ranged from 13 to 25 °C and was highly negatively correlated with plot elevation at individual- and basal area-weighted approach, demonstrating the efficacy of the methodology and the relationship between tree-reported temperature and measured temperature, and the importance of the thermal niche in controlling species distributions (Individual-weighted: r = -0.98, P ˂ 0.001; basal area-weighted: r = -0.98, P ˂ 0.001; SI Appendix, Fig. S3). We observed that changes in

CTI varied over time across the plots and along the gradient, however, with some of the plots showing a consistent positive (e.g. TAM-02) or negative (e.g. CUZ-01) CTI trajectory since 1979 for both, individual- and basal area-weighted approach, while others were mixed (Fig. 1a, b; SI Appendix, Fig. S5a, S5b).

Annualized rates of thermophilization varied widely among the plots along the

Amazon to the Andes elevation gradient, ranging from -0.0208 to +0.0257 °C yr-1 in terms of number of individuals and from -0.0113 to +0.0204 °C yr-1 in terms of basal area (Fig.

2a, b; SI Appendix, Table S2). The overall annual community thermophilization rate

-1 (TRplot) across the plots for the entire gradient was +0.0035 °C yr (95% CI = +0.0006 -

143

+0.0065 °C yr-1) for the individual-weighted approach and +0.0023 °C yr-1 (95% CI =

+0.00003 - +0.0047 °C yr-1) for the basal area-weighted approach. Assuming the adiabatic lapse rate of 5.5 °C km−1 (Bush et al. 2004), plot-level mean upward migration corresponds to +0.64 m yr-1 (95% CI = +0.12 - +1.17 m yr-1) and +0.43 m yr-1 (95% CI = +0.004 - +0.85 m yr-1) for individuals- and basal area-weighted basis respectively (SI Appendix, Table

S2). We did not observe a correlation between TRplot with number of censuses (Individual- weighted: r = -0.05, P = 0.78; basal area-weighted: r = -0.02, P = 0.89), which indicates that the overall TRplot are not influenced by the number of censuses, but long-term monitoring takes into account the fluctuation of tree demography over time.

Of the 38 plots with more than three censuses, 66 % (n = 25, individual-weighted) and 68% (n = 26, basal area-weighted) of the plots had positive TRplot. Additionally, 73 %

(n = 16, individual-weighted) and 70 % (n = 14, basal area-weighted) of those plots with positive TRplot significantly increased the thermophilization rates (Consistently increased in CTI; Fig. 1a, b) over time (Fig. 2a, b; SI Appendix, Table S2). This indicates that those plots are increasing in the relative abundances of tree species from relative warmer climates, consistent with the thermophilization hypothesis. The overall number of plots with positive TRplot (n = 25, individuals-weighted; n = 26, basal area-weighted) along the gradient was marginally more than expected under the null expectation of equal proportion of the plots with positive and negative TRplot under random fluctuations on the relative population change and tree species composition (binomial probability; P = 0.04, CI = 0.20

– 0.51 for individuals basis and P = 0.02, CI = 0.18 – 0.49 for basal area basis). The forest plots with the negative TRplot along the gradient were mainly located in the lowland

Amazonian forests with the lowest TRplot values found in the lowland floodplain and

144 treeline forests for both the individual- and basal area-weighted approach (Fig. 2a, b; SI

Appendix, Table S2).

Looking at the TRplot for Amazonian (n = 17 plots) and Andean (n =21 plots) forest plots separately, we observed that the TRplot in the Amazonian plots was significantly lower than their Andean counterparts [Mann-Whitney-Wilcoxon test; P = 0.03 for individuals- weighted (26-fold difference) and P = 0.04 for basal area-weighted basis (21-fold difference); SI Appendix, Fig. S6]. The mean TRplot for Amazonian plots (< 500 m) was essentially zero for both individual- and basal area-weighted measures [+0.00024 °C yr-1

(95% CI = -0.0024 - +0.0028 °C yr-1) in terms of individuals and -0.00021 °C yr-1 (95% CI

= -0.0021 - +0.0017 °C yr-1) for basal area-weighted measures]. Across to the Amazonian sites, the average TRplot for floodplain plots was approximately zero [mean TRplot

-1 -1 individual-weighted = -0.0019 °C yr (95% CI = -0.0098 - +0.0060 °C yr ); mean TRplot basal area-weighted = -0.0027 °C yr-1 (95% CI = -0.0076 - +0.0021 °C yr-1)] compared to

-1 the terra firme plots [mean TRplot individual-weighted = +0.0014 °C yr (95% CI = -0.0004

-1 -1 - +0.0033 °C yr ); mean TRplot basal area-weighted = +0.0012 °C yr ; (95% CI = -0.0004

- +0.0027 °C yr-1); Inset in Fig. 2a, b and SI Appendix, Table S2].

Thermophilization rates observed in the Andean plots were +0.0062 °C yr-1 (95%

CI = +0.0015 - +0.0109 °C yr-1) for individual-weighted and was 25 % higher than the

-1 basal area-weighted, with a mean TRplot of +0.0044 °C yr (95% CI = -0.0006 - +0.0083

°C yr-1). This implies that the Amazonian tree communities are in stasis with respect to

-1 -1 thermophilization [TRplot = +0.04 m yr (95% CI = -0.43 - +0.51 m yr ) in terms of individuals, and -0.04 m yr-1 (95% CI = -0.39 - +0.31 m yr-1) for basal area-weighted].

Andean tree communities showed evidence of weak upslope migration [+1.13 m yr-1 (95%

145

CI = +0.27, +1.99 m yr-1) in terms of individuals and +0.80 m yr-1 (95% CI = +0.10, +1.50 m yr-1) in terms of basal area].

Mortality, growth, and recruitment effects on community thermophilization

The changes in TRplot was driven primarily by tree mortality (TRmor) events (F =

224.5, df = 36, p < 0.001, inset in Fig. 3) as opposed to tree recruitment (TRrec) and growth

(TRgro) (Fig. 3). In 20 of 38 plots, TRmor accounted for a larger proportion of observed thermophilization (53 % of the plots, n = 20 plots) when compared to TRrec (11 %, n = 4 plots) and TRgro (37 %, n = 14 plots). TRplot was significant and positive correlated with tree mortality (r= 0.64, P ˂ 0.001) and moderately correlated with growth (r= 0.38, P ˂

0.05), but did not show a correlation with tree recruitment (r= -0.15, P = 0.36; SI Appendix,

Table S7). High tree mortality, in particular in Andean plots, suggests that the observed thermophilizaton is the result of the mortality of more cold-adapted species at a site rather than changes in tree growth or the recruitment of new individuals into the plots.

Migration of individual taxa

Genus-level thermal migration rates (TMRgenus) were variable along the gradient.

However, we observed the higher TMRgenus in the upper Andean montane forest with both the individual- and basal area-weighted approach (Fig. 4a, b; SI Appendix, Table S3, S4).

For our selected dominant tree genera (n = 78) that occurred at least in 2 of the 38 forest

-1 plots along the gradient, the mean TMRgenus was +0.00593 °C yr (95% CI = +0.0022 -

146

+0.0097 °C yr-1) in terms of individuals and was almost the same in terms of basal area,

-1 -1 with a mean TMRgenus of +0.00594 °C yr (95% CI = +0.0018 - +0.0100 °C yr ). The

TMRgenus was positive in 63 % (n = 49, individuals-weighted) and 60 % (n = 47, basal area- weighted) of the total genera, but only 35 % (n = 17, individuals-weighted) to 43 % (n =

20, basal area-weighted) significantly increased in TMRgenus over time (Fig. 4a, b; SI

Appendix, Table S3, S4). The overall number genera with positive TMRgenus along the gradient was slightly more than expected under the null expectation of equal proportion of the genera with positive and negative TMRgenus (binomial probability; P = 0.02, CI = 0.26

– 0.49 for individuals basis and P = 0.04, CI = 0.29 – 0.51 for basal area basis), though there was high heterogeneity in rates (Fig. 4a, b).

Thermal migration rates calculated at species-level (TMRspecies) showed broad variability in TMRspecies along the gradient with no relationship with elevation. Although we observed widespread variability in TMRspecies, our results also showed an overall mean positive thermal migration rates along the gradient, though the magnitude of the migration was small. TMRspecies for the selected species (n = 79) showed a mean TRspecies of +0.0025

°C yr-1 (95% CI = +0.0011 - +0.0038 °C yr-1) on an individual basis and +0.0024 °C yr-1

(95% CI = +0.0006 - +0.0042 °C yr-1) on a basal area basis. Sixty-seven percent of our species (n = 53) showed a positive TMRspecies for the individuals-weighted and 62 % (n =

49) for the basal area-weighted approach, exceeding the null expectation of equal proportions of the species with positive and negative TMRspecies (binomial probability; P =

0.001, CI = 0.23 – 0.44 for individuals basis and P = 0.02, CI = 0.27 – 0.49 for basal area basis). Even though we observed an overall positive TMRspecies along the gradient, only 30

% (n = 16, individuals-weighted) and 19 % (n = 10, basal area-weighted) significantly

147 increased in TMRspecies over time (Fig. 5a, b; SI Appendix, Table S5, S6). It was not observe a correlation between the tree species with significant positive thermal migration and elevation (Individual-weighted: r = -0.38, P = 0.15; basal area-weighted: r = -0.45, P =

0.18). On the other hand, 31 % (n = 8, individuals-weighted) and 10 % (n = 3, basal area- weighted) of the species showed a significant negative TMRspecies.

Discussion

Our results showed that individual plots in Amazonian-Andean forests are becoming more thermophilic over the last 38 years, but barely so. The overall rates of thermophilization reported here showed an average of +0.0035 °C yr-1 (individuals- weighted) to +0.0023 °C yr-1 (basal area-weighted) are vastly slower compared to results of previous studies across the tropics, being 65 % lower (Feeley et al. 2013, Duque et al.

2015, Fadrique et al. 2018). Moreover, thermophilization as defined here can be caused by species with warmer mean distributions recruiting into a community, growing more within that community, or species with cooler mean temperatures dying within the community. “Migration” of taxa can similarly be due to increased recruitment, growth, or mortality at the upper (lower) edges of ranges. Our results show that both community thermophilization and taxon-based migration is mostly explained by the increase in tree mortality rather than “true” shifts of entire ranges upslope, calling into question the very term migration when applied to these communities. When viewed in the context of predicted climate change the result is striking as it means that by ~2100 the Amazonian-

Andean tree communities will have only changed by around a quarter of a degree Celsius

( +0.28 - +0.19 °C, based in our current migration rates) while the ambient temperature at

148 any plot is predicted to have changed by +2 - +7 °C (Urrutia and Vuille 2009), depending on the position along the elevational gradient. The observed lag in tree migration is out of equilibrium with current climate change by over an order of magnitude and is insufficient, increasing the “climatic debt” or migration that must happen for species to remain in equilibrium with climate (Bertrand et al. 2011, Svenning and Sandel 2013) in tropical forests. Moreover, given the absence of evidence for increased recruitment towards the thermal equilibrium (upslope), this climatic debt will likely increase over the coming decades.

High thermophilization of Andean forests in contrast to Amazonian forests

The overall thermophilization rates across all plots were positive on average and similar whether examined using individual- or basal area-weighted approaches, but with highly heterogeneous rates from plot-to-plot along the gradient (Fig. 2a, b). The velocity of Andean community thermophilization in response to the ongoing climate change was significantly higher than those of Amazonian communities (SI Appendix, Fig. S6). The overall mean Amazonian TRplot over time was no-different from zero (+0.00024 and -

0.00021 °C yr-1 for individuals- and basal area-weighed respectively; SI Appendix, Fig.

S6, Table S2), in other words, we observed no overall thermophilization for Amazonian tree communities. In contrast, the Andean forest increased the TRplot over time (+0.0062 and +0.0044 °C yr-1 for individuals- and basal area-weighed respectively; SI Appendix,

Fig. S6a, b, Table S2). These TRplot for the Amazon and Andean forests are an order of magnitude slower than that required to remain in equilibrium with the observed increase

149 in temperature of 0.03-0.05 oC yr-1 observed over the last half-century in the western

Amazon. Thus, these results indicate that the Amazonian tree community’s response to warming is absent (in terms of plant migration or community thermophilization), and slow but present, at least on average, for Andean tree communities, though highly variable among species, with many showing no detectable migration. Indeed, the signal appears to be more one of climatic disruption (increased mortality rates) rather than climatic migration.

Several plausible factors may explain the lack of thermophilization of Amazonian tree communities in contrast to the Andean forests. First, Amazonian trees may have greater tolerance to climate warming through local historical adaptation and persistence, and also because Amazonian trees have truncated niches, with less niche space on the warm side of their ranges as the MAT of Amazonia is relatively uniform, and temperature change due to atmospheric lapse rate is bounded at the upper end by sea level (Feeley and Silman 2010). In addition, the compensatory changes in demographic rates (Doak and Morris 2010) in the Amazonian forest may buffer population dynamics against the negative effects of global warming causing non-thermophilization. For example, higher growth of tree-individuals can compensate for the lower survival and recruitment rates allowing persistence of Amazonian populations, this trend can be observed in most of the floodplain plots in which tree growth has a negative effect in

TRplot (Figure 3, SI Appendix, Table S2). Second, an additional factor that may be driving the forces preventing thermophilization in the Amazonian forest could be due to the physiographic differences between the floodplain and terra firme forests that could affect tree demography. Floodplain forests are characterized to be underwater for a large

150 portion of the year while the adjacent terra firme stands are on old and highly weathered soils (Pitman et al. 1999). We observed that tree mortality was the cause of positive

TRplot for 73 % of the terra firme plots, in contrast, recruitment was the cause of negative

TRplot in 67 % of the floodplain plots (Fig. 3). This can be explained by the increase of mortality of wet-affiliated taxa (Esquivel-Muelbert et al. 2018) in terra firme forest. In addition, flooded areas could buffer the negative effects of droughts decreasing tree mortality in floodplain forests. Third, multiple droughts are driving slow but directional shifts in species composition across the Amazon in the last three decades (Esquivel-

Muelbert et al. 2018). Likewise, these drought events could also promote longer periods of tree growth in floodplain forests because the roots will potentially be closer to the water table extending the growing season, which can be reflected in high recruitment. For example, the CUZ-01, CUZ-02 plots when analyzed at individuals basis showed the lowest TRplot in comparison with the basal area (Inset in Fig. 2a, b), indicating that the high growth of small size stems is mirrored in the increments of new recruits in the plots.

This current inertia of Amazonian tree communities in response to climate change could lead to future lowland biotic attrition (Colwell et al. 2008).

Mortality causes ‘apparent’ plant migration

Here, we showed that studying tree demographic processes are crucial to comprehend the effects of ongoing warming in Amazonian-Andean tree communities, making long-term forest monitoring imperative to observe the fluctuations of tree populations from the year-to-decades basis. For example, Feeley et al. 2011 reported an

151 upslope tree migration of 2.5 - 3.5 m yr-1 in four years of study (2003-2007), where the high rates of migration were explained by the disbalance in mortality and recruitment that could be the reflection of any climatic or disturbance factors. In this study we monitored

38 years of tree demography capturing the variability of mortality and recruitment over time. First, we observed that tree mortality (TRmor) was the primary driver in changes in

TRplot across the plots (Fig. 3). Second to that was changes in growth. Taken together, changes in growth and mortality accounted for 89% of thermophilization. To turn it around, population shifts or recruitment favoring cooler areas are currently accounting for only 11% of the observed changes in species ranges of community thermophilization.

The pattern has several explanations. First, it may be that the process of mortality and recruitment could be expected to be decoupled in time, particularly with respect to trees reaching a diameter of 10 cm. Trees of this size can be decades to centuries old, having recruited in cooler temperatures than currently found in sites. In this case, thermophilization signals from growth may be a good indicator of future plant performance and recruitment, particularly as mortality is a fast process, while recruitment is a slow process in tree demography. If this is true, an examination of the seedling and small sapling layer should show an accelerated thermophilization or migration rate, with changes concentrated in recruitment as opposed to mortality. Second, the prevalence of positive TRmor cases along the gradient indicates the increase in mortality of cold-adapted species (less-thermophilic) in contrast to warm-adapted species (thermophilic) in response to warming, consistent with the findings from Costa Rica and Colombia (Feeley et al. 2013, Duque et al. 2015). The high tree mortality can be explained by the incapability of species to persist in areas where temperature increase exceeds species

152 thermal tolerance causing dieback along species ranges. In addition, heat-induced mortality can also be caused by an increase in temperature alone, increasing water stress independent of precipitation amounts (Barber et al. 2000).

An alternative is that drought-induced tree mortality—as discussed above—could be causing shifts in species composition and function in the Amazonian forest by killing trees preferentially in the warmer (lower) portions of their ranges. This is supported by the fact that there have been three major drought events in the Amazon basin in the last two decades, increasing tree mortality and reshuffling species composition (Saatchi et al.

2013, Esquivel-Muelbert et al. 2018). As previously stated, tree mortality was the driver of TRplot along the gradient, with the highest TRplot found around the estimate cloud base

(~1500-2000 m, Fig. 2a, b) where high tree mortality was reported (Farfan Rios 2011).

These results contrast with those of Fadrique et al. 2018, who suggested that the cloud base is a barrier for species migrations and who reported negative TRplot across the

Andes. Our results from the Manu-Tambopata transect showed that the cloud base zone is shifting in species composition to a more heat-tolerant species assembly. This trend may be explained due to the interplay of drought events and heat stress are increasing tree mortality globally leading to a future forest die-off in response to climate change (Allen et al. 2009) causing species range contraction, and therefore an inexorable reshuffling in species composition.

153

Past and current climate change shuffle species distribution

Past climate changes have been restructuring forests across the Andes-Amazon system as long as the Andes and Amazon have existed (Garzione et al. 2008, Hoorn et al.

2010) and what sets modern climate change with respect to temperature apart is the pace

(Loarie et al. 2009). Even the fastest-warming during the 5 - 6 oC warming that marked the change from the Last Glacial Maximum to Holocene in Amazonia occurred over ~8 - 10 thousand years (van der Hammen and Absy 1994, Urrego et al. 2005). Ongoing warming is expected to reach a similar magnitude in the system within a century to years (Lenssen et al. 2019), a rate unprecedented in the paleoecological record for the Amazon and forested

Andes (Bush et al. 2004). While modern temperature change may be without analog in the system (Lenssen et al. 2019), modern drought is actually less than that experienced by the system during the mid-Holocene (Baker et al. 2001) and the paleoecological record may provide good analogs for the rapidity at which systems in the Andes-Amazon can be restructured by precipitation in climate change (Malhi et al. 2008).

As discussed above, tree community thermophilization in response to climate change is based in demography, which by definition is a species-level population process

(Harper 1967), but it has not looked in that way in regional and global studies. The lack of studies looking at species-level population changes to understand the mechanisms of species migrations along environmental gradients makes tropical trees excluded from global synthesis studies in species migrations (Chen et al. 2011, Freeman et al. 2018) because long-term tree species-specific data are not available for the majority of tropical species. Thus, here we attempted to show the species-specific responses of tropical trees to ongoing climate change. Although we observed high variability in the thermal migration

154 rates at the species-level, 24 % and 20 % (individuals- and basal area-weighted approach respectively) of our species with positive TMRspecies showed significant positive migration rates (SI Appendix, Table S5, S6). For instance, the hyperdominance of Iriartea deltoideia in the Amazon basin has risen within the last 3000 years (Bush and McMichael 2016) and under the current climate change our results showed a significant upslope migration of this species (SI Appendix, Table S5, S6). These observed variable species-level migration rates suggest the idiosyncratic response of each taxon to climate change. However, we still need to understand the different underlying causes of tree demography—in particular mortality—along specific-species elevational ranges. Accordingly, we can hypothesize high mortality in the lower range-limit of the fast upslope migrators (positive TMRspecies) and high mortality in the upper range-limit for the downslope migrators (negative

TMRspecies). In addition, it is important to understand the role of disturbance (e.g. gaps and landslides) in tree mortality and species movement as potential corridors for species migration.

Here we provide comprehensive evidence of how Amazonian and Andean forests are most probably responding to warming and droughts, shifting their distributions in time and space mainly because of mortality events. The slow thermophilization rates for Andean forests and the absence of thermophiliation from Amazonian forests shows they are in disequilibrium with current climate change. The wide variability of tropical forest response to global change at community- and species-level makes challenging the predictions of winners (upslope migrators) and losers (non or downslope migrators) within specific communities in response to climate change. Forecasting the winners can contribute to

155 forest restoration practices and to most effective strategies for climate change mitigation, in a world with extra room for 0.9 billion hectares of continuous forest (Bastin et al. 2019).

156

Acknowledgements

This paper is a product of the Andes Biodiversity and Ecosystem Research Group

(ABERG; http://www.andesconservation.org/) with contributions for lowland plots data from John Terborgh, Percy Nunez and affiliated networks RAINFOR, GEM, and the

ForestPlots.net data management utility for permanent plots. Data included in this study is the result of an extraordinary effort by a large team in Peru specially from the

Universidad Nacional de San Antonio Abad de Cusco. Special thanks go to Luis Imunda for his assistance in the field sampling campaigns. SERNANP and personnel of Manu

National Park - Peru provided assistance with logistics and permission to work in the protected area. Pantiacolla Tours and the Amazon Conservation Association provided logistical support. Funding came from the Gordon and Betty Moore Foundation’s Andes to Amazon initiative and the US National Science Foundation (NSF) DEB 0743666 and

NSF Long-Term Research in Environmental Biology (LTREB) 1754647. The research was also supported by the National Aeronautics and Space Administration (NASA)

Terrestrial Ecology Program grant # NNH08ZDA001N-TE/ 08-TE08-0037. Support for

RAINFOR and ForestPlots.net plot monitoring in Peru has come from a European

Research Council (ERC) Advanced Grant (T‐FORCES, “Tropical Forests in the

Changing Earth System”, 291585), Natural Environment Research Council grants

(including NE/F005806/), NE/D005590/1, and NE/N012542/1), and the Gordon and

Betty Moore Foundation.

157

Literature cited

Allen, C. D., A. K. Macalady, H. Chenchouni, D. Bachelet, N. McDowell, M. Vennetier,

T. Kitzberger, A. Rigling, D. D. Breshears, E. H. Hogg, P. Gonzalez, R. Fensham,

Z. Zhang, J. Castro, N. Demidova, J. H. Lim, G. Allard, S. W. Running, A. Semerci,

and N. Cobb. 2009. A global overview of drought and heat-induced tree mortality

reveals emerging climate change risks for forests. Forest Ecology and Management

259:660–684.

Baker, P. A., G. O. Seltzer, S. C. Fritz, R. B. Dunbar, M. J. Grove, P. M. Tapia, S. L.

Cross, H. D. Rowe, and J. P. Broda. 2001. The history of South American tropical

precipitation for the past 25,000 years. Science (New York, N.Y.) 291:640–3.

Barber, V. A., G. P. Juday, and B. P. Finney. 2000. Reduced growth of Alaskan white

spruce in the twentieth century from temperature-induced drought stress. Nature

405:668–673.

Bertrand, R., J. Lenoir, C. Piedallu, G. Riofrío-Dillon, P. de Ruffray, C. Vidal, J.-C.

Pierrat, and J.-C. Gégout. 2011. Changes in plant community composition lag

behind climate warming in lowland forests. Nature 479:517–520.

Boyle, B., N. Hopkins, Z. Lu, J. A. Raygoza Garay, D. Mozzherin, T. Rees, N. Matasci,

M. L. Narro, W. H. Piel, S. J. McKay, S. Lowry, C. Freeland, R. K. Peet, and B. J.

Enquist. 2013. The taxonomic name resolution service: an online tool for automated

standardization of plant names. BMC bioinformatics 14:16.

Bush, M. B. 2002. Distributional change and conservation on the Andean flank: A

158

palaeoecological perspective. Global Ecology and Biogeography 11:463–473.

Bush, M. B., and C. N. H. McMichael. 2016. Holocene Variability of an Amazonian

Hyperdominant. Journal of Ecology.

Bush, M. B., M. R. Silman, and D. H. Urrego. 2004. 48,000 years of climate and forest

change in a biodiversity hot spot. Science (New York, N.Y.) 303:827–9.

Cano, A., K. R. Young, B. Leon, and R. B. Foster. 1995. Composition and diversity of

flowering plants in the upper montane forest of Manu National Park, Southern Peru.

Pages 271–280 in S. P. Churchill, H. Balslev, E. Forero, and J. L. Luteyn, editors.

Biodiversity and Conservation of Neotropical Montane Forests: Proceedings of the

Neotropical Montane Forest. New York Botanical Garden Pr Dept, United states of

America.

Chase, M. W., M. J. M. Christenhusz, M. F. Fay, J. W. Byng, W. S. Judd, D. E. Soltis, D.

J. Mabberley, A. N. Sennikov, P. S. Soltis, P. F. Stevens, B. Briggs, S. Brockington,

A. Chautems, J. C. Clark, J. Conran, E. Haston, M. Möller, M. Moore, R. Olmstead,

M. Perret, L. Skog, J. Smith, D. Tank, M. Vorontsova, and A. Weber. 2016. An

update of the Angiosperm Phylogeny Group classification for the orders and

families of flowering plants: APG IV. Botanical Journal of the Linnean Society

181:1–20.

Chen, I.-C., J. K. Hill, R. Ohlemüller, D. B. Roy, and C. D. Thomas. 2011. Rapid range

shifts of species associated with high levels of climate warming. Science (New

York, N.Y.) 333:1024–6.

159

Colwell, R. K., G. Brehm, C. L. Cardelús, A. C. Gilman, and J. T. Longino. 2008. Global

warming, elevational range shifts, and lowland biotic attrition in the wet tropics.

Science (New York, N.Y.) 322:258–61.

Davis, M. B. 1969. Climatic changes in southern Connecticut recorded by pollen

deposition at Rogers Lake. Ecology 50:409–422.

Doak, D. F., and W. F. Morris. 2010. Demographic compensation and tipping points in

climate-induced range shifts. Nature 467:959–962.

Duque, A., P. R. Stevenson, and K. J. Feeley. 2015. Thermophilization of adult and

juvenile tree communities in the northern tropical Andes. Proceedings of the

National Academy of Sciences.