TESIS Salgado-Negret

EFECTOS DE LA FRAGMENTACIÓN SOBRE LA DISTRIBUCIÓN DE ESPECIES ARBÓREAS EN EL PARQUE NACIONAL FRAY JORGE: IMPORTANCIA DE LOS ATRIBUTOS ECOFISIOLÓGICOS

PONTIFICIA UNIVERSIDAD CATÓLICA DE CHILE FACULTAD DE CIENCIAS BIOLÓGICAS PROGRAMA DOCTORADO EN CIENCIAS BIOLÓGICAS MENCIÓN ECOLOGÍA

EFECTOS DE LA FRAGMENTACIÓN SOBRE LA DISTRIBUCIÓN DE ESPECIES ARBÓREAS EN EL PARQUE NACIONAL FRAY JORGE: IMPORTANCIA DE LOS ATRIBUTOS ECOFISIOLÓGICOS

Por

BEATRIZ EUGENIA SALGADO NEGRET

Tesis presentada a la Facultad de Ciencias Biológicas de la Pontificia Universidad Católica de Chile para optar al grado académico de Doctor en Ciencias Biológicas mención Ecología

Dirigida por: Dr. Juan José Armesto Dra. Fernanda Pérez

Noviembre, 2013 Santiago, Chile

III

Agradecimientos

Quiero comenzar agradeciendo a la Comisión Nacional de Investigación Científica y Tecnológica (CONICYT, Chile) y al Instituto de Ecología y Biodiversidad (IEB) por el apoyo financiero para realizar este doctorado y el trabajo de investigación.

Quiero agradecer a mi tutor principal Juan Armesto, por apostar a ciegas y permitirme ser parte de su equipo, por sus invaluables enseñanzas y apoyo incondicional. A todos los miembros de su Laboratorio por todas las tertulias e increíbles discusiones.

A Fernanda Pérez por ser una excelente guía, por las eternas discusiones teóricas, por todas las salidas de campo, pero sobretodo por convertirse en una gran amiga y confidente. Fefita, eres de los grandes regalos que me llevo de Chile…gracias por todo!

A Fernando Valladares y su equipo por adoptarme por meses en su laboratorio y acogerme como un miembro más del equipo. Gracias por todos los análisis y discusiones que mejoraron este documento.

A Pablo Marquet, Javier Figueroa y Martín Carmona por sus aportes y comentarios que mejoraron esta propuesta desde sus inicios.

A Mylthon Jimenez-Castillo y su equipo por enseñarme el mundo de la hidráulica. Gracias especiales a Paulina Lobos.

A Juan Monardez, por su fiel compañía, por las deliciosas cenas y discusiones en compañía del mejor vino. Juan mil gracias por presentarme un ecosistema maravilloso.

A Aurora Gaxiola y Daniel Stanton por las múltiples charlas planteando hipótesis y discutiendo resultados…. sus comentarios mejoraron enormemente los manuscritos y su compañía fue un gran apoyo.

A Felipe Albornoz, Rafaella Canessa, Carmen Ossa, Daniel Salinas, Patricio Valenzuela e Isabel Mujica por su invaluable apoyo en campo y laboratorio y por hacer de las salidas de campo paseos repletos de risas y complicidad. A Mariela Aguilera y Ximena Alvarez por todas las reuniones, discusiones y tertulias alrededor de la fisiología de las plantas…hubo momentos brillantes…gracias queridas!

A mis compañeras de batalla y familia en Chile: Lidia Mansur, Sabrina Clavijo, Daniela Rivera y Carmencha Ossa…. no habría sido lo mismo sin ustedes. A Leo por todo su apoyo durante tantos años. IV

A Carolina Alcázar, Olga Caro y Carolina Useche por ser mis terapeutas en la distancia…gracias por todo el apoyo.

A mi familia en Colombia….por darme la libertad de soñar y por estar al pie del cañón….mil gracias por estar siempre tan cerca a pesar de la distancia.

A todas aquellas personas que no he nombrado pero que hicieron parte de este logro. Muchas gracias.

Tabla de contenidos

Lista de abreviaturas ...... VIII

Resumen ...... IX

Introducción General ……………….………………………..………1

Estructura de la tesis………………………… ...... 6

Área de estudio ...... 7

Visión general…...... 9

Referencias ...... 11

Capítulo I

Estrategias divergentes de tolerancia a la sequía explican la distribución de especies arbóreas a través de un gradiente de humedad dependiente de neblina en un bosque lluvioso templado

Abstract ...... 17

Introduction ...... 18

Materials and Methods ...... 20

Results ...... 25

Discussion...... 27

Acknowledgements ………………………………………………31

References ...... 32

Tables...... 38

Figures ...... 41 VI

Capítulo II

Variación en rasgos funcionales explica la distribución de Aextoxicon punctatum a través de un fuerte gradiente de humedad en un bosque fragmentado dependiente de neblina

Abstract ...... 47

Introduction ...... 48

Materials and Methods ...... 51

Results ...... 54

Discussion ...... 56

Acknowledgements…………………………………………………….. 59

References ...... 60

Tables...... 67

Figures ...... 70

Online supplemental materials ……………………………………74

Conclusiones Generales

Conclusiones ...... 78

Anexo I

Simetría de los parches de bosque depende de la dirección de los recursos limitantes

Abstract ...... 84

Introduction ...... 85

Materials and Methods ...... 88

Results ...... 90

Discussion ...... 91

Acknowledgements……………………..………………………………. 95 VII

References ...... 96

Tables ...... 100

Figures ...... 102

VIII

Lista de abreviaturas

AMAX = Photosynthetic rate; Tasa fotosíntesis gs = Stomatal conductance; Conductancia estomática

Hv = Huber value; Valor Huber

Ks = Sapwood-specific hydraulic conductivity

LA = Leaf area; Área foliar

LMA = Leaf mass area; Relación masa: área de la hoja

PLC = Percentage of loss conductivity; Porcentaje de pérdida de conductividad

RWC tlp = Relative water content at turgor loss point; Contenido relativo de agua al punto

de pérdida de turgor

SD = Stomatal density; Densidad estomática

TD = Trichome density; Densidad de tricomas

VD = Vessel density; Densidad de vasos

VDi = Vessel diameter, Diámetro de vasos

π0 = Solute potential at full turgor; Potencial de solutes a full turgor

πtlp = Water potential at turgor loss; Potencial hídrico al punto de pérdida de turgor

ɛ = Bulk modulus of elasticity; Modulo de elasticidad

ψPD = Leaf water potentials predawn; Potencial hídrico al amanecer

ψMD = Leaf water potentials at midday; Potencial hídrico al medio día

Resumen

El estudio de los rasgos funcionales y mecanismos fisiológicos que determinan la tolerancia de las especies a la sequía y su habilidad para competir por agua es fundamental para entender su distribución a través de gradientes de humedad y predecir su respuesta al cambio global, donde la fragmentación del hábitat y el cambio de uso del suelo son los principales motores de cambio. En este sentido, los bosques dependientes de neblina en las regiones semiáridas del mundo son un buen modelo de estudio para entender las respuestas de las especies al incremento en la aridez y la fragmentación del hábitat.

Se identificó un continuo de estrategias en el uso de agua explicando la distribución de las especies a través del gradiente de humedad a pequeña escala. Drimys winteri , una especie restringida al núcleo húmedo, mostró rasgos que permiten un eficiente transporte de agua y ganancia de carbono; en contraste, Myrceugenia correifolia , especie que domina los bordes secos de sotavento, presentó rasgos que promueven la conservación del agua y menores tasas de intercambio de gases, así como menor potencial hídrico al punto de pérdida de turgor. La especie con amplia distribución Aextoxicon punctatum , mostró valores de rasgos intermedios, pero se observó variación de las medias, magnitud e integración fenotípica a través de las zonas dentro de los fragmentos. Así, árboles creciendo en los bordes secos presentaron mayor LMA, densidad de estomas y tricomas que los árboles del núcleo húmedo y el borde barlovento. En contraste, rasgos de la anatomía del xilema no variaron produciendo pérdida de la conductividad hidráulica en los bordes más secos. También se detectaron mayores niveles de integración fenotípica y variabilidad en los bordes secos.

Los resultados mostraron que el particionamiento del pronunciado gradiente de humedad a pequeña escala entre las especies arbóreas está determinado por las tolerancias diferenciales de las especies a la sequía, y esas diferencias indican que las especies tienen habilidades contrastantes para lidiar con futuros cambios climáticos.

I. Introducción General

La disponibilidad de agua es el principal factor que determina la distribución de las especies arbóreas a través de gradientes de lluvia a gran escala así como en gradientes topográficos a pequeña escala (Gentry 1988, Wright 1992, Condit 1998, Bongers et al. 1999, Pyke et al. 2001, Condit et al. 2002, Engelbrecht et al. 2007). El estudio de los rasgos funcionales asociados al comportamiento de las especies bajo condiciones particulares de humedad del suelo ayuda a explicar la distribución de las especies (Poorter 2007, Markesteijn et al. 2011, Sterck et al. 2011), donde el éxito en el establecimiento y sobrevivencia en ambientes o épocas secas estará determinado por su habilidad para competir por agua y tolerar la sequía (Markesteijn et al. 2011). La capacidad de respuesta de las especies a la sequía y a cualquier variable ambiental está determinada por sus rasgos funcionales, los cuales son todas las características morfológicas, fisiológicas o fenológicas medidas a nivel individual (Viollé et al. 2007). Es bien conocido el trade off entre la adquisición y conservación de recursos que le permite a las especies especializarse a lo largo de esos gradientes ambientales (Reich et al. 2003, Diaz et al. 2004, Wright et al. 2004). Así, plantas que crecen en ambientes secos generalmente tienen hojas pequeñas, baja conductancia estomática, alta área foliar específica (e.j. Fahn 1986, Baldini et al. 1997, Niinemets 2001), pero presentan bajas tasas fotosintéticas y tasas de crecimiento (Reich et al. 2003). A nivel hidráulico también existen ciertos rasgos que determinan el establecimiento de las especies en determinados ambientes. Por ejemplo, especies que crecen en ambientes secos generalmente tienen vasos conductores más pequeños y densos con pequeños poros en las membranas que les permiten conducir agua bajo condiciones de baja disponibilidad hídrica disminuyendo el riesgo de embolismo. Estos rasgos incrementan la resistencia al flujo de agua y reducen la eficiencia hidráulica de las especies, afectando el suministro de agua a las hojas (Hacke et al. 2001, Choat et al. 2005, Markesteijn et al. 2011a,b). Según la combinación de rasgos funcionales, las plantas pueden estar ubicadas a través de un gradiente de estrategias (Reich et al. 2003, Díaz et al. 2004): en un extremo especies con rasgos que favorecen la conservación de los recursos (conservativas) a especies con rasgos que promueven la rápida captura de recursos (adquisitivas). Entender las estrategias y mecanismos que tienen las especies para sobrevivir a la sequía es crítico para predecir las consecuencias ecológicas de futuras alteraciones en la humedad del suelo debido a motores del cambio global como la fragmentación, el cambio de uso del suelo o el cambio climático. 3

La mayoría de los estudios relacionados con distribución de especies a través de gradientes de humedad, han examinado la variación de rasgos funcionales a nivel interespecifico (Pockman y Sperry 2000, Cornwell y Ackerly 2009, Engelbrecht et al. 2007, Choat et al. 2012, Salgado-Negret et al. 2013), considerando únicamente los valores promedio de los rasgos para cada especie, ignorado la importancia de la variabilidad intraespecífica. Esto puede responder a que las tendencias en comunidades diversas son principalmente el resultado del recambio de especies más que de la variación a nivel de especie (Cornwell y Ackerly 2009; Albert et al. 2010a,b; Hulshof y Swenson 2010). En ambientes con limitaciones hídricas, se ha propuesto una disminución de la variabilidad (coeficiente de variación) de los rasgos funcionales a nivel intraespecífico, debido a que solo individuos con un rango restringido de valores de rasgos es capaz de sobrevivir bajo esas condiciones ambientales (Cornwell y Ackerly 2009). A través de los gradientes de humedad del suelo también pueden variar las respuestas de los rasgos individuales (media y coeficiente de variación) y por lo tanto los patrones de correlación entre ellos (Pigliucci y Kolodynska 2002; Sardans, Penuelas y Roda 2006), y aunque en la literatura son bien conocidas las correlaciones entre rasgos foliares (Wright et al. 2004), rasgos hidráulicos (Chavé et al. 2009, Zanne et al. 2010) y entre ambos módulos vegetativos (Brodribb y Field 2000; Brodribb et al. 2002; Santiago et al. 2004; Wright et al. 2006; Meinzer et al. 2008; Baraloto et al. 2010), existe poca información acerca de cómo el ambiente puede alterar los patrones de correlación fenotípica entre rasgos de foliares y de madera (Nicotra et al. 1997, Wright et al. 2006). Uno de los ecosistemas con mayores limitaciones hídricas son los bosques dependientes de neblina encontrados en las regiones semiáridas del mundo (Hildebrandt y Eltahir 2006, del-Val et al. 2006, Katata et al. 2010). Estos bosques son relictos de periodos pasados cuando las condiciones fueron más húmedas, por lo cual son ecosistemas especialmente sensibles a los cambios actuales en la producción y distribución de la neblina. Se predice que alteraciones en la frecuencia e intensidad de la niebla ocurrirán debido a cambios en la temperatura superficial del mar y la altura de la capa de inversión térmica (Cereceda et al. 2002), pérdida de áreas de bosque y fragmentación o cambios en la estructura de los bosques afectando la captura de niebla (Hildebrandt y Eltahir 2006). En esos fragmentos de bosque, la intercepción de la niebla por las plantas es la principal o incluso la única fuente de agua durante la mayor parte del año (Dawson 1998, del-Val et al . 2006, Ewing et al . 2009). La intercepción por parte 4 de la vegetación crea pronunciados gradientes de agua y nutrientes desde el borde barlovento (entrada de niebla) al borde sotavento de los parches (Weathers et al . 2000, del-Val et al . 2006, Ewing et al . 2009), con fuertes contrastes en cortas distancias (Ewing et al . 2009). Estudiar la respuesta de las especies a la variación en la humedad del suelo a cortas escalas espaciales generadas por gradientes topográficos o de fragmentación en estos ecosistemas, nos permite direccionar preguntas acerca las condiciones críticas para el mantenimiento de especies arbóreas bajo estrés por sequía debido a cambio climático. Un interesante ejemplo de bosques dependientes de neblina se encuentra en la región semiárida en Chile (30°S), donde un mosaico de más de 180 parches de bosque persiste en las montañas costeras rodeado por una matriz de vegetación xerofítica (Barbosa et al. 2010). Este bosque tuvo una distribución continua, pero el incremento en la aridez en el Terciario tardío dividió su distribución (Villagrán et al. 2004). Como consecuencia, este tipo de bosque quedó restringido al rango montañoso costero de la región Mediterránea en Chile inundado por niebla (Villagrán et al. 2004), la cual duplica la precipitación efectiva de esta zona (del-Val et al. 2006). Los fuertes gradientes de humedad generados por la intercepción de la neblina afectan la distribución y dinámica de las especies. Las especies arbóreas dominantes en esos parches son: Aextoxicon punctatum (Aextoxicaceae), que ocurre en todos los bosques pero prefiere el borde barlovento que recibe directamente la entrada de la neblina; Drimys winteri (Winteraceae) que tiende a estar agregada en el núcleo de los grandes parches de bosque; y Myrceugenia correifolia (Myrtaceae) que es más común en los parches pequeños y está normalmente confinada en los bordes sotavento más secos (del-Val et al. 2006; Gutiérrez et al. 2008). Estas distribuciones están determinadas por contrastantes patrones de regeneración y mortalidad dentro de los parches. El reclutamiento está concentrado en los bordes húmedos en barlovento y es tres veces mayor que en sotavento, mientras que la mortalidad es mayor en el borde sotavento opuesto al ingreso de la neblina costera (del-Val et al. 2006). Los contrastantes patrones de distribución de estas especies arbóreas ofrecen una gran oportunidad para valorar los mecanismos subyacentes a su habilidad para tolerar las condiciones secas y la variación en esos mecanismos a lo largo de gradientes de humedad espacial determinados por la entrada de la niebla. Las preguntas e hipótesis que se abordarán en esta tesis son las siguientes: 5

1) ¿La variación de los rasgos foliares e hidráulicos relacionados con la tolerancia a la sequía explican los patrones de distribución contrastantes de tres especies arbóreas dominantes a través de gradientes de humedad a pequeña escala en los fragmentos de bosque del Parque Nacional Fray Jorge? Se espera que especies que crecen en sotavento bajo condiciones de déficit hídrico presenten un grupo de rasgos que favorezcan la conservación del agua (menor conductancia estomática) y que reduzcan el riesgo de cavitación (vasos angostos) con el costo de una menor efi ciencia hidráulica y fotosintética. 2) ¿Qué adaptaciones o mecanismos le permiten a los individuos de A. punctatum y M. correifolia crecer en los fragmentos pequeños o en los bordes secos de sotavento para lidiar con el déficit hídrico en comparación con individuos conespecíficos que creen en los núcleos húmedos de los fragmentos? Se espera que individuos que crecen en el borde seco en sotavento presenten rasgos fisiológicos que favorezcan la tolerancia a la sequía como menor πtlp y π0, en comparación con individuos conespecíficos que crecen en los núcleos húmedos de los fragmentos. Se espera que los individuos que crecen en fragmentos pequeños y en los bordes secos de sotavento tengan rasgos que favorezcan la conservación del agua como mayor densidad de tricomas y LMA y que reduzcan el riesgo de cavitación disminuyendo el diámetro de sus vasos conductores. 3) ¿La variabilidad e integración fenotípica incrementan en sotavento con mayor variabilidad ambiental y menor disponibilidad de agua? Se espera que la variabilidad e integración fenotípicas incrementen en los bordes de sotavento debido a la mayor variabilidad ambiental e incremento en el déficit hídrico. Las especies arbóreas que viven en los bosques de neblina del Parque Nacional Fray Jorge han estado expuestas a un incremento en la aridez debido a cambios climáticos ocurridos por periodos extendidos de tiempo (Villagrán et al. 2004; Gutiérrez et al. 2008), y han enfrentado cambios estacionales en la producción de la neblina que generan pronunciados gradientes de humedad dentro de los fragmentos (del-Val et al. 2006). Este estudio revela algunos de los mecanismos clave que explican el éxito de esas especies para coexistir dadas las variaciones pasadas y actuales en la disponibilidad de humedad del suelo. Los resultados se discuten a la luz de las posibles consecuencias de futuros cambios climáticos y sus efectos sobre la distribución y coexistencia de especies. 6

Estructura de la tesis En consideración a lo expuesto, en este proyecto de tesis se plantea como objetivo central estudiar los mecanismos fisiológicos que ayudan a explicar los patrones contrastantes de distribución y abundancia observados en las tres especies arbóreas dominantes en los fragmentos de bosque del Parque Nacional Fray Jorge: Aextoxicon punctatum , Drimys winteri y Myrceugenia correifolia . La intercepción de la niebla por parte de la vegetación en los bosques del Parque Nacional Fray Jorge genera fuertes gradientes de humedad del suelo, donde las zonas sotavento son más secas que los otros dos microhábitats, mientras que la humedad del suelo en las zonas barlovento (ingreso de la neblina) es comparable con los núcleos de los fragmentos (véase capitulo 1). Las especies arbóreas dominantes están distribuidas diferencialmente a través de este gradiente de humedad del suelo. Así, Aextoxicon punctatum (Aextoxicaceae), ocurre en todas las zonas de los parches pero prefiere el borde barlovento que recibe directamente la entrada de la neblina, Drimys winteri (Winteraceae) tiende a estar agregada en el núcleo de los grandes parches de bosque y Myrceugenia correifolia (Myrtaceae) es más común en los parches pequeños y está normalmente confinada en los bordes sotavento más secos (del-Val et al. 2006; Gutiérrez et al. 2008). Primero se estudiaron los rasgos foliares (área foliar, área foliar específica, tasa fotosintética, conductancia estomática) e hidráulicos (diámetro y densidad de vasos conductores, conductividad hidráulica específica de la madera y valor Huber) relacionados con la tolerancia a la sequía en las tres especies arbóreas (véase capitulo 1). Adicionalmente, se realizaron curvas presión-volumen para los individuos que crecen en el borde seco sotavento y en el núcleo húmedo, con el objetivo de entender cuáles eran los mecanismos de las especies para lidiar con el déficit hídrico (ajuste osmótico o incremento en la elasticidad celular) en sotavento en comparación con individuos conespecíficos que creen en los núcleos húmedos de los fragmentos (véase capitulo 1). Para entender la habilidad de Aextoxicon punctatum para sobrevivir a través del gradiente de humedad del suelo, primero se estudió la magnitud y variabilidad de los rasgos foliares (área foliar específica, densidad de estomas y de tricomas) e hidráulicos (diámetro y densidad de vasos conductores y conductividad hidráulica específica de la madera) relacionados con la tolerancia a la sequía a través de las tres zonas en los parches (ver capítulo 2); y segundo se estudió la integración fenotípica entre los rasgos 7 funcionales evaluados en cada una de las zonas de humedad del suelo de los parches (ver capítulo 2). Esta tesis revela mecanismos fisiológicos clave que ayudan a explicar la distribución contrastante de las especies arbóreas a través de zonas y parches en el Parque Nacional Fray Jorge, y aporta información útil para intentar predecir la respuesta de estas especies a futuros cambios globales como la fragmentación y el cambio climático.

Área de estudio El área de estudio está ubicada en el Parque Nacional Fray Jorge, localizado en la región de Coquimbo (Chile) (30°40´S, 71°35´W) (Figura 1). El clima es Mediterráneo árido caracterizado por veranos cálidos y secos e inviernos húmedos y fríos (Di Castri y Hajek 1976). La temperatura promedio es de 13.6°C y la precipitación promedio es de 147 mm concentradas en los meses de Junio a Agosto (López-Cortez y López 2004). Durante los meses de Octubre a Enero hay mayor incidencia de la niebla costera, la cual puede aportar anualmente alrededor de 250 mm adicionales a las precipitaciones (del Val et al. 2006). Esta neblina está asociada con el agua fría generada por la corriente de Humboldt e inversión producidos por la subsidencia Anticiclón Pacífico Sur (Cereceda et al. 2002). Los fragmentos de bosque de neblina varían entre 0,1 y 36 ha (Barbosa et al. 2010) y se encuentran rodeados por una matriz de vegetación xerofítica y cactáceas. Están localizados entre los 400 y 600 m de altitud, representando el límite norte de distribución del bosque templado dominado por Aextoxicon punctatum , el cual tiene una distribución continua cerca de 1000 km hacia el sur del país (37°-43°S) (Smith-Ramírez et al. 2005). Florísticamente los fragmentos de bosque están dominados en su estrato arbóreo por Aextoxicon punctatum , género monotípico de una familia endémica de los bosques templados de Sudamérica (Aextoxicaceae), y otras especies como Myrceugenia correifolia (Myrtaceae), Rhaphithamnus spinosus (Verbenaceae), Drymis winteri (Winteraceae) y Azara microphylla (Flacourtiaceae) (Villagrán et al. 2004). Tienen importantes trepadoras y epífitas leñosas y herbáceas como Griselinia scandens (Griseliniaceae), Sarmienta repens (Gesneriaceae) y Mitraria coccinea (Gesneriaceae) (Villagrán et al. 2004), incluyendo helechos como Polypodium e Hymenophyllum . 8

Figura 1. Ubicación de los fragmentos de bosque dependientes de neblina en el Parque Nacional Fray Jorge a los 30°S (Barbosa et al. 2010). Para esta investigación se seleccionaron cuatro fragmentos: dos fragmentos pequeños (<0.5 ha) y dos fragmentos grandes (> 20 ha) (Tabla 1). Los fragmentos tienen similar exposición y edad (Gutiérrez et al. 2008), están separados por una distancia mínima de 400 metros entre sí y no están afectados por la presencia de otros fragmentos que pudieran alterar la captura de neblina (Barbosa 2005).

Tabla 1. Caracterización de los fragmentos de bosque, valores medios de las variables microclimáticas y área basal relativa para los individuos vivos (>5 cm dap) para los cuatro fragmentos estudiados (Gutiérrez et al. 2008, Barbosa et al. 2010).

F1 F2 F5 F6

Área del fragmento 0.21 0.28 36.08 23.76

Altitud (m) 529 566 635 639

Pendiente (%) 1 11 42 38

Throughfall (mm) 31.10 ± 21.31 49.91 ± 43.16 29.56 ± 18.05 37.38 ± 22.55

Stemflow (mm) 0.10 ± 0.06 0.25 ± 0.06 0.69 ± 1.07 1.00 ± 0.99

Temperatura media (°C) 11.8 ± 1.6 11.46 ± 1.75 11.29 ± 1.51 10.95 ± 1.42

Humedad relative media (%) 91.33 ± 4.00 94.98 ± 3.04 95.96 ± 3.73 95.12 ± 4.63

Área basal arbórea (m 2 ha -1) 61.64 49.41 125.12 102.61

Área basal A. punctatum (%) 49 75.7 46.4 88.8

Área basal D. winteri (%) 0 0 52 10.8

Área basal M. correifolia (%) 50.8 21.6 0.3 0.2

Área basal otras especies (%) 0.2 2.7 1.3 0.3

Visión general

Uno de los objetivos de la ecología es entender los procesos que estructuran las comunidades naturales, donde los estudios a través de gradientes ambientales han tenido gran relevancia. En las comunidades forestales de las regiones áridas del mundo, la disponibilidad de agua es uno de los principales factores que determina la distribución de las especies, y los patrones observados han sido frecuentemente atribuidos a las diferencias entre especies en sus tolerancias a la sequía y habilidades para competir por agua. Entender cómo los rasgos funcionales relacionados a la tolerancia a la sequía varían a través de gradientes a pequeña escala es importante para predecir la respuesta de las especies a futuros cambios climáticos.

En esta tesis se estudiaron los mecanismos fisiológicos que explican los patrones contrastantes de distribución observados a través de gradientes de humedad generados por la neblina costera en las tres principales especies arbóreas Aextoxicon punctatum , Drimys winteri y Myrceugenia correifolia que coexisten en los fragmentos de bosque del Parque Nacional Fray Jorge, en la región semiárida en Chile. Se encontró un 10 continuo de estrategias en el uso de agua que permitieron explicar la distribución de las especies a través del gradiente de humedad: en un extremo la especie Drimys winteri , con rasgos favoreciendo la eficiencia hidráulica y fotosintética; mientras que en el extremo opuesto la especie Myrceugenia correifolia, con rasgos favoreciendo la conservación del agua y reduciendo el riesgo a la cavitación. La especie con amplia distribución Aextoxicon punctatum , mostró valores de rasgos intermedios, con variación en los rasgos foliares y ausencia de variación en la anatomía del xilema a través de las zonas dentro de los fragmentos. En Aextoxicon punctatum se detectaron mayores niveles de integración fenotípica y variabilidad en los bordes secos.

Referencias

Ackerly DD, Cornwell WK. 2007. A trait-based approach to community assembly: partitioning of species trait values into within- and among-community components. Ecology Letters 10: 135-145. Albert CH, Thuiller W, Yoccoz NG, Soudant A, Boucher F, Saccone P, Lavorel S. 2010a. Intraspecific functional variability: extent, structure and sources of variation. Journal of Ecology 98: 604-613. Albert CH, Thuiller W, Yoccoz NG, Douzet R, Aubert S, Lavorel S. 2010b A multi- trait approach reveals the structure and the relative importance of intra- vs. interspecific variability in plant traits. Functional Ecology 24: 1192-1201. Baldini E, Facini O, Nerozzi F, Rossi F, Rotondi A. 1997. Leaf characteristics and optical properties of different woody species. - Trees 12: 73-81. Baraloto C, Paine CET, Poorter L, Beauchene J, Bonal D, Domenach AM, Herault B, Patiño S, Roggy JC, Chave J. 2010. Decoupled leaf and stem economics in rain forest trees. Ecology Letters 13: 1338-1347. Barbosa O, Marquet PA, Bacigalupe LD, Christie DA, del-Val E, Gutiérrez AG, Jones CG, Weathers KC, Armesto JJ. 2010. Interactions among patch area, forest structure and water fluxes in a fog-inundated forest ecosystem in semi-arid Chile. Functional Ecology 24: 909-917. Bongers F, Poorter L, Van Rompaey RSAR, Parren MPE (1999) Distribution of twelve moist forest canopy tree species in Liberia and Cote d’Ivoire: response curves to a climatic gradient. Journal of Vegetation Science 10: 371–382 Brodribb TJ, Feild TS. 2000. Stem hydraulic supply is linked to leaf photosynthetic capacity: evidence from New Caledonian and Tasmanian rainforests. Plant, Cell and Environment 23: 1381-1388. Brodribb TJ, Holbrook NM, Gutiérrez MV. 2002. Hydraulic and photosynthetic coordination in seasonally dry tropical forest trees. Plant, Cell and Environment 25: 1435–1444. Cereceda P, Osses P, Larraín H, Farías M, Lagos M, Pinto R, Schemenauer RS (2002) Advective, orographic and radiation fog in the Tarapacá region, Chile. Atmosphere Research 64: 261–271. Chave J, Coomes D, Jansen S, Lewis SL, Swenson NG, Zanne AE. 2009. Towards a worldwide wood economics spectrum. Ecology Letters 12: 351–366. Choat B, Marilyn CB, Luly JG, Holtum JAM. 2005. Hydraulic architecture of deciduous and evergreen dry rainforest tree species from north-eastern Australia. Trees-Struct Funct 19:305-311. Choat B, Jansen S, Brodribb TJ, Cochard H, Delzon S, Bhaskar R, Bucci SJ, Field TS, Gleason SM, Hacke UG, Jacobsen AL, Lens F, Maherali H, Martínez-Vilalta J, Mayr S, Mencuccini M, Mitchell PJ, Nardini A, Pittermann J, Pratt RB, Sperry JS, Westoby M, Wright IJ, Zanne AE. 2012. Global convergence in the vulnerability of forests to drought. Science 491: 752-756. Condit R. 1998. Ecological implications of changes in drought patterns: shift in forest composition in Panama. Climate Change 39: 413–427. 12

Condit R, Pitman N, Leigh EG Jr, Chave J, Terborgh J, Foster RB, Nuñez-V P, Aguilar S, Valencia R, Villa G, Muller-Landau HC, Losos E, Hubbell SP. 2002. Beta- diversity in tropical forest trees. Science 295: 666–669. Cornwell WK, Ackerly DD. 2009. Community assembly and shifts in plant trait distributions across an environmental gradient in coastal California. Ecological Monographs 79: 109-126. Dawson TE. 1998. Fog in the California redwood forest: ecosystem inputs and use by plants. Oecologia 117: 476–485. Díaz S, Hodgson JG, Thompson K, Cabido M, Cornelissen JHC, Jalili A, Montserrat- Martí G, Grime JP, Zarrinkamar F, Asri Y, Band SR, BasconceloS1, Castro- Díez P, Funes G, Hamzehee B, Khoshnevi M, Pérez-Harguindeguy N, Pérez- Rontomé MC, Shirvany FA, Vendramini F, Yazdani S, Abbas-Azimi R, Bogaard A, Boustani S, Charles M, Dehghan M, de Torres-Espuny L, FalczukV, Guerrero-Campo J, Hynd A, Jones G, Kowsary E, Kazemi-Saeed F, Maestro- Martínez M, Romo-Díez A, Shaw S, Siavash B, Villar-Salvador P, Zak MR. 2004. The plant traits that drive ecosystems: Evidence from three continents. Journal of Vegetation Science 15: 295-304. del-Val E, Armesto JJ, Barbosa O, Christie DA, Gutie´rrez AG, Jones CG, Marquet PA, Weathers KC. 2006. Rain forest islands in the Chilean semiarid region: fog- dependency, ecosystem persistence and tree regeneration. Ecosystems 9: 598– 608. Di Castri F, Hajek E. 1976. Bioclimatología de Chile. Universidad Católica de Chile, Santiago Engelbrecht BMJ, Comita LS, Condit R, Kursar TA, Tyree MT, Turner BL, Hubbell SP. 2007. Drought sensitivity shapes species distribution patterns in tropical forests. Nature 447: 80–83. Ewing HA, Weathers KC, Templer PH, Dawson TE, Firestone MK, Elliott AM, Boukili VKS. 2009. Fog water and ecosystem function: heterogeneity in a California redwood forest. Ecosystems 12: 417–433. Fahn, A. 1986. Structural and functional properties of trichomes of xeromorphic leaves. Annals of Botany 57: 631-637 Gentry AH. 1988. Changes in plant community diversity and ﬂoristic composition on environmental and geographic gradients. Annals of Missouri Botanical Garden 75:1–34. Gutiérrez AG, Barbosa O, Christie DA, del-Val E, Ewing HA, Jones CG, Marquet PA, Weathers KC, Armesto JJ. 2008. Regeneration patterns and persistence of the fog dependent Fray Jorge forest in semiarid Chile during the past two centuries. Global Change Biology 14: 161–176. Hacke UG, Sperry JS, Pockman WT, Davis SD, McCulloh K. 2001. Trends in wood density and structure are linked to prevention of xylem implosion by negative pressure. Oecologia 126: 457-461. Hildebrandt A, Eltahir EAB. 2006. Forest on the edge: seasonal cloud forest in Oman creates its own ecological niche. Geophysical Research Letter 33: L11401. 13

Hulshof CM, Swenson NG. 2010. Variation in leaf functional trait values within and across individuals and species: an example from a Costa Rican dry forest. Functional Ecology 24: 217–223. Katata G, Nagai H, Kajino M, Ueda H, Hozumi Y. 2010. Numerical study of fog deposition on vegetation for atmosphere–land interactions in semi-arid and arid regions. Agricicultural Forest Meteoroly 150: 340–353 López-Cortés F, López D. 2004. Antecedentes bioclimáticos del Parque Nacional Bosque Fray Jorge. In: Squeo FA, Gutiérrez JR, Hernández IR (eds) Historia natural del parque nacional bosque Fray Jorge. Ediciones Universidad de La Serena, Chile, pp 45–60. Markesteijn L, Poorter L, Bongers F, Paz H, Sack L. 2011a. Hydraulics and life history of tropical dry forest tree species: coordination of species’ drought and shade tolerance. New Phytology 191:480-495. Markesteijn L, Poorter L, Paz H, Sack L, Bongers F. 2011b. Ecological differentiation in xylem cavitation resistance is associated with stem and leaf structural traits. Plant Cell Environ 34:137-148. Meinzer FC, Woodruff DR, Domec JC, Goldstein G, Campanello PI, Gatti MG, Villalobos-Vega R. 2008. Coordination of leaf and stem water transport properties in tropical forest trees. Oecologia 156: 31–41. Nicotra AB, Chazdon RL, Schlichting DS. 1997. Patterns of genotypic variation and phenotypic plasticity of light response in two tropical Piper (Piperaceae) species. American Journal of Botany 84: 1542-1552. Niinemets, U. 2001. Global-scale climatic controls of leaf dry mass per area, density and thickness in trees and shrub. Ecology 82: 453-469. Pigliucci M, Kolodynska A. 2003. Phenotypic plasticity and integration in resonse to flooded conditions in natural accessions of Arabidopsis thaliana (L.) Heynh (Brassicaceae). Annals of Botany 90: 199-207. Poockman WT, Sperry JS. 2000. Vulnerability to xylem cavitation and the distribution of Sonoran Desert. American Journal of Botany 87: 1287-1299. Poorter L. 2007. Are species adapted to their regeneration niche, adult niche, or both? Am Nat 169: 433–442 Pyke CR, Condit R, Aguilar S, Lao S. 2001. Floristic composition across a climatic gradient in a neotropical lowland forest. Journal of Vegetation Science 12: 553– 566. Reich PB, Wright IJ, Cavender-Bares J, Craine JM, Oleksyn J, Westoby M, Walters MB. 2003. The evolution of plant functional variation: traits, spectra, and strategies. International Journal of Plant Sciences 164: S143–S164. Salgado-Negret B, Pérez F, Markesteijn L, Jiménez-Castillo M, Armesto JJ. 2013. Diverging drought-tolerance strategies explain tree species distribution along a fog-dependent moisture gradient in a temperate rain forest. Oecologia. DOI 10.1007/s00442-013-2650-7 Santiago LS, Goldstein G, Meinzer FC, Fisher JB, Machado K, Woodruff D, Jones T. 2004. Leaf photosynthetic traits scale with hydraulic conductivity and wood density in Panamanian forest canopy trees. Oecologia 140: 543-550. 14

Sardans J, Peñuelas J, Rodá F. 2006. Plasticity of leaf morphological traits, leaf nutrient content, and water capture in the Mediterranean evergreen oak Quercus ilex subsp. ballota in response to fertilization and changes in competitive conditions. Ecoscience 13: 258-270. Smith-Ramírez C, Rovere AE, Núñez-Avila MC, Armesto JJ. 2007. Habitat fragmentation and reproductive ecology of Embothirum coccineum, Eucryphia cordifolia and Aextoxicon punctatum in southern temperate rainforests. In: (A.C. Newton, ed), “Biodiversity Loss and Conservation in Fragmented Forest Landscapes: The Forests of Montane Mexico and Temperate South America ", pp. 102-119. Sterck F, Markesteijn L, Schieving F, Poorter L. 2011. Functional traits determine trade-offs and niches in a tropical forest community. Proceedings of the National Academy of Scienes 108: 20627-20632. Villagrán C, Armesto JJ, Hinojosa LF, Cuvertino J, Pérez C, Medina C. 2004. El enigmático origen del bosque relicto de Fray Jorge. In ‘Historia natural del parque nacional bosque Fray Jorge’. (Eds FA Squeo, JR Gutiérrez, IR Hernández) pp. 3–43. Weathers KC, Lovett GM, Likens GE, Caraco NFM. 2000. Cloud-water inputs of nitrogen to forest ecosystems in southern Chile: forms, ﬂuxes and sources. Ecosystems 3: 590–595. Wright SJ. 1992. Seasonal drought, soil fertility and the species density of tropical forest plant communities. Trends in Ecology and Evolution 7: 260–263. Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers F, Cavender-Bares J, Chapin T, Cornelissen JH, Diemer M, Flexas J, Garnier E, Groom PK, Gulias J, Hikosaka K, Lamont BB, Lee T, Lee W, Lusk C, Midgley JJ, Navas ML, Niinemets U, Oleksyn J, Osada N, Poorter H, Poot P, Prior L, Pyankov VI, Roumet C, Thomas SC, Tjoelklaas MG, Veneklaas EJ, Villar R. 2004. The worldwide leaf economics spectrum. Nature 428: 821-827. Wright IJ, Falster DS, Pickup M, Westoby M. 2006. Cross-species patterns in the coordination between leaf and stem traits, and their implications for plants hydraulics. Physiologia Plantarum 127: 445-456. Zanne AE, Westoby M, Falster DS, Ackerly DD, Loarie SR, Arnold SEJ, Coomes DA. 2010. Angiosperm wood structure: global patterns in vessel anatomy and their relation to wood density and potential conductivity. American Journal of Botany 97: 207-215.

CAPÍTULO I

II. Estrategias divergentes de tolerancia a la sequía explican la distribución de especies arbóreas a través de un gradiente de humedad dependiente de neblina en un bosque lluvioso templado

Salgado-Negret B, Pérez F, Markesteijn L, Jimenez-Castillo M, Armesto JJ. 2013. Diverging drought tolerance strategies explain tree species distribution along a fog- dependent moisture gradient in a temperate rain forest. Oecologia DOI 10.1007/s00442- 013-2650-7

Diverging drought tolerance strategies explain tree species distribution along a fog- dependent moisture gradient in a temperate rain forest

Beatriz Salgado Negret 1,2,* , Fernanda Pérez 1,2 , Lars Markesteijn 3, Mylthon Jiménez Castillo 4,5 , Juan J. Armesto 1,2

1Departamento de Ecología, Pontificia Universidad Católica de Chile, Casilla 114-D, Santiago, Chile; 2Instituto de Ecología y Biodiversidad, Casilla 653, Santiago, Chile; 3Departamento de Biogeografía y Cambio Global, Museo Nacional de Ciencias Naturales, Consejo Superior de Investigaciones Científicas (CSIC), Serrano 115 dpdo, E-28006, Madrid, Spain; 4Instituto de Ciencias Ambientales y Evolutivas, Universidad Austral de Chile, Casilla 567, Valdivia-Chile; 5Jardín Botánico Universidad Austral de Chile, Facultad de Ciencias, Universidad Austral de Chile, Casilla 567, Valdivia-Chile.

*Author for correspondence:

Beatriz Salgado Negret [email protected]

56-2-3542637

Abstract

The study of functional traits and physiological mechanisms determining species´ drought tolerance is important to predict their responses to climatic change. Fog-dependent forest patches in semiarid regions are a good study system to understand species responses to increasing aridity and patch fragmentation.

Here we measured leaf and hydraulic traits for three dominant species with contrasting distributions within patches in relict, fog-dependent forests in semiarid Chile. In addition, we assessed pressure-volume curve parameters in trees growing at dry leeward edge and wet patch core.

We predicted species would display contrasting suites of traits according to local water availability: from one end favoring water conservation and reducing cavitation risk, to opposite end favoring photosynthetic and hydraulic efficiency. Consistent with our hypothesis, we identified a continuum of water use strategies explaining species distribution along small-scale moisture gradient. Drimys winteri, a tree restricted to the humid core, showed traits allowing efficient water transport and high carbon gain; in contrast, Myrceugenia correifolia , a tree that occurs in the drier patch edges, exhibited traits promoting water conservation and lower gas exchange rates, as well low water potential at turgor loss point. The most widespread species, Aextoxicon punctatum, showed intermediate trait values. Osmotic compensatory mechanism was detected in M. correifolia , but not in A. punctatum .

We show that partitioning of the pronounced soil moisture gradients from patch cores to leeward edges among tree species is driven by differential drought tolerance. Such differences indicate that trees have contrasting abilities to cope with future reductions in soil moisture.

Keywords

Climate change, fog-dependent forest, local water gradient, species distribution, plant hydraulic traits.

Introduction

Water availability is a major factor influencing species distribution in forest communities across large-scale rainfall gradients as well as small-scale topographic gradients (Gentry 1988; Wright 1992; Condit 1998; Bongers et al. 1999; Pyke et al. 2001; Condit et al. 2002; Engelbrecht et al. 2007). Species’ distribution may be explained by functional trait divergence associated with performance under particular conditions of soil humidity (Poorter 2007; Markesteijn et al. 2011a; Sterck et al. 2011). Understanding the bases of such differentiation among forest trees may be critical for predicting the ecological consequences of future alteration of soil moisture gradients due to climate change.

Fog-dependent forests, found in semiarid regions of the world (Hildebrandt and Eltahir 2006; del-Val et al. 2006; Katata et al. 2010), are thought to be relicts from past periods when conditions were more humid, and thus these ecosystems might be especially sensitive to current changes in fog water supply. Alterations in fog frequency and intensity are predicted to occur due to changes in sea-surface temperature and the height of the temperature inversion layer (Cereceda et al. 2002), loss of forest patch area and fragmentation, or changes in forest structure affecting fog capture (Hildebrandt and Eltahir 2006). In these patchy forests, fog interception by plants is the primary or even the only source of water during most of the year (Dawson 1998; del-Val et al . 2006; Ewing et al . 2009). The fog interception by trees creates pronounced water and nutrient gradients from windward to leeward edges in forest patches (Weathers et al . 2000; del-Val et al . 2006; Ewing et al . 2009), with strong contrasts over short distances, depending on wind direction (Ewing et al . 2009). Studying tree species responses to soil moisture variation at short spatial scales, due to topographic and/or patch fragmentation gradients in these fog- dependent ecosystems, allows us to address questions about the critical conditions for sustaining tree species under increasing drought stress due to changing climate.

Our study site, the Fray Jorge forest in central Chile, is a striking example of such a fog dependent ecosystem, where the strong water (and possibly nutrient) gradients inside the isolated forest patches affect the distribution and regeneration dynamics of tree species (del-Val et al. 2006). The patches are dominated by species characteristic of temperate and Mediterranean forests in Chile: Aextoxicon punctatum (in the monotypic family

Aextoxicaceae) is found in all-size patches but it is more frequent in humid windward edges, directly facing the incoming fog; Drimys winteri (Winteraceae) tends to be aggregated in the interior of the largest forest patches and is not found in small patches; finally, Myrceugenia correifolia (Myrtaceae) is more common along the edges of small patches, including the drier leeward edge (del-Val et al. 2006; Gutiérrez et al. 2008). Such contrasting distribution patterns, and the pronounced short-distance, environmental gradients related to moisture supply by fog, offer a great opportunity to investigate the physiological mechanisms that explain tree species ability to respond to abrupt and pronounced changes in climate due global warming.

Convergence in leaf traits reducing water loss by transpiration, as well as hydraulic traits favouring safety at the expense of hydraulic efficiency, has been reported for plants that are periodically exposed to severe water deficit (Mitchell et al. 2008; Markesteijn et al. 2011a,b). Such plants usually show narrower and shorter vessels with small pit pores, which are more resistant to drought-induced cavitation, but at the same time have an increased flow resistance and a lower hydraulic efficiency (Hacke et al. 2001; Choat et al. 2005; Mitchell et al. 2008; Markesteijn et al. 2011a,b), affecting leaf water supply. The capacity to maintain leaf turgor in response to decreasing soil moisture availability is also an important mechanism that favours drought tolerance (Kozlowski and Pallardy 2002; Baltzer et al. 2008; Kursar et al. 2009; Bartlett et al. 2012). Water potential at loss turgor point ( πtlp ) is a critical physiological determinant of a plant’s tolerance to water stress

(Bartlett et al. 2012). Plants can reduce πtlp by accumulating osmotically active compounds in the cells (osmotic adjustment) or by increased cell wall ﬂexibility (elasticity, ε). However, recently Bartlett et al. (2012) showed no direct role for ɛ in driving differences in

πtlp across species, instead, elastic adjustments acted to maintain relative water content at turgor loss point (RWC tlp ) despite very negative water potentials at full turgor ( π0) and πtlp.

Here, we measured leaf and hydraulic traits of the three main tree species occurring in fog-inundated rain forest patches of Fray Jorge (semiarid Chile), which show contrasting distribution patterns along the soil moisture gradient produced by fog influx. We also compared pressure-volume curves traits of individuals growing at windward and leeward edges of forest patches.

Specifically, we addressed the following questions: 1) How does the variation in functional traits related to drought tolerance explain species distribution along small-scale moisture gradients? 2) What mechanisms allow individuals growing along the drier leeward edges to cope with reduced water availability (such as osmotic adjustment or increased cell elasticity) in comparison with conspecific individuals growing in wetter patch core habitats? We expect that species growing in small patches and leeward patch edges would display a suite of leaf traits favoring water conservation (such a reduced stomatal conductance) and a suite of hydraulic traits reducing cavitation risk (such as narrow vessels), at the expense of photosynthesis and hydraulic efficiency. We also predict that individuals growing at leeward patch edge would have pressure-volume traits values favoring drought tolerance (such as lower πtlp and π0) in comparison with conspecific individuals growing in wetter patch core.

Tree species occurring in the fog forest of Fray Jorge are exposed to increased aridity due to climatic changes over an extended period of time (Villagrán et al. 2004; Gutiérrez et al. 2008), facing seasonal changes in fog influx that drive pronounced moisture gradients within patches (del-Val et al. 2006). This study aims to reveal some of the basic mechanisms underlying the relative success of these species to coexist given past and current variations in moisture availability. Here, we will further discuss results in the light of the possible consequences of future climate change and its effects on species’ distribution and coexistence.

Materials and methods

Study site and species

Fray Jorge National Park (30°40´S. 71°30´W) comprises the northernmost patches of Chilean temperate rainforests, dominated by broad-leaved evergreen tree species, which exhibit remarkable ﬂoristic afﬁnities with temperate forests located some 1000 km to the south (Villagrán et al. 2004). The area contains a mosaic of about 180 forest patches ranging in size from 0.1 to 36 ha, located on the summits of coastal mountains at an elevation of 450 to 660 m, surrounded by a matrix of semiarid scrub vegetation (Barbosa et

21 al. 2010). The regional climate is Mediterranean-arid with a mean annual rainfall of 147 mm concentrated during the cool winter months from May to August and a mean annual temperature of 13.6°C (López-Cortés and López 2004). Fog is a prominent and constant feature of the landscape above 400 m elevation especially during spring and summer months, when fragments can receive an additional input of at least 200 mm of cloud water annually via throughfall and stemflow (del-Val et al. 2006).

A large 36 ha patch was selected for this study because it was the only one where all three focal tree species coexist. Additional details on the structure and physical gradients of patches are given by Barbosa et al. (2010). The forest patch studied was located at an altitude of 635 m, with average air temperatures inside the patch varying from 9.2°C in spring (October to December) to 13.3°C in winter (July to September) and relative air humidity varying between 83.6% in winter and 99.6% in spring-summer.

The forest canopy is dominated by A. punctatum (Aextoxicaceae), with juveniles occurring more frequently along the edge directly receiving fog influx (windward), but adults found throughout patch, and co-dominated by D. winteri (Winteraceae), which tends to be aggregated inside the patch. M. correifolia (Myrtaceae) is occasionally represented in the canopy of the forest patch (0.3% basal area) but it is confined to the drier leeward edge (Gutiérrez et al. 2008). Volumetric soil moisture varies substantially in both small and large patches. Leeward edges are drier than the other two microhabitats, while soil moisture at the windward edges is comparable with patch core (25 measurements per zone in A. punctatum individuals): small patches; windward: 10.43% ± 1.01; core: 12.13% ± 1.12; leeward: 5.02% ± 0.49 and large patches; windward: 9.25% ± 0.62; core: 14.59% ± 0.72; leeward: 4.72% ± 0.30) (Salgado-Negret unpublished data). Volumetric soil moisture for our species measured at 20 cm depth, varied accordingly across sites occupied by the different tree species (30 measurements per species): D. winteri (22.9% ± 2.66), A. punctatum (13.4% ± 1.7) and M. correifolia (5.3% ± 0.53) (p<0.0001; F=23.01; d.f.=2) (Salgado-Negret unpublished data).

The three species have a different phytoclimatic distribution in Chile: A. punctatum is a tree species endemic of western South America and it is broadly distributed in coastal forests from 30 - 43°S; D. winteri is distributed from Fray Jorge and central Chile to Sub-

Antarctic forest in Tierra del Fuego at 55°S (Villagrán et al. 2004). Finally, M. correifolia is restricted to central Chile with a Mediterranean climate subjected to a cool rainy winter and a summer drought period of 2– 3 months (Di Castri and Hajek 1976).

Leaf traits

We measured leaf traits for six individuals (dbh >10 cm) of each tree species using mature, fully expanded leaves without herbivore damage. All measurements were done on the same six individuals. CO 2 assimilation curves were constructed using the CO 2 reference concentration of 380 ppm, 50% relative humidity, and a temperature of 25° C.

Photosynthesis (A MAX ) and stomatal conductance (g s) were measured in M. correfolia , A. punctatum and D. winteri at 700, 500 and 700 umol m -2 s-1 respectively, with an open portable photosynthesis system (CIRAS-2 CRS068, PP Systems, Amesbury, USA) equipped with a LED light. Measurements were conducted between 10:00 and 13:00 h. After measurements of gas exchange, leaves were cut and leaf water potentials at midday were measured ( ψMD , MPa) using a pressure chamber (Scholander-type, Model 1000 PMS).

We also measured predawn leaf water potentials ( ψPD , MPa) between 5:00 and 7:00 h for the same six individuals per species.

After measurement, leaves were scanned (EPSON Stylus TX200) and analysed using ImageJ software (http://imagej.nih.gov/ij/) to determine leaf area (LA). Finally, leaves were dried for 48h at 65°C to obtain leaf dry mass (g) and calculate leaf mass per area (LMA; g cm -2) (Cornelissen et al. 2003).

Pressure-volume curves

Pressure-volume curves were constructed for six individuals per species. One shoot was cut from each individual and the shoots were hydrated with distilled water in plastic bags to bring leaves to full turgor. Tissue rehydration is necessary to ensure that all samples are near saturation thus allowing for construction of the entire moisture release curve (Baltzer et al. 2008). After 24h of rehydration, we constructed pressure-volume curves following the Sack and Pasquet-Kok protocol ( www.prometheuswiki.com ). Water potentials of the leaves

23 were measured with a Scholander-type pressure chamber (PMS, Model 1000) and the tissue was weighed immediately after measurement. The tissue was dehydrated slightly at room temperature, before re-weighing the leaf mass and re-measuring the water potential. This process was repeated until the tissue reached constant mass. When there was no further decrease in mass, leaves were dried for 48h at 80° C to determine dry mass. The following traits were estimated from the pressure-volume curves: solute potential at full turgor ( π0;

MPa), solute potential at turgor loss point ( πtlp ; MPa), relative water content at turgor loss point (RWC tlp ; %), and the bulk modulus of elasticity ( ɛ; MPa).

Hydraulic traits

Maximum vessel length - One branch (2.5 – 10 mm diameter) was cut from the outer crown of each of six individuals per species and transported to the ﬁeld station. Here, maximum vessel lengths were estimated cutting branches approximately 1 m from the distal apex and applying air pressure (approx. 60 Kpa) (cf. Ewers and Fisher 1989) to the cut end of the branch. The distal end of the branch was then trimmed back approximately 1 cm at a time until air bubbles were seen emerging from vessel ends (Brodribb and Feild 2000). The remaining branch length at this point was then measured as an estimate of Maximum vessel length (MVL; cm).

Sapwood-specific hydraulic conductivity - A second collection of branches was made from the same six individuals per species to measure hydraulic conductivity (water flux through -1 -1 -1 a unit length of stem over a pressure gradient; K h, in kg m s MPa ) following Sperry et al. (1988). In the field station, branches were recut under water to avoid the induction of new embolisms. Distal ends were trimmed with a razor blade to clear any accidentally blocked vessels and about 1 cm of the bark at each side of the branch was removed. While submerged, the shaved end of the branch was wrapped in Paraﬁlm. All branches used for hydraulic conductivity measurements were cut to the same length (approx. 30 cm). The branch was connected to a fluid column fed by a reservoir elevated to a height of 1 m, providing a constant pressure of 9.8 KPa. An electronic balance registered KCl solution flux as an increase in sample mass each 30 seconds. Measurements were taken when an

24 approximately constant flow was observed for at least 3 min. Afterwards, the stems were flushed with KCl solution at a pressure of ≈170 KPa for 10-15 minutes to remove emboli (Sperry et al. 1987) and hydraulic conductivity was measured again at its maximum capacity. We divided K h by the cross-sectional area of the conductive xylem (see methods Hydraulic anatomy below), to standardise the flow of water per unit sapwood area and -1 -1 -1 obtain sapwood specific hydraulic conductivity ( Ks; kg MPa m s ). As such, hydraulic conductivity was made comparable among segments of different diameters.

Hydraulic anatomy . The same stems were then perfused with safranin dye to visualize the conductive wood area. A cross-sectional area of the upper distal end of the stem was photographed with a digital camera mounted on a microscope, at 10x magnificacion and the image was processed using the imaging software SigmaScan Pro 5 (SPSS Inc.) to determine vessel diameter (VD; µm) and density (VDi; vessels mm -2). For each branch, we calculated the Huber value (Hv; cm 2 cm -2) as the cross-sectional sapwood area of the upper distal end of the stem divided by the total supported leaf area. Finally, for each species, vessel diameters were divided into 5 µm size classes to construct frequency histograms. In line with the Hagan-Poiseuille law, the vessel ratios in each size class were raised to the fourth power and summed to determine the relative contribution of each vessel size class to overall hydraulic conductance (Choat et al. 2005).

Data analysis

Differences in leaf traits (LMA, LA, g s and A MAX ), hydraulic traits (vessel diameter and density, Ks and Hv), and traits derived from pressure-volume curves ( π0, πtlp , RWC tlp and ɛ) were contrasted among three tree species using a multivariate analysis of variance (MANOVA). Because MANOVA showed significant species effects, we conducted a series of univariate ANOVAs followed by post-hoc Tukey´s tests to identify individual responses of each trait. Overall multivariate relations and trait differences among species were further explored using a principal components analysis (PCA). Differences in traits derived from pressure-volume curves between leeward and core zones from A. punctatum and M.

25 correifolia individuals were analysed with independent-samples t-tests. Statistical analyses were performed using InfoStat (Di Rienzo et al. 2011).

Results

Species differences in leaf and hydraulic traits

Leaf and hydraulic traits, as well pressure-volume curve related traits, differed substantially among the three coexisting tree species in Fray Jorge forest (MANOVA; Willk´s = 9.9 x E- 05; F = 33.11; p < 0.0001). Trait differentiation among species is best described by principal component analysis. The ﬁrst component, which explained 53% of trait variation, showed an even contribution of variables with a magnitude of 0.3, and it clearly separated M. correifolia from D. winteri , placing A. punctatum at an intermediate position (Fig. 1). This component was negatively correlated with leaf traits that increased water transpiration and carbon gain (LA, gs, Amax), as well as with the solute potential at full turgor ( π0) and the potential at turgor loss point ( πtlp ) (Table 1). Then, higher values along the ﬁrst PCA component reflect stronger ability to conserve water and tolerate to drought, but lower gas exchange rates. PCA component 1 was also positively correlated with vessel diameter (VDi) and negatively correlated with vessel density (VD) (Table 1). The second component explained an additional 25.3% of the total variance and it separated A. punctatum from the other two species. This component was dominated by higher values of RWC tlp and lower values of Hv (Table 1).

Significant differences in leaf traits among species were additionally detected using separate ANOVAs (Table 2). Accordingly, we found that D. winteri , a tree restricted to the moist cores of large patches, exhibited a higher stomatal conductance and photosynthetic rates than the other two species, although its average LMA did not differ from that of A. punctatum . In turn, we found that M. correifolia , a tree that occurs primarily in the drier leeward edges, had the smallest leaf area and lower stomatal conductance and photosynthetic rates. Finally, the most widespread tree species in these patches, A. punctatum, did not differ in stomatal conductance and photosynthetic rates from M. correifolia (Table 2, Fig. 2).

Clear differences among the three species in traits derived from pressure-volume curves were also found (Table 2). The two species with more sclerophyllous leaves, A. punctatum and M. correifolia , showed the lowest πtlp and π0 values, and A. punctatum had the lowest RWC tlp (Table 2). The latter species also had the lowest ɛ, while values between the other two tree species did not differ.

Predawn and midday leaf water potentials varied strongly among species (Table 2). In the summer season, presumably the warmer and drier period of the year, predawn leaf water potentials ( ψPD ) ranged from -0.075 MPa to -0.144 MPa, while midday water potential ( ψMD ) ranged from -0.28 to -0.35 MPa across the three species. Midday leaf water potentials never dropped below the turgor loss point, suggesting that species did not suffer from drought stress during the period of study.

We found significant differences in hydraulic traits among tree species (Table 2). Hydraulic conductivity and vessel densities were higher and vessel diameters were smaller for D. winteri than for the other two species (Table 2, Fig 2). Contrary to our predictions, M. correifolia , the species that is most restricted to the semiarid Mediterranean-climate region, and presumably better adapted to summer drought, had larger vessel diameters than the other two species. A. punctatum , a predominantly coastal tree species, with a broad latitudinal distribution in Chilean forests and in the Fray Jorge forest patch mosaic, showed the lowest hydraulic conductivity, with intermediate vessel diameters and densities (Table 2, Fig. 2). According to the Hagan-Poiseuille law which states that in theory a vessel’s hydraulic conductance is proportional to the fourth power of its radius, D. winteri and A. punctatum hydraulic conductivity depended strongly on the lower vessel size classes (10 to 20 µm), 92.7% and 56.6% respectively (Fig. 3), while M. correifolia showed greater range of diameter classes and had 52% of its hydraulic conductivity accounted for by the wider vessel size class (20 to 30 µm) (Fig. 3).

Trait differences between patch core and leeward edge individuals

We compared traits derived from pressure-volume curves between individuals growing in the patch core (away from edges) and in the leeward edge of the same patches; this comparison was only possible for A. punctatum and M. correifolia as these species co-

27 occur in these two microhabitats. We did not have comparative data for D. winteri , because it was never found in patch edges. Most physiological traits obtained from the pressure- volume curves did not differ between A. punctatum trees in the core and leeward trees (Table 3), except for parameter ɛ. In the latter case, trees on the leeward edge of patches had a lower bulk modulus of elasticity than patch core trees. In contrast, M. correifolia showed clear differences in several attributes between trees sampled in the patch core and in the drier leeward edge. For this species, πtlp and RWC tlp values were lower at the leeward edge than at the patch core (Table 3). In the case of M. correifolia , ɛ did not vary between trees in the core and leeward edge of patches. Significant differences in ψPD and ψMD between trees in patch core and those in the leeward edge were found for both species, with the lowest values found for trees at the leeward edge (Table 3). In contrast to M. correifolia , for A. punctatum trees found at the leeward edge, ψMD dropped below πtlp .

Discussion

Our results indicate that evergreen tree species were able to partition small-scale, but strong soil moisture gradients, fog-dependent forest patches, due to their differential ability to use soil water and tolerate drought-related habitat differences. For the three species dominating the canopy of fog-inundated patches in this semiarid region, we identified a continuous gradient of water-use strategies. Ecophysiological strategies varied between a set of plant traits that allows efficient water transport and high carbon gain, at the one end, to traits that enhance water conservation at the cost of lower gas exchange rates, at the opposite end. At one end of the continuum we find D. winteri , a tree species restricted to wet microhabitats in the core of large forest patches, which has high Ks, leaf area, photosynthetic rates and stomatal conductance. The opposite end of this gradient is occupied by M. correifolia, a species that is typically found in drier microhabitats of the leeward edges and in small forest patches, showing traits that imply increased drought tolerance, such as a small leaf area, reduced stomatal conductance and hydraulic conductivity, and low water potentials at turgor loss point. Finally, A. punctatum , the most abundant and widespread species in different microhabitats of Fray Jorge forest patches, displays intermediate values for the drought-tolerance traits investigated. The morphological and physiological differences

28 detected among tree species in this ecosystem are likely to be important in shaping species- specific responses to future reductions in water availability as produced by reductions in fog frequency and rainfall, that are predicted for this and other semiarid regions in the coming decades (Johnstone and Dawson 2010).

In this forest, D. winteri showed the broadest leaf area, highest photosynthetic rates and greatest stomatal conductance, which are associated with the highest KS. High conductivity contributes to a more efficient water supply to the leaves, supporting greater carbon assimilation (Meinzer et al. 1995; Sperry 2000; Brodribb and Feild 2000; Santiago et al. 2004). Still, in contrast with the former suite of traits, D. winteri had the smallest vessel diameters and the highest vessel density among species. D. winteri is an angiosperm, but belongs to the very primitive family Winteraceae, which does not have true vessels, but instead tissues that are very similar to the tracheids of coniferous species. Species with such vesselless wood are known to have up to 21 times lower inter-element pit resistance than eudicot vessels, and therefore their wood is highly conductive despite the short length and narrow diameter of tracheids (Hacke et al. 2006, 2007; Sperry et al. 2007). Despite its high

Ks, large leaf area, and high stomatal conductance, D. winteri has a reduced ability to regulate water loss (Feild et al. 1998). Low stomatal control in D. winteri is probably associated with its hydrophobic granular plug, which consists of a porous, granular material that fills the stomatal cavity above the guard cells preventing them from fully closing (Feild et al. 1998; Feild and Holbrook 2000). This seems to be an adaptation to humid environments, where it precludes the formation of a permanent water film on the leaf surface that would obstruct CO 2 diffusion into the leaf (Feild et al. 1998). Consequently, a reduced ability to regulate water loss in D. winteri implies a greater hydraulic demand that cannot be satisfied under the drier conditions that characterize small forest patches or patch edges in Fray Jorge. Species, such as D. winteri , will be more vulnerable to increased moisture stress at patch edges, as created by fragmentation. This will be further accentuated by the regional reductions in rainfall or fog inputs and will likely reduce the possibility that this species are able to maintain a viable population in the future.

By contrast, M. correifoli a, which is typically found in drier microhabitats in Fray Jorge, showed an opposite suite of traits compared to D. winteri , including smaller leaf areas, higher LMA, and a reduced stomatal and hydraulic conductance. The combination of

29 these traits will enhance water conservation under water stress, but have a cost on gas exchange rates. A MAX measured in the field in M. correifolia was two times lower than in the less-stress tolerant D. winteri . M. correifolia also showed a greater range of vessel diameter classes than the other two species, implying greater functional diversity for this trait. Wider vessels are more efficient in water transport and could be useful in wetter habitats and wetter periods of the year, while in the drier season or drier habitats, when wider vessels are more prone to cavitation, M. correifolia can use its narrower vessels to maintain water transport. The wider range of vessel sizes exhibited by M. correifolia likely explains the ability of this species to cope with the strong fluctuations in water availability that characterize small patches and leeward edges (Barbosa et al. 2010), where it is found. Accordingly, among the three species studied, M. correifolia is the most capable of tolerating a substantial increment of climatic variability and more extreme droughts as expected from global climate change in this region. M. correifolia is thus most likely to profit from the altered climate conditions as expected for this region.

Finally , A. punctatum , the most abundant and widespread species in Fray Jorge forest patches had similar levels of stomatal and hydraulic conductance and photosynthetic rates as M. correifolia , even though A. punctatum is a temperate tree species with sclerophyllous leaves that generally occurs in areas of higher rainfall at higher latitudes in south-central Chile. We suggest that the unexpectedly low values of πtlp recorded in this species could be a response to the strong effects of oceanic salt spray over most of its coastal distribution (Pérez and Villagrán 1985), which is intercepted by the crown foliage and branches, and conducted to the soil via throughfall and stemﬂow (Ponette-González et al. 2009). The high salt content of marine spray and rain water in Chilean coastal forests (Hedin et al. 1995) can reduce soil osmotic potential and thus soil water potential, forcing tree species to limit leaf water potential as a mechanism to sustain soil water absorption and transport. Individuals of A. punctatum growing in the forest patch core have lower ψPD values than the more drought tolerant M. correifolia , which may have deeper root systems, which allows a better access to deeper soil water reserves. This could also explain why this species showed a greater capacity to rehydrate overnight than the other two evergreen tree species in Fray Jorge.

Despite interspecific differences in ψPD and ψMD , all three species showed ψMD values higher than πtlp when growing in the forest patch core, confirming that frequent summer fog in Fray Jorge represents an effective physical buffer against diurnal temperature ﬂuctuations and desiccation that characterize the semiarid surrounding vegetation (del-Val et al. 2006; Ewing et al. 2009). Trees of M. correifolia and A. punctatum occurring at the leeward edge of Fray Jorge forest patches had lower ψPD values than those trees occurring in the patch core, showing that trees along edges have more limited access to soil moisture and lower capacity to rehydrate and recover leaf water status overnight. For the more drought resistant M. correifolia, ψMD values were higher than πtlp values, but in the case of A. punctatum they were lower. These results indicate that in leeward patch edges, A. punctatum , but not M. correifolia , experiences more water stress, and therefore it might be unable to recover its leaf water status overnight, after losing substantial water by transpiration during the day. Interspecific differences can be best explained by pressure-volume curves. Trees of M. correifolia in the leeward edge had the most negative osmotic potentials at full turgor and at turgor loss point, and lower cell water content at turgor loss point than trees of the same species in the patch core. In turn, the modulus of elasticity did not vary between habitats. According to these results, M. correifolia appears to be able to tolerate (rather than avoid) drought (Bartlett et al. 2012) by adjusting its osmotic potential at cell level, as reflected in a reduced π0. Such compensatory osmotic mechanisms have been described in south-central Chile for the trees Kageneckia oblonga (Cabrera 2002) and Eucryphia cordifolia (Figueroa et al. 2010). In contrast to M. correifolia , A. punctatum did not show much variation in pressure-volume parameters between trees in patch core habitats and leeward edges, except for ɛ, resulting in ψMD values lower than πtlp values, and therefore, significant water stress at leeward edges. This suggests that over longer time periods, increased water stress can result in a negative water balance for A. punctatum trees that occur in the leeward edge of patches. This might also explain the increased mortality rates and lower regeneration of A. punctatum observed in leeward edges compared to windward edges and patch cores (del-Val et al. 2006). Considering that global change scenarios for this region of the world predict increased patch fragmentation (Sala et al. 2000), and therefore enhanced edge effects in forested landscapes, A. punctatum trees will be at increased risk of mortality due to drought conditions along patch edges. In

Fray Jorge, the disruption of the canopy of A. punctatum in forest patches, due to enhanced drought or lower fog inputs, may substantially reduce the fog interception capacity of patches. In turn, this could modify the hydrological balance of the forest and affect the regeneration and persistence of other tree species that dependent on the fog capture by the A. punctatum canopy (Gutiérrez et al. 2008).

Overall, our findings support the broader concept that along pronounced soil moisture gradients driven by fog interception in forest patches, tree functional diversity is strongly linked to interspecific differences in drought tolerance and/or efficiency of water use. We emphasize that plant hydraulic traits play a fundamental role in explaining niche differentiation among species in patch center-to-edge habitats and their quantitative understanding is key to predict how forests will respond to future scenarios of land use and climate change.

Acknowledgements

We would like to express our gratitude to Leonardo Ramirez, Felipe Albornoz, Rafaella Canessa, Aurora Gaxiola, Paulina Lobos, Juan Monardez, Carmen Ossa, Daniel Salinas, Daniel Stanton and Patricio Valenzuela for their invaluable assistance in the ﬁeld and useful discussions and comments on the manuscript. This work was supported by CONICYT fellowship 24110074 to B.S-N., and grants Fondecyt 1110929 to F.P., ICM P05-002 and PFB-23 from CONICYT to the Institute of Ecology and Biodiversity. This is a contribution to the LINC-Global and Research Program of the Chilean LTSER network at Fray Jorge National Park.

References

Baltzer JL, Davies SJ, Bunyavejchewin S, Noor NSM (2008) The role of desiccation tolerance in determining tree species distributions along the Malay–Thai Peninsula. Funct Ecol 22:221-231. doi: 10.1111/j.1365-2435.2007.01374.x Barbosa O, Marquet PA, Bacigalupe LD, Christie DA, del-Val E, Gutiérrez AG, Jones CG, Weathers KC, Armesto JJ (2010) Interactions among patch area, forest structure and water fluxes in a fog-inundated forest ecosystem in semi-arid Chile. Funct Ecol 24:909-917. doi: 10.1111/j.1365-2435.2010.01697.x Bartlett M, Scoffoni C, Sack L. (2012) The determinants of leaf turgor loss point and prediction of drought tolerance of species and biomes: a global meta-analysis. Ecol Lett 15:393-405. doi: 10.1111/j.1461-0248.2012.01751.x Bongers F, Poorter L, Van Rompaey RSAR, Parren MPE (1999) Distribution of twelve moist forest canopy tree species in Liberia and Cote d’Ivoire: response curves to a climatic gradient. J Veg Sci 10:371-382. doi: 10.2307/3237066 Brodribb TJ, Feild TS (2000) Stem hydraulic supply is linked to leaf photosynthetic capacity: evidence from New Caledonian and Tasmanian rainforests. Plant Cell Environ 23:1381-1388. doi: 10.1046/j.1365-3040.2000.00647.x Cabrera HM (2002) Ecophysiological responses of plants in ecosystems with Mediterranean-like climate and high mountain environments. Rev Chil Hist Nat 75:625-637 Cereceda P,Osses P, Larraín H, Farías M, Lagos M, Pinto R, Schemenauer RS (2002) Advective, orographic and radiation fog in the Tarapacá region, Chile. Atmos Res 64:261-271 Choat B, Marilyn CB, Luly JG, Holtum JAM (2005) Hydraulic architecture of deciduous and evergreen dry rainforest tree species from north-eastern Australia. Trees-Struct Funct 19:305-311. doi: 10.1007/s00468-004-0392-1 Condit R (1998) Ecological implications of changes in drought patterns: shift in forest composition in Panama. Climatic Change 39:413-427. doi: 10.1023/A:1005395806800

Condit R, Pitman N, Leigh Jr EG, Chave J, Terborgh J, Foster RB, Nuñez-V P, Aguilar S, Valencia R, Villa G, Muller-Landau HC, Losos E, Hubbell SP (2002) Beta-diversity in tropical forest trees. Science 295:666-669. doi: 10.1126/science.1066854 Cornelissen JHC, Lavorel S, Garnier E, Díaz S, Buchmann N, Gurvich DE, Reich PB, ter Steege H, Morgan HD, van der Heijden MGA, Pausas JG, Poorter H (2003) A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Aust J Bot 51:335-380. doi: 10.1071/BT02124 Dawson TE (1998) Fog in the California redwood forest: ecosystem inputs and use by plants. Oecologia 117:476-485. doi: 10.1007/s004420050683 del-Val E, Armesto JJ, Barbosa O, Christie DA, Gutiérrez AG, Jones CG, Marquet PA, Weathers KC (2006) Rain forest islands in the Chilean semiarid region: fog- dependency, ecosystem persistence and tree regeneration. Ecosystems 9:598-608. doi: 10.1007/s10021-006-0065-6 Di Castri F, Hajek E (1976) Bioclimatología de Chile. Universidad Católica de Chile, Santiago, Chile Di Rienzo JA, Casanoves F, Balzarini MG, Gonzalez L, Tablada M, Robledo CW. InfoStat versión 2011. Grupo InfoStat, FCA, Universidad Nacional de Córdoba, Argentina. URL http://www.infostat.com.ar Engelbrecht BMJ, Comita LS, Condit R, Kursar TA, Tyree MT, Turner BL, Hubbell SP (2007) Drought sensitivity shapes species distribution patterns in tropical forests. Nature 447:80-83. doi:10.1038/nature05747 Ewers FW, Fisher JB (1989) Techniques for measuring vessel lengths and diameters in 17 stems of woody plants. Am J Bot 76:645-656 Ewing HA, Weathers KC, Templer PH, Dawson TE, Firestone MK, Elliott AM, Boukili VKS (2009) Fog water and ecosystem function: heterogeneity in a California Redwood forest. Ecosystems 12:417-433. doi: 10.1007/s10021-009-9232-x Feild TS, Zwieniecki MA, Donoghue MJ, Holbrook NM (1998) Stomatal plugs of Drimys winteri (Winteraceae) project leaves from mist but not drought. Proc Natl Acad Sci USA 95:14256-14259. doi: 10.1073/pnas.95.24.14256

Feild TS, Holbrook NM (2000) Xylem sap flow and stem hydraulics of the vesselless angiosperm Drimys granadensis (Winteraceae) in a Costa Rican elfin forest. Plant Cell Environ 23: 1067-1077. doi: 10.1046/j.1365-3040.2000.00626.x Figueroa JA, Cabrera HM, Queirolo C, Hinojosa LF (2010) Variability of water relations and photosynthesis in Eucryphia cordifolia Cav. (Cunoniaceae) over the range of its latitudinal and altitudinal distribution in Chile. Tree Physiol 30:574-585. doi: 10.1093/treephys/tpq016 Gentry AH (1988) Changes in plant community diversity and floristic composition on environmental and geographic gradients. Ann Mo Bot Gard 75:1-34 Gutiérrez AG, Barbosa O, Christie DA, del-Val E, Ewing HA, Jones CG, Marquet PA, Weathers KC, Armesto JJ (2008) Regeneration patterns and persistence of the fog dependent Fray Jorge forest in semiarid Chile during the past two centuries. Glob Change Biol 14:161-176. doi: 10.1111/j.1365-2486.2007.01482.x Hacke UG, Sperry JS, Pockman WT, Davis SD, McCulloh K (2001) Trends in wood density and structure are linked to prevention of xylem implosion by negative pressure. Oecologia 126:457-461. doi: 10.1007/s004420100628 Hacke UG, Sperry JS, Wheeler JK, Castro L (2006) Scaling of angiosperm xylem structure with safety and efficiency. Tree Physiol 26:689-701. doi: 10.1093/treephys/26.6.689 Hacke UG, Sperry JS, Feild TS, Sano Y, Sikkema EH, Pittermann J (2007) Water transport in vesselless angiosperms: conducting efficiency and cavitation safety. Int J Plant Sci 168:1113-1126. doi: 10.1086/520724 Hedin LO, Armesto JJ, Johnson AH (1995) Patterns of nutrient loss from unpolluted, old- growth temperate forests: evaluation of biogeochemical theory. Ecology 76:493-509 Hildebrandt A, Eltahir EAB (2006) Forest on the edge: seasonal cloud forest in Oman creates its own ecological niche. Geophys Res Lett 33:L11401. doi: 10.1029/2006GL026022 Johnstone JA, Dawson TE (2010) Climatic context and ecological implications of summer fog decline in the coast redwood region. Proc Natl Acad Sci USA 107:4533–4538. doi: 10.1073/pnas.0915062107

Katata G, Nagai H, Kajino M, Ueda H, Hozumi Y (2010) Numerical study of fog deposition on vegetation for atmosphere–land interactions in semi-arid and arid regions. Agric For Meteorol 150:340-353 Kozlowski TT, Pallardy SG (2002) Acclimation and adaptive responses of woody plants to environmental stresses. The Botanical Review 68:270-334. doi: 10.1663/0006-8101 Kursar TA, Engelbrecht BMJ, Burke A, Tyree MT, Omari BE, Giraldo JP (2009) Tolerance to low leaf water status of tropical tree seedlings is related to drought performance and distribution. Funct Ecol 23:93-102. doi: 10.1111/j.1365-2435.2008.01483.x López-Cortés F, López D (2004) Antecedentes bioclimáticos del Parque Nacional Bosque Fray Jorge. In: Squeo FA, Gutiérrez JR, Hernández IR (eds). Historia natural del parque nacional bosque Fray Jorge. Ediciones Universidad de La Serena, Chile, pp 45-60 Markesteijn L, Poorter L, Bongers F, Paz H, Sack L (2011a) Hydraulics and life history of tropical dry forest tree species: coordination of species’ drought and shade tolerance. New Phytol 191:480-495. doi: 10.1111/j.1469-8137.2011.03708.x Markesteijn L, Poorter L, Paz H, Sack L, Bongers F (2011b) Ecological differentiation in xylem cavitation resistance is associated with stem and leaf structural traits. Plant Cell Environ 34:137-148. doi: 10.1111/j.1365-3040.2010.02231.x Meinzer FC, Goldstein G, Jackson P, Holbrook NM, Gutiérrez MV, Cavelier J (1995) Environmental and physiological regulation of transpiration in tropical forest gap species: the influence of boundary layer and hydraulic properties. Oecologia 101:514-522. doi: 10.1007/BF00329432 Mitchell PJ, Veneklaas EJ, Lambers H, Burguess SSO (2008) Using multiple trait associations to deﬁne hydraulic functional types in plant communities of south- western Australia. Oecologia 158:385-397. doi: 10.1007/s00442-008-1152-5 Pérez C, Villagrán C (1985) Distribution of species abundances in relict forests of the Mediterranean zone of Chile. Rev Chil Hist Nat 58:157-170 Ponette-González A, Weathers KC, Curran LM (2009) Water inputs across a tropical montane landscape in Veracruz, Mexico: synergistic effects of land cover, rain and fog seasonality, and interannual precipitation variability. Glob Change Biol 16:46- 49. doi: 10.1111/j.1365-2486.2009.01985.x

Poorter L (2007) Are species adapted to their regeneration niche, adult niche, or both? Am Nat 169:433-442 Pyke CR, Condit R, Aguilar S, Lao S (2001) Floristic composition across a climatic gradient in a neotropical lowland forest. J Veg Sci 12:553-566. doi: 10.2307/3237007 Sala OE, Chapin FS III, Armesto JJ, Berlow E, Bloomfield J, Dirzo R, Huber-Sanwald E, Huenneke LF, Jackson RB, Kinzig A, Leemans R, Lodge DM, Mooney HA, Oesterheld M, Poff NL, Sykes MT, Walker BH, Walker M, Wall DH (2000) Global biodiversity scenarios for the year 2100. Science 287:1770-1774. doi: 10.1126/science.287.5459.1770 Santiago LS, Goldstein G, Meinzer FC, Fisher JB, Machado K, Woodruff D, Jones T (2004) Leaf photosynthetic traits scale with hydraulic conductivity and wood density in Panamanian forest canopy trees. Oecologia 140:543-550. doi: 10.1007/s00442-004-1624-1 Sperry JS, Holbrook MN, Zimmermann M, Tyree MT (1987) Spring filling of xylem vessels in wild grapevine. Plant Physiol 83:414-417 Sperry JS, Donnelly JR, Tyree MT (1988) A method for measuring hydraulic conductivity and embolism in xylem. Plant Cell Environ 11:35-40. doi: 10.1111/j.1365- 3040.1988.tb01774.x Sperry JS (2000) Hydraulic constraints on plant gas exchange. Agric For Meteorol 104:13- 23 Sperry JS, Hacke UG, Feild TS, Sano Y, Sikkema EH (2007) Hydraulic consequences of vessel evolution in angiosperms. Int J Plant Sci 168:1127-1139. doi: 10.1086/520726 Sterck F, Markesteijn L, Schieving F, Poorter L (2011) Functional traits determine trade- offs and niches in a tropical forest community. Proc Natl Acad Sci USA 108:20627- 20632. doi: 10.1073/pnas.1106950108 Villagrán C, Armesto JJ, Hinojosa LF, Cuvertino J, Pérez C, Medina C (2004) El enigmático origen del bosque relicto de Fray Jorge. In: Squeo FA, Gutiérrez JR, Hernández IR (eds). Historia natural del parque nacional bosque Fray Jorge. Ediciones Universidad de La Serena, Chile, pp 3-43

Weathers KC, Lovett GM, Likens GE, Caraco NFM (2000) Cloudwater inputs of nitrogen to forest ecosystems in southern Chile: Forms, fluxes and sources. Ecosystems 3:590-595. doi: 10.1007/s100210000051 Wright SJ (1992) Seasonal drought, soil fertility and the species density of tropical forest plant communities. Trends Ecol Evol 7:260-263

Table 1. Eigenvector scores of leaf and hydraulic traits in two main PCA axes. Values in parentheses indicate the percentage of total variance accounted by each axis. Traits are abbreviated as; LA = Leaf area, LMA = Leaf mass area; A MAX = Photosynthetic rate; g s =

Stomatal conductance; π0 = Solute potential at full turgor; πtlp = Water potential at turgor loss; RWC tlp = Relative water content at turgor loss point; ɛ = Bulk modulus of elasticity;

VD = Vessel density; VDi = Vessel diameter; K s = Sapwood-specific hydraulic conductivity; Hv = Huber value. Variables PCA 1 (52.7%) PCA 2 (25.3%) LA -0.34 -0.06 LMA 0.34 0.21

AMAX -0.31 0.23 gs -0.32 0.23

π0 -0.35 0.00

πtlp -0.32 0.13

RWC tlp -0.04 0.51 ɛ 0.22 0.42 VD -0.35 -0.15 VDi 0.37 0.11

Ks -0.2 0.39 Hv -0.03 -0.45

Table 2. Among-species variation in leaf and hydraulic traits. Values indicate the mean ±

SE for each trait per species (n=6). F (2,15) and p values come from univariate ANOVA. Letters represent statistical differences between species for each trait according to a post- hoc Tukey test ( α = 0.05). For trait abbreviations, see Table 1. M. correifolia A. punctatum D. winteri F p Leaf traits LA 6.12 ± 0.38 (a) 30.04 ± 3.07 (b) 40.20 ± 4.48 (b) 30.93 <0.0001 LMA 0.02 ± 0.0006 (b) 0.0098 ± 0.0005 (a) 0.0094 ± 0.0003 (a) 256.20 <0.0001

AMAX 3.48 ± 0.12 (a) 4.34 ± 0.45 (a) 7.58 ± 0.47 (b) 31.59 <0.0001 gs 44.08 ± 3.98 (a) 47.83 ± 5.41 (a) 77.42 ± 3.69 (b) 17.03 0.0001 Traits derived from pressure-volume curve

π0 -0.89 ± 0.05 (a) -0.69 ± 0.07 (b) -0.57 ± 0.03 (b) 11.63 0.0009

πtlp -1.18 ± 0.05 (a) -1.05 ± 0.10 (a) -0.81 ± 0.02 (b) 8.61 0.0032

RWC tlp 96.51 ± 0.23 (b) 95.26 ± 0.23 (a) 96.98 ± 0.23 (b) 15.02 0.0003 ɛ 24.66 ± 1.87 (b) 12.13 ± 1.13 (a) 18.81 ± 1.83 (b) 14.48 0.0003

ψPD -0.075 ± 0.01 (a) -0.144 ± 0.01 (b) -0.138 ± 0.012 (b) 25.6 <0.0001

ψMD -0.35 ± 0.05 (a) -0.32 ± 0.04 (a) -0.28 ± 0.01 (a) 0.94 0.4127 Hydraulic traits VD 158.57 ± 10.41 (a) 294.46 ± 14.69 (b) 310.92 ± 9.58 (b) 50.43 <0.0001 VDi 21.42 ± 0.42 (c) 16.55± 0.55 (b) 15.01 ± 0.11 (a) 67.98 <0.0001

Ks 0.44 ± 0.04 (a) 0.38 ± 0.03 (a) 0.59 ± 0.02 (b) 11.12 0.0011 Hv 5 x 10 -6 ± 1 x 10 -6 (a) 1.5 x 10 -5 ± 3 x 10 -6 (b) 4 x 10 -6 ± 2.42 x 10 -7(a) 10.33 0.0015

Table 3. Differences in pressure-volume curve traits between individual trees growing in the patch cores (core) and leeward edge (edge) for two dominant tree species; Aextoxicon punctatum and Myrceugenia correifolia . T (10) and p values result from t-test. For trait abbreviations, see Table 1. Species Traits Core Edge t-test p

π0 -0.69 ± 0.07 -0.84 ± 0.05 1.85 0.0939

πtlp -1.05 ± 0.10 -1.18 ± 0.05 1.14 0.2805

A. punctatum RWC tlp 95.26 ± 0.23 95.18 ± 0.58 0.13 0.8992 ɛ 12.13 ± 1.13 17.94 ± 1.17 -3.57 0.0051

ψPD -0.144 ± 0.01 -0.77 ± 0.05 -8.42 <0.0001

ψMD -0.32 ± 0.04 -1.39 ± 0.04 -10.82 <0.0001

π0 -0.89 ± 0.05 -1.15 ± 0.16 1.58 0.1658

πtlp -1.18 ± 0.05 -1.58 ± 0.15 2.47 0.0487

RWC tlp 96.51 ± 0.23 95.51 ± 0.23 3.12 0.0108 M. correifolia ɛ 24.66 ± 1.87 24.51 ± 3.05 0.04 0.9673

ψPD -0.075 ± 0.01 -0.32 ± 0.02 -14.44 <0.0001

ψMD -0.35 ± 0.05 -0.97 ± 0.05 -5.39 0.0007

Figure caption

Figure 1. Principal Component Analysis (PCA) of hydraulic and leaf traits of tree individuals of three species in Fray Jorge forest patches. Eigenvector scores of all traits along PCA axes are given in Table 1. Species are abbreviated as: Ap = Aextoxicon punctatum ; Dw = Drimys winteri ; Mc = Myrceugenia correifolia .

Figure 2. Differences in leaf and hydraulic traits for three tree species in the Fray Jorge forest patches in semiarid Chile. a) Leaf area, b) Leaf mass area, c) Photosynthetic rate, d) Stomatal conductance, e) Vessel/tracheid density, f) Vessel/tracheid diameter, g) Sapwood- speciﬁc hydraulic conductivity, and h) Huber value. Bars represent means ± SE. Letters above the bars represent statistical differences between species for each trait resulting from univariate ANOVA with a post-hoc Tukey tests (see table 2 for statistics).

Figure 3. Frequency distributions of xylem vessel diameters in cross sections of branches of three tree species in Fray Jorge forest patches: a) A. puncatum ; b) M. correifolia ; c) D. winteri . Plots show the number of vessels in 5 µm size classes as percentages of the total number of vessels in a given cross sectional area (black bars) and the contribution of each size class to the theoretical hydraulic conductance ( Ʃd4) of the branch (following the Hagan-Poiseuille Law) (grey bars). Bars represent mean values ± SE (n=6).

5.00

2.50 Dw. 3 Dw. 5 Mc. 3 Dw. 6 Mc. 5 Mc. 4 Dw. 1 Mc. 2 Dw. 4 Mc. 6 Dw. 2 0.00 Mc. 1 Ap. 4 Ap. 3 Ap. 1 PCA axis PCA (25.3%) 2 Ap. 2 -2.50 Ap. 5 Ap. 6

-5.00 -5.00 -2.50 0.00 2.50 5.00 PCA axis 1 (52.7%)

Fig. 1.

50 ) 350 a) b -2 e) b b 300 40 p < 0.0011p < 0.0011 ) p < 0.0001 b

2 250 30 200 a 20 150 100

Leaf area (cm Leafarea 10 a 50

0 (mm density Vessel/tracheid 0 0.030 25 m)

) b) f) c µ

-1 b p < 0.0011 0.025 20 b 0.020 p < 0.0001 a 15 0.015 a a 10 0.010

0.005 5 Leafmass (g cm area

0.000 ( diameter Vessel/tracheid 0 10 0.7 ) c) g) b -1 b ) 0.6 p = 0.0011 s -1

8 s -2 p < 0.0001 a

-1 0.5 m

m a 2 6 a -1 0.4 0.3 4 a molCO

µ 0.2 (Kg MPa (Kg

2 s

K 0.1 Amax( 0 0.0 100 2e-5

) h) ) d) b

b -2 -1 p = 0.0001

s 80 2e-5 p = 0.0015 cm -2 2 60 a Om a 2 1e-5 40 a molH

µ 5e-6 a 20 gs ( gs Huber value (cm value Huber 0 0 i a m m r u te tu olia n ifoli ta if tat winteri c wi . re n . orre D u D . c . punc M A M. cor A. p Species Species Fig. 2.

70 a) 60 50 40

(%) 30 20 10 0 70 b) 60 50 40

(%) 30 20 10 0 70 c) 60 50 40

(%) 30 20 10 0 10 - 15 15 -10 20 -15 25 -20 30 -25 35 -30 40 -35 45 -40 50 -45 Diameter class ( µm)

Fig. 3.

CAPÍTULO II

III. Variación en rasgos funcionales explica la distribución de Aextoxicon punctatum a través de un fuerte gradiente de humedad en un bosque fragmentado dependiente de neblina

Salgado-Negret B, Pérez F, Canessa R, Valladares F, Armesto JJ. Variation in functional traits explains the distribution of Aextoxicon punctatum across a strong moisture gradient in a fragmented fog dependent forest. American Journal of Botany (Submitted)

Variation in functional traits explains the distribution of Aextoxicon punctatum across a strong moisture gradient in a fragmented fog dependent forest

Beatriz Salgado Negret 1,2* , Fernanda Pérez 1,2,3 , Rafaella Canessa 1, Fernando Valladares 3 and Juan J. Armesto 1,2,3

1Departamento de Ecología, Pontificia Universidad Católica de Chile, Casilla 114-D, Santiago, Chile.

2Instituto de Ecología y Biodiversidad, Casilla 653, Santiago, Chile.

3LINCGlobal, Museo Nacional de Ciencias Naturales, CSIC, Serrano 115 dpdo, E-28006 Madrid, Spain.

*Corresponding author:

Beatriz Salgado Negret

[email protected]

56-2-3542637

Abstract

- Premise of the study : Climate change and fragmentation are major threats to world forests. Understanding how functional traits related to drought tolerance change across small-scale, pronounced moisture gradients in fragmented forests is important to predict species’ responses to these threats. - Methods : In the case of Aextoxicon punctatum, a dominant canopy tree in fog- dependent rain forest patches in semiarid Chile, we explored how the magnitude, variability and correlation patterns of leaf and hydraulic traits varied across pronounced soil moisture gradients established within and among forest patches of different size, which are associated the differences in tree establishment and mortality patterns. - Key results: Leaf traits varied across soil-moisture gradients produced by fog interception from windward to leeward edges of patches. At drier leeward edges trees showed higher LMA, trichome and stomatal densities than trees from the wetter patch core and windward zones. In contrast, xylem anatomy traits did not vary causing loss of hydraulic conductivity at drier leeward edges. We also detected higher phenotypic integration and variability at the drier leeward edges. - Conclusions : The ability of A. punctatum to modify leaf traits in response to differences in soil moisture availability established over short distances (<500 m) facilitates its persistence in contrasting microhabitats within forest patches. However, xylem anatomy showed limited plasticity, which increases cavitation risk at leeward edges. Greater patch fragmentation, together with fluctuations in irradiance and soil moisture in small patches, could result in higher risk of drought-related tree mortality, with profound impacts on hydrological balances at the ecosystem scale. Intensification of drought due to increasing fragmentation and enhanced edge effects can seriously threaten the future persistence of many tree species in a warming world.

Key-words Climate change; fog-dependent forest; fragmentation; hydraulic traits; intraspecific phenotypic variability; leaf traits; moisture gradient; phenotypic integration.

Introduction

Reductions in precipitation expected under climate change and increasing forest fragmentation are major threats to temperate forests worldwide (Breshears et al., 2005; Echeverría et al., 2006; Choat et al., 2012). Particularly sensitive are forests located in the boundary with drier formations, where drought intensification may be fundamentally important for the persistence of forest communities (Pockman and Sperry, 2000; Engelbrecht et al., 2007; Choat et al., 2012; Salgado-Negret et al., 2013). As a consequence, improved understanding functional trait variation in relation to drought tolerance becomes critical for modeling and predicting tree species responses to future climate change (Anderegg et al., 2013). Variation in functional traits can derive from phenotypic plasticity, genetic variation, developmental instability, and direct effects of stress on plant performance, or a combination of these mechanisms (Matesanz et al., 2010; Valladares et al., 2007; Gianoli and Valladares, 2012). In recent years, interest in drought-resistance trait variation at the intraspecific level has increased (Choat et al., 2007; Cornwell and Ackerly, 2009; Figueroa et al., 2010; Fajardo and Piper, 2012), because of its relevance to understanding plant species responses to drought stress and the maintenance of biodiversity (Violle et al., 2012). Studies have often focused on species distributions across broad geographic ranges and stress conditions (Choat et al., 2007; Figueroa et al., 2010; Fajardo and Piper, 2012). However, intraspecific variation of functional plant traits across pronounced environmental gradients at small spatial scales can also provide clues to identifying species responses to key environmental factors, such as water availability, and their interactions with widespread global change threats such as fragmentation (Matesanz et al., 2009). In semiarid regions, forests that depend on coastal fogs for water supply (Hildebrandt and Eltahir, 2006; del-Val et al., 2006; Katata et al., 2010) represent an interesting case with respect to acute moisture gradients. In such forests, fog interception by trees is the primary or even the only source of moisture during prolonged dry periods (Dawson 1998; del-Val et al., 2006; Ewing et al., 2009). Fog influx creates pronounced asymmetries between windward to leeward edges of forest patches (Weathers et al., 2000; del-Val et al., 2006; Ewing et al., 2009; Stanton et al. 2013), as well as among different-size patches with contrasting edge effects. Fragmentation enhances sensitivity to current and future changes in fog water supply

(Gutierrez et al., 2008; Hildebrandt and Eltahir, 2008; Johnstone and Dawson, 2010). Changes in fog frequency and intensity are predicted to occur in these areas due to changes in sea-surface temperature and the height of the temperature inversion layer (Cereceda et al., 2002, Garreaud et al., 2008), together with changes in other forest features affecting fog capture (Hildebrandt and Eltahir, 2006). An emblematic example of fog-dependent forests found in semiarid Chile (30°S) is the northernmost extension of temperate rainforest on coastal hilltops of the semiarid region. Here, a mosaic of rain forest patches of different sizes occurs immersed in a xerophytic shrubland matrix (Barbosa et al., 2010). The dominant tree species in all forest patches is the southern South American endemic Aextoxicon punctatum Ruiz and Pav, belonging to the monotypic and isolated family Aextoxicaceae. This species is broadly distributed in temperate rain forests of western South America. In fray Jorge, it occurs in forest patches of all sizes and throughout the soil moisture gradient produced by fog influx from windward to leeward edges (del-Val et al., 2006). Moreover, population genetic studies suggest that gene flow via seed dispersal across neighboring patches in this patchy landscape has been highly signiﬁcant (Fst < 0.05) during recent history (Nuñez-Ávila et al. 2013). Patterns of tree radial growth and regeneration dynamics of A. punctatum in this forest have shown constant growth and continuous regeneration for 200 years, despite a declining trend in rainfall during the last century. This suggests that this species can survive extreme temporal fluctuations in water availability (Gutiérrez et al., 2008). Understanding the ability of A. punctatum to withstand spatial and temporal fluctuations in water availability requires improved knowledge of the mechanisms involved in drought tolerance and vulnerability to the combined effects of increased water shortage and forest fragmentation. Plants often respond to water deficit by modifying leaf traits and decreasing transpirational water losses through reductions in stomatal size and density, greater trichome density (Fahn, 1986; Baldini et al., 1997), and enhanced leaf mass per unit area (LMA) (Niinemets, 2001). We know less about hydraulic features conferring drought tolerance, but it has been reported that a large number of short, narrow, vessels per unit area are adaptive under arid conditions (Carlquist, 2001) and reduced the chances of hydraulic embolism (Markesteijn et al., 2011). The above-cited studies have generally focused on changes in mean trait values, while changes in trait variability (measured by the coefficient of variation) have received less attention (Violle et al., 2012). Likewise, comparative studies across moisture gradients have often ignored

50 coordinated trait responses (Nicotra et al. 2007). Coordinated variation of morphological traits can result from genetic, developmental and functional relationships among traits, combined in the concept of phenotypic integration (Murren, 2002; Pigliucci, 2003). Correlations between leaf (Wright et al., 2004) and hydraulic traits (Chave et al., 2009; Zanne et al., 2010) have been documented by several recent studies (Brodribb and Field, 2000; Brodribb et al., 2002; Santiago et al., 2004; Wright et al., 2006; Meinzer et al., 2008; Baraloto et al., 2010). However, we lack information about how the environment can alter patterns of phenotypic integration (Nicotra et al., 1997; Nicotra et al. 2007; Wright et al., 2006). Nevertheless, studies of other groups of traits indicate that phenotypic integration should increase with environmental stress (Schlichting, 1989; Gianoli, 2004; Godoy et al., 2012). This study explores the magnitude, variability and correlation patterns of leaf and hydraulic traits of the rain forest tree A. punctatum across contrasting soil moisture conditions, which occur within fog-dependent forest patches in semiarid Chile. A striking asymmetric pattern in these patches is that tree mortality increases towards the leeward edge and regeneration is enhanced towards windward edges (del Val et al. 2006). Specifically, we addressed the following two hypotheses: (1) Individuals that occur in drier leeward edges of forest patches may display traits that favor water conservation (lower stomatal and higher and trichome density, and higher LMA) and minimize cavitation risk (lower vessel diameter, higher vessel density, and enhanced hydraulic conductivity); (2) Phenotypic variation and integration may increase in leeward edges due to greater environmental variability and increased water shortage. Aextoxicon punctatum populations in this northern outpost of temperate forests have confronted climate change over an extended period of increasing aridity (Villagrán et al., 2004; Gutierrez et al., 2008). Accordingly, studying drought tolerance strategies in this species will be of great value to understand and predict the consequences of future changes in climate and forest fragmentation at the margins of distribution of temperate forests. In addition, increased fragmentation of temperate forests, due to human land use, and expansion of edge habitats combined with climate change, are likely to enhance desiccation effects and cause increased mortality of forest trees (Choat et al., 2012), unless trees can accommodate to drier edge environments (Breda et al., 2006). The analysis of A. punctatum responses to pronounced microhabitat differences within and between patches in fog-dependent forests could provide clues to understanding tree responses to changing water stress gradients.

Materials and methods

Study site and species -Fray Jorge National Park (30°40´S. 71°30´W) comprises the northernmost patches of Chilean temperate rainforests under the direct influence of maritime fog. A mosaic of about 180 forest patches ranging in size from 0.1 to 36 ha are spread out on the summits of coastal mountains at an elevation of 450 to 660 m (Fig 1) (Barbosa et al., 2010). Forest patches are surrounded by a matrix of semiarid shrub vegetation, in correspondence with the Mediterranean-arid regional climate, with a mean annual rainfall of 147 mm concentrated during the winter months (May to August) and a mean annual temperature of 13.6° C (López-Cortés and López, 2004). Fog is the major water input above 400 m elevation, especially during spring and summer months, such that fragments receive at least an additional 200-400 mm of water annually via throughfall and stemflow (del-Val et al., 2006). In these fog-dependent forests, soil moisture is spatially heterogeneous due to fog interception by trees, creating an asymmetric soil moisture distribution from windward to leeward edges (Stanton et al. 2013). This within-patch environmental gradient has important effects on the dynamics of tree species, yielding an asymmetric distribution of tree regeneration and mortality from windward to leeward edge of patches (del-Val et al., 2006; Gutiérrez et al., 2008).

Sampling design and soil moisture-To assess intraspecific variation in leaf and hydraulic traits of A. punctatum across the soil moisture gradient produced within patches by fog influx, we sampled four forest patches separated by at least 200 m from one another. For logistic reasons, due to the number of simultaneous measurements per patch, we selected two small (< 1 ha) and two large patches (> 20 ha) corresponding to the extremes of the distribution of patch sizes in the mosaic studied (Table 1) (Barbosa et al. 2010). Patches were subdivided into three zones according to spatial variation in fog influx: windward edge, patch core, and leeward edge, and five individuals of A. punctatum (dbh >10 cm) per zone per patch were sampled (n=60). Five measurements of volumetric soil moisture were recorded during spring and summer (between November and January 2010) for each tree using a hand-held TDR probe (Fieldscout TDR 100, Spectrum Technologies, Illinois, USA). Measurements were collected after clearing away leaf litter and subaerial roots directly beneath the tree crown. To assess the real water status of plants, we measured leaf water potentials at predawn ( ψPD , MPa) and at midday ( ψMD , MPa) using a pressure chamber (Scholander-type, model 1000

PMS). Measures were conducted between 0500 and 0700 hours and between 1100 and 1300 hours respectively.

Leaf traits-Ten mature, fully expanded leaves without herbivore damage were taken from each of five sample trees per patch zone (n=10*60=600). Leaves were scanned (EPSON Stylus TX200) and analyzed using ImageJ software (http://imagej.nih.gov/ij/) to determine leaf area (LA), and then dried for 48 h at 65°C to obtain leaf dry mass (g) and then calculate leaf mass per area (LMA) in g cm-2 (Cornelissen et al., 2003). One leaf per individual was prepared to determinate trichome and stomatal density. Leaves were kept in Je ﬀrey solution (chromic acid at 10% and nitric acid at 10% in equal parts) for 48 h, until the epidermis could be easily separated from the mesophyll. Later, the epidermis was dyed in diluted methylene blue and stomatal and trichome densities were measured on one spot of 1 mm diameter located halfway along the length of the leaf using ImageJ software (http://imagej.nih.gov/ij/).

Hydraulic traits and conductivity - A sample of branches, each 10-15 mm in diameter was collected from the outer crown of sampled trees to measure hydraulic conductivity, i.e., water flux through a unit length of stem divided by the pressure gradient (K s, in kg m-1 s-1 MPa -1), following Sperry et al. (1988). Samples were cut in the morning (between 6:00 to 9:00 am) and immediately after cutting they were re-cut under water about 0.2 m higher. Branches were subsequently transported inside dark bags to the field station, located 45 min from the place of collection. Hydraulic conductivity was measured in the field station within five hours after cutting as follows. Distal ends of each branch were trimmed under water with a razor blade (to clear any accidentally blocked vessels), and a segment of around 30 cm in length was obtained. Segments were larger than the maximum conduit length, which was previously estimated from a separate collection of branches taken from the same individuals (see below). While submerged, the basal end of the branch was connected to a fluid column fed by a reservoir of 10 mM KCl solution elevated to a height of 1 m (providing a constant pressure of 9.8 KP), while the apex end of the branch was wrapped with parafilm. An electronic balance recorded KCl solution flux as increase in sample mass every 15 seconds. Measurements were made when an approximately constant flow was observed for at least 3 min. Afterwards, a subset of branches was flushed with KCl solution at a

53 pressure of 170 kPa for 10-15 min to remove embolism (Sperry et al. 1987) and hydraulic conductivity was measured again at its maximum capacity. To standardize the flow of water per unit sapwood area and obtain sapwood specific hydraulic -1 -1 -1 conductivity ( Ks, kg MPa m s ), we divided K h by the cross-sectional area of the conductive xylem (see hydraulic anatomy below). Thus, hydraulic conductivity was made comparable among segments of different diameters. Ks was compared with sapwood specific hydraulic conductivity at maximum capacity to obtain the percentage of loss of conductivity (PLC), estimated as (max Ks-ﬁeld Ks)/max Ks. These data were available for a subset of three individuals per zone in only two patches.

Hydraulic anatomy -- To visualize the conductive wood area, the same stems were perfused with safranin dye using positive pressure by syringe connected to the cut end of the branch to introduce the dye into stems. A cross-sectional area of the upper distal end of the stem was photographed with a digital camera mounted on a microscope, at 10x and the image processed using the imaging software SigmaScan Pro 5 (SPSS Inc.) to determine vessel diameter (µm) and density (vessels mm -2).

Data analysis- Differences across forest patch zones (windward and leeward edges and core) in leaf and hydraulic traits were explored using principal component analysis (PCA). We also performed split-plot two-way ANOVA model with zone (Z) and patch size (S) as factors and estimated the interaction between them. Stomatal and trichome densities and leaf water potentials at predawn and midday were incorporated into the model using an exponential function for the relationship between the variance and the mean to conform to assumptions of heteroscedasticity. Because ANOVA showed significant interactions between patch zone and patch area, we analyzed large and small patches separately using one-way ANOVA with Z as factor, followed by post-hoc Tukey´s tests to identify individual responses of each trait. Given that we did not find clear differences in mean values among patch zones in small forest patches, we examined shifts in the spread (coefficient of variation) and phenotypic integration among zones only in the large patches. Because the two large patches studied showed similar patterns in mean values, we pooled these data for further analyses (10 individuals per zone). In order to compare the level of variation of leaf and hydraulic traits among zones within patches, we obtained 95% confidence intervals

(CI 95% ) by bootstrapping the original data using Poptools (Hood, 2010).

To assess phenotypic integration, we constructed 5*5 correlation matrices with morphological traits for each zone (MCs) and for all individuals using Pearson’s correlation coefficients to test the relationships for every pair of traits. The magnitude of character integration (INT) for each zone and for large patch data was estimated from the variance of eigenvalues of each correlation matrix (Wagner, 1984; Cheverud et al., 1989). A 95% confidence interval of INT was estimated by bootstrapping the original log-transformed data.

Results

Within-patch moisture gradient and leaf water potential - In both small and large patches volumetric soil moisture varied substantially among zones, with leeward edges significantly drier than the other two microhabitats (small patches: F=16.42, p<0.0001, large patches: F=73.77, p<0.0001) (Table 2, see Supplemental data with the online version of this article). Differences in soil moisture among patch zones were reflected in lower ψPD at leeward edges in large forest patches (small: F=1.33, p=0.2806, large:

F=98.78, p<0.0001) and lower ψMD at leeward edges in both small and large patches

(small: F=8.63, p=0.0013, large: F=54.51, p<0.0001) (Table 2). The ψPD and ψMD estimated in the patch cores were comparable to windward edge values in both patch sizes (Table 2, see Supplemental data with the online version of this article).

Shifts in mean trait values across zones within patches - Hydraulic and leaf traits also differed among patch zones. The first PCA axis, which explained 41% of trait variation, clearly separated leeward edge from other two wetter zones (Fig. 2). This axis was positively correlated with traits related to water conservation strategy (trichome and stomatal density and LMA) and negatively correlated with K (see Supplemental data with the online version of this article). Then, higher values along the ﬁrst PCA axis reflect stronger ability to conserve water and tolerate drought, but decreased water transport efficiency. The second PCA component explained an additional 27% of the total variance and it was dominated by the tradeoff between vessels diameter and density. However, it did not separate trees in different patch zones (see Supplemental data with the online version of this article). Similar results were detected when each trait was analyzed separately. ANOVAs show significant differences among patch zones for

55 mean values of leaf traits and hydraulic conductivity, but no for vessel density and vessel diameter (Table 2, see Supplemental data with the online version of this article). These analyses also provided evidence that trait variation was more pronounced in large than in small forest fragments. In the latter, within-patch differences in soil moisture and water potentials were less accentuated. Stomatal density was higher for trees in the leeward edge than in the wetter windward and core zones of large forest patches (F=21.48, p<0.0001), but did not differ among zones in the small patches (F=2.77, p=0.08). Trichome density and leaf mass area were higher for trees in the leeward edge than in the wetter core zone in both small (trichomes: F=14.31, p=0.0001, LMA: F=5.11, p=0.01) and large forest patches (trichomes: F=31.43, p<0.0001, LMA: F=17.32, p<0.0001), but did not show differences with the windward edge in small patches (Table 2). Likewise, K was lower for trees in the leeward edges than in the wetter windward edges in both small (F=17.98, p=0.0001) and large patches (F=28.95, p<0.0001) (Table 2). To assess whether reduction in Ks at leeward edges reflected higher levels of embolism, we estimated K at maximum capacity and the percentage of loss conductivity (PLC) in three to five individuals per zone for two patches. As expected, we found higher PLC values for trees in leeward edges of both small (F=10.52, p=0.01) and large patches (F=32.63, p<0.001)

Shifts in the spread of trait values across zones within patches - Three of the six traits evaluated showed significant trends in relation to forest patch zones (Fig 3). Coefficients of variation (CV) of stomatal density, trichome density and hydraulic conductance differed significantly among zones within patches as revealed by the non- overlapping 95% confidence intervals, which are 1.8 to 4.6 times higher in the leeward edge than in the wetter core and windward zones (Fig 3). LMA and xylem anatomical traits did not show statically significant differences in CV (Fig 3).

Shifts in trait correlations and the extent of phenotypic integration within patches - Phenotypic correlation matrices varied among zones within forest patches (Table 3). The most divergent matrix was that of the windward zone, showing similarity indices of -0.12 and 0.11 with respect to the leeward and core matrices. Phenotypic matrices of these last two zones (leeward and core) were more similar (similarity index=0.71, p=0.02), but often correlation coefficients were stronger in the drier leeward zone. Whereas mean r 2 value for characters of trees in leeward areas was 0.40, this parameter

56 was only 0.14 and 0.18 for trees in windward and core zones respectively. Integration values were also higher in the drier leeward zone (INT = 1.9, 95%, Confidence interval (CI): 1.32-3.39) than in windward (INT = 0.64, 95%, CI 0.52-1.99) or core (INT = 0.92, 95%, CI: 0.90-2.30) zones, but differences were not statically significant.

Discussion

The temperate rainforest tree Aextoxicon punctatum showed considerable variation in leaf traits across soil-moisture gradients produced by fog interception by the tree canopy. Notably, leaf trait variation within structurally asymmetric forest fragments (del Val et al., 2006; Stanton et al., 2013) at spatial scales of 100 meters or less was greater than variation observed between fragments of contrasting size, and even greater than differences between trees in the Fray Jorge patch mosaic and Aextoxicon populations located 1500 km to the south, where precipitation is ten times higher (Salgado-Negret, unpublished data). In contrast to foliar traits, those related to xylem anatomy (vessel diameter and density) did not vary significantly within forest fragments of A. punctatum in Fray Jorge or in populations located 1500 km to the south (Salgado-Negret, unpublished data), and were decoupled from the observed variation in leaf traits. Lack of variability in xylem anatomy of Aextoxicon trees in Fray Jorge forest patches was associated with lower hydraulic conductivity in the drier leeward edge, as K was four times lower for trees in the core or windward zones of patches. Reduced conductivity at the leeward edge might be explained by higher levels of embolism, because PLC values at leeward edges were five times higher than PLC measured at core zones in both small and large patches. To determine whether leaf trait differentiation among patch zones is due plastic responses or to local adaptation is necessary to compare among trees experimentally grown in common gardens. Indirect evidence based on distances between patches and dispersal distances of Aextoxicon seeds dispersed by birds suggest that gene flow should occur among zones within forest patches as well as among patches (Nuñez-Ávila et al. 2013), and hence differences among trees in different patch zones are likely due to plasticity. Thus, leaf phenotypic plasticity in response to within patch differences in water availability is likely involved in the persistence of this tree species across a range of habitats. Trees with higher LMA, trichome and stomatal densities grew more often in leeward edges, where water availability was two to three times lower than in the patch

57 core zone and windward edges of patches. These differences in leaf traits can be related to water conservation strategies (Chapin, 1980). Leaves that are more dense and rigid (higher LMA) have smaller transpiring surfaces, hence reducing wilting and water requirements (Poorter et al., 2009). Greater leaf pubescence increases boundary layer resistance, decreases transpirational water losses (Fahn, 1986; Baldini et al., 1997), and also enlarges the surface available for water uptake by leaves (Savé et al., 2000; Grammatikopoulos and Manetas, 1994). The observed increment in leaf stomatal density in trees growing at the drier leeward edge of patches is less intuitive, because greater stomatal densities are often associated with higher transpiration and water loss. However, stomata in A. punctatum leaves are sunken and located in the abaxial epidermis. Sunken stomata generally reduce leaf transpiration (Jordan et al., 2008) and facilitate CO 2 diffusion in thick, hard leaves (Hassiotou et al., 2009). High stomatal densities in the drier leeward edge may probably compensate for the greater internal resistance to CO 2 uptake by thicker and denser leaves (with higher LMA). In addition, long-lived leaves with higher LMA can exhibit higher stomatal densities as a ‘backup’ mechanism, in case that some stomata become inactive, i.e., dust blocked (Hassiotou et al., 2009). Three of the four leaf traits that showed differences in mean values across zones within fragments, also showed differences in their degree of variability. For stomatal and trichome densities, and K, the coefficient of variation was greater for trees in the drier leeward edge. This patch zone is not only drier but also subjected to higher fluctuations in irradiance and temperature and therefore soil moisture compared to core and windward zones of patches. These results agree with other studies showing increasing number of alternative phenotypes with increasing resource heterogeneity (Sultan, 1987; Lortie and Aarssen, 1996; Balaguer et al., 2001). In the case of A. punctatum , the higher coefficient of variation for stomatal and trichome densities of trees in leeward habitats may be related to successive generations of leaves experiencing contrasting environments and therefore promoting alternative phenotypes. In contrast, the uniformity of hydraulic traits may indicate high environmental canalization, due to the strong connection of hydraulic properties with water transport and survival, which enables organisms to maintain the highest possible level of fitness across environments (Debat and David, 2001). High canalization of hydraulic anatomy across all within- patch zones could lead to high K variability at leeward edges.

We also found stronger correlations among leaf traits and greater level of phenotypic integration at decreasing levels of soil water availability within patches. Other studies of phenotypic integration also showed greater correlation values in heterogeneous environments (Schlichting, 1989; Nicotra et al., 1997; Gianoli, 2004), but the functional benefits or constraints on this pattern for plants have not been clearly established (Gianoli, 2004; Matesanz et al., 2010). Notably, we found that in the case of Aextoxicon punctatum leaf traits varied rather independently of hydraulic traits, except for trees in the leeward edge, where LMA, vessel density and vessel diameter were correlated. In this heterogeneous and variable environment, which characterizes fragmented forests, functional coordination between stem conductive capacity and leaf hydraulic properties might be essential. Our results on this point contrast with other studies reporting coordinated variation of leaf and stem traits in forest trees (Brodribb and Field, 2000; Brodribb et al., 2002; Santiago et al., 2004; Wright et al., 2006; Meinzer et al., 2008; but see Baraloto et al., 2010), and highlight the need to examine patterns of phenotypic integration across different environmental gradients. Overall, this study demonstrates that Aextoxicon punctatum leaf traits, but not xylem anatomy, vary within forest patches under contrasting soil-moisture conditions produced by fog interception patterns and also vary among patches due to differences in forest patch size. The absence of similar plasticity in xylem traits of trees was correlated with a reduction in hydraulic conductance at the drier leeward edge and it evidenced higher water stress expressed by more negative ψPD in the leeward edges with respect to other zones. Although vessel diameters recorded for A. punctatum stems are in the smaller range of those reported for tree species in the literature (Ewers and Fisher, 1989; Zanne et al., 2003; Chave et al., 2009), they could not prevent cavitation, revealing that soil water availability at the leeward edge of patches is insufficient to maintain a constant flux along stems. Indeed, we previously reported that hydraulic potential at midday

(ψMD ) for individuals of A. punctatum growing at leeward edges frequently fell below the turgor loss point ( πtlp ), suggesting intense water stress (Salgado-Negret et al., 2013). Recent climate change scenarios for Chile (CONAMA, 2006) predict enhanced interannual variability in rainfall, greater intervals between extremely wet and dry years, and particularly a decline in winter rainfall (concentrating >80% of annual rainfall) in the study area. However, rain contributes only a fraction (about 50% during low rainfall years) of the annual water budget in Fray Jorge forests and future changes in fog

59 frequency over time, are uncertain (Gutiérrez et al. 2008). Reductions in rainfall and fog inputs coupled to increasing patch fragmentation (Sala et al., 2000), will decrease water budget of these forests because lower water capture surfaces and higher environmental variability. This scenario will expose A. punctatum trees to greater water stress in this patch mosaic. The inability of this tree species to modify xylem anatomy traits, associated with its problem to maintain leaf turgor in the face of decreasing soil moisture at leeward edges (Salgado-Negret et al., 2013) and the narrow hydraulic safety margins for tree species around the world (Choat et al., 2012) could seriously impair the ability of A. punctatum to supply water to leaves for photosynthetic gas exchange. This mechanism could eventually lead to negative water balance and increased tree mortality along exposed patch edges and small size patches. Higher tree mortality would alter hydrologic balance of fragmented forests, affecting regeneration and persistence of other species that depend on ecosystem integrity. In particular, fog capture by the A. punctatum canopy may be impaired due to disruption of hydrologic balance in small patches and leeward edge of patches. Global change, expressed in reductions of forest cover, increased fragmentation and more intense edge effects are likely to have strong negative impacts, on forest ecosystems worldwide, and on these fog-dependent ecosystems in particular, because of ecophysiological limitations and drought effects on the performance and survival of the dominant tree species.

Acknowledgements

We express our gratitude to Leonardo Ramirez, Felipe Albornoz, Juan Monardez, Carmen Ossa, Daniel Salinas and Patricio Valenzuela for their invaluable assistance in the ﬁeld. We thank to Daniel Stanton for useful discussions and comments on the manuscript and to Fernando Casanoves for statistical support. Work was supported by CONICYT fellowship 24110074 to B.S-N., and grants Fondecyt 1110929 to F.P., P05- 002 from Millennium Scientific Initiative and PFB-23 from CONICYT to the Institute of Ecology and Biodiversity, Chile. This is a contribution to LINC-Global (Chile-Spain) and to the Research Program of the Chilean LTSER network at Fray Jorge National Park.

References

ANDEREGG, W. R., J. M. KANE, AND L. D. ANDEREGG.2013. Consequences of widespread tree mortality triggered by drought and temperature stress. Nature Climate Change 3: 30-36. BALAGUER, L., E. MARTÍNEZ-FERRI, F. VALLADARES, M. E. PÉREZ- CORONA, F. J. BAQUEDANO, F. J. CASTILLO, AND E. MANRIQUE. 2001. Population divergence in the plasticity of the response of Quercus coccifera to the light environment. Functional Ecology 15: 124-135. BALDINI, E., O. FACINI, F. NEROZZI, F. ROSSI, AND A. ROTONDI. 1997. Leaf characteristics and optical properties of different woody species. Trees 12: 73- 81. BARALOTO, C., C. E. T. PAINE, L. POORTER, J. BEAUCHENE, D. BONAL, A. M. DOMENACH, B. HÉRAULT, ET AL. 2010. Decoupled leaf and stem economics in rain forest trees. Ecology Letters 13: 1338-1347. BARBOSA, O., P. A. MARQUET, L. D. BACIGALUPE, D. A. CHRISTIE, E. DEL- VAL, A. G. GUTIERREZ, C. G. JONES, ET AL. 2010. Interactions among patch area, forest structure and water fluxes in a fog-inundated forest ecosystem in semi-arid Chile. Functional Ecology 24: 909-917.

BRÉDA , N., R. HUC , A. GRANIER , AND E. DREYER . 2006. Temperate forest trees and stands under severe drought: a review of ecophysiological responses, adaptation processes and long-term consequences. Annals of Forest Sciences 63: 625-644. BRESHEARS, D.D., N. S. COBB, P. M. RICH, K. P. PRICE, D. C. ALLEN, R. G. BALICE, W. H. ROMME, ET AL. 2005. Regional vegetation die-off in response to global-change-type drought. Proceedings of the National Academy of Sciences of the United States of America 102: 15144-15148. BRODRIBB, T.J., AND T. S. FIELD. 2000. Stem hydraulic supply is linked to leaf photosynthetic capacity: evidence from New Caledonian and Tasmanian rainforest. Plant, Cell and Environment 23: 1381–1388. BRODRIBB, T.J., N. M. HOLBROOK, AND M. V. GUTIÉRREZ. 2002. Hydraulic and photosynthetic co-ordination in seasonally dry tropical forest trees. Plant, Cell and Environment 25: 1435–1444.

CARLQUIST, S. 2001. Comparative wood anatomy: systematic, ecological and evolutionary aspects of dicotyledon wood . 2nd Edn. Springer-Verlag. Berlin Heidelberg. Germany. CERECEDA, P., P. OSSES, H. LARRAÍN, M. FARÍAS, M. LAGOS, R. PINTO, AND R. S. SCHEMENAUER. 2002. Advective, orographic and radiation fog in the Tarapacá region, Chile. Atmospheric Research 64: 261-271. CHAPIN, F.S. III. 1980. The mineral nutrition of wild plants. Annual Review of Ecology and Systematics 11: 233-260. 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. CHEVERUD, J.M., G. P. WAGNER, AND M. M. DOW. 1989. Methods for the comparative analysis of variation patterns. Systematic Zoology 38: 201-213. CHOAT, B., L. SACK, AND N. M. HOLBROOK. 2007. Diversity of hydraulic traits in nine Cordia species growing in tropical forests with contrasting precipitation. New Phytologist 175: 686-698. CHOAT, B., S. JANSEN, T. J. BRODRIBB, H. COCHARD, S. DELZON, R. BHASKAR, S. J. BUCCI, ET AL. 2012. Global convergence in the vulnerability of forests to drought. Nature 491: 752-755. CONAMA. 2006. Estudio de la variabilidad climática en Chile para el siglo XX1. Comisión Nacional de Medio Ambiente. Santiago, Chile. CORNELISSEN, J.H.C., S. LAVOREL, E. GARNIER, S. DÍAZ, N. BUCHMANN, D. E. GURVICH, P. B. REICH, ET AL. 2003. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian Journal of Botany 51: 335-380. CORNWELL, W.K., AND D. D. ACKERLY. 2009. Community assembly and shifts in plant trait distributions across an environmental gradient in coastal California. Ecological Monographs 79: 109-126. DAWSON, T.E. 1998. Fog in the California redwood forest: ecosystem inputs and use by plants. Oecologia 117: 476-485. DEBAT, V., AND P, DAVID. 2001. Mapping phenotypes: canalization, plasticity and developmental stability. Trends in Ecology and Evolution 16: 555-561. DEL-VAL, E., J. J. ARMESTO, O, BARBOSA, D. A. CHRISTIE, A. G. GUTIÉRREZ, P. A. MARQUET, C. G. JONES, AND K. C. WEATHERS. 2006. Rain forest

islands in the Chilean semiarid region: fog-dependency, ecosystem persistence and tree regeneration. Ecosystems 9: 598-608. ECHEVERRIA, C., D. COOMES, J. SALAS, J. M. REY-BENAYAS, A. LARA, AND A. NEWTON. 2006. Rapid deforestation and fragmentation of Chilean Temperate Forests. Biological Conservation 130: 481-494. ENGELBRECHT, B.M.J., L. S. COMITA, R. CONDIT, T. A. KURSAR, M. T. TYREE, B. L. TURNER, AND S. P. HUBBELL. 2007. Drought sensitivity shapes species distribution patterns in tropical forests. Nature 447: 80-83. EWERS, F.W., AND J. B. FISHER. 1989. Techniques for measuring vessel lengths and diameters in stems of woody plants. American Journal of Botany 76: 645-656. EWING, H.A., K. C. WEATHERS, P. H. TEMPLER, T. E. DAWSON, M. K. FIRESTONE, A. M. ELLIOTT, AND V. K. S. BOUKILI. 2009. Fog water and ecosystem function: heterogeneity in a California Redwood forest. Ecosystems 12: 417-433. FAHN A. 1986. Structural and functional properties of trichomes of xeromorphic leaves. Annals of Botany 57: 631-637. FAJARDO, A., AND F. I. PIPER. 2011. Intraspeciﬁc trait variation and covariation in a widespread tree species ( Nothofagus pumilio ) in southern Chile. New Phytologist 189: 259-271. FIGUEROA, J.A., H. M. CABRERA, C. QUEIROLO, AND L. F. HINOJOSA. 2010. Variability of water relations and photosynthesis in Eucryphia cordifolia Cav. (Cunoniaceae) over the range of its latitudinal and altitudinal distribution in Chile. Tree Physiology 30: 574-585. GARREAUD, R., J. BARICHIVICH, C. CHRISTIE, AND A. MALDONADO. 2008. Interannual variability of the coastal fog at Fray Jorge relict forests in semiarid Chile. Journal of Geophysical Research doi:10.1029/2008JG000709 GIANOLI, E. 2004. Plasticity of traits and correlations in two populations of Convolvulus arvensis (Convolvulaceae) differing in environmental heterogeneity. International Journal of Plant Sciences 165: 825-832. GIANOLI, E., AND F. VALLADARES. 2012. Studying phenotypic plasticity: the advantages of a broad approach. Biological Journal of the Linnean Society 105: 1-7.

GODOY, O., F. VALLADARES, AND P. CASTRO-DÍEZ. 2012. The relative importance for plant invasiveness of trait means, and their plasticity and integration in a multivariate framework. New Phytologist 195: 912-922. GRAMMATIKOPOULOS, G., AND Y. MANETAS. 1994. Direct absorption of water by hairy leaves of Phlomis fruticosa and its contribution to drought avoidance. Canadian Journal of Botany 72: 1805-1811. GUTIÉRREZ, A.G., O. BARBOSA, D. A. CHRISTIE, E. DEL-VAL, H. A. EWING, C. G. JONES, P. A. MARQUET, ET AL. 2008. Regeneration patterns and persistence of the fog dependent Fray Jorge forest in semiarid Chile during the past two centuries. Global Change Biology 14: 161-176. HASSIOTOU, F., J. R. EVANS, M. LUDWIG, AND E. J. VENEKLAAS. 2009.

Stomatal crypts may facilitate diffusion of CO 2 to adaxial mesophyll cells in thick sclerophylls. Plant, Cell and Environment 32: 1596-161. HILDEBRANDT, A., AND E. A. B. ELTAHIR. 2006. Forest on the edge: seasonal cloud forest in Oman creates its own ecological niche. Geophysical Research Letter 33: L11401. HOOD, G. M. 2010. PopTools version 3.2.5. Available on the internet. URL http://www.poptools.org JOHNSTONE, J.A., AND T. E. DAWSON. 2010. Climatic context and ecological implications of summer fog decline in the coast redwood region. Proceedings of the National Academy of Sciences of the United States of America 107: 4533- 4538. JORDAN, G.J., P. H. WESTON, R. J. CARPENTER, R. A. DILLON, AND T. J. BRODRIBB. 2008. The evolutionary relations of sunken, covered, and encrypted stomata to dry habitats in Proteaceae. American Journal of Botany 95: 521-530. KATATA, G., H. NAGAI, M. KAJINO, H. UEDA, AND Y. HOZUMI. 2010. Numerical study of fog deposition on vegetation for atmosphere–land interactions in semi-arid and arid regions. Agricultural and Forest Meteorology 150: 340-353. LÓPEZ-CORTÉS, F., AND D. LÓPEZ. 2004. Antecedentes bioclimáticos del Parque Nacional Bosque Fray Jorge. In: Squeo, F.A., Gutiérrez, J.R. & Hernández, I.R. (eds.) Historia natural del parque nacional bosque Fray Jorge . Ediciones Universidad de La Serena, Chile. 45-60.

LORTIE, C., AND L. W. AARSSEN. 1996. The specialization hypothesis for phenotypic plasticity in plants. International Journal of Plant Science 157: 484- 487. MARKESTEIJN, L., L. POORTER, F. BONGERS, H. PAZ, AND L. SACK. 2011a. Hydraulics and life history of tropical dry forest tree species: coordination of species’ drought and shade tolerance. New Phytologist 191: 480-495. MARKESTEIJN, L., L. POORTER, H. PAZ, L. SACK, AND F. BONGERS. 2011b. Ecological differentiation in xylem cavitation resistance is associated with stem and leaf structural traits. Plant Cell and Environment 34: 137-148. MATESANZ, S., A. ESCUDERO, AND F. VALLADARES. 2009. Impact of three global change drivers on a Mediterranean shrub. Ecology 90: 2609-2621. MATESANZ, S., E. GIANOLI, AND F. VALLADARES. 2010. Global change and the evolution of phenotypic plasticity in plants. Annals of the New York Academy of Sciences 1206: 35–55. MEINZER, F. C., D. R. WOODRUFF, J. C. DOMEC, G. GOLDSTEIN, P. I. CAMPANELLO, M. G. GATTI, AND R. VILLALOBOS-VEGA. 2008. Coordination of leaf and stem water transport properties in tropical forest trees. Oecologia 156: 31-41. MURREN, C.J. 2002. Phenotypic integration in plants. Plant Species Biology 17: 89- 99. NICOTRA, A.B., R. L. CHAZDON, AND A. D. SCHLICHTING. 1997. Patterns of genotypic variation and phenotypic plasticity of light response in two tropical Piper (Piperaceae) species. American Journal of Botany 84: 1542-1552. NICOTRA, A.B., J.P. HERMES, C.S. JONES, AND C.D. SCHLICHTING. 2007. Geographic variation and plasticity to water and nutrients in Pelargonium austral. New Phytologist 176: 136–149. NIINEMETS, U. 2001. Global-scale climatic controls of leaf dry mass per area, density and thickness in trees and shrub. Ecology 82: 453-469. NUÑEZ-ÁVILA, M. C., URIARTE, M., MARQUET, P. A., AND J. J. ARMESTO. 2013. Decomposing recruitment limitation for an avian-dispersed rain forest tree in an anciently fragmented landscape. Journal of Ecology doi: 10.1111/1365- 2745.12148 PIGLIUCCI, M. 2003. Phenotypic integration: studying the ecology and evolution of complex phenotypes. Ecology Letters 6: 265-272.

POCKMAN, W.T., AND J. S. SPERRY. 2000. Vulnerability to xylem cavitation and the distribution of Sonoran Desert. American Journal of Botany 87: 1287-1299. POORTER, H., U. NIINEMETS, L. POORTER, I. J. WRIGHT, AND R. VILLAR. 2009. Causes and consequences of variation in leaf mass per area (LMA): a meta-analysis. New Phytologist 182: 565-588. SALA, O.E., F. S. III. CHAPIN, J. J. ARMESTO, E. BERLOW, J. BLOOMFIELD, R. DIRZO, E. HUBER-SANWALD, ET AL. 2000. Global biodiversity scenarios for the year 2100. Science 287: 1770-1774. SALGADO-NEGRET, B., F. PÉREZ, L. MARKESTEIJN, M. JIMENEZ-CASTILLO, AND J. J. ARMESTO. 2013. Diverging drought tolerance strategies explain tree species distribution along a fog-dependent moisture gradient in a temperate rain forest. Oecologia. doi 10.1007/s00442-013-2650-7 SANTIAGO, L.S., G. GOLDSTEIN, F. C. MEINZER, J. B. FISHER, K. MACHADO, D. WOODRUFF, AND T. JONES. 2004. Leaf photosynthetic traits scale with hydraulic conductivity and wood density in Panamanian forest canopy trees. Oecologia 140: 543-550. SAVÉ, R., C. BIEL, AND F. DE HERRALDE. 2000. Leaf pubescence, water relations and chlorophyll fluorescence in two subspecies of Lotus creticus L. Biologia Plantarum 43: 239-244. SCHLICHTING, C. D. 1989. Phenotypic plasticity in Phlox. II. Plasticity of character correlations. Oecologia 78: 496-501. SPERRY, J. S., N. M. HOLBROOK, M. H. ZIMMERMANN, AND M. T. TYREE. 1987. Spring filling of xylem vessels in wild grapevine. Plant Physiology 83: 414-417. SPERRY, J.S., J. R. DONNELLY, AND M. T. TYREE. 1988. A method for measuring hydraulic conductivity and embolism in xylem. Plant Cell and Environment 11: 35-40. STANTON, D. E., B. SALGADO-NEGRET, J. J. ARMESTO, AND L. O. HEDIN. 2013. Forest patch symmetry depends on direction of limiting resource delivery. Ecosphere http://dx.doi.org/10.1890/ES13-00064.1 SULTAN, S. E. 1987. Evolutionary implications of phenotypic plasticity in plants. Evolutionary Biology 21: 127-178. VALLADARES, F., E. GIANOLI, AND J. M. GÓMEZ. 2007. Ecological limits to plant phenotypic plasticity. Tansley review. New Phytologist 176: 749-763.

VILLAGRÁN, C., J. J. ARMESTO, L. F. HINOJOSA, J. CUVERTINO, C. PÉREZ, AND C. MEDINA. 2004. El enigmático origen del bosque relicto de Fray Jorge. In: Squeo, F.A., Gutiérrez, J.R. & Hernández, I.R. (eds.) Historia natural del parque nacional bosque Fray Jorge . Ediciones Universidad de La Serena, Chile. 3-43. VIOLLE, C., B. J. ENQUIST, B. J. MCGILL, L. JIANG, C. H. ALBERT, C. HULSHOF, V. JUNG, AND J. MESSIER. 2012. The return of the variance: intraspecific variability in community ecology. Trends in Ecology & Evolution 27: 244–252. WAGNER, G.P. 1984. On the eigenvalue distribution of genetic and phenotypic dispersion matrices: Evidence for a nonrandom organization for quantitative character variation. Journal of Mathematical Biology 21: 77-95. WEATHERS, K.C., G. M. LOVETT, G. E. LIKENS, AND N. F. M. CARACO. 2000. Cloudwater inputs of nitrogen to forest ecosystems in southern Chile: Forms, fluxes and sources. Ecosystems 3: 590-595. WRIGHT, I.J., P. B. REICH, M. WESTOBY, D. D. ACKERLY, Z. BARUCH., F. BONGERS, J. CAVENDER-BARES, ET AL. 2004. The worldwide leaf economics spectrum. Nature 428: 821-827. WRIGHT, I.J., D. S. FALSTER, M. PICKUP, AND M. WESTOBY. 2006. Cross- species patterns in the coordination between leaf and stem traits, and their implications for plants hydraulics. Physiologia Plantarum 127: 445-456. 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: 207-215.

Table 1. Characterization of forest patches, including differences in mean values for microclimatic variables and relative basal area for all live stems (>5 cm dbh) (Gutierrez et al., 2008, Barbosa et al., 2010). P1 P2 P5 P6 Patch area 0.21 0.28 36.08 23.76 Altitude (m) 529 566 635 639 Slope (%) 1 11 42 38 Throughfall (mm) 31.10 ± 21.31 49.91 ± 43.16 29.56 ± 18.05 37.38 ± 22.55 Stemflow (mm) 0.10 ± 0.06 0.25 ± 0.06 0.69 ± 1.07 1.00 ± 0.99 Mean temperature ( °C) 11.8 ± 1.6 11.46 ± 1.75 11.29 ± 1.51 10.95 ± 1.42 Mean relative humidity (%) 91.33 ± 4.00 94.98 ± 3.04 95.96 ± 3.73 95.12 ± 4.63 Tree basal area (m 2 ha -1) 61.64 49.41 125.12 102.61 Basal area A. punctatum (%) 49 75.7 46.4 88.8

Table 2. Differences in soil moisture, leaf water potential, leaf and hydraulic traits among individuals of Aextoxicon punctatum growing in different zones of small and

large forest patches in Fray Jorge. Traits are abbreviated as, SOIL= soil moisture, ψPD =

leaf water potential at predawn, ψMD = leaf water potential at midday, SD = Stomatal density, TD = Trichome density, LMA = Leaf mass area, VD = Vessel density, VDi = Vessel diameter, K = Sapwood-specific hydraulic conductivity, PLC = % loss conductivity. Different letters above mean values represent statistical differences between zones for each trait from univariate ANOVA and post-hoc Tukey tests. PLC data are available only for a subset of three individuals per zone in two patches.

Variables Small patches Large patches Windward Core Leeward Windward Core Leeward SOIL 10.43 ± 1.01 12.13 ± 5.02 ± 9.25 ± 14.59 ± 4.72 ± (%) (b) 1.12 (b) 0.49 (a) 0.62 (b) 0.72 (c) 0.30 (a)

ΨPD 0.55 ± 0.10 0.57 ± 0.74 ± 0.14 ± 0.10 ± 0.02 0.82 ± (MPa) (a) 0.09 (a) 0.10 (a) 0.03 (a) (a) 0.06 (b)

ΨMD 1.27 ± 0.10 1.42 ± 1.81 ± 0.72 ± 0.71 ± 0.04 1.71 ± (MPa) (a) 0.08 (a) 0.11 (b) 0.04 (a) (a) 0.12 (b) SD 143.70 ± 130.00 ± 147.00 ± 107.90 ± 123.00 ± 170.70 ± (N/mm 2) 4.82 (a) 4.84 (a) 6.44 (a) 1.59 (a) 4.35 (b) 11.34 (c) TD 9.30 ± 0.52 6.50 ± 9.50 ± 4.70 ± 5.00 ± 0.26 10.70 ± (N/mm 2) (b) 0.45 (a) 0.34 (b) 0.30 (a) (a) 0.97 (b) LMA 0.02 ± 0.02 ± 0.02 ± 0.02 ± 0.01 ± 0.03 ± (g/cm 2) 0.0007 (ab) 0.001 (a) 0.001 (b) 0.001 (a) 0.002 (a) 0.002 (b) VD (#/mm 2) 330.76 ± 330.55 ± 290.57 ± 284.96 ± 302.00 ± 309.28 ± 10.25(a) 16.92 (a) 17.54 (a) 10.55 (a) 11.37 (a) 12.03 (a) VDI (µm) 16.62 ± 0.24 16.65 ± 17.66 ± 17.97 ± 16.68 ± 16.89 ± (a) 0.33 (a) 0.41 (a) 0.36 (a) 0.49 (a) 0.36 (a) K 0.40 ± 0.04 0.24 ± 0.13 ± 0.41 ± 0.37 ± 0.03 0.10 ± (c) 0.02 (b) 0.02 (a) 0.04 (b) (b) 0.02 (a) PLC (%) 0.063 ± 0.075 ± 0.502 ± 0.023 ± 0.045 ± 0.251 ± 0.048 (a) 0.043 (a) 0.116 (b) 0.014 (a) 0.014 (a) 0.032 (b)

Table 3. Pearson correlation coefficients between leaf and hydraulic traits for trees from three zones that differ in soil moisture, within four forest patches in Fray Jorge. Significant correlations among traits (p<0.05) are indicated in bold. Zone Traits TD LMA VD VDi Windward SD 0.55 (0.09) -0.03 (0.93) -0.56 (0.09) 0.16 (0.65) TD -0.22 (0.54) -0.76 (0.0009) 0.11 (0.74) LMA 0.00 (0.99) -0.27 (0.44) VD -0.16 (0.65) Core SD 0.34 (0.34) -0.87 (0.001) 0.04 (0.91) 0.01 (0.96) TD -0.25 (0.48) 0.04 (0.90) 0.05 (0.88) LMA -0.18 (0.61) 0.06 (0.85) VD -0.9 (0.0004) Leeward SD 0.74 (0.01) -0.75 (0.01) 0.32 (0.37) -0.29 (0.40) TD -0.76 (0.01) 0.41 (0.23) -0.46 (0.17) LMA -0.75 (0.01) 0.63 (0.05) VD -0.73 (0.01)

Figure caption

Figure 1. a. Location of rain forest patches in Fray Jorge National Park, Chile, at 30°S

(Barbosa et al. 2010). b. Directionality of fog and atmospheric resource inputs to forest patches in Fray Jorge (Stanton et al. 2013).

Figure 2. Principal Component Analysis (PCA) of hydraulic and leaf trait variation for

Aextoxicon punctatum trees among forest patches (P1: circle, P2: rhombus, P5: triangle,

P6: square) and zones within patches (windward edge: black, patch core: grey, leeward edge: white) in Fray Jorge.

Figure 3. Coefficient of variation of leaf and hydraulic traits of A. punctatum trees among zones (windward: black, core: grey, leeward: white) within large forest patches in semiarid Chile. Bars represent means ± 1 SE (n = 10). Different letters above the bars represent statistically significant differences between zones.

71 a

Windward Core Leeward

Fig 1.

4.00

2.00

Ks 0.00 SD TD

PCA PCA 2 (26.8%) LMA

-2.00

VDi

-4.00 -4.00 -2.00 0.00 2.00 4.00 PCA 1 (40.6%)

Fig 2.

50 200

40 150 30

K 100 20 CV (%) CV (%) CV 50 10 Stomataldensity 0 0

80 30

60 20 40 CV (%) CV (%) CV 10 20

Vesselsdensity Trichomedensity 0 0

80 25

60 20 15 40 10 LMA

(%) CV 20 5 CV (%) CV 0 0

Patch zones Vesselsdiameter Patch zones

Fig. 3

Online Supplemental Materials

Appendix S1. Differences in soil moisture, leaf water potential, leaf and hydraulic traits among individuals of Aextoxicon punctatum growing in different zones of small and large forest patches. Traits are abbreviated as, SOIL= soil moisture, ψPD = leaf water potential at predawn, ψMD = leaf water potential at midday, SD = Stomatal density, TD = Trichome density, LMA = Leaf mass area, VD = Vessel density, VDi = Vessel diameter, K = Sapwood-specific hydraulic conductivity, PLC = % loss conductivity (available only for a subset of three individuals per zone for two patches).

Traits df F p SOIL (%) Patch size (S) 1 0.20 0.6677 Zone (Z) 2 66.43 <0.0001 S x Z 2 3.30 0.0458

ΨPD (-MPa) Patch size (S) 1 54.50 0.0001 Zone (Z) 2 27.65 <0.0001 S x Z 2 4.91 0.012

ΨMD (-MPa) Patch size (S) 1 86.75 <0.0001 Zone (Z) 2 47.37 <0.0001 S x Z 2 4.13 0.022 SD (number mm -2) Patch size (S) 1 15.42 0.0044 Zone (Z) 2 24.28 <0.0001 S x Z 2 10.59 <0.0001 TD (number mm -2) Patch size (S) 1 32.15 0.0002 Zone (Z) 2 37.82 <0.0001 S x Z 2 13.57 <0.0001 LMA (g cm -1) Patch size (S) 1 0.49 0.5047 Zone (Z) 2 20.19 <0.0001 S x Z 2 9.83 0.0003 VD (number mm -2) Patch size (S) 1 2.85 0.1296 Zone (Z) 2 0.74 0.4829 S x Z 2 3.09 0.0553

VDi (µm) Patch size (S) 1 0.39 0.5484 Zone (Z) 2 1.95 0.1543 S x Z 2 4.32 0.0191 K (KgMPa -1m-1s-1) Patch size (S) 1 2.24 0.1727 Zone (Z) 2 43.39 <0.0001 S x Z 2 3.51 0.0380 PLC (%) Patch size (S) 1 5.35 0.0540 Zone (Z) 2 21.91 0.0034 S x Z 2 2.44 0.1819

Appendix S2. Eigenvector scores of leaf and hydraulic traits in two main PCA axes. Values in parentheses indicate the percentage of total variance accounted by each axis. Traits are abbreviated as: SD = Stomatal density; TD = Trichome density; LMA = Leaf mass area; Ks = Hydraulic conductivity; VD = Vessels density; VDi Vessels diameter. Variables PCA 1 (40.6%) PCA 2 (26.8%) SD 0.53 -0.06 TD 0.54 -0.07 LMA 0.42 -0.23 K -0.43 0.14 VD 0.16 0.68 VDi -0.17 -0.68

V. Conclusiones Generales

Conclusiones

Los rasgos funcionales ayudan a explicar la distribución de las especies a través de gradientes de humedad del suelo, debido a que determinan la habilidad de las especies para competir por agua y tolerar la sequía. El estudio de los rasgos funcionales y mecanismos fisiológicos que determinan la tolerancia a la sequía de las especies es importante para predecir sus respuestas a motores de cambio global como cambios climáticos y fragmentación del hábitat.

En la primera parte de la tesis se evaluaron rasgos foliares e hidráulicos relacionados con la tolerancia a la sequía en tres especies arbóreas con patrones contrastantes de distribución dentro de los parches dependientes de neblina en el Parque Nacional Fray Jorge. Los resultados mostraron que la distribución contrastante de las especies a través del gradiente de humedad del suelo a pequeña escala es explicada por las diferentes estrategias de uso del agua y carbono: Drimys winteri , especie restringida al núcleo húmedo de los grandes fragmentos, presentó rasgos que permiten un eficiente transporte de agua y ganancia de carbono; por el contrario, Myrceugenia correifolia , especie que domina los bordes secos de los fragmentos, exhibió rasgos que promueven la conservación del agua y bajas tasas fotosintéticas, así como menores punto de pérdida de turgor. Aextoxicon punctatum , la especie ampliamente distribuida entre zonas y fragmentos mostró valores intermedios de rasgos. Los datos demostraron que el particionamiento del gradiente de humedad desde el núcleo a los bordes más secos entre las especies arbóreas es dirigido por la tolerancia diferencial a la sequía, lo que implica habilidades contrastantes para lidiar con las futuras reducciones en humedad del suelo.

En la segunda parte de la tesis se estudió la variación de la magnitud (media), variabilidad (coeficiente de variación) y patrones de integración fenotípica de rasgos foliares e hidráulicos en Aextoxicon punctatum a través de las zonas de humedad del suelo en bosques de diferente tamaño en el Parque Nacional Fray Jorge. Los resultados variaron según los rasgos evaluados: individuos creciendo en los bordes secos mostraron mayores valores de LMA y densidad de tricomas y estomas que los individuos creciendo en las zonas más húmedas del núcleo y el borde barlovento. En contraste, los rasgos de la anatomía del xilema (diámetro y densidad de vasos conductores) no variaron entre zonas o tamaños de fragmentos, produciendo pérdida de conductividad hidráulica en las zonas más secas (sotavento). También se detectó mayor

79 integración fenotípica y variabilidad en sotavento. La habilidad de A. punctatum para modificar los rasgos foliares en respuesta a la disponibilidad de agua en el suelo facilita su persistencia en un amplio rango de microhábitats dentro de los fragmentos. Sin embargo, su limitada plasticidad en la anatomía xilemática amenaza el flujo de agua e incrementa el riesgo de cavitación en sotavento.

Los resultados obtenidos en esta tesis demuestran la importancia del estudio de rasgos funcionales para explicar patrones de coexistencia y distribución espacial de las especies a través de gradientes ambientales. Adicionalmente, son un insumo clave para predecir la respuesta de las especies a futuros cambios en el clima, información que debería ser incluida en los modelos de distribución de las especies bajo diferentes escenarios de cambio climático.

Con el desarrollo de esta tesis surgieron algunas preguntas que sería interesante responder a futuro:

El mantenimiento de la vegetación de los fragmentos de bosque de Fray Jorge está determinado por el balance entre la niebla en primavera – verano y la precipitación en temporada invernal. Este estudio se realizó durante la temporada primavera-verano donde los fragmentos dependen exclusivamente de la neblina costera y se genera el mayor gradiente de humedad. Sin embargo, sería clave monitorear el comportamiento de las especies arbóreas a través de las distintas temporadas del año y a través de años Niño y Niña que generan fuertes cambios en la precipitación y niebla. Este monitoreo permitirá tener un panorama más claro sobre el estrés hídrico que experimentan las especies y sobre las posibles tendencias climáticas en la zona.

Los rasgos foliares e hidráulicos relacionados con la tolerancia a la sequía ayudaron a explicar la distribución de las especies arbóreas a través del gradiente de humedad del suelo. Sin embargo, sería interesante explorar que otros factores podrían afectar estos patrones de distribución. Por ejemplo evaluar si existen limitaciones en la dispersión de las semillas o en el reclutamiento debido a la depredación de frutos, semillas o plántulas. Adicionalmente, sería importante incluir la medición de rasgos radiculares, por su importancia para la adquisición de agua-

A. puncatum y D. winteri son especies con amplia distribución en Chile. A futuro sería interesante entender cómo varían los rasgos foliares e hidráulicos a través de su rango de distribución y estudiar si las posibles diferencias son generadas por

80 plasticidad fenotípica o adaptación local. Si son adaptaciones generadas por presiones selectivas en largos periodos de tiempo, los cambios climáticos acelerados pueden ser más rápidos que la capacidad de las poblaciones a adaptarse. Así, esta información se convierte en un insumo clave para tratar de predecir la respuesta de las especies a variaciones ambientales.

V. Anexo I

Simetría de los parches de bosque depende de la dirección de los recursos limitantes

Stanton DE, Salgado-Negret B, Armesto JJ, Hedin LO. 2013. Forest patch symmetry depends on direction of limiting resource delivery. Ecosphere http://dx.doi.org/10.1890/ES13-00064.1

Forest patch symmetry depends on direction of limiting resource delivery.

Daniel E. Stanton 1,2,3 , Beatriz Salgado-Negret 2,4 , Juan J. Armesto 2,4 , Lars O. Hedin 1

1Department of Ecology and Evolutionary Biology, Princeton University, Princeton, New Jersey 08544 USA

2Departamento de Ecología, Pontificia Universidad Católica de Chile, Santiago, Chile

3Ecology, Evolution and Behavior Department, University of Minnesota, St. Paul, Minnesota 55108 USA

4Institute of Ecology and Biodiversity (IEB), Santiago, Chile

5Present address: Division of Plant Sciences, Research School of Biology, the Australian National University, Acton, ACT 0200, Australia.

E-mail: [email protected]

Abstract

Edge effects are a major concern in the study and conservation of forest patches. The traditional perspective, derived from patches formed by fragmentation, considers forest edges as intermediates in a gradient between interior and exterior conditions, symmetrically distributed around the core of the patch. We present a more general conceptual model that shows that this perspective is only one of several possible environmental gradients across forest patches. When resources are delivered horizontally (e.g., fog, surface runoff), environmental parameters and species composition are expected to have very different, asymmetric, distributions within forest patches. We conducted transect surveys characterizing environmental conditions (light, soil moisture, soil nutrients), vegetation structure and species composition in fog-fed patches of relict temperate forest in northern Chile. Windward edges differed most from the surrounding scrubland, whereas the core merely represented an intermediate between windward and leeward edges. Community composition changed drastically from temperate forest specialists on the windward edge to mediterranean shrub species leeward. The simple edge-core model is shown to be inadequate for describing spatial patterns in fog-influenced forests: a more universal model including the directionality of external resource inputs and internal dynamics must be considered when evaluating forest patch dynamics.

Keywords asymmetry; community composition; ecosystems; fog; forest patch; matorral; temperate rainforest; vegetation banding.

Introduction

Concerns over increasing forest fragmentation have drawn attention to the par ticularities of forest patches. The edge of a forest will be affected by the surrounding matrix outside the forest, and thus differ considerably from the interior. A consistent focus of the literature has been to evaluate how far into a forest these ‘edge effects’ penetrate. Within this c on text, forest patches have often bee n portrayed as an edge (bearing the influence of the area outside of the patch) surrounding a core unaffected by the external matrix (Murcia 1995). Often implicit in this understanding of forest patches is the assumption that small patches were formerly parts of a larger continuous forested area.

This idealization is inherently radially symmetrical when viewed from above. The width of the edge may be variable but the nucleus is conceived as a core around which approximately symmetrical sides extend (Fig. 1A). Intrinsic to this concept of the patch is an assumption that resources are delivered either vertically or diffuse horizontally from all sides equally. Edge effect s such as slanted light, wind and non- forest animals diffuse inwards towards the core from all sides. Vertically delivered resources (light and rain) are delivered approximately evenly to the entire top layer of the forest. Ecosystem properties that are tied to these resources, such as soil moisture, soil nutrients and plant height can therefore be expected to also show a radial symmetry around the core (Fig. 1B), as competition for the m will occur along the vertical axis.

Although often the case in antropogenically modified landscape s, patchiness is not necessarily formed by fragmentation (Rietkerk and Van De Koppel 2008). Vegetation patches can also arise by self-organization through local facilitation. For example, Klausmeier (1999) showed theoretically that forest patches can arise from directional surface-runoff in semi-arid ecosystems, and similar pat terns have been empirically been demonstrated to occur in a wide range of ecosystems, from fog-fed bromeliad fields in the Atacama (Borthagaray et al. 2010) through semi-arid shrublands (Klausmeier 1999, Van De Koppel and Rietkerk 2004, Saco et al. 2007, Kéfi et al. 2010) to Sphagnum aggregations in fens (Eppinga et al. 2008, 2009, Manor and Shnerb 2008) and tree islands in the Everglades (Wetzel et al. 2008). In all of these cases limiting resources such as water and nutrients are delivered in a horizontally asymmetric

86 manner (Fig. 1C). This strong directionality of resource delivery is likely to lead to asymmetrical distribution of related ecosystem properties (Fig. 1D).

We propose a single general conceptual model that unites symmetrical and asymmetric al forest patch structures. These two models are not so different when we consider that light is also a directional limiting resource in forests. Although light competition is considered a fundamental aspect of plant community structuring, it is rarely considered in the context of generating asymmetry along the axis of delivery, even if such a situation is evidently implied. This is qualitatively different from the reported difference s between north- and south- facing edges of forests (Wales 1972, Matlack 1993, 1994, Chen et al. 1996, Hylander 2005), which are attributed to effects of insolation (i.e ., response to microclimatic differences rather than growth towards light). The difference between vertically and horizon tally delivered resources is therefore best considered in terms of directionality relative to the axis of plant growth, either parallel (Fig. 1A) or perpendicular (Fig. 1C). Although light competition may be the most commonly known form vertical competition, rainwater and nitrogen can also be subject to ‘vertical processing’ (Ewing et al. 2009).

Many resources are effectively co-limiting at the ecosystem scale. For example, if we consider banded forests in semi-arid environments, surface run-off water (perpendicular) determines the presence and scale of forest patches, but light may structure vegetation within t he patch (parallel). As such, while some ecosystem properties (e.g., soil moisture) may be horizontally asymmetrical, others may be horizontally symmetrical (e.g., vegetation structure) (Fig . 1E). Spatial patterns in plant communities composition, which are driven by colimitation and trade-off between both paralleland perpendicular resources, will reflect an intersection of these bidirectional effects (Fig. 1F).

The consideration of directionality of resource delivery additionally challenges the static view of forest patches. In a vertically structured forest patch the dynamics of light competition will lead to upward growth and tradition al forest succesion (Horn 1971). If critical resources are delivered horizontally, we expect competition to occur along the horizontal axis, for example ups lope for water and nutrients from surface runoff (Saco et al. 2007), windward for resources from fog (Borthagaray et al. 2010) or leeward when wind is a cause of mortality (Watt 1947, Sprugel and Bormann 1981, Sato and Iwasa 1993). Since horizontal expansion is not constrained by the

87 biomechanical costs of overcoming gravity and retaining access to soil resources, it should become apparent as a directional progression of forest patches across a landscape, with considerable differences in community composition and ecosystem processes between leading and lagging edges.

To test this conceptual model of spatial distribution of resources within forest patches, we measured a number of above- and below-ground environmental variables across fog-dependent forest patches in northern Chile. These forest patches contain temperate rainforest trees far outside of their main climatic range in the midst of Mediterranean semi-arid matorral (Squeo et al. 2004). The fog-water inputs are strongly directional, and lead to large differences in tree recruitment and mortality between wind-ward and leeward patch edge s (del Val et al. 2006). This directionality makes these patches an ideal system in which to test whether ecosystem properties are more strongly associated with directionality of resource deliver y or simply symmetrically determined by distance from forest edge.

The spacing and width of fog-created banding is also dependent on slope: steep slopes decrease the strength of horizontal competition for fog, leading to broader bands or even continuous plant cover, whereas flatter ground encourage s the formation of narrow, widely spaced bands (Borthagaray et al. 2010). Fog forest relics occur in areas of highly variable topography, and larger patches tend to be associated with steeper slopes (Barbosa et al. 2010), and so the effects of directionality on resource distribution might be expected to weaken with increasing slope and patch size.

Considering that the primary source of water is horizontally driven fog, we predicted below-ground ecosystem properties controlled by water availability (soil moisture and nutrient availability) to be horizontally asymmetric al (Fig. 1 D), where as above ground properties (plant height, understory light availability, litter depth) to be driven by light competition, and therefore horizontally symmetrical (Fig. 1B). Plant community composition, which is expected to be driven by competition for light, water and nutrients, was predicted to reflect both vertical and horizontal influences (Fig. 1F).

We hypothesized that the fog influenced forest patches would not show a symmetrical resource distribution (Core > Windward Edge = Leeward Edge > Matrix; Fig. 1B) but rather an entirely asymmetrical (Windward Edge > Core > Leeward Edge > Matrix; Fig. 1D) or mixed (Windward Edge = Core > Leeward Edge > Matrix; Fig. 1F).

Furthermore, we predicted asymmetries to be stronger in the small patches, in which horizontal competition is expected to be stronger than in the larger patches.

Materials and methods

Site description

Research was conducted in Fray Jorge National Park, IVth Region, Chile (30°40´S, 71°83´W). A large number (370) of small patches of forest form a mosaic embedded in a xerophytic matorral scrubland (Squeo et al. 2004, del Val et al. 20 06, Gutiérrez et al. 2008, Barbosa et al. 2010). The persistence of these forest patches, whose species composition closely resembles Valdivian temperate rainforest (Villagrán et al. 2004), despite very low rain fall (147 mm annually) at Fray Jorge has been attributed to fogwater inputs (del Val et al. 2006). Forest patches span a wide range of sizes, from 0.1 to 36 ha (Barbosa et al. 2010) and in some areas form bands perpendicular to the predominant wind direction (Fig. 2).

Sampling design

Transects were established perpendicular to forest patch edges and parallel to the dominant wind direction. The length of each transect depended on the width of the forest patch crossed, and was chosen to extend at least 3 sampling points beyond both lee- and windward border. The ‘borders’ of the patch were determined as the first and last point along each transect at which at least one plant exceeded 3 m in height. Ten transects were conducted crossing a total of 14 patches, with several transects crossing more than one patch. The patches sampled had been identified as representative of the range of patch size s by previous studies (Barbosa et al. 2010), and can be roughly categorized as small (width < 30 m, area < 1 ha), medium (30 m < width < 100 m, 1 ha < area < 10 ha) and large (width > 100, area > 10 ha).

Light environments, vegetation structure and composition was assessed at 2-m intervals along each transect (4-m intervals in medium and large patches). Measurements of photosynthetically active radiation (PAR) were made using aquantum sensor (LI-1905B; LI-COR, Lincoln, Nebraska, USA) under uniformly cloudy conditions, and expressed as a percentage of the above-canopy PAR, which was

89 simultaneously recorded using a second quantum sensor that had been placed in a nearby forest clearing. PAR measurements along transects were taken 1m above the ground to represent the light environment of small saplings. The area surrounding the sampling point was divided into 4 equal quadrats. The height and species of the canopy overlying the sampling point, as well as the height and identity of the nearest woody plant species in each quadrat were recorded.

Volumetric soil moisture at 2-m intervals was recorded in situ using a hand-held TDR probe (Field scout TDR 100, Spectrum Technologies, Illinois, USA). Five measurements were recorded for each sampling point, after clearing away leaf litter and sub aerial roots. The depth of leaf litter was recorded when present.

Soil samples for soil nutrient content were collected at intervals of 2 m (small patches), 4 m (medium patches) or 8 m (large patches). Approximately 20 g of soil were collected, homogenized and oven-dried at 60 8 C to constant weight in the Biogeochemistry lab of the Pontificia Universidad Católica de Chile, Santiago, Chile. Subsamples (3 g) were sieved through 1-mm mesh and sent to the Hedin Lab, Prince ton University, New Jersey, US A, for analysis. Samples were ground by mortar and pestle and oven-dried at 60 °C for 3 days prior to carbon and nitrogen analysis using an elemental analyzer (Carlo Erba 4500, Costech, California, USA).

Data analysis

The points within each transect were partitioned according to location within the transect as one of four zones: core, leeward edge, windward edge, matrix. The core of each patch was defined as the region in which average plant height at each sampling point > 3 m. Edges were defined as those points within the patch (at least one plant > 3 m tall) but not contained in the core. The matrix was considered to be all points within a transect in which no plants exceeded > 3 m in height, corresponding to scrubland rather than forest.

All statistical analyses were performed using the open-source statistical soft ware program R (R Development Core Team 2012). The distributions of above and below-ground variables were evaluated by linear mixed effects models maximising log- likelihood using the function lme in R package nlme (Pinheiro et al. 2013). Soil moisture, plant height, leaf litter and light availability data were square-root transformed

90 for the analysis to conform to assumptions of normality and heteroscedasticity. Patch identity and zone were used as random effects and patch size, zone and the interaction of patch size and zone applied as fixed effects. Likelihood ratio tests were used to determine the best model for each. Single fixed effect models are compared to the null model, interaction models are compared to the relevant significant single fixed effect model. Data available was insufficient to fit full inter action models for soil carbon and soil nitrogen. To determine the pattern underlying significant fixed effects we conducted Tukey multiple comparisons applied to the fullest significant LME using R package multcomp (Hothorn et al. 2008).

Vegetation community composition was analyzed using Principal Coordinates Analysis (PCO). We computed floristic similarity between locations using a Sorensen dissimilarity and computed the two first axes of the PCO projection using R package labdsv (Roberts 2012). The first axis of the PCO provided a single variable descriptor of the community assemblage of each transect point. Linear mixed models using the PCO 1st axis as the dependent variable, patch size and/or zone as fixed effects and patch identity and zone as random effects were created and tested as above to determine the distribution of vegetation across patch zones and patch sizes. Furthermore, the proportional distribution of individual species across patch zones was evaluated.

Results

Spatial patterns of abiotic variables

All but one (C:N ratio) of the above and below-ground abiotic variables measured varied significantly with zone within the landscape (Table 1). Furthermore there were interactions between zone and patch size for soil moisture, leaf litter depth and understory light. Plant height was symmetrically distributed around the core (Fig. 3A), which is partly driven by the height–based definition of the zonation. Leaf litter depth did not differ within patches but was significantly greater than in the surrounding matrix (z=-5, P< 0.0001; Fig. 3B), the only interaction with patch size being the significance of the difference between the patch and the surroundings. No variables were found to be completely asymmetrically distributed (Windward Edge > Core > Leeward Edge > Matrix; Fig. 1D), however soil moisture, light, soil C and soil N all showed mixed symmetry (Windward Edge = Core > Leeward Edge > Matrix; Fig. 1F, Fig. 3C, D, Fig.

4A, B). Windward edges and cores were wettest and most shaded in the small patches, but not in the medium and large patch (Fig. 4A). Small patches showed asymmetrical distributions, with the degree of symmetry decreasing with size. Large patches showed the most symmetrical within-patch distributions (Fig. 4). Contrary to predictions soil C and N were marginally asymetrical (z = -2.550, P < 0.05217 and z = -2.408, P < 0.07464, respectively) with the greatest values found at the leeward edge and core (Fig. 3C, D).

Plant community composition

The PCO first axis was able to repre sent 38.5% of the variance in plant community composition. The woody plant community showed a significant response to patch zone (Table 1). Although patches always differed from the surrounding matrix, the within- patch distributions varied with patch size, from a symmetrical in small patches to symmetrical around the core in medium and large patches (Fig. 4C). When individual species are considered the patter ns are even more clearly pronounced. Species with strong temperate wet forest affinities (Villagrán et al. 2004) were predominantly found inside patches (Table 2). In small patches they showed asymmetric al preferences for the windward edge and core with a reduced presence at the leeward edge. In large patches the distribution was more frequently symmetrical, centering around the core of the patch for the trees Aextoxicon punctatum , Drimys winteri and Raphithamnus spinosus , but not for sclerophyllous trees Azara microphylla and Myrceugenia correifolia and woody vine Griselinia scandens .

Discussion

Above-ground variables

The fog-fed forest patches were poorly described by the tradition al symmetrical model, and showed strong directionality in several ecosystem properties. Many environmental variables showed horizontally asymmetric al distributions (Table 1). Although some of these distributions matched those predicted by our conceptual model, others differed from prediction either in symmetry or in the form of asymmetry.

Differences in tree survival and foliage retention are also a likely explanation for the striking asymmetry in understory light availability (Salgado-Negret et al. 2013). Although we predicted that the vertical competition for light within patches would lead to a symmetrical distribution (Fig. 1A, B), understory light availability actually appears to show an inverse response to soil water content (Fig. 4). We observed considerably denser living vegetation at windward than leeward edges, and high drought-induced mortality and leaf loss probably allow for far greater light penetration. The increased light penetration would then create a positive feedback by increasing evaporation from the soil surf ace. The reduced insolation on the wind ward edge should also translate into lowered soil temperatures and reduced vaporation rates from the soil, which may translate into reduce d drought and greater canopy density.

Below-ground variables

Soil characteristics were distinctly asymmetrical along a windward to leeward axis, as predicted, however the details of the distributions differed markedly from our hypotheses. Soil moisture, which is strongly influenced by fog water inputs, was expected to be greatest at the windward edge and decrease a cross the patch due to the progressive ‘filtering’ out of fog-droplets from the air by trees, as described in simpler fog-influenced banded vegetation (Borthagaray et al. 2010). However, although soil water content was high at the windward edge, it was comparable or greater in the patch core (Fig. 4A). Trees were significantly taller in the patch core, which may allow them to access fog water unavailable to shorter trees, thereby partially escaping the interference effects of upwind competitors. Soil carbon and nitrogen increased greatly from the windward edge to the core, before decreasing again more gradually to leeward. This suggests that the availability of soil nutrients is not a simple function of moisture and litter inputs, and may instead be indicative of more complex ecosystem dynamics, as discussed below.

Fundamental differences between patch types

Small and large patches differed considerably in their spatial structure, both above- and below-ground. Barbosa et al. (2010) characterized the microclimatic and structural characteristics of forest patches (including a subset of those sampled in the present

93 study) representative of different sizes. One of the traits reported but not commented on is that small patches occur on flat ground where as most medium and large patches are found on steep slopes (30 – 45°S). Windflow over flat areas will be strongly affected by the boundary layer created by a forest edge, and forest patches will leave a long wakes in which little to no fog water is available, until airflow (and fog water) are replenished downstream (Oke 1987 ). These ‘fog-shadows’ (del Val et al. 2006) are likely to be far less pronounced or potentially absent on steep slopes (Borthagaray et al. 2010), reducing or eliminating the competition for fog-water between trees. This difference in topography may explain reduced asymmetry in large and medium patches (Fig. 4). In the large and medium patches topography may still play some role: the leeward edge is always associated with the flattening out of the terrain at a ridge crest, where as the small patches are topographically homogeneous and flat throughout.

It is also important to clarify that several of the small patches sampled in this study (but not in Barbosa et al. 2010) do occur on steep slopes. However, they are located such that the slope does not interact with wind direction (see Fig. 2), and therefore there is little to no sloping along the actual windpath. This observation supports our interpretation the asymmetries are due to directional fog inputs rather than by the differing solar radiation that can be create d by sloping terrain (e.g., Tian et al. 2001, Allen et al. 2006). Variations on incoming solar radiation associate d to slope steepness and orientation may favor moist retention and most likely plays an important role in ecosystem dynamics, however, in the present study it is unlikely to be the primary factor in the formation of patch asymmetries.

If water availability is indeed a primary driver of spatial distributions of other ecosystem properties, then it is perhaps unsurprising that small patches, in which competition for fog water will be strongly asymmetric al, show far more marked differences between windward edge and core than do the larger patches, in which windward and core trees likely have access to comparable water inputs. This fundamental biophysical difference leads to a reinterpretation of Barbosa et al.’s (2010) fin ding that small patch microclimates were more strongly impacted by edge effects. Flat- ground (small) patches will have greater depletion of fog water by the windward edge, such that the patch interior and leeward edges will be dryer than in larger patches. This effect will amplify the edge effect (in the usual sense of the term) of the mostly dead leeward edge allowing for increased insolation of the patch interior.

Plant community as an integration of co-limiting factors

Vascular plant communities are often structured by competition for numerous potentially limiting resources, such as light, water and nutrients. Having predicted that these different resources would have different spatial distributions driven by directionality of resource delivery, we hypothesized that plant communities would reflect overlapping effects of vertical inputs (light and rain fall, Fig. 1A, B) and horizontal inputs (fog water and nutrients, Fig. 1C, D). Principal components ordination clearly distinguished between forest and matorral plant communities (Fig. 4C). Contrary to our predictions, understory light availability was strongly horizontally asymmetrical in all patches, and itself possibly driven by positive feedbacks with asymmetric soil water availability (Fig. 4A, B). As such, plant community composition was also strikingly asymmetrical across patches, especially in small patches. Larger patches were symmetrical in nature, with some more arid adapted shrubs ( Myceugenia correifolia , Kageneckia oblongata ) present at both edges (Table 2). The differential microclimatic conditions across these patches may also lead to ecophysiological differences between individuals in those species that span the patches (Salgado-Ne gret et al. 2013).

Forest patches as self-organizing ecosystems

The spatial asymmetries of soil carbon and nitrogen content (Fig. 3C, D), while differing from those predicted, they are in line with a dynamic view of forest patches. Del Val et al. (2006) have proposed, based on the strong asymmetry in recruitment and mortality between edges, that forest patches at Fray Jorge may be progressively moving windward across the landscape. Under such a scenario, windward edges would be the youngest, and considering that matorral soils are very carbon and nitrogen-poor, the greater carbon and nitrogen content in core and leeward soils (Fig. 3C, D) may in fact reflect the greater accumulation of organic matter and nutrients. The transition from matorral to forest soils and communities across very small spatial scales (at times <5 m) may therefore reflect the build-up of water and nutrient cycling facilitated by fog collection. Such a self-organization of forest patches will leave a trail of modified above-and below-ground ecosystem attributes in its leeward wake. The presence of such

95 a ‘foot-print’ of forests past can indeed be identified, and will be the subject of a forthcoming paper (Stanton et al., unpublished manuscript).

Several forest associated species, such as Aextoxicon punctatum , Griselinia scandens and Myrceugenia correifolia were occasionally found outside of the forest patches (Table 2). These individuals, when not just windward of the forest edge, formed small ‘mini-patches’ that may be incipient forest patches. The long-term persistence of windward migrating forest patches requires the regeneration of patches downwind. The mechanisms for formation of these patches are unknown, and may be associated with exceptional weather events, such as large El Niño-Southern Oscillation (ENSO) events, as is the case for tree recruitment in other semi-arid locations (Holmgren et al. 2006).

It is well understood that species will assort along environmental gradients such as those found across forest patches. However, general models of how these gradients themselves form are more often overlooked or implicitly assumed. We have shown that the direction of delivery of limiting resources drives the spatial asymmetries in forest structure. Symmetrical forest patches consisting of a core and periphery are but a special (albeit widespread) case of forest patch structure, in which t he primary directional limiting resources (water and light) are delivered vertically. Not all natural ecosystems incorporating a mosaic of small forest patches may show the same directionality. For example, Silva and Anand (2011) studied Araucaria forest patches that exhibited asymmetries, but without the strong directionality that we have documented in Fray Jorge. In such cases feedbacks from the surrounding matrix (e.g., fire, competition with shrubs or grasses, different water and nutrient availability, soil microbial communities) may act to stabilize patches. In yet other ecosystems the driver of asymmetries may be seed rain, nutrient deposition (Weathers 1999, Ewing et al. 2009), run off (Klausmeier 1999, Van De Koppel and Rietkerk 2004, Saco et al. 2007, Kéfi et al. 2010) or frost damage (Watt 1947, Sprugel and Bormann 1981, Sato and Iwasa 1993). The conceptual frame work and empirical confirmation presented here are a step towards a more inclusive understanding of forest patches and their internal and external dynamics.

Acknowledgements

This research was funded by NSF DDIG award # 0909984 to L. H. and D. S.; Princeton Latin American Studies Travel Grants and a Princeton President’s Award to D. S. and

CONICYT fellowship 24110074 to B.S-N. Research in Chile was conducted under CONAF research permit 06/08. We would like to extend special thanks to Patricio Valenzuela, María Fernanda Pérez and the CONAF staff at Fray Jorge for support in the field, Aurora Gaxiola, Pablo Marquet, Adam Wolf, Carla Staver and members of the Armesto and Perez labs for their support and discussion of ideas as well as Madhur Anand and two anonymous reviewers for suggestions that have greatly improved the manuscript.

References

Allen, R. G., R. Trezza, and M. Tasumi. 2006. Analytical integrated functions for daily solar radiation on slopes. Agricultural and Forest Meteorology 139:55-73.

Barbosa, O., P. A. Marquet, L. D. Bacigalupe, D. A. Christie, E. Del-Val, A. G. Gutierrez, C. G. Jones, K. C. Weathers, and J. J. Armesto. 2010. Interactions among patch area, forest structure and water fluxes in a fog-inundated forest ecosystem in semi-arid Chile. Functional Ecology 24:909-917.

Borthagaray, A. I., M. A. Fuentes, and P. A. Marquet. 2010. Vegetation pattern formation in a fog-dependent ecosystem. Journal of Theoretical Biology 265:18- 26.

Chen, J., J. Franklin, and J. Lowe. 1996. Comparison of abiotic and structurally defined patch patterns in a hypothetical forest landscape. Conservation Biology 10:854- 862. del Val, E., J. Armesto, O. Barbosa, D. Christie, A. Gutiérrez, C. Jones, P. Marquet, and K. Weathers.2006. Rain forest islands in the Chilean semiarid region: fog- dependency, ecosystem persistence and tree regeneration. Ecosystems 9:598- 608.

Eppinga, M., M. Rietkerk, W. Borren, E. Lapshina, W. Bleuten, and M. Wassen. 2008. Regular surface patterning of peatlands: confronting theory with field data. Ecosystems 11:520-536.

Eppinga, M. B., P. C. De Ruiter, M. J. Wassen, and M. Rietkerk. 2009. Nutrients and hydrology indicate the driving mechanisms of peatland surface patterning. American Naturalist 173:803-818.

Ewing, H. A., K. C. Weathers, P. H. Templer, T. E. Dawson, M. K. Firestone, A. M. Elliott, and V. K. S. Boukili. 2009. Fog water and ecosystem function: heterogeneity in a California redwood forest. Ecosystems 12:417-433.

Gutiérrez, A., O. Barbosa, D. Christie, E. Del-Val, H. Ewing, C. Jones, P. Marquet, K. Weathers, and J. Armesto. 2008. Regeneration patterns and persistence of the fog-dependent Fray Jorge forest in semiarid Chile during the past two centuries. Global Change Biology 14:161-176.

Holmgren, M., B. López, J. Gutiérrez, and F. Squeo. 2006. Herbivory and plant growth rate determine the success of El Niño Southern Oscillation-driven tree establishment in semiarid South America Global Change Biology 12:2263-2271.

Horn, H. S. 1971. The adaptive geometry of trees. Princeton University Press, Princeton, New Jersey, USA.

Hothorn, T., F. Bretz, and P. Westfall. 2008. Simultaneous inference in general parametric models. Biometrical Journal 50:346-363.

Hylander, K. 2005. Aspect modifies the magnitude of edge effects on bryophyte growth in boreal forests. Journal of Applied Ecology 42:518-525.

Kéfi, S., M. B. Eppinga, P. C. Ruiter, and M. Rietkerk. 2010. Bistability and regular spatial patterns in arid ecosystems. Theoretical Ecology 3:257-269.

Klausmeier, C. 1999. Regular and irregular patterns in semiarid vegetation. Science 284:1826-1828.

Manor, A., and N. Shnerb. 2008. Facilitation, competition, and vegetation patchiness: From scale free distribution to patterns. Journal of Theoretical Biology 253:838- 842.

Matlack, G. R. 1993. Microenvironment variation within and among forest edge sites in the eastern United States. Biological Conservation 66:185-194.

Matlack, G. R. 1994. Vegetation dynamics of the forest edge: trends in space and successional time. Journal of Ecology 82:113-123.

Murcia, C. 1995. Edge effects in fragmented forests: implications for conservation. Trends in Ecology and Evolution 10:58-62.

Oke, T. R. 1987. Boundary layer climates. Routledge, Oxon, UK.

Pinheiro, J., D. Bates, S. DebRoy, D. Sarkar, and R Core Team. 2013. nlme: Linear and nonlinear mixed effects models. http://CRAN.R-project.org/package=nlme .

R Development Core Team. 2012. R: A language and environment for statistical computing. R Founda-tion for Statistical Computing, Vienna, Austria.

Rietkerk, M., and J. Van De Koppel. 2008. Regular pattern formation in real ecosystems. Trends in Ecology and Evolution 23:169-175.

Roberts, D. W. 2012. labdsv: Ordination and multivariate analysis for ecology. http://CRAN.R-project . org/package¼labdsv

Saco, P., G. Willgoose, and G. Hancock. 2007. Ecogeomorphology of banded vegetation patterns in arid and semi-arid regions. Hydrology and Earth System Sciences 11:1717-1730.

Salgado-Negret, B., F. Pérez, L. Markesteijn, M. J. Castillo, and J. J. Armesto. 2013. Diverging drought-tolerance strategies explain tree species distribution along a fog-dependent moisture gradient in a temperate rain forest. Oecologia doi: 10.1007/s00442-013-2650 -7

Sato, K., and Y. Iwasa. 1993. Modeling of wave regeneration in subalpine Abies forests: population dynamics with spatial structure. Ecology 74:1538-1550.

Silva, L. C. R., and M. Anand. 2011. Mechanisms of Araucaria (Atlantic) forest expansion into southern Brazilian grasslands. Ecosystems 14:1354-1371.

Sprugel, D., and F. Bormann. 1981. Natural disturbance and the steady state in high- altitude balsam fir forests. Science 211:390-393.

Squeo, F., A. Gutiérrez, and I. Hernández, editors. 2004. Historia Natural del Parque Nacional Bosque Fray Jorge. Ediciones Universidad de La Serena, Chile.

Tian, Y. Q., J. Davies-Colley, P. Gong, and B. W. Thorrold. 2001. Estimating solar radiation on slopes of arbitrary aspect. Agricultural and Forest Meteorology 109:67-74.

Van De Koppel, J., and M. Rietkerk. 2004. Spatial interactions and resilience in arid ecosystems. American Naturalist 163:113-121.

Villagrán, C., J. Armesto, F. Hinojosa, J. Cuvertino, F. L. Pérez, and C. Medina. 2004. Historia Natural del Parque Nacional Bosque Fray Jorge, Chapter El enigmático

origen del bosque relicto de Fray Jorge. Ediciones Universidad de La Serena, La Serena, Chile.

Wales, B. A. 1972. Vegetation analysis of north and south edges in a mature oak- hickory forest. Ecological Monographs 42:451-471.

Watt, A. 1947. Pattern and process in the plant community. Journal of Ecology 35:1–22.

Weathers, K. 1999. The importance of cloud and fog in the maintenance of ecosystems. Trends in Ecology and Evolution 14:214-215.

Wetzel, P. R., A. G. Valk, S. Newman, C. A. Coronado, T. G. Troxler-Gann, D. L. Childers, W. H. Orem, and F. H. Sklar. 2008. Heterogeneity of phosphorus distribution in a patterned landscape, the Florida Everglades. Plant Ecology 200:83-90

100

Table 1. Effects of forest patch size and location with patch (zone) on soil moisture, plant height, leaf litter depth, understory light availability, soil carbon, soil nitrogen and woody plant community composition in linear mixed effects models (LME).

Dependent Fixed variable df AIC ∆AIC Likelihood p Pattern variable ratio

Soil Patch size 6 4419.1 6.3 10.304 0.0058 moisture Zone 7 4427.1 4.8 42.569 <0.0001 Mixed

Patch size x Zone 15 4461.4 -34.3 25.472 0.0013 Mixed-Sym

Plant height Patch size 6 9163.8 -2 2.051 0.3587

Zone 7 9107.1 54.7 60.724 <0.0001 Sym

Patch size x Zone 15 9115.6 -8.5 7.516 0.4821

Leaf litter Patch size 6 673.1 -1.6 2.360 0.3073 depth Zone 7 644.8 -1.6 32.690 <0.0001 Sym

Patch size x Zone 15 643.9 26.7 18.929 0.0309 Sym

Understory Patch size 6 623.0 -2.6 1.402 0.4962 light Zone 7 602.2 18.3 24.205 <0.0001 Mixed

Patch size x Zone 15 595.1 7.1 23.202 0.0031 Mixed-Sym

Soil carbon Patch size 6 1000.9 -1.9 2.117 0.3470

7.18 13.110

Zone 7 991.9 7.18 13.110 0.0044 Mixed

Soil Patch size 6 140.6 -1.9 2.161 0.3394 nitrogen Zone 7 133.4 5.3 11.302 0.0102 Mixed

C:N Patch size 6 663.6 -3.3 0.687 0.7093

Zone 7 661.4 -1.1 4.812 0.1861

Community Patch size 6 -1194.9 -2.1 1.911 0.3846 composition Zone 7 -1246.7 49.7 55.796 <0.0001 Mixed

Patch size x Zone 15 -1250.1 3.4 19.363 0.0130 Mixed-Sym

101

Table 2. Species and distribution of woody vascular plants (and the comparably sized bromeliad Puya) large found in forest patch transects.

Species Family Small patches (%) Large patches (%)

WE C LE M WE C LE M

Aextoxicon punctatum Aextoxicaceae 32 49 3 15 6 76 17 1

Ageratina glechonophylla Asteraceae 50 0 17 33 11 0 0 89

Azara microphylla Salicaceae 0 89 0 11 36 19 17 28

Baccharis linearis Asteraceae 0 0 0 100 0 0 0 0

Baccharis vernalis Asteraceae 5 3 5 87 0 0 9 91

Bahia ambrosoides Asteraceae 0 0 0 100 0 0 0 0

Berberis actinacantha Berberidaceae 20 0 0 80 0 0 0 100

Calceolaria integrifolia Calceolariaceae 0 0 0 0 0 0 0 100

Colletia spinosa Rhamnaceae 0 0 0 100 0 0 0 100

Colliguaja odorifera Euphorbiaceae 0 0 0 0 100 0 0 0

Drimys winteri Winteraceae 0 0 0 0 5 92 3 0

Echinopsis chilensis Cactaceae 0 0 0 100 0 0 0 0

Erigeron luxurians Asteraceae 0 0 2 98 0 0 0 100

Eupatorium salvia Asteraceae 0 0 8 92 6 4 2 88

Fuchsia lysioides Onagraceae 0 0 0 100 0 0 0 0

Griselinia scandens Griselinaceae 26 32 16 26 10 28 47 16

Haplopappus foliosus Asteraceae 0 0 0 100 0 0 0 0

Kageneckia oblonga Rosaceae 0 0 0 100 0 0 0 100

Myrceugenia correifolia Myrtaceae 22 37 18 22 9 31 38 23

Puya chilensis Bromeliaceae 0 0 0 100 0 0 0 0

Raphithamnus spinosus Verbenaceae 8 58 8 25 8 69 18 5

Ribes punctatum Grossulariaceae 0 7 0 93 0 0 29 71

Senecio planiflorus Asteraceae 9 0 0 91 0 0 0 100

102

Figure caption

Figure 1. Hypothetical resource distributions predicted according the directionality of resource input: (A, B) vertical inputs only (e.g., rainfall, parallel to the direction of plant growth); (C, D) horizontal only (e.g., fog or slope runoff, perpendicular to the direction of plant growth), and (E, F) both vertical and horizontal (e.g., fog and rainwater inputs, bidirectional). The principal axis of plant growth is illustrated by the dotted line. Soil based resources (e.g., %C, %N) are controlled by water availability, and thus indirectly controlled by water input direction (upward arrows in panels A, C and E).

Figure 2. Small forest patches in Fray Jorge National Park, IVth Region, Chile (30°84´ S, 71°30´ W), as seen from the leeward side. The temperate forest patches are easily distinguished from the surrounding arid matorral. The asymmetry of the patches is also clearly visible, with leeward plants primarily dead and windward plants with full foliage. The arrow indicates the primary direction of fog entering from the nearby coast. Photo by D. Stanton.

Figure 3. Boxplots of distributions of (A) plant height, (B) leaf litter depth, (C) total soil carbon, and (D) total soil nitrogen with location within patches (windward edge, core, leeward edge and surrounding matrix). Thick lines represent the median, boxes represent the interquartile range, whiskers represent maxima and minima within 1.5 times the interquartile range and open circles show outliers. Letters indicate significantly different groups (p < 0.05) as determined by Tukey HSD multiple comparisons applied to an LME model with zone as a fixed effect (see Methods and Table 2).

Figure 4. Boxplots of distributions (A) of soil moisture, (B) understory light availability, and (C) woody plant community composition with patch size (small, medium, large) and location within patches (windward edge, core, leeward edge and surrounding matrix). Thick lines represent the median, boxes represent the interquartile range, whiskers represent maxima and minima within 1.5 times the interquartile range and open circles show outliers. Letters indicate significantly different groups (p < 0.05) as

103 determined by Tukey HSD multiple comparisons applied to the LME model of the Patch Size x Patch Zone interaction (see Methods and Table 2).

104

Fig. 1.

105

Fig. 2.

106

Fig. 3.

107

Fig. 4.