Understanding Tropical Biodiversity to Improve Reforestation Outcomes

Home , Alstonia macrophylla, Artocarpus odoratissimus, Defaunation, Ficus nota, Lansium parasiticum, Sandoricum

Understanding tropical biodiversity to improve reforestation outcomes: examples from the Philippines and Northern Australia Jarrah Dominic Wills Bachelor of Applied Science (Ecology), Queensland University of Technology

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2016 School of Agriculture and Food Sciences Abstract

Tropical regions such as the Philippines and the Wet Tropics (WTs) bioregion of Northern Australia have seen dramatic declines in forest cover, primarily since respective Spanish, American and English colonization events. This has led to declines in biodiversity and ecosystem services, both of which support the livelihoods of a vast number of people. For example, drastic declines in forest cover in the Philippines and Northern Australia have seen large reductions in timber resources. In view of this, different reforestation types, including the management of ‘secondary forest’ regeneration, are being applied across the tropics in order to mitigate biodiversity loss and contribute to ecosystem services including timber production and biodiversity enhancement. To date, the ecology of these novel forests within the study regions have not been thoroughly investigated, from either conservation or a socio-economic perspectives. This research applies functional and phylogenetic approaches to gain insights into complex ecological processes such as regeneration and growth dynamics within different forest types. The forest types studied were established as small-scale community-based plantations located on the Island of Leyte in the Philippines, and included monocultures of Swietenia macrophylla King (mahogany), one of the most valuable timber species in the tropics; and mixed-species plantations established as a reforestation system called ‘Rainforestation Farming’ (here after, Rainforestation). Naturally regenerating selectively logged native forest located within Leyte and the WTs Bioregion was also included in this study. Within the Philippines, species, functional and phylogenetic diversity of understorey recruitment beneath the different forest types were compared and contrasted, in order to identify species or groups that were favoured or limited, and to compare community assembly processes. Within the WTs bioregion, I utilized a unique historical dataset, to investigate if leaf functional traits, such as Specific Leaf Area (SLA) or leaf nutrient contents, obtained from regenerating selectively logged forest trees, can be used to predict tree growth over time and how this relationship may vary depending on the position in the canopy. An improved understanding of the influence of these leaf traits could help to inform practitioners’ design of reforestation projects. Understorey recruitment, including species, functional and phylogenetic diversity, was influenced by the reforestation type in the Philippines. Generally, monocultures had the lowest diversity and regenerating selectively logged forest the highest, with the Rainforestation showing intermediate understorey diversity. Species within the Moraceae family were common across all forest types and wind-dispersed emergent-tree species were absent within the monoculture forest types. These trees represent a limited group of species, and their absence has consequences for future ecosystem function. Using phylogenetic relatedness and leaf traits, seedling communities were assembled II more by environmental filtering within monoculture forest types. Both environmental filtering and competitive exclusion seemed to be operating within the seedling communities of regenerating selectively logged forests. Human-assisted dispersal increased the levels of understorey diversity within monoculture forest types, indicating simple monocultures can provide some biodiversity benefits, and that these benefits are valued by local communities. Within regenerating selectively logged forests of the WTs bioregion, leaf functional trait and growth rate analysis revealed that leaf traits are weakly correlated to growth for trees that have reached the top layer of the canopy. However, for sub-canopy trees, leaf traits and growth relationships become stronger, suggesting that soft leaf traits are a stronger determinant of growth and success in the sub-canopy, whereas in the canopy other traits become more important. In particular leaf phosphorus alone positively explained the variation in Periodic Annual Increment (PAI). This highlights that a more nuanced outlook on leaf traits is required if they are going to provide insights into growth rates across multiple life-stages, and in order for these traits to have practical reforestation applications. Collectively, my research highlights the value of analysing additional aspects of biodiversity, in order to gain insights into ecological questions relating to reforestation outcomes. Specific implications of my study include an increased understanding of the conservation and reforestation significance of wind-dispersed emergent tree species, and species that are phylogenetically and functionally distinct (in terms of the mean and variation in leaf traits), resulting in greater ecosystem functioning; the conservation and socio-economic value of small-scale community monoculture plantations within human-dominated tropical landscapes; and that a more detailed understanding of leaf trait and growth relationships is needed if this theoretical research discipline is to benefit practical reforestation efforts. Some important future research questions identified as a result of my research include: What are the environmental factors (e.g., cyclone frequency, rainfall, elevation etc.) that influence the relative biomass of wind- versus animal-dispersed tree species? How does this vary between tropical continents? Can functional trait ecology provide more detailed groupings beyond pioneer versus non-pioneer species? Can these groupings be replicated across tropical forest ecosystems?

III

Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis.

Publications during candidature

Peer-reviewed papers Wills, J., Herbohn, J., Moreno, M.O.M., Avela, M.S. & Firn, J. (2016) Next-generation tropical forests: reforestation type affects recruitment of species and functional diversity in a human- dominated landscape. Journal of Applied Ecology, doi:10.1111/1365-2664.12770.

Conference abstracts Jarrah Wills, Jennifer Firn and John Herbohn, 2014. Understory species and functional diversity beneath different reforestation methods. Accepted by The Association for Tropical Biology and Conversation (ATBC) Conference, Cairns, Australia 20-24 July 2014.

Jarrah Wills, Jennifer Firn and John Herbohn, 2015. Understory species, functional and phylogenetic diversity beneath different reforestation methods. Accepted by The IUFRO 3.08 Small-Scale and Community Forestry and the Changing Nature of Forest Landscapes, Sunshine Coast, Australia, 11-15 October 2015.

Publications included in this thesis I have chosen to incorporate one publication into my thesis as per UQ policy (PPL 4.60.07 Alternative Thesis Format Options).

Wills, J., Herbohn, J., Moreno, M.O.M., Avela, M.S. & Firn, J. (2016) Next-generation tropical forests: reforestation type affects recruitment of species and functional diversity in a human- dominated landscape. Journal of Applied Ecology, doi:10.1111/1365-2664.12770 Incorporated as Chapter 3.

Contributor Statement of contribution Author Jarrah Wills (Candidate) Designed experiment (50%) Wrote the paper (90%) Conducted field and lab work (60%) Author Jennifer Firn Designed experiment (40%) Wrote and edited paper (15%) Author John Herbohn Designed experiments (10%) Wrote and edited paper (5%)

Author Ofelia Moreno Conducted field and lab work (20%)

Author Mayet Avela Conducted field and lab work (20%)

Contributions by others to the thesis

In addition to my supervisors Associate Professor Jennifer Firn and Professor John Herbohn who contributed throughout my thesis; Dr Jessie Wells provided feedback and advice for chapter 4, Dr Robert Congdon and Professor Angela Almendras-Ferraren provided input for chapters 3 and 4. Dr Don Butler also provided feedback for chapter 5.

Statement of parts of the thesis submitted to qualify for the award of another degree

None.

VII

Acknowledgements

I would like to thank those who have directly or indirectly contributed to my thesis.

I am very grateful to Associate Professor Jennifer Firn and Professor John Herbohn for their support and guidance throughout my PhD, which has made this research possible. Specifically I would like to thank Drs Arturo Pasa and Nestor Gregoria, and Professor Angela Ferraen from Visayas State University, and ACIAR project staff Val Solano, Nayar Adanarg, Chris Solano, Mayet Avela, Rogelio Tripoli, Nova Parcia, Jufamar Fernandez and Ofelia Moreno for their help and guidance. Dr Robert Congdon’s knowledge of analytical chemistry and botany was also greatly appreciated for both the Philippines and Australian projects. Jun Zhang and Sharif Mukul also made my stay in the Philippines very enjoyable. Professors Robin Chazdon, David Lamb, Susanne Schimdt, Jerry Vanclay, and Drs Jack Baynes, Bill McDonald, Huong Nguyen and Tony Page, and Grahame Applegate and Mark Annandale also contributed to my thesis by providing insightful discussions and advice, often on rainy days on field trips. This work has been made possible by an Australian Postgraduate Award scholarship and with financial assistance from the Australian Centre for International Agricultural Research (ACIAR- ASEM/2010/50). A special thanks to my colleagues and friends Jing Hu, Md Shawkat I. Sohel, Sharif Mukul, Liz Ota, Kurt Von Kleist, Gene Mason and John Meadows for their support and company throughout my PhD. Finally I would like to thank my family and friends for their help in conducting field work and for their support and friendship. I could not have finished this work without their help.

VIII

Keywords tropical forest, reforestation, recruitment, biodiversity conservation, Tropical North Queensland, phylogenetic comparative ecology, functional ecology, Philippines, functional traits

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 050102, Ecosystem Function, 60% ANZSRC code: 050104, Landscape Ecology, 20% ANZSRC code: 060202, Community Ecology, 20%

Fields of Research (FoR) Classification

FoR code: 0501, Ecological Applications, 70% FoR code: 0502, Environmental Science and Management, 30%

Table of Contents List of Supplementary Material ...... 9 Chapter 1: Introduction ...... 15 1.1 Background ...... 15 1.2 Research problem and Questions ...... 16 1.3 Justification ...... 18 1.4 Methodology ...... 18 1.5 Limitations ...... 19 1.6 Structure and outline of the thesis ...... 20 References ...... 22 Chapter 2: Literature Review ...... 24 2.1 Tropical forests ...... 24 2.2 Dependency of people on tropical forests ...... 26 2.2.1 Situation within the Philippines: forest biodiversity and sustainable livelihoods ...... 27 2.2.2 Situation within Tropical North Queensland: forest biodiversity and sustainable livelihoods ...... 28 2.3 Tropical reforestation ...... 31 2.3.1 Monocultures ...... 32 2.3.2 Mixed-species plantations ...... 33 2.3.3 Human-modified tropical landscapes...... 34 2.4 Community assembly underneath tropical forests: Regeneration and successional dynamics ... 37 2.4.1 Dispersal limitations to recruitment ...... 39 2.5 Human impacts on plant dispersal (subsistence hunting and habitat fragmentation) ...... 42 2.6 Functional trait approaches to understanding ecological processes...... 42 2.6.1 Which functional traits to use and why? ...... 43 2.7 Examining evolution and ecological traits together; advantages to understanding community assembly...... 45 References ...... 48 Chapter 3. Next-generation tropical forests: reforestation type affects recruitment of species and functional diversity in a human-dominated landscape ...... 63 3.1 Introduction ...... 63 3.2. Materials and methods ...... 65 3.2.1. Study Sites...... 65 3.2.2. Data Collection ...... 68 3.2.3. Data analysis ...... 69 3.3. Results ...... 70

3.3.1. Comparison of the richness and abundance of recruited species under three forest types ...... 70 3.3.2. Compositional differences in seedling community assemblages under three forest types ...... 73 3.3.3. Compositional differences in the functional traits of community assemblages under three forest types ...... 74 3.4. Discussion ...... 77 3.4.1. Seedling species diversity and composition beneath each forest type ...... 77 3.4.2. Species and functional traits favoured across forest types ...... 78 3.4.3. Species and traits limited across forest types ...... 79 3.5. Conclusion ...... 79 References ...... 80 Chapter 4. Reforestation methods influence community assembly in tropical forest understoreys: insights using functional traits and phylogenetic diversity ...... 108 4.1. Introduction ...... 108 4.2. Methods ...... 111 4.2.1. Study sites and Data Collection ...... 111 4.2.2. Functional traits...... 111 4.2.3. Community phylogeny ...... 112 4.2.4. Data analysis ...... 113 4.3. Results ...... 115 4.3.1. What is the phylogenetic and functional trait structure of seedling communities beneath the different forest types? ...... 115 4.3.2. SLA and variation beneath the different forest types and between common or obligate clades ...... 120 4.4. Discussion ...... 123 4.4.1. Phylogenetic and functional structure beneath forest types ...... 125 4.4.2. Intraspecific variation in SLA, comparing functional groups and forest types ...... 126 4.5. Conclusion and implications for tropical reforestation ...... 127 References ...... 128 Chapter 5. Leaf traits predict growth rates of sub-canopy trees but not canopy trees in Australian regenerating selectively logged tropical forests ...... 142 5.1. Introduction ...... 142 5.2. Materials and Methods ...... 144 5.2.1. Study Sites...... 144 5.2.2. Historical Data and Site History ...... 144 5.2.3. Tree selection and sampling ...... 145 5.2.4. Data Analysis ...... 146 5.3. Results ...... 148 5.3.1. Are trees growth rates correlated with leaf traits? ...... 148

5.3.2. Relationships between growth and leaf traits for different canopy strata ...... 150 5.3.3. Are relationships between leaf traits and between leaf traits and tree height consistent between canopy strata? ...... 152 5.4. Discussion ...... 153 5.4.1. Are tree growth rates related to leaf traits? ...... 154 5.4.2. Growth rate and leaf trait relationships between canopy strata ...... 154 5.4.3. Are relationships between leaf traits consistent between canopy strata? ...... 155 5.5. Implications for reforestation ...... 156 Chapter 6. General Conclusion ...... 172 6.1 Summary ...... 172 6.2 A synthesis of understorey recruitment beneath contrasting reforestation types ...... 172 6.2.1. Summary of key outcomes ...... 172 6.2.2. Limitations and further work ...... 173 6.3 A synthesis of predicting growth rates using simple morphological leaf traits ...... 174 6.3.1. Summary of key outcomes ...... 174 6.3.2. Limitations and further work ...... 175 6.4 Final conclusion ...... 176 References ...... 176

List of Figures

Figure 3.1. Leyte Island, Philippines. Symbols indicate study sites and shaded areas 66 approximately represent regenerating selectively logged forest. Figure 3.2. Pictures of the three forest types; monoculture forests (a), Rainforestation 68 forests (b) and regenerating selectively logged forests (c).

Figure 3.3. Higher order fixed effects from LMEM terms where the response variable 73 is seedling species richness of all growth forms (a-d) and limited to trees and shrubs (e- h), and how it varies with soil phosphorus (c and g), soil nitrogen (b and f) and LAI (d and h). Error bars and shaded regions represent ± standard error. * denotes significant relationships, using F-statistics (Table S3.1).

Figure 3.4. Non-metric multidimensional (nMDS) ordinations (using Bray-Curtis 75 similarity index) of species richness (a, stress: 0.16) and species abundance (b, stress: 0.17) (all growth forms). Colours represent plots surveyed in monoculture forests (light grey), Rainforestation forests (dark grey) and regenerating selectively logged forests (black).

Figure 3.5. Higher order fixed effects from LMEM terms where the response variable 77 is seedling richness of dispersal modes (native species), and how it varies with soil nitrogen (a), soil phosphorus (b), and LAI (c). Error bars and shaded regions represent ± standard error. * denotes significant relationships, using F-statistics (Table S3.1).

Figure 4.1. Phylogentic diversity in different forest types. (a) Phylogenetic diversity 116 (PD) was compared to null model distrubutions (SES) of understories beneath monoculture, Rainforestation and regenerating selectivly logged forests. Positive values indicate overdispersion whereas negative values indicate clustering, showing higher order fixed effects from LMEM terms. Variation in PD is also shown across variation in soil nitrogen (b), soil phosphorus (c, * denotes significance (P<0.05), using F- statistics) and LAI (d).

Figure 4.2. Understorey phylogenetic and leaf trait diversity beneath monoculture, 117 Rainforestation and regenerating selectively logged forests. Phylogenetic and leaf trait structure were measured against standardised effect sizes (SES) for (a) non-weighted mean nearest taxon phylogenetic distance (MNTD) for all species, (b) non-weighted

5 mean pairwise phylogenetic distance (MPD) for tree and shrub species in isolation, (c) weighted mean pairwise phylogenetic distance (MPD) for tree and shrub species in isolation, (d) non-weighted mean pairwise functional distances (MFD) of SLA, (e) non- weighted MFD of LNC and (f) non-weighted MFD of log transformed LPC. The phylogeny incorporating Bayesian estimates of divergence times and a functional trait dendrogram were used as the basis of the displayed metrics. Positive values indicate overdispersion and negative values indicate clustering compared to the null model expectations that used species richness to randomise the phylogeny and dendrogram. Figure 4.3. Higher order fixed effects from LMEMs for understorey functional 119 diversity (life-form, potential plant height, dispersal type, SLA, LNC and LPC), measured as standard effect sizes of abundance weighted MFD, for all species (a) and for native species in isolation (b), beneath the different forest types. A Gower distance matrix was constructed as it allows for categorical and missing data. Figure 4.4. Higher order fixed effects for mean coefficient of variation (CV) for SLA 121 values of all individuals at the plot level (a), MFD and MNFD weighted by abundances for species with 5 or greater SLA replicates (b, c), between monoculture, Rainforestation and regenerating selectively logged forests seedling communities. * denotes a significant relationship overall (P<0.05). Figure 4.5. Summarizing community assembly processes indicated by analysing 124 evolutionary, leaf trait (SLA, LNC and LPC), discrete trait (potential height, dispersal and life-form) and within-species SLA data, for regenerating selectively logged forest (Regenerating), Rainforestation and monoculture forest types. Species richness (SR) and phylogenetic diversity (PD) were highest within regenerating selectively logged forest and lowest within monoculture forest types, Rainforestation was intermediate. Species within the family Moraceae (green) were common across forest types and tall, wind-dispersed native species (red) were limited to regenerating selectively logged forest and Rainforestation forest types. Species that exhibited high and low variation in SLA (brown gradient) were also restricted to regenerating selectively logged forest seedling communities. Exotic human-dispersed herbs, shrubs and trees (blue) increased all measures of seedling diversity within monoculture forest types. Figure 5.1. Map of study sites and stem maps of the four plots studied. The size of the 146 circles is proportional to DBH size in 2015. Black filled dots represent the trees that were sampled for leaf functional traits. Figure 5.2. Tree height (a) and C: N (b) explained the variation in Periodic Annual 148

Increment (PAI), using trees from all canopy strata. The line represents the linear regression between the two variables. Figure 5.3. Periodic Annual Increment (PAI) versus leaf phosphorus (P), nitrogen (N), 150 specific leaf area (SLA) and carbon to nitrogen ratio (C:N). P was the strongest predictor of PAI for trees occurring within the sub-canopy (a). SLA was positively correlated with PAI in the sub-canopy (c) and N and C:N were positively and negatively correlated with PAI respectively, in both canopy layers (b, c). Figure 5.4. Specific leaf area (SLA) was significantly negatively correlated with tree 152 height for canopy trees (a). The correlations between SLA and nitrogen (N), and phosphorus (b, c) were stronger for sub-canopy trees than for canopy trees. The correlation between N and P was also stronger for sub-canopy trees than for trees occurring within the canopy (d).

List of Tables

Table 2.1. Forest values for conservation and socio-economic services, between 30 Australian WTs bioregion and Philippine tropical forests.

Table 2.2. Importance of animal dispersal for regeneration 40

Table 3.1. Site characteristics; for Rainforestation (Rain.), monocultures (Mono.) and 67 regenerating selectively logged forest (Regen.), including: Leaf Area Index (LAI), soil N and P, elevation, slope and aspect were estimated at each plot. Geology, soil type, density and canopy diversity were obtained from previous studies (Nguyen et al., 2012, Le et al., 2014, Le et al., 2012, Milan et al., 2004). Some information was unavailable (na). Table 4.1. The coefficient of variation (CV) for clades that represent common and 122 obligate groups between monoculture (M), Rainforestation (R) and regenerating selectively logged forest types (S). Plus (+) signs represent a statistical overabundance of those clades within the corresponding community compared to a null model that randomly assigned the same number of species from the same species pool (nodesig statistic) (Webb et al., 2008). Minus (-) signs represent a lack of that species or family within the corresponding forest type. Table 5.1. Site information for the four plots studied, including the QFS experiment 147 name and plot number, the area of the plots and plot environmental variables, and the year of first measurement. Table 5.2. Results from simple bivariate correlations between leaf traits and Basal 148 Area Increment (BAI), Periodic Annual Increment (PAI) and Relative Growth Rate (RGR), and whether the result was expected or unexpected based on theoretical predictions for one or more of the growth measurements. “-“= no significant relationship. Table 5.3. Results from simple bivariate correlations between canopy tree leaf traits 149 and Basal Area Increment (BAI), Periodic Annual Increment (PAI) and Relative Growth Rate (RGR), and whether the result was expected or unexpected based on theoretical predictions for one or more of the growth measurements. “-“= no significant relationship.

Table 5.4. Results from simple bivariate correlations between sub-canopy tree leaf 150 traits and Basal Area Increment (BAI), Periodic Annual Increment (PAI) and Relative Growth Rate (RGR), and whether the result was expected or unexpected based on theoretical predictions for one or more of the growth measurements. “-“= no significant relationship. Table 5.5. The three most important variables derived from the LMEMs for all canopy 151 strata and when restricted to the canopy and sub-canopy trees. * denotes if that variable was retained in the simplest model within Δ AICc = 4, for explaining the variation in Periodic Annual Increment (PAI), Basel Area Increment (BAI) and Relative Growth Rate (RGR).

List of Supplementary Material Table S3.1. Analyses of deviance table for linear mixed effect models examining 86 change in richness and diversity within different life forms depending on forest type, soil phosphorus, soil nitrogen and leaf area index. Random effects were plots nested within sites. Table S3.2. Analyses of deviance table for linear mixed effect models examining 87 change in richness and abundance depending on functional groups, forest type, soil phosphorus, soil nitrogen and leaf area index. Random effects were plots nested within sites. Table S3.3. Proportion of trait types found within each forest type. 88 Figure S3.1. Expected mean species richness for accumulated plots within each forest 88 type (a). Accumulated individuals within each forest type, using “Rarefaction” method (b). Figure S3.2. Soil phosphorus was significantly higher in monoculture forest types (a), 89 soil nitrogen and LAI were not significantly higher in regenerating selectively logged forests (b) and (c). Figure S3.3. Comparison of species richness (a), number of individuals (b), Shannon’s 90 diversity index (c) and Simpson’s diversity index (d), for the different growth forms (a, c and d) between monoculture, Rainforestation and regenerating selectively logged forests types. Error bars represent 95% confidence intervals. Figure S3.4. Higher order fixed effects from LMEM terms where the response 91 variable is seedling species richness of herbs, ferns and graminoids recruited into the different forest types and how it varies with soil phosphorus (c ), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S1. Figure S3.5. Higher order fixed effects from LMEM terms where the response 92 variables are number of exotics (a-d) and native (e-h) seedling species recruited into the different forest types and how exotic and native richness varies depending on soil phosphorus (c and g), soil nitrogen (b and f) and LAI (d and h). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S1. Figure S3.6. Higher order fixed effects from LMEM terms where the response 93 variable is number of individual seedlings recruited into the different forest types and how it varies with soil phosphorus (c), soil nitrogen (b) and LAI (d). Error bars and

10 shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S1. Figure S3.7. Higher order fixed effects from LMEM terms where the response 94 variable is Shannon’s diversity (all growth forms) of seedlings recruited into the different forest types and how it varies with soil phosphorus (c), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S1. Figure S3.8. Higher order fixed effects from LMEM terms where the response 95 variable is Shannon’s diversity of tree and shrub seedling species recruited into the different forest types and how it varies with soil phosphorus (c), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S1. Figure S3.9. Higher order fixed effects from LMEM terms where the response 96 variable is Simpson’s diversity (all growth forms) of seedlings recruited into the different forest types and how it varies with soil phosphorus (c), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S1. Figure S3.10. Higher order fixed effects from LMEM terms where the response 97 variable is Simpson’s diversity of tree and shrub seedling species recruited into the different forest types and how it varies with soil phosphorus (c ), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S1. Figure S3.11. Non-metric multidimensional (nMDS) ordinations (based on Bray- 98 Curtis similarity index) of functional richness for dispersal (a, stress: 0.15), fruit type (c, stress: 0.16), fruit size (e, stress: 0.15) and seed size (g, stress: 0.2). Functional abundance of dispersal (b, stress: 0.12), fruit type (d, stress: 0.17), fruit size (f, stress: 0.16) and seed size (h, stress: 0.22) of all growth forms. Colours represent monoculture forests (light grey), Rainforestation forests (dark grey) and regenerating selectively logged forests (black). Figure S3.12. Higher order fixed effects from LMEM terms where the response 99 variable is richness of dispersal modes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.2.

Figure S3.13. Higher order fixed effects from LMEM terms where the response 100 variable is abundance of dispersal modes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.2. Figure S3.14. Higher order fixed effects from LMEM terms where the response 101 variable is abundance of dispersal modes, when restricted to native species recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.2. Figure S3.15. Higher order fixed effects from LMEM terms where the response 102 variable is richness of fruit types recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and fruit type code corresponds to achene (a), cone (b), syncarp (c), berry (d), capsule (e), drupe (f), follicle (g), legume (h), nutlet (i) and samara (j). Figure S3.16. Higher order fixed effects from LMEM terms where the response 103 variable is abundance of fruit types recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and fruit type code corresponds to achene (a), cone (b), syncarp (c), berry (d), capsule (e), drupe (f), follicle (g), legume (h), nutlet (i) and samara (j). Figure S3.17. Higher order fixed effects from LMEM terms where the response 104 variable is richness of fruit sizes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and fruit size dimensions correspond to 1 = “<2mm x <2mm”, 2 = “2-5mm x 2-5mm”, 3 = “6-15mm x 6-15mm”, 4 = “16-25mm x 16-25mm”, 5 = “26-100mm x 26-100mm” and 6 = “>100mm in any dimension”. Figure S3.18. Higher order fixed effects from LMEM terms where the response 105 variable is abundance of fruit sizes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and fruit size dimensions correspond to 1 = “<2mm x <2mm”, 2 = “2-5mm

12 x 2-5mm”, 3 = “6-15mm x 6-15mm”, 4 = “16-25mm x 16-25mm”, 5 = “26-100mm x 26-100mm” and 6 = “>100mm in any dimension”. Figure S3.19. Higher order fixed effects from LMEM terms where the response 106 variable is richness of seed sizes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and seed size dimensions correspond to 1 = “0-1mm x 0-1mm”, 2 = “1.1- 3mm x 1.1-3mm”, 3 = “4-8mm x 4-8mm”, 4 = “9-12mm x 9-12mm” and 5 = “>13mm in any dimension”. Figure S3.20. Higher order fixed effects from LMEM terms where the response 107 variable is abundance of seed sizes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and seed size dimensions correspond to 1 = “0-1mm x 0-1mm”, 2 = “1.1- 3mm x 1.1-3mm”, 3 = “4-8mm x 4-8mm”, 4 = “9-12mm x 9-12mm” and 5 = “>13mm in any dimension”. Table S4.1. Phylogenetic diversity of seedling communities depending on the higher 135 order fixed effects of forest type, soil nitrogen (N) and phosphorus (P) and leaf area index (LAI). Faith’s phylogenetic diversity (PD), the mean nearest taxon distance (MNTD), and mean pairwise distance (MPD) were used to assess phylogenetic diversity using different species pools and phylogenies that used different estimates of divergence times.

Table S4.2. Functional Diversity of seedling communities depending on forest type, 136 soil nitrogen (N) and phosphorus (P) and leaf area index (LAI). Mean functional distance (MFD) and mean nearest functional distance (MNFD) were used to assess the functional diversity for different traits and combinations, and for different species pools. Table S4.3. Intraspecific variation in SLA for species with five or more measurements, 138 using the mean nearest functional distance (MNFD) and the mean functional distance (MFD). Table S4.4. Phylogenetic signal of categorical traits 138 Figure S4.1. Coefficient of variation (CV) of SLA for species with 5 or more 139 individuals plotted against the phylogeny that included Bayesian estimates of

13 divergence times. Figure S4.2. Phylogeny incorporating Bayesian estimates of divergence times, with 140 the occurrence of each species indicated by circles for monoculture (Mono), Rainforestation (Rain.) and regenerating selectively logged forest (Regen.). Traits including dispersal type (abiotic = large circle and biotic = small circle), potential plant height, mean species SLA and within species variation in SLA are plotted as tip labels, and are scaled to the size of the circle. Green represents species within the Moraceae family, which were favoured across forest types, and red represents tall wind-dispersed trees that were limited within monoculture forest types. Blue represents human- dispersed plant species. Black dots represent species that show high and low variation and mean SLA values, which co-occurred within regenerating selectively logged forest. Table S5.1. Models within Δ AICc = 4 for explaining the variation in Periodic Annual 160 Increment (PAI). Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Random effects were species/exp/plot, representing our sampling design. Table S5.2. Models within Δ AICc = 4 for explaining the variation in Basel Area 161 Increment (BAI). Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Table S5.3. Models within Δ AICc = 4 for explaining the variation in Relative Growth 162 Rate (RGR). Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Table S5.4. Models within Δ AICc = 4 for explaining the variation in Periodic Annual 164 Increment (PAI), for canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Table S5.5. Models within Δ AICc = 4 for explaining the variation in Basel Area 164 Increment (BAI), for canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Table S5.6. Models within Δ AICc = 4 for explaining the variation in Relative Growth 165 Rate (RGR), for canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N.

Table S5.7. Models within Δ AICc = 4 for explaining the variation in Periodic Annual 166 Increment (PAI), for sub-canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Table S5.8. Models within Δ AICc = 4 for explaining the variation in Basel Area 167 Increment (BAI), for sub-canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Table S5.9. Models within Δ AICc = 4 for explaining the variation in Relative Growth 168 Rate (RGR), for sub-canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Figure S5.1. Principle Components Analysis (PCA) of tree leaf traits (a) and structural 169 characteristics (b). Labels for a) correspond to chemical element, except for LA.N: leaf area to nitrogen ratio, C.N: carbon to nitrogen ratio and SLA.ind: Specific Leaf Area (SLA) of the individual tree. PCA a) indicates that SLA, N, P, C and C: N explains the majority of variation in leaf traits. For b) nn: nearest neighbour, DBH: Diameter at Breast Height, CII: Canopy Illumination Index and ht: tree height. PCA b) indicates that CII and ht explain similar amounts of variation. Figure S5.2. Important variables for explaining the variation in BAI (a) and RGR (b 170 and c), using trees from all canopy strata. Height was positively correlated with BAI and RGR, and initial DBH was positively associated with RGR. Figure S5.3. Important variables in explaining the variation in growth measured as 171 Periodic Annual Increment (PAI), Relative Growth Rate (RGR) and Basel Area

Increment (BAI). Tree height and initial DBH (DBH0) were positively associated with growth (a-c), and foliar P concentrations were positively associated with PAI for sub- canopy trees (d).

Chapter 1: Introduction

1.1 Background

Tropical forests provide critical services for human wellbeing, particularly for the world’s most vulnerable people (Scherr et al., 2004). Due to the large decline in tropical forest cover over the last century (Hansen et al., 2013), services such as biodiversity conservation, the provision of clean water, the production of food and timber, and carbon sequestration are all declining. These declines are negatively impacting human wellbeing at both local and global scales. Tropical reforestation is now being recognised as an important tool in addressing the environmental and socio-economic declines associated with forest loss (Lamb et al., 2005).

Different types of reforestation vary in the degree to which they contribute to restoring or enhancing biodiversity and supporting the socio-economic needs of local communities. To date, this has seen the establishment of large areas of plantations using only a small number of timber species that are typically exotic to the locations where they are planted. These species are largely representatives of the genera Acacia, Eucalyptus, Gmelina and Swietenia. These plantations have largely been established with the aim of maximizing productivity, often at the expense of biodiversity and other important forest products. The financial rewards of this method of reforestation have often gone to large foreign-owned companies, contributing little to the livelihoods of local communities. Alternative reforestation methods are increasingly being researched for their benefits. Such methods include mixed-species plantations and other forms of small-scale community-based reforestation. These alternative methods are often aimed at simultaneously increasing biodiversity and socio- economic benefits that can be generated for local communities (Le et al., 2014, Herbohn et al., 2014).

Both the Philippines and the Wet Tropics (hereafter, WTs) bioregion of Australia offer unique biodiversity values, but have also experienced many threats to these values. The Philippines is considered one of three biodiversity hotspots within South East Asia, attributed to the countries exceptional levels of vascular plant and vertebrate endemism (Myers et al., 2000, Sodhi et al., 2010). Similarly, the WTs bioregion of Australia is recognized globally as a mega-diverse region, with the second highest number of endemic genera per unit area in the world (Goosem and Tucker, 2013). Both countries have experienced very high rates of deforestation (Lasco et al., 2001, Goosem and Tucker, 2013). With colonisation came land clearing and the establishment of sedentary agriculture, which has led to declines in productivity and subsequent farmland degradation and

16 abandonment. This sequence of events has resulted in large areas of former farmland now being potentially available for reforestation. In the Philippines, there is estimated to be five million hectares of land that is now covered with exotic cogon (Imperata cylindrica (L.) Beauv.) grassland, and within the WTs bioregion there is approximately 30,000 ha that is considered unsustainable for productive agriculture (Shea, 1992, Garrity et al., 1996). Clearly, there are large areas of land requiring active or passive reforestation in order to address declining biodiversity values and the socio-economic conditions of local communities.

Both the Philippine and Australian governments have recognised that reforestation can help to address many biodiversity and socio-economic issues relating to forest use. The Philippines government has launched the National Greening Program (NGP) that aims to plant 1.5 million hectares of degraded land between 2011 and 2016 (Le et al., 2014). Within the WTs bioregion, several reforestation schemes aimed at generating biodiversity and socio-economic outcomes have been initiated since logging operations in native forest were terminated in 1989 (Vanclay, 2006). Reforestation schemes in both countries have had mixed-success, with many of the failures mostly attributed to a lack of technical knowledge but also managerial short comings.

Knowledge regarding ecological processes operating within forests can help to improve the design and outcomes of reforestation projects. This knowledge includes information about the regeneration and successional dynamics occurring within monoculture plantations, mixed-species plantations and regenerating selectively logged native forests. Improved understanding and management of these dynamics can be used to better integrate targeted ecosystem functions that can improve services including livelihood support into reforestation projects. Ultimately, such better-designed and managed reforestation projects can be of greater community value from both biodiversity conservation and socio-economic perspectives.

1.2 Research problem and Questions

The overarching research problem considered in this thesis is outlined below. Research Problem: Can a better understanding of ecological processes inform practical reforestation and conservation efforts of abandoned agricultural land?

The first and second research questions describe understory species and functional diversity under alternative reforestation methods across a human-dominated landscape on the Island of Leyte in the Philippines. These forests have any extremely close affiliation with local people as means to provide basic provisions like firewood, fruit and fibre. Therefore, the ecology is intrinsically linked to the socio-economic value of these forest communities. Research Question One: How do different reforestation methods influence understory species and functional diversity and composition?

The second research question deepens the understanding of ecological processes operating beneath different reforestation methods by using a wide range of ecological sampling and analytical techniques, including evolutionary relatedness and multiple functional traits. Research Question Two: How do different reforestation methods influence the evolutionary and ecological functioning of understory plant communities, and what does this tell us about community assembly?

The third research question relates simple-to-measure plant functional leaf traits to tree growth within regenerating selectively logged forest in the WTs bioregion of Australia. The purpose of this research question is to test if these simple leaf traits can be used to guide reforestation techniques in diverse tropical forests.

Research Question Three: Can simple leaf traits predict plant growth rates within regenerating selectively logged forest?

Simple leaf traits refer to the morphological traits specific leaf area (SLA) and leaf nutrient contents. These traits are commonly applied to ecological problems and are linked to difficult-to- measure ecosystem processes, such as plant resource acquisition strategies (Funk et al., 2016).

1.3 Justification

My research questions concern the rapid loss of ecological and socio-economic systems as a result of tropical deforestation and degradation, and how these systems can be restored. This investigation is important for four main reasons. Firstly, it draws information from multiple ecological and socio- economic systems within the Philippines and Australian WTs in order to improve reforestation outcomes. Reforestation in both countries is gaining momentum as a means of stemming the rapid loss of biodiversity and also to provide improved livelihood benefits to local communities. Secondly, it is one of the few ecological field-based studies to quantitatively evaluate the multiple reforestation methods currently being implemented within the Philippines and more generally across the tropics. Thirdly, the ecological methods utilised help to inform practitioner concerns with how to quantify and compare biodiversity in different forms of reforestation. The methods can help to refine our understanding of what aspects of biodiversity are of importance from both a conservation and socio-economic perspective. Finally, the thesis also value adds to larger research initiatives and data-sets by providing information that can be used to directly influence on-ground applications.

To address the research questions, I utilise modern ecological sampling and analytical techniques in order to make recommendations regarding how different reforestation methods could be implemented in order to restore key ecological functions and contribute to ecosystem services. By flagging groups of species or ecological interactions that are advantaged or disadvantaged within human-dominated landscapes, and how this will impact on ecological services, I combine applied ecological and forestry theory to understand the broader implications of different reforestation techniques in degraded tropical landscapes.

1.4 Methodology

In order to answer research questions one and two, I undertook plot-based field sampling over a three month period on the Island of Leyte in the Philippines. Leyte Island offers a unique number of small-scale community-based reforestation plantings, which range from simple exotic monocultures to complex native mixed-species plantations, and passive regenerating selectively logged natural forests. The accessibility to the many replications of these contrasting approaches to the restoration of degraded areas and to the previous research conducted on them, as well as the accessibility to

19 experts from Visayas State University (VSU) that have been involved in the local reforestation initiatives makes Leyte Island an ideal study location to examine the proposed research questions. Three months were also spent collecting data on plant functional traits. Discrete traits were collected from literature and online data-bases, and continuous traits were analysed within a laboratory facility at VSU.

For research question three, I sampled study sites located within Danbulla National Park and State Forest, Atherton Tableland, North Queensland, Australia. The plots used in this study form part of a permanent plot network established by the Queensland Department of Forestry as part of their research into sustainable yields from selectively harvested WTs forests (referred to as the QFS permanent plots). The plots were initiated in 1948 and 1969, with demographic rates (growth, mortality and recruitment) recorded at regular intervals since establishment. These plots, initially established for forestry production purposes, offer a unique opportunity to answer contemporary ecological questions. Historical data was matched with field-based sampling to re-establish these plots and to obtain long-term growth data. Trees were identified within the field and functional trait sampling was undertaken, to examine the relationships between leaf traits and growth rates, within regenerating selectively logged natural forest. Further laboratory analysis was conducted at the Commonwealth Scientific and Industry Research Organization (CSIRO) facility in Atherton, North Queensland.

1.5 Limitations

Regarding research questions one and two; despite considerable spatial replication of reforestation types including five sites of each, all located on the same volcanic soil type, and of consistent age (established, circa 1995), elevation and spacing, the sampling was only undertaken over one sampling period. Ultimately, temporal replication is lacking in the investigation of both of these research questions. This is known to influence results due to the masking of natural seasonal variation, which can influence seedling leaf traits, survival and successional dynamics. Temporal replication is also lacking for research question three, which could be masking seasonal changes in

20 growth and leaf trait relationships. Throughout this thesis, conclusions and implications are given in the context of these limitations, which are discussed in more depth in the final chapter.

1.6 Structure and outline of the thesis

My PhD research is presented in six chapters. Chapter 2 provides a review of the key literature. It discusses the socio-economic relationships people have with tropical forests, what approaches to reforestation have been undertaken, and how these approaches can influence biodiversity and sustainable livelihoods. The review then discusses the factors influencing the regeneration and successional dynamics of tropical forests. This includes highlighting of the key influential plant- animal and human interactions. I then discuss how biodiversity can be evaluated through both functional and evolutionary lenses, how this can help to improve our understanding of ecological processes, and how this can lead to improvements in the conservation and sustainable use of tropical forest communities.

Chapter 3 addresses research question 1. In this study, species and trait diversity were measured within the understories of three different forest types on the Island of Leyte: mixed-species plantations, Swietenia macrophylla monocultures plantations, and regenerating selectively logged native forests. All plants less than two meters in height were identified, and information on their dispersal type, fruit type, and seed and fruit size were extracted from literature and online databases. We found that overall seedling richness and diversity were lower within the monoculture forests compared with the regenerating selectively logged forests, with the mixed-species forests showing intermediate seedling diversity, including trait diversity. Monoculture understories had a higher proportion of large-fruited domesticated species that are likely dispersed by people, and a significantly lower proportion of wind-dispersed native seedlings than the other forest types. Higher understorey diversity was generally negatively correlated with soil nutrients and positively correlated with increased leaf area index, i.e. more canopy cover. Our results confirm that mixed- species plantations and regenerating selectively logged forests recruit higher species diversity, but we also found evidence that monocultures can recruit diverse species in the understorey. However, monoculture understories were depauperate of native wind-dispersed traits that are often important emergent native species in tropical rainforests. Overall, our results show that simply having trees in a cleared landscape provides some conservation value, but if monocultures are used as less costly and technically simpler solutions for initiating recruitment, then wind-dispersed native species (e.g.,

21 species from the Dipterocarpaceae family) in addition to species with other limited functional traits (e.g., large-seeded species) should be planted to enhance the long-term term survival of ecologically significant native tree populations.

Chapter 4 deepens our understanding of seedling recruitment by assessing the phylogenetic and functional structure of seedling communities beneath the different forest types, taking into consideration Specific Leaf Area (SLA, including within-species variation), leaf nitrogen and phosphorus, plant life-form, potential plant height and dispersal type. Our analyses based on phylogenetic relatedness (and to a lesser extent, those based on leaf traits of SLA and nutrients) revealed clustering of seedling communities beneath monoculture plantations, and overdispersion of those within regenerating selectively logged forest. Seedling communities beneath monocultures appear to assemble primarily through environmental filtering, especially through dispersal limitation of predictable functional guilds. However, this limitation is frequently overcome by human-assisted dispersal, increasing the diversity of trait combinations (height, life-form and dispersal type). Comparing SLA values (species means and within-species variances) between forest types revealed that regenerating selectively logged forests recruit seedlings with both high and low variation and means, leading to a higher total diversity of leaf traits. Seedling communities beneath regenerating selectively logged forests also showed signs of environmental filtering, but based on within-species variation of SLA rather than on other traits. These understorey conditions appear to have favoured species that show a high intra-specific variation in SLA values, but these conditions also provide habitat for later successional seedlings that show a lower intra-specific variation in SLA. This diversity of traits suggests that limiting similarity may be operating or competitive exclusion may be reduced because of niche differences, allowing species with different traits to co-exist.

Results from chapter 4 suggest that greater niche differentiation and less vacant niche space occur in the seedling communities beneath regenerating selectively logged forest. These analyses also highlighted groups of species that can be targeted for reforestation efforts and contribute to increasing phylogenetic and functional diversity across tropical landscapes. Based on these findings, we recommend that reforestation projects can maximise their conservation value by including species that are phylogenetically and functionally distinct (in terms of the mean and variation in leaf traits), resulting in greater fulfilment of niche space and ecosystem functioning.

Chapter 5 addresses research question 3, by investigating if simple leaf traits can reveal complex ecological processes such as plant growth in regenerating selectively logged natural forest within the WTs bioregion of Australia. This study used a rich historical data-set to quantify tree growth within plots located at Danbulla National Park and State Forest on the Atherton Tableland. Leaf traits were collected from trees that have exhibited fast or slow growth over the last ~50 years of measurement. Leaf traits were found to be poor predictors of tree growth for trees that have entered the canopy; however for sub-canopy trees leaf traits had a stronger association with growth rates. Leaf phosphorus concentrations were the strongest predictor of Periodic Annual Increment (PAI) for trees growing within the sub-canopy. Sub-canopy tree leaves also exhibited stronger trade-offs between leaf traits and adhere to theoretical predictions more so than for canopy trees. From this chapter I suggest that in order for leaf traits to be more applicable to reforestation size-dependence of traits and growth relationships needs to be more carefully considered, particularly when reforestation practitioners assign mean trait values to tropical tree species from multiple canopy strata.

Chapter 6 provides a summary of the study’s overall findings. This includes details of the study’s limitations, with an emphasis on describing the next steps to address these limitations. The chapter concludes by noting questions arising from the results that could guide future research.

References Funk, J. L., Larson, J. E., Ames, G. M., Butterfield, B. J., Cavender-Bares, J., Firn, J., Laughlin, D. C., Sutton-Grier, A. E., Williams, L. & Wright, J. 2016. Revisiting the Holy Grail: using plant functional traits to understand ecological processes. Biological Reviews, DOI: 10.1111/brv.12275. Garrity, D. P., Soekardi, M., Noordwijk, M., Cruz, R., Pathak, P. S., Gunasena, H. P. M., So, N., Huijun, G. & Majid, N. M. 1996. The Imperata grasslands of tropical Asia: area, distribution, and typology. Agroforestry Systems, 36, 3-29. Goosem, S. & Tucker, N. I. J. 2013. Repairing the rainforest (second edition). Wet Tropics Management Authority and Biotropica Australia Pty. Ltd. , Cairns. Hansen, M. C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. A., Tyukavina, A., Thau, D., Stehman, S. V., Goetz, S. J., Loveland, T. R., Kommareddy, A., Egorov, A., Chini, L., Justice, C. O. & Townshend, J. R. G. 2013. High-Resolution Global Maps of 21st-Century Forest Cover Change. Science, 342, 850-853. Herbohn, J. L., Vanclay, J., Ngyuen, H., Le, H. D., Baynes, J., Harrison, S. R., Cedamon, E., Smith, C., Firn, J., Gregorio, N. O., Mangaoang, E. & Lamarre, E. 2014. Inventory Procedures for

Smallholder and Community Woodlots in the Philippines: Methods, Initial Findings and Insights. Small-scale Forestry, 13, 79-100. Lamb, D., Erskine, P. D. & Parrotta, J. A. 2005. Restoration of Degraded Tropical Forest Landscapes. Science, 310, 1628-1632. Lasco, R., Visco, R. & Pulhin, J. 2001. Secondary forests in the Philippines: formation and transformation in the 20th century. Journal of Tropical Forest Science, 13, 652-670. Le, H. D., Smith, C. & Herbohn, J. 2014. What drives the success of reforestation projects in tropical developing countries? The case of the Philippines. Global Environmental Change, 24, 334-348. Myers, N., Mittermeier, R. A., Mittermeier, C. G., Da Fonseca, G. a. B. & Kent, J. 2000. Biodiversity hotspots for conservation priorities Nature, 403. Scherr, S. J., White, A. & Kaimowitz, A. 2004. A new Agenda for Forest Conservation and Poverty Reduction: Making Markets Work for low-income producers. Forest Trends. Washington DC. Shea, G. M. 1992. New timber industry based on valuable cabinetwoods and hardwoods. Queensland Forest Service, Consultancy report for Councils of the Wet Tropics Region, Brisbane Sodhi, N. S., Posa, M. R. C., Lee, T. M., Bickford, D., Koh, L. P. & Brook, B. W. 2010. The state and conservation of southeast asian biodiversity. Biodiversity Conservation, 19, 317-328. Vanclay, J. K. 2006. Can lessons from the Community Rainforest Reforestation Program in eastern Australia be learned? International Forestry Review, 8, 256-264.

Chapter 2: Literature Review

2.1 Tropical forests

Tropical forests are the most diverse and ecologically complex of all terrestrial biomes (Harrison, 2005). They are located on 7% of the world’s land surface, support roughly half of all the biota on earth, and are responsible for around a third of earth’s net primary productivity (Laurance, 1999, Wright et al., 2010b, Schnitzer and Bongers, 2011). An estimated quarter of the world’s most vulnerable people rely on tropical forests for their socio-economic needs (Scherr et al., 2004). Large increases in population densities and a dominant move to sedentary agriculture has meant that tropical forests accounted for 32% of global forest loss with an increase to 2100 km2/year between 2000 to 2012 (Hansen et al., 2013). Some of this land surface remains productive agricultural land, however much of it is considered degraded as ecosystem functions such as soil nutrients and hydrological stability are commonly lost after trees are cleared (Lamb, 2011, Mukul et al., 2016). Approximately 350 million ha of former tropical forest are estimated to now be degraded land (ITTO, 2002). These high rates of deforestation and subsequent land degradation have major implications for ecological and socio-economic systems (Scherr et al., 2004).

Globally, tropical forests sustain a disproportionate amount of biodiversity (Laurance, 1999). For example, the number of tropical tree species can reflect a large proportion of all species within the tropics, due to the large amount of tree-dependent taxa known to occur in tropical forests, e.g. arthropods (Degaard, 2000). It is estimated that there are between 40,000 to 53,000 tree species found within the tropics, with highest numbers occurring within the Indo-Pacific and Neotropical regions (~19,000-25,000 species), (Slik et al., 2015). For comparison, within temperate regions a ten-fold decrease in the number of tree species is estimated (e.g., 124 tree species in Europe and 253 tree species in eastern North America) (Rees et al., 2001, Latham and Ricklefs, 1993). These large latitudinal gradients are common across the web of life and clearly the conservation of tropical habitats can heavily reduce global species extinctions. Further to rises in species extinctions, deforestation in the tropics can negatively influence key ecosystem functions; such as through soil erosion, and changes in fire regimes, soil fertility, carbon storage, and water quality and quantity (Lamb, 2011). This degradation of ecosystem function has implications globally (Hansen et al., 2013, Laurance, 1999). For example, deforestation is estimated to contribute between 12% and 20% of global greenhouse emissions. Haltering deforestation rates and restoring forests are therefore important climate change mitigation strategies (Plugge et al., 2013).

Similarly to the high diversity found within a tropical forest at fine scales, there is also high diversity across larger scales. This larger scale diversity arises due to differences in latitude, altitude, and climatic and soil conditions, resulting in sub-categories of tropical forests (Richards, 1952). One of the main divisions of tropical forests is generally between the Old World tropics- Africa, Asia and Australia, and the New World tropics- Caribbean, Central and South America, which have very few species in common (Richards, 1952). There are however, some families that are emphasised in the major tropical forest regions. For example tropical forest within South East Asia are dominated by species occurring within Dipterocarpaceae family (Richards, 1952). In northern Australia and southeastern Brazil Myrtaceae species are common (Russell-Smith, 1991), in Africa Leguminosae species dominate, and in many New World forests Leguminosae and Lecythidaceae species are common (Richards, 1952, Ter Steege et al., 1993). Lowland tropical forest generally refers to forest growing below 1000 m in altitude, while montane forest or cloud forest refers to forest growing above this. However, this division may co-vary with other factors such as climate (Richards, 1952).

Biodiversity can be defined in a number of different ways; for example, a simple count of the number of species, and measures of abundance and composition of species, communities, populations, genotypes, functional types and landscape units, all of which directly or indirectly impact on ecosystem functions and associated goods and services (Hooper et al., 2005). Many of these ecosystem services are vital for human wellbeing from local to global scales and include:

 Regulation of climate and hydrological systems;

 Provision of resources such as timber, food and pharmaceutical compounds;

 Supporting pollinators and nutrient cycling; and,

 Supporting cultural or spiritual practices (MEA, 2005).

In some parts of the world, policy makers have responded to the increased forest loss and degradation by increasing the amount of conservation land gazetted as National Parks and protected regions, increasing the efficiency and productivity of cleared agricultural land, decreasing the amount of damaging logging techniques and reforesting cleared land in order to increase ecosystem services (Lamb, 2011). Within the Philippines, National Park management has been decentralized to local governments, with the goals of transferring management to rural people who have a greater knowledge of local needs, and to promote more efficient and equitable provisioning of resources. Commonly, community-based conservation initiatives have supported a small number of local elite

26 and further marginalized indigenous communities (Dressler et al., 2006). People living within these novel landscapes have often been displaced from their traditional homelands, and their livelihood and cultural practices have been adversely impacted (Mukul and Herbohn, 2016, Laurance, 1999). Given the broad functional role of tropical forests, the conservation of tropical forest species necessitates a balance with socio-economic functions if either are likely to be maintained in coming decades.

In this literature review, I discuss the intimate relationships people have with tropical forests, in particular focusing on forest biodiversity and sustainable livelihoods within the Philippines and Northern Australia. Then I will discuss the approaches to reforestation and land management have been undertaken within the tropics, and what abiotic and biotic factors affect regeneration and successional dynamics beneath these different forest types by highlighting plant-animal interactions and dispersal abilities important within tropical plant communities, and how these have been influenced by human disturbance. I will then discuss how biodiversity can be assessed through both functional and evolutionary lenses, and how understanding these can improve our understanding of ecological processes including productivity, leading to improvements in the conservation and management of tropical forest communities.

2.2 Dependency of people on tropical forests The livelihood of around a billion people globally are entirely or in part dependent on forests (Scherr et al., 2004). This occurs through a variety of avenues such as subsistence alternative incomes, major sources of cash income, capital assets and employment opportunities (Scherr et al., 2004). Rural communities have suffered from resource exploitation because they are often denied access to the more valuable products, and in many cases are denied legal access to their traditional lands when timber companies are provided with exclusive access (Sunderlin et al., 2005).

Large-scale logging operations, public protected areas and industrial plantations (e.g., monocultures) in most cases do not allow local communities to traditionally manage land that was once available for their sole use, but may not have been officially owned by the community. These forest management regimes typically benefit few and contribute little to rural livelihoods’ (Scherr et al., 2004, Lamb, 2011, Erskine et al., 2006). Colonization has in most instances brought industrialization at a government level, promoted by extraction of the countries’ natural resources and with revenues invested in national economic growth (Sunderlin et al., 2005, Scherr et al., 2004). This has come at the expense of local community livelihoods and biodiversity conservation (Lamb,

2011), while often not assisting to curb illegal logging and systemic corruption, which can exacerbate negative impacts (Scherr et al., 2004).

2.2.1 Situation within the Philippines: forest biodiversity and sustainable livelihoods The Philippines is situated within Tropical East Asia and represents the northern extremity of the Malesia biogeographic floristic zone, which comprises 20% of the world’s flora and 2% of the land area. The Malesia biogeographic region extends from the Malay Peninsula in the west to New Guinea in the east (Myers et al., 2000, Sniderman and Jordan, 2011). A lack of botanical knowledge within the region means diversity estimates are likely inaccurate (Johns, 2008).

There have been 23 biodiversity hotspots identified globally, based on their level of vascular plant endemism, vertebrate endemism and rates of habitat loss (Myers et al., 2000), and there are over a billion people living within these hotspots (Scherr et al., 2004). The Philippines is one of three biodiversity hotspots identified within South East Asia (Sodhi et al., 2010, Myers et al., 2000). The Philippines is estimated to contain 5,832 endemic plant species and 518 endemic vertebrate species across its relatively small land area. These numbers equate to approximately 1.9% of the total numbers of both the plant and vertebrate species found across the globe (Myers et al., 2000).

South East Asia, and in particular the Malesia floristic zone, continues to exhibit some of the highest rates of deforestation (Sodhi et al., 2010). For example, globally, Indonesia experienced the highest rate of deforestation in 2011. From 2000 to 2003 it was estimated at approximately 10,000 km2/year, which increased to a high of 20,000 km2/year in 2011 to 2012 (Hansen et al., 2013). Due to only 10% of the land area within South East Asia being under varying types of protection; predictions of species extinctions of 13% to 42% by 2100 are estimated (Sodhi et al., 2010). Specifically, in the Philippines, deforestation has a long tradition; forest cover was estimated at 92% when the Spanish colonizers arrived in 1521. Before this time the forests served as a vital part of the traditional cultural belief systems held by the indigenous communities. By the time America recognized the independence of the Philippines in 1946, forest cover was estimated to be between 49% to 56% (Maohong, 2012). During the 1900s and 2000s, the Philippines became a large exporter of tropical hardwood into the global market (Maohong, 2012, Lamb, 2011, Kettle, 2010). More recently, the Philippines is recognised as having the highest level of human-dominated land within the Asia-Pacific Region, estimated at 87% of the land area. The country has lost 100% of its ‘frontier forest’ cover; with frontier forest being defined as ‘large, ecologically intact, and relatively undisturbed natural forests’ (Bryant et al., 1997), and it has approximately five million ha of exotic cogon grassland (Imperata cylindrica (L.) Beauv.), which is the largest proportion within the region (Garrity et al., 1996).

As a result of this deforestation, environmental and socio-economic systems in the Philippines are thought by many to have collapsed. For example, it is postulated that deforestation within the Philippines and the associated increase in steep deforested slopes has substantially increased the susceptibility of the land to catastrophic landslides (Maohong, 2012).

Similar to other South East Asian countries, many indigenous communities within the Philippines have transitioned from traditional shifting cultivation to sedentary and intensified crop production (Mukul et al., 2016). This has been driven by local governments and through international pressure, often derived from old colonial beliefs that shifting cultivation was primitive and unsustainable. However, traditional shifting cultivation is recognised as an important cultural practice, which has maintained diverse social and ecological practices for upland indigenous communities. These upland communities have moved to lowland areas, and sedentary agriculture, agroforestry and conservation are now promoted, at the expense of traditional cultural and ecological knowledge (Dressler and Pulhin, 2010). Today, shifting cultivation is still used for subsistence food production.

Given the amount of degraded land now present within the Philippines and the intimate dependency many of the nation’s people have with forests for multiple uses, reforestation can serve as an important management option. The need for reforestation is recognised by the Philippine government with the launch of the National Greening Program that aims to plant 1.5 billion trees from 2011 to 2016. However, choosing the planting method most appropriate to fulfil the desired landscape objective remains a challenge. Methods available range from simple monocultures to full ecological restoration, as well as passive management of secondary or regrowth forest (Lamb et al., 2005, Chazdon, 2008).

2.2.2 Situation within Tropical North Queensland: forest biodiversity and sustainable livelihoods Tropical forests within Australia were once widespread, however, following prolonged periods of aridity it is now limited to moist refuges, predominately within Northern Queensland. This area, known as the Wet Tropics (hereafter, WTs) bioregion extends for 400 km along Australia’s North East coast and is between 20-80 km wide (WTMA, 2014), with an area of 8,340 km2 (Goosem and Tucker, 2013). The region represents only 0.3% of the land surface of Australia but contains a disproportionately large amount of Australia’s biodiversity and is recognized globally as a mega- diverse region (Goosem and Tucker, 2013). For example, it harbours 18% of Australia’s vascular plants and 50% of Australian birds and, notably, the region has the second highest number of

29 endemic genera per unit area in the world (second only to New Caledonia) (Goosem and Tucker, 2013).

European settlement of the WTs bioregion during the 1860s-80s brought the clearing of approximately one-quarter of the vegetation cover, particularly on the lowland plains and in areas with volcanic soils (Wilson et al., 2002, Erskine, 2002). This clearing was primarily driven by timber extraction and agricultural development of arable land (Erskine, 2002). In 1988, around 900,000 ha of forested land in the WTs was declared World Heritage (WTMA, 2014, Goosem and Tucker, 2013). Prior to this, approximately half of this forest was managed for multiple uses, including the production of timber through a selective logging regime (Crome et al., 1992, Horne and Hickey, 1991). The Queensland Forest Service supervised these logging operations within the WTs bioregion. It implemented a sustainable forest management system, primarily through conservative silvicultural practices and the minimisation of road construction associated with timber harvesting and extraction (Vanclay, 1994). With World Heritage protection came the banning of logging within the now protected forests (Vanclay, 1994). This sparked intense debate due to the reduction in local employment associated with the supply of high-value timber species now protected by the World Heritage listing (Erskine, 2005). Today some forest clearing and selective harvesting of tropical forest continues on private land.

The sequence of events of land clearing, conversion to agriculture and then declines in productivity leading to abandonment and degradation often seen in tropical forests globally, is also evident in the WTs bioregion (Laurance, 1999). For example Shea (1992) estimated that 30,000 ha of land that was unsustainable for productive agricultural land could be suitable for reforestation. In order to compensate former timber workers, and the potential market opportunities of providing high-value timber, the Queensland government established a new timber industry, that aimed to restore socio- economic and biodiversity services. Since World Heritage listing in 1988, numerous reforestation projects were established through individual landholders, independent advisory bodies, non- government and government agencies, and these have varied to the degree to which they successfully provided both economic and biodiversity benefits (Catterall and Harrison, 2006). The Community Rainforest Reforestation Program (CRRP) aimed to maintain wood production and employment (post World Heritage listing), but also to provide biodiversity benefits. From 1993- 2000 the CRRP forested 1,782 ha of land within Eastern Australia, at an estimated cost of AUD $10 million. Despite this, the socio-economic and environmental goals were not fulfilled and the program was considered unsuccessful by many, with technical and managerial factors attributed to the failure of the project (Vanclay, 2006). Clearly, the balance between addressing conservation and

30 socio-economic needs is not restricted to developing countries, and parallels are evident between the Philippines and the WTs of Australia (Table XX).

Table 2.1. Forest values for conservation and socio-economic services, between Australian WTs bioregion and Philippine tropical forests.

Forest Conservation Socio-economic Forest events Reforestation location value value (colonisation-land initiative’s clearing-sedentary (conservation and agriculture-declines in socio-economic productivity- aims) abandonment- degradation)

WTs High Medium - high Yes Often failed Australia -very high -logging ban 1988 -more recent (~200y) -due technical and endemism managerial factors -tourism -plant families -cultural with primitive origins

Philippines High High Yes Often failed

-high endemism -many people -longer time period -due technical and directly dependant (~400y) managerial factors -global on forests biodiversity hotspot -cultural (swidden)

-timber resource

The forests of the WTs bioregion have provided livelihoods and were influenced by human habitation for more than 40,000 years. In fact, the region had one of the highest population densities within Australia, pre-European arrival, with indigenous communities within this region representing diverse language groups (WTMA, 2014).

The degree to which hunter-gatherer peoples have managed to sustain themselves purely within tropical forests is debated, primarily due to the perceived lack of a year round carbohydrate rich plant food source (Hill and Baird, 2003). In order to help sustain their large population densities within the WTs forests, a unique abundance of endemic large-seeded tree species were utilized as a carbohydrate resource, particularly during the wet season (Hill and Baird, 2003).

The extent to which indigenous communities within the WTs modified their environment through fire management is heavily debated (Bowman, 1998). In particularly, there is contention over the degree and scale in which tropical forest expansion and contraction was driven through anthropogenic fire management or ‘natural’ environmental variation. Indigenous burning has been widely practiced within Australian ecosystems, but since colonisation indigenous burning practices as well as cultural knowledge have been reduced. Consequently, the lack of fire frequency has been reported as a key factor in explaining vegetation transitions from grassland and savanna, to more closed forest, including tropical forest (Fairfax et al., 2009).

2.3 Tropical reforestation Although clearing of tropical forests is substantial and challenges regarding the conservation of ecological goods and services derived from natural forest remain (Barlow et al., 2007, Barlow et al., 2016), some tropical countries have transitioned in the last decade from net deforestation to net reforestation (Meyfroidt and Lambin, 2011). The rate of reforestation within the tropics from 1990 to 2000; shifting from non-forested to ‘natural closed forest’ (>40% tree cover) was highest within the Pantropical countries (0.35 M ha/year) and lowest in African and Asian countries (0.11 M ha/year) (Meyfroidt and Lambin, 2011). Despite being influenced by relevant criteria and the definition of terms (Brown and Lugo, 1990, Chazdon, 2008), this transition has seen a large increase in research regarding the type, scale and timing of tropical reforestation. For example a literature search within Web of Science using the keywords ‘tropical’ and ‘reforestation’ reveals that from 2006-2016 on average 47 articles relating to these search terms were published per year, compared to 13 from 1996-2006.

Options for tropical reforestation are determined by the degree to which land managers wish to restore or enhance biodiversity, to supply financial and livelihood benefits and the state of forest degradation. Chazdon (2008) conceptualisers reforestation options with the ‘restoration staircase’, which trades off time and cost of reforestation, the desired biodiversity and ecosystem service outcomes and the current state of degradation. These trade-offs have generally favoured the productivity of a limited number of forest products derived from a small number of exotic genera

32 that have been specifically bred and introduced worldwide. For example, Pinus, Acacia, Eucalyptus, Gmelina, Tectona and Swietenia are the most commonly grown genera, and these are commonly grown in monocultures and under relatively short rotations. Governments and large companies have promoted monoculture plantations in order to provide predictable income and consistent products such as veneer, pulpwood, particleboard and sawlogs (Erskine et al., 2006, Lamb, 2011). These homogenous forests have generally come at the expense of biodiversity; however in degraded landscapes exotic monocultures can act as a form of reclamation by stabilising soil and nutrient cycling and facilitating the establishment of a wider range of native forest products and species, providing some remnant forest and ecological systems remain within the surrounding landscape (Lamb et al., 2005, Parrotta, 1992, Lamb, 1998, Potvin and Gotelli, 2008, Ashton et al., 2001).

In contrast, full ecological restoration aimed at re-establishing the original ecosystem, generally trades-off increases in biodiversity at the expense of large volumes of commercially viable goods. The ecological services full restoration provides has a limited market and does not attract the financial incentives at a temporal scale relevant for commercial operations (Lamb et al., 2005).

Monoculture plantations and full ecological restoration generally do not account for the diversity of landscape objectives that tropical forests provide. An alternative is to rehabilitate degraded tropical landscapes by establishing plantations using commercially viable exotic or native species, which can potentially provide a greater diversity of ecological services including increasing biodiversity and socio-economic values (Lamb et al., 2005, Erskine et al., 2006).

2.3.1 Monocultures A valuable timber species that is often grown as monoculture plantations in many tropical regions is Swietenia macrophylla King (Family: Meliaceae, Common name: Mahogany). This species is endemic to large parts of South America, stretching from the Gulf Coast of Mexico to Amazonian Bolivia where it is found naturally at low densities, is considered naturally depleted and is relatively difficult to establish in its native range (Grogan et al., 2014). It is considered a shade-tolerant but long-lived pioneer and has been planted on approximately 100,000 ha of tropical land as a forestry resource (Wadsworth and González, 2008, Lamb, 2011). It is a popular species for monocultures because it has a fast growth rate and beneficial wood properties such as a high stability and a natural resistance to dry-wood termites (Wadsworth and González, 2008). Within the Philippines, Mahogany is commonly established as small-scale (~1ha) community-based plantations, on a range of sites (Le et al., 2014).

Monoculture plantations can be used to kick start understory development and increase biodiversity and longer-term sustainability. However, this depends on a number of management strategies including whether harvesting users a clear-cut or a more selective approach, as well as the frequency and type of other stand management practices, and the connectivity of plantings with other forests fragments, particularly remnant forest (Farwig et al., 2009).

2.3.2 Mixed-species plantations Mixed-species plantations offer a means to enhance biodiversity, and at the same time increase the number of forest products derived from a plantation, over a longer time period. Due to these benefits, the mixing of species within plantations has occurred as a reforestation tool across the tropics, although typically at a small-scale (e.g., Australia (Kanowski et al., 2003), China (Forrester and Tang, 2016) and Brazil (Montagnini et al., 1995)).

An example of using mixed-species plantations to reforest degraded tropical landscapes was developed on the Island of Leyte, in the Philippines, in 1992 and this method has become more widely known as ‘Rainforestation Farming’ (hereafter Rainforestation). The premise behind this approach is ‘the closer a farming system in the humid tropics is to a natural rainforest ecosystem, the more sustainable it is’ (Sales-Come, 2010). This approach to reforestation aimed to replace traditional forms of shifting agriculture known as ‘kaingin’, to form buffer zones around natural vegetation, to maintain hydrological cycles, to conserve/enhance biodiversity, and to improve the livelihoods of local farmers (Milan et al., 2004).

Rainforestation used mostly native tree species that were collected from nearby mother trees and cultivated in local nurseries, as well as some exotic species, fruit trees, and other local and exotic species that produce non-timber forest products (NTFPs). Approximately 100 native plant species were used (Nguyen et al., 2012). Primary species were initially planted to facilitate the survival of shade tolerate apex species that were planted later (Milan et al., 2004). In line with ecological restoration, planting densities were high, estimated at 5,000 trees per ha (Nguyen et al., 2012). The use of early successional species aimed to provide an early cash flow into local communities due to the harvesting of timber and/or fruit. Initial research screened suitable native species based on economic quality and basic successional status. Pilot plantings were undertaken at 28 locations over a six year period and were approximately 1ha in size.

In a social context, the Rainforestation and the mixed-species approach in general is often targeted at engaging community members at all stages of the forest development process, including seedling choice and propagation, and plantation location, design and implementation. Within tropical North Queensland mixed-species plantations of high value cabinet timber species were established on private land in order to provide a financial asset as well as for biodiversity benefits (Vanclay, 2006, Kanowski et al., 2003).

The age, locations and replicates of the Rainforestation plantations make them ideal study sites to examine a large number of ecological, economic and social factors that can be assessed against the original project aims (Nguyen et al., 2012, Kanowski et al., 2003, Forrester et al., 2011, Vanclay, 2006). Using these plantations, Nguyen et al. (2014) suggest that wood density values for native tree species can aid in the design of mixed-species plantations for both environmental and socio- economic outcomes.

2.3.3 Human-modified tropical landscapes Human-modified landscapes within tropical regions are increasing and are an important landscape type for both biodiversity conservation and socio-economic outcomes (Gardner et al., 2009, Brown and Lugo, 1990). As little as 10% of tropical forests are allocated within protected areas (Schmitt et al., 2008), with existing protected areas often heavily influenced by human activities. Barlow et al. (2016) have suggested that in addition to deforestation, Amazonian conservation initiatives should consider disturbance to primary forests, as this disturbance has led to dramatic declines in conservation values of the existing primary forest.

In general, the term ‘secondary forests’ refers to forest that was at some stage in its development influenced by human activity (Brown and Lugo, 1990); however, the term has a wide array of meanings in terms of the extent, type and frequency of these activities (McNamara et al., 2012). Common human activities that lead to secondary forests include shifting or sedentary agriculture (Mukul and Herbohn, 2016), fuel wood collection, grazing, burning or selective logging, all of which differ in the type of secondary forest they produce (Brown and Lugo, 1990). In many tropical countries, the area of secondary forests far exceeds the area of primary forest (Brown and Lugo, 1990). For example, secondary forests and brush-lands in the Philippines are estimated at five million ha and primary forest is estimated at 2.9 million ha (Lasco et al., 2010). In North Queensland, Australia, approximately half of the World Heritage listed tropical forest (~450,000

35 ha) was previously used for timber extraction through a selective logging regime (Crome et al., 1992, Horne and Hickey, 1991), and can therefore also be considered secondary forest.

Consequently selectively-logged forests are a common secondary forest type within the Philippines and tropical Australia. The applied logging regimes have varied in terms of the amount of time the forest has been subject to logging and the specific diameter limits for tree selection (Ashton and Peters, 1999). Immediate impacts of selective logging operations within rainforests include the opening of canopy gaps, an overall reduction in canopy cover, soil surface disturbance and subsequent changes in microclimate (Medjibe et al., 2011, Reich et al., 2001). The influence of longer-term effects of selective logging on measures of biodiversity were reviewed in two meta- analyses conducted by Clark and Covey (2012) and Putz et al. (2012). Logging activities were found to significantly reduce tree species richness relative to natural forests but 85-100% of mammal, bird, invertebrate and plant species are retained post logging. Comparisons of species biodiversity measures, including pure species richness between logged and unlogged forests have yielded variable results partly attributed to the spatial scale used, the requirements and resilience of the study species and the method of evaluation (Putz et al., 2012, Clark and Covey, 2012, Imai et al., 2012, Laufer et al., 2013). Some studies have illustrated an increase in the diversity and abundance of early successional plant species or species that were not harvested and were competing with the large targeted species, and that this increased measures of biodiversity relative to reference sites (Hector et al., 2011, Horne and Hickey, 1991, Verburg and Eijk-Bos, 2003). These findings are supported in the results of faunal studies that often find an idiosyncratic response of biodiversity measures to logging disturbances (Laufer et al., 2013, Baraloto et al., 2012b). However, differences in species composition irrespective of diversity metrics have large consequences for ecosystem functions and associated services (Baraloto et al., 2012b).

A widely reported and longer-term influence of selective logging is the alteration of size class distributions, namely the reduction in large-girthed emergent tree species (e.g., wind-dispersed species such as emergent Dipterocarps, Flindersia spp. etc.) (Horne and Hickey, 1991, Hector et al., 2011, Ashton and Peters, 1999). The loss of emergent trees from the canopy has ramifications for ecosystem structure and function, including a reduction in the amount of above-ground biomass and subsequent carbon stocks (Hector et al., 2011, Putz et al., 2012, Edwards et al., 2014a). These reductions in biomass can, however, be remediated over time through the increase in growth rates of recruiting individuals, but this is highly dependent on the method of selective logging used

(Medjibe et al., 2011, Edwards et al., 2014a, Putz et al., 2008). A longer-lasting influence of selective logging is a change in the relative composition of species and functional types within the canopy and emergent layers both temporally through successional change and spatially through the natural variability of species spacing within rainforests (Horne and Hickey, 1991, Verburg and Eijk- Bos, 2003).

A key question, relating to the sustainability of tropical forest logging and the methods used is how disturbances caused by selective logging vary to that of other disturbance regimes found within tropical forests? Specifically within the Philippines, shifting cultivation (known locally as ‘kaingin’) is a major land-use type and has been over a longer time period than selective logging (Mukul et al., 2016). In addition, forests of the Philippines are regularly impacted by typhoons (e.g., Super Typhoon ‘Yolanda’ in 2013). Within tropical Australia, the WTs bioregion is subject to cyclones (Webb, 1958). Forests within both of these regions have a long history of recovering following large-scale disturbances, and may be more adapted to a selective logging regime than other forested regions. However, it is suggested that a ‘sustainable’ selective logging regime within the WTs bioregion of Australia for example, is not likely to be viable in the current economic and political environment (Lamb, 2016).

The consequences of the proliferation of regenerating selectively logged forests, particularly within the Philippines and tropical North Queensland, means that ‘pristine’ forest ecosystems are often unavailable for ecological and conservation comparative studies. However, differences in forestry practices between Australian and the Philippines are vast. Specifically, within Australia, the monitoring of forests after selective logging and silvicultural treatments has a long history, dating back to the early 1900s. Due to this a rich historic dataset is available to understand tropical Australian forest growth and regeneration dynamics over the last 100 years. Comparatively, within the Philippines, logging operations have had a much longer history, through different colonization events and as such forestry practices were more ad hoc, and forests continue to be an integral part many people’s livelihoods; therefore the impacts of selective logging are ongoing.

Examining growth and regeneration dynamics of biodiversity within these different reforestation types would provide a greater understanding of their role in conservation and the provision of ecological services to local communities (McNamara et al., 2012).

2.4 Community assembly underneath tropical forests: Regeneration and successional dynamics Plant regeneration and community assembly are central paradigms within forest ecology. Within tropical forests the classical example of these processes occurs following disturbance, with the most common and well-studied being that of tree fall gaps (Attiwill, 1994, Rees et al., 2001). A notable consequence of a tree fall is an increase in the amount of direct sunlight that can reach the forest understory. Consequently, a central axis of variation for tropical plant species is how they can survive and reproduce across light gradients (Valladares and Niinemets, 2008). A convenient concept for this variation has a long history in plant ecology; whereby ‘pioneer’, shade-intolerant or early-successional species that require direct sunlight for germination, growth and reproduction establish first, and subsequently allow for the germination, growth and reproduction of ‘climax’, shade-tolerant or later-successional species (Swaine and Whitmore, 1988). It has emerged that these are not discrete categorisations, but are influenced by ontogenetic stages, such as seedlings or saplings, and different light response strategies also occur within the groups. Trade-offs are also thought to occur between biotic (e.g. facilitative or competitive interactions) and abiotic (e.g. topography, geology and climate) influences, both within and among species (Bloor and Grubb, 2004, Gibert et al., 2016, Rees et al., 2001).

More broadly, the spectrums that underpin community assembly within tropical forests, such as species’ divergent responses to changes in light gradients (i.e. ‘pioneer’ versus ‘climax’ species), are thought to occur along a continuum (Gibert et al., 2016). Higher resource environments assist species to invest more in initial growth, at the expense of longevity, whereas species occurring in lower resource environments trade-off high initial growth for survival and longevity (Poorter and Bongers, 2006, Wright et al., 2004, Chave et al., 2009). These differences are thought to increase niche differentiation and microhabitat specialization through both space and time, allowing greater coexistence of highly diverse tropical plant species. Within gap-phase and successional dynamics, ‘climax’ or late-successional species often exist as a seed, seedling or sapling bank when the gap is created and represent a ‘storage effect’ that is often used to explain tropical plant species coexistence (Rees et al., 2001, Usinowicz et al., 2012). Further, maintenance of high diversity tropical systems is quantitatively supported by negative density-dependence operating at the seedling recruitment level, meaning that seedlings of a species surrounded by seedlings of the same species experience lower survival than if they were surrounded by different species (i.e. rare-species advantage) (Harms et al., 2000, Connell, 1971, Janzen, 1970, Hautier et al., 2010). These ‘stabilizing’ mechanisms tend to strengthen negative intraspecific interactions compared with interspecific interactions, promoting coexistence (Chesson, 2000).

Given the high number of tropical tree species, it is difficult to assume that complete differences in their niche and life history occur for all species, with at least some showing functional overlap or even redundancy. For these species, ecological drift may play a large role in governing their coexistence within a community (Hubbell, 2001). Lastly, historical inertia, regional speciation and extinction, dispersal, and biogeographic affiliations also play a role in community assembly within tropical forests (Ricklefs, 1987, Weiher et al., 2011). The influence of the niche, ecological drift and historical factors on community assembly within tropical forests at varying spatial and temporal scales, is currently a research focus (Ricklefs, 2008). Knowledge of these ecological processes and differences in niches is increasingly being used to guide management and reforestation in tropical regions e.g., Verdú et al. (2012) and Schweizer et al. (2015).

Understanding the regeneration dynamics of a forest is also essential when determining the long- term sustainability of a management regime or reforestation project. This long-term sustainability can influence the ability of a forest to provide additional ecosystem services under a range of landscape objectives (White, 2004). Regeneration can occur through the germination and persistence of seedlings, supplementary/enrichment planting or via asexual reproduction (suckering or re-sprouting) (Mwavu and Witkowski, 2009, Lamb, 1998). Within tropical forest regions, the establishment of seedlings is often hindered by the ability of mostly exotic grasses to out-compete native species or increase the frequency of fires making the conditions unsuitable for the germination of seedlings (Farwig et al., 2009). Other factors such as changes in seed predation, herbivory, frugivory, microclimate, fire (absence or presence) and seed influxes are also known to reduce forest regeneration (Zanne and Chapman, 2001). The reforestation of degraded lands via human intervention can mimic understory conditions found within natural systems and therefore release this arrested stage of succession, although this most likely apparent following canopy closure (Farwig et al., 2009, Zanne and Chapman, 2001). Whether the reforested site is aiming for ecological restoration or serving as an economic resource, the ability to regenerate is of importance to land managers and to the long-term sustainability of the contrived habitat.

Recruitment within plantations can provide benefits such as mitigating top soil erosion, more efficient nutrient cycling, provision of a greater number of forest products, and increased habitat for faunal communities, which in turn contributes to sustainable forest management (Firn et al., 2007, Lamb, 1998, Lindenmayer and Hobbs, 2004, Simonetti et al., 2013). Negative consequences of recruitment include reductions in desired species growth, and disrupting forest management practices, including harvesting (Lindenmayer and Hobbs, 2004, Lamb, 1998). Forest types, ages and locations can drive differences in recruitment dynamics. Research into these differences is lacking. Such research can inform management strategies to improve recruitment of desirable

39 species including combinations of species or species with sought-after traits and this improved recruitment can lead to enhanced ecosystem services (Mayfield et al., 2006).

Studies investigating regeneration and successional dynamics in the tropics are generally located within the Neotropics, with far fewer studies being conducted in South East Asia, Africa and Australia. In a relevant study located in the WTs bioregion, Firn et al. (2007) found a negative relationship between canopy diversity and productivity, but also suggest that monoculture plantations, if not harvested, can recruit diverse understories and restore some ecosystem functions. However, the capacity of a monoculture to achieve these multiple-use outcomes depends on the characteristics of the species planted, including whether it is endemic to a region, its growth rates, its functional traits and its ability to provide suitable habitat for a range of animals.

2.4.1 Dispersal limitations to recruitment A species’ ability to disperse across a landscape is critical in evaluating how the species will respond to habitat fragmentation, biodiversity loss and climate change impacts (Booth and Williams, 2012, Corlett, 2011a, Corlett, 2011b, Markl et al., 2012). Temporal scales are an important consideration when evaluating a species dispersal distances; at an evolutionary time- scale, rare dispersal events can create population founder events, but at a reforestation time-scale (e.g. multi-decadal), common dispersal distances are more appropriate (Corlett, 2009, Booth and Williams, 2012). ‘Recruitment foci’ points can occur around older isolated individuals that have dispersed an uncommonly large distance perhaps leading to establishment of a new forest in which a species assemblage is filtered to only include species with a high dispersal ability (White, 2004). Therefore, seed dispersal distances have a major influence on plant species persistence within an anthropogenic landscape and also in regeneration of degraded sites (Corlett, 2009). Sampling seedlings to infer average dispersal distances can be done in rare events (large distance dispersal), however, the Janzen-Connell Hypothesis effect on seedlings is likely to underestimate seed dispersed especially at high densities, close to mother trees (Corlett, 2009).

Results in tropical forests to date, on the relative importance of biotic dispersal compared to abiotic dispersal (wind, water and gravity) are equivocal. Biotic dispersal is a common strategy within tropical forests, with estimates of between 35-90% (Table XX) (Corlett, 1998). Generally, when compared to canopy or emergent trees, sub-canopy trees show a greater proportion of biotic- dispersed species relative to abiotic dispersed species. Lianas and canopy tree species are more

40 likely to be dispersed abiotically. Gradients regarding the relative proportion of biotic versus abiotic dispersal are apparent; with evidence suggesting that wetter habitats have more fleshy-fruited biotic dispersed species, and dryer and higher-elevated forests have more abiotic dispersed species, presumably due to increases in water availability and wind strength and frequency (Howe and Smallwood, 1982, Butler et al., 2007).

Table 2.2. Importance of animal dispersal for regeneration

Study Animal dispersal

Zanne and Chapman (2001) 35-46%

Howe and Smallwood (1982) >50-75%

Corlett (2011c) 34.4-54.4% (tree species: 69.8%)

Markl et al. (2012) 90% (woody plant species)

Within tropical East Asia, abiotic dispersal via wind, water and gravity is also common (Corlett, 2011c, Corlett, 2009). Explosive release occurs in some species of trees, however this dispersal mode is often followed by secondary dispersal by ants (Corlett, 2011a). The relative proportion of seed predation to seed dispersal is unknown for the majority of species. Dispersal using the movement of water is also known to occur within some riparian species and more commonly within herbaceous plants within landscapes dominated by abandoned rice paddies (Corlett, 2011b). Wind dispersal is common in lowland tropical forests, which are some of the most diverse forests. This dispersal mechanism mostly occurs in exposed microhabitats such as forest canopies. Ground orchids often exhibit wind dispersal of small-seeds (Corlett, 2011c). The family Dipterocarpaceae and other genera within families that use dispersal via wind usually have larger sized winged seeds and a routine dispersal maximum of 100 m. Strong winds, such as during a typhoon event, are circumstantially reported to disperse seeds of the Dipterocarpaceae family for kilometers (Corlett, 2009). This has implications at a broader spatial and temporal scale, such as for the establishment of new plant populations (e.g., population founder events); however, for the purpose of assessing the family’s dispersal potential across a human-modified landscape, the common maximum dispersal distance of 100 m is more appropriate.

Scatter-hoarding rodents disperse often-wingless fruits in usually winged genera and can facilitate secondary dispersal in winged species (Corlett, 2009). The role of rodents in disturbed environments is generally understudied (Beckman and Rogers, 2013). Common dispersal distances of between 10-100 m are reported and rodents are known to be the main dispersal agents for the family Fagaceae (Corlett, 2009). Corlett (2011b) has proposed that small rodent species may also play an important role in dispersing small-seeded shrubs across landscapes dominated by grassland.

Macaques, which are a common primate within tropical Asia, can disperse seeds via ingestion or simple transportation while the flesh is being swallowed within the mouth and the seed discarded (Corlett and Lucas, 1990). Small-seeded species or species with seeds that is difficult to separate from the flesh of the fruit are usually ingested. These seeds have a common maximum dispersal distance of 100 m to 1 km. Species with a larger seed (>4mm) are dispersed smaller distances of 10-100 m (Corlett, 2009).

Fruit bats are a common dispersal agent within tropical Asia and show a similar dispersal pattern as Macaques. Large seeds that cannot be swallowed are taken to nearby roosting trees usually within 80 m of the fruiting tree and dropped (Corlett, 1998). Smaller seeds that can be swallowed (< 2-3 mm) are transported greater distances of 1-10 km (Corlett, 2009). Canopy bats also have larger daily flight distances than sub-canopy species, thereby increasing their common seed dispersal distances (Corlett, 1998). In remnant forests of Hong Kong, bats were found to eat a variety of native and cultivated fruits, however, the majority of endemic species eaten were found within the genus Ficus (Corlett, 2011a).

Gape width of bird species generally increases with body mass and gape width dictates the size of seed that a bird species can disperse seed (Galetti et al., 2013). Compositions of frugivorous passerine birds have impacts on the composition of plant species in tropical forests (Galetti et al., 2013, Corlett, 2009). Using gut passage times and daily home range distances, Corlett (2009) estimated most bird species would routinely disperse seeds between 100-1000 m. The maximum fruit size that bird species are able to swallow ranges from <8 mm to >20 mm and depends on their gape size (Corlett, 2011b). Bulbuls (Pycnonotus spp.), White-eyes (Zosterops spp.) and Starlings (Lamprotornis spp.) are wide spread bird species within tropical Asia and accounted for the majority of seed dispersal across different habitats within Hong Kong and possibly within other parts of the region. Canopy birds such as Hornbills, fruit pigeons (Dacula spp., Ptilinopus spp.) and the Hill Myna (Gracula religiose) are known to have large home ranges (>1km) and long gut passage times. Common dispersal distance therefore increases in some cases to well over 1 km (Corlett, 2009). Seasonal latitudinal or altitudinal migrations have been reported in larger tropical

42 bird species. This has large implications for plant community assembly between different habitats such as between montane and lowland, and between biogeographically distinct islands (Beckman and Rogers, 2013).

Isolation from seed sources such as primary or secondary forests is shown to decrease the abundance and diversity of natural recruitment (White, 2004, Lamb, 1998). What species or dispersal vectors are favored or limited across a landscape can have large consequences for forest regeneration and successional dynamics, and ultimately the provision of ecosystem services. This is because many of the recruited seedlings will later enter into the canopy, fulfilling their ecological and possibly a socio-economic role.

2.5 Human impacts on plant dispersal (subsistence hunting and habitat fragmentation) Subsistence hunting is a common anthropogenic disturbance within tropical Asia and often targets large-bodied species particularly those above 1 kg (Peres, 2000). These large-bodied species tend to have a large gape and also require large amounts of intact habitat making them sensitive to habitat fragmentation (Galetti et al., 2013). Often small remnant patches of vegetation are more prone to increased rates of subsistence hunting (Peres, 2000), thereby further reducing large frugivorous bird populations. This reduces large-seeded tree species’ ability to disperse in terms of both the amount of seed removed and the distance the seed travels (Galetti et al., 2013, Markl et al., 2012). Small bird species are known to be more resilient to large-scale disturbances and are more often habitat generalists. In a recent study, a palm species in Brazil responded to a decrease in large frugivorous bird populations by producing smaller seeds (Galetti et al., 2013). This evolutionary pressure to smaller seed sizes is postulated as potentially creating negative cascading effects on ecosystem functioning, eventually resulting in reduced reproductive fitness of the palm population. More broadly, reductions in dispersal agents able to disperse large seeds over long distances can lead to lower seedling recruitment, local plant extirpation and ultimately lower plant diversity across a landscape (McConkey et al., 2012, Beckman and Rogers, 2013).

2.6 Functional trait approaches to understanding ecological processes The high species richness of tropical forests means that grouping species or individuals based on a commonality of form or function can help address questions regarding regeneration, successional dynamics and productivity. Traditional community ecology has focused more on taxonomic diversity and nomenclature, leading to a loss of generality and predictability between ecological

43 systems (McGill et al., 2006). Given this, the field of functional trait ecology has gained impetus for questions regarding the mechanisms behind ecological processes (Westoby and Wright, 2006).

In addition to taxonomic diversity, many studies have assigned species of flora to broad functional groups (e.g. ‘pioneer’ or ‘climax’ species). This approach has provided information regarding species’ richness, abundance and composition, giving ecologists coarse insights into how assemblages of these groups respond to environmental gradients. Despite being a convenient method to deal with the great diversity of flora found within tropical forests; using multiple continuous traits is possibly a better way to relate plant distributions and performance to abiotic and biotic gradients (Kraft and Ackerly, 2010, McGill et al., 2006). Functional traits can quantify a species’ or an individual’s effect on ecosystem processes and/or their response to an abiotic and biotic environment (Lavorel and Garnier, 2002). Multiple continuous traits also account for within- species or functional group variation associated with phenotypic plasticity, which is inherent within plant ecological communities (Richards et al., 2006, Bloor and Grubb, 2004).

Following the surge in studies using continuous plant functional traits, a large amount of data has accumulated that uses species-specific mean trait values derived from multiple within species measurements; e.g., the TRY database which has compiled 5 million trait records for 100,000 plant species (Garamszegi and Møller, 2010, Violle et al., 2012, Kattge et al., 2011). These values are assumed to be biologically meaningful, despite the fact that many of these traits show large amounts of within-species variation derived from measurement error, genetic and ontogenetic differences, and variation arising from environmental heterogeneity. Several authors now suggest that incorporating intraspecific variation in meaningful functional traits can improve the detection of community assembly mechanisms and the quantification of functional diversity (Jung et al., 2010, Albert et al., 2012). Additionally, traits directly quantifying a species’ function within an ecosystem might only be accounting for a subset of the species’ total influence on ecosystem function and ignoring species evolutionary relatedness (i.e. phylogeny).

2.6.1 Which functional traits to use and why? Insights into how ecosystem processes are operating within natural systems or in response to human activities have been derived from grouping species based on how they function, and this has had a long history within plant ecology (Grime, 1979). However, in the late 1990s a surge of research was conducted that aimed to identify how functional traits reflect and predict changes in environment, and how species traits can influence ecological processes (Lavorel and Garnier, 2002, Díaz et al., 2003). Since then the use of functional traits has become more nuanced and targeted, with specific

44 experimental questions now being answered with the use of functional traits that represent multiple dimensions of several important ecosystem functions (Lasky et al., 2014, Funk et al., 2016)

For questions regarding regeneration dynamics in tropical forests several traits have been used to gain insights into ecological processes. Dispersal and fruit type are used in many studies investigating regeneration dynamics (Katovai et al., 2012, White, 2004, Wills et al., 2016). Dispersal and fruit type have large consequences on the distance, direction and final location of seed and consequently seedlings, thereby influencing community assembly (Howe and Smallwood, 1982, Pérez-Harguindeguy et al., 2013). The level of classification of dispersal and fruit type depends largely on the ecological question and availability of data (Pérez-Harguindeguy et al., 2013). Fruit and seed size also have large consequences for regeneration dynamics. Fruit size relates to dispersal type and can determine the ‘reward’ a biotic dispersal vector can obtain from a species. Seed size can determine how a seedling will persist within an environment, with larger-seeded species often better able to tolerate shaded environments and herbivory. These species tend to be later successional ‘climax’ species. However small-seeds require less reproductive output per individual, and can be beneficial for seed bank persistence and mass reproductive events (Pérez- Harguindeguy et al., 2013).

Specific leaf area (SLA) is defined as the one-sided area of the newest fully expanded leaf, divided by its dry mass. Leaf nitrogen and phosphorus concentrations (LNC and LPC) can be defined as a percent of dry leaf mass. SLA, LNC and LPC have been shown to reflect growth rates, competitive ability and successional dynamics within tropical forests (Lasky et al., 2014).

A central premise behind trait-based ecology is that cheap, simple to measure morphological traits or ‘soft’ traits reflect physiological functions and that these in turn provide insights into ecosystem processes (Funk et al., 2016, Garnier et al., 2001). For example, SLA and LNC have been able to predict the photosynthetic and respiration rates of a diverse set of plant species, across multiple biomes (Reich et al., 1997). Formalising these large-scale trade-offs,Wright et al. (2004) proposed a worldwide leaf economic spectrum (LES), which was developed on the idea of economy of resources (Bloom et al., 1985) and places plants on a continuum from slow to fast returns on investments, and being either resource acquisitionists or conservationists (Reich et al., 1997, Wright et al., 2004). At either end of the spectrum a plant’s leaves tend to have a suite of characteristics that are predictable across different phylogenies, growth forms and at varying scales (Reich, 2014, Shipley et al., 2006). At one end of the spectrum, resource acquisitionist plants, invest less in carbon when growing leaves, have shorter leaf longevity with quicker returns on resources, invest more in LNC and LPC and have a higher leaf turnover. At the opposite end, resource

45 conservationists, leaf characteristics’ are reversed with higher investments in carbon when growing leaves, longer leaf longevity with slower returns on resources, lower investments in LNC and LPC and less leaf turnover. Another related hypothesis used to generalise plant growth and function is known as the leaf-height-seed (LHS) strategy (Westoby, 1998), which describes similar fundamental trade-offs between the cost of investments and returns on investments associated with height at maturity, investment in leaf tissue and provisioning of seed resources.

More recently the LES and LHS have been used as theoretical frameworks to develop part of a broader whole plant economic spectrum that incorporates characteristics associated with other plant tissue and functions (Poorter et al., 2014, Edwards et al., 2014b, Reich, 2014, Kraft et al., 2008). Broadly, these spectrums represent a growth versus survival trade-off, with set constraints on how plants function. A more detailed emergence of different plant strategies, and insights into processes governing community assembly and how forest ecosystems function have all supported a trait-based approach to plant ecology (Reich et al., 1997, Edwards et al., 2014b, Kraft et al., 2015b).

An ability to gain insights into demographic rates such as, growth, mortality and recruitment, from using simple and cheap to obtain functional traits has been seen as a research priority, particularly for the many species found within tropical forests. To date, evidence suggests that leaf traits can predict growth rates for many tree seedling species, but that this relationship weakens as the seedlings develop and other factors become a stronger determinant of growth (Poorter et al., 2008, Sterck and Bongers, 2001). However, how these growth and leaf trait relationships vary depending on light availability and ontogenetic stages within regenerating selectively logged forest has not been sufficiently determined in order to draw conclusions about the applications to reforestation outcomes.

2.7 Examining evolution and ecological traits together; advantages to understanding community assembly. Species within a community share an evolutionary history, through adaptations to historic environments that influence the modern ecology and traits of the species within a community. Phylogenetic comparative methods (PCMs) were developed to account for this non-independence, derived from shared ancestry. These methods use evolutionary relatedness as a proxy for similarities or differences in ecological function (Webb, 2000, Verdú et al., 2012). Importantly, PCMs can incorporate potentially difficult to measure traits and traits unidentified as being important determinants of plant community assembly (Swenson and Enquist, 2009, Kraft and Ackerly, 2010, Baraloto et al., 2012a). Recently, the number of studies using PCMs has

46 significantly increased and subsequently large advancements in methods and tools have allowed more researchers to incorporate this aspect of biodiversity. For example, a literature search within Web of Science using the keywords ‘phylogenetic’ and ‘ecology’ and ‘forest’ demonstrated that from 2006-2016 on average 67 articles were published per year, compared to 13 per year from 1996-2006, with a clear increasing number within more recent years.

The linking of evolutionary processes to patterns of species diversity, functional diversity, and coexistence within tropical forests has generally focused on successional development (Meiners et al., 2014). Several studies have corroborated the successional dynamic model; phylogenetic clustering (i.e., community members are more closely related than expected by chance) is observed at earlier stages of succession and in younger aged cohorts; and overdispersion (i.e., community members are more distantly related than expected by chance) is observed at later successional stages and in older-aged cohorts (Letcher, 2010a, Li et al., 2015, Whitfeld et al., 2012, Mo et al., 2013). Generally, these results are interpreted as environmental filtering influencing species coexistence at younger successional stages and competitive exclusion operating at older successional stages (Cavender-Bares et al., 2009, Cornwell and Ackerly, 2009). Alternatively, positive biotic interactions (e.g. mutualism and facilitation) can influence species coexistence (Shooner et al., 2015, Valiente-Banuet and Verdú, 2007, Verdú et al., 2012), or negative density- dependence (particularly at the seedling stage) may also have a strong influence on coexistence within tropical forests (Webb et al., 2006, Paine et al., 2012). Using both phylogenetic and a wide range of functional traits (including within species variation), Baraloto et al. (2012a) found that environmental filtering was the strongest mechanism influencing the community assembly of trees (>10 cm DBH), within a tropical forest of French Guiana.

In order to decipher community assembly processes from PCMs, the spatial, temporal and phylogenetic scales should be defined, and the results should be analysed and interpreted in context (Cadotte, 2014, Swenson et al., 2007). For example, community assembly mechanisms appear to vary spatially; Parmentier et al. (2014) found that environmental filtering was a key process in tropical forest communities at most spatial scales, however at finer scales (<1m) competitive exclusion was suggested to be a stronger mechanism limiting the occurrence of more closely related species. The relative influence of assembly mechanisms also appears to vary over time (Purschke et al., 2013). Comprising distantly related taxa within an analysis can have a large influence on the phylogenetic diversity within a system, and hinder meaningful conclusions regarding community assembly (Cadotte, 2014). PCMs should be used on different subsets of the regional species pool (e.g. excluding gymnosperms and/or monocots and natives etc.) in order to decipher their influence on the observed results and conclusions regarding community assembly (Schweizer et al., 2015).

Due to the large increase in the use of PCMs in addressing questions regarding conservation and community ecology, a large number of phylogenetic diversity metrics and methods have been developed to answer specific questions (Pausas and Verdú, 2010). Due to this, the ability to synthesize and create generalisations between studies and ecosystems has been hampered (Tucker et al., 2016, Pausas and Verdú, 2010). Despite the large number of metrics that quantify phylogenetic diversity, Tucker et al. (2016) show that there is a large degree of mathematical redundancy among them. They recommend that for questions relating to how much phylogenetic diversity occurs at a site (i.e. α-diversity), researchers should use Faith’s Phylogenetic Diversity (PD) (Faith, 1992). For determining how much divergence in phylogenetic diversity occurs between sites, it is recommended that the Mean Pairwise Phylogenetic Diversity (MPD) metric be used, particularly when researchers are attempting to identify patterns of environmental filtering and describing branching divergences deep within a phylogeny. The Mean Nearest Taxon Phylogenetic Distance (MNTD) has also been suggested as a strong indicator of phylogenetic divergence between sites. The MNTD displays a greater ability to detect competitive interactions and better represents divergences occurring more terminally within a phylogeny (Pausas and Verdú, 2010, Tucker et al., 2016, Kraft et al., 2007, Webb et al., 2002). These metrics can also be used when analysing pure presence-absence data or when including species’ relative abundances (Webb et al., 2002). The influence of uneven numbers of species and abundances between experimental treatments can be accounted for by the use of null models. Null models can be generated a number of different ways, including randomly shuffling the tips of the phylogeny or by randomizing the species identify at the plot level; the type of null model chosen should reflect the research questions (Tucker et al., 2016, Webb et al., 2002). The topology and the accuracy of divergence times within a phylogeny can also have a large influence on the observed phylogenetic diversity, and subsequent conclusions regarding community assembly (Cadotte, 2014).

To date, PCMs within tropical forest have shown that evolutionary relatedness, as a proxy for direct ecological functions, can reveal additional information, without revealing the exact mechanisms (Verdú et al., 2012, Webb et al., 2002, Flynn et al., 2011). Some authors have postulated that phylogenies provided little additional information regarding current ecological processes and have instead encouraged the use of multiple ecologically-meaningful traits (Purschke et al., 2013). The combination of both functional trait and phylogenetic information is now considered important in examining complex ecological processes such as regeneration and successional dynamics.

References Albert, C. H., De Bello, F., Boulangeat, I., Pellet, G., Lavorel, S. & Thuiller, W. 2012. On the importance of intraspecific variability for the quantification of functional diversity. Oikos, 121, 116-126. Ashton, M. S., Gunatilleke, C. V. S., Singhakumara, B. M. P. & Gunatilleke, I. a. U. N. 2001. Restoration pathways for rain forest in southwest Sri Lanka: a review of concepts and models. Forest Ecology and Management, 154, 409-430. Ashton, M. S. & Peters, C. M. 1999. Even-aged silviculture in tropical rainforests of Asia: Lessons learned and myths perpetuated. Journal of Forestry, 97, 14. Attiwill, P. M. 1994. The disturbance of forest ecosystems: the ecological basis for conservative management. Forest Ecology and Management, 63, 247-300. Baraloto, C., Hardy, O. J., Paine, C. E. T., Dexter, K. G., Cruaud, C., Dunning, L. T., Gonzalez, M.- A., Molino, J.-F., Sabatier, D., Savolainen, V. & Chave, J. 2012a. Using functional traits and phylogenetic trees to examine the assembly of tropical tree communities. Journal of Ecology, 100, 690-701. Baraloto, C., Hérault, B., Paine, C. E. T., Massot, H., Blanc, L., Bonal, D., Molino, J.-F., Nicolini, E. A. & Sabatier, D. 2012b. Contrasting taxonomic and functional responses of a tropical tree community to selective logging. Journal of Applied Ecology, 49, 861-870. Barlow, J., Gardner, T. A., Araujo, I. S., Ávila-Pires, T. C., Bonaldo, A. B., Costa, J. E., Esposito, M. C., Ferreira, L. V., Hawes, J., Hernandez, M. I. M., Hoogmoed, M. S., Leite, R. N., Lo- Man-Hung, N. F., Malcolm, J. R., Martins, M. B., Mestre, L. a. M., Miranda-Santos, R., Nunes-Gutjahr, A. L., Overal, W. L., Parry, L., Peters, S. L., Ribeiro-Junior, M. A., Da Silva, M. N. F., Da Silva Motta, C. & Peres, C. A. 2007. Quantifying the biodiversity value of tropical primary, secondary, and plantation forests. Proceedings of the National Academy of Sciences, 104, 18555-18560. Barlow, J., Lennox, G. D., Ferreira, J., Berenguer, E., Lees, A. C., Nally, R. M., Thomson, J. R., Ferraz, S. F. D. B., Louzada, J., Oliveira, V. H. F., Parry, L., Ribeiro De Castro Solar, R., Vieira, I. C. G., Aragão, L. E. O. C., Begotti, R. A., Braga, R. F., Cardoso, T. M., Jr, R. C. D. O., Souza Jr, C. M., Moura, N. G., Nunes, S. S., Siqueira, J. V., Pardini, R., Silveira, J. M., Vaz-De-Mello, F. Z., Veiga, R. C. S., Venturieri, A. & Gardner, T. A. 2016. Anthropogenic disturbance in tropical forests can double biodiversity loss from deforestation. Nature, 535, 144-147. Beckman, N. G. & Rogers, H. S. 2013. Consequences of Seed Dispersal for Plant Recruitment in Tropical Forests: Interactions Within the Seedscape. Biotropica, 45, 666-681.

Bloom, A. J., Chapin Iii, F. S. & Mooney, H. A. 1985. Resource limitation in plants - an economic analogy. Annual review of ecology and systematics. Vol. 16, 363-392. Bloor, J. M. G. & Grubb, P. J. 2004. Morphological Plasticity of Shade-Tolerant Tropical Rainforest Tree Seedlings Exposed to Light Changes. Functional Ecology, 18, 337-348. Booth, T. H. & Williams, K. J. 2012. Developing biodiverse plantings suitable for changing climatic conditions 1: Underpinning scientific methods ecological management and restoration, 13, 267. Bowman, D. M. J. S. 1998. The impact of Aboriginal landscape burning on the Australian biota. New Phytologist, 140, 385-410. Brown, S. & Lugo, A. E. 1990. Tropical Secondary Forests. Journal of Tropical Ecology, 6, 1-32. Bryant, D., Nielson, D. & Tangley, L. 1997. The last frontier forests. Issues in Science and Technology, 14, 85-87. Butler, D. W., Green, R. J., Lamb, D., Mcdonald, W. J. F. & Forster, P. I. 2007. Biogeography of seed-dispersal syndromes, life-forms and seed sizes among woody rain-forest plants in Australia's subtropics. Journal of Biogeography, 34, 1736-1750. Cadotte, M. W. 2014. Including distantly related taxa can bias phylogenetic tests. Proceedings of the National Academy of Sciences, 111, E536. Catterall, C. P. & Harrison, D. A. 2006. RAINFOREST RESTORATION ACTIVITIES IN AUSTRALIA'S TROPICS AND SUBTROPICS Rainforest CRC and Environmental Sciences , Griffith University Cooperative Research Centre for Tropical Rainforest Ecology and Management. Cavender-Bares, J., Kozak, K. H., Fine, P. V. A. & Kembel, S. W. 2009. The merging of community ecology and phylogenetic biology. Ecology Letters, 12, 693-715. Chave, J., Coomes, D., Jansen, S., Lewis, S. L., Swenson, N. G. & Zanne, A. E. 2009. Towards a worldwide wood economics spectrum. Ecology Letters, 12, 351-366. Chazdon, R. L. 2008. Beyond Deforestation: Restoring Forests and Ecosystem Services on Degraded Lands. Science, 320, 1458-1460. Chesson, P. 2000. Mechanisms of Maintenance of Species Diversity. Annual Review of Ecology and Systematics, 31, 343-366. Clark, J. A. & Covey, K. R. 2012. Tree species richness and the logging of natural forests: A meta- analysis. Forest Ecology and Management, 276, 146-153. Connell, J. H. 1971. On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. In: Dynamics of Populations (eds P. J. den Boer & G. Gradwell) . , Pudoc, Wageningen., pp. 298-312.

Corlett, R. 1998. Frugivory and seed dispersal by vertebrates in the Oriental (Indomalayan) Region. Biological Reviews, 73, 413-448. Corlett, R. 2011a. Seed dispersal in Hong Kong, China: past, present and possible futures. Intergrative Zoology, 6, 97-109. Corlett, R. T. 2009. Seed dispersal distances and plant migration potential in tropical East Asia Biotropica 41: 592–598. Corlett, R. T. 2011b. Impacts of warming on tropical lowland rainforests. Trends in Ecology & Evolution, 26, 606-613. Corlett, R. T. 2011c. Impacts of warming on tropical lowland rainforests. . Trends in Ecology and Evolution, 26, 606-613. Corlett, R. T. & Lucas, P. W. 1990. Alternative Seed-Handling Strategies in Primates: Seed- Spitting by Long-Tailed Macaques (Macaca fascicularis). Oecologia, 82, 166-171. Cornwell, W. K. & Ackerly, D. D. 2009. Community Assembly and Shifts in Plant Trait Distributions across an Environmental Gradient in Coastal California. Ecological Monographs, 79, 109-126. Crome, F. H. J., Moore, L. A. & Richards, G. C. 1992. A study of logging damage in upland rainforest in north Queensland. Forest Ecology and Management, 49, 1-29. Degaard, F. 2000. How many species of arthropods? Erwin's estimate revised. Biological Journal of the Linnean Society, 71, 583-597. Díaz, S., Symstad, A. J., Chapin Iii, F. S., Wardle, D. A. & Huenneke, L. F. 2003. Functional diversity revealed by removal experiments. Trends in Ecology and Evolution, 18, 140-146. Dressler, W. & Pulhin, J. 2010. The shifting ground of swidden agriculture on Palawan Island, the Philippines. Agriculture and Human Values, 27, 445-459. Dressler, W. H., Kull, C. A. & Meredith, T. C. 2006. The politics of decentralizing national parks management in the Philippines. Political Geography, 25, 789-816. Edwards, D. P., Tobias, J. A., Sheil, D., Meijaard, E. & Laurance, W. F. 2014a. Maintaining ecosystem function and services in logged tropical forests. Trends in Ecology & Evolution, 29, 511-520. Edwards, E. J., Chatelet, D. S., Sack, L. & Donoghue, M. J. 2014b. Leaf life span and the leaf economic spectrum in the context of whole plant architecture. Journal of Ecology, 102, 328- 336. Erskine, P. D. 2002. Land clearing and forest rehabilitation in the Wet Tropics of north Queensland, Australia. Ecological Management & Restoration, 3, 135-152.

Erskine, P. D., Lamb, D. & Bristow, M. 2006. Tree species diversity and ecosystem function: Can tropical multi-species plantations generate greater productivity? Forest Ecology and Management, 233, 205-210. Erskine, P. D., Lamb, D., Bristow M., 2005. Reforestation in the Tropics and Subtropics of Australia Rainforest CRC: Rural Industries Research and Development Corporation. Fairfax, R., Fensham, R., Butler, D., Quinn, K., Sigley, B. & Holman, J. 2009. Effects of multiple fires on tree invasion in montane grasslands. Landscape Ecology, 24, 1363-1373. Faith, D. P. 1992. Conservation evaluation and phylogenetic diversity. Biological Conservation, 61, 1-10. Farwig, N., Sajita, N. & Böhning-Gaese, K. 2009. High seedling recruitment of indigenous tree species in forest plantations in Kakamega Forest, western Kenya. Forest Ecology and Management, 257, 143-150. Firn, J., Erskine, P. D. & Lamb, D. 2007. Woody Species Diversity Influences Productivity and Soil Nutrient Availability in Tropical Plantations. Oecologia, 154, 521-533. Flynn, D. F. B., Mirotchnick, N., Jain, M., Palmer, M. I. & Naeem, S. 2011. Functional and phylogenetic diversity as predictors of biodiversity- Ecosystem-function relationships. Ecology, 92, 1573-1581. Forrester, D., Vanclay, J. & Forrester, R. 2011. The balance between facilitation and competition in mixtures of Eucalyptus and Acacia changes as stands develop. Oecologia, 166, 265-272. Forrester, D. I. & Tang, X. L. 2016. Analysing the spatial and temporal dynamics of species interactions in mixed-species forests and the effects of stand density using the 3-PG model. Ecological Modelling, 319, 233-254. Funk, J. L., Larson, J. E., Ames, G. M., Butterfield, B. J., Cavender-Bares, J., Firn, J., Laughlin, D. C., Sutton-Grier, A. E., Williams, L. & Wright, J. 2016. Revisiting the Holy Grail: using plant functional traits to understand ecological processes. Biological Reviews, DOI: 10.1111/brv.12275. Galetti, M., Guevara, R., Côrtes, M. C., Fadini, R., Von Matter, S., Leite, A. B., Labecca, F., Ribeiro, T., Carvalho, C. S., Collevatti, R. G., Pires, M. M., Guimarães, P. R., Brancalion, P. H., Ribeiro, M. C. & Jordano, P. 2013. Functional Extinction of Birds Drives Rapid Evolutionary Changes in Seed Size. Science, 340, 1086-1090. Garamszegi, L. Z. & Møller, A. P. 2010. Effects of sample size and intraspecific variation in phylogenetic comparative studies: a meta-analytic review. Biological Reviews, 85, 797-805.

Gardner, T. A., Barlow, J., Chazdon, R., Ewers, R. M., Harvey, C. A., Peres, C. A. & Sodhi, N. S. 2009. Prospects for tropical forest biodiversity in a human-modified world. Ecology Letters, 12, 561-582. Garnier, E., Laurent, G., Bellmann, A., Debain, S., Berthelier, P., Ducout, B., Roumet, C. & Navas, M. L. 2001. Consistency of species ranking based on functional leaf traits. New Phytologist, 152, 69-83. Garrity, D. P., Soekardi, M., Noordwijk, M., Cruz, R., Pathak, P. S., Gunasena, H. P. M., So, N., Huijun, G. & Majid, N. M. 1996. The Imperata grasslands of tropical Asia: area, distribution, and typology. Agroforestry Systems, 36, 3-29. Gibert, A., Gray, E. F., Westoby, M., Wright, I. J. & Falster, D. S. 2016. On the link between functional traits and growth rate: meta-analysis shows effects change with plant size, as predicted. Journal of Ecology, n/a-n/a. Goosem, S. & Tucker, N. I. J. 2013. Repairing the rainforest (second edition). Wet Tropics Management Authority and Biotropica Australia Pty. Ltd. , Cairns. Grime, J. P. 1979. Plant stratagies and vegetation processes, New York, USA, Wiley. Grogan, J., Landis, R. M., Free, C. M., Schulze, M. D., Lentini, M. & Ashton, M. S. 2014. Big-leaf mahogany Swietenia macrophylla population dynamics and implications for sustainable management. Journal of Applied Ecology, 51, 664-674. Hansen, M. C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. A., Tyukavina, A., Thau, D., Stehman, S. V., Goetz, S. J., Loveland, T. R., Kommareddy, A., Egorov, A., Chini, L., Justice, C. O. & Townshend, J. R. G. 2013. High-Resolution Global Maps of 21st-Century Forest Cover Change. Science, 342, 850-853. Harms, K. E., Wright, S. J., Calderon, O., Hernandez, A. & Edward Allen, H. 2000. Pervasive density-dependent recruitment enhances seedling diversity in a tropical forest. Nature, 404, 493-5. Harrison, R. D. 2005. Figs and the Diversity of Tropical Rainforests. Bioscience, 55, 1053-1064. Hautier, Y., Saner, P., Philipson, C., Bagchi, R., Ong, R. C. & Hector, A. 2010. Effects of Seed Predators of Different Body Size on Seed Mortality in Bornean Logged Forest. PLoS ONE, 5, e11651. Hector, A., Philipson, C., Saner, P., Chamagne, J., Dzulkifli, D., O'brien, M., Snaddon, J. L., Ulok, P., Weilenmann, M., Reynolds, G. & Godfray, H. C. J. 2011. The Sabah Biodiversity Experiment: a long-term test of the role of tree diversity in restoring tropical forest structure and functioning. Philosophical Transactions of the Royal Society B: Biological Sciences, 366, 3303-3315.

Hill, R. & Baird, A. 2003. Kuku—Yalanji Rainforest Aboriginal People and Carbohydrate Resource Management in the Wet Tropics of Queensland, Australia. Human Ecology, 31, 27-52. Hooper, D. U., Chapin, F. S., Iii, Ewel, J. J., Hector, A., Inchausti, P., Lavorel, S., Lawton, J. H., Lodge, D. M., Loreau, M., Naeem, S., Schmid, B., Setälä, H., Symstad, A. J., Vandermeer, J. & Wardle, D. A. 2005. Effects of Biodiversity on Ecosystem Functioning: A Consensus of Current Knowledge. Ecological Monographs, 75, 3-35. Horne, R. & Hickey, J. 1991. Ecological sensitivity of Australian rainforests to selective logging. Australian Journal of Ecology, 16, 119-129. Howe, H. F. & Smallwood, J. 1982. Ecology of Seed Dispersal. Annual Review of Ecology and Systematics, 13, 201-228. Hubbell, S. P. 2001. The Unified Neutral Theory of Biodiversity and Biogeography Princeton, NJ., Princeton University Press Imai, N., Seino, T., Aiba, S.-I., Takyu, M., Titin, J. & Kitayama, K. 2012. Effects of selective logging on tree species diversity and composition of Bornean tropical rain forests at different spatial scales. Plant Ecology, 213, 1413-1424. Itto 2002. Guidelines for the restoration, management and rehabilitation of degraded and secondary tropical forests.: International Tropical Timber Organization Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests. The American Naturalist, Vol. 104, No. 940. (Nov. - Dec., 1970), pp. 501-528., 104. Johns, R. J. 2008. Malesia- An Introduction. . Curtis's Botanical Magazine, 12, 52-62. Jung, V., Violle, C., Mondy, C., Hoffmann, L. & Muller, S. 2010. Intraspecific variability and trait- based community assembly. Journal of Ecology, 98, 1134-1140. Kanowski, J., Catterall, C. P., Wardell-Johnson, G. W., Proctor, H. & Reis, T. 2003. Development of forest structure on cleared rainforest land in eastern Australia under different styles of reforestation. Forest Ecology and Management, 183, 265-280. Katovai, E., Burley, A. L. & Mayfield, M. M. 2012. Understory plant species and functional diversity in the degraded wet tropical forests of Kolombangara Island, Solomon Islands. biological conservation, 145, 214-224. Kattge, J., Díaz, S., Lavorel, S., Prentice, I. C., Leadley, P., Bönisch, G., Garnier, E., Westoby, M., Reich, P. B., Wright, I. J., Cornelissen, J. H. C., Violle, C., Harrison, S. P., Van Bodegom, P. M., Reichstein, M., Enquist, B. J., Soudzilovskaia, N. A., Ackerly, D. D., Anand, M., Atkin, O., Bahn, M., Baker, T. R., Baldocchi, D., Bekker, R., Blanco, C. C., Blonder, B., Bond, W. J., Bradstock, R., Bunker, D. E., Casanoves, F., Cavender-Bares, J., Chambers, J.

Q., Chapin Iii, F. S., Chave, J., Coomes, D., Cornwell, W. K., Craine, J. M., Dobrin, B. H., Duarte, L., Durka, W., Elser, J., Esser, G., Estiarte, M., Fagan, W. F., Fang, J., Fernández- Méndez, F., Fidelis, A., Finegan, B., Flores, O., Ford, H., Frank, D., Freschet, G. T., Fyllas, N. M., Gallagher, R. V., Green, W. A., Gutierrez, A. G., Hickler, T., Higgins, S. I., Hodgson, J. G., Jalili, A., Jansen, S., Joly, C. A., Kerkhoff, A. J., Kirkup, D., Kitajima, K., Kleyer, M., Klotz, S., Knops, J. M. H., Kramer, K., Kühn, I., Kurokawa, H., Laughlin, D., Lee, T. D., Leishman, M., Lens, F., Lenz, T., Lewis, S. L., Lloyd, J., Llusià, J., Louault, F., Ma, S., Mahecha, M. D., Manning, P., Massad, T., Medlyn, B. E., Messier, J., Moles, A. T., Müller, S. C., Nadrowski, K., Naeem, S., Niinemets, Ü., Nöllert, S., Nüske, A., Ogaya, R., Oleksyn, J., Onipchenko, V. G., Onoda, Y., Ordoñez, J., Overbeck, G., Ozinga, W. A., et al. 2011. TRY – a global database of plant traits. Global Change Biology, 17, 2905-2935. Kettle, C. 2010. Ecological considerations for using dipterocarps for restoration of lowland rainforest in Southeast Asia. Biodiversity and Conservation, 19, 1137-1151. Kraft, N. J. B. & Ackerly, D. D. 2010. Functional trait and phylogenetic tests of community assembly across spatial scales in an Amazonian forest. Ecological Monographs, 80, 401- 422. Kraft, Nathan j. B., Cornwell, William k., Webb, Campbell o. & Ackerly, David d. 2007. Trait Evolution, Community Assembly, and the Phylogenetic Structure of Ecological Communities. The American Naturalist, 170, 271-283. Kraft, N. J. B., Godoy, O. & Levine, J. M. 2015. Plant functional traits and the multidimensional nature of species coexistence. Proceedings of the National Academy of Sciences, 112, 797- 802. Kraft, N. J. B., Valencia, R. & Ackerly, D. D. 2008. Functional Traits and Niche-Based Tree Community Assembly in an Amazonian Forest. Science, 322, 580-582. Lamb, D. 1998. Large-scale Ecological Restoration of Degraded Tropical Forest Lands: The Potential Role of Timber Plantations. Restoration Ecology, 6, 271-279. Lamb, D. 2011. Regreening the Bare Hills, Tropical Forest Restoration in the Asia-Pacific Region, London, United Kingdom, Springer Lamb, D. 2016. RE: Personal Communication Lamb, D., Erskine, P. D. & Parrotta, J. A. 2005. Restoration of Degraded Tropical Forest Landscapes. Science, 310, 1628-1632. Lasco, R., Evangelista, R. & Pulhin, F. 2010. Potential of Community-Based Forest Management to Mitigate Climate Change in the Philippines. Small-scale Forestry, 9, 429-443.

Lasky, J. R., Uriarte, M., Boukili, V. K. & Chazdon, R. L. 2014. Trait-mediated assembly processes predict successional changes in community diversity of tropical forests. Proceedings of the National Academy of Sciences of the United States of America, 111, 5616-5621. Latham, R. E. & Ricklefs, R. E. 1993. Global Patterns of Tree Species Richness in Moist Forests: Energy-Diversity Theory Does Not Account for Variation in Species Richness. Oikos, 67, 325-333. Laufer, J., Michalski, F. & Peres, C. A. 2013. Assessing sampling biases in logging impact studies in tropical forests. Tropical Conservation Science, 6, 16-34. Laurance, W. F. 1999. Reflections on the tropical deforestation crisis. Biological Conservation, 91, 109-117. Lavorel, S. & Garnier, E. 2002. Predicting changes in community composition and ecosystem functioning from plant traits: revisiting the Holy Grail. Functional Ecology, 16, 545-556. Le, H. D., Smith, C. & Herbohn, J. 2014. What drives the success of reforestation projects in tropical developing countries? The case of the Philippines. Global Environmental Change, 24, 334-348. Letcher, S. G. 2010. Phylogenetic structure of angiosperm communities during tropical forest succession. Li, S.-P., Cadotte, M. W., Meiners, S. J., Hua, Z.-S., Jiang, L. & Shu, W.-S. 2015. Species colonisation, not competitive exclusion, drives community overdispersion over long-term succession. Ecology Letters, 18, 964-973. Lindenmayer, D. B. & Hobbs, R. J. 2004. Fauna conservation in Australian plantation forests – a review. Biological Conservation, 119, 151-168. Maohong, B. 2012. Deforestation in the Philippines 1946-1995. Ateneo de Manila University, 60, 117-130. Markl, J. S., Schleuning, M., Michel Forgit, P., Lambert, J. E., Traveset, A., Wright, S. J. & Bohning-Gease, K. 2012. Meta-Analysis of the Effects of Human Disturbance on Seed Dispersal by Animals. Conservation Biology 26, 1072-1081. Mayfield, M. M., Ackerly, D. & Daily, G. C. 2006. The diversity and conservation of plant reproductive and dispersal functional traits in human-dominated tropical landscapes. Journal of Ecology, 94, 522-536. Mcconkey, K. R., Prasad, S., Corlett, R. T., Campos-Arceiz, A., Brodie, J. F., Rogers, H. & Santamaria, L. 2012. Seed dispersal in changing landscapes. Biological Conservation, 146, 1-13.

Mcgill, B. J., Enquist, B. J., Weiher, E. & Westoby, M. 2006. Rebuilding community ecology from functional traits. Trends in Ecology & Evolution, 21, 178-185. Mcnamara, S., Erskine, P. D., Lamb, D., Chantalangsy, L. & Boyle, S. 2012. Primary tree species diversity in secondary fallow forests of Laos. Forest Ecology and Management, 281, 93-99. Mea 2005. Ecosystems and Human Well-Being In: A, S. J. A. W. (ed.) A Report of the Millennium Ecosytem Assessment Washington, DC: Millennium Ecosystem Assessment Medjibe, V. P., Putz, F. E., Starkey, M. P., Ndouna, A. A. & Memiaghe, H. R. 2011. Impacts of selective logging on above-ground forest biomass in the Monts de Cristal in Gabon. Forest Ecology and Management, 262, 1799-1806. Meiners, S. J., Cadotte, M. W., Fridley, J. D., Pickett, S. T. A. & Walker, L. R. 2014. Is successional research nearing its climax? New approaches for understanding dynamic communities. Functional Ecology, n/a-n/a. Meyfroidt, P. & Lambin, E. F. 2011. Global Forest Transition: Prospects for an End to Deforestation. Annual Review of Environment and Resources, 36, 343-371. Milan, P. P., Ceniza, M. J. C., Asio, V. B., Bulayog, S. B. & Napiza, M. D. 2004. Evaluation of Silvicultural Management, Ecological Changes and Market Study of Products of Existing Rainforestation demo and cooperation farms. Institute of tropical ecology terminal report Mo, X.-X., Shi, L.-L., Zhang, Y.-J., Zhu, H. & Slik, J. W. F. 2013. Change in Phylogenetic Community Structure during Succession of Traditionally Managed Tropical Rainforest in Southwest China. PLoS ONE, 8, e71464. Montagnini, F., Fanzeres, A. & Davinha, S. G. 1995. The potentials of 20 indigenous tree species for soil rehabilitation in the Atlantic forest region of Bahia, Brazil. Journal of Applied Ecology, 32, 841-856. Mukul, S. A. & Herbohn, J. 2016. The impacts of shifting cultivation on secondary forests dynamics in tropics: A synthesis of the key findings and spatio temporal distribution of research. Environmental Science & Policy, 55, Part 1, 167-177. Mukul, S. A., Herbohn, J. & Firn, J. 2016. Tropical secondary forests regenerating after shifting cultivation in the Philippines uplands are important carbon sinks. Scientific Reports, 6, 22483. Mwavu, E. N. & Witkowski, E. T. F. 2009. Seedling regeneration, environment and management in a semi-deciduous African tropical rain forest. Journal of Vegetation Science, 20, 791-804. Myers, N., Mittermeier, R. A., Mittermeier, C. G., Da Fonseca, G. a. B. & Kent, J. 2000. Biodiversity hotspots for conservation priorities Nature, 403.

Nguyen, H., Firn, J., Lamb, D. & Herbohn, J. 2014. Wood density: A tool to find complementary species for the design of mixed species plantations. Forest Ecology and Management, 334, 106-113. Nguyen, H., Herbohn, J., Firn, J. & Lamb, D. 2012. Biodiversity–productivity relationships in small-scale mixed-species plantations using native species in Leyte province, Philippines. Forest Ecology and Management, 274, 81-90. Paine, C. E. T., Norden, N., Chave, J., Forget, P.-M., Fortunel, C., Dexter, K. G. & Baraloto, C. 2012. Phylogenetic density dependence and environmental filtering predict seedling mortality in a tropical forest. Ecology Letters, 15, 34-41. Parmentier, I., Réjou-Méchain, M., Chave, J., Vleminckx, J., Thomas, D. W., Kenfack, D., Chuyong, G. B. & Hardy, O. J. 2014. Prevalence of phylogenetic clustering at multiple scales in an African rain forest tree community. Journal of Ecology, 102, 1008-1016. Parrotta, J. A. 1992. The role of plantation forests in rehabilitating degraded tropical ecosystems. Agriculture, Ecosystems & Environment, 41, 115-133. Pausas, J. G. & Verdú, M. 2010. The Jungle of Methods for Evaluating Phenotypic and Phylogenetic Structure of Communities. BioScience, 60, 614-625. Peres, C. A. 2000. Effects of Subsistence Hunting on Vertebrate Community Structure in Amazonian Forests. Conservation Biology, 14, 240-253. Pérez-Harguindeguy, N., Díaz, S., Garnier, E., Lavorel, S., Poorter, H., Jaureguiberry, P., Bret- Harte, M. S., Cornwell, W. K., Craine, J. M., Gurvich, D. E., Urcelay, C., Veneklaas, E. J., Reich, P. B., Poorter, L., Wright, I. J., Ray, P., Enrico, L., Pausas, J. G., De Vos, A. C., Buchmann, N., Funes, G., Quétier, F., Hodgson, J. G., Thompson, K., Morgan, H. D., Ter Steege, H., Van Der Heijden, M. G. A., Sack, L., Blonder, B., Poschlod, P., Vaieretti, M. V., Conti, G., Staver, A. C., Aquino, S. & Cornelissen, J. H. C. 2013. New handbook for standardised measurement of plant functional traits worldwide. Australian Journal of Botany, 61, 167-234. Plugge, D., Baldauf, T. & Köhl, M. 2013. The global climate change mitigation strategy REDD: monitoring costs and uncertainties jeopardize economic benefits. Climatic Change, 119, 247-259. Poorter, H., Lambers, H. & Evans, J. R. 2014. Trait correlation networks: a whole-plant perspective on the recently criticized leaf economic spectrum. New Phytologist, 201, 378-382. Poorter, L. & Bongers, F. 2006. Leaf Traits are Good Predictors of Plant Performance Across 53 Rain Forest Species. Ecology, 87, 1733-1743.

Poorter, L., Wright, S. J., Paz, H., Ackerly, D. D., Condit, R., Ibarra-Manríquez, G., Harms, K. E., Licona, J. C., Martínez-Ramos, M., Mazer, S. J., Muller-Landau, H. C., Peña-Claros, M., Webb, C. O. & Wright, I. J. 2008. Are Functional Traits Good Predictors of Demographic Rates? Evidence From Five Neotropical Forests. Ecology, 89, 1908-1920. Potvin, C. & Gotelli, N. J. 2008. Biodiversity enhances individual performance but does not affect survivorship in tropical trees. Ecology Letters, 11, 217-223. Purschke, O., Schmid, B. C., Sykes, M. T., Poschlod, P., Michalski, S. G., Durka, W., Kühn, I., Winter, M. & Prentice, H. C. 2013. Contrasting changes in taxonomic, phylogenetic and functional diversity during a long-term succession: insights into assembly processes. Journal of Ecology, 101, 857-866. Putz, F. E., Sist, P., Fredericksen, T. & Dykstra, D. 2008. Reduced-impact logging: Challenges and opportunities. Forest Ecology and Management, 256, 1427-1433. Putz, F. E., Zuidema, P. A., Synnott, T., Peña-Claros, M., Pinard, M. A., Sheil, D., Vanclay, J. K., Sist, P., Gourlet-Fleury, S., Griscom, B., Palmer, J. & Zagt, R. 2012. Sustaining conservation values in selectively logged tropical forests: the attained and the attainable. Conservation Letters, 5, 296-303. Rees, M., Condit, R., Crawley, M., Pacala, S. & Tilman, D. 2001. Long-Term Studies of Vegetation Dynamics. Science, 293, 650-655. Reich, P. B. 2014. The world-wide 'fast-slow' plant economics spectrum: A traits manifesto. Journal of Ecology, 102, 275-301. Reich, P. B., Bakken, P., Carlson, D., Frelich, L. E., Friedman, S. K. & Grigal, D. F. 2001. Influence of Logging, Fire, and Forest Type on Biodiversity and Productivity in Southern Boreal Forests. Ecology, 82, 2731-2748. Reich, P. B., Walters, M. B. & Ellsworth, D. S. 1997. From tropics to tundra: Global convergence in plant functioning. Proceedings of the National Academy of Sciences of the United States of America, 94, 13730-13734. Richards, C. L., Bossdorf, O., Muth, N. Z., Gurevitch, J. & Pigliucci, M. 2006. Jack of all trades, master of some? On the role of phenotypic plasticity in plant invasions. Ecology Letters, 9, 981-993. Richards, P. W. 1952. The tropical rain forest. An ecological study., Cambridge University Press, Cambridge. Ricklefs, R. E. 1987. Community Diversity: Relative Roles of Local and Regional Processes. Science, 235, 167-171.

Ricklefs, Robert e. 2008. Disintegration of the Ecological Community. The American Naturalist, 172, 741-750. Russell-Smith, J. 1991. Classification, species richness, and environmental relations of monsoon rain forest in northern Australia. Journal of Vegetation Science, 2, 259-278. Sales-Come, R. 2010. Variability and Grouping of Tree Leaf Traits in Multi-Species Reforestation Gottingen University. Scherr, S. J., White, A. & Kaimowitz, A. 2004. A new Agenda for Forest Conservation and Poverty Reduction: Making Markets Work for low-income producers. Forest Trends. Washington DC. Schmitt, C. B., Belokurov, A., Besanc¸on, C., Boisrobert, L., Burgess,, N.D., C., A., Coad, L., Fish, L., Gliddon, D., Humphries,, K., K., V., Loucks, C., Lysenko, I., Miles, L., Mills, C.,, Minnemeyer, S., Pistorius, T., Ravilious, C., Steininger, M. & & Winkel, G. 2008. Global Ecological Forest Classification and Forest Protected Area Gap Analysis. Analyses and recommendations in view of the 10% target for forest protection under the Convention on Biological Diversity (CBD). In. University of Freiburg, Freiburg. Schnitzer, S. A. & Bongers, F. 2011. Increasing liana abundance and biomass in tropical forests: emerging patterns and putative mechanisms. Ecology Letters, 14, 397-406. Schweizer, D., Machado, R., Durigan, G. & Brancalion, P. H. S. 2015. Phylogenetic patterns of Atlantic forest restoration communities are mainly driven by stochastic, dispersal related factors. Forest Ecology and Management, 354, 300-308. Shea, G. M. 1992. New timber industry based on valuable cabinetwoods and hardwoods. Queensland Forest Service, Consultancy report for Councils of the Wet Tropics Region, Brisbane Shipley, B., Lechowicz, M. J., Wright, I. & Reich, P. B. 2006. Fundamental Trade-offs Generating the Worldwide Leaf Econmic Spectrum Ecology, 87, 535-541. Shooner, S., Chisholm, C. & Davies, T. J. 2015. The phylogenetics of succession can guide restoration: an example from abandoned mine sites in the subarctic. J Appl Ecol, 52: 1509– 1517. doi:10.1111/1365-2664.12517. Simonetti, J. A., Grez, A. A. & Estades, C. F. 2013. Providing Habitat for Native Mammals through Understory Enhancement in Forestry Plantations. Conservation Biology, 27, 1117-1121. Slik, J. W. F., Arroyo-Rodríguez, V., Aiba, S.-I., Alvarez-Loayza, P., Alves, L. F., Ashton, P., Balvanera, P., Bastian, M. L., Bellingham, P. J., Van Den Berg, E., Bernacci, L., Da Conceição Bispo, P., Blanc, L., Böhning-Gaese, K., Boeckx, P., Bongers, F., Boyle, B.,

Bradford, M., Brearley, F. Q., Breuer-Ndoundou Hockemba, M., Bunyavejchewin, S., Calderado Leal Matos, D., Castillo-Santiago, M., Catharino, E. L. M., Chai, S.-L., Chen, Y., Colwell, R. K., Chazdon, R. L., Clark, C., Clark, D. B., Clark, D. A., Culmsee, H., Damas, K., Dattaraja, H. S., Dauby, G., Davidar, P., Dewalt, S. J., Doucet, J.-L., Duque, A., Durigan, G., Eichhorn, K. a. O., Eisenlohr, P. V., Eler, E., Ewango, C., Farwig, N., Feeley, K. J., Ferreira, L., Field, R., De Oliveira Filho, A. T., Fletcher, C., Forshed, O., Franco, G., Fredriksson, G., Gillespie, T., Gillet, J.-F., Amarnath, G., Griffith, D. M., Grogan, J., Gunatilleke, N., Harris, D., Harrison, R., Hector, A., Homeier, J., Imai, N., Itoh, A., Jansen, P. A., Joly, C. A., De Jong, B. H. J., Kartawinata, K., Kearsley, E., Kelly, D. L., Kenfack, D., Kessler, M., Kitayama, K., Kooyman, R., Larney, E., Laumonier, Y., Laurance, S., Laurance, W. F., Lawes, M. J., Amaral, I. L. D., Letcher, S. G., Lindsell, J., Lu, X., Mansor, A., Marjokorpi, A., Martin, E. H., Meilby, H., Melo, F. P. L., Metcalfe, D. J., Medjibe, V. P., Metzger, J. P., Millet, J., Mohandass, D., Montero, J. C., De Morisson Valeriano, M., Mugerwa, B., Nagamasu, H., Nilus, R., Ochoa-Gaona, S., et al. 2015. An estimate of the number of tropical tree species. Proceedings of the National Academy of Sciences, 112, 7472-7477. Sniderman, J. M. K. & Jordan, G. J. 2011. Extent and timing of floristic exchange between Australian and Asian rain forests. Journal of Biogeography, 38, 1445-1455. Sodhi, N. S., Posa, M. R. C., Lee, T. M., Bickford, D., Koh, L. P. & Brook, B. W. 2010. The state and conservation of southeast asian biodiversity. Biodiversity Conservation, 19, 317-328. Sterck, F. J. & Bongers, F. 2001. Crown development in tropical rain forest trees: patterns with tree height and light availability. Journal of Ecology, 89, 1-13. Sunderlin, W. D., Angelsen, A., Belcher, B., Burgers, P., Nasi, R., Santoso, L. & Wunder, S. 2005. Livelihoods, forests, and conservation in developing countries: An Overview. World Development, 33, 1383-1402. Swaine, M. D. & Whitmore, T. C. 1988. On the definition of ecological species groups in tropical rain forests. Vegetatio, 75, 81-86. Swenson, N. G. & Enquist, B. J. 2009. Opposing Assembly Mechanisms in a Neotropical Dry Forest: Implications for Phylogenetic and Functional Community Ecology. Ecology, 90, 2161-2170. Swenson, N. G., Enquist, B. J., Thompson, J. & Zimmerman, J. K. 2007. The Influence of Spatial and Size Scale on Phylogenetic Relatedness in Tropical Forest Communities. Ecology, 88, 1770-1780.

Ter Steege, H., Jetten, V. G., Polak, A. M. & Werger, M. J. A. 1993. Tropical rain forest types and soil factors in a watershed area in Guyana. Journal of Vegetation Science, 4, 705-716. Tucker, C. M., Cadotte, M. W., Carvalho, S. B., Davies, T. J., Ferrier, S., Fritz, S. A., Grenyer, R., Helmus, M. R., Jin, L. S., Mooers, A. O., Pavoine, S., Purschke, O., Redding, D. W., Rosauer, D. F., Winter, M. & Mazel, F. 2016. A guide to phylogenetic metrics for conservation, community ecology and macroecology. Biological Reviews, n/a-n/a. Usinowicz, J., Wright, S. J. & Ives, A. R. 2012. Coexistence in tropical forests through asynchronous variation in annual seed production. Ecology, 93, 2073-2084. Valiente-Banuet, A. & Verdú, M. 2007. Facilitation can increase the phylogenetic diversity of plant communities. Ecology Letters, 10, 1029-1036. Valladares, F. & Niinemets, Ü. 2008. Shade Tolerance, a Key Plant Feature of Complex Nature and Consequences. Annual Review of Ecology, Evolution, and Systematics, 39, 237-257. Vanclay, J. K. 1994. Contrasts between biologically-based process models and management- oriented growth and yield models Sustainable timber harvesting: simulation studies in the tropical rainforests of north Queensland. Forest Ecology and Management, 69, 299-320. Vanclay, J. K. 2006. Can lessons from the Community Rainforest Reforestation Program in eastern Australia be learned? International Forestry Review, 8, 256-264. Verburg, R. & Eijk-Bos, C. V. 2003. Effects of Selective Logging on Tree Diversity, Composition and Plant Functional Type Patterns in a Bornean Rain Forest. Journal of Vegetation Science, 14, 99-110. Verdú, M., Gómez-Aparicio, L. & Valiente-Banuet, A. 2012. Phylogenetic relatedness as a tool in restoration ecology: a meta-analysis. Proceedings of the Royal Society B: Biological Sciences, 279, 1761-1767. Violle, C., Enquist, B. J., Mcgill, B. J., Jiang, L., Albert, C. H., Hulshof, C., Jung, V. & Messier, J. 2012. The return of the variance: intraspecific variability in community ecology. Trends in Ecology & Evolution, 27, 244-252. Wadsworth, F. H. & González, E. 2008. Sustained mahogany (Swietenia macrophylla) plantation heartwood increment. Forest Ecology and Management, 255, 320-323. Webb, C. O. 2000. Exploring the Phylogenetic Structure of Ecological Communities: An Example for Rain Forest Trees. The American Naturalist, 156, 145-155. Webb, C. O., Ackerly, D. D., Mcpeek, M. A. & Donoghue, M. J. 2002. Phylogenies and Community Ecology. Annual Review of Ecology and Systematics, 33, 475-505. Webb, C. O., Gilbert, G. S. & Donoghue, M. J. 2006. Phylodiversity-Dependent Seedling Mortality, Size Structure, and Disease in a Bornean Rain Forest. Ecology, 87, S123-S131.

Webb, L. 1958. Cyclones as an ecological factor in tropical lowland rain-forest, North Queensland. Australian Journal of Botany, 6, 220-228. Weiher, E., Freund, D., Bunton, T., Stefanski, A., Lee, T. & Bentivenga, S. 2011. Advances, challenges and a developing synthesis of ecological community assembly theory. Philosophical Transactions of the Royal Society B: Biological Sciences, 366, 2403-2413. Westoby, M. 1998. A leaf-height-seed (LHS) plant ecology strategy scheme. Plant and Soil, 199, 213-227. Westoby, M. & Wright, I. J. 2006. Land-plant ecology on the basis of functional traits. Trends in Ecology & Evolution, 21, 261-268. White, E., Tucker, N., Meyers, N. And Wilson, J. 2004. Seed dispersal to revegetated isoloated rainforest patches in North Queensland Forest Ecology and Management Whitfeld, T. J. S., Kress, W. J., Erickson, D. L. & Weiblen, G. D. 2012. Change in community phylogenetic structure during tropical forest succession: evidence from New Guinea. Ecography, 35, 821-830. Wills, J., Herbohn, J., Moreno, M. O. M., Avela, M. S. & Firn, J. 2016. Next-generation tropical forests: reforestation type affects recruitment of species and functional diversity in a human- dominated landscape. J Appl Ecol. doi:10.1111/1365-2664.12770. Wilson, B. A., Neldner, V. J. & Accad, A. 2002. The extent and status of remnant vegetation in Queensland and its implications for statewide vegetation management and legislation. The Rangeland Journal, 24, 6-35. Wright, I. J., Reich, P. B., Westoby, M., Ackerly, D. D. & Et Al. 2004. The worldwide leaf economics spectrum. Nature, 428, 821-7. Wright, S. J., Kitajima, K., Kraft, N. J. B., Reich, P. B., Wright, I. J., Bunker, D. E., Condit, R., Dalling, J. W., Davies, S. J., Díaz, S., Engelbrecht, B. M. J., Harms, K. E., Hubbell, S. P., Marks, C. O., Ruiz-Jaen, M. C., Salvador, C. M. & Zanne, A. E. 2010. Functional traits and the growth—mortality trade-off in tropical trees. Ecology, 91, 3664-3674. Wtma 2014. Wet Tropics Management Authority Cairns, Queensland Zanne, A. E. & Chapman, C. A. 2001. Expediting Reforestation in Tropical Grasslands: Distance and Isolation from Seed Sources in Plantations. Ecological Applications, 11, 1610-1621.

Chapter 3. Next-generation tropical forests: reforestation type affects recruitment of species and functional diversity in a human-dominated landscape

3.1 Introduction Tropical rainforests cover just 7% of the world’s land surface but support roughly half of all biota on earth (Laurance, 1999) and provide essential socio-economic services to a quarter of the world’s most vulnerable people (Scherr et al., 2004). Between 2000 and 2012, there was an increase in the loss of tropical forests to approximately 2100 km2/year (Hansen et al., 2013). In the Philippines, 100% of its ‘frontier forest’ cover has been cleared (Bryant et al., 1997), with 17% of this land area being replaced by pasture dominated by the exotic grass Imperata cylindrical (cogon grass) (Garrity et al., 1996). Consequently, ecological goods and services are reduced, and this reduction has heavily impacted the livelihoods of the most vulnerable communities (Le et al., 2012). Although clearing of tropical forests is substantial, some tropical countries have transitioned in the last decade from net deforestation to net reforestation (Meyfroidt and Lambin, 2011). Often monoculture plantations are established for providing timber products but lack the diversity of products obtained from the prior forest landscape (Lamb et al., 2005). Mixed-species plantations are increasingly being established to meet conservation and socio-economic needs (Nguyen et al., 2014). Secondary regrowth forests offer a potentially low-cost passive reforestation method, where forests are allowed the time to re-grow with little if any management actions (Brown and Lugo, 1990). Seedling recruitment within plantations is important (Parrotta, 1992) because diversity can provide additional benefits such as more efficient nutrient cycling (Firn et al., 2007); mitigate top soil erosion (Lamb, 1998); maintain a greater number of forest products (Naeem and Wright, 2003); provide habitat for faunal communities (Simonetti et al., 2013); and in turn contribute to long-term sustainability (Lamb, 1998). There are also potential negative consequences of understorey development including reductions in stand productivity and disruption to management techniques during harvesting (Lindenmayer and Hobbs, 2004, Zanne and Chapman, 2001). The extent to which these negatives deter land managers from allowing understorey development depends on their goals and reforestation techniques. The needs of small-holder tree-farmers within the tropics differ to those of an industrial owner. In particular, smallholders are often willing to trade-off maximising productivity and consistency of timber products for a wider range of forest products (Le et al., 2012, Le et al., 2014, Herbohn et al., 2014). Consequently, management techniques used in small-scale community-based plantation projects differ to an industrial timber estate (Harrison et al., 2002). In particular harvesting of timber

64 products for sawmills occurs more sporadically and silvicultural practices such as thinning and weed control are often neglected (Herbohn et al., 2014); therefore, understories often develop within these plantations. This increases the potential of these plantations to contribute to conservation, including increasing species diversity at a site and landscape. Increased diversity can then provide additional non-timber forest products to local people that may result in extending the rotation of these forests as their socio-economic value is not solely reliant on timber (Herbohn et al., 2014). Understanding seedling recruitment within small-scale community-based reforestation is, therefore, an important research goal for both biodiversity and socio-economic outcomes. The reproductive and dispersal traits of recruiting species can indicate important ecosystem functions. Traits that characterise understories can vary depending on the characteristics of the overstorey and the type, intensity and magnitude of anthropogenic disturbance (McConkey et al., 2012, Katovai et al., 2012). Studies have shown animals are the main mode of dispersal for ~80% of tropical woody plant species (Corlett, 2011a, Markl et al., 2012). Declines in large-bodied animal populations due to habitat decline and subsistence hunting have generally reduced large-seeded plant species from recruitment pools (exceptions being,Wright et al. (2007) and Dirzo et al. (2007), and consequently led to declines in tree diversity (Peña-Domene et al., 2013, Harrison et al., 2013). Conversely plant species dispersed by abiotic mechanisms (e.g., wind, water and gravity) or by non- targeted animals, usually with a smaller seed are thought to be dispersal advantaged, regarding seed numbers (Ingle, 2003, Nunez-Iturri et al., 2008) and distances (Corlett, 2009). Many emergent tree species of rainforests (e.g., Octomeles sumatrana Miq.) and some of the most valued timber trees (e.g., Shorea contorta S.Vidal) are dispersed by abiotic mechanisms.

Here we investigate seedling diversity in the understorey of different forest types, i.e., monocultures, mixed-species plantations and regenerating selectively logged forests, across a highly modified agricultural landscape on Leyte Island, Philippines. We specifically aimed to: 1. Measure and compare seedling abundance and composition of species and traits beneath each forest type; 2. Identify species or functional traits that are favoured beneath each forest type; and 3. Identify species or functional traits limited in their reproductive ability and therefore, represent conservation significant species or traits.

We expected to find that seedling diversity would be significantly lower within the understorey of the monoculture forests, and higher and similar in the mixed-species and regenerating selectively logged forests (Barlow et al., 2007). We also hypothesized that large-seeded species would be absent and small-seeded species would dominate the recruitment pool beneath the monoculture forests, while in contrast animal-dispersed species would dominate in the understorey of the more

65 diverse mixed-species plantations and the regenerating selectively logged forests.

3.2. Materials and methods

3.2.1. Study Sites The study was conducted at 15 sites on the Island Leyte, Philippines, which lies between 124ᵒ17’ and 125ᵒ18’ east longitude and between 9ᵒ55’ 11ᵒ48’north latitude. This included five exotic monocultures, five mixed-species forests and five regenerating selectively logged forests (Figure 3.1). All sites had elevations of less than 600 m a.s.l. and were previously lowland forests (~<800 m a.s.l.) dominated by the family Dipterocarpaceae. All sites occurred on soils of volcanic origin with one exception, Pomponanan, where the soil is derived from limestone and volcanic origin (Table 3.1). Study sites are located on the western side of the dividing mountain range with an average annual rainfall of 2753mm and average annual temperature of 27.5ᵒC (Figure 3.1) (Nguyen et al., 2012).

Figure 3.1. Leyte Island, Philippines. Symbols indicate study sites and shaded areas approximately represent regenerating selectively logged forest.

Table 3.1. Site characteristics; for Rainforestation (Rain.), monocultures (Mono.) and regenerating selectively logged forest (Regen.), including: Leaf Area Index (LAI), soil N and P, elevation, slope and aspect were estimated at each plot. Geology, soil type, density and canopy diversity were obtained from previous studies (Nguyen et al., 2012, Le et al., 2014, Le et al., 2012, Milan et al., 2004). Some information was unavailable (na).

Site Name Type Plots Est. Elevation Estimated dist. LAI Geology Soil type Mean Mean Slope Aspect Density Canopy Canopy Canopy (year) (masl) to secondary soil P soil N angle (trees/ha) richness Shannon's Simpson's forests(km) index index

Catmon Rain. 3 1998 42 5.5 2.3 Volcanic Clay loam 0.122 0.083 1 N 5000* 6*** 1.71*** 0.91***

Milagro Rain. 2 1996 600 0.34 0.51 Volcanic Clay loam 0.113 0.375 2 SW 5000* 5*** 1.5*** 0.88***

Marcos Rain. 2 1995 40 0.25 0.87 Volcanic Clay loam 0.053 0.2 13 N 5000* 7.5*** 1.92*** 0.95***

Mailhi Rain. 3 1996 380 0.73 1.31 Volcanic Clay to clay 0.127 0.267 45 W 5000* 5*** 1.29*** 0.68*** loam Cienda Rain. 3 1996 89 0.16 2.94 Volcanic Clay to clay 0.087 0.2 1 S 5000* 5.3*** 1.54*** 0.83*** loam Pomponanan Mono. 1 1998 53 0.18 1.25 Vol/lime na 0.06 0.25 25 E na 2 na na

Mahaplag Mono. 3 1997 56 0.76 0.91 Volcanic na 0.127 0.167 1 E na 1 na na

Mahayag Mono. 2 1998 15 0.48 1.85 Volcanic na 0.178 0.175 1 W na 1 na na

Dolores Mono. 2 2000 250 1.07 1.18 Volcanic na 0.11 0.325 2 SSE na 3 na na

Hikgop Mono. 2 1995 15 0.55 0.82 Volcanic na 0.16 0.325 1 S na 2 na na

Pangasugan Regen. 3 na 172 0 1.61 Volcanic na 0.065 0.183 40 NNE 460** 2.61**** 0.61**** 0.29****

Hubasan Regen. 2 na 360 0 2.64 Volcanic na 0.07 0.275 30 N na na na na

Balinsasayao Regen. 2 na 50 0 1.66 Volcanic na 0.053 0.175 45 S na na na na

Cienda A Regen. 2 na 135 0 1.575 Volcanic na 0.08 0.475 45 E na na na na

Cienda B Regen. 3 na 130 0 2.64 Volcanic na 0.06 0.333 40 SW na na na na

* indicates density at planting, *** indicates canopy diversity measures at the same plots, **, **** indicates estimation of tree density and diversity (>5 cm DBH) scaled to the current plot size, with current plots sampled within larger 0.6ha permanent plot.

The monoculture plantations of Swietenia macrophylla King (mahogany), one of the most valuable timber species in the tropics, were between 13 to 18 years old at the time of sampling and established as small-scale community-based forestry projects (Figure 3.2a and Table 3.1).

Figure 3.2. Pictures of the three forest types; monoculture forests (a), Rainforestation forests (b) and regenerating selectively logged forests (c).

The mixed-species plantations were established as a reforestation system called ‘Rainforestation Farming’ (here after Rainforestation), which was funded by the European Nature Heritage Fund in the 1990s (Figure 3.2b). The system used approximately 100 mainly native timber tree species, fruit trees, and non-timber forest products (NTFPs), which were collected from nearby mother trees and cultivated in local nurseries (Nguyen et al., 2014). Shade intolerant species were initially planted to facilitate the survival of shade tolerate apex species, at an estimated density of 5000 trees per

68 hectare (Nguyen et al., 2012). Rainforestation plantations were established at 28 locations over a period of 6 years and were approximately 1 ha in size. At the time of sampling, the Rainforestation plantations were aged between 15 and 18 years (Table 3.1).

We also sampled regenerating selectively logged forests, which were logged within the last ~20 years and are regularly used to gather NTFPs by local communities (Figure 3.2c) (Avela, 2013). Plot locations were chosen by randomly assigning a number to a compass bearing and distance within suitable regenerating selectively logged forests. Elevation and average slope angle for plots located within regenerating selectively logged forest were higher than at the other forest types (Table 3.1).

Distances to regenerating selectively logged forests (Table 3.1) and the shaded areas in Figure 3.1 were estimated using Google Earth and from discussions with local contacts from Visayas State University. Distance to regenerating selectively logged forests between monoculture and

Rainforestation sites did not differ significantly (F1, 8 = 1.56, P = 0.23), and compositionally, species pools between regenerating selectively logged forests sites, showed greater similarity to each other than to other forest sites (Figure 3.4).

3.2.2. Data Collection

Between two and four plots were established per site, depending on the size of the forest, to prevent edge effects. In total across all 15 sites, we sampled 35 plots (Table 3.1) using methods based on those set out in Herbohn et al. (2014). In the case of the Rainforestation sites, five plots were used that were originally established by Nguyen et al. (2012). In three cases, the plots established by Nguyen et al. (2012) were either; too close to the edge (within ~5m), significantly disturbed (harvested or grazed) or the location could not be confirmed to be able to replicate. In these cases, only two plots were sampled. Plots were circular with a 5 m radius (78m2) extending from the centre point.

All individuals below 2 m in height were sampled. Tree and shrub species height and either diameter at base or diameter at breast height (individuals >130cm) were measured and recorded using a 1 m measuring stick and callipers. Herbs, vines and ferns were identified and either number of individuals or cover was recorded. The life-form of each individual, province: native or exotic, and whether planted or recruited naturally (wilding) were also recorded. Planting rows were identified to determine if the individual seedling was a planting or wilding and to help identify the species. Local botanists from Visayas State University assisted with species identification. Leaf area index (LAI) of the canopy was measured as the average of three readings taken at random locations

69 within each plot using a CID Bio-Science CI-110 Plant Canopy Imager. Soil nitrogen (N) and phosphorus (P) samples were collected from three random locations within each plot and then bulked at the plot level. Samples were prepared with a single digestion method and analysed with a colorimetric determination of N using the salicylate-hypochlorite method developed by Baethgen and Alley (1989) and P using an adaptation of Murphy and Riley (1962) single solution method (Anderson and Ingram, 1989).

Data on discrete traits and taxonomic classification for each species were sourced from books, literature sources, databases and personal communication with experts from Visayas State University. These functional traits included dispersal type, fruit type, fruit size and seed size. We used a range for fruit and seed size classes consistent with Mayfield et al. (2006) and Katovai et al. (2012), with fruit size dimensions of 1 = “<2mm x <2mm”, 2 = “2-5mm x 2-5mm”, 3 = “6-15mm x 6-15mm”, 4 = “16-25mm x 16-25mm”, 5 = “26-100mm x 26-100mm”, 6 = “>100mm in any dimension”; and seed size dimensions of 1 = “0-1mm x 0-1mm”, 2 = “1.1-3mm x 1.1-3mm”, 3 = “4-8mm x 4-8mm”, 4 = “9-12mm x 9-12mm”, 5 = “>13mm in any dimension”.

3.2.3. Data analysis Most analyses were conducted using R statistical computing version 3.1.1 (R Core Team, 2013), except Figure S3.2, which was constructed in Microsoft Excel and Figure S3.4, which was constructed in SigmaPlot version 12.5. To test if understorey species diversity was adequately sampled we constructed species accumulation curves using the Specaccum function in the vegan package using two different methods: 1) the “exact” method that estimates species accumulation curves using sample-based rarefaction where plots were sampled without replacement, and is also known as Mao Tau estimates (Colwell et al., 2012); and, 2) “rarefaction” method that estimates expected species richness by sampling individuals instead of sites, and does this by calculating average number of individuals per plot using rarefy (Oksanen et al., 2015). We investigated if soil phosphorus, soil nitrogen, and leaf area index (LAI), varied depending on forest types using linear mixed effect models (LMEMs), estimated with maximum likelihood (ML). Random effects were plots nested within sites as representative of our sampling design. All LMEMs were fitted using the nLME package (Pinheiro et al., 2016). Understorey diversity was quantified using several indices: species richness, Shannon’s diversity index, Simpson’s diversity index, mean number of seedling individuals and functional traits (i.e. dispersal, fruit type, fruit size and seed size) (Magurran, 2004). To assess the relationship between

70 understorey diversity and forest type, and abiotic conditions e.g. soil phosphorus, soil nitrogen and LAI (proxy for light availability), we again used LMEMs estimated with ML and a random effects structure of plots nested within sites (Pinheiro et al., 2016). Because our study has a balanced design, we used F-statistics to assess the significance of fixed effects as explanations of variation in the response variables (Pinheiro and Bates, 2001); and the effects package (Fox, 2016) was used to graph the higher-order fixed effects in the LMEMs. We log transformed the response variable, number of individuals, due to breaches of normality. Non-metric multi-dimensional scaling (nMDS) was used to compare both understorey species richness and abundance (to compare evenness across forest types), based on Bray-Curtis similarity matrix, as it performs most satisfactorily with datasets with high numbers of 0 values (Faith et al., 1987). We also developed nested permutational ANOVAs with 1000 permutations using the BiodiversityR package (Anderson and Walsh, 2013). Trait richness and abundance data was log (x+1) transformed due to breaches of normality.

3.3. Results Overall, 2898 individuals were sampled, comprising 217 species of which 188 were native and 31 were non-native. Functional traits including, dispersal and fruit types were identified for 113 species, fruit size for 98 and seed size for 59. Forest types were adequately sampled after approximately nine plots, although less plots for monoculture forests as shown with species and individual accumulation curves (Figure S3.1).

When we compared abiotic conditions, soil phosphorus varied between forest types (F2, 12 = 6.32, P < 0.02) with monocultures having the highest soil phosphorus content and regenerating selectively logged forests the lowest (Figure S3.2a). Soil nitrogen and LAI did not vary significantly between forest types (soil N: F2, 12 = 0.57, P = 0.58, and LAI: F2, 12 = 1.85, P = 0.199; Figure S3.3b and c).

3.3.1. Comparison of the richness and abundance of recruited species under three forest types Understorey species richness differed depending on forest type with regenerating selectively logged forests having the highest species richness and monocultures the lowest (F2, 12 = 30.02, P < 0.0001; Figures 3.3 and S3.3a). Understorey species richness showed a significant negative correlation with both soil nitrogen and soil phosphorus (soil N: F1, 17 = 11.07, P < 0.004 and soil P: F1, 17 = 4.99, P < 0.04), but no significant correlation was found with LAI (Figures 3.3a-d and Table S3.1). Similar patterns were evident when the analyses were restricted to tree and shrub species (forest type: F2, 12

= 18.68, p = 0.002, soil N: F1, 17 = 6.13, P < 0.02, soil P: F1, 17 = 5.17, P < 0.04; Figures 3.3e-h,

S3.3a and Table S3.1). Herbs, ferns and graminoid richness did not differ significantly between forest types, soil phosphorus or LAI, but was significantly negatively correlation with soil nitrogen

(soil N: F1, 17 = 6.27, P < 0.03; Figures S3.3a, S3.4 and Table S3.1). Exotic seedling species richness differed significantly depending on forest type, with monocultures having the highest richness of exotics and the regenerating selectively logged forests the lowest

(forest type: F2, 12 = 23.95, P <0.001; Figures S3.5). Richness of exotic seedlings was not explained by variation in soil nitrogen, soil phosphorus or LAI (Figure S3.5 and Table S3.1). Native seedling species richness was significantly lower within the monoculture forest type (F2, 12 = 33.56, P < 0.001), and showed a significant negative correlation with soil nitrogen and soil phosphorus (soil N:

F1, 17 = 8.57, P <0.009, soil P: F1, 17 = 5.56, P =< 0.03; Figures S3.5 and Table S3.1).

Figure 3.3. Higher order fixed effects from LMEM terms where the response variable is seedling species richness of all growth forms (a-d) and limited to trees and shrubs (e-h), and how it varies with soil phosphorus (c and g), soil nitrogen (b and f) and LAI (d and h). Error bars and shaded regions represent ± standard error. * denotes significant relationships, using F-statistics (Table S3.1).

Total number of individuals (mean per plot but excluding graminoids) varied significantly depending on forest type with the Rainforestation plots having the highest number of individuals and monocultures and regenerating selectively logged forests having similar numbers of individuals (Fig. S6a). Number of individuals also showed a significant positive correlation with LAI (forest type: F2, 12 = 8.77, P = 0.005, LAI: F1, 17 = 11.23, P = 0.004; Figure S3.6 and Table S3.1).

Shannon’s diversity index (all growth forms) differed significantly between forest types (F2, 12 = 7.27, P = 0.009) with regenerating selectively logged forests showing the highest level of Shannon’s diversity and monocultures the lowest (Figure S3.7a). Shannon’s diversity in the understorey did not show a significant correlation with soil nitrogen, soil phosphorus or LAI (Figure S3.7 b-d and Table 3.1). Shannon’s diversity of tree and shrub species followed the same trends (forest type: F2, 12 = 11.02, P = 0.002; Figure S3.8 and Table 3.1). Simpson’s diversity index did not differ between forest types, soil nitrogen, soil phosphorus or LAI (Figures S3.9, S3.10 and Table S3.1).

3.3.2. Compositional differences in seedling community assemblages under three forest types Both understorey species richness and abundance were distinctly clustered within the regenerating selectively logged forest plots and monoculture forest plots respectively using an nMDS (Figure 3.4, all growth forms). The Rainforestation forest plots are less distinctly clustered, overlapping with two monoculture forest plots and one regenerating selectively logged forest plot. Understorey species richness (Figure 3.4a) and abundance (Figure 3.4b) differed significantly between forest types using a non-parametric multivariate analysis, PERMANOVA (richness:

Pseudo F2 = 6.34, P < 0.001 and abundance: Pseudo F2 = 3.67, P < 0.0001).

Figure 3.4. Non-metric multidimensional (nMDS) ordinations (using Bray-Curtis similarity index) of species richness (a, stress: 0.16) and species abundance (b, stress: 0.17) (all growth forms). Colours represent plots surveyed in monoculture forests (light grey), Rainforestation forests (dark grey) and regenerating selectively logged forests (black).

3.3.3. Compositional differences in the functional traits of community assemblages under three forest types Animals were the most common dispersers of seedlings across all forest types, followed by wind, multiple dispersal types and water (Table S3.3). Drupe’s and syconiums represented the most common fruit types. The most common fruit size classes were 5 = “26-100mm x 26-100mm” and 3 = “6-15mm x 6-15mm”, with the most common seed size classes being 3 = “4-8mm x 4-8mm” and 4 = “9-12mm x 9-12mm” (Table S3.3). Multivariate analysis of functional diversity including richness and abundance of dispersal type, fruit type, fruit size and seed size showed no clustering in nMDS ordinations (Figure S3.11). Numerically using a PERMANOVA, the richness of traits differed significantly with forest types

(dispersal: Pseudo F2 = 6.42, P = 0.001, fruit type: Pseudo F2 = 6.65, P = 0.001, fruit size: Pseudo F2

= 9.73, P = 0.001, and seed size: Pseudo F2 = 3.16, P = 0.003), and incorporating relative abundances (dispersal: Pseudo F2 = 5.29, P = 0.003, fruit type: Pseudo F2 = 4.92, P = 0.001, fruit size: Pseudo F2 = 6.59, P = 0.001, and seed size: Pseudo F2 = 2.27, P = 0.01). Regenerating selectively logged forests and Rainforestation forests showed significantly higher richness of animal-dispersed species and wind-dispersed species (when only considering native species) than the monoculture forests (F6, 96 = 5.2, P < 0.0001, natives species only: F6, 96 = 5.73, P < 0.0001; Figure S3.12 and Table S3.2). Richness of dispersal modes also decreased with increasing soil nitrogen levels (soil N: F1, 17 = 7.5, P < 0.008, Figure S3.12 and Table S3.2). When only native species were considered richness of dispersal modes decreased with increasing soil phosphorus (soil

P: F1, 17 = 4.77, P < 0.04; Figure 3.5 and Table S3.2). The abundance of animal-dispersed species was highest at the Rainforestation forests, whereas monocultures and regenerating selectively logged forest showed similar abundance levels across dispersal modes (F6, 96 = 4.96, P = 0.0002; Figure S3.13 and Table S3.2), and these trends remained consistent when only considering native species (F6, 96 = 4.5, P < 0.001; Figure S3.14 and Table

S3.2). Dispersal mode abundance was higher in plots with a higher LAI (F1, 17 = 13.54, P < 0.002, natives only: F1, 17 = 13.02, P < 0.003; Figures S3.13, S3.14 and Table S3.2). Drupes and syncarps had the highest richness in Rainforestation forests and capsules, drupes, and syncarps in monocultures; whereas, drupes had the highest richness in regenerating selectively logged forests. Monocultures overall had the lowest drupe richness (F20, 320 = 4.72, P < 0.0001; Figure S3.15 and Table S3.2). Richness of fruit types also decreased significantly with increasing soil phosphorus (F1, 17 = 5.98, P < 0.03; Figure S3.15b and Table S3.2). Regenerating selectively logged forests showed similar abundance levels across fruit types. Capsules were the most abundant fruit type in monocultures and drupes and syncarps in Rainforestation forests (F10, 320 = 7.64, P < 0.0001; Figure S3.16d and Table S3.2). Abundance of fruit types also increased significantly with

LAI (F1, 17 = 16.35, P < 0.001; Figure S3.16c and Table S3.2). Monocultures had the highest richness and abundance of the largest fruit size class of six (>100mm in any dimension); whereas both Rainforestation and regenerating selectively logged forests showed similar levels of richness between fruit size classes (F10, 160 = 4.72, P < 0.0001, soil N: F1, 17 = 6.48,

P = 0.02; Figure S3.17), and these trends held when considering fruit type abundance (F10, 160 = 2.9, P = 0.003; Figure S3.18). As fruit size class richness increased soil nitrogen levels significantly decreased (F1, 17 = 6.48, P = 0.02; Figure S3.17a). The abundance of fruit size classes was positively correlated with LAI (F1, 17 = 15.41, P = 0.001; Figure S3.18c).

Regenerating selectively logged forests showed similar richness and abundance levels across the different seed class sizes (Figures S3.19d and S3.20d). Rainforestation forests showed a higher richness of seed class size 3 (8mm x 4-8mm), (richness seed size: F8, 128 = 3.01, P = 0.004; Figures S3.19d). Monoculture forest had higher abundance of seed class size 4 (9-12mm x 9-12mm),

(abundance seed size: F8, 128 = 2.51, P = 0.015; Figure S3.20d and Table S3.2). The abundance of seed class sizes also varied negatively with soil phosphorus (F1, 17 = 6.12, P = 0.024; Figure S3.20b) and positively with LAI (F1, 17 = 15.83, P = 0.001; Figure S3.20c).

Figure 3.5. Higher order fixed effects from LMEM terms where the response variable is seedling richness of dispersal modes (native species), and how it varies with soil nitrogen (a), soil phosphorus (b), and LAI (c). Error bars and shaded regions represent ± standard error. * denotes significant relationships, using F-statistics (Table S3.1).

3.4. Discussion Human-modified habitats are increasingly being recognised as important for conservation purposes, particularly in the tropics where only a small amount of land is protected (Barlow et al., 2007). We expected to find that the regenerating selectively logged forests and Rainforestation forests would have considerably higher seedling species richness and functional trait diversity because of their diverse canopy tree species compositions. This pattern was weaker than expected with monoculture forests also recruiting a relatively diverse number of seedling species, although regenerating selectively logged forests did recruit significantly higher amounts of understorey diversity (species richness and Shannon’s Diversity Index) than monoculture forests. Overall, we found that despite compositional differences, monocultures in fragmented landscapes can recruit some understorey diversity. We found that the understories of the different forest types shared some common species, e.g., Glochidion album Boerl., Fagreae racemose Jack, and Ficus septica Lour. We also found consistent trends with seedling diversity, functional diversity and canopy species diversity being negatively correlated with both soil nitrogen and phosphorus levels. The mechanism responsible for this relationship is unclear. However, it may be the result of differences in pre-planting soil treatments including fertiliser additions between forest types, litter nutrient differences depending on canopy composition and higher diversity forests being more efficient at nitrogen and phosphorus cycling compared with less diverse systems (Richards and Schmidt, 2010).

3.4.1. Seedling species diversity and composition beneath each forest type The choice of the overstorey species impacts heavily on understorey seedling recruitment. As such, the characteristics of the species selected should be carefully considered if planting is designed to achieve biodiversity and production objectives (Le et al., 2014). The species used within our monoculture forests; Swietenia macrophylla, is a wind-dispersed, shade intolerant but long-lived pioneer, with fruit that is of limited attractiveness to animal-dispersers, however, this species can provide canopy refuge (Grogan et al., 2014, Slocum and Horvitz, 2000). Elsewhere, Parrotta (1992) found the highest recruitment beneath monocultures of Leucaena leucocephala, possibly because of its overall structure and leaf litter traits (Slocum and Horvitz, 2000), but did not recommend this species for planting because of invasion potential. Overstorey fruit bearing species (e.g. Ficus) can also accelerate understorey development within plantations, even when located long distances from seed sources (Peña-Domene et al., 2013). Other studies have found little relationship between overstorey composition and recruited seedling richness and number (Farwig et al., 2009).

3.4.2. Species and functional traits favoured across forest types The family Moraceae maintained similar levels of diversity across forest types. These were species belonging to the genera’s Artocarpus and Ficus, which were described by the comparable richness of the syncarpous fruit type between forest types (Laliberté et al., 2010). Other studies have found a similar prevalence of this functional group in particular genera’s Artocarpus (Quimio, 1999) and Ficus (Ingle, 2003). Due to the associations between gape width, body mass and gut passage times, influences of the composition and abundance of frugivorous passerines on plant dispersal and subsequent recruitment, have been reported throughout the tropics (Galetti et al., 2013, Corlett, 2009, Corlett, 2011a). Common bird species within the Philippines study region include the chestnut munia (Lonchura atricapilla atricapilla) and Philippine bulbul (Hypsipetes philippinus) (Ingle, 2003, Vallejo et al., 2008). These birds are habitat generalists and likely account for a significant proportion of animal-dispersal, particularly of smaller-seeded species across human-modified landscapes (Markl et al., 2012, Corlett, 2011a). In the Philippines subsistence hunting of animals such as wild pigs (Sus cebifrons) (Lacuna- Richman, 2002) and various bird species (Shively, 1997) is an important income and food source for the rural poor. Loss of these animals negatively impacts dispersal systems across all forest types. Dispersal of later successional species has been shown to be more specialized and often involve larger forest interior animal species within the Philippines (Hamann and Curio, 1999). Therefore, the regenerating selectively logged forests cannot be used as an ecological base line regarding natural dispersal regimes, but rather considered a less impacted system. Several large-seeded animal-dispersed species were only found in the understorey of the regenerating selectively logged forests (e.g., Canarium luzonicum Miq. and Lithocarpus llanosii Rehder); this number may increase if more pristine forests were available for comparison. Growing evidence suggests that human activities have a strong legacy on forest structure and composition within the tropics (Bhagwat et al., 2008). Monoculture understories had more species belonging to the largest fruit size class (6 = “>100mm”). These species included Mangifera indica Blume (mango), Theobroma cacao L. (cacao), Citrus maxima (Burm.) Merr. (pomelo), Artocarpus altilis (Parkinson) Fosberg (bread fruit), Chrysophyllum cainito L. (star apple), Psidium guajava L. (guava) and Swietenia macrophylla. People likely dispersed these species into plantations and the vast majority of the individuals observed were wildings. This suggests, that exotic species can fulfil a ‘utilitarian role’ within small-scale forestry by regenerating and producing NTFPs once the canopy species has been selectively harvested, and that mahogany monocultures can provide conditions conducive for seedling recruitment.

3.4.3. Species and traits limited across forest types Abiotic dispersal (e.g. wind, water, and gravity) in the tropics is less common than biotic dispersal and is estimated at ~20% of tree species (Beckman and Rogers, 2013). Wind-dispersed tree species often occur in exposed microhabitats such as emergent layers (Corlett 2011). We found wind- dispersed richness in the understorey of the monoculture forest types was threefold less than in the regenerating selectively logged forests. Only one individual of a wind-dispersed native species (Alstonia macrophylla Wall.) was found within the monoculture forests. The Dipterocarpaceae family is abiotically dispersed and ecologically dominant, particularly within Asia (Oshima et al., 2015). This family is also economically important, as it accounted for a quarter of the global consumption of tropical hardwood between 2006 and 2007 (Kettle, 2010). Dipterocarps are known for their irregular mast fruiting events; therefore, low populations of mature trees and irregular fruiting are likely impediments restricting colonization irrespective of seed size (Oshima et al., 2015). The high value of these species has led to high rates of logging and resulted in a low number of mother trees. Monocultures did not recruit any Dipterocarpaceae species in the understorey, but four species were found in the Rainforestation forest types and five species in the regenerating selectively logged forest types. Dipterocarpaceae species have been estimated to have a routine dispersal distance of ~100 m, with some experimental data suggesting the majority of seeds are dispersed less than 40 m from the mother tree (Corlett, 2009, Ingle, 2003). In the present study, possible seed sources occurred at distances between 160 m to 5.5 km from the plantation locations (Table 3.1). These distances are likely too large for wind-dispersed tree species to recruit into the understorey of plantations, at a time scale relevant for small-scale reforestation projects.

3.5. Conclusion Seedling recruitment beneath small-scale community and smallholder plantations is an important factor in determining the longer-term sustainability and success of reforestation efforts (Le et al., 2014). This study highlights the value of reforestation in general, providing that exotic monocultures can recruit diverse understories that contain some species valued by the forest dependent local communities. However, from a conservation perspective, it is a concern that wind- dispersed seedlings are absent, as this functional trait included ecologically important tree species such as Octomeles sumatrana, Cratoxylum sumatranum (Jack) Blume, etc. and several species of Dipterocarpaceae. Our results suggest that more active means of facilitating the establishment of species with this functional trait is required to sustain recruitment of future emergent trees in diverse tropical forests.

References Anderson, M. J. & Walsh, D. C. I. 2013. PERMANOVA, ANOSIM, and the Mantel test in the face of heterogeneous dispersions: What null hypothesis are you testing? Ecological Monographs, 83, 557-574. Anderson, S. E. & Ingram, J. S. I. 1989. Tropical Soil Biology and Fertility: A handbook of Methods. p. 171. C.A.B. International, Aberystwyth. Avela, M. 25/02/2013 2013. RE: Personal Communication. Baethgen, W. E. & Alley, M. M. 1989. A manual colorimetric procedure for measuring ammonium nitrogen in soil and plant Kjeldahl digests. Communications in Soil Science and Plant Analysis, 20, 961-969. Bank, T. W. 2001. Recommended revisions to OP 4.36: Proposals for discussion. . Washington, D.C. : The World Bank. Barlow, J., Gardner, T. A., Araujo, I. S., Ávila-Pires, T. C., Bonaldo, A. B., Costa, J. E., Esposito, M. C., Ferreira, L. V., Hawes, J., Hernandez, M. I. M., Hoogmoed, M. S., Leite, R. N., Lo- Man-Hung, N. F., Malcolm, J. R., Martins, M. B., Mestre, L. a. M., Miranda-Santos, R., Nunes-Gutjahr, A. L., Overal, W. L., Parry, L., Peters, S. L., Ribeiro-Junior, M. A., Da Silva, M. N. F., Da Silva Motta, C. & Peres, C. A. 2007. Quantifying the biodiversity value of tropical primary, secondary, and plantation forests. Proceedings of the National Academy of Sciences, 104, 18555-18560. Beckman, N. G. & Rogers, H. S. 2013. Consequences of Seed Dispersal for Plant Recruitment in Tropical Forests: Interactions Within the Seedscape. Biotropica, 45, 666-681. Bhagwat, S. A., Willis, K. J., Birks, H. J. B. & Whittaker, R. J. 2008. Agroforestry: a refuge for tropical biodiversity? Trends in Ecology & Evolution, 23, 261-267. Brown, S. & Lugo, A. E. 1990. Tropical Secondary Forests. Journal of Tropical Ecology, 6, 1-32. Bryant, D., Nielson, D. & Tangley, L. 1997. The last frontier forests. Issues in Science and Technology, 14, 85-87. Colwell, R. K., Chao, A., Gotelli, N. J., Lin, S.-Y., Mao, C. X., Chazdon, R. L. & Longino, J. T. 2012. Models and estimators linking individual-based and sample-based rarefaction, extrapolation and comparison of assemblages. Journal of Plant Ecology, 5, 3-21. Corlett, R. 2011. Seed dispersal in Hong Kong, China: past, present and possible futures. Intergrative Zoology, 6, 97-109. Corlett, R. T. 2009. Seed dispersal distances and plant migration potential in tropical East Asia Biotropica 41: 592–598.

Dirzo, R., Mendoza, E. & Ortíz, P. 2007. Size-Related Differential Seed Predation in a Heavily Defaunated Neotropical Rain Forest. Biotropica, 39, 355-362. Faith, D. P., Minchin, P. R. & Belbin, L. 1987. Compositional dissimilarity as a robust measure of ecological distance. Vegetatio, 69, 57-68. Farwig, N., Sajita, N. & Böhning-Gaese, K. 2009. High seedling recruitment of indigenous tree species in forest plantations in Kakamega Forest, western Kenya. Forest Ecology and Management, 257, 143-150. Firn, J., Erskine, P. D. & Lamb, D. 2007. Woody Species Diversity Influences Productivity and Soil Nutrient Availability in Tropical Plantations. Oecologia, 154, 521-533. Fox, J. 2016. Effect Displays for Linear, Generalized Linear, and Other Models. . http://www.r- project.org, http://socserv.socsci.mcmaster.ca/jfox/. Galetti, M., Guevara, R., Côrtes, M. C., Fadini, R., Von Matter, S., Leite, A. B., Labecca, F., Ribeiro, T., Carvalho, C. S., Collevatti, R. G., Pires, M. M., Guimarães, P. R., Brancalion, P. H., Ribeiro, M. C. & Jordano, P. 2013. Functional Extinction of Birds Drives Rapid Evolutionary Changes in Seed Size. Science, 340, 1086-1090. Garrity, D. P., Soekardi, M., Noordwijk, M., Cruz, R., Pathak, P. S., Gunasena, H. P. M., So, N., Huijun, G. & Majid, N. M. 1996. The Imperata grasslands of tropical Asia: area, distribution, and typology. Agroforestry Systems, 36, 3-29. Grogan, J., Landis, R. M., Free, C. M., Schulze, M. D., Lentini, M. & Ashton, M. S. 2014. Big-leaf mahogany Swietenia macrophylla population dynamics and implications for sustainable management. Journal of Applied Ecology, 51, 664-674. Hamann, A. & Curio, E. 1999. Interactions among Frugivores and Fleshy Fruit Trees in a Philippine Submontane Rainforest. Conservation Biology, 13, 766-773. Hansen, M. C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. A., Tyukavina, A., Thau, D., Stehman, S. V., Goetz, S. J., Loveland, T. R., Kommareddy, A., Egorov, A., Chini, L., Justice, C. O. & Townshend, J. R. G. 2013. High-Resolution Global Maps of 21st-Century Forest Cover Change. Science, 342, 850-853. Harrison, R. D., Tan, S., Plotkin, J. B., Slik, F., Detto, M., Brenes, T., Itoh, A. & Davies, S. J. 2013. Consequences of defaunation for a tropical tree community. Ecology Letters, 16, 687-694. Harrison, S., Herbohn, J. & Niskanen, A. 2002. Non-industrial, smallholder, small-scale and family forestry: What’s in a name? Small-scale Forest Economics, Management and Policy, 1, 1- 11. Herbohn, J. L., Vanclay, J., Ngyuen, H., Le, H. D., Baynes, J., Harrison, S. R., Cedamon, E., Smith, C., Firn, J., Gregorio, N. O., Mangaoang, E. & Lamarre, E. 2014. Inventory Procedures for

Smallholder and Community Woodlots in the Philippines: Methods, Initial Findings and Insights. Small-scale Forestry, 13, 79-100. Ingle, N. 2003. Seed dispersal by wind, birds, and bats between Philippine montane rainforest and successional vegetation. Oecologia, 134, 251-261. Katovai, E., Burley, A. L. & Mayfield, M. M. 2012. Understory plant species and functional diversity in the degraded wet tropical forests of Kolombangara Island, Solomon Islands. biological conservation, 145, 214-224. Kettle, C. 2010. Ecological considerations for using dipterocarps for restoration of lowland rainforest in Southeast Asia. Biodiversity and Conservation, 19, 1137-1151. Lacuna-Richman, C. 2002. The socioeconomic significance of subsistence non-wood forest products in Leyte, Philippines. Environmental Conservation, 29, 253-262. Laliberté, E., Wells, J. A., Declerck, F., Metcalfe, D. J., Catterall, C. P., Queiroz, C., Aubin, I., Bonser, S. P., Ding, Y., Fraterrigo, J. M., Mcnamara, S., Morgan, J. W., Merlos, D. S., Vesk, P. A. & Mayfield, M. M. 2010. Land-use intensification reduces functional redundancy and response diversity in plant communities. Ecology Letters, 13, 76-86. Lamb, D. 1998. Large-scale Ecological Restoration of Degraded Tropical Forest Lands: The Potential Role of Timber Plantations. Restoration Ecology, 6, 271-279. Lamb, D., Erskine, P. D. & Parrotta, J. A. 2005. Restoration of Degraded Tropical Forest Landscapes. Science, 310, 1628-1632. Laurance, W. F. 1999. Reflections on the tropical deforestation crisis. Biological Conservation, 91, 109-117. Le, H. D., Smith, C. & Herbohn, J. 2014. What drives the success of reforestation projects in tropical developing countries? The case of the Philippines. Global Environmental Change, 24, 334-348. Le, H. D., Smith, C., Herbohn, J. & Harrison, S. 2012. More than just trees: Assessing reforestation success in tropical developing countries. Journal of Rural Studies, 28, 5-19. Lindenmayer, D. B. & Hobbs, R. J. 2004. Fauna conservation in Australian plantation forests – a review. Biological Conservation, 119, 151-168. Magurran, A. E. 2004. Measuring Biological Diversity Blackwell Publishing Markl, J. S., Schleuning, M., Michel Forgit, P., Lambert, J. E., Traveset, A., Wright, S. J. & Bohning-Gease, K. 2012. Meta-Analysis of the Effects of Human Disturbance on Seed Dispersal by Animals. Conservation Biology 26, 1072-1081.

Mayfield, M. M., Ackerly, D. & Daily, G. C. 2006. The diversity and conservation of plant reproductive and dispersal functional traits in human-dominated tropical landscapes. Journal of Ecology, 94, 522-536. Mcconkey, K. R., Prasad, S., Corlett, R. T., Campos-Arceiz, A., Brodie, J. F., Rogers, H. & Santamaria, L. 2012. Seed dispersal in changing landscapes. Biological Conservation, 146, 1-13. Meyfroidt, P. & Lambin, E. F. 2011. Global Forest Transition: Prospects for an End to Deforestation. Annual Review of Environment and Resources, 36, 343-371. Milan, P. P., Ceniza, M. J. C., Asio, V. B., Bulayog, S. B. & Napiza, M. D. 2004. Evaluation of Silvicultural Management, Ecological Changes and Market Study of Products of Existing Rainforestation demo and cooperation farms. Institute of tropical ecology terminal report Murphy, J. & Riley, J. P. 1962. A modified single solution method for the determination of phosphate in natural waters. Analytica Chimica Acta, 27, 31-36. Naeem, S. & Wright, J. P. 2003. Disentangling biodiversity effects on ecosystem functioning: deriving solutions to a seemingly insurmountable problem. Ecology Letters, 6, 567-579. Nguyen, H., Firn, J., Lamb, D. & Herbohn, J. 2014. Wood density: A tool to find complementary species for the design of mixed species plantations. Forest Ecology and Management, 334, 106-113. Nguyen, H., Herbohn, J., Firn, J. & Lamb, D. 2012. Biodiversity–productivity relationships in small-scale mixed-species plantations using native species in Leyte province, Philippines. Forest Ecology and Management, 274, 81-90. Nunez-Iturri, G., Olsson, O. & Howe, H. F. 2008. Hunting reduces recruitment of primate-dispersed trees in Amazonian Peru. Biological Conservation, 141, 1536-1546. Oksanen, J. F., Blanchet, G., Kindt, R., Legendre, P., Peter, R., Minchin, R., O'hara, B., Simpson, G. L., Solymos, P. M., Stevens, H. H. & Wagner, H. 2015. vegan: Community Ecology Package. R package version [Online]. http://CRAN.R-project.org/package=vegan. Oshima, C., Tokumoto, Y. & Nakagawa, M. 2015. Biotic and abiotic drivers of dipterocarp seedling survival following mast fruiting in Malaysian Borneo. Journal of Tropical Ecology, 31, 129- 137. Parrotta, J. A. 1992. The role of plantation forests in rehabilitating degraded tropical ecosystems. Agriculture, Ecosystems & Environment, 41, 115-133. Parrotta, J. A. 1995. Influence of Overstory Composition on Understory Colonization by Native Species in Plantations on a Degraded Tropical Site. Journal of Vegetation Science, 6, 627- 636.

Peña-Domene, D. L. M., Martínez-Garza, C. & Howe, H. F. 2013. Early recruitment dynamics in tropical restoration. Ecological Applications, 23, 1124-1134. Pinheiro, J. & Bates, D. 2001. Mixed-Effects Models in S and S-PLUS. Technometrics, 43, 113- 114. Pinheiro, J., Bates, D., Debroy, S., Sarkar, D. & Team, R. C. 2016. nlme: Linear and Nonlinear Mixed Effects Models [Online]. http://CRAN.R-project.org/package=nlme. Quimio, J. M. 1999. Plant diversity in coconut plantation of Baybay, Leyte, Philippines., Leyte, Philippines, Proceedings of the national workshop on local knowledge and biodiversity conservation in forestry practice and education, FARMI training hall, Leyte, Philippines. R Core Team. 2013. R: A language and environment for statistical computing. [Online]. R Foundation for Statistical Computing, Vienna, Austria. Available: URL http://www.R-project.org/. Richards, A. E. & Schmidt, S. 2010. Complementary resource use by tree species in a rain forest tree plantation. Ecological Applications, 20, 1237-1254. Scherr, S. J., White, A. & Kaimowitz, A. 2004. A new Agenda for Forest Conservation and Poverty Reduction: Making Markets Work for low-income producers. Forest Trends. Washington DC. Shively, G. E. 1997. Poverty, technology, and wildlife hunting in Palawan. Environmental Conservation, 24, 57-63. Simonetti, J. A., Grez, A. A. & Estades, C. F. 2013. Providing Habitat for Native Mammals through Understory Enhancement in Forestry Plantations. Conservation Biology, 27, 1117-1121. Slocum, M. & Horvitz, C. 2000. Seed arrival under different genera of trees in a neotropical pasture. Plant Ecology, 149, 51-62. Vallejo, B., Aloyab, A., Ong, P., Tamino, A. & Villasper, J. 2008. Spatial Patterns of Bird Diversity and Abundance in an Urban Tropical Landscape: The University of the Philippines (UP) Diliman Campus. Wright, S. J., Stoner, K. E., Beckman, N., Corlett, R. T., Dirzo, R., Muller-Landau, H. C., Nuñez- Lturri, G., Peres, C. A. & Wang, B. C. 2007. The Plight of Large Animals in Tropical Forests and the Consequences for Plant Regeneration. Biotropica, 39, 289-291. Zanne, A. E. & Chapman, C. A. 2001. Expediting Reforestation in Tropical Grasslands: Distance and Isolation from Seed Sources in Plantations. Ecological Applications, 11, 1610-1621.

Supplementary material Table S3.1. Analyses of deviance table for linear mixed effect models examining change in richness and diversity within different life forms depending on forest type, soil phosphorus, soil nitrogen and leaf area index. Random effects were plots nested within sites. Species diversity Fixed DF (num./den.) F value P value effects richness forest type 2/12 30.02 < 0.001 richness soil N 1/17 11.07 0.004 richness soil P 1/17 4.99 0.039 richness LAI 1/17 0.98 0.336 richness (trees and shrubs) forest type 2/12 18.68 0.002 richness (trees and shrubs) soil N 1/17 6.13 0.024 richness (trees and shrubs) soil P 1/17 5.17 0.036 richness (trees and shrubs) LAI 1/17 0.82 0.377 richness (herbs, ferns and graminoids) forest type 2/12 2.4 0.133 richness (herbs, ferns and graminoids) soil N 1/17 6.27 0.023 richness (herbs, ferns and graminoids) soil P 1/17 0.02 0.9 richness (herbs, ferns and graminoids) LAI 1/17 0.0 0.998 richness (exotics) forest type 2/12 23.95 0.001 richness (exotics) soil N 1/17 0.0 0.978 richness (exotics) soil P 1/17 2.64 0.123 richness (exotics) LAI 1/17 2.19 0.157 richness (natives) forest type 2/12 33.56 < 0.001 richness (natives) soil N 1/17 8.57 0.0094 richness (natives) soil P 1/17 5.56 0.031 richness (natives) LAI 1/17 1.35 0.262 number of individuals forest type 2/12 8.77 0.005 number of individuals soil N 1/17 2.21 0.155 number of individuals soil P 1/17 2.08 0.168 number of individuals LAI 1/17 11.23 0.004 Shannon’s diversity forest type 2/12 7.27 0.009 Shannon’s diversity soil N 1/17 4.14 0.058 Shannon’s diversity soil P 1/17 0.8 0.383 Shannon’s diversity LAI 1/17 0.11 0.745 Shannon’s diversity (trees and shrubs) forest type 2/12 11.02 0.002 Shannon’s diversity (trees and shrubs) soil N 1/17 2.96 0.103 Shannon’s diversity (trees and shrubs) soil P 1/17 1.93 0.182 Shannon’s diversity (trees and shrubs) LAI 1/17 0.1 0.752 Simpson’s diversity forest type 2/12 1.36 0.29 Simpson’s diversity soil N 1/17 1.65 0.216 Simpson’s diversity soil P 1/17 0.033 0.858 Simpson’s diversity LAI 1/17 0.03 0.857 Simpson’s diversity (tree and shrubs) forest type 2/12 2.21 0.153 Simpson’s diversity (tree and shrubs) soil N 1/17 0.56 0.46 Simpson’s diversity (tree and shrubs) soil P 1/17 1.26 0.278 Simpson’s diversity (tree and shrubs) LAI 1/17 0.37 0.553

Table S3.2. Analyses of deviance table for linear mixed effect models examining change in richness and abundance depending on functional groups, forest type, soil phosphorus, soil nitrogen and leaf area index. Random effects were plots nested within sites. Functional diversity Fixed effects DF (num./den.) F value P value

Dispersal richness forest type 2/12 7.5 0.0077 dispersal 3/96 138.8 <0.0001 soil N 1/17 7.5 0.0077 soil P 1/17 4.24 0.055 forest type:dispersal 6/96 5.2 <0.0001 LAI 1/17 0.33 0.574 abundance forest type 2/12 3.07 0.084 dispersal 3/96 21.27 <0.0001 soil N 1/17 0.22 0.643 soil P 1/17 3.35 0.085 forest type:dispersal 6/96 4.96 0.0002 LAI 1/17 13.54 0.0019 Dispersal richness forest type 2/12 10.36 0.0024 (natives) dispersal 3/96 94.15 <0.0001 soil N 1/17 3.53 0.078 soil P 1/17 4.77 0.04 forest type:dispersal 6/96 5.73 <0.0001 LAI 1/17 0.58 0.455 abundance forest type 2/12 6.14 0.015 dispersal 3/96 22.97 <0.0001 soil N 1/17 0.15 0.701 soil P 1/17 1.36 0.259 forest type:dispersal 6/96 4.5 0.0005 LAI 1/17 13.02 0.0022

Fruit types richness forest type 2/12 10.43 0.0024 fruit type 3/96 138.8 <0.0001 soil N 1/17 4.29 0.054 soil P 1/17 5.98 0.026 forest type:fruit type 6/96 5.2 <0.0001 LAI 1/17 0.42 0.5239 abundance forest type 2/12 3.71 0.056 fruit type 10/320 7.64 <0.0001 soil N 1/17 0.19 0.672 soil P 1/17 3.87 0.0655 forest type:fruit type 20/320 3.45 <0.0001 LAI 1/17 16.35 0.0008 Fruit sizes richness forest type 2/12 6.18 0.014 fruit size 5/160 36.48 <0.0001 soil N 1/17 6.48 0.021 soil P 1/17 2.74 0.116 forest type:fruit size 10/160 4.72 0.0001 LAI 1/17 0.01 0.908 abundance forest type 2/12 3.57 0.061 fruit size 5/160 5.7 0.0001

soil N 1/17 0.49 0.4946 soil P 1/17 3.52 0.078 forest type:fruit size 10/160 2.87 0.0026 LAI 1/17 15.41 0.0011 Seed sizes richness forest type 2/12 5.04 0.0258 seed size 4/128 6.49 0.0001 soil N 1/17 1.2 0.2886 soil P 1/17 2.24 0.1525 forest type:seed size 8/128 3.01 0.0040 LAI 1/17 0.32 0.5778 abundance forest type 2/12 0.61 0.557 seed size 4/128 4.87 0.0011 soil N 1/17 0.01 0.9160 soil P 1/17 6.12 0.0242 forest type:seed size 8/128 2.51 0.0145 LAI 1/17 15.83 0.0010

Table S3.3. Proportion of trait types found within each forest type. Trait Trait state Monoculture Rainforestation Regenerating Dispersal (%) Animal 67 72 71 Wind 27 21 24 Water 0 3 3 Multiple 6 4 2

Dispersal (% Animal 91 77 68 natives) Wind 1 16 26 Water 0 2 3 Multiple 8 6 2 Fruit types (%) Achene 4 4 2 Berry 13 14 13 Capsule 21 14 18 Drupe 29 28 26 Follicle 4 5 6 Legume 8 12 6 Nutlet 0 5 9 Samara 0 0 2 Cone 0 0 2 Syncarp 21 18 18 Fruit sizes (%) <2mmx<2mm 0 2 0 2-5mmx2-5mm 0 11 6 6-15mmx6-15mm 17 29 32

16-25mmx16-25mm 13 20 22 26-100mmx26-100mm 29 21 30 >100mm 42 18 11 Seed sizes (%) 0-1mmx0-1mm 13 13 24 1.1-3mmx1.1-3mm 13 7 17 4-8mmx4-8mm 38 47 17 9-12mmx9-12mm 25 20 17 >13mm 13 13 24

Figure S3.1. Expected mean species richness for accumulated plots within each forest type (a). Accumulated individuals within each forest type, using “Rarefaction” method (b).

Figure S3.2. Soil phosphorus was significantly higher in monoculture forest types (a), soil nitrogen and LAI were not significantly higher in regenerating selectively logged forests (b) and (c).

All growth form Trees and shrubs Herbs, ferns and graminoids

Monocultur Rainforestatio Regeneratin Monocultur Rainforestatio Regeneratin e n g e n g Forest type

Figure S3.3. Comparison of species richness (a), number of individuals (b), Shannon’s diversity index (c) and Simpson’s diversity index (d), for the different growth forms (a, c and d) between monoculture, Rainforestation and regenerating selectively logged forests types. Error bars represent 95% confidence intervals.

Figure S3.4. Higher order fixed effects from LMEM terms where the response variable is seedling species richness of herbs, ferns and graminoids recruited into the different forest types and how it varies with soil phosphorus (c ), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.1.

Figure S3.5. Higher order fixed effects from LMEM terms where the response variables are number of exotics (a-d) and native (e-h) seedling species recruited into the different forest types and how exotic and native richness varies depending on soil phosphorus (c and g), soil nitrogen (b and f) and LAI (d and h). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.1.

Figure S3.6. Higher order fixed effects from LMEM terms where the response variable is number of individual seedlings recruited into the different forest types and how it varies with soil phosphorus (c), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.1.

Figure S3.7. Higher order fixed effects from LMEM terms where the response variable is Shannon’s diversity (all growth forms) of seedlings recruited into the different forest types and how it varies with soil phosphorus (c), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.1.

Figure S3.8. Higher order fixed effects from LMEM terms where the response variable is Shannon’s diversity of tree and shrub seedling species recruited into the different forest types and how it varies with soil phosphorus (c), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.1.

Figure S3.9. Higher order fixed effects from LMEM terms where the response variable is Simpson’s diversity (all growth forms) of seedlings recruited into the different forest types and how it varies with soil phosphorus (c), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.1.

Figure S3.10. Higher order fixed effects from LMEM terms where the response variable is Simpson’s diversity of tree and shrub seedling species recruited into the different forest types and how it varies with soil phosphorus (c ), soil nitrogen (b) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.1.

Figure S3.11. Non-metric multidimensional (nMDS) ordinations (based on Bray-Curtis similarity index) of functional richness for dispersal (a, stress: 0.15), fruit type (c, stress: 0.16), fruit size (e, stress: 0.15) and seed size (g, stress: 0.2). Functional abundance of dispersal (b, stress: 0.12), fruit type (d, stress: 0.17), fruit size (f, stress: 0.16) and seed size (h, stress: 0.22) of all growth forms. Colours represent monoculture forests (light grey), Rainforestation forests (dark grey) and regenerating selectively logged forests (black).

Figure S3.12. Higher order fixed effects from LMEM terms where the response variable is richness of dispersal modes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.2.

100

Figure S3.13. Higher order fixed effects from LMEM terms where the response variable is abundance of dispersal modes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics, see Table S3.2.

101

Figure S3.14. Higher order fixed effects from LMEM terms where the response variable is abundance of dispersal modes, when restricted to native species recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F- statistics, see Table S3.2.

102

Figure S3.15. Higher order fixed effects from LMEM terms where the response variable is richness of fruit types recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and fruit type code corresponds to achene (a), cone (b), syncarp (c), berry (d), capsule (e), drupe (f), follicle (g), legume (h), nutlet (i) and samara (j).

103

Figure S3.16. Higher order fixed effects from LMEM terms where the response variable is abundance of fruit types recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and fruit type code corresponds to achene (a), cone (b), syncarp (c), berry (d), capsule (e), drupe (f), follicle (g), legume (h), nutlet (i) and samara (j).

104

Figure S3.17. Higher order fixed effects from LMEM terms where the response variable is richness of fruit sizes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and fruit size dimensions correspond to 1 = “<2mm x <2mm”, 2 = “2-5mm x 2-5mm”, 3 = “6-15mm x 6-15mm”, 4 = “16-25mm x 16-25mm”, 5 = “26-100mm x 26-100mm” and 6 = “>100mm in any dimension”.

105

Figure S3.18. Higher order fixed effects from LMEM terms where the response variable is abundance of fruit sizes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and fruit size dimensions correspond to 1 = “<2mm x <2mm”, 2 = “2-5mm x 2-5mm”, 3 = “6-15mm x 6-15mm”, 4 = “16- 25mm x 16-25mm”, 5 = “26-100mm x 26-100mm” and 6 = “>100mm in any dimension”.

106

Figure S3.19. Higher order fixed effects from LMEM terms where the response variable is richness of seed sizes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and seed size dimensions correspond to 1 = “0-1mm x 0-1mm”, 2 = “1.1-3mm x 1.1-3mm”, 3 = “4-8mm x 4-8mm”, 4 = “9-12mm x 9-12mm” and 5 = “>13mm in any dimension”.

107

Figure S3.20. Higher order fixed effects from LMEM terms where the response variable is abundance of seed sizes recruited into the different forest types and how it varies with soil phosphorus (b), soil nitrogen (a) and LAI (d). Error bars and shaded regions represent ± standard error. * denotes significant relationships were found using F-statistics and seed size dimensions correspond to 1 = “0-1mm x 0-1mm”, 2 = “1.1-3mm x 1.1-3mm”, 3 = “4-8mm x 4-8mm”, 4 = “9- 12mm x 9-12mm” and 5 = “>13mm in any dimension”.

108

Chapter 4. Reforestation methods influence community assembly in tropical forest understoreys: insights using functional traits and phylogenetic diversity

4.1. Introduction Understanding how and why some plant species coexist is essential information when attempting to restore degraded plant communities, particularly in areas once occupied by highly diverse ecosystems, like tropical forests (Funk et al., 2016). Finding the optimal restoration strategies in tropical forests is not straightforward, with strategies ranging from the planting of one or just a few species at a site to initiate understorey recruitment of native biodiversity, to the planting of highly diverse ecological restoration plantings. In highly modified tropical landscapes, a combination of monocultures and mixed-species plantations, high diversity restoration plantings, and natural regeneration of secondary forests have all been found to provide some benefits for biodiversity and ecosystem function (Barlow et al., 2007, Wills et al., 2016). What is yet to be understood is how the restoration strategy used impacts on the mechanisms that drive plant community assembly. Unpacking the mechanisms acting on the recruitment of plant species in the understories of, for example, monocultures versus more diverse restoration focused strategies may provide a mechanistic underpinning for future attempts at regrowing tropical forests (Cadotte et al., 2017).

Environmental filtering and interactions between species are key mechanisms acting on the assembly of plant communities (Silvertown, 2004). Environmental conditions, such as rainfall, soil nutrient and light availability, provide conditions that potentially favour some species over others, essentially acting to select for species with traits that are suited for the abiotic conditions, which leads to trait similarity. Whereas biotic interactions such as competition can limit how similar traits are between species as they will be in competition for the same niche space (Grubb, 1977). The relative importance of environmental filtering and competitive interactions can potentially be teased apart by measuring the phylogenetic and functional structure of plant communities (Ricklefs, 2008, Baraloto et al., 2012a).

If functional traits are conserved amongst related species or inherited over evolutionary timescales, a community’s phylogenetic and functional structure may be found to be either clustered (i.e., community members are more closely related and display a higher similarity in trait values than would be expected by chance alone) or overdispersed (i.e., community members are more distantly related and display a higher dissimilarity than would be expected by chance alone) (Webb, 2000, Kraft and Ackerly, 2010, Cavender-Bares et al., 2004). Niche theory predicts that clustered communities are explained by environmental filtering, as a reflection of species adaptations to their

109 shared environment (Cornwell et al., 2006, Webb et al., 2002, Pausas and Verdú, 2010); and overdispersed communities are explained by competitive interactions between species that have similar resource requirements and growing habits, resulting in competitive exclusion and greater niche differentiation (Pausas and Verdú, 2010). However, competitive interactions may not simply lead to overdispersion in either traits or phylogenetic structure, because of the opposing effects of niche differences (altering the balance between intra- and inter-specific competition, and tending to stabilise coexistence), versus differences in competitive ability (which could involve a narrow range of trait values and actually lead to clustering) (Mayfield and Levine, 2010).

Many phylogenetic and functional trait comparative studies assume that species mean trait values are ecologically meaningful (Garamszegi and Møller, 2010). However, plant community assembly can also be influenced by intraspecific variation in traits that arise because of genotypic diversity and/or phenotypic plasticity (Burns and Strauss, 2012, Violle et al., 2012). Intraspecific variation can influence the detection of environmental filtering by allowing a species with high intraspecific variation to adapt and grow at sites filtered by a range of different environmental conditions (Jung et al., 2010). It can also affect competitive exclusion by allowing a species to adjust to its neighbours and potentially reduce the impacts of limiting similarity (the tendency of co-occurring species to differ in their resource acquisition traits) (Burns and Strauss, 2012, Pfennig et al., 2006). Detection of the mechanisms of community assembly can benefit then from considering intraspecific variation in functional trait analysis (Albert et al., 2012).

Several recent studies that have assessed successional change through time using phylogenetic and/or functional trait approaches have found clustering in early successional or more disturbed plant communities, and increasing overdispersion as successional processes continue within less disturbed communities (Li et al., 2015, Whitfeld et al., 2012, Mo et al., 2013, Letcher, 2010b). This suggests that environmental filtering might play a more important role in the recruitment of species during the early stages of assembly in disturbed communities and competitive exclusion gradually takes over in importance as plant community’s age.

Phylogenetic approaches have been applied very rarely to understanding reforestation methods (but see Hipp et al. (2015) and Verdú et al. (2012)) and studies of reforestation have focused more on positive biotic interactions, such as facilitation that can initiate successional development (Shooner et al., 2015). Valiente-Banuet and Verdú (2007) found that regeneration niches are conserved across evolutionary time, and they argue that positive interactions occur between phylogenetically distant species and that facilitation can lead to phylogenetic overdispersion. Seedling phylogenetic diversity can also have a positive influence on survival, via phylogenetically-correlated pathogen

110 susceptibilities, which can lead to density-dependent selection (Webb et al., 2006). In a meta- analysis Verdú et al. (2012) show that co-occurring plant species that are more similar in life-form are more likely to survive if they are distantly related. Here we measure the relative importance of environmental filtering and competitive interactions on the recruitment of seedlings beneath different reforestation methods across a degraded tropical forest landscape on the Island of Leyte in the Philippines. The reforestation methods we compare ranged from low-to high-diversity forests, in the form of monoculture plantations of Swietenia macrophylla King, mixed-species plantations and regenerating selectively logged native forests. We analysed both phylogenetic and functional traits, including intraspecific variability in specific leaf area (SLA). We specifically address the following three questions:

1. What is the phylogenetic and functional trait structure (SLA, leaf nutrients, life-form, potential plant height and dispersal type) of seedling communities beneath different reforestation types? And how do these phylogenetic and functional trait structures vary depending on site-level edaphic conditions? 2. What is the intraspecific variation of SLA observed in the different forest types, and between species and functional groups that are common or obligate across forest types? 3. What do the phylogenetic structure, functional trait structure and intraspecific variation in SLA tell us about how seedling communities assemble in the different forest types? We expected to find a shift in the main assembly mechanisms between seedling communities beneath the different reforestation methods, from environmental filtering under monoculture forests (where species colonisation is limited by abiotic conditions) to competitive exclusion within regenerating selectively logged forest, which are more diverse in terms of species and microclimates. We also expected to find that seedlings beneath regenerating selectively logged forests and mixed-species plantations will show greater intraspecific variation in SLA. This is because of the greater environmental and biotic variation in these forests (e.g., in light levels, topography and leaf litter composition) (Poorter, 1999).

111

4.2. Methods

4.2.1. Study sites and Data Collection The study was undertaken on the Island of Leyte in the Philippines (between 124ᵒ17’ and 125ᵒ18’ east longitude, and between 9ᵒ55’ and 11ᵒ48’north latitude). Leyte has an average annual rainfall of 2753 mm and an average annual temperature of 27.5 ᵒC. All plants below 2 m in height were sampled within a total of 35 circular plots (individual plot area = 78 m2) that were spread across 15 sites. These sites included five mahogany (Swietenia macrophylla King) monoculture plantations, five mixed-species plantations (known locally and hereafter as ‘Rainforestation’) and five regenerating selectively logged native forests. Sites occurred at elevations of less than 600 m a.s.l. and, with one exception, had soils of volcanic origin. Plantations were between 13 and 18 years of age at the time of sampling. Regenerating selectively logged native forest sites had higher average slope angles and elevations than plantations, were logged relatively recently (~20 years) and, at the time of sampling, were frequently used by nearby communities for the harvesting of non-timber forest products (NTFPs). The plantations were located at similar distances to potential seed sources as the regenerating selectively logged native forests (Wills et al., 2016, Nguyen et al., 2016). Plant identification was verified at Visayas State University. A CID Bio-Science CI-110 Plant Canopy Imager was used to measure Leaf area index (LAI) and the average of three readings per plot were used in the analysis.

4.2.2. Functional traits Three continuous traits were measured: SLA (cm2/g), leaf nitrogen concentration (hereafter, LNC, % dry leaf mass) and leaf phosphorus concentration (hereafter, LPC, % dry leaf mass), generally following the protocols set out by Pérez-Harguindeguy et al. (2013). We collected a minimum of two leaves per seedling for all tree and shrub species recorded in the plots. The youngest mature fully expanded leaves were collected, but in some cases (~20 individuals) leaf traits were not collected as it were judged to be detrimental to the individual’s survival because of only a few leaves being available. Collected leaves were placed into a paper bag, labelled and scanned using a CID Bio-Science CI-203 Laser Area Meter. Leaf area scanning was conducted either onsite or in the afternoon of the same day. Leaf samples were oven dried at 65 ᵒC for 48 hours, and weighed to calculate SLA. To determine LNC and LPC, samples were prepared with a single digestion method and analysed with a colorimetric determination of LNC using the salicylate-hypochlorite method developed by Baethgen and Alley (1989) and LPC using an adaptation of Murphy and Riley (1962) single solution method (Anderson and Ingram, 1989).

112

SLA was sampled for 856 individuals, representing 91 identified species. LNC and LPC were analysed on a subset of the species used for the calculation of SLA, which included 127 individuals representing 53 species. For intraspecific variation of SLA, initially all species with more than one SLA measurement were used, and a minimum of five individuals per species are presented within the results, limiting the number of species that were analysed to 39. Data on three discrete traits were extracted from open databases and primary literature; dispersal type (abiotic or biotic), potential plant height was coded with 4 levels (1 = understory “0 m-5 m”, 2 = mid-canopy “6 m-15 m”, 3 = canopy “16 m-30 m”, 4 = emergent “30 m +”), and life-form was also coded with 5 levels (1 = herb, 2 = vine, 3 = palm, 4 = shrub, 5 = tree).

4.2.3. Community phylogeny The regional species pool can have significant influences on the local phylogenetic structure (Lessard et al., 2012), statistical inferences and subsequent conclusions regarding community assembly processes (Pigot and Etienne, 2015). Therefore, we contextualize the present study within broader evolutionary temporal and spatial scales. The Philippine flora, in particular the Mindanao- Eastern Visayas tropical forest (that includes the island of Leyte) has a dominant affiliation with the Asia/Malesia floristic province, but also includes Gondwanan relicts such as the southern gymnosperms (e.g. Podocarpus rumphii and Agathis philippinensis) (Sniderman and Jordan, 2011, Langenberger et al., 2006). These Gondwanan lineages have extremely long divergence times relative to all other lineages. The community phylogeny was constructed using a species pool consisting of all 125 seed-plant species recorded in the understorey of all forest types. (Five non-seed plants were excluded due to their extremely early divergence relative to all other species, which would have eclipsed the phylogenetic distances among seed plants). The 125 seed plants consisted of 124 angiosperms and one gymnosperm (Agathis philippinensis). This gymnosperm was included in traits analyses, but excluded from phylogenetic diversity analysis due to its long phylogenetic distance compared to other species (Cadotte, 2014, Cavender-Bares et al., 2006). Of the total 125 species used (124 angiosperms for phylogenetic diversity), 95 were native and 30 were classified as recently introduced. The phylogenetic structure was analysed using three taxonomic subsets: (1) all angiosperm species (i.e., excluding the gymnosperm), (2) all angiosperm species excluding recently introduced species, and (3) all angiosperm species excluding monocot species (i.e., tree and shrub species). These subsets were used to decipher their influence on the phylogenetic structure of the different seedling communities (Table S4.1).

113

A Newick formatted tree was constructed using the latest seed plant super tree (R20120829) from the Phylomatic command (Version 3). Dating of internal nodes was initially performed using the Phylocom 4.2 package and its evolutionary ‘ages’ file derived from Wikström et al. (2001). Due to syntax and nomenclature errors between the Newick formatted tree and the wikstrom.ages file, we used the corrected ‘ages’ file provided by Gastauer and Meira-Neto (2013) in their supplementary material. The ages provided by Wikström et al. (2001) are minimum ages (earliest known fossils) and therefore systematically underestimate the true ages of these divergences. Therefore, we also performed analyses using Bayesian estimates of divergence times for shared internal nodes from a new molecular phylogeny of angiosperms in Australia’s Wet Tropics (Wells, 2012). These estimates gave the posterior mean ages based on a rigorous set of fossil calibrations and probability distributions for the true age of corresponding nodes using Bayesian relaxed clock methods (BEAST) – (Drummond and Rambaut, 2007). The molecular phylogeny from Australia’s Wet Tropics contained 14 internal nodes common to our Leyte island phylogeny. Phylocom’s Bladj algorithm was used to evenly space divergence times between internal nodes of ‘known’ ages in each tree, i.e. one based solely on Wikström et al. (2001) ages, and one based on the Bayesian estimates. Both chronograms are, therefore, approximate or ‘pseudo’ chronograms because single fixed points were assigned for ‘known’ ages, and all other divergences were assumed to be evenly distributed between them. However, the Bayesian means do estimate the true ages of the ‘known’ divergences (rather than only their logical minima), and should provide a more accurate estimation of branch lengths than the tree based on the Wikstrom minimum ages.

4.2.4. Data analysis Analyses were conducted using R statistical computing version 3.1.1 (R Core Team, 2013) and Phylomatic command Version 3 (Webb et al., 2008). Phylogenetic structure was quantified using phylogenetic diversity (PD), mean pairwise phylogenetic distance (MPD) and mean nearest taxon phylogenetic distance (MNTD) for both incidence- and abundance- based methods (Webb et al., 2002). The functional structure was quantified using the same metrics as the phylogenetic structure; mean pairwise functional traits distance (MFD) and mean nearest functional traits distance (MNFD). Trait distances were constructed using a Euclidian distance matrix for continuous traits both individually and together. For categorical traits (that included missing data) a Gower distance matrix was constructed within the FD package in R (Gower, 1971, Laliberté, 2014, Laliberté and Legendre, 2010). To test for differences in the seedling phylogenetic and functional structure beneath different forest types, taking into consideration unequal sample sizes and species richness, we compared analyses to a null

114 model that randomized the species identify at the plot level with species drawn from the regional species pool, using 1000 null iterations. To do this, we created standard effect sizes (SES) in the picante R package (Kembel et al., 2010). We used linear mixed effect models (LMEMs), estimated using maximum likelihood, to compare the phylogenetic and functional structure between forest types and how this varied depending on the abiotic conditions (e.g. soil phosphorus, soil nitrogen and LAI). To account for our sampling design, random effects were structured as plots nested within sites. The R package nLME was used to fit all LMEMs and because the experimental design is largely balanced, Wald F-statistics were used to assess the significance of the fixed effects (Pinheiro et al., 2016). The effects package (Fox, 2016) from R was used to display the higher-order fixed effects. We also applied a Holm Sidàk correction to account for the potential inflation of Type 1 errors (Abdi, 2010) to LMEMs conducted on the influence of forest type, soil phosphorus, soil nitrogen, and LAI for each of the response variables (Firn et al., 2017). We used the nodesigl command from the Phylocom 4.2 package to identify the patterns of phylogenetic structure between forest types. This function estimates which nodes in each forest type contributed disproportionally more or less to the phylogenetic structure. The function does this by comparing observed frequencies to that of a random null model, which is created by drawing the number of taxa within a forest type randomly from the regional species pool. To test relationships between continuous traits (SLA, LNC and LPC) taking into consideration phylogenetic covariance, phylogenetic generalized least squares (pGLS) models were constructed and tested for significance using ANOVAs. In order to test the relationships between discrete traits (potential plant height, dispersal and life-form), we developed phylogenetic logistic regression models within the R package phylolm (Ho and Ané, 2014). Potential plant height and life-form were considered the independent variables and dispersal was treated as a binary (abiotic or biotic) dependent variable (Ives and Garland, 2010). Phylogenetic signal was tested for each trait using Blomberg’s K statistic (Blomberg et al., 2003), using mean SLA, LNC and LPC values for each species. Significance of the phylogenetic signals were determined by comparing the observed K values with those obtained from null models that were based on a Brownian motion model of trait evolution. In order to test for phylogenetic signal of SLA taking into consideration sampling error and/or intraspecific variation, we used the approach outlined in Ives et al. (2007) and for species with only one individual we used the mean variance for all species with multiple measurements. Then to test for phylogenetic signals of discrete traits including potential plant height, life-form and dispersal type, we used the ‘Fixed Tree, Character randomly Reshuffled’ methods proposed by Maddison and Slatkin (1991).

115

4.3. Results

4.3.1. What is the phylogenetic and functional trait structure of seedling communities beneath the different forest types?

Phylogenetic diversity Overall, seedling communities were phylogenetically overdispersed within regenerating selectively logged forests and clustered within monoculture forest types (Figure 4.1 and 4.2). This pattern was robust to different null models, species pools (e.g. all species, natives, trees and shrubs) and pseudo- chronograms, i.e., one pseudo-chronogram based solely on Wikström et al. (2001) ages and one incorporating Bayesian estimates (Table S4.1).

Observed PD was higher within the regenerating selectively logged forest sites (F2, 12 = 9.0, P = 0.03). PD between forest types also differed from the null model expectations when all species were included in the models using standard P values, however these patterns were non-significant using corrected P values (Figure 4.1). The PD of tree and shrub species differed between forest types (F2,

12 = 9.26, P = 0.03). The abundance-weighted and non-weighted MPD and weighted MNTD did not differ from the null model expectations between forest types, soil nitrogen, soil phosphorus or LAI (Table S1), using the entire species pool (i.e. including exotics and natives, and all growth forms). The non-weighted MNTD did not differ from the null model expectations for all species and natives, but did when restricted to trees and shrubs (all species: F2, 12 = 3.95, P = 0.3, tree and shrub:

F2, 12 = 9.74, P = 0.02 and natives: F2, 12 = 6.56, P = 0.08), with regenerating selectively logged forest seedlings being overdispersed and monoculture and Rainforestation plantations being more similar and clustered (Figure 2). When the analysis was restricted to tree and shrub species, significant overdispersion was detected within regenerating selectively logged forest and clustering was detected within the monoculture and Rainforestation seedling communities, for the MPD non- weighted (F2, 12 = 9.49, P = 0.02) (Figure 4.2).

116

Figure 4.1. Phylogentic diversity in different forest types. (a) Phylogenetic diversity (PD) was compared to null model distrubutions (SES) of understories beneath monoculture, Rainforestation and regenerating selectivly logged forests. Positive values indicate overdispersion whereas negative values indicate clustering, showing higher order fixed effects from LMEM terms. Variation in PD is also shown across variation in soil nitrogen (b), soil phosphorus (c, * denotes significance (P<0.05), using F-statistics) and LAI (d).

117

Figure 4.2. Understorey phylogenetic and leaf trait diversity beneath monoculture, Rainforestation and regenerating selectively logged forests. Phylogenetic and leaf trait structure were measured against standardised effect sizes (SES) for (a) non-weighted mean nearest taxon phylogenetic distance (MNTD) for all species, (b) non-weighted mean pairwise phylogenetic distance (MPD) for tree and shrub species in isolation, (c) weighted mean pairwise phylogenetic distance (MPD) for tree and shrub species in isolation, (d) non-weighted mean pairwise functional distances (MFD) of SLA, (e) non-weighted MFD of LNC and (f) non-weighted MFD of log transformed LPC. The phylogeny incorporating Bayesian estimates of divergence times and a functional trait dendrogram were used as the basis of the displayed metrics. Positive values indicate overdispersion and negative values indicate clustering compared to the null model expectations that used species richness to randomise the phylogeny and dendrogram.

118

Associations between phylogenetic groups and forest types Using the nodesigl function from Phylocom 4.2, the family Meliaceae (Swietenia macrophylla, Sandoricum koetjape, Lansium domesticum and Dysoxylum gaudichaudianum) contributed significantly more taxa to seedling communities beneath the monoculture forest types than by chance. The family Moraceae (Ficus septica, F. pseudopalma, F. nota, Artocarpus odoratissimus and A. blancoi) also contributed more taxa than by chance within monoculture forest types, when restricting the analysis to native species. There was also an overabundance of seedlings from the genus Ficus (that included seven species) within regenerating selectively logged forests.

Functional diversity Generally, functional diversity showed weaker trends across forest types, compared to phylogenetic diversity and did not differ from the null model expectations (Table S4.2). Using SLA, abundance- weighted and non-weighted MFD displayed overdispersion for regenerating selectively logged forests, while monoculture and Rainforestation sites were similar and clustered (Figure 2). However, this trend was not evident using MNFD of SLA seedling values. Using seedling LNC, abundance-weighted and non-weighted MFD, and MNFD displayed a similar trend to SLA. Abundance-weighted and non-weighted MFD and MNFD of seedling SLA, LPC and LNC displayed random patterns, where forest types had similar SES values. Overall, understory leaf traits suggests similar trends to phylogenetic structure, with overdispersion being apparent under the regenerating selectively logged forest and clustering within monoculture forest seedling communities. However, this relationship was weak, and could not be statistically differentiated from random patterns. In contrast, analysis of the functional structure incorporating categorical traits of potential plant height, dispersal type and life-form suggests a reverse in the patterns generally found using phylogenies and to a lesser extent leaf traits. The weighted SES.MFD using all traits tended towards overdispersion within monoculture forests and random to clustering within the Rainforestation and regenerating selectively logged forests (Figure 3a). When restricting the analysis to native species, there was a weakening in the patterns seen when incorporating exotics. Specifically, the regenerating selectively logged forests were less clustered and monoculture forests were less overdispersed, than when non-native species were also included (Figure 4.3b). Within the monoculture forests, weighting by species’ relative abundances using the MFD and MNFD of all traits resulted in an increase towards overdispersion compared to the null model.

119

Figure 4.3. Higher order fixed effects from LMEMs for understorey functional diversity (life-form, potential plant height, dispersal type, SLA, LNC and LPC), measured as standard effect sizes of abundance weighted MFD, for all species (a) and for native species in isolation (b), beneath the different forest types. A Gower distance matrix was constructed as it allows for categorical and missing data.

Phylogenetic and functional trait relationships Taking into account phylogenetic covariance, the tallest trees and native species were significantly more likely to be wind-dispersed (trees: z = 2.98, P = 0.003 and natives: z = 2.289, P = 0.02).

However, this trend was not detected when using the entire species pool (all species: z = 0.958, p =

0.34). SLA had a positive relationship with LNC after accounting for phylogenetic covariance (F1, 51 = 24.47, P = < 0.0001). SLA and LPC were not significantly correlated after accounting for phylogenetic covariance (F1, 51 = 0.5, P = 0.48). LPC and LNC exhibited significant phylogenetic signals compared to a Brownian motion model of evolution (P < 0.05), with LPC showing higher K values (K = 0.528) than both LNC (K = 0.419) and SLA (K = 0.285, non-significant). Due to the high number of replicates for SLA, we analysed phylogenetic signal taking into account sampling error and/or intraspecific variation, and this increased the K value for SLA considerably (K = 0.436). Significant phylogenetic signals (P<0.05) were also found for life-form, dispersal type and potential plant height (Table S4.3).

120

4.3.2. SLA and variation beneath the different forest types and between common or obligate clades The seedling populations beneath all forest types showed an average a coefficient of variation (CV) for SLA values, of around18. Average seedling SLA values at the plot level were significantly higher in the monoculture forest type than in the regenerating selectively logged forests, while the Rainforestation forest types were intermediate (F2, 12 = 10.3, P = 0.003). The mean SLA weighted by species’ relative abundances showed a similar pattern, however it was not significant (F2, 12 = 1.02, P= 0.4). The plot level CV was significantly higher within regenerating selectively logged forests than monoculture forests and the Rainforestation forest type was again intermediate (F2, 12 = 6.28, P = 0.01) (Figure 4.4a). Taking into account differences in species richness and only using species with adequate replication (≥5 individuals per species) for SLA measurements; we found regenerating selectively logged forests included species with both high and low intraspecific variation in SLA (e.g., higher variation in SLA: Pterocarpus indicus, Koordersiodendron pinnatum and Neolitsea vidalii, lower variation in SLA: Canarium luzonicum, Diplodiscus paniculatus, Ficus balete and Shorea contorta) (Figure

S1). This resulted in a significant overdispersion of CV in the weighted SESMNFD (F2, 12 = 15.32,

P = 0.007) and SESMFD (F2, 12 = 14.91, P > 0.007) within regenerating selectively logged forest seedling communities, and clustering within the monoculture and Rainforestation forest types (Figure 4.4b and c). These also varied depending on soil nitrogen (Table S4.4) The Moraceae family was common across forest types and generally showed higher than average CV values for SLA (~18 to ~34). The monoculture grown species Swietenia macrophylla showed a relatively lower than average CV in SLA (~16). Species within the family Dipterocarpaceae that were absent within the monoculture forests showed a relatively low CV in SLA (~14-16). Wind- dispersed species had representatives that showed both a higher variation in SLA (e.g. Pterocarpus indicus) and a lower variation in SLA (e.g. Dipterocarpaceae species) (Table 4.1 and S4.4).

121

Figure 4.4. Higher order fixed effects for mean coefficient of variation (CV) for SLA values of all individuals at the plot level (a), MFD and MNFD weighted by abundances for species with 5 or greater SLA replicates (b, c), between monoculture, Rainforestation and regenerating selectively logged forests seedling communities. * denotes a significant relationship overall (P<0.05).

122

Table 4.1. The coefficient of variation (CV) for clades that represent common and obligate groups between monoculture (M), Rainforestation (R) and regenerating selectively logged forest types (S). Plus (+) signs represent a statistical overabundance of those clades within the corresponding community compared to a null model that randomly assigned the same number of species from the same species pool (nodesig statistic) (Webb et al., 2008). Minus (-) signs represent a lack of that species or family within the corresponding forest type. Forest type Family Species SLA- Species present n CV Common Moraceae (+M) Ficus septica 21 18.4 (+S) Common Ficus 25 33.8 pseudopalma (+S) Common Ficus nota (+S) 20 28.1 Common Artocarpus 39 18.4 odoratissimus Common Artocarpus 24 23.5 blancoi Obligate (M) Meliaceae (+M) Swietenia 49 16.2 macrophylla Common Sandoricum 5 23.9 koetjape Obligate (M Lansium 5 35 and R) domesticum Obligate (M Dysoxylum 6 10.8 and S) gaudichaudianum Obligate (R Dipterocarpaceae Shorea contorta 43 13.6 and S) (-M) Obligate (R Hopea plagata 21 15.9 and S) Obligate (R Fabaceae Pterocarpus 16 36.5 and S) indicus (-M)

Obligate (S) Clusiaceae Calophyllum 6 15.7 inophyllum(-M) Obligate (S) Burseraceae Canarium 7 7.9

123

luzonicum(-M)

4.4. Discussion Overall, we found overdispersion within regenerating selectively logged forest seedling communities and clustering within monoculture seedling communities, using both a phylogenetic approach and leaf traits, though leaf traits showed weaker patterns. Rainforestation sites in general were not statistically different phylogenetically or functionally from the other forest types (see Figure 4.5 for a summary of community assembly processes acting on the different forest types). These results support our initial predictions that community assembly in monoculture seedling communities is more strongly influenced by environmental filtering likely, because of the low diversity and even-aged canopy facilitating a more homogenous understorey environment, particularly regarding light availability, and due to the lack of seed sources and limited dispersal. Regenerating selectively logged forest, which had higher-diversity canopies resulting in more complex canopy structures, showed evidence of competitive interactions between species recruiting into the understorey, as well as overdispersion of phylogenetic relatedness and trait values. We also found an unexpected result when we included in the analyses potential plant height, life-form and dispersal type. Monoculture forest seedling communities displayed high overdispersion and this was driven by the introduction of functionally distinct, human-dispersed species, such as wind- dispersed herbs and shrubs; Crotalaria spp., Chromolaena odorate (Siam weed), Sphagneticola trilobata (Singapore daisy), and smaller stature trees; Theobroma cacao (cacao), Psidium guajava (guava). When restricting the analysis to native species or just tree species, taking phylogenetic covariance into consideration, we found that wind-dispersed tree species generally displayed the tallest potential heights. This result is consistent with a previous study on the same forest types, where we found that native wind-dispersed tree species are limited across the studied forest types and are the tallest trees at maturity, further supporting their occurrence as important emergent tree species within tropical forests (Wills et al., 2016). Wind-dispersed tree species are recruitment-limited within these monoculture forests and this might result in a reduction in forest canopy height across landscapes as these novel forests mature and succession proceeds, which has long-term implications for animal habitat and forest microclimatic conditions (Wills et al., 2016). We expected the regenerating selectively logged forest seedling communities to show a higher variation in SLA, due to more complex canopy structures and therefore understory species being exposed to varied environmental and biotic conditions. This was supported with the regenerating

124 selectively logged forest understoreys comprising species with a high SLA variation (e.g., Pterocarpus indicus, Neolitsea vidalii and Palaquium foxworthyi), but also species with a lower variation in SLA (e.g., Canarium luzonicum, Diplodiscus paniculatus, Ficus balete and Shorea contorta). These results extend our initial predictions, indicating both environmental filtering and competitive exclusion are operating within regenerating selectively logged forest seedling communities; whereas within monoculture seedling communities, environmental filtering is more prominent, but human-assisted recruitment appears to overcome this filtering in some cases (Figure 4.5 and S4.2).

Figure 4.5. Summarizing community assembly processes indicated by analysing evolutionary, leaf trait (SLA, LNC and LPC), discrete trait (potential height, dispersal and life-form) and within- species SLA data, for regenerating selectively logged forest (Regenerating), Rainforestation and monoculture forest types. Species richness (SR) and phylogenetic diversity (PD) were highest within regenerating selectively logged forest and lowest within monoculture forest types, Rainforestation was intermediate. Species within the family Moraceae (green) were common across forest types and tall, wind-dispersed native species (red) were limited to regenerating selectively logged forest and Rainforestation forest types. Species that exhibited high and low variation in SLA (brown gradient) were also restricted to regenerating selectively logged forest seedling communities. Exotic human-dispersed herbs, shrubs and trees (blue) increased all measures of seedling diversity within monoculture forest types.

125

4.4.1. Phylogenetic and functional structure beneath forest types We saw some convergence in the results obtained from analyses of phylogenetic structure and functional trait structure, i.e., in patterns of dispersion and clustering in the different forest types, although leaf traits showed some inconsistent patterns. Leaf traits showed significant phylogenetic signals, particularly for LPC and LNC, and to a lesser extent SLA after accounting for intraspecific variation and/or sampling error. In line with our results, several other studies have found that these leaf traits are important axes for revealing ecosystem processes (Pérez-Harguindeguy et al., 2013, Jung et al., 2010). However, phylogenetic diversity can represent additional traits without revealing what these traits are, or their underlying importance for the assembly of plant communities (Cadotte et al., 2009, Flynn et al., 2011). In contrast to our findings but in grasslands, Purschke et al. (2013) found divergent responses between phylogenetic and functional structure, with a lack of phylogenetic signal in measured traits, which resulted in the conclusion that phylogenies provided little information about community assembly processes, particularly if multiple ecologically- meaningful traits are used instead. Collectively, our leaf trait results concur with several authors, with our results supporting the importance of environmental filtering of leaf functional traits in shaping community assemblages (Lemoine et al., 2015, Kraft and Ackerly, 2010, Paine et al., 2012). The trend of phylogenetic clustering within more disturbed sites and at earlier stages of succession, developing towards overdispersion within more natural communities and at later stages of succession, has been commonly observed within tropical landscapes (Swenson et al., 2007, Purschke et al., 2013, Mo et al., 2013, Whitfeld et al., 2012, Letcher, 2010b). Our results generally support this, however, it should be noted that competition can also lead to clustering (Mayfield and Levine, 2010) and abiotic and biotic factors may not be mutually exclusive. Instead, these factors likely interact to explain community assembly (Kraft et al., 2015a). The understories within the monoculture forest type contained related species that possess traits that enable them to overcome dispersal limitations. For example, the family Moraceae typically have very small seeds and are dispersed by habitat-generalist bird species. This role for dispersal limitation has been illustrated across the tropics (Corlett, 2006), and more recently at the sites utilised in this study, see Wills et al. (2016). Our findings have revealed that the family Meliaceae also exhibited an overabundance within the monoculture forest type, likely from human-assisted dispersal of species with a large fruit size (e.g. Sandoricum koetjape and Lansium domesticum). Primates are known to be important dispersal agents in tropical forests and are responsible for the

126 dispersal of many large-fruited species’ (Chapman, 1989). Here, we find that humans are also key dispersal agents for exotic large-fruited species in highly modified tropical landscapes. The mechanism behind the overdispersion in categorical traits (potential plant height, life-form and dispersal type) within monoculture seedling communities is driven by the introduction of functionally distinct exotic species, both directly (via human-dispersal of seeds) and indirectly (invasion of species/traits due to habitat modification and vacant niche space). Interestingly, this result provides support for the hypothesis proposed by Li et al. (2015) that colonisation by functionally distinct species drives overdispersion through environmental modification. However, in the present study this is within the most unnatural forest communities (monoculture forests) through the introduction of exotic species due to the close physical relationship these small-scale community-managed forests have with local people. The disparity in results between incorporating categorical traits and phylogenetic or leaf trait structure reveals the importance of quantifying the different aspects of biodiversity and deciding on ecologically-appropriate traits. Lemoine et al. (2015) and Kraft and Ackerly (2010) suggest that community assembly mechanisms can show different patterns depending on which traits are included in the analysis. Our results provide support for this, with seedling leaf and categorical traits showing divergent patterns across different reforestation methods; specifically, random to clustering of leaf trait structure and overdispersion in categorical trait structure within monoculture forest types. Therefore, our results should be considered in the context of the trait(s) used in determination of functional diversity and what aspect of functional diversity is of importance to socio-economic and conservation outcomes. One possible explanation is that the continuous leaf traits (SLA, LNC and LPC) measured within this study show the influence of the conditions found within these forests at the time of sampling, while the categorical traits derived from dispersal events reflect future characteristics of the forest, as these species mature. Assessing the value of seedling biodiversity beneath monoculture forest types depends largely on the desires of landholders and the relative importance of conservation and socio-economic values. Non-native understorey species contribute to ecosystem function and subsequent services, but if dominated by few species these understoreys may lack the ecological complexity and ecosystem services associated with increased leaf trait diversity and phylogenetic diversity (e.g. ecosystem resilience to invasion and to a changing climate).

4.4.2. Intraspecific variation in SLA, comparing functional groups and forest types It is widely recognised that there is a large amount of intraspecific variation in key plant functional traits including SLA (Swenson and Enquist, 2009, Cavender-Bares et al., 2006, Messier et al.,

127

2010), and that this variation can have implications for coexistence and improvements in detecting underlying community assembly processes (Ashton et al., 2010, Burns and Strauss, 2012, Jung et al., 2010). Our results suggest that incorporating information on intraspecific variation in SLA can extend our understanding of seedling community assembly. Our findings revealed that regenerating selectively logged forests recruit species with both high and low variation in SLA. Therefore, these forests appear to provide habitat for later successional species with conservative leaf economies and low variation in SLA (Walters and Reich, 1999) as well as species that have higher variation in SLA, due to genetic diversity or environmental acclimation. Previous studies have found variation in SLA is due to both abiotic conditions (e.g., soil and light) and competitive interactions (Burns and Strauss, 2012, Bloor and Grubb, 2004). Our results support these findings, and with this relationship likely explained by the more varied abiotic and biotic conditions found in the understory of regenerating selectively logged forests compared to the other forest types. Species in the Moraceae family, and particularly the genus Ficus, were abundant across all of the studied forest types. Members of this family showed a larger than average variation in SLA, in all forest types. This may reflect the ability of species within this genus to overcome biotic and abiotic filters, or may reflect their pre-establishment traits such as seed size. Native wind-dispersed species, which were previously identified as limited in their recruitment ability across the studied forest types (Wills et al., 2016), show both high and low variation in SLA. The leguminous species Pterocarpus indicus showed a very high SLA variation. In contrast, other wind-dispersed native species including Calophyllum inophyllum and species within the family Dipterocarpaceae all showed below-average species-specific mean SLA and variation in SLA. This likely reflects their later successional status and a more conservative resource acquisition strategy on the leaf economic spectrum (Walters and Reich, 1999, Wright et al., 2004).

4.5. Conclusion and implications for tropical reforestation Overall, we found that seedling communities beneath mahogany monocultures show influences of environmental filtering, with co-occurring seedling species being more closely related than by chance, and having more similar leaf traits than expected by chance. In contrast, seedling communities within regenerating selectively logged forest, show overdispersion in both leaf traits and phylogenetic relatedness, indicating stronger influences of competitive processes, with co- occurring seedling species being more distantly related and displaying more divergent trait values than would be expected by chance. However, when considering intraspecific variation in SLA, we

128 find that the assembly of these seedling communities (beneath regenerating selectively logged forests) is likely explained by both environmental filtering and competitive interactions (see Figure 4.5 for a summary conceptual diagram of community assembly processes).

Our results highlight that overall, a greater fulfilment of niche space occurs within seedling communities in regenerating selectively logged forests, compared to beneath monocultures. The Rainforestaion forests showed intermediate fulfilment of niche space between the other forest types, suggesting that more diverse plantations do not necessarily capture true differences in species effects on ecosystem functions. Within the regenerating selectively logged forests there is a greater occurrence of more phylogenetically- and functionally-divergent seedling species, and these species show both high and low abilities to adjust their SLA values.

In conclusion, we highlight the need for incorporating greater phylogenetic and functional diversity in reforestation projects, which may not equate to increasing pure species numbers, but incorporates actual differences in how species influence ecosystem function, within human-modified tropical landscapes. This could result in reforestation interventions that lead to a greater niche fulfilment by seedling communities regenerating under planted forests, and therefore more functionally rich future forests that can better adapt to future environmental conditions. To do this, we recommend the promotion of phylogenetically and functionally broader ranges of seed or seedling stocks in reforestation schemes across degraded tropical landscapes. In particular, efforts should be made to include native emergent, wind-dispersed tree species (Wills et al., 2016), species from other limited functional groups (e.g., large-seeded species), and species with a broader range of mean SLA values and levels of intraspecific variation in SLA.

References Abdi, H. (2010) Holm's sequential Bonferroni procedure. Pages 573–577 in N. Salkind, editor. Encyclopedia of research design. Sage, Thousand Oaks, California, USA. Albert, C. H., de Bello, F., Boulangeat, I., Pellet, G., Lavorel, S. & Thuiller, W. (2012) On the importance of intraspecific variability for the quantification of functional diversity. Oikos, 121, 116-126. Anderson, S. E. & Ingram, J. S. I. (1989) Tropical Soil Biology and Fertility: A handbook of Methods. p. 171. C.A.B. International, Aberystwyth. Ashton, I. W., Miller, A. E., Bowman, W. D. & Suding, K. N. (2010) Niche complementarity due to plasticity in resource use: plant partitioning of chemical N forms. Ecology, 91, 3252-3260.

129

Baethgen, W. E. & Alley, M. M. (1989) A manual colorimetric procedure for measuring ammonium nitrogen in soil and plant Kjeldahl digests. Communications in Soil Science and Plant Analysis, 20, 961-969. Baraloto, C., Hardy, O. J., Paine, C. E. T., Dexter, K. G., Cruaud, C., Dunning, L. T., Gonzalez, M.-A., Molino, J.-F., Sabatier, D., Savolainen, V. & Chave, J. (2012) Using functional traits and phylogenetic trees to examine the assembly of tropical tree communities. Journal of Ecology, 100, 690-701. Barlow, J., Gardner, T. A., Araujo, I. S., Ávila-Pires, T. C., Bonaldo, A. B., Costa, J. E., Esposito, M. C., Ferreira, L. V., Hawes, J., Hernandez, M. I. M., Hoogmoed, M. S., Leite, R. N., Lo-Man-Hung, N. F., Malcolm, J. R., Martins, M. B., Mestre, L. A. M., Miranda-Santos, R., Nunes-Gutjahr, A. L., Overal, W. L., Parry, L., Peters, S. L., Ribeiro-Junior, M. A., da Silva, M. N. F., da Silva Motta, C. & Peres, C. A. (2007) Quantifying the biodiversity value of tropical primary, secondary, and plantation forests. Proceedings of the National Academy of Sciences, 104, 18555-18560. Blomberg, S. P., Garland, T., Ives, A. R. & Crespi, B. (2003) Testing for Phylogenetic Signal in Comparative Data: Behavioral Traits are more Labile Evolution, 57, 717-745. Bloor, J. M. G. & Grubb, P. J. (2004) Morphological Plasticity of Shade-Tolerant Tropical Rainforest Tree Seedlings Exposed to Light Changes. Functional Ecology, 18, 337-348. Burns, J. H. & Strauss, S. Y. (2012) Effects of competition on phylogenetic signal and phenotypic plasticity in plant functional traits. Ecology, 93, S126-S137. Cadotte, M. W. (2014) Including distantly related taxa can bias phylogenetic tests. Proceedings of the National Academy of Sciences, 111, E536. Cadotte, M. W., Barlow, J., Nuñez, M. A., Pettorelli, N. & Stephens, P. A. (2017) Solving environmental problems in the Anthropocene: the need to bring novel theoretical advances into the applied ecology fold. Journal of Applied Ecology, 54, 1-6. Cadotte, M. W., Cavender-Bares, J., Tilman, D. & Oakley, T. H. (2009) Using Phylogenetic, Functional and Trait Diversity to Understand Patterns of Plant Community Productivity. PLoS ONE, 4, e5695. Cavender-Bares, J., Ackerly, D. D., Baum, D. A. & Bazzaz, F. A. (2004) Phylogenetic Overdispersion in Floridian Oak Communities. The American Naturalist, 163, 823-843. Cavender-Bares, J., Keen, A. & Miles, B. (2006) Phylogenetic Structure of Floridian Plant Communities Depends on Taxonomic and Spatial Scale. Ecology, 87, S109-S122.

130

Chapman, C. A. (1989) Primate Seed Dispersal: The Fate of Dispersed Seeds. Biotropica, 21, 148- 154. Corlett, R. T. (2006) Figs (Ficus, Moraceae) in Urban Hong Kong, South China. Biotropica, 38, 116- 121. Cornwell, W. K., Schwilk, D. W. & Ackerly, D. D. (2006) A Trait-Based Test for Habitat Filtering: Convex Hull Volume. Ecology, 87, 1465-1471. Drummond, A. J. & Rambaut, A. (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology, 7, 1-8. Firn, J., Schütz, M., Nguyen, H. & Risch, A. C. (2017) Herbivores sculpt leaf traits differently in grasslands depending on life form and land-use histories. Ecology, 98, 239-252. Flynn, D. F. B., Mirotchnick, N., Jain, M., Palmer, M. I. & Naeem, S. (2011) Functional and phylogenetic diversity as predictors of biodiversity- Ecosystem-function relationships. Ecology, 92, 1573-1581. Fox, J. (2016) Effect Displays for Linear, Generalized Linear, and Other Models. . http://www.r- project.org, http://socserv.socsci.mcmaster.ca/jfox/. Funk, J. L., Larson, J. E., Ames, G. M., Butterfield, B. J., Cavender-Bares, J., Firn, J., Laughlin, D. C., Sutton-Grier, A. E., Williams, L. & Wright, J. (2016) Revisiting the Holy Grail: using plant functional traits to understand ecological processes. Biological Reviews, DOI: 10.1111/brv.12275. Garamszegi, L. Z. & Møller, A. P. (2010) Effects of sample size and intraspecific variation in phylogenetic comparative studies: a meta-analytic review. Biological Reviews, 85, 797-805. Gastauer, M. & Meira-Neto, J. A. A. (2013) Avoiding inaccuracies in tree calibration and phylogenetic community analysis using Phylocom 4.2. Ecological Informatics, 15, 85-90. Gower, J. C. (1971) A General Coefficient of Similarity and Some of Its Properties. Biometrics, 27, 857-871. Grubb, P. J. (1977) The Maintenance of Species-richness in Plant Communities: the importance of the Regeneration Niche. Biological Reviews, 52, 107-145. Hipp, A. L., Larkin, D. J., Barak, R. S., Bowles, M. L., Cadotte, M. W., Jacobi, S. K., Lonsdorf, E., Scharenbroch, B. C., Williams, E. & Weiher, E. (2015) Phylogeny in the Service of Ecological Restoration. American Journal of Botany, 102, 647-648. Ho, L. S. T. & Ané, C. (2014) A linear-time algorithm for Gaussian and non-Gaussian trait evolution models. Systematic Biology.

131

Ives, A. R. & Garland, T. (2010) Phylogenetic Logistic Regression for Binary Dependent Variables. Systematic Biology, 59, 9-26. Ives, A. R., Midford, P. E. & Garland, T. (2007) Within-Species Variation and Measurement Error in Phylogenetic Comparative Methods. Systematic Biology, 56, 252-270. Jung, V., Violle, C., Mondy, C., Hoffmann, L. & Muller, S. (2010) Intraspecific variability and trait- based community assembly. Journal of Ecology, 98, 1134-1140. Kembel, S. W., Cowan, P. D., Helmus, M. R., Cornwell, W. K., Morlon, H., Ackerly, D. D., Blomberg, S. P. & Webb, C. O. (2010) Picante: R tools for integrating phylogenies and ecology. Bioinformatics, 26, 1463-1464. Kraft, N. J. B. & Ackerly, D. D. (2010) Functional trait and phylogenetic tests of community assembly across spatial scales in an Amazonian forest. Ecological Monographs, 80, 401- 422. Kraft, N. J. B., Adler, P. B., Godoy, O., James, E. C., Fuller, S. & Levine, J. M. (2015) Community assembly, coexistence and the environmental filtering metaphor. Functional Ecology, 29, 592-599. Laliberté, E. & Legendre, P. (2010) A distance-based framework for measuring functional diversity from multiple traits. Ecology, 91, 299-305. Laliberté, E., Legendre, P., and Shipley, B. (2014) FD: measuring functional diversity from multiple traits, and other tools for functional ecology. R package version 1.0-12. Langenberger, G., Martin, K. & Sauerborn, J. (2006) Vascular Plant Species Inventory of a Philippine Lowland Rain Forest and its Conservation Value. Biodiversity & Conservation, 15, 1271- 1301. Lemoine, N. P., Shue, J., Verrico, B., Erickson, D., Kress, W. J. & Parker, J. D. (2015) Phylogenetic relatedness and leaf functional traits, not introduced status, influence community assembly. Ecology, 96, 2605-2612. Lessard, J.-P., Belmaker, J., Myers, J. A., Chase, J. M. & Rahbek, C. (2012) Inferring local ecological processes amid species pool influences. Trends in Ecology & Evolution, 27, 600-607. Letcher, S. G. (2010) Phylogenetic Structure of Angiosperm Communities during Tropical Forest Succession. Proceedings: Biological Sciences, 277, 97-104. Li, S.-p., Cadotte, M. W., Meiners, S. J., Hua, Z.-s., Jiang, L. & Shu, W.-s. (2015) Species colonisation, not competitive exclusion, drives community overdispersion over long-term succession. Ecology Letters, 18, 964-973.

132

Maddison, W. P. & Slatkin, M. (1991) Null Models for the Number of Evolutionary Steps in a Character on a Phylogenetic Tree. Evolution, 45, 1184-1197. Mayfield, M. M. & Levine, J. M. (2010) Opposing effects of competitive exclusion on the phylogenetic structure of communities. Ecology Letters, 13, 1085-1093. Messier, J., McGill, B. J. & Lechowicz, M. J. (2010) How do traits vary across ecological scales? A case for trait-based ecology. Ecology Letters, 13, 838-848. Mo, X.-X., Shi, L.-L., Zhang, Y.-J., Zhu, H. & Slik, J. W. F. (2013) Change in Phylogenetic Community Structure during Succession of Traditionally Managed Tropical Rainforest in Southwest China. PLoS ONE, 8, e71464. Murphy, J. & Riley, J. P. (1962) A modified single solution method for the determination of phosphate in natural waters. Analytica Chimica Acta, 27, 31-36. Nguyen, H., Vanclay, J., Herbohn, J. & Firn, J. (2016) Drivers of Tree Growth, Mortality and Harvest Preferences in Species-Rich Plantations for Smallholders and Communities in the Tropics. PLOS ONE, 11, e0164957. Paine, C. E. T., Norden, N., Chave, J., Forget, P.-M., Fortunel, C., Dexter, K. G. & Baraloto, C. (2012) Phylogenetic density dependence and environmental filtering predict seedling mortality in a tropical forest. Ecology Letters, 15, 34-41. Pausas, J. G. & Verdú, M. (2010) The Jungle of Methods for Evaluating Phenotypic and Phylogenetic Structure of Communities. BioScience, 60, 614-625. Pérez-Harguindeguy, N., Díaz, S., Garnier, E., Lavorel, S., Poorter, H., Jaureguiberry, P., Bret-Harte, M. S., Cornwell, W. K., Craine, J. M., Gurvich, D. E., Urcelay, C., Veneklaas, E. J., Reich, P. B., Poorter, L., Wright, I. J., Ray, P., Enrico, L., Pausas, J. G., de Vos, A. C., Buchmann, N., Funes, G., Quétier, F., Hodgson, J. G., Thompson, K., Morgan, H. D., ter Steege, H., van der Heijden, M. G. A., Sack, L., Blonder, B., Poschlod, P., Vaieretti, M. V., Conti, G., Staver, A. C., Aquino, S. & Cornelissen, J. H. C. (2013) New handbook for standardised measurement of plant functional traits worldwide. Australian Journal of Botany, 61, 167-234. Pfennig, D. W., Rice, A. M. & Martin, R. A. (2006) Ecological Opportunity and Phenotypic Plasticity Interact to Promote Character Displacement and Species Coexistence. Ecology, 87, 769- 779. Pigot, A. L. & Etienne, R. S. (2015) A new dynamic null model for phylogenetic community structure. Ecology Letters, 18, 153-163.

133

Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. & Team, R. C. (2016) nlme: Linear and Nonlinear Mixed Effects Models. R package version 3.1-124. http://CRAN.R- project.org/package=nlme. Poorter, L. (1999) Growth responses of 15 rain-forest tree species to a light gradient: the relative importance of morphological and physiological traits. Functional Ecology, 13, 396-410. Purschke, O., Schmid, B. C., Sykes, M. T., Poschlod, P., Michalski, S. G., Durka, W., Kühn, I., Winter, M. & Prentice, H. C. (2013) Contrasting changes in taxonomic, phylogenetic and functional diversity during a long-term succession: insights into assembly processes. Journal of Ecology, 101, 857-866. R Core Team (2013) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Ricklefs, Robert E. (2008) Disintegration of the Ecological Community. The American Naturalist, 172, 741-750. Shooner, S., Chisholm, C. & Davies, T. J. (2015) The phylogenetics of succession can guide restoration: an example from abandoned mine sites in the subarctic. J Appl Ecol, 52: 1509– 1517. doi:10.1111/1365-2664.12517. Silvertown, J. (2004) Plant coexistence and the niche. Trends in Ecology & Evolution, 19, 605-611. Sniderman, J. M. K. & Jordan, G. J. (2011) Extent and timing of floristic exchange between Australian and Asian rain forests. Journal of Biogeography, 38, 1445-1455. Swenson, N. G. & Enquist, B. J. (2009) Opposing Assembly Mechanisms in a Neotropical Dry Forest: Implications for Phylogenetic and Functional Community Ecology. Ecology, 90, 2161-2170. Swenson, N. G., Enquist, B. J., Thompson, J. & Zimmerman, J. K. (2007) The Influence of Spatial and Size Scale on Phylogenetic Relatedness in Tropical Forest Communities. Ecology, 88, 1770- 1780. Valiente-Banuet, A. & Verdú, M. (2007) Facilitation can increase the phylogenetic diversity of plant communities. Ecology Letters, 10, 1029-1036. Verdú, M., Gómez-Aparicio, L. & Valiente-Banuet, A. (2012) Phylogenetic relatedness as a tool in restoration ecology: a meta-analysis. Proceedings of the Royal Society B: Biological Sciences, 279, 1761-1767. Violle, C., Enquist, B. J., McGill, B. J., Jiang, L., Albert, C. H., Hulshof, C., Jung, V. & Messier, J. (2012) The return of the variance: intraspecific variability in community ecology. Trends in Ecology & Evolution, 27, 244-252.

134

Walters, M. B. & Reich, P. B. (1999) Low-light carbon balance and shade tolerance in the seedlings of woody plants: do winter deciduous and broad-leaved evergreen species differ? New Phytologist, 143, 143-154. Webb, C. O. (2000) Exploring the Phylogenetic Structure of Ecological Communities: An Example for Rain Forest Trees. The American Naturalist, 156, 145-155. Webb, C. O., Ackerly, D. D. & Kembel, S. W. (2008) Phylocom: software for the analysis of phylogenetic community structure and trait evolution. Bioinformatics, 24, 2098-2100. Webb, C. O., Ackerly, D. D., McPeek, M. A. & Donoghue, M. J. (2002) Phylogenies and Community Ecology. Annual Review of Ecology and Systematics, 33, 475-505. Webb, C. O., Gilbert, G. S. & Donoghue, M. J. (2006) Phylodiversity-Dependent Seedling Mortality, Size Structure, and Disease in a Bornean Rain Forest. Ecology, 87, S123-S131. Wells, J. A. (2012) Phylogeny and inter-relations of ecological traits and seed dispersal in rainforest plants: exploring aspects of functional diversity in primary and secondary rainforests in Australia’s Wet Tropics PhD Thesis The University of Queensland Whitfeld, T. J. S., Kress, W. J., Erickson, D. L. & Weiblen, G. D. (2012) Change in community phylogenetic structure during tropical forest succession: evidence from New Guinea. Ecography, 35, 821-830. Wikström, N., Savolainen, V. & Chase, M. W. (2001) Evolution of the angiosperms: calibrating the family tree. Proceedings of the Royal Society B: Biological Sciences, 268, 2211-2220. Wills, J., Herbohn, J., Moreno, M. O. M., Avela, M. S. & Firn, J. (2016) Next-generation tropical forests: reforestation type affects recruitment of species and functional diversity in a human-dominated landscape. J Appl Ecol. doi:10.1111/1365-2664.12770. Wright, I. J., Reich, P. B., Westoby, M., Ackerly, D. D. & et al. (2004) The worldwide leaf economics spectrum. Nature, 428, 821-7.

135

Supplementary material Table S4.1. Phylogenetic diversity of seedling communities depending on the higher order fixed effects of forest type, soil nitrogen (N) and phosphorus (P) and leaf area index (LAI). Faith’s phylogenetic diversity (PD), the mean nearest taxon distance (MNTD), and mean pairwise distance (MPD) were used to assess phylogenetic diversity using different species pools and phylogenies that used different estimates of divergence times. Phylogenetic diversity Fixed DF F P P effects (num./den.) value (corrected) PD (observed) forest type 2/12 8.99 0.004 0.028 PD (observed) soil N 1/17 5.51 0.031 0.172 PD (observed) soil P 1/17 0.81 0.381 0.909 PD (observed) LAI 1/17 0.68 0.42 0.887 PD (Z-value) forest type 2/12 4.6 0.033 0.209 PD (Z-value) soil N 1/17 0.01 0.91 1.000 PD (Z-value) soil P 1/17 4.9 0.041 0.189 PD (Z-value) LAI 1/17 2.53 0.130 0.427 PD-natives (observed) forest type 2/12 18.9 <0.001 0.007 PD-natives (observed) soil N 1/17 6.15 0.024 0.136 PD-natives (observed) soil P 1/17 2.9 0.102 0.416 PD-natives (observed) LAI 1/17 0.03 0.85 0.999 MNTD-natives non-weighted (Z-value) forest type 2/12 6.56 0.012 0.081 MNTD-natives non-weighted (Z-value) soil N 1/17 0.04 0.845 1.000 MNTD-natives non-weighted (Z-value) soil P 1/17 0.32 0.578 0.987 MNTD-natives non-weighted (Z-value) LAI 1/17 0.44 0.519 0.946 MNTD-natives weighted (Z-value) forest type 2/12 1.98 0.181 0.753 MNTD-natives weighted (Z-value) soil N 1/17 0.002 0.967 1.000 MNTD-natives weighted (Z-value) soil P 1/17 0.002 0.966 1.000 MNTD-natives weighted (Z-value) LAI 1/17 1.19 0.292 0.749 PD-tree and shrub (Z-value) forest type 2/12 9.26 0.004 0.028 PD-tree and shrub (Z-value) soil N 1/17 0.41 0.529 0.989 PD-tree and shrub (Z-value) soil P 1/17 1.75 0.203 0.678 PD-tree and shrub (Z-value) LAI 1/17 6.62 0.02 0.078

MPD-tree and shrub non-weighted (Z-value) forest type 2/12 9.49 0.003 0.021 MPD-tree and shrub non-weighted (Z-value) soil N 1/17 0.77 0.392 0.949 MPD-tree and shrub non-weighted (Z-value) soil P 1/17 2.02 0.174 0.615 MPD-tree and shrub non-weighted (Z-value) LAI 1/17 2.03 0.172 0.530 MPD-tree and shrub weighted (Z-value) forest type 2/12 4.39 0.037 0.232 MPD-tree and shrub weighted (Z-value) soil N 1/17 0.89 0.358 0.930 MPD-tree and shrub weighted (Z-value) soil P 1/17 0.91 0.352 0.886 MPD-tree and shrub weighted (Z-value) LAI 1/17 0.90 0.357 0.829 MNTD-tree and shrub non-weighted (Z-value) forest type 2/12 9.74 0.003 0.021 MNTD-tree and shrub non-weighted (Z-value) soil N 1/17 1.02 0.326 0.906 MNTD-tree and shrub non-weighted (Z-value) soil P 1/17 0.29 0.599 0.990 MNTD-tree and shrub non-weighted (Z-value) LAI 1/17 3.87 0.066 0.239 MNTD-tree and shrub weighted (Z-value) forest type 2/12 2.59 0.116 0.578 MNTD-tree and shrub weighted (Z-value) soil N 1/17 0.22 0.644 0.998 MNTD-tree and shrub weighted (Z-value) soil P 1/17 0.00 0.992 1.000 MNTD-tree and shrub weighted (Z-value) LAI 1/17 0.48 0.5 0.938 MPD non-weighted (Z-value) forest type 2/12 3.00 0.088 0.475 MPD non-weighted (Z-value) soil N 1/17 0.00 0.966 1.000 MPD non-weighted (Z-value) soil P 1/17 2.85 0.11 0.442 MPD non-weighted (Z-value) LAI 1/17 1.18 0.292 0.749 MNTD non-weighted (Z-value) forest type 2/12 4.44 0.036 0.226 MNTD non-weighted (Z-value) soil N 1/17 0.24 0.631 0.997 MNTD non-weighted (Z-value) soil P 1/17 0.99 0.333 0.868 MNTD non-weighted (Z-value) LAI 1/17 2.0 0.173 0.532 MNTD weighted (Z-value) forest type 2/12 0.88 0.439 0.983 MNTD weighted (Z-value) soil N 1/17 0.04 0.85 1.000 MNTD weighted (Z-value) soil P 1/17 0.07 0.796 1.000 MNTD weighted (Z-value) LAI 1/17 1.32 0.266 0.710 MPD weighted (Z-value) forest type 2/12 0.93 0.421 0.978 MPD weighted (Z-value) soil N 1/17 0.76 0.395 0.951 MPD weighted (Z-value) soil P 1/17 1.25 0.278 0.804 MPD weighted (Z-value) LAI 1/17 0.62 0.442 0.903 MPD-wickstrom non-weighted (Z-value) forest type 2/12 1.39 0.285 0.904 MPD-wickstrom non-weighted (Z-value) soil N 1/17 0.02 0.879 1.000 MPD-wickstrom non-weighted (Z-value) soil P 1/17 1.85 0.192 0.656

136

MPD-wickstrom non-weighted (Z-value) LAI 1/17 2.34 0.144 0.463 MPD-wickstrom weighted (Z-value) forest type 2/12 0.57 0.579 0.998 MPD-wickstrom weighted (Z-value) soil N 1/17 0.48 0.499 0.984 MPD-wickstrom weighted (Z-value) soil P 1/17 0.81 0.379 0.908 MPD-wickstrom weighted (Z-value) LAI 1/17 0.083 0.375 0.847 MNTD-wickstrom non-weighted (Z-value) forest type 2/12 1.57 0.248 0.864 MNTD-wickstrom non-weighted (Z-value) soil N 1/17 0.00 0.961 1.000 MNTD-wickstrom non-weighted (Z-value) soil P 1/17 0.07 0.792 1.000 MNTD-wickstrom non-weighted (Z-value) LAI 1/17 1.34 0.263 0.705 MNTD-wickstrom weighted (Z-value) forest type 2/12 0.6 0.566 0.997 MNTD-wickstrom weighted (Z-value) soil N 1/17 0.11 0.748 1.000 MNTD-wickstrom weighted (Z-value) soil P 1/17 0.28 0.602 0.990 MNTD-wickstrom weighted (Z-value) LAI 1/17 1.23 0.283 0.736 MPD-Bayesian non-weighted (Z-value) forest type 2/12 1.72 0.220 0.824 MPD-Bayesian non-weighted (Z-value) soil N 1/17 0.09 0.773 1.000 MPD-Bayesian non-weighted (Z-value) soil P 1/17 2.9 0.107 0.432 MPD-Bayesian non-weighted (Z-value) LAI 1/17 1.63 0.219 0.628 MPD-Bayesian weighted (Z-value) forest type 2/12 0.74 0.496 0.992 MPD-Bayesian weighted (Z-value) soil N 1/17 0.73 0.404 0.955 MPD-Bayesian weighted (Z-value) soil P 1/17 1.25 0.279 0.805 MPD-Bayesian weighted (Z-value) LAI 1/17 0.56 0.463 0.917 MNTD-Bayesian non-weighted (Z-value) forest type 2/12 3.95 0.048 0.291 MNTD-Bayesian non-weighted (Z-value) soil N 1/17 0.25 0.623 0.997 MNTD-Bayesian non-weighted (Z-value) soil P 1/17 0.49 0.492 0.966 MNTD-Bayesian non-weighted (Z-value) LAI 1/17 1.00 0.330 0.798 MNTD-Bayesian weighted (Z-value) forest type 2/12 0.84 0.454 0.986 MNTD-Bayesian weighted (Z-value) soil N 1/17 0.04 0.846 1.000 MNTD-Bayesian weighted (Z-value) soil P 1/17 0.12 0.747 0.999 MNTD-Bayesian weighted (Z-value) LAI 1/17 1.23 0.283 0.736

Table S4.2. Functional Diversity of seedling communities depending on forest type, soil nitrogen (N) and phosphorus (P) and leaf area index (LAI). Mean functional distance (MFD) and mean nearest functional distance (MNFD) were used to assess the functional diversity for different traits and combinations, and for different species pools Functional diversity Fixed DF F value P P effects (num./den.) (corrected) SLA (mean/plot) forest type 2/12 9.0 0.004 0.028 SLA (mean/plot) soil N 1/17 0.05 0.821 1.000 SLA (mean/plot) soil P 1/17 0.00 0.996 1.000 SLA (mean/plot) LAI 1/17 0.72 0.410 0.879 MFD-SLA non-weighted (Z-value) forest type 2/12 1.92 0.189 0.769 MFD-SLA non-weighted (Z-value) soil N 1/17 3.84 0.067 0.340 MFD-SLA non-weighted (Z-value) soil P 1/17 0.46 0.508 0.971 MFD-SLA non-weighted (Z-value) LAI 1/17 2.61 0.125 0.414 MFD-SLA weighted (Z-value) forest type 2/12 2.96 0.09 0.483 MFD-SLA weighted (Z-value) soil N 1/17 2.14 0.161 0.651 MFD-SLA weighted (Z-value) soil P 1/17 0.04 0.844 1.000 MFD-SLA weighted (Z-value) LAI 1/17 1.09 0.311 0.775 MNFD-SLA non-weighted (Z-value) forest type 2/12 1.7 0.224 0.831 MNFD-SLA non-weighted (Z-value) soil N 1/17 4.7 0.044 0.237 MNFD-SLA non-weighted (Z-value) soil P 1/17 1.02 0.327 0.862 MNFD-SLA non-weighted (Z-value) LAI 1/17 1.3 0.269 0.714 MNFD-SLA weighted (Z-value) forest type 2/12 0.27 0.77 1.000 MNFD-SLA weighted (Z-value) soil N 1/17 3.63 0.074 0.370 MNFD-SLA weighted (Z-value) soil P 1/17 1.41 0.251 0.764 MNFD-SLA weighted (Z-value) LAI 1/17 0.32 0.577 0.968 MFD-N, P non-weighted (Z-value) forest type 2/12 3.00 0.088 0.475 MFD-N, P non-weighted (Z-value) soil N 1/17 0.00 0.966 1.000 MFD-N, P non-weighted (Z-value) soil P 1/17 2.85 0.11 0.442 MFD-N, P non-weighted (Z-value) LAI 1/17 1.18 0.292 0.749 MNFD-P weighted (Z-value) forest type 2/12 1.49 0.264 0.883 MNFD-P weighted (Z-value) soil N 1/17 0.01 0.911 1.000 MNFD-P weighted (Z-value) soil P 1/17 0.04 0.842 1.000 MNFD-P weighted (Z-value) LAI 1/17 1.32 0.267 0.711 MNFD-P non-weighted (Z-value) forest type 2/12 0.32 0.734 1.000

137

MNFD-P non-weighted (Z-value) soil N 1/17 0.18 0.671 0.999 MNFD-P non-weighted (Z-value) soil P 1/17 0.97 0.339 0.874 MNFD-P non-weighted (Z-value) LAI 1/17 3.05 0.099 0.341 MFD-P non-weighted (Z-value) forest type 2/12 0.10 0.902 1.000 MFD-P non-weighted (Z-value) soil N 1/17 0.00 1.000 1.000 MFD-P non-weighted (Z-value) soil P 1/17 0.86 0.367 0.898 MFD-P non-weighted (Z-value) LAI 1/17 0.49 0.494 0.934 MFD-P weighted (Z-value) forest type 2/12 1.64 0.234 0.845 MFD-P weighted (Z-value) soil N 1/17 0.04 0.839 1.000 MFD-P weighted (Z-value) soil P 1/17 0.86 0.366 0.898 MFD-P weighted (Z-value) LAI 1/17 0.12 0.735 0.995 MNFD-N weighted (Z-value) forest type 2/12 2.23 0.150 0.679 MNFD-N weighted (Z-value) soil N 1/17 1.31 0.268 0.846 MNFD-N weighted (Z-value) soil P 1/17 0.46 0.505 0.970 MNFD-N weighted (Z-value) LAI 1/17 0.86 0.368 0.840 MNFD-N non-weighted (Z-value) forest type 2/12 1.57 0.249 0.865 MNFD-N non-weighted (Z-value) soil N 1/17 6.47 0.021 0.120 MNFD-N non-weighted (Z-value) soil P 1/17 0.21 0.656 0.995 MNFD-N non-weighted (Z-value) LAI 1/17 2.27 0.15 0.478 MFD-N non-weighted (Z-value) forest type 2/12 0.44 0.653 0.999 MFD-N non-weighted (Z-value) soil N 1/17 2.11 0.165 0.661 MFD-N non-weighted (Z-value) soil P 1/17 0.45 0.511 0.972 MFD-N non-weighted (Z-value) LAI 1/17 1.08 0.313 0.777 MFD-N weighted (Z-value) forest type 2/12 0.3 0.748 1.000 MFD-N weighted (Z-value) soil N 1/17 1.22 0.285 0.866 MFD-N weighted (Z-value) soil P 1/17 0.05 0.829 1.000 MFD-N weighted (Z-value) LAI 1/17 1.48 0.24 0.666 MFD-all traits non-weighted (Z-value) forest type 2/12 0.41 0.673 1.000 MFD-all traits non-weighted (Z-value) soil N 1/17 4.53 0.049 0.260 MFD-all traits non-weighted (Z-value) soil P 1/17 1.72 0.207 0.686 MFD-all traits non-weighted (Z-value) LAI 1/17 1.24 0.281 0.733 MFD-all traits weighted (Z-value) forest type 2/12 1.25 0.323 0.935 MFD-all traits weighted (Z-value) soil N 1/17 1.72 0.207 0.751 MFD-all traits weighted (Z-value) soil P 1/17 0.03 0.875 1.000 MFD-all traits weighted (Z-value) LAI 1/17 0.26 0.618 0.979 MNFD-all traits non-weighted (Z-value) forest type 2/12 0.95 0.415 0.977 MNFD-all traits non-weighted (Z-value) soil N 1/17 3.22 0.091 0.436 MNFD-all traits non-weighted (Z-value) soil P 1/17 0.08 0.781 0.999 MNFD-all traits non-weighted (Z-value) LAI 1/17 6.26 0.023 0.089 MNFD-all traits weighted (Z-value) forest type 2/12 3.55 0.062 0.361 MNFD-all traits weighted (Z-value) soil N 1/17 4.05 0.060 0.310 MNFD-all traits weighted (Z-value) soil P 1/17 0.29 0.6 0.990 MNFD-all traits weighted (Z-value) LAI 1/17 4.12 0.058 0.213 MNFD-all traits-natives weighted (Z-value) forest type 2/12 0.2 0.823 1.000 MNFD-all traits-natives weighted (Z-value) soil N 1/17 1.26 0.279 0.860 MNFD-all traits-natives weighted (Z-value) soil P 1/17 4.69 0.047 0.214 MNFD-all traits-natives weighted (Z-value) LAI 1/17 2.98 0.105 0.358 MFD-all traits-natives non-weighted (Z-value) forest type 2/12 0.34 0.717 1.000 MFD-all traits-natives non-weighted (Z-value) soil N 1/17 1.17 0.297 0.879 MFD-all traits-natives non-weighted (Z-value) soil P 1/17 1.06 0.321 0.856 MFD-all traits-natives non-weighted (Z-value) LAI 1/17 0.03 0.856 1.000

138

Table S4.3. Intraspecific variation in SLA for species with five or more measurements, using the mean nearest functional distance (MNFD) and the mean functional distance (MFD). Functional diversity Fixed DF F value P P effects (num./den.) (corrected) CV (plot) forest type 2/12 6.59 0.012 0.079 CV (plot) soil N 1/17 3.74 0.07 0.353 CV (plot) soil P 1/17 0.69 0.42 0.997 CV (plot) LAI 1/17 0.15 0.71 0.478

MNFD-CV weighted (Z-value) forest type 2/12 15.32 <0.001 0.007 MNFD-CV weighted (Z-value) soil N 1/17 16.62 <0.001 0.006 MNFD-CV weighted (Z-value) soil P 1/17 1.34 0.264 0.784 MNFD-CV weighted (Z-value) LAI 1/17 5.76 0.029 0.111

MFD-CV weighted (Z-value) forest type 2/12 14.91 <0.001 0.007 MFD-CV weighted (Z-value) soil N 1/17 10.24 0.006 0.035 MFD-CV weighted (Z-value) soil P 1/17 1.66 0.215 0.702 MFD-CV weighted (Z-value) LAI 1/17 3.89 0.066 0.239 MNFD-CV non-weighted (Z-value) forest type 2/12 5.99 0.016 0.107 MNFD-CV non-weighted (Z-value) soil N 1/17 1.88 0.19 0.718 MNFD-CV non-weighted (Z-value) soil P 1/17 2.01 0.176 0.620 MNFD-CV non-weighted (Z-value) LAI 1/17 0.18 0.679 0.989 MFD-CV non-weighted (Z-value) forest type 2/12 5.82 0.017 0.113 MFD-CV non-weighted (Z-value) soil N 1/17 2.68 0.121 0.539 MFD-CV non-weighted (Z-value) soil P 1/17 1.64 0.219 0.709 MFD-CV non-weighted (Z-value) LAI 1/17 0.11 0.746 0.996

Table S4.4. Phylogenetic signal of categorical traits Categorical traits P Life-form (Herb, Vine, Palm, Shrub, Tree) <0.001 Dispersal (Biotic, Abiotic) 0.001 Potential height (Understorey, Mid-canopy, <0.001 Canopy, Emergent)

139

Figure S4.1. Coefficient of variation (CV) of SLA for species with 5 or more individuals plotted against the phylogeny that included Bayesian estimates of divergence times.

140

Mono Rain. Regen. Dispersal Height SLA SLA (biotic/abiotic) (variation)

Moraceae

Psidium guajava

Theobroma cacao

Dipterocarpaceae

Citrus maxima Lansium parasiticum Sandoricum koetjape

Mangifera indica

Chrysophyllum cainito

141

Figure S4.2. Phylogeny incorporating Bayesian estimates of divergence times, with the occurrence of each species indicated by circles for monoculture (Mono), Rainforestation (Rain.) and regenerating selectively logged forest (Regen.). Traits including dispersal type (abiotic = large circle and biotic = small circle), potential plant height, mean species SLA and within species variation in SLA are plotted as tip labels, and are scaled to the size of the circle. Green represents species within the Moraceae family, which were favoured across forest types, and red represents tall wind-dispersed trees that were limited within monoculture forest types. Blue represents human- dispersed plant species. Black dots represent species that show high and low variation and mean SLA values, which co-occurred within regenerating selectively logged forest.

142

Chapter 5. Leaf traits predict growth rates of sub-canopy trees but not canopy trees in Australian regenerating selectively logged tropical forests

5.1. Introduction Easy to measure morphological plant traits, sometimes referred to as ‘soft traits’ are commonly used to inform practitioners about hard to measure ecological processes (Funk et al., 2016). The interest of this trait-based approach is increasingly being applied to the reforestation of degraded tropical forests. A central tenet in all reforestation projects is how to accelerate plant growth and how, depending on tree size, individual tree stem growth measures will vary depending on species. The linking of tropical tree functional traits to growth rates have yielded some generalized applications; however, for some traits this link has been idiosyncratic or equivocal (Rüger et al., 2012, Poorter et al., 2008, Wright et al., 2010a). Due to the high number and longevity of tropical tree species, planting trials that assess growth rates through time are often unfeasible for many species (Martínez-Garza et al., 2005). Some tropical tree species are also known to exhibit changes in their ranked performance as they mature and increase in size (Martínez-Garza et al., 2013). For example, pioneer tree species might have fast initial growth rates within light gaps, but as the light gaps are filled growth rates might reduce to levels less than shade-tolerate species (Brokaw, 1985). This change in performance over time likely occurs both with pioneer species (generally shade-intolerant, fast-growing species with shorter lifespans) and also secondary species (generally shade-tolerant, slow-growing species with long lifespans). Understanding these changes in performance could lead to a more nuanced understanding of community assembly and the implications for designing reforestation strategies. Failing the establishment of field trials for multiple species and combinations, one of the few ways to gain an understanding of tree performance for species-rich ecosystems is to establish permanent sample plots (PSPs) and to monitor tree dynamics regularly through time. This information can be used to gain insights into many ecosystem processes from local to global scales (Bowman et al., 2013). Proxy measures of tree performance have also received considerable attention, due to the large resources required for establishing field trials and PSPs. One of the most feasible proxies for depicting complex ecological processes like growth is the use of simple to obtain and measure leaf functional traits. To date, evidence of the correlation between leaf functional traits and growth in complex and diverse tropical forests has been equivocal (Wright et al., 2010a). The Leaf Economic Spectrum (LES), based on the idea of economy of resources (Bloom et al., 1985), reflects variation in plant growth rates (Wright et al., 2004), through the costs associated with leaf tissue construction, trading off with tissue longevity. Typically, it is expected that plants

143 with low construction costs (e.g., high Specific Leaf Area, SLA) and short tissue longevity will have fast growth rates. For tree seedlings/saplings, leaf traits such as SLA have been shown to be positively related to growth rates and support the LES growth-survival trade-off (Wright and Westoby, 1999, Lambers and Poorter, 2004, Wright et al., 2010a, Hérault et al., 2011). Given that studies have found the LES traits explain a high proportion of variation of plant strategies across ecosystems, it is expected that mature fast-growing trees would show higher SLA and leaf nutrient values compared to slow-growing trees. However, this does not always hold in studies conducted in structurally complex and diverse tropical forests (Poorter et al., 2008, Wright et al., 2010a, Gibert et al., 2016). Mature tree leaf trait and growth relationships show shifts through ontogeny, derived from size dependence and changes in abiotic conditions (e.g., light) (Iida et al., 2014). Theoretical predictions regarding tree growth using leaf traits have often been supported at the seedling and sapling stage, with relationships becoming absent as trees mature and the relative importance of other traits on growth increases (e.g., wood density, number of active meristems) (King, 1999, Hérault et al., 2011). However, when and how growth and leaf trait relationships change and the mechanisms behind this are less known, particularly for tropical forests. If leaf traits can be used to predict the performance of tree species in diverse mixtures, traits may provide land managers with more information about how forests are functioning including how seedlings will perform during the early and later stages of their life cycle. Growth rates based on seedling performance under favourable conditions have found support for differential growth strategies that are reflected in leaf trait syndromes, such as SLA (Poorter and Markesteijn, 2008). However, a recent study found that the correlations between growth and leaf traits are not consistent (Gibert et al., 2016). The practical application of using averaged leaf trait values to design reforestation projects might be hindered by these differences in trait-growth relationships through plant development or through secondary cross-correlations with other evolutionary or environmental factors (e.g., light). Given that previous studies have revealed that leaf trait and growth relationships are complicated, particularly as trees mature, this may limit the applicability of adult tree leaf traits for revealing useful ecosystem functions for practitioners. Here, we test this assumption by using long-term growth data (i.e., an ~50 year dataset) and measured leaf traits within a diverse tropical regenerating selectively logged natural forest, and ask the following questions: 1. Are tree growth rates correlated with soft and easy to measure leaf traits, e.g. SLA and leaf nutrients? 2. Are relationships between tree growth rates and leaf traits consistent depending on the positioning in the canopy strata?

144

3. Are relationships between leaf traits consistent between canopy strata?

We predict that tree growth rates will correlate positively with SLA, N and P, and have a negative relationship with C and the ratio of C : N (C:N). We also predict that these associations will become weaker as trees mature (Gibert et al., 2016).

5.2. Materials and Methods

5.2.1. Study Sites The study sites are located within Danbulla National Park and State Forest, Atherton Tableland, North Queensland, Australia (-ᵒ17.17’S, 145ᵒ58’E). Surrounding forest and plantations cover an area of more than 12,000ha. The elevation ranges from 680 to 790 m.a.s.l. and mean annual rainfall from 1370 to 1650 mm. In this study we use a total of four plots, which we refer to here as experiment 625 plots 4 and 5 and experiment 78 plots 1 and 2, this is done in order to remain consistent with historic data described below. Experiment 625 plots 4 and 5 have soil of granitic origin and experiment 78 plot 1of volcanic origin and in the case of plot 2 being metamorphic (Table 5.1). The vegetation has been classified under the Wet Tropics Management Authority vegetation classification (WTMA., 2009) as Simple Notophyll Vine Forests (SNVF) type 10a. More specifically the floristics within experiment 78 is comprised of species, such as Argyrodendron peralatum and Aleurites moluccana, which occur across a range of elevations (0-1000m.a.s.l). Within experiment 625 species such as Flindersia pimenteliana and Flindersia bourjotiana occur, which prefer the slightly higher elevations found at these plots. Experiment 78 is arranged as a paired plot, with two identical plots measuring 200 m long and 20 m wide. A 20 m buffer zone separates the two plots. Plots within Experiment 625 are geographically distinct and are 40 m long by 50 m wide (Figure 5.1).

5.2.2. Historical Data and Site History The plots used in this study form part of a permanent plot network established by the Queensland Department of Forestry (referred to as the QFS permanent plots). These plots were established for research into sustainable yield. Tree species and diameter at breast height (DBH) were measured for all trees ≥10 cm DBH at regular periods dating back to 1948 for experiment 78 and 1968 for experiment 625 (Table 5.1). Experiment 78 was measured at intervals that ranged from 1-17 years, and experiment 625 ranged from 5-17 years. In both cases the latest remeasure was the longest interval (17 years). Plots were located using a combination of historic maps and personal communication. Individual trees were found by first identify trees that were within the plot using

145 steel tags or finding old paint. For approximately 50% of the trees, tree numbers could be identified using the tree tag or painted number. For the remaining trees the number sequence, species identity and size were used to identify the individual tree. This study uses the historical datasets for a subset of the trees within each experiment (Figure 5.1). Experiment 78 was established as a paired observation plot that aimed to measure the effects of silvicultural treatment on regeneration and growth responses. Plot one was established as a control plot that received no silvicultural treatment and plot two received a silvicultural treatment in 1949, prescribed by the 1946 rules, consisting of the thinning of non-commercial species. Experiment 625 was established in order to provide Basel Area (BA) data for logged forests. Five plots were established as detailed yield plots, two of which are used within the current study (plots four and five). In experiment 78, plot one was selectively logged in 1943 and plot two in 1931. In experiment 625, plot four was selectively logged in 1952 and plot five in 1970.

5.2.3. Tree selection and sampling Sampled trees were identified by firstly re-measuring all trees ≥10 cm in each plot and aligning with the historical data obtained from the QFS database. Historical data for experiment 78 was measured at 18 intervals since 1948 and for experiment 625 at 7 intervals since 1968. Time between measurement intervals varied between 1 and 17 years. All individual trees’ growth rates (measured as Basal Area Increment (BAI), Periodic Annual Increment (PAI) and Relative Growth Rate (RGR)) were analysed in order to identify individual trees that have exhibited fast growth or slow growth over the measurement periods. BA was calculated using the formula퐵퐴 = 0.00007854 × 퐷퐵퐻(푐푚)2, and BAI is defined here by the change in BA between measurement periods, divided by the number of years within the measurement period (Da Cunha et al., 2016). PAI is defined here by the change in DBH between measurement periods, divided by the number of years within the growing period. RGR was calculated using the standard formula, 푅퐺푅 = ln(퐷퐵퐻2) − ln(퐷퐵퐻1) ÷

(푦푒푎푟2 − 푦푒푎푟1). A minimum of three individuals per species are presented, in order for adequate replication. Each tree’s measurements included its height, distance to nearest neighbour and canopy illumination index (CII) (as per Clark and Clark (1992)) , and the slope and aspect of its position were measured. Canopy leaves were collected using smaller traditional sling shots and for large individual trees a Big Shot line launcher was used. Five fully expanded sun leaves were collected per tree following the protocols by Pérez-Harguindeguy et al. (2013). The leaves were then labelled and placed into paper bags. The leaf area was measured on the same day as sampling using a CID Bio-Science CI-203 Laser Area Meter. These leaf samples were then oven dried at 65ᵒC for 48

146 hours, then weighed to calculate Specific Leaf Area (SLA, cm2/g) for each of the five leaves. SLA values presented here are the average from the 5 leaves collected per individual. To obtain leaf nutrients (in % dry leaf mass), the five leaves were bulked, milled and later analysed using the carbon and nitrogen combustion analysis and the Inductively Coupled Plasma Optical Emission Spectroscopy analysis (ICP-OES). For each tree, 16 leaf traits were measured; however using two axes from a principal component analysis and correlation coefficients SLA, Nitrogen (N) and Phosphorus (P), significantly explained the variation in Calcium (Ca), Boron (B), Iron (Fe), Potassium (K), Copper (Cu), Zinc (Zn), Magnesium (Mg) and Sodium (Na). Carbon (C) and Carbon to Nitrogen ratio (C: N) were significantly negatively correlated with the aforementioned nutrients and in addition to Leaf Area to Nitrogen ratio (LA: N) were retained in the analysis. These results indicate that SLA, N, P, C, C: N and LA: N explains the majority of variation in leaf traits (Figure S5.1).

5.2.4. Data Analysis All analyses were conducted within R statistical computing version 3.1.1 (R Core Team, 2013). We first analysed the relationship between leaf traits and growth rates using simple bivariate spearman correlations within the Hmisc package (Hollander and Wolfe, 1973). Generalized linear mixed models (GLMM) were then used to further investigate leaf trait and growth rate relationships. These relationships were estimated using maximum likelihood, in which nested random effects were plots within experiments, within species, as representing our sampling design. Models were fit using the LME4 package (Bates et al., 2015) and, nlme package (Pinheiro et al., 2016), and model selection was undertaken using the MUMIN package (Barton, 2013). Within the GLMM models, BAI, PAI and RGR were considered response variables. The traits SLA, N, P, C, LA: N, C: N, tree height and initial DBH were considered explanatory variables. The response variables PAI, BAI and RGR and the explanatory variables P, SLA and LA: N had non-normal distributions and were log transformed. Linear regressions were used to test correlations between the response and explanatory variables and these relationships are shown graphically.

147

Figure 5.1. Map of study sites and stem maps of the four plots studied. The size of the circles is proportional to DBH size in 2015. Black filled dots represent the trees that were sampled for leaf functional traits.

148

Table 5.1. Site information for the four plots studied, including the QFS experiment name and plot number, the area of the plots and plot environmental variables, and the year of first measurement. Experiment Plot Area Soil Parent Altitude Aspect Slope Rainfall First (ha) Material (masl) (deg) (mm/year) measurement 78 1 0.4047 Volcanic 680 NNW 5 1320 1948 78 2 0.4047 Metamorphic 680 NNW 5 1320 1948 625 4 0.2023 Coarse 790 E 15 1650 1968 Granite 625 5 0.2023 Coarse 730 SE 20 1650 1968 Granite

5.3. Results Overall, 162 trees were sampled for their structural characteristics, growth rates, SLA and leaf nutrients. The sampled trees comprised 54 species, of which 33 species had between three and six sampled individuals, and 21 species had two or less individuals. The trees ranged from 10 to 80 cm DBH and from 3 to 28 m in height.

5.3.1. Are trees growth rates correlated with leaf traits? Using trees from all canopy strata, tree height and C: N together best explained the variation in PAI (Table S5.1); height was positively correlated with PAI (Figure 5.2a) and negatively correlated with leaf C: N (Figure 5.2b). Height was positively correlated with BAI, and height and initial DBH together explained the variation in RGR and again was positively correlated (Figure S5.2a-c and Tables S5.2 and S5.3). Generally, using simple bivariate correlations, leaf traits for all trees showed absent to moderate relationships with growth rates. C: N was significantly negatively correlated with PAI (Figure 2b). SLA, C and LA: N was not significantly correlated with PAI, however N and P was significantly positively correlated with PAI. BAI was not significantly correlated with N, P, C, C: N or LA: N, but was significantly negatively correlated with SLA. RGR was significantly negatively correlated with SLA, and positively correlated with C. However, N, P, C: N and LA: N did not significantly correlate with RGR.

149

Table 5.2. Results from simple bivariate correlations between leaf traits and Basal Area Increment (BAI), Periodic Annual Increment (PAI) and Relative Growth Rate (RGR), and whether the result was expected or unexpected based on theoretical predictions for one or more of the growth measurements. “-“= no significant relationship. Leaf traits BAI PAI RGR Expected ( ) r P r P r P Unexpected ( )

SLA -0.30 <0.001 -0.05 0.52 -0.32 <0.001 N 0.05 0.555 0.38 <0.001 -0.003 0.96 P 0.06 0.436 0.29 <0.001 0.08 0.3 C 0.09 0.248 0.01 0.937 0.24 <0.001 C: N -0.09 0.979 -0.35 <0.001 0.06 0.47 LA: N 0.13 0.404 0.13 0.739 0.09 0.243 -

Figure 5.2. Tree height (a) and C: N (b) explained the variation in Periodic Annual Increment (PAI), using trees from all canopy strata. The line represents the linear regression between the two variables.

150

5.3.2. Relationships between growth and leaf traits for different canopy strata Correlations between leaf traits and growth rates for canopy trees For canopy trees, height significantly explained the variation in PAI, BAI and RGR and the relationship was positive. However, for RGR initial DBH was also retained in the simplest model and was positively associated (Figure S3a-c). Similarly to trees from all canopy strata, for canopy trees in isolation PAI significantly negatively correlated with C: N and N was positively correlated. SLA, P, C and leaf area: N was not significantly correlated to PAI. BAI and RGR were not significantly correlated to any of the measured leaf traits (Table XX).

Table 5.3. Results from simple bivariate correlations between canopy tree leaf traits and Basal Area Increment (BAI), Periodic Annual Increment (PAI) and Relative Growth Rate (RGR), and whether the result was expected or unexpected based on theoretical predictions for one or more of the growth measurements. “-“= no significant relationship. Trait BAI PAI RGR Expected ( ) r P r P r P Unexpected ( ) SLA -0.15 0.176 -0.02 0.885 -0.15 0.175 - N 0.21 0.054 0.38 <0.001 0.17 0.111 P 0.10 0.35 0.17 0.12 0.16 0.149 - C -0.01 0.913 -0.09 0.404 -0.04 0.711 - C: N -0.18 0.09 -0.35 <0.001 -0.16 0.148 LA: N 0.17 0.112 0.14 0.188 0.11 0.325 -

Correlations between leaf traits and growth rates for sub-canopy trees PAI of sub-canopy trees significantly positively correlated with P; however, P was not significantly correlated with PAI when only canopy trees were considered (Figure 3a). P was also the strongest predictor of PAI for sub-canopy trees. For the BAI of sub-canopy trees, height and initial DBH positively explained the variation and initial DBH alone explained the variation in sub-canopy RGR (Figure SXXX and Tables S3a-c). PAI was significantly negatively correlated with C: N, and positively with SLA, N and P (Figure 3a-d). BAI and RGR were again not significantly correlated to any of the measured leaf traits. Table 5.4. Results from simple bivariate correlations between sub-canopy tree leaf traits and Basal Area Increment (BAI), Periodic Annual Increment (PAI) and Relative Growth Rate (RGR), and

151 whether the result was expected or unexpected based on theoretical predictions for one or more of the growth measurements. “-“= no significant relationship. Leaf traits BAI PAI RGR Expected ( ) r P r P r P Unexpected ( )

SLA -0.02 0.881 0.24 0.034 -0.07 0.561 N 0.05 0.659 0.51 <0.001 -0.06 0.637 P 0.18 0.11 0.52 <0.001 0.07 0.526 C 0.03 0.83 -0.17 0.135 0.12 0.289 - C: N -0.05 0.694 -0.50 <0.001 0.08 0.504 LA: N 0.04 0.712 0.09 0.431 0.01 0.91 -

Figure 5.3. The relationships between Periodic Annual Increment and (a) P, (b) N, (c) SLA and C: N. P and SLA were significantly positively correlated with PAI for sub-canopy trees, but not for canopy trees.

152

Table 5.5. The relative importance for the first three variables derived from the LMEMs for all canopy strata and when restricted to the canopy and sub-canopy trees. * denotes if that variable was retained in the simplest model within Δ AICc = 4, for explaining the variation in Periodic Annual Increment (PAI), Basel Area Increment (BAI) and Relative Growth Rate (RGR).

Forest strata Growth Height DBH0 P C:N N LA:N SLA All PAI 1.00* 0.72 - 0.82* - - - BAI 1.00* 0.87 - 0.46 - - - RGR 1.00* 1.00* - - 0.78 - - Canopy PAI 0.99* 0.57 - 0.58 - - - BAI 1.00* 0.68 - - - - 0.30 RGR 1.00* 1.00* 0.40 - - - - Sub-canopy PAI - - 0.89* - 0.71 0.67 - BAI 0.94* 0.47* 0.52 - - - - RGR 0.83 1.00* - - - - 0.32

5.3.3. Are relationships between leaf traits and between leaf traits and tree height consistent between canopy strata? Overall, for all canopy strata, growth rates were positively correlated with height. For all canopy strata and canopy trees in isolation, SLA and height were significantly negatively correlated (all strata: r = -0.49, P < 0.001 and canopy: r = -0.49, P < 0.001). But when considering only sub- canopy trees, SLA and height were not significantly correlated (r = -0.07, P = 0.561) (Figure 5.4a). SLA and leaf N were significantly positively correlated when considering all canopy strata and when only considering the canopy (all strata: r = 0.39, P < 0.001 and canopy: r = 0.27, P = 0.013); however, SLA and N had a stronger association within the sub-canopy (r = 0.56, P < 0.001) (Figure 5.4c). SLA and P were significantly positively correlated and for both the canopy and sub-canopy, with the correlation being stronger with sub-canopy trees (all strata: r = 0.42, P < 0.001, canopy: r = 0.43, P < 0.001 and sub-canopy: r = 0.50, P < 0.001) (Figure 5.4.b). N and P were strongly positively correlated when considering both all strata and canopy and sub-canopy trees respectively; again the relationship was strongest within the sub-canopy trees (all strata: r = 0.66, P < 0.001, canopy: r = 0.53, P < 0.001 and sub-canopy: r = 0.79, P < 0.001) (Figure 5.4d).

153

Figure 5.4. Specific leaf area (SLA) was significantly negatively correlated with tree height for canopy trees (a). The correlations between SLA and nitrogen (N), and phosphorus (b, c) were stronger for sub-canopy trees than for canopy trees. The correlation between N and P was also stronger for sub-canopy trees than for trees occurring within the canopy (d).

5.4. Discussion Using a 50-year dataset of tree growth within regenerating selectively logged native tropical forest, we evaluate whether leaf traits can be used to predict growth rates within canopy and sub-canopy life-stages. We find that leaf traits are stronger predictors of growth rates for trees growing within the light-limiting sub-canopy than for canopy trees. Our results build on other studies that have found weak to absent relationships between adult tree leaf traits and growth rates (Poorter et al., 2008, Wright et al., 2010a). Our study also highlights that light availability is likely a key factor influencing leaf trait and growth relationships, and trade-offs between leaf traits. These results reflect the myriad of internal and external factors that interact and change throughout a trees’ development to ultimately determine its growth. Canopy trees are positioned to receive maximum light levels and soil nutrientsthey are in strong positions competitively within tropical forests; where, sub-canopy trees are shaded by canopy trees and therefore likely to be less competitive. The different measures of growth (PAI, BAI and RGR) also displayed different relationships to leaf traits. Generally, when considering canopy or sub-canopy trees in isolation, PAI was more frequently significantly correlated to leaf traits, and this relationship was generally more supported by theoretical predictions (Tables 5.2, 5.3 and 5.4).

154

5.4.1. Are tree growth rates related to leaf traits? Our initial expectations that SLA, N and P would correlate positively with growth rates, and that C will correlate negatively with growth rates, were partially supported. This was because overall, the leaf traits showed weak associations with growth, particularly when considering trees that were found in the canopy. However, the ratio of C: N supported our initial predictions, by correlating negatively with one of our measures of growth (i.e., PAI) (Figure 5.2), and along with height this ratio was also retained within the simplest model explaining variation in PAI. Changes in leaf numbers, overall leaf area, level of self-shading, number of active meristems, and partitioning of support tissue of mature trees have all been suggested as factors confounding mature tree leaf trait and growth rate relationships (King, 1999, Poorter et al., 2008, Wright et al., 2010a, Hérault et al., 2011). The overall leaf area of a canopy tree is influenced by the number of active meristems rather than by adjusting SLA. Whether at the seedling life-stage rather than the canopy life-stage , changes in SLA occurs in order to adjust overall light interception and therefore could more directly link to growth (Sterck and Bongers, 2001). The level of self-shading within the canopy of trees is also suggested as an important factor in leaf trait and growth relationships, indicating that in the outer canopy, fully expanded leaves may be intercepting light in order for optimal light interception of more internal leaves. Variation in partitioning of support tissue within trees compared to seedlings has also been suggested as another factor confounding leaf trait and growth relationships (King, 1999, Hérault et al., 2011).

5.4.2. Growth rate and leaf trait relationships between canopy strata In general, we found that leaf trait and growth relationships are stronger when considering sub- canopy trees compared to trees that have gained dominance in the canopy. These findings therefore suggest that leaf traits are better predictors of growth in light-limiting environments. For example similar relationships between leaf traits and size dependency have been found in Malaysian forests (Iida et al. (2014). When light was not limiting (e.g., even below the canopy) wood density was not as important a predictor of growth and mortality, than when light was limiting. When isolating growth and leaf trait relationships between canopy strata, we found for canopy trees, tree height and initial DBH were the strongest predictors of all measures of growth. However, for sub-canopy trees P alone positively correlated with PAI and BAI (Table 2). This result supported our initial prediction that P would correlate positively with growth rates, but only for trees growing within the low-light sub-canopy. This association between growth and P, particularly within light-limiting environments, has also been observed in other tropical forest ecosystems. For example, P was the only leaf nutrient to correlate positively with sapling relative growth rates within central Amazonian forest (Marenco et

155 al., 2015). Foliar P concentrations are comprised of P fractions derived from biochemical compounds (e.g. lipids and nucleic acids) and inorganic P. In our study, the relative concentrations and proportions P fractions determining overall foliar P were not measured. Trees growing on low P soils have been found to have overall reduced foliar P requirements as they have adapted by reducing metabolic and nuclei acid P concentrations (Hidaka and Kitayama, 2011). Tropical soils (including in parts of the Wet Tropics bioregion of North Queensland) are generally deficient in P, perhaps suggesting that variation in leaf P within our samples is predominantly explained by variations in metabolic and nuclei acid P fractions, due to soil P limitations (Prescott, 1941). SLA was positively correlated with PAI for trees occurring within the sub-canopy, but was not significantly correlated with the growth of canopy trees. These results suggest that increasing SLA becomes a more important strategy to increase photosynthesis and growth in plants that are occurring within light-limiting environments. Producing leaves with a higher SLA would provide plants growing under low light conditions with more leaf mass per area to intercept light and reduce metabolic costs associated with deploying new leaves (Sterck and Bongers, 2001). In comparison within canopy trees where light is not limiting, the total number of leaves and the leaf and crown area are possibly more important traits to increase photosynthesis and growth (King, 1999, Sterck and Bongers, 2001). Our results differ to Iida et al. (2014) who found that within a subtropical forest of Taiwan there was a negative correlation between SLA and growth (RGR) at smaller tree sizes and a positive correlation at larger tree sizes. However, Poorter and Bongers (2006) also found a positive relationship between SLA and growth, but did not consider the size dependence of this relationship. SLA values for canopy trees were consistently lower than for trees occurring within the sub-canopy. This finding is supported by other studies (Sterck and Bongers, 2001, Houter and Pons, 2012, Verbeeck et al., 2014) and suggests that plants adapt to light-limiting environments by increasing their SLA values. Within the canopy trees may adapt to high light levels by reducing their SLA values in order to maintain a water balance. Similar to foliar P concentrations, SLA values can arise from variation in multiple fractions including chlorophyll or secondary defence production. This highlights the limitations in using simple morphological traits for predicting tree growth in complex forests like those found in the tropics (Ellsworth and Reich, 1993).

5.4.3. Are relationships between leaf traits consistent between canopy strata? The relationships between SLA, N and P within our study revealed strong correlations for trees found in the sub-canopy (Figure 5.4b-d). Theoretical predictions based on the LES appear to be more strongly supported within sub-canopy conditions, and are less fundamental as trees enter the

156 canopy environment, where other un-accounted for traits may become important (Wright et al., 2010a). Variation in the relationship between leaf traits between canopy strata is likely a result of a change in the restraints imposed on plants by light limitation, and may be beneficial for maximising the amount of photosynthetic tissue available for light interception. Given the generality of these correlations across biomes (Reich et al., 1999), it is not surprising that we find similar correlations in fundamental growth-survival trade-offs in leaf function (Shipley et al., 2006). What is surprising from this study is that these correlations are consistently stronger for trees occurring within the sub-canopy environment. What is unclear from these results is whether the changes in leaf correlations between canopy strata are determined by internal developmental processes, are environmentally induced (through variations in microclimate) and in response to higher competition for light within the sub-canopy, or if the changes are a combination of all of these factors. We consider the trends are most likely explained by the latter. How these correlations scale up to whole plant trade-offs operating at different stages of tropical trees’ development requires further research. These results, however, highlight that competition for light within the sub- canopy is a strong determinant of leaf trait relationships and that these relationships are less constrained in the canopy where competition for light is reduced.

5.5. Implications for reforestation Leaf traits can predict growth trajectories at early stages of plant development (e.g., seedling and sapling), but become less indicative of demographic rates as trees mature. Traits associated with stem economics such as potential plant height and wood density have been shown to better predict growth rates across multiple size classes within tropical forests (Iida et al., 2014, Hérault et al., 2011, Rüger et al., 2012), and might provide practitioners with better predictive power than canopy tree leaf traits. Stem and leaf traits offer different axes of plant strategies and consequently offer unique insights into plant ecological strategies, at different ontogenetic stages. Leaf traits for trees growing in complex ecosystems like tropical forests provide mixed insights into complex ecological processes such as growth, especially once trees have entered the canopy, where their predictive ability is weak. However, our results also confirm that leaf traits (particularly P) correlate with growth rates within the sub-canopy, where light availability is low. Overall the relative importance of different leaf traits on growth rates changes systematically with tree size (Iida et al., 2014), and due to this, leaf traits (particularly SLA) may not be useful for describing growth strategies of tropical tree species in isolation (Rüger et al., 2012). We recommend that in order for leaf traits to be more informative for reforestation design, size-dependence of leaf traits and growth relationships should be more carefully considered, particularly when reforestation practitioners

157 assign mean trait values to tropical tree species from multiple light environments (e.g., canopy strata).

References

Barton, K. 2013. MuMIn: multi-model inference, R package version 1.9.13 [Online]. http://CRAN.R-project.org/package=MuMIn. Bates, D., Mächler, M., Bolker, B. & Walker, S. 2015. Fitting Linear Mixed-Effects Models Using lme4. 2015, 67, 48. Bloom, A. J., Chapin Iii, F. S. & Mooney, H. A. 1985. Resource limitation in plants - an economic analogy. Annual review of ecology and systematics. Vol. 16, 363-392. Bowman, D. M. J. S., Brienen, R. J. W., Gloor, E., Phillips, O. L. & Prior, L. D. 2013. Detecting trends in tree growth: not so simple. Trends in Plant Science, 18, 11-17. Brokaw, N. V. L. 1985. Gap-Phase Regeneration in a Tropical Forest. Ecology, 66, 682-687. Clark, D. A. & Clark, D. B. 1992. Life History Diversity of Canopy and Emergent Trees in a Neotropical Rain Forest. Ecological Monographs, 62, 315-344. Da Cunha, T. A., Finger, C. a. G. & Hasenauer, H. 2016. Tree basal area increment models for Cedrela, Amburana, Copaifera and Swietenia growing in the Amazon rain forests. Forest Ecology and Management, 365, 174-183. Ellsworth, D. S. & Reich, P. B. 1993. Canopy structure and vertical patterns of photosynthesis and related leaf traits in a deciduous forest. Oecologia, 96, 169-178. Funk, J. L., Larson, J. E., Ames, G. M., Butterfield, B. J., Cavender-Bares, J., Firn, J., Laughlin, D. C., Sutton-Grier, A. E., Williams, L. & Wright, J. 2016. Revisiting the Holy Grail: using plant functional traits to understand ecological processes. Biological Reviews, DOI: 10.1111/brv.12275. Gibert, A., Gray, E. F., Westoby, M., Wright, I. J. & Falster, D. S. 2016. On the link between functional traits and growth rate: meta-analysis shows effects change with plant size, as predicted. Journal of Ecology, n/a-n/a. Hérault, B., Bachelot, B., Poorter, L., Rossi, V., Bongers, F., Chave, J., Paine, C. E. T., Wagner, F. & Baraloto, C. 2011. Functional traits shape ontogenetic growth trajectories of rain forest tree species. Journal of Ecology, 99, 1431-1440.

158

Hidaka, A. & Kitayama, K. 2011. Allocation of foliar phosphorus fractions and leaf traits of tropical tree species in response to decreased soil phosphorus availability on Mount Kinabalu, Borneo. Journal of Ecology, 99, 849-857. Hollander, M. & Wolfe, D. A. 1973. Nonparametric Statistical Methods. New York: Wiley.

Press WH, Flannery BP, Teukolsky SA, Vetterling, WT (1988): Numerical Recipes in C. Cambridge:

Cambridge University Press. Houter, N. C. & Pons, T. L. 2012. Ontogenetic changes in leaf traits of tropical rainforest trees differing in juvenile light requirement. Oecologia, 169, 33-45. Iida, Y., Kohyama, T. S., Swenson, N. G., Su, S.-H., Chen, C.-T., Chiang, J.-M. & Sun, I. F. 2014. Linking functional traits and demographic rates in a subtropical tree community: the importance of size dependency. Journal of Ecology, 102, 641-650. King, D. A. 1999. JUVENILE FOLIAGE AND THE SCALING OF TREE PROPORTIONS, WITH EMPHASIS ON EUCALYPTUS. Ecology, 80, 1944-1954. Lambers, H. & Poorter, H. 2004. Inherent Variation in Growth Rate Between Higher Plants: A Search for Physiological Causes and Ecological Consequences. Advances in Ecological Research. Academic Press. Marenco, R. A., Magalhães, N. D. S., Gouvêa, P. R. D. S. & Antezana-Vera, S. A. 2015. Juvenile tree growth correlates with photosynthesis and leaf phosphorus content in central Amazonia. Revista Ceres, 62, 175-183. Martínez-Garza, C., Bongers, F. & Poorter, L. 2013. Are functional traits good predictors of species performance in restoration plantings in tropical abandoned pastures? Forest Ecology and Management, 303, 35-45. Martínez-Garza, C., Peña, V., Ricker, M., Campos, A. & Howe, H. F. 2005. Restoring tropical biodiversity: Leaf traits predict growth and survival of late-successional trees in early- successional environments. Forest Ecology and Management, 217, 365-379. Pérez-Harguindeguy, N., Díaz, S., Garnier, E., Lavorel, S., Poorter, H., Jaureguiberry, P., Bret- Harte, M. S., Cornwell, W. K., Craine, J. M., Gurvich, D. E., Urcelay, C., Veneklaas, E. J., Reich, P. B., Poorter, L., Wright, I. J., Ray, P., Enrico, L., Pausas, J. G., De Vos, A. C., Buchmann, N., Funes, G., Quétier, F., Hodgson, J. G., Thompson, K., Morgan, H. D., Ter Steege, H., Van Der Heijden, M. G. A., Sack, L., Blonder, B., Poschlod, P., Vaieretti, M. V., Conti, G., Staver, A. C., Aquino, S. & Cornelissen, J. H. C. 2013. New handbook for standardised measurement of plant functional traits worldwide. Australian Journal of Botany, 61, 167-234.

159

Pinheiro, J., Bates, D., Debroy, S., Sarkar, D. & Team, R. C. 2016. nlme: Linear and Nonlinear Mixed Effects Models [Online]. http://CRAN.R-project.org/package=nlme. Poorter, L. & Bongers, F. 2006. Leaf Traits are Good Predictors of Plant Performance Across 53 Rain Forest Species. Ecology, 87, 1733-1743. Poorter, L. & Markesteijn, L. 2008. Seedling Traits Determine Drought Tolerance of Tropical Tree Species. Biotropica, 40, 321-331. Poorter, L., Wright, S. J., Paz, H., Ackerly, D. D., Condit, R., Ibarra-Manríquez, G., Harms, K. E., Licona, J. C., Martínez-Ramos, M., Mazer, S. J., Muller-Landau, H. C., Peña-Claros, M., Webb, C. O. & Wright, I. J. 2008. Are Functional Traits Good Predictors of Demographic Rates? Evidence From Five Neotropical Forests. Ecology, 89, 1908-1920. Prescott, J. A. 1941. The soils of tropical Australia. Australian Geographer, 4, 16-19. R Core Team. 2013. R: A language and environment for statistical

computing. [Online]. R Foundation for Statistical Computing, Vienna, Austria. Available: URL http://www.R-project.org/. Reich, P. B., Ellsworth, D. S., Walters, M. B., Vose, J. M., Gresham, C., Volin, J. C. & Bowman, W. D. 1999. Generality of Leaf Trait Relationships: A Test Across Six Biomes Ecology, 80, 1955-1969. Rüger, N., Wirth, C., Wright, S. J. & Condit, R. 2012. Functional traits explain light and size response of growth rates in tropical tree species. Ecology, 93, 2626-2636. Shipley, B., Lechowicz, M. J., Wright, I. & Reich, P. B. 2006. Fundamental Trade-offs Generating the Worldwide Leaf Econmic Spectrum Ecology, 87, 535-541. Sterck, F. J. & Bongers, F. 2001. Crown development in tropical rain forest trees: patterns with tree height and light availability. Journal of Ecology, 89, 1-13. Verbeeck, H., Betehndoh, E., Maes, W. H., Hubau, W., Kearsley, E., Buggenhout, L., Hufkens, K., Huygens, D., Vanacker, J., Beeckman, H., Mweru, J. P. M., Boeckx, P. & Steppe, K. 2014. Functional Leaf Trait Diversity of 10 Tree Species In Congolese Secondary Tropical Forest. Journal of Tropical Forest Science, 26, 409-419. Wright, I. J., Reich, P. B., Westoby, M., Ackerly, D. D. & Et Al. 2004. The worldwide leaf economics spectrum. Nature, 428, 821-7. Wright, I. J. & Westoby, M. 1999. Differences in seedling growth behaviour among species: trait correlations across species, and trait shifts along nutrient compared to rainfall gradients. Journal of Ecology, 87, 85-97. Wright, S. J., Kitajima, K., Kraft, N. J. B., Reich, P. B., Wright, I. J., Bunker, D. E., Condit, R., Dalling, J. W., Davies, S. J., Díaz, S., Engelbrecht, B. M. J., Harms, K. E., Hubbell, S. P.,

160

Marks, C. O., Ruiz-Jaen, M. C., Salvador, C. M. & Zanne, A. E. 2010. Functional traits and the growth–mortality trade-off in tropical trees. Ecology, 91, 3664-3674. Wtma., W. T. M. A. 2009. Vegetation Mapping of the Wet Tropics of Queensland Bioregion. In: AUTHORITY, W. T. M. (ed.). Cairns.

Supplementary Material

Table S5.1. Models within Δ AICc = 4 for explaining the variation in Periodic Annual Increment (PAI). Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Random effects were species/exp/plot, representing our sampling design.

Component models df logLik AICc Delta Weight 2+3+4 8 -200.71 418.36 0.00 0.09 2+3+4+5 9 -200.18 419.54 1.18 0.05 2+3+4+7 9 -200.31 419.81 1.45 0.04 2+4 7 -202.57 419.87 1.51 0.04 2+3+4+6 9 -200.43 420.04 1.68 0.04 1+2+3+4 9 -200.44 420.07 1.71 0.04 2+3+4+8 9 -200.63 420.44 2.08 0.03 3+4+8 8 -201.78 420.51 2.15 0.03 1+2+3+4+5 10 -199.63 420.73 2.37 0.03 2+3+4+5+6 10 -199.93 421.32 2.96 0.02 2+3+4+5+7 10 -199.95 421.36 3.00 0.02 2+4+6 8 -202.21 421.37 3.01 0.02 2+3+4+5+8 10 -200.01 421.48 3.12 0.02 1+2+3+4+7 10 -200.01 421.49 3.13 0.02 2+4+5 8 -202.31 421.56 3.19 0.02 1+2+4 8 -202.33 421.61 3.25 0.02 2+4+7 8 -202.35 421.64 3.28 0.02 3+4+6+8 9 -201.28 421.74 3.38 0.02

161

3+4+5+8 9 -201.31 421.80 3.44 0.02 1+2+3+4+6 10 -200.17 421.80 3.44 0.02 2+3+4+6+7 10 -200.19 421.83 3.47 0.02 3+4+7+8 9 -201.42 422.02 3.66 0.01 2+3+4+7+8 10 -200.28 422.03 3.66 0.01 2+4+8 8 -202.56 422.07 3.71 0.01 2+3+4+6+8 10 -200.35 422.16 3.80 0.01 1+2+3+4+8 10 -200.42 422.30 3.94 0.01

Table S5.2. Models within Δ AICc = 4 for explaining the variation in Basel Area Increment (BAI). Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N.

Component models df logLik AICc Delta Weight 3+4 7 -207.32 429.37 0.00 0.09 2+3+4+8 9 -205.46 430.11 0.74 0.06 2+3+4 8 -206.79 430.52 1.15 0.05 3+4+5 8 -206.89 430.71 1.34 0.04 3+4+6 8 -206.90 430.74 1.37 0.04 1+2+3+4+8 10 -204.97 431.40 2.03 0.03 3+4+7 8 -207.26 431.47 2.10 0.03 3+4+8 8 -207.31 431.55 2.18 0.03 1+3+4 8 -207.32 431.58 2.21 0.03 2+3+4+5+8 10 -205.17 431.79 2.42 0.03 2+3+4+6+8 10 -205.32 432.09 2.72 0.02 3+4+5+6 9 -206.56 432.31 2.94 0.02 2+3+4+7+8 10 -205.43 432.32 2.95 0.02 2+3+4+5 9 -206.62 432.42 3.05 0.02 1+2+3+4 9 -206.66 432.51 3.14 0.02 2+3+4+6 9 -206.69 432.56 3.19 0.02

162

2+3+4+7 9 -206.78 432.75 3.38 0.02 3+4+5+7 9 -206.81 432.81 3.44 0.02 3+4+6+8 9 -206.85 432.88 3.51 0.02 4 6 -210.19 432.91 3.54 0.01 3+4+5+8 9 -206.87 432.92 3.55 0.01 1+3+4+6 9 -206.88 432.94 3.57 0.01 1+3+4+5 9 -206.89 432.96 3.59 0.01 3+4+6+7 9 -206.90 432.98 3.61 0.01 1+2+3+4+5+8 11 -204.81 433.38 4.01 0.01 1+2+3+4+6+8 11 -204.83 433.42 4.05 0.01 1+2+3+4+7+8 11 -204.91 433.59 4.22 0.01 1+3+4+7 9 -207.26 433.71 4.34 0.01 3+4+7+8 9 -207.26 433.71 4.34 0.01 2+3+4+5+6+8 11 -204.98 433.72 4.35 0.01 1+3+4+8 9 -207.31 433.80 4.43 0.01 2+3+4+5+7+8 11 -205.05 433.87 4.50 0.01

Table S5.3. Models within Δ AICc = 4 for explaining the variation in Relative Growth Rate (RGR). Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N.

Component models df logLik AICc Delta Weight 1+2+3+4+8 10 283.83 -546.20 0.00 0.06 3+4+5+6+8 10 283.78 -546.09 0.10 0.06 1+2+3+4+6+8 11 284.90 -546.03 0.17 0.06 2+3+4+5+6+8 11 284.85 -545.95 0.25 0.06 2+3+4+8 9 282.53 -545.87 0.33 0.05 2+3+4+6+8 10 283.61 -545.77 0.42 0.05 2+3+4+5+8 10 283.49 -545.53 0.66 0.04 1+2+3+4+5+6+8 12 285.69 -545.29 0.91 0.04 3+4+6+8 9 282.22 -545.27 0.93 0.04 1+2+3+4+5+8 11 284.39 -545.02 1.17 0.03 3+4 7 279.72 -544.70 1.49 0.03

163

1+3+4+5+6+8 11 283.91 -544.05 2.14 0.02 3+4+5+8 9 281.56 -543.94 2.26 0.02 1+2+3+4+7+8 11 283.83 -543.89 2.30 0.02 1+2+3+4+6+7+8 12 284.98 -543.87 2.32 0.02 3+4+5+6+7+8 11 283.78 -543.80 2.39 0.02 3+4+8 8 280.35 -543.76 2.43 0.02 2+3+4+6+7+8 11 283.76 -543.76 2.44 0.02 2+3+4+5+6+7+8 12 284.86 -543.63 2.57 0.02 2+3+4+7+8 10 282.54 -543.61 2.58 0.02 3+4+6 8 280.27 -543.59 2.61 0.02 1+3+4+6+8 10 282.50 -543.55 2.65 0.02 3+4+5 8 280.20 -543.47 2.73 0.02 3+4+6+7+8 10 282.40 -543.35 2.85 0.02 2+3+4+5+7+8 11 283.53 -543.31 2.89 0.01 1+3+4 8 280.07 -543.20 3.00 0.01 1+2+3+4+5+6+7+8 13 285.69 -542.93 3.27 0.01 3+4+7 8 279.90 -542.86 3.34 0.01 3+4+6+7 9 280.98 -542.78 3.41 0.01 1+2+3+4+5+7+8 12 284.43 -542.78 3.42 0.01 1+3+4+6 9 280.90 -542.62 3.57 0.01 2+3+4 8 279.72 -542.49 3.70 0.01 3+4+5+6 9 280.63 -542.08 4.12 0.01

164

Canopy trees Table S5.4. Models within Δ AICc = 4 for explaining the variation in Periodic Annual Increment (PAI), for canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N. Component models df logLik AICc Delta Weight 2+3+4 8 -112.31 242.48 0.00 0.08 2+4 7 -113.97 243.37 0.88 0.05 3+4 7 -114.09 243.62 1.14 0.04 2+3+4+8 9 -111.88 244.13 1.65 0.03 4 6 -115.63 244.33 1.84 0.03 2+4+8 8 -113.24 244.35 1.86 0.03 2+3+4+6 9 -112.01 244.38 1.90 0.03 3+4+8 8 -113.33 244.52 2.04 0.03 1+2+3+4 9 -112.29 244.96 2.47 0.02 1+3+4 8 -113.54 244.96 2.47 0.02 2+3+4+5 9 -112.31 244.98 2.50 0.02 2+3+4+7 9 -112.31 244.98 2.50 0.02

Table S5.5. Models within Δ AICc = 4 for explaining the variation in Basel Area Increment (BAI), for canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N.

Component models df logLik AICc Delta Weight 3+4 7 -113.36 242.16 0.00 0.11 3+4+5 8 -112.96 243.79 1.63 0.05 2+3+4 8 -112.97 243.82 1.66 0.05 4 6 -115.40 243.87 1.71 0.05 3+4+7 8 -113.05 243.96 1.80 0.04 3+4+8 8 -113.25 244.38 2.22 0.04 3+4+6 8 -113.36 244.59 2.43 0.03 1+3+4 8 -113.36 244.60 2.43 0.03

165

Table S5.6. Models within Δ AICc = 4 for explaining the variation in Relative Growth Rate (RGR), for canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N.

Component models df logLik AICc Delta Weight 3+4 7 148.20 -280.96 0.00 0.13 3+4+6 8 149.04 -280.21 0.75 0.09 3+4+5 8 148.62 -279.38 1.59 0.06 2+3+4 8 148.47 -279.07 1.90 0.05 1+3+4+6 9 149.58 -278.80 2.16 0.04 1+3+4 8 148.30 -278.74 2.22 0.04 3+4+7 8 148.30 -278.72 2.24 0.04 3+4+8 8 148.27 -278.67 2.29 0.04 3+4+5+6 9 149.43 -278.49 2.47 0.04 3+4+6+8 9 149.06 -277.75 3.21 0.03 3+4+6+7 9 149.05 -277.72 3.24 0.03 2+3+4+6 9 149.04 -277.71 3.25 0.03 1+2+3+4 9 148.94 -277.51 3.45 0.02 2+3+4+5 9 148.74 -277.11 3.85 0.02 3+4+5+7 9 148.73 -277.08 3.88 0.02 2+3+4+8 9 148.69 -277.01 3.95 0.02 1+3+4+5 9 148.67 -276.97 3.99 0.02 3+4+5+8 9 148.63 -276.88 4.08 0.02 1+3+4+5+6 10 149.82 -276.70 4.26 0.02 1+3+4+8 9 148.51 -276.64 4.32 0.01 2+3+4+7 9 148.48 -276.59 4.37 0.01 1+3+4+7 9 148.47 -276.57 4.39 0.01 1+2+3+4+6 10 149.70 -276.47 4.49 0.01 3+4+7+8 9 148.32 -276.27 4.69 0.01 1+3+4+6+8 10 149.59 -276.25 4.71 0.01 3+4+5+6+8 10 149.59 -276.24 4.72 0.01 1+3+4+6+7 10 149.58 -276.23 4.73 0.01 2+3+4+5+6 10 149.47 -276.00 4.96 0.01

166

Sub-canopy results

Table S5.7. Models within Δ AICc = 4 for explaining the variation in Periodic Annual Increment (PAI), for sub-canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N.

Component models df logLik AICc Delta Weight 5+6+8 8 -84.59 187.33 0.00 0.10 4+5+6+8 9 -83.97 188.67 1.34 0.05 1+5+6+8 9 -84.28 189.29 1.96 0.04 3+5+6+8 9 -84.36 189.45 2.12 0.03 2+5+6+8 9 -84.53 189.79 2.46 0.03 3+4+5+6+8 10 -83.26 189.90 2.57 0.03 5+6+7+8 9 -84.59 189.91 2.58 0.03 6+8 7 -87.14 189.92 2.59 0.03 6 6 -88.77 190.77 3.44 0.02 1+4+5+6+8 10 -83.71 190.81 3.48 0.02 5+8 7 -87.62 190.89 3.56 0.02 4+5+6+7+8 10 -83.90 191.19 3.86 0.01 2+4+5+6+8 10 -83.93 191.25 3.92 0.01 1+3+5+6+8 10 -83.95 191.29 3.96 0.01 2+5+6 8 -86.59 191.33 4.00 0.01 4+6+8 8 -86.61 191.37 4.04 0.01 6+7+8 8 -86.79 191.73 4.40 0.01 2+6 7 -88.05 191.75 4.42 0.01 3+4+6+7 9 -85.53 191.78 4.45 0.01 1+3+4+5+6+8 11 -82.84 191.81 4.48 0.01 1+2+5+6 9 -85.55 191.82 4.49 0.01 4+6 7 -88.09 191.82 4.49 0.01

167

Table S5.8. Models within Δ AICc = 4 for explaining the variation in Basel Area Increment (BAI), for sub-canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N.

Component models df logLik AICc Delta Weight 3+4 7 -94.71 205.06 0.00 0.06 4+6 7 -94.84 205.33 0.28 0.05 4 6 -96.06 205.34 0.28 0.05 4+6+8 8 -93.68 205.51 0.45 0.05 3+4+6 8 -93.78 205.71 0.65 0.04 3+4+6+8 9 -92.91 206.55 1.49 0.03 4+7 7 -95.58 206.81 1.75 0.02 2+4+6 8 -94.39 206.93 1.88 0.02 3+4+7 8 -94.54 207.23 2.17 0.02 4+5+6 8 -94.59 207.33 2.27 0.02 3+4+5 8 -94.63 207.42 2.36 0.02 2+3+4 8 -94.68 207.51 2.45 0.02 1+3+4 8 -94.70 207.54 2.48 0.02 3+4+8 8 -94.70 207.55 2.49 0.02 1+4+6 8 -94.71 207.57 2.51 0.02 2+3+4+6 9 -93.45 207.62 2.57 0.02 4+5 7 -96.00 207.65 2.59 0.02 1+4 7 -96.03 207.71 2.65 0.02 2+4 7 -96.04 207.73 2.67 0.02 4+8 7 -96.04 207.74 2.68 0.02 4+6+7 8 -94.80 207.74 2.68 0.02 3+4+5+6 9 -93.54 207.81 2.75 0.01 1+4+6+8 9 -93.56 207.86 2.80 0.01 4+6+7+8 9 -93.57 207.87 2.81 0.01 2+4+6+8 9 -93.60 207.92 2.87 0.01 4+5+6+8 9 -93.68 208.08 3.03 0.01 1+3+4+6 9 -93.70 208.12 3.07 0.01 3+4+6+7 9 -93.78 208.28 3.22 0.01

168

4+7+8 8 -95.38 208.91 3.85 0.01 3 6 -97.86 208.94 3.88 0.01 1+3+4+6+8 10 -92.83 209.05 3.99 0.01 2+3+4+6+8 10 -92.85 209.09 4.03 0.01

Table S5.9. Models within Δ AICc = 4 for explaining the variation in Relative Growth Rate (RGR), for sub-canopy trees. Models included the following fixed effects: 1= C, 2 = C: N ratio, 3 = initial DBH, 4 = height, 5 = leaf area: N ratio, 6 = log(P), 7 = log(SLA) and 8 = N.

Component models df logLik AICc Delta Weight 3+4 7 147.90 -280.15 0.00 0.12 3+4+8 8 148.32 -278.50 1.65 0.05 3+4+5 8 148.22 -278.28 1.87 0.05 3+4+7 8 148.21 -278.27 1.88 0.05 1+3+4 8 148.14 -278.12 2.03 0.04 3+4+6 8 148.00 -277.85 2.30 0.04 2+3+4 8 147.99 -277.83 2.32 0.04 3+4+6+8 9 149.17 -277.62 2.54 0.03 3+4+5+7 9 148.63 -276.53 3.62 0.02 2+3+4+8 9 148.58 -276.43 3.72 0.02 3+4+6+7 9 148.57 -276.40 3.75 0.02 3 6 144.81 -276.40 3.75 0.02 1+3+4+8 9 148.49 -276.25 3.90 0.02 3+7 7 145.95 -276.25 3.90 0.02 1+3+4+5 9 148.47 -276.21 3.95 0.02 3+4+7+8 9 148.45 -276.17 3.98 0.02 1+3+4+7 9 148.41 -276.10 4.05 0.02 3+4+5+6 9 148.41 -276.08 4.07 0.02 3+4+5+8 9 148.38 -276.03 4.12 0.0

169

Figure S5.1. Principle Components Analysis (PCA) of tree leaf traits (a) and structural characteristics (b). Labels for a) correspond to chemical element, except for LA.N: leaf area to nitrogen ratio, C.N: carbon to nitrogen ratio and SLA.ind: Specific Leaf Area (SLA) of the individual tree. PCA a) indicates that SLA, N, P, C and C: N explains the majority of variation in leaf traits. For b) nn: nearest neighbour, DBH: Diameter at Breast Height, CII: Canopy Illumination Index and ht: tree height. PCA b) indicates that CII and ht explain similar amounts of variation.

170

(a) (b)

B RGR

Height (m) Height (m)

DBH0 (mm) Figure S5.2. Important variables for explaining the variation in BAI (a) and RGR (b and c), using trees from all canopy strata. Height was positively correlated with BAI and RGR, and initial DBH was positively associated with RGR.

171

Figure S5.3. Important variables in explaining the variation in growth measured as Periodic Annual Increment (PAI), Relative Growth Rate (RGR) and Basel Area Increment (BAI). Tree height and initial DBH (DBH0) were positively associated with growth (a-c), and foliar P concentrations were positively associated with PAI for sub-canopy trees (d).

172

Chapter 6. General Conclusion

6.1 Summary Overall, in my PhD research I have aimed to understand ecological processes in order to better inform reforestation and the conservation of degraded tropical forest. I did this by analysing different aspects of biodiversity from two tropical regions. Chapters 3 and 4 aimed to understand the ecological process of understorey plant recruitment beneath contrasting reforestation types on Leyte Island in the Philippines. Chapter 5 aimed to test how leaf traits can be used to understand tree growth within regenerating selectively logged natural forest of the WTs bioregion of Australia.

6.2 A synthesis of understorey recruitment beneath contrasting reforestation types

6.2.1. Summary of key outcomes The different ways to reforest cleared land vary in the degree to which they provide biodiversity and socio-economic outcomes (Lamb et al., 2005). Small-scale community-based reforestation projects have a different set of goals and a unique set of management techniques that markedly differ from monoculture plantings. These include increasing biodiversity and socio-economic values by mixing plantation species and using selective harvesting approaches. Understorey development can then be seen as value-adding, in terms of socio-economics, biodiversity and longer-term sustainability. In particular, the understorey can be of value if it can be left to mature and extend the rotation times of these plantations. Chapter 3 highlights the value of reforestation in general. The results of this study showed that even establishing exotic monocultures can provide understorey conditions conducive to the recruitment of species favoured by local communities (e.g., Theobroma cacao) and a sub-set of native forest species (e.g., Ficus septica). However, it also provides evidence that native wind-dispersed tree species are limited across human-dominated landscapes. This group of species form important emergent layers within tropical forest and also provide valuable timber resources throughout Asia. We recommend that this functional group, in addition to other limited functional groups such as large-seeded species, be emphasized within reforestation planting designs and conservation initiatives in order to enhance their future biodiversity and socio-economic values. Chapter 4 increases our understanding of understorey recruitment beneath the same contrasting reforestation types, by using continuous leaf traits, discrete traits and phylogenetic diversity. The findings show that the mechanisms of community assembly are operating differently depending on the reforestation type. For example, monoculture understoreys assemble by environmental filtering, primarily through dispersal limitation or human-assisted dispersal. The understories of regenerating

173 selectively logged forests comprise of species with both high and low variation and mean SLA values, leading to higher total leaf trait diversity; indicating that both competitive exclusion and environmental filtering seem to be operating. Chapter 4 also extends our knowledge of the importance of wind-dispersed tree species by providing evidence that these species tend to be the tallest trees within these forests, taking into consideration phylogenetic non-independence through shared ancestry. The main conclusion from this chapter is that reforestation projects can increase their potential conservation and socio-economic values by planting species that are phylogenetically and functionally distinct (in terms of the mean and variation in leaf traits), resulting in greater fulfilment of niche space and ecosystem functioning.

6.2.2. Limitations and further work A limitation of chapters 3 and 4 is the lack of temporal and spatial replication of understorey recruitment data. In particular, the long-term survival of recruited individuals is unknown. Therefore, this data represents a snap-shot of understorey recruitment at the time of sampling. The continuous plant functional traits analysed in chapter 4 are also known to vary seasonally and through ontogenetic development (Gibert et al., 2016). Spatially, the recruitment data is replicated considerably on Leyte Island, including five replicates of each forest type. However, the data is not replicated across multiple tropical forest regions. Conducting similar studies within different tropical regions, and testing if the trends in functional diversity are consistent between continents would be a logical development of this research. Within chapters 3 and 4, ‘pristine’ forest was not available for comparison. Recruitment diversity within regenerating selectively logged native forest was instead used as a ‘baseline’ for comparison. This problem is not unique to the Philippines and highlights the difficulties associated with finding undisturbed tropical forest systems in order to perform comparative ecological studies (Laurance, 1999). A next-step of this work would be to re-sample within the same plantations and assess shifts in species and functional trait diversity, and composition as the understorey matures. The sampled plantations represent small-holder community-based reforestation projects, and in some cases at the time of writing this thesis, have already been harvested or will be in the future. The influence of overstorey harvesting and the techniques used (e.g., selective or clear fell) on understorey development would likely yield important insights into the persistence of different species and traits into the next generation overstorey. In addition, since the time of sampling the Island of Leyte was exposed to Super Typhoon Yolanda, which was the most intense tropical typhoon on record. The

174 influence of this climatic event on the plantation overstorey and subsequent impacts on the understorey would also provide insights relating to tropical forest recovery. Functional diversity patterns can also vary considerably in traits and syndromes that were not accounted for within these studies, and could possibly indicate alternative mechanisms of community assembly. These could include the prevalence of fruiting events, including mast fruiting and seed dormancy, influencing the temporal and spatial arrangement of reproductive output across the study region (Buoro and Carlson, 2014). We selected the traits used within chapters 3 and 4 to represent multiple facets of a plantations’ understorey ecology. The traits relate to species colonization and dispersal abilities, to competitive interactions, and to growth rates (Pérez- Harguindeguy et al., 2013). The phylogeny that was used within chapter 4 includes species that were found at all forest types to construct the regional species pool, which was the ‘baseline’ to compare phylogenetic diversity metrics found within each of the contrasting reforestation types. Despite being a commonly used method to determine the potential phylogenetic diversity at a site, perhaps it would be more insightful to use a phylogeny that included all species found on the Island of Leyte (Lessard et al., 2012). This information could be gathered from the literature, which includes vegetation sampling on Leyte. This could potentially give a more accurate representation of the phylogenetic diversity that existed in the past. Another further research inquiry arising from chapters 3 and 4 is to quantify the relative frequency of wind- and animal-dispersed tree species across tropical regions. Animal-dispersed including human-dispersed tree species comprise the majority of tropical trees; however, within many tropical regions including the Philippine and WTs bioregions wind-dispersed species often dominate the canopy and the total basal area within the forests of these regions. Which factors such as biogeographic affiliations and environmental variables such as rainfall, elevation, cyclone history and disturbance regime that influence the relative frequency of animal- versus wind-dispersed tree species could be compared from different tropical continents? The biomass or carbon stocks stored within animal- or wind-dispersed tree species could also be compared between canopy strata and tropical forest regions.

6.3 A synthesis of predicting growth rates using simple morphological leaf traits

6.3.1. Summary of key outcomes Understanding the ability of morphological plant functional traits to predict demographic rates (e.g., growth, mortality and recruitment) within diverse tropical forests is an important goal of ecological research (Iida et al., 2014). This is because demographic rates for most tropical tree species are difficult to determine due to the high number and longevity of tropical trees (Bowman et al., 2013).

175

Permanent sample plots (PSP) offer a method to measure demographic rates, however, PSPs involve a large amount of resources because of the costs of maintenance and the requirement for repeated sampling occurring over multiple decades (Bowman et al., 2013). Therefore the ability to predict demographic rates using simple to obtain and measure morphological traits would greatly contribute to the knowledge of tropical tree species. This knowledge can then be used to better design and manage reforestation projects across the tropics. In chapter 5, I used historical data (dating back to as early as 1948) derived from PSPs within regenerating selectively logged natural forest of tropical North Queensland, to test if leaf traits can predict growth rates. I found that leaf traits are generally weak predictors of growth rates for trees once they reach the canopy layer. However, within the light limiting sub-canopy, leaf traits are a stronger predictor of growth rates. In particular leaf phosphorus was an important predictor of growth for trees occurring within the sub-canopy. Correlations based on theoretical predictions between leaf traits were also stronger for sub-canopy trees than for canopy trees, where light availability is not a limiting factor. The key outcome for reforestation applications from this chapter is that mean leaf traits derived from multiple canopy strata and used to design reforestation projects may not be the most reliable predictor of how to grow mixtures of species. The size-dependence of functional traits and demographic relationships, such as growth rates, should also be considered when making predictions of community assembly and forest succession.

6.3.2. Limitations and further work Temporal replication in leaf traits is also lacking in chapter 5, which can mask seasonal variations in leaf trait and growth relationships (McKown et al., 2013). This chapter also highlights a limitation of using leaf traits of canopy trees to inform ecologists about demographic rates. Therefore, we find that the life stage when leaf traits are collected does matter if these traits are to have practical use. Leaf traits collected from species while they are seedlings or saplings may better inform growth rates for reforestation with diverse mixtures. This finding has important implications for the generic use of leaf trait values from trait databases like TRY (Kattge et al., 2011). Further work within chapter five’s research plots could extend to the linking of additional traits such as wood density and potential plant height. These traits have been shown to be better predictors of demographic rates across multiple life-stages, to growth, mortality or recruitment (Hérault et al., 2011, Nguyen et al., 2014). The collection of seedling leaf traits could also quantify how and when species leaf traits change through ontogenetic development, and how this relates to seedling, sapling and tree demographic rates. More broadly, the use of functional traits to reveal information about tropical species ecology would greatly benefit if these traits could provide additional information beyond the ‘pioneer’ or

176

‘non-pioneer’ classifications. Approximately 20% of tropical tree species are considered pioneers, with the other non-pioneers being an extremely diverse group. Knowledge of life-history about pioneer versus non-pioneer species needs to be more nuanced than practical knowledge of the species’ life-history. Functional trait data from this thesis depicts the division of trait values between pioneer and non-pioneer species. However, for functional traits to be more applicable to reforestation and conservation applications, they should provide greater detail of a species’ ecology within these groups. This could potentially assist in the mixing of species and traits in order to provide optimum ecosystem services.

6.4 Final conclusion It is difficult to reforest and manage degraded tropical forests in order to provide the diversity of services required for both biodiversity and livelihood benefits. Historically, this has seen monoculture plantations established and, the financial rewards often going to large multinational companies. This reforestation approach operates at the expense of local biodiversity and livelihoods. More recently, alternative reforestation strategies have been implemented and researched for their potential to provide a greater diversity of ecosystem services (Le et al., 2014). Ecologists who study the links between biodiversity and ecosystem function can provide insights into the mechanisms that underpin the success of tropical reforestation from both an environmental and socio-economic perspective (Diaz, 2005). The results of my PhD research supports this, with several practical recommendations, highlighted above, developing a stronger understanding of ecological theory and explicitly applying it to on-ground problems. However, it is unclear if these recommendations will be implemented into reforestation initiatives within tropical regions. Some evidence suggests that reforestation initiatives are often politicised in order to benefit government or foreign aid objectives (Mayda, 1986). The ‘on-ground’ outcomes can be vastly different to these objectives, and improving the application of ecological research into these outcomes may see a greater success in the reforestation of degraded tropical forest.

References

Bowman, D. M. J. S., Brienen, R. J. W., Gloor, E., Phillips, O. L. & Prior, L. D. 2013. Detecting trends in tree growth: not so simple. Trends in Plant Science, 18, 11-17. Buoro, M. & Carlson, S. M. 2014. Life-history syndromes: Integrating dispersal through space and time. Ecology Letters, 17, 756-767.

177

Diaz, S., Tilman, D. And Fargiome, J. 2005. Ecosystem and Human Well-being: Current State and Trends, Biodiversity Regulation of Ecosystem Services Island Press, Washington DC, USA, Ch11, 297-322. Gibert, A., Gray, E. F., Westoby, M., Wright, I. J. & Falster, D. S. 2016. On the link between functional traits and growth rate: meta-analysis shows effects change with plant size, as predicted. Journal of Ecology, n/a-n/a. Hérault, B., Bachelot, B., Poorter, L., Rossi, V., Bongers, F., Chave, J., Paine, C. E. T., Wagner, F. & Baraloto, C. 2011. Functional traits shape ontogenetic growth trajectories of rain forest tree species. Journal of Ecology, 99, 1431-1440. Iida, Y., Kohyama, T. S., Swenson, N. G., Su, S.-H., Chen, C.-T., Chiang, J.-M. & Sun, I. F. 2014. Linking functional traits and demographic rates in a subtropical tree community: the importance of size dependency. Journal of Ecology, 102, 641-650. Kattge, J., Díaz, S., Lavorel, S., Prentice, I. C., Leadley, P., Bönisch, G., Garnier, E., Westoby, M., Reich, P. B., Wright, I. J., Cornelissen, J. H. C., Violle, C., Harrison, S. P., Van Bodegom, P. M., Reichstein, M., Enquist, B. J., Soudzilovskaia, N. A., Ackerly, D. D., Anand, M., Atkin, O., Bahn, M., Baker, T. R., Baldocchi, D., Bekker, R., Blanco, C. C., Blonder, B., Bond, W. J., Bradstock, R., Bunker, D. E., Casanoves, F., Cavender-Bares, J., Chambers, J. Q., Chapin Iii, F. S., Chave, J., Coomes, D., Cornwell, W. K., Craine, J. M., Dobrin, B. H., Duarte, L., Durka, W., Elser, J., Esser, G., Estiarte, M., Fagan, W. F., Fang, J., Fernández- Méndez, F., Fidelis, A., Finegan, B., Flores, O., Ford, H., Frank, D., Freschet, G. T., Fyllas, N. M., Gallagher, R. V., Green, W. A., Gutierrez, A. G., Hickler, T., Higgins, S. I., Hodgson, J. G., Jalili, A., Jansen, S., Joly, C. A., Kerkhoff, A. J., Kirkup, D., Kitajima, K., Kleyer, M., Klotz, S., Knops, J. M. H., Kramer, K., Kühn, I., Kurokawa, H., Laughlin, D., Lee, T. D., Leishman, M., Lens, F., Lenz, T., Lewis, S. L., Lloyd, J., Llusià, J., Louault, F., Ma, S., Mahecha, M. D., Manning, P., Massad, T., Medlyn, B. E., Messier, J., Moles, A. T., Müller, S. C., Nadrowski, K., Naeem, S., Niinemets, Ü., Nöllert, S., Nüske, A., Ogaya, R., Oleksyn, J., Onipchenko, V. G., Onoda, Y., Ordoñez, J., Overbeck, G., Ozinga, W. A., et al. 2011. TRY – a global database of plant traits. Global Change Biology, 17, 2905-2935. Lamb, D., Erskine, P. D. & Parrotta, J. A. 2005. Restoration of Degraded Tropical Forest Landscapes. Science, 310, 1628-1632. Laurance, W. F. 1999. Reflections on the tropical deforestation crisis. Biological Conservation, 91, 109-117.

178

Le, H. D., Smith, C. & Herbohn, J. 2014. What drives the success of reforestation projects in tropical developing countries? The case of the Philippines. Global Environmental Change, 24, 334-348. Lessard, J.-P., Belmaker, J., Myers, J. A., Chase, J. M. & Rahbek, C. 2012. Inferring local ecological processes amid species pool influences. Trends in Ecology & Evolution, 27, 600- 607. Mayda, J. 1986. Forest management and the environment: Worldwide trends in legislation and institutional arrangements. Forest Ecology and Management, 14, 241-257. Mckown, A. D., Guy, R. D., Azam, M. S., Drewes, E. C. & Quamme, L. K. 2013. Seasonality and phenology alter functional leaf traits. Oecologia, 172, 653-665. Nguyen, H., Firn, J., Lamb, D. & Herbohn, J. 2014. Wood density: A tool to find complementary species for the design of mixed species plantations. Forest Ecology and Management, 334, 106-113. Pérez-Harguindeguy, N., Díaz, S., Garnier, E., Lavorel, S., Poorter, H., Jaureguiberry, P., Bret- Harte, M. S., Cornwell, W. K., Craine, J. M., Gurvich, D. E., Urcelay, C., Veneklaas, E. J., Reich, P. B., Poorter, L., Wright, I. J., Ray, P., Enrico, L., Pausas, J. G., De Vos, A. C., Buchmann, N., Funes, G., Quétier, F., Hodgson, J. G., Thompson, K., Morgan, H. D., Ter Steege, H., Van Der Heijden, M. G. A., Sack, L., Blonder, B., Poschlod, P., Vaieretti, M. V., Conti, G., Staver, A. C., Aquino, S. & Cornelissen, J. H. C. 2013. New handbook for standardised measurement of plant functional traits worldwide. Australian Journal of Botany, 61, 167-234.

179