biogeochemical cycles in contrastingéén na grootste tropisch regenwoud ter wereld. in dit...

Promotoren Prof. dr. ir. Pascal Boeckx Department of Green Chemistry and Technology Faculty of Bioscience Engineering Ghent University, Belgium Prof. dr. ir. Hans Verbeeck Department of Environment Faculty of Bioscience Engineering Ghent University, Belgium Decaan Prof. dr. ir. Mark Van Meirvenne Rector Prof. dr. ir. Rik Van de Walle

Biogeochemical cycles in contrasting tropical forests of the Congo Basin

Marijn Bauters

Thesis submitted in fulfillment of the requirements for the degree of Doctor (PhD) in Applied Biological sciences

Dutch translation of the title:

Biogeochemische cyclussen in contrasterende tropische bossen in het Congo-bekken

Citation:

Bauters, M. (2018) Biogeochemical cycles of contrasting tropical forests of the Congo basin.

PhD thesis, Ghent University, Belgium

Front cover art is courtesy of Iris Desmaricaux (www.desiris.be). All rights reserved.

ISBN: 978-94-6357-070-1

The author and promoters give the authorization to consult and copy parts of this work for

personal use only. Every other use is subject to copyright laws. Permission to reproduce any

material contained in this work should be obtained from the author.

i

Summary Tropical forests play a central role in global biogeochemical cycles due to the high exchange of carbon between biosphere and atmosphere. Nutrient cycles control these exchanges via their impact on forest growth and dynamics. However, the process knowledge of how nutrients cycle through atmosphere, biosphere and pedosphere in tropical forests is still poor. This hinders both the mechanistic understanding of tropical forest biogeochemistry and the implementation of nutrient cycling in Earth system models. Additionally, there is an important research bias towards the Amazon basin, which leaves an important knowledge gap for the Congo basin, the second largest contiguous block of tropical forest. In this thesis, we use observational studies at different locations in the Congo basin to respectively assess: the input of nutrients via atmosphere-biosphere interaction, the effect of pedosphere-biosphere interaction on nutrient cycling, and the resulting nutrient budgets in contrasting forest types of the Congo basin. In a first part of the thesis, the pathways of nitrogen (N) input via the atmosphere-biosphere continuum are assessed by combining a study on biological nitrogen fixation along forest succession, with field-based measurements of total nitrogen deposition. For the former, in chapter 2, plant inventories on Fabaceae members were carried out along a forest succession gradient with four successional stages. Subsequently, roots were dug up from the inventoried plants to check nodulation. The results clearly showed that, although Fabaceae members remain present in ageing forests, nodulation is clearly downregulated with forest age in the Congo basin. Secondly, chapter 3 reports on a study were field data of total nitrogen deposition was collected in a lowland tropical forest site at approximately 300 km from the latter study site. We specifically targeted the characterization of organic nitrogen via high-resolution FT-ICR spectrometry. We subsequently combined those data with a crossed dataset of satellite-derived fire pixels and model-simulated backwards wind trajectories. Overall, the study provides convincing evidence of a high nitrogen deposition on central African forests, mainly derived from the seasonal wildfires in the savanna borders of the continent. Overall, against expectations, the first part shows that deposition is much higher than expected, which resolves the prevailing N-paradox for tropical forests in central Africa. In a second part, the interaction of the functional composition of the biosphere and the pedosphere and its effects on tropical forest biogeochemistry is assessed. The vexing problem with this interaction is that on a large scale, the soil and environment filter the species composition in the biosphere, but the biosphere in turn also affects the soil chemistry on a local scale. Firstly, in chapter 4 and 5 a nearly 80-year-old experimental forest plantation - a legacy from the colonial Belgian researchers in the Congo - is used as a setup to assess the effects of functional identity of the biosphere on carbon stocks and soil biogeochemistry. Both chapters jointly quantify the effects of a different functional identity of species on aboveground and below ground carbon stocks, and other soil geochemical parameters. The species with the more conservative nutrient strategy showed higher aboveground carbon stocks, and increased the soil carbon and nutrient content, the latter presumably via soil acidification. Secondly, in chapter 6, the shifts in functional composition of natural forests along a much larger environmental gradient are quantified to assess pedosphere effects on biosphere. Using a standardized approach, two altitudinal gradients were set up in South-American and African tropical forests. Functional and stoichiometric trait shifts in the forest and the δ15N signal of soil and canopy were assessed along both elevational gradients. The study demonstrated that on both continents the forest composition shifts to more nitrogen conservative communities with elevation, with shifts in community traits and canopy and soil δ15N being parallel on both continents. In the final part, both part I and part II are integrated in the overall nitrogen balance of different, contrasting forests of the Congo basin. Full nutrient budgets are key to interpret input and

ii

output loads, to identify potential knowledge gaps and to parameterize the recently implemented Earth system models. Hence, chapter 7 reports on intensive N flux monitoring in two types of lowland tropical forest and two types of montane tropical forests, with contrasting biotic (mycorrhizal association and plant community) and abiotic (location: lowland-highland) environments. N deposition, throughfall, litterfall, leaching and export were monitored and the measurements were complemented with gross soil N dynamics in the different sites. The variability of N input and N losses was high over all forest types, even within location. Lowland and montane forests showed distinct gross soil N dynamics, with little differences between forest types within the location. The results revealed that lowland forests are characterized by a tight soil N cycle, while montane forests have a more open N cycle directly linked to high leaching of predominantly NO3

-. For the lowland non-ectomycorrhizal forest, the losses were dominated by dissolved organic nitrogen (DON), while the ectomycorrhizal forest showed predominantly NO3

- losses. This was at odds with the mineralization-immobilization and nitrification-mineralization ratios being close to 1, and suggests that ectomycorrhizal play a dominant role in the N uptake. Overall, and in strong contrast with the Neotropical forests, the budget of central African forests were imbalanced by a higher input than output. This suggests that unmeasured outputs such as N2 emissions play a major role in the N balance of central African tropical forests.

iii

Samenvatting Tropische bossen spelen een centrale rol in de globale biogeochemische cycli als gevolg van de hoge uitwisseling van koolstof tussen biosfeer en atmosfeer. Nutriëntencycli controleren deze uitwisselingen via hun impact op de groei en dynamiek van bossen. De proceskennis van hoe nutriënten in de tropische bossen door atmosfeer, biosfeer en pedosfeer stromen, is echter nog steeds mager. Dit beperkt zowel het mechanistisch begrijpen van biogeochemie in tropische bossen als de implementatie van nutriëntencycli in Earth-system modellen. Daarnaast is de meeste kennis gebaseerd op onderzoek in het Amazonebekken, en is er bijgevolg heel weinig geweten over die processen in het bos van het Congobekken, het op één na grootste tropisch regenwoud ter wereld. In dit proefschrift gebruiken we observatiestudies op verschillende locaties in het Congobekken om het volgende te onderzoeken: de aanvoer van nutriënten via de interactie tussen atmosfeer en biosfeer, het effect van de interactie tussen de biosfeer en de pedosfeer op de nutriëntencyclus en de resulterende nutriëntenbudgetten in contrasterende bostypes van het Congobekken. In een eerste deel van het proefschrift worden de verschillende kanalen van stikstof (N) toevoer via het atmosfeer-biosfeer continuüm beoordeeld door de combinatie van een studie op biologische stikstoffixatie langsheen een successie gradient van bos, met veldmetingen van totale stikstofdepositie in oud bos. In hoofdstuk 2, werden plantinventarisaties op planten van de Fabaceae-familie uitgevoerd in vier successiestadia. Vervolgens werden de wortels van de geïnventariseerde planten opgegraven om de nodulatie te controleren. De resultaten toonden duidelijk aan dat, hoewel Fabaceae soorten aanwezig blijven in verouderde bossen, de nodulatie van wortels verdwijnt met de leeftijd van het bos. Ten tweede rapporteert hoofdstuk 3 over een onderzoek waarbij veldgegevens van de totale stikstofdepositie werden verzameld in een laagland tropisch bosgebied op ongeveer 300 km van de studie in hoofdstuk 2. Meer specifiek wordt hier de organische stikstof gekarakteriseerd via hoge resolutie FT-ICR-spectrometrie. Vervolgens hebben we deze gegevens gecombineerd met een dataset van brandhaarden die via satelliet gedecteerd worden en modelgesimuleerde windtrajecten. In zijn geheel levert dit onderzoek overtuigend bewijs voor een hoge stikstofdepositie op de bossen van Centraal-Afrika, voornamelijk afkomstig van de seizoensgebonden bosbranden in de savannezones van het continent. Over het algemeen laat het eerste deel, tegen de verwachting in, zien dat de depositie veel hoger is dan verwacht, wat de heersende N-paradox voor tropische bossen in centraal-Afrika oplost. In een tweede deel wordt de interactie van de functionele samenstelling van de biosfeer en de pedosfeer en de effecten ervan op de biogeochemie van tropische bossen beoordeeld. Het moeilijke met deze interactie is dat op grote schaal de bodem en het milieu de soortensamenstelling van de bosvegetatie filteren, maar de biosfeer beïnvloedt op zijn beurt ook de bodemchemie op lokale schaal. Ten eerste wordt in hoofdstuk 4 en 5 een bijna 80 jaar oude experimentele bosaanplanting - een erfenis van de koloniale Belgische onderzoekers in Congo - gebruikt als een opstelling om de effecten van functionele identiteit van de bosvegetatie op koolstofvoorraden en bodem biogeochemie te beoordelen. Beide hoofdstukken kwantificeren gezamenlijk de effecten van een verschillende functionele identiteit van soorten op bovengrondse en ondergrondse koolstofvoorraden en andere geochemische bodemparameters. De soorten met de meer conservatieve strategie voor nutriënten vertoonden hogere bovengrondse koolstofvoorraden en verhoogden het koolstof- en nutriëntengehalte in de bodem, vermoedelijk door bodemverzuring. Ten tweede worden in hoofdstuk 6 de verschuivingen in functionele samenstelling van natuurlijke bossen langs een veel grotere milieu-gradiënt gekwantificeerd om de effecten van de omgeving op de vegetatie te bepalen. Met behulp van een gestandaardiseerde aanpak werden twee hoogtetransecten opgezet in Zuid-Amerikaanse en Afrikaanse tropische bossen. Functionele en stoichiometrische verschuivingen in het boskronendak en het δ15N-signaal van grond en boskroon werden gekwantificeerd langs beide hoogtetransecten. De studie toonde aan dat op

iv

beide continenten de bossamenstelling verschuift naar meer stikstofconservatieve gemeenschappen met toenemende hoogte, met parallelle verschuivingen in chemische samenstelling van het kronendak en het δ15N-signaal van zowel bodem en kronendak op beide continenten. In het laatste deel zijn zowel deel I als deel II geïntegreerd in de algemene stikstofbalans van verschillende, contrasterende bossen van het Congobekken. Volledige nutriëntenbudgetten zijn essentieel om invoer- en uitvoer te interpreteren, mogelijke kennishiaten te identificeren en de onlangs geïmplementeerde cycli in Earth-system modellen te parametriseren. Vandaar dat hoofdstuk 7 rapporteert over intensieve N-fluxmonitoring in twee soorten laagland-tropisch bos en twee soorten bergbossen met contrasterende biotische (mycorrhiza-associatie en plantengemeenschap) en abiotische (locatie: laagland-hoogland) omgevingen. N depositie, doorval, bladval, uitloging en export werden gemonitord en de metingen werden aangevuld met een kwantificatie van de bodem N-cyclus in de verschillende bossen. De variabiliteit van N-invoer en N-verliezen was hoog over alle bostypen, zelfs binnen éénzelfde locatie. Laagland- en bergbossen vertoonden een duidelijke verschillende bodem-N-cyclus, met weinig verschillen tussen bostypen op de locatie. De resultaten onthulden dat laaglandbossen worden gekenmerkt door een gesloten N-bodemcyclus, terwijl bergbossen een meer open N-cyclus hebben die direct is gekoppeld aan hoge uitloging van overwegend NO3

-. Voor het laagland niet-ectomycorrhizale bos werden de verliezen gedomineerd door opgeloste organische stikstof (DON), terwijl het ectomycorrhizale bos overwegend NO3

-verliezen vertoonde. Dit was tegenstrijdig met de mineralisatie-immobilisatie en nitrificatie-mineralisatie verhouding in beide laagland bostypes, die telkens dichtbij 1 was voor beide ratios. Dit suggereert dat ectomycorrhiza een dominante rol spelen bij de opname van N in het ectomycorrhizale bos. Over het geheel genomen, en in sterk contrast met de neotropische bossen, was het budget van de bossen in Centraal-Afrika onevenwichtig door een hogere input dan de output. Dit suggereert dat niet-gemeten outputs zoals N2-emissies een grote rol spelen in de N-balans van Centraal-Afrikaanse tropische bossen.

v

Résumé Les forêts tropicales jouent un rôle central dans les cycles biogéochimiques mondiaux en raison des échanges importants de carbone entre la biosphère et l'atmosphère. Les cycles des nutriments controlent ces échanges via leur impact sur la croissance et la dynamique forestière. Cependant, la compréhension de ce processus et de la façon dont les nutriments circulent dans l'atmosphère, la biosphère et la pédosphère en forêts tropicales est encore faible. Cela entrave à la fois la compréhension mécaniste de la biogéochimie des forêts tropicales et l’implémentation du cycle des nutriments dans la modélisation du système terrestre. De plus, il existe un important biais de recherche vers le bassin Amazonien, ce qui laisse un important déficit de connaissances pour le bassin du Congo, le deuxième plus grand bloc contigu de forêt tropicale. Dans cette thèse, nous utilisons des études observationnelles en différents lieux du bassin du Congo afin d’ évaluer respectivement: l'apport de nutriments par interaction atmosphère-biosphère, l'effet de l'interaction pédosphère-biosphère sur le cycle des nutriments et les bilans nutritifs des forêts du bassin du Congo. Dans une première partie de la thèse, les voies d'apport d'azote (N) via le continuum atmosphère-biosphère sont évaluées en combinant une étude sur la fixation biologique de l'azote le long d’un gradient de succession forestière, a l’aide de mesures de terrain du dépôt total d'azote. Dans le premier cas, dans le chapitre 2, des inventaires de plantes de la famille des Fabaceae ont été effectués le long d'un gradient de succession forestière à quatre stades de succession. Par la suite, les racines des individus inventoriés ont été déterrées afin de vérifier la nodulation. Les résultats ont clairement montré que, bien que les membres de la famille des Fabaceae restent présents dans les forêts vieillissantes, la nodulation est clairement régulée négativement avec l'âge de la forêt dans le bassin du Congo. Deuxièmement, le chapitre 3 fait le raport sur une étude où les dépôts totaux d'azote d’une forêt primaire ont été recueillis dans un site de forêt tropicale de basse altitude à environ 300 km du dernier site d'étude. Nous avons spécifiquement ciblé la caractérisation de l'azote organique par spectrométrie FT-ICR haute résolution. Nous avons ensuite combiné ces données avec une base de données de pixels de feu dérivés de données satellites et une autre base des données présentant les trajectoires du vent obtenues à l’aide d’ un modèle. Dans l'ensemble, l'étude fournit des preuves convaincantes d'un fort dépôt d'azote dans les forêts d'Afrique centrale, provenant principalement des feux de friches saisonniers dans les zones de savane du continent. Globalement, contre toutes attentes, la première partie montre que les dépôts sont beaucoup plus élevés que prévu, ce qui résout le paradoxe-Azote dominant pour les forêts tropicales en Afrique centrale. Dans une deuxième partie, l'interaction de la composition fonctionnelle de la biosphère et de la pédosphère et ses effets sur la biogéochimie des forêts tropicales est évaluée. Le problème épineux de cette interaction est qu'à grande échelle, le sol et l'environnement filtrent la composition des espèces dans la biosphère, mais la biosphère affecte également la chimie du sol à l'échelle locale. Premièrement, dans les chapitres 4 et 5, une plantation forestière expérimentale de près de 80 ans - un legs des chercheurs coloniaux belges au Congo - est utilisée pour évaluer les effets de l'identité fonctionnelle de la biosphère sur les stocks de carbone et la biogéochimie du sol. Les deux chapitres quantifient conjointement les effets d'une identité fonctionnelle différente des espèces sur les stocks de carbone aériens et souterrains et d'autres paramètres géochimiques du sol. Les espèces ayant la stratégie nutritive la plus conservatrice présentaient des stocks de carbone plus élevés au-dessus du sol et augmentaient la teneur en carbone et en éléments nutritifs du sol, cette dernière étant vraisemblablement due à l'acidification du sol. Deuxièmement, au chapitre 6, les changements dans la composition fonctionnelle des forêts naturelles le long d'un gradient environnemental beaucoup plus large sont quantifiés pour évaluer les effets de la pédosphère sur la biosphère. En utilisant une approche standardisée, deux gradients altitudinaux sont mis en place dans les forêts tropicales sud-américaines et africaines. Les changements de traits fonctionnels et

vi

stœchiométriques dans la forêt et le signal δ15N du sol et de la canopée ont été évalués selon les deux gradients d'élévation. L'étude a démontré que sur les deux continents, la composition de la forêt se déplace vers des communautés plus conservatrices d'azote avec l’élévation, avec des changements dans les traits de la communauté et la canopée et le δ15N du sol étant parallèles sur les deux continents. Dans la dernière partie, la partie I et la partie II sont intégrées dans le bilan azoté global de différentes forêts contrastées du bassin du Congo. Les bilans nutritionnels complets sont essentiels pour interpréter les charges d'entrée et de sortie, identifier les lacunes potentielles dans les connaissances et paramétrer les modèles du système terrestre récemment mis en œuvre. Ainsi, le chapitre 7 décrit le suivitintensif des flux de N dans deux types de forêts tropicales de basse altitude et deux types de forêts tropicales montagnardes, avec des environnements biotiques (association mycorhizienne et communauté végétale) et abiotiques (emplacement: basses terres-hautes terres). Les dépôts d'azote, les chutes, les chutes de litière, la lixiviation et l'exportation ont été suivit et les mesures ont été complétées par la dynamique brute de l'azote dans le sol dans les différents sites. La variabilité des pertes et de gains d'azote était élevée dans tous les types de forêts, même à l’intérieur d’un même site. Les forêts des basses terres et des montagnes ont montré une dynamique brute nette de l'azote dans le sol, avec peu de différences entre les types de forêts à l'intérieur du même site. Les résultats ont révélé que les forêts de basse altitude sont caractérisées par un cycle de N serré, tandis que les forêts montagnardes ont un cycle de N plus ouvert, directement lié à une forte lixiviation de NO3

-. Pour la forêt non ectomycorhizienne des basses terres, les pertes étaient dominées par l'azote organique dissous (DON), tandis que la forêt ectomycorhizienne présentait principalement des pertes de NO3

-. Ceci était en contradiction avec les rapports minéralisation-immobilisation et nitrification-minéralisation proches de 1, et suggère que les ectomycorhizes jouent un rôle dominant dans l'absorption de N. Dans l'ensemble, et contrairement aux forêts néotropicales, le budget des forêts d'Afrique centrale a été déséquilibré par un apport supérieur à celui de la production. Ceci suggère que des extrants non mesurés tels que les émissions de N2 jouent un rôle majeur dans l'équilibre des N des forets tropicaux d'Afrique centrale.

vii

Abbreviations

AGC above-ground carbon

AIC akaike information criterion

BApl basal area of planted trees

BELSPO Belgian science policy

BNF biological nitrogen fixation

can canopy

COBIMFO Congo basin integrated monitoring for forest carbon mitigation and biodiversity

CUEe carbon use efficiency

DBH diameter breast-height

DRC Democratic republic of Congo

GPP gross primary production

ICP inductively-coupled plasma

IFA institute facultaire d'agronomie de Yangambi

INERA institution nationale de recherche agronome

IPCC intergovernmental panel on climate change

IRMS isotopic ratio mass spectrometry

KW Kruskal-Wallis

lit litter layer

LLP long-lived pioneers

LSM land surface models

N nitrogen

NEP net ecosystem productivity

NPP net primary production

P phosphorus

PCA principal component analysis PE PST

poly-ethylene partial shade tolerants

R2adj,cond conditional R2

adjusted

R2adj,marg marginal R2

adjusted

SI supplementary information

SLP short-lived pioneers

SOC soil organic carbon

ST shade tolerants

UCB université Catholique de Bukavu

UNIKIS université de Kisangani

USDA United States department of agriculture

WD wood density

WRB world reference base

viii

Table of Contents

Summary ............................................................................................................................................. i

Samenvatting .................................................................................................................................... iii

Résumé ............................................................................................................................................... v

Abbreviations .................................................................................................................................... vii

Table of Contents ........................................................................................................................... viii

Introduction ........................................................................................................................................ 1

Chapter 1: Introduction ................................................................................................................ 2

I Nitrogen input ................................................................................................................................ 11

Chapter 2: Down-regulation of facultative nitrogen fixation by legumes in the central

Congo Basin ................................................................................................................................ 12

Chapter 3: High fire-derived nitrogen deposition on central African forests ...................... 17

II Biogeochemical interaction of vegetation and environment ................................................. 27

Chapter 4: Functional identity explains carbon sequestration in a 77-year-old

experimental tropical plantation ................................................................................................ 28

Chapter 5: Functional composition of tree communities changed topsoil properties in an

old experimental tropical plantation ......................................................................................... 40

Chapter 6: Parallel functional and stoichiometric trait shifts in South-American and

African forest communities with elevation ............................................................................... 53

III Nitrogen budgets of contrasting forests in the Congo Basin ............................................... 65

Chapter 7: The nitrogen cycle of African tropical forests: a shift in paradigms ................. 66

Conclusions and future prospects ................................................................................................ 83

Chapter 8: General conclusions and future prospects .......................................................... 84

Afterword .......................................................................................................................................... 90

Literature cited................................................................................................................................. 91

Appendices .................................................................................................................................... 117

Appendix A................................................................................................................................. 118

Appendix B................................................................................................................................. 120

Appendix C ................................................................................................................................ 123

Appendix D ................................................................................................................................ 127

Appendix E................................................................................................................................. 133

Appendix F ................................................................................................................................. 141

Acknowledgements ...................................................................................................................... 149

Academic Curriculum vitae ......................................................................................................... 152

1

Introduction

Chapter 1: Introduction

2


1.1. The importance of this research (background)

Global change, per definition, encompasses changes in the global environment that may alter the capacity of the Earth to sustain life. This includes climate change, but also land productivity changes, alterations in oceans, atmospheric chemistry and in ecological systems (NRC 2012). In the past, interactions between biosphere and atmosphere have been determining atmospheric carbon contents and thereby also the earth temperature. However, human activities have caused the atmospheric CO2 levels to increase from 277 ppm in 1750 to >400 ppm nowadays, as measured in the Mauna Loa station (Scripps 2014). Anno 2015, about 10% of these ‘anthropogenic’ emissions are caused by land use changes, and the remainder by fossil fuel combustion and cement industry (respectively estimated to be 1.3 and 9.9 Gt C yr-1). Subsequently, the fate of this excess CO2

is partitioned between an increased atmospheric carbon stock (44% or 4.4 Gt C yr-1), an ocean sink (26% or 2.6 Gt C yr-1) and a residual terrestrial sink (31% or 3.1 Gt C yr-1; Le Quéré et al. 2016). This shows that the changes are almost exclusively human-induced, but that the abiotic sinks (ocean and terrestrial) are affected via biotic – natural - interactions with this changed set of environmental parameters. The global carbon cycle compromises carbon exchange between six major carbon reservoirs globally; the atmosphere (720 Gt), oceans (38 400 Gt), the lithosphere (>75 000 000 Gt), the terrestrial biosphere (2 000 Gt), the aquatic biosphere (1-2 Gt) and fossil fuels (4 130 Gt) (Falkowski et al. 2000). Although the pools range across eight orders of magnitude, the biggest pools are not necessarily dominating the exchange processes towards the atmosphere. On the contrary, and as stated above, the dominating processes are fossil fuel and land use change emissions, and ocean and terrestrial ecosystem uptake (Ciais et al. 2013). Living plant biomass sequesters atmospheric carbon by the assimilation of sugars via photosynthesis in the green plant parts (Prentice et al. 2001): the gross primary production (GPP) on an ecosystem-scale. This GPP is split up in net primary productivity (NPP; i.e. build-up of organic material in the ecosystem) and the autotrophic respiration. Hence, NPP is the net amount of carbon that is fixed from the atmosphere into new organic matter, per unit of time (Clark et al. 2001). Of the global NPP budget, forest ecosystems globally cycle about 4 Pg C annually (Table 1.1). Tropical forests in particular are the biome with both the highest NPP, GPP and NPP:GPP ratio (Field et al. 1998, Zhang et al. 2009, Beer et al. 2010), and they store more than one third of the global forest carbon stock (Table 1.1). Hence understanding this biome in particular is priority for a profound biogeochemical understanding of system Earth. Additionally, there are other than only biogeochemical reasons for increased research efforts in the tropics. Biodiversity is high in the tropics compared to higher latitudes (Gaston 2000), and out of 25 identified hotspots to prioritize conservation, 16 are situated in the tropical belt, and 15 compromise tropical forest (Myers et al. 2000). Apart from the intrinsic value of biodiversity, it is also a vital component of the ecosystem functioning, and hence also the capacity of ecosystems to provide human society with goods and services (Loreau et al. 2001, Hooper et al. 2005, Cardinale et al. 2012).


3

Table 1.1. Global stocks (1994) and fluxes (2000-2007) of forests based on Dixon et al. 1994 and Pan et al. 2011. Living biomass comprises estimated aboveground and belowground living biomass.

The fluxes indicate the C stock changes on ‘forest land remaining forest land’ and ‘new forest land’.

1.2. The problem in short

Poor mechanistic understanding

Implementation of forest ecosystems in different model types has traditionally been

very carbon-oriented, where upscaling from leaf-level photosynthesis to ecosystem-

level NEP is in essence what is done. However, that generation of global land surface

models (LSM’s) have failed to explain some of recent empirical observations in tropical

forest carbon dynamics (Brienen et al. 2015). Additionally, recent data analysis and

modeling efforts have shown that N and phosphorus (P) play a crucial role in the

carbon balance of forests (Peñuelas et al. 2013, Fernandez-Martinez et al. 2014,

Wieder et al. 2015). For this reason, models have recently started to incorporate N

and P dynamics (Zaehle and Friend 2010, Wang et al. 2010, Goll et al. 2012, Yang et

al. 2014, Reed et al. 2015), the two most essential nutrients for both plant and

microbial life. Some first simulations show that including only the nitrogen cycle

already hugely affects model results (Bonan 2008a).

The parameterization of models depends on the available empirical data, and the

nitrogen cycle is complex with high rate dynamics between different compounds and

different in- and outputs (Figure 1.2), which makes correct quantification of all

processes difficult and a labor-intensive job. However, the problem extends well

beyond this lack of empirical data to parameterize. There are still substantial

knowledge gaps in the mechanistic understanding of how nutrient and carbon cycles

interact at different levels of the ecosystem.

Geographic bias

The tropical belt geographically subdivided by three regions; Latin-America (the so-

called Neotropics), Africa and South-East Asia (jointly called the Paleotropics). All

three regions have vast amounts of tropical forest, but the region-specific carbon

fluxes are differing. As such, in terms of total carbon stock, the Neotropical forest is

the most important, followed by respectively Africa and South-east Asia. The fluxes of

carbon uptake in intact tropical forest, however, are highest in Africa (Figure 1.1).

Nevertheless, the African tropical forest has received less attention due to a research

bias towards neotropical forest (Verbeeck et al. 2011, Trimblet and van Aarde 2012).

Additionally, recent work has shown that African and South-American tropical forest

Boreal Temperate Tropical Global

Living biomass 88 59 212 359

Stocks (Pg C) Soil 471 100 216 787

Total 559 159 428 1146

Flux (Tg C yr-1)

Living biomass 120 454 2367 2941

Dead wood 132 42 98 273

Soil 74 156 226 456

Total 499 777 2740 4017

Annual change 0.09% 0.49% 0.64% 0.35%


4

show important differences in structure (Banin et al. 2012), biomass (Lewis et al. 2013)

and species composition (Slik et al. 2015a), and hence function potentially different.

Consequently, this raises questions on the universality of biogeochemical and

ecological functioning of forests on both continents, and subsequently their response

to future global change scenarios. This forms an important bottleneck to further

optimize Earth system models explicitly including the African continent (Ciais et al.

2011b).

Figure 1.1 Map of the three tropical regions with their respective estimated carbon stocks (black) and fluxes (arrows), with carbon uptake in regrowth forest (green), carbon uptake in old-growth tropical forest (blue) and carbon emission via deforestation (red). All numbers are expressed in Pg C (for stocks) and Pg C yr-1 for fluxes. Map was composed based on Dixon et al. 1994; Pan et al. 2011 and Hansen et al. 2013. The stocks include vegetation and soil carbon stocks. The coloration on the map shows wet tropical forest (dark green), moist deciduous tropical forest (light green) and montane tropical forest (red).

This knowledge gap is apparent in different fields of biogeochemistry and ecology, and by consequence central Africa always shows up as a blind spot in global syntheses or maps (see e.g. Baldocchi et al. 2000; Jia et al. 2016). Therefore, it is sure African tropical forests play an important role in the global carbon cycle and hence in global change, but the current state-of-the-art is limited. By consequence, more research on the functioning and biogeochemistry of these forests is needed to reduce the existing uncertainties on process-knowledge and to optimize Earth system model parameterization.

1.3. The problem in long and a framework to solve the problem (state-of-the-art)

This thesis aims to tackle both problems, i.e. to contribute to both the mechanistic

understanding of nutrient cycling in tropical forests, and to also provide data from a

very poorly documented region.

Atmosphere, pedosphere and biosphere

From a simplified representation of the N (Figure 1.2) and P cycle (Figure 1.3) it is apparent that ecosystem biogeochemistry can be summarized as molecular exchanges between athmosphere, pedosphere and biosphere. However, it’s important to note that the biogeochemical controls over both the N and P cycle is different (McGill and Cole 1981, Vitousek et al. 2010). N is accumulated from the atmosphere via either atmospheric deposition or biological nitrogen fixation. This leads to the accumulation of N, or increase in N availability, with age of the soil and the ecosystem. On the other hand, P is primarily derived from rock weathering, and in many cases a large part of the P pool is adsorbed to mineral surfaces, reducing the


5

mobility of P along with the availability for plants. Hence, compared to N, the P cycle in general shows low inputs and losses in natural ecosystems, resulting in a closed nutrient cycle (Vitousek et al. 2010). Due to the lack of high inputs other than rock weathering, the P pool in the soil cannot readily be replenished. This introduces the concept or P retro-gradation, i.e. the depletion of P with soil aging. Hence very old and weathered soils, such as the oxisols commonly found in tropics, have undergone net P losses, leaving tropical lowland forests with very low P-availability (Walker and Syers 1976). Finally, for biogeochemistry of both nutrients, both the atmosphere and the soil matter.

Figure 1.2 The nitrogen (N) cycle in forests, with inputs in red, output in blue, and ‘internal’ N cycle

in dark green. A simplified internal soil N cycle is depicted in the right panel (e).

Figure 1.3 The phosphorus (P) cycle in forests, with inputs indicated in red, outputs in blue, and

internal P cycling in green, and the soil P cycle in detail (e).


6

The nutrient availability that results from atmosphere and soil factors will subsequently drive interactions of the organisms in the ecosystem. Plants and microbial life are at the forefront here, because they are the main producers and consumers of the ecosystem (Geider et al. 2001, Falkowski et al. 2008, Townsend et al. 2008). Consequently, an important challenge is to determine how the variety of plants and microbial life affect ecosystems biochemistry. As an example, members of the Fabaceae can hugely affect ecosystem N accumulation by fixing atmospheric N2 via symbiosis with Rhizobia (Batterman et al. 2013). However, tropical forests host a minimum of 40.000 tree species, comprising a wide variety in form and function (Slik et al. 2015b). This undermines the applicability of a taxonomic framework, since this would imply profound knowledge on biogeochemical impacts of at the very least the most abundant species. This is not only overly extensive, but it also limits the generality and transferability of research results across regions. If we link back to ecology, the life history or strategy within an ecosystem can be inferred from the so called functional identity of species (Díaz et al. 2015). This functional identity results from a projection species from the classical taxonomic framework into a functional framework, which can be done via assessments of readily measurable plant traits. Does the species fix nitrogen? What is the leaf chemistry of the tree species (Reich et al. 1997, Wright et al. 2004, Poorter et al. 2009)? How is the wood composed (Chave et al. 2009)? This new functional approach creates a tool that abstracts the functional information from a species, i.e. what we want to know, and enables us to transfer, generalize and even calculate with species data. This tree functioning is in part determined by the soil environment while at the same time affecting it (Binkley and Giardina 1998, Laliberte et al. 2014, Waring et al. 2015). Furthermore, not only trees matter. The soil N (Figure 1.2) and P (Figure 1.3) cycle are dominated by the microbial community (Falkowski et al. 2008). As such N mineralization has long been perceived as control step for N availability for plants. This view was based on two widely accepted paradigms 1) that plants only use inorganic N that is not needed to sustain microbial growth and 2) that decomposition is a microbial process which produces excess NH4

+. This paradigm has shifted last decade, and the believe now is that depolymerization of organic matter (OM) is the controlling step in nutrient availability. As such the breakdown of OM into organic monomers by microbes/mycorrhizae regulates nutrient availability, which is then prone to competition between plants and the microbial community (Schimel and Bennett 2004). In essence, ecosystem biogeochemistry is determined by the integration of exchange of molecules between atmosphere, pedosphere and biosphere. The high degree of complexity of these exchange reactions calls for a integrated approach. Hence, instead of focussing on atmosphere, pedosphere and biosphere separately, the interactions need to be taken into account via an integrated assessments. As such, we split up the state-of-the-art of the atmosphere-biosphere continuum, the pedosphere-biosphere continuum and conclude with the integration of both in the ecosystem biogeochemistry (Figure 1.4.).


7

Figure 1.4 Schematic overview of the thesis with the different parts and chapters. The black

arrows represent the biogeochemical exchanges.

Part I: Atmospheric nitrogen input

Lowland humid tropical forests are in general assumed to be N-rich environments, an assumption which is supported by the observation of very high rates of N build-up, recycling and export (Vitousek 1984, Vitousek and Sanford 1986, Hedin et al. 2003, Davidson et al. 2007). The question arises which N input can sustain these high N export rates? Biological nitrogen fixation (BNF), i.e. the conversion of atmospheric N2 to reactive N in the soil via bacteria, has been inferred as explanation. However, there is clear evidence that BNF is downregulated in N-rich environments (Fujikake et al. 2002, Pearson and Vitousek 2008, Menge et al. 2009). This gives rise to the so-called N paradox of tropical forests, i.e. the seemingly N richness of tropical lowland forests which in turn implies the downregulation of the mechanism which causes/sustains the N richness (Hedin et al. 2009). The most viable hypothesis that was postulated to resolve this paradox was that BNF still operates in the a so-called leaky nitrostat forest (Hedin et al. 2009). This model infers ecosystem heterogeneity as a matrix where BNF can still take place in N-poor niches in the forest as 1) an upregulated symbiotic fixation and 2) heterotrophic asymbiotic fixation by free-living bacteria. Such N-poor niches might be hotspots for fixation in the topsoil, the litter layer and on canopy leaf surfaces (Menge and Hedin 2009, Reed et al. 2011, Menge, Duncan and Levin 2017). Another hypothesis is that atmospheric N deposition can also contribute to total ecosystem N input. N deposition is in general assumed to be low in remote locations with low industrial activity and in absence of intensive agriculture, with numbers < 5 kg ha-1 yr-1 (Galloway et al. 2004, Dentener et al. 2006). However, empirical measurements of deposition in tropical forests are few and of varying quality. Additionally, there is an increased awareness of the importance of organic N in the N cycle (Cape et al. 2011, Kanakidou et al. 2016). Traditionally, both field studies and simulations only take into account inorganic


8

nitrogen forms, while it has been shown that organic N can even dominate the deposition fluxes (Cornell et al. 2003). This problem of lack of understanding of N input is indicative for our knowledge on the N cycle, and needs to be assessed in depth.

Part II: Biogeochemical interaction of vegetation and environment

The pedosphere-biosphere continuum of molecular exchanges in ecosystems poses

yet another challenge. Research has since long evidenced the functional link between

soil properties and plant communities in forests (Zinke et al. 1962, Binkley and

Giardina 1998, Finzi et al. 1998c, Reich et al. 2005). A large amount of studies

observed the covariation of soil properties, plant traits and communities on a larger

scale in natural forests (Zinke 1962, Finzi et al. 1998c, John et al. 2007, Toledo et al.

2012, Quesada et al. 2012, Condit et al. 2013, Laughlin et al. 2015). Although well-

documented, these correlations between soil and plant communities lack the ability to

infer causality in the plant-soil interaction since the effects go both ways: (1) the soil

and local climate can act as an environmental filter for species communities, and (2)

the tree species can change their environment (Laughlin et al. 2015). Although the

environmental filtering concept is accepted to be more important on larger scale

(Fortunel et al. 2014), research also starts to focus on the ‘afterlife’ effects of this biotic

spectrum on its abiotic environment (Freschet and others 2012). The current

understanding of mechanisms through which plant traits can alter their environment

and the soil in particular, is primarily based on studies of monocultures in a ‘common

garden‘ setting (Reich et al. 2005, Vesterdal et al. 2008). This approach has the

advantage that a variety of species are planted in adjacent blocks, so that climate,

geology, previous land use, topography and hydrology are relatively homogeneous,

and that the observed differences in soil properties in function of time can be inferred

to a tree species effect. However, important time and funding related challenges

prevail in these experimental approaches, due to the long-term commitments that are

needed to do experiments in woody ecosystems. For temperate forests, it has been

shown that both direct (leaf chemistry) and indirect (e.g. via earthworm abundance)

effects can be considerable (Reich et al. 2005, Mueller et al. 2012, Vesterdal et al.

2013). In the tropics, only very few of those experiments have been done, and all

efforts imply short terms of forest development (Fisher 1995, Stanley and Montagnini

1999, Russell et al. 2007). Nevertheless, since a large part of the lowland tropical

forest grows on oxisols or ultisols, a long-term study on these soils are highly relevant

for global biogeochemical cycles. In conclusion, to what extent plants affect their own

environment or the environment affect plant communities is still vexing question in

tropical ecosystem biogeochemistry.

Part III: Nitrogen budgets of contrasting forests in the Congo Basin Nutrient budgets integrate biogeochemical exchanges along the atmosphere-biosphere and

the pedosphere-biosphere continuum. However, full N balances are rare in tropical forests

(Vitousek 1984, Bruijnzeel 1991, Hedin et al. 2003, Taylor et al. 2015). The current literature

rather looks at inputs and losses separately and infers the N availability or N cycle metrics

from those observations. As such, there have been reports on high hydrological losses of N

in tropical forests (Brookshire et al. 2012a, 2012b, Gücker et al. 2016), as well as on high

gaseous losses (Houlton et al. 2006, Bai and Houlton 2009), but one could comment that ‘high’


9

is a relative measure which can actually only be used when the full balance of the ecosystem

is known. Likewise, it is not well known what constrains the composition of hydrological losses

from tropical forest catchments (Perakis and Hedin 2002, Brookshire et al. 2012b, Taylor et

al. 2015) in part because studies either focus on one forest type, or have a landscape-scale

approach where the included types are not well characterized. In addition to this input-output

problem, our understanding of N cycling in tropical forest is also challenged by our current

understanding of the soil N cycle (van Groenigen et al. 2015). Studies have traditionally

approached these soil N dynamics by using proxies such as net soil transformation rates

(Vitousek and Matson 1988), but in fact theses metrics offer limited insight and do not

necessarily correlate with the gross soil N rates (Davidson et al. 1992). Furthermore, the last

decade has clearly shown that these gross rates need to be assessed in-situ in intact soil

cores to avoid disturbance effects (Booth et al. 2006, Arnold 2008, Gütlein et al. 2016).

Gerschlauer et al. (2016) combined those constraints while looking at published rates, and

came to the conclusion that only a very limited amount of rates have been published, with a

high range in rates.

1.4 Outline of the thesis and specific research questions The thesis is split up in the three parts as outlined in the state-of-art.

Part I: Atmospheric nitrogen input

The aim of the first part is to identify the contributions of both BNF and atmospheric deposition

to total ecosystem N input in central African forests. We first assess the role of BNF in different

stages of forest succession by assessing the nodulation and abundance of Fabaceae

members (Chapter 2). Subsequently, we assessed nitrogen deposition in two different forest

types and open field in central DRC over the course of one hydrological year (Chapter 3).

Specific research questions:

- What is the overall role of BNF/deposition in total ecosystem N input?

- Can we transfer the observed BNF downregulation with forest age from the Amazon?

- How do the N deposition loads in the field compare to simulations when including organic

N? Could this resolve the N paradox for central African forests?

Part II: Biogeochemical interaction of vegetation and environment

The second part tries to integrate an assessment of the vegetation-soil interaction. First, we

look into tree species composition effects on local biogeochemical cycles. For this, a unique

nearly 80-year-old tree plantation in the hearth of the DRC is used. We link the aboveground

carbon-accrual and the topsoil development with the functional identity of the planted species

(Chapter 4 and 5). Secondly, wider environmental gradients along elevation are used (both in

Rwanda and in Ecuador) to determine the direction and magnitude of the forest compositional

response on climatic (indirect) and soil (direct) changes (Chapter 6).


- What is the impact of functional identity of species on carbon stocks after 80 years of

tree growth?

- What are the impacts of functional identity of species on soil chemistry after 80 years of

tree growth?

- What can we infer from the shifts in functional composition in natural forest over larger

environmental gradients?


10

Part III: Nitrogen budgets of contrasting forests in the Congo basin

In the final part, we build further on the results from part I and part II, respectively the

atmosphere-biosphere input continuum and the pedosphere-biosphere continuum. As such,

we integrate the basis that has been laid out in both parts and use it to interpret the ecosystem

biogeochemistry in different forest types of the Congo Basin. Via an intensive monitoring

network, we aim to make budgets and quantify the fluxes in and between compartments

(Chapter 7).


- How are nutrient cycles affected by atmosphere, pedosphere and biosphere?

- What are the consequences of those effects for budgets?

11

I Nitrogen input

Chapter 2: Down-regulation of facultative nitrogen fixation by legumes in the central Congo Basin

12

Chapter 2: Down-regulation of facultative nitrogen fixation by

legumes in the central Congo Basin After: Bauters, M., Mapenzi, N., Kearsley, E., Vanlauwe, B. and Boeckx, P.: Facultative

nitrogen fixation by legumes in the central Congo basin is downregulated during late successional stages, Biotropica, 48(3), 281–284, 2016.

Abstract The contribution of the legume community to the nitrogen cycle in natural forest regeneration remains poorly understood. We systematically assessed the changes in abundances and nodulation status of all legumes, across taxa and plant types, in a forest succession gradient in the Equateur province in DR Congo. Our results clearly show that symbiotic N2 fixation is down-regulated in later-successional stages.

Chapter 2 Cover. Djolu is located in the very heart of the DRC, four days of motorbike from the closest airport, on small dirt tracks straight through the forest and some small settlements. The best driver in the DRC is Héritier Ololo Fundji (left). The coloration on the map shows wet tropical forest (darkgreen), moist deciduous tropical forest (light green) and montane tropical forest (red).


13

2.1. Introduction The role of symbiotic nitrogen fixation in tropical forests has been debated recently (Barron et al. 2011, Batterman et al. 2013). Tropical forests have traditionally been considered nitrogen (N) - rich, and rather limited by the availability of phosphorus (e.g. Vitousek & Sanford 1986). Coupled to this, there have been observations of high losses of bio-available N in these ecosystems, via leaching and denitrification (Brookshire et al. 2012a). The combination of both suggests that a rather high N input should exist, sustaining this leaky N cycle. This has pushed forward the key role of biological nitrogen fixation (BNF) as a vital process in these ecosystems. Although the rhizobia symbiosis traditionally received more attention (e.g. Houlton et al. 2008), the role of canopy legumes has recently been questioned due to an apparent paradox in this theory (Barron et al. 2011). Experimental results have demonstrated that symbiotic N2 fixation is down-regulated in an N-rich environment (Fujikake et al. 2002), thus indicating a facultative rather than an obligate strategy. Additionally, the process causes an energetic cost for the plant, and an obligate symbiosis would thus negatively affect the competitiveness of N2 fixers in an environment with high amounts of bio-available N. This contradicts the abundance of canopy legumes in some natural forests (Losos and Leigh 2004). Although, it remains difficult to test directly whether this down-regulation is also occurring in more complex, natural systems because of the labor intensiveness of checking root nodulation status at a large scale, convincing empirical proof and model results have recently shown that this is indeed valid for some older tropical forests in the Neotropics (Barron et al. 2011, Batterman et al. 2013). However, to our knowledge, there are no systematic studies addressing the nodulation status on a forest chronosequence in African rainforests, despite important differences in the rainforest structure and composition between South American and African rainforests (Lewis et al. 2013, Slik et al. 2015a). In particular for legumes, a family closely linked to the N cycle of forests, large datasets on forest inventories have made clear that family and subfamily abundances are different across continents (Losos and Leigh 2004, Sprent 2009). While South America has high abundances of nodulating species from all three subfamilies (Mimosoideae, Papilionoideae and Caesalpinioideae), the Caesalpinoideae species in Africa are more abundant, but less likely to be nodulated (Sprent 2009). Additionally, there is a unique monodominance of a Caesalpinioideae species over large areas in the African rainforest (i.e. Gilbertiodendron dewevrei; Gérard 1960). This led Sprent (2005) to point out that much more work needs to be done in African rainforest, and that the clear cross-continental difference in taxonomic composition of the legume community raises question marks about their role in the N dynamics of African rainforests. In addition to this geographic bias in knowledge, in general little is known about the community dynamics through which legumes contribute to the N cycle of tropical forest regeneration. 2.2. Material and Methods For this study, we surveyed legumes in a forest chronosequence near Djolu, in the Equateur province of the DR Congo (00°40'26" N, 22°27'33" E). This region has a tropical rainforest climate, Af-type according to the Köppen climate classification, with an annual rainfall of 1750 mm and a temperature of 24.5°C throughout the year (Peel et al. 2007b). On site, we delineated four different stages of forest development at six different locations: 0-5 yr. old fallow, 5-10 yr. old fallow, secondary forest and primary


14

forest. The classification of the stages of forest succession was done in the field by expert judgement. In each of these stages, six separate, straight transects of 500 m were delineated, separated by a minimum distance of 5 km, and maximum 60 km. On these transects, we marked three equally separated points (250 meter), which formed the center of three circular inventory plots with radius of 10 meter (i.e. total inventory area per transect: 943 m2), where all individual plants belonging to the Fabaceae were inventoried, and subsequently categorized by subfamily, plant type and nodulating capacity. Additionally, we surveyed their nodulation status by following roots of legumes and checking for nodules down to 0.25 m for the seedling trees and herbaceous plants, and to 0.5 m for the trees with a diameter breast-height (DBH) >0.1 m. We used non-parametric Kruskal-Wallis rank sum test (KW) and subsequent Mann-Whitney two sample tests to test the differences of subfamily and plant type on abundance and nodulation status. All statistical analyses were conducted with R version 3.1.1 (R Development Core Team 2015). 2.3. Results and discussion In total, we found 43 woody and herbaceous species of Fabaceae. The number of species occurring for each stage was, respectively, 21 species in the 0-5 yr. fallow, 19 in the 5-10 yr. fallow, 14 in the secondary forest and 16 in the primary forest (Appendix A1). The subfamily abundances differed significantly for all three subfamilies across successional stages, but trends differed among subfamilies. The relative abundance (RA) of Caesalpinioideae increased from 17% to 75% over the forest succession (P =0.007), while Papilionoideae RA decreased from 73% to 22% (P <0.001). Mimosoideae (P =0.003) stayed more or less constant in the first stages, but was absent in the primary forest (Figure 2.1a, Appendix A2). The decrease in Papilionoideae abundance was mainly due to the very high abundance of Teramnus labialis (L.f.) Spreng. and Desmodium adscendens (Sw.) DC., which each constituted 17 percent of the total number of individuals in the first succession stage. The latter are both herbaceous species that are common in the region. The differences in nodulated fraction of each subfamily across successional stages were significant for Mimosoideae (P < 0.001) and Papilionoideae (P < 0.001), but not significant for Caesalpinioideae (P =0.09) (Figure 2.1b). Caesalpinioideae had a very low fraction of nodulated individuals, with no clear decline in the first three phases of the succession. Individuals belonging to either Papilionoideae or Mimosoideae, showed high nodulation in the early-successional stages, but almost no nodules were present in the secondary forest. None of the three subfamilies showed nodules in the primary forest and no nodules were found on the trees with DBH > 10 cm in the field. The abundance of all four plant types also differed significantly over the succession stages (P <0.001). The legume community was dominated by herbs in the earliest successional stage, while absent in later stages. Shrub and liana abundances diminished significantly over succession, while trees (mainly consisting of seedlings) became significantly more abundant (Figure 2.1c). For each plant type, the fraction nodulated decreased with successional stage. (P <0.001) (Figure 2.1d). Furthermore, species with the potential to nodulate decreased as succession progresses, constituting 80% of all legumes in 0-5 yr. fallow with 71% effectively nodulated, to 40% in the primary forest, where no nodulation was found. For non-nodulating species, comparable abundances were found in earliest and latest successional stage, with a minimum in secondary forest. This gave rise to a ratio of nodulating to non-nodulating


15

species of respectively 4.1, 2.1, 5.3 through the first three stages of succession, and only 0.7 in the primary forest (Table 2.1). Table 2.1. Abundance (n ha-1) and percentage nodulation through succession, and the ratio between nodulating specimens versus non-nodulating specimens. The occurring Fabaceae-species (see Appendix A1) were subdivided into two groups, nodulating and non-nodulating, based on reports from the literature.

0-5 Fallow 5-10 Fallow Secondary Forest Primary Forest

Nodulating species 24469 ± 1614

(71%) 7084 ± 1428

(13%) 7968 ± 1299

(1%) 3585 ± 1133

(0%)

Non-nodulating species 5979 ± 132

(0%) 3358 ± 481

(0%) 1505 ± 333

(0%) 5238 ± 702

(0%)

Ratio 4.1 2.1 5.3 0.7

To our knowledge, this study is the first to assess legume nodulation systematically along a chronosequence in central Africa. Our results are in line with other studies in both African and South American rainforest. Bonnier (1958) assessed the nodulation of forest legumes in the Yangambi Man and Biosphere reserve, only 200 km from our study site, and reported that only very few mature trees were nodulated, but that nodulation was far more common in disturbed soils. The same has been observed in the lowland rainforest of Panama, with higher symbiotic fixation rates in gaps and disturbed forests (Barron et al. 2011). Sprent (2005) pointed out that mature trees are expected to recycle much of their N, and thus have little need for nodulation. Accordingly, no nodules were found for big trees in this experiment, but neither for present herbs or seedlings in the later successional stages. The decrease in the fraction of potentially nodulating species that were nodulated suggests that the N cycle of mature forests in the central Congo basin no longer depend on N inputs from rhizobia symbiosis. Similar observations of decrease in nodulation along succession gradients were carried out by Pearson and Vitousek (2008), when they compared even-aged stands of 6 yr. and 20 yr. of one N2 fixing species, and more recently by Batterman et al. (2013) in Panama. As in the latter studies, we conclude that symbiotic N2 fixation plays an important role for forest N input only in the earlier regeneration phases, when biomass accumulation is high. The sustained presence of potentially N2 fixing trees in later successional stages, although less abundant, supports the hypothesis of a facultative fixation strategy at the individual level. However, the decline in abundance of nodulating species, along with the higher abundance of non-nodulating species in primary versus secondary forest, shows that there is also a species-turnover effect, which is to be expected during succession (Van Breugel et al. 2007). Although we have no further details about functional composition at each stage, our data indicates that pioneers are more likely to be potentially nodulating legumes, while their share diminishes in the climax vegetation (Table 2.1). These results suggest that, despite ecological and biogeographical differences between the African and South American rainforest (Corlett and Primack 2006), the facultative down regulation of symbiotic N2 fixation is very likely universal in this biome.


16

Figure 2.1. Abundances (a, c) and fraction of nodulated individuals (b, d) for the three Fabaceae subfamilies (a, b) and for the different plant types (c, d) in the forest chronosequence in the Equateur province DR Congo. The bars indicate the mean value, with standard deviations, for all six replicated transects per succession stage; significant differences of abundance per subfamily and plant type along the chronosequence are indicated by different letters per succession stage (P < 0.05).

Acknowledgements

Data were collected by N.M., and analyzed by M.B. All authors contributed to the writing of this paper. M.B. and E.K. are currently funded by the Belgian Science Policy Office (BELSPO). Supporting Information Additional Supporting Information may be found in Appendix A. Appendix A1 List of species occurrence and nodulation in the different forest succession stages in the study area. Appendix A2 Legume composition of the different successional stages.

Chapter 3: High fire-derived nitrogen deposition on central African forests

17

Chapter 3: High fire-derived nitrogen deposition on central

African forests After: Bauters, M., Drake, T. W., Verbeeck, H., Bodé, S., Hervé-Fernandez, P., Zito, P.,

Podgorski, D. C., Boyemba, F., Makelele, I., Ntaboba, L. C., Spencer, R. G. M. and Boeckx, P.: High fire-derived nitrogen deposition on central African forests, Proceedings of the National Academy of Sciences of the United States, early view, 2018

Abstract Atmospheric nitrogen (N) deposition is an important determinant of N availability for natural ecosystems worldwide. Increased anthropogenic N deposition shifts the stoichiometric equilibrium of ecosystems, with direct and indirect impacts on ecosystem functioning and biogeochemical cycles. Current simulation data suggest that remote tropical forests still receive low atmospheric N deposition due to a lack of proximate industry, low rates of fossil fuel combustion, and absence of intensive agriculture. We present field-based N deposition data for forests of the central Congo Basin, and use ultrahigh-resolution mass spectrometry to characterize the organic N fraction. Additionally, we use satellite data and modeling for atmospheric N source apportionment. Our results indicate that these forests receive 18.2 kg N ha-1 yr-1 as wet deposition with dry deposition via canopy interception adding considerably to this flux. We also show that roughly half of the N deposition is organic, which is often ignored in N deposition measurements and simulations. The source of atmospheric N is predominantly derived from intensive seasonal burning of biomass on the continent. This high N deposition has important implications for the ecology of the Congo Basin and for global biogeochemical cycles more broadly.

Chapter 3 Cover. Yoko is roughly 25 km south of Kisangani and quite easy to reach. The forests in the Congo basin are subjected to high nitrogen deposition, caused by annual biomass burning on the African continent. Drone picture by Adam Amir: a view on the Congo Basin from the Republic of Congo (Brazzaville) with a burning wildfire. The coloration on the map shows wet tropical forest (darkgreen), moist deciduous tropical forest (light green) and montane tropical forest (red).


18

3.1. Introduction Human activities, such as the production of ammonium-based fertilizers and the burning of fossil fuels, have led to dramatic changes in the global N cycle over the past century (Galloway et al. 2008). Increased levels of reactive N in the atmosphere have clear ramifications for ecosystems worldwide. Nitrogen is tightly coupled to carbon (C) in global biogeochemical cycles due to the metabolic need of plants and (micro)-organisms. Furthermore, recent data analyses and modeling activities have shown that carbon dioxide (CO2) uptake by terrestrial ecosystems strongly depends on nutrient availability (Peñuelas et al. 2012, Fernandez-Martinez et al. 2014, Wieder et al. 2015). This finding is important for tropical forests, which dominate the terrestrial C cycle and account for approximately one-third of global terrestrial gross primary productivity (Beer et al. 2010). This nutrient control on forests highlights the importance of quantifying atmospheric N deposition in order to understand and predict the global C cycle. This has prompted the Earth system modeling community to improve the implementation of nutrient cycles in their models, with varying consequences for the predicted effects of N deposition on C cycling in tropical forests (Sokolov et al. 2008, Thornton et al. 2009, Zaehle and Friend 2010, Thomas et al. 2015). Indeed, model parameterization largely depends on the quantity and quality of empirical data used to constrain the simulations. Recently, global empirical N deposition data mapped from different monitoring networks showed a distinct lack of field-based estimates for the tropics (Jia et al. 2016). More importantly, the studies that do exist report typically solely the inorganic N inputs, overlooking the organic component in deposition. Recent reviews call attention to the importance of including this atmospheric organic N component, while pointing to a complete lack of data from Africa (Cape et al. 2011, Cornell 2011). They conclude that the effect of N deposition can only be understood if the organic component is consistently included and considered. In short, there is no empirical organic N deposition data for African forests and, consequently, no insight into its potential sources. Some studies have suggested that a large proportion of organic N is sourced from the atmosphere and derived from biomass burning (Mace et al. 2003, Ito et al. 2015). Satellite data coupled to biogeochemical models have identified the African continent as a key region for biomass burning, comprising 65% of the total global burnt area (i.e. 222 Mha yr-1) (van der Werf et al. 2006). Most of this biomass burning occurs as large seasonal wildfires in the savannas, with an estimated 78% of the tropical reactive N from savanna fire emissions sourced within the African savanna (4.3 out of 5.5 Tg N yr-1) (Chen et al. 2010). Additionally, it was found that the net transfer of N from savannas to forests around the equator was larger in Africa than in South-America and Asia due to the prevailing equatorward winds that carry emissions into the inter-tropical convergence zone. These models hint at the potential magnitude and composition of total N deposition on these remote forests but have not been corroborated by field-based observational data to-date. In this study, we directly measured N deposition on the ground in a remote forest in the central Congo Basin for the first time. Specifically, we collected bi-weekly water samples in the open field and under the forest canopy of two forest types. To unravel the origin of the N, we used both ultra-high resolution mass spectrometry and modeling techniques.


19

3.2. Material and methods Study site The study was carried out in an old-growth tropical forest of Yoko, roughly 30 km south of Kisangani (N00°17’; E25°18’), in the Democratic Republic of the Congo (DRC). Vegetation at the study site is classified as semi-deciduous rainforest, and the climate falls within the Af-type (tropical rainforest climate), following the Köppen-Geiger classification. Soils in the region are typical deeply weathered and nutrient-poor Ferralsols, with very limited elevation differences and gentle slopes. The site has two dominant forest types: mixed lowland tropical forest and Gilbertiodendron-dominated forest, where >60% of the basal area consists of one species Gilbertiodendron dewevrei. Throughfall was sampled at three study plots of 40x40 m per forest type. Additionally, we collected bulk precipitation in one location in the open field at a nearby deforested site. Data collection Samples were collected from 19/09/2015 – 10/09/2016. Throughfall and bulk precipitation was collected using polyethylene funnels with a diameter of 15 cm supported by a wooden pole of 1.5 m height, on which a polyethylene (PE) tube was attached to drain into a 5-L PE container. A nylon mesh was placed in the neck of the funnel to avoid contamination by large particles. The container was buried in the soil and covered with black plastic to prevent the growth of algae and to keep the samples cool to minimize microbial activity. In each plot, we installed eight throughfall collectors in two rows of four, with approximately 8 m distance between all collectors. All plots were sampled every fortnight. On each sampling occasion, the water volume in each collector was measured in the field, and used recipients, funnels and mesh were replaced by new ones, rinsed with distilled water. A volume-weighted composite sample was made per plot. The volume-weighted biweekly composite samples were filtered using a nylon membrane filter of 0.45 µm before freezing, and then immediately stored in a freezer. Finally, the samples were sent in several batches to Ghent University (Belgium) for chemical analysis. Chemical analysis and empirical data analysis After transport to Belgium, NH4

+ was determined colorimetrically by the salycilate–nitroprusside method (Mulvaney 1996) on an auto-analyzer (AA3, Bran and Luebbe, Germany). NO3

- was determined colorimetrically using the same auto-analyzer as NO2 after reduction of NO3

- in a Cd–Cu column, followed by the reaction of the NO2 with N-1-napthylethylenediamine to produce a chromophore. Dissolved organic nitrogen (DON) can only be determined indirectly via measurement of total dissolved nitrogen (TDN) and subtraction of dissolved inorganic nitrogen (DIN), leading to additional analytical uncertainty. As discussed elaborately by Cornell et al. (2003) there are three methods that are often used to determine TDN in water samples: Kjeldahl digestion, persulfate oxidation, and high-temperature catalytic oxidation. At least for marine samples, no systematic biases have been found in method inter-comparisons (Sharp et al. 2002, Cornell et al. 2003). However, there is a risk of underestimating DON, because the main source of error of TDN determination is the incomplete conversion of DON to DIN (Cornell et al. 2003). We used persulfate oxidation for the determination of TDN in the water sample. For this, a 1:1 oxidizing solution of NaOH, H3BO3 and K2S2O8 is added to the sample, which is subsequently placed in an autoclave at 121°C for 1 hour in order to convert NH4

+ and dissolved DON into NO3- (Lachouani et al.

2010). In addition, a subset of samples was taken for bulk precipitation (n=12), for one


20

of the mixed lowland plots (n=23) and for one of the monodominant plots (n=19) for further analysis by electrospray ionization coupled with Fourier transform ion cyclotron resonance mass spectrometry (ESI-FT-ICR MS). Prior to extraction for FT-ICR MS, dissolved organic carbon (DOC) was analysed. Therefore samples were thawed under refrigeration (4ºC), acidified with 12N HCl to a pH of approximately 2, and measured via high-temperature catalytic oxidation on a TOC-analyzer (Shimadzu, Japan). DOC concentrations were determined as the mean of at least three replicate injections, for which the coefficient of variation was < 2% (Mann et al. 2012). Samples were prepared for ESI-FT-ICR MS analysis by the solid-phase extraction of dissolved organic mater (SPE-DOM) method (Dittmar et al. 2008). Briefly, filtered samples were acidified to pH 2.3 with HPLC-grade HCl before solid phase extraction on 100 mg Bond Elut PPL (Agilent Technologies) cartridges. Sample volume was adjusted depending on original DOC concentration to obtain 40 µg C per mL of methanol eluate. Negative ions produced by ESI from direct infusion of the methanol extracts at a concentration of 50 µg C mL-1 were analyzed with a 21 T FT-ICR mass spectrometer (FSU, Tallahassee, Florida, USA) (Hendrickson et al. 2015). Each signal >6σ RMS baseline noise was assigned a molecular formula with EnviroOrg ©,™ (Corilo 2015) and classified as condensed aromatic (modified aromaticity index (AImod) ≥ 0.67), polyphenol (0.67 > AImod > 0.5), unsaturated, low oxygen (AImod < 0.5, H/C < 1.5, O/C < 0.5), unsaturated, high oxygen (AImod < 0.5, H/C < 1.5, O/C ≥ 0.5), aliphatic (H/C ≥ 1.5, N = 0), peptide (H/C ≥ 1.5, N > 0) (Koch and Dittmar 2006, Spencer et al. 2014), as defined in the python script used in the analysis (Hemingway 2017). We used plot-averaged values for the bulk rainfall and throughfall volume data. The water flux for bulk rainfall and throughfall was calculated by dividing the average water volume by the surface area of the collector. Nitrogen deposition (TDN, DIN and DON) was calculated by multiplying the water volume with the element concentration in that volume. One of the most widely used methods to disentangle dry deposition from canopy leaching is the ‘canopy budget model’, which assumes that a natural tracer ion (most commonly Na+) is inert when deposited on the canopy. Additionally, the method is based on the calculation of an ion charge balance based cation deposition and computes canopy exchange interactions of NH4

+, NO3- (Staelens et al. 2008).

However, several assumptions underlying this model seem to be challenged in tropical forest canopies, not at least the apparent uptake of Na+ in some cases (Tobón et al. 2004). Additionally, the model offers no theoretical basis for the calculation of organic N compounds from the canopy, other than the assumption that they are equally deposited on the canopy as the tracer ion. Given these factors, we chose not to apply this model here (Hofhansl et al. 2011). Satellite and wind trajectory analysis We calculated a fire load index (FLI), representing the theoretical number of fires that were passed by the wind parcels that arrive above the experimental site. For this, we crossed a daily fire pixel dataset with backwards wind trajectories ending at the site. The fire pixel dataset covering the entire African continent was obtained for the study period from NASA’s MODIS Collection 6 NRT (NASA n.d.). The backwards wind trajectories were generated using the Hybrid Single Particle Lagrangian Integrated Trajectories (HYSPLIT) model provided by the National Oceanic and Atmospheric Administration Air Resources Laboratory (NOAA ARL) (Draxler 1997, 1999, Draxler


21

and Hess 1998, Stein et al. 2015) with the GDAS half-degree archive meteorological dataset. For the entire study period, we generated one trajectory every 6 hours ending above the study site pixel for five different target altitudes (respectively 40, 500, 1000, 2500 and 5000 meters above ground level), going back one week. To cross the datasets, we counted fires within a 25 km buffer around the modelled air parcels at each time point going back 24 hours. The 24 hour period was considered to take into account smoke from smouldering fires that might not have been detected by the VIIRS system at the time point in question. These fire counts were cumulated along the trajectory, resulting in a fire count per trajectory, and hence four fire counts per day for the site. We then accumulated the daily fire counts for every fortnight, resulting in the fire load index (FLI) for the concurrent sampling dates. Additionally, we extracted the hourly simulated black carbon column mass density (BCCMD) from above our study site from the Modern-era Retrospective Analysis for Research and Applications version 2 (MERRA-2) from the Goddard Earth Observing System Model, version 5 (Global Modeling and Assimilation Office n.d.). We subsequently calculated the average over this time series for all sample dates. 3.3. Results and Discussion Monitoring N deposition throughout a hydrological year in the remote forest of the central Congo Basin, we found a high total dissolved nitrogen (TDN) open field (bulk) deposition of 18.2 kg N ha-1 yr-1, of which 70% was organic N. Throughfall, the precipitation collected under the forest’s canopy, amounted to 53.1 ± 3.2 kg N ha-1 yr-

1 in mixed forest, and 37.5 ± 4.2 kg N ha-1 yr-1 in monodominant forest, with roughly 50% organic N in both cases (mean ± SD, Appendix B4). The time series of deposition showed a consistent series of peaks at the onset of the rainy seasons in March and September, suggesting a strong contribution from previous dry deposition (Figure 3.1). This elevated deposition was not expected at this remote site, because there is no industrial activity or intensive agriculture in the vicinity. Moreover, these measured N deposition rates already surpass the simulated deposition rates for 2030 (Reay et al. 2008).

To gain further insight into the source and composition of the organic fraction (50-70% of total N deposition (Appendix B4)), we characterized the compound classes of the organic components using ultra-high resolution Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS). The composition of dissolved organic matter (DOM) in the three sample types (i.e. open field rainfall, mixed forest and monodominant throughfall) showed a striking similarity (Appendix B1). For each sample type, the absolute and sub-group relative abundances of N-containing compounds were consistent (Appendix B2a-c). The annual average molecular composition plotted in van Krevelen space also yielded, overall, very similar distributions (Appendix B2d-f, Appendix B1). We found a high number of assigned molecular formulae in all sample types, with an average of 17,096, 16,955 and 15,627 for the open field rainfall, mixed forest, and monodominant forest throughfall, respectively. Of these formulae, 52% contained at least one N atom on average, resulting in a large number of N containing compounds with similar composition in all sample types (Appendix B4, Appendix B1). One question that arises from this data is to what extent the increased organic N load in throughfall is caused by dry deposition versus canopy leaching. The relative abundances of compound classes in the three sample types, with a high percentage of unsaturated phenolics, closely resembles a previous study of glacier DOM on the Tibetan plateau, where atmospheric deposition


22

was identified as the main source (Spencer et al. 2014). The overall molecular similarity of the two forest types suggests that the throughfall enrichment of N is due to canopy interception of dry deposition rather than canopy leaching, as canopy leaching would be expected to result in a divergent compositional signature between the two forest types. In fact, we found only 2,811 and 2,459 unique formulae in the mixed and monodominant forest throughfall samples, respectively, that were not found in the open field rainfall samples. Of these formulae, only 1,380 and 1,039 contained at least one N atom. These unique formulae only contributed 0.94 and 0.86% relative abundance to the total DOM signature for the mixed and monodominant forest samples, respectively (Fig 3.2d-f). Collectively, this qualitative compositional evidence suggests that the vast majority of N found in our forest throughfall samples was from atmospheric deposition.

Quantitatively measuring the contribution of canopy exchange and dry deposition in throughfall samples has proven to be a challenging task, especially in the tropics where appropriate methods still need to be developed (Hofhansl et al. 2011). Rather than using Na+ or Cl- as an inert tracer (see Methods section), we calculated an average condensed aromatic content using FT-ICR-MS data as proxy for fire-derived N deposition component of the organic fraction. Condensed aromatics are formed only through combustion and are therefore not a canopy leaching product (Golberg 1985). The DON deposition loads increased linearly with increasing average condensed aromatic content, which suggests that fire-derived dry deposition controls organic N loads in the throughfall (Appendix B2). The difference in N deposition and condensed aromatics content between mixed lowland and monodominant forests might be due to the increased heterogeneity of the mixed forest, which would lead to differing levels of aerosol interception of the canopies. Gilbertiodendron dewevrei, the monodominant species constituting a minimum of 60% of the basal area, is characterized by large leaves. Hypothetically, the combination of a homogenous canopy and large leaves might render them less efficient for dry deposition removal from the air, much like the described differences in fine dust removal efficiency between tree species (Chen et al. 2017). Altogether, this suggests that organic N deposition found at all sites is predominantly sourced from biomass burning and not the forest canopy itself. We expect that at least part of the inorganic deposition comes from dry deposition as well, but there are no satisfactory techniques to further confirm this using the dataset.

Although the abundance of fires on the savanna borders has been shown before (Chen et al. 2010), our data suggests that the impact on central African forests in terms of N addition is much higher than expected. To investigate this link, we crossed existing datasets of daily fire pixels (recorded by MODIS satellites and processed through NASA’s FIRMS system) with daily backwards wind-trajectories as simulated for the study period by the HYSPLIT model of NOAA (Movie S1). This gave rise to an empirical fire load index (FLI, Figure 3.1a), which represents the number of fires that the winds directly above our sites passed over during the week prior to the sampling. A striking seasonality in this FLI showed that the local dry season is the peak season for fire arrivals for the central African rainforests. This was confirmed by extracting the black carbon column mass density (BCCMD) from NASA’s Merra 2 model (Figure 3.1a). Black carbon is a known combustion product, originating from either fossil fuel combustion or biomass burning (Golberg 1985), and hence can also serve as a proxy for the fire load of the atmosphere.


23

Figure 3.1. Seasonality of total dissolved nitrogen (TDN) deposition, rainfall and biomass burning-derived compound arrival. The graphs show (a) biweekly deposition of total dissolved nitrogen (TDN) in the open field (red), mixed lowland tropical forest (green) and monodominant lowland forest (blue), showing mean ± SD values for each forest types, with additionally (b) the rainfall seasonality (mm) in the open field (red), mixed lowland tropical forest sites (green) and monodominant lowland forest sites (blue); evolution of our empirical Fire Load Index (FLI) (c) above our study site, i.e., the number of fires that the winds directly above our sites passed over during the week prior to the sampling (x1000) at different altitudes (colors), along with the evolution of time-averaged black carbon column mass density (BCCMD) above the study site as simulated by MERRA-2 (black line). The gray and white zones denote the rainy and dry season, respectively. Lower maps (d) show the monthly averaged BCCMD for, from left to right, October 2015, and February, April and July 2016, with increasing intensity from yellow to brown. Throughout the year, low pressure areas move from the northern savanna zone, over the equatorial forest, to the southern savannas and back, causing shifts in rainfall patterns around the inter-tropical convergence zone (ITCZ, approximate position indicated as a black line on the maps). Both natural fires, as well as those resulting from farmers preparing their land for the coming growing season, cause biomass burning products to arrive in the equatorial tropical forest during the dry season, with our study site indicated by the red dot on the maps. The coloration on the same map shows the ecosystem type delineation, with tropical rainforest in dark green.


24

The inter-tropical convergence zone (ITCZ) shifts from the southern savanna border in January to the northern savanna border in July, inducing shifts in the rainy seasons for the savanna borders. As a result, the dry seasons oscillate inversely from the north to the south, in January – July, and back. During these shifts, the ITCZ passes over the central Congo Basin twice, inducing a bi-modal rainfall pattern. It has been shown that the location of the ITCZ and the interoceanic confluence on the African continent causes biomass combustion products from the savanna borders to be transported in the direction of the equator during the dry season, and shifts of the ITCZ introduce seasonality in fire regime with latitude (Figure 3.1)(Cachier and Ducret 1991). These novel, field-based observations, combined with molecular signatures of fire-derived organic compounds that correspond well with model-based fire inputs from the atmosphere, raise substantial questions as to the overall impact of high N deposition in central African forests. The fire abundance on the African continent has been chronically high during the last millennia (Marlon et al. 2009, Archibald et al. 2012), suggesting that N deposition to the Congo Basin has also been high over this period. Additionally, given the agreement between the FLI and the BCCMD (Figure 3.1), annual and monthly deposition maps and aerosol transport models suggest that the entire Congo Basin is subject to similar, if not higher, deposition loads (Appendix B3, Appendix B6). Currently available global simulation maps, which are used by modelers for reactive nitrogen loads, predict values for central Africa between 4-10 kg DIN ha-1 yr-1 (Dentener et al. 2006, Vet et al. 2014, Wang et al. 2017). While these estimates correspond well with the DIN loads we found in the open field, they dramatically underestimate total N deposition since organic N is not accounted for, which was also shown by recent simulations (Kanakidou et al. 2016). Although the bioavailability of this organic N is still being studied (Bronk et al. 2007, Cornell 2011), the discrepancy between simulations and our data raises questions about the consequences of this high N deposition on forest functioning, biodiversity and the regional C budget of central African forests. The latter is of particular importance given that African forests currently represent one of the largest sources of uncertainty for the global CO2 budget (Ciais et al. 2011a). It is generally assumed that productivity in lowland tropical forests is not N limited (Vitousek and Sanford 1986, Hedin et al. 2009). Moreover, symbiotic N fixation by canopy trees is usually down-regulated in mature central African forests, confirming that N is cycled in excess (Chapter 2). However, it remains unclear to what extent reactive N loads affect N and C cycling and biodiversity in tropical forests (Janssens et al. 2010, Hietz et al. 2011). It is possible that the delivery of reactive N impacts forest structure, functioning and biodiversity via effects on taxa, functional groups, and size classes (Obbink et al. 2010, Alvarez-Clare et al. 2013). We therefore suggest an urgent expansion of the monitoring efforts in the Congo Basin, in order to adjust and improve global and regional estimates of reactive nitrogen loads.


25

Figure 3.2. Molecular characterization of rainfall and throughfall samples, based on a subset of samples from the open field and from one of the monodominant and one of the mixed forest plots respectively. Relative abundance of different compound classes (a) of dissolved organic molecular formulae in the open field rainfall and throughfall of monodominant and mixed lowland tropical forest samples, with (b) the relative abundance of nitrogen containing compounds in the samples, and (c) the relative abundance of nitrogen containing compounds within each separate compound class; molecular formulae (e-f) of the open rainfall, monodominant forest and the mixed forest, respectively, plotted in the van Krevelen space, using the Oxygen to Carbon ratio and the Hydrogen to Carbon ratio on respectively x and y-axis. Relative abundance is the mean intensity of the plotted formula. Bars and error bars in a-c represent the mean over the samples per forest type ± SD.

Acknowledgements Some analyses and visualizations used in this paper were produced with the Giovanni online data system, developed and maintained by the NASA GES DISC. We also acknowledge the use of data and imagery from LANCE FIRMS operated by the NASA/GSFC/Earth Science Data and Information System (ESDIS) with funding provided by NASA/HQ. This research has been supported by the Belgian


26

Development Cooperation through VLIR-UOS, both through a personal scholarship of M.B. and project funding. VLIR-UOS supports partnerships between universities and university colleges in Flanders (Belgium) and the South looking for innovative responses to global and local challenges. Visit www.vliruos.be for more information. Acquisition of mass spectra supported by the National Science Foundation Division of Materials Research (DMR-11-57490). P.H.-F. is funded by Programa de Formación de Personal Avanzado CONICYT, BECAS Chile. Supporting Information Additional Supporting Information may be found in Appendix B. Appendix B1 Molecular characterization of rainfall and throughfall samples in central Congo Basin. Appendix B2 Linear regression of mean annual relative abundance of condensed aromatics and annual organic N deposition Appendix B3 Annual-averaged black carbon column mass density over the African continent. Appendix B4 Total dissolved nitrogen (TDN), dissolved inorganic nitrogen (DIN), dissolved organic nitrogen (DON) and average molecular characteristics (mean ± SD) Appendix B5 Backwards wind trajectories (white dot lines) and fire pixels (red dots) per day, looped through the study period. Appendix B6 Monthly averaged Black Carbon Column Mass Density over the last 5 years over the African continent, with increasing intensity from yellow to brown.

27

II Biogeochemical interaction of vegetation and

environment

Chapter 4: Functional identity explains carbon sequestration in a 77-year-old experimental tropical plantation

28

Chapter 4: Functional identity explains carbon sequestration in a 77-year-old experimental tropical plantation After: Bauters, M., Ampoorter, E., Huygens, D., Kearsley, E., De Haulleville, T., Sellan, G.,

Verbeeck, H., Boeckx, P. and Verheyen, K.: Functional identity explains carbon sequestration in a 77-year-old experimental tropical plantation, Ecosphere, 6, 1–11, 2015.

Abstract Planting forests is an important practice for climate change mitigation, especially in the tropics where the carbon (C) sequestration potential is high. Successful implementation of this mitigation practice requires knowledge of the role of species identity and diversity on carbon accrual of plantations. Despite this need, solid data on the long-term development of forest plantations are still very scarce. Monospecific and two species mixture plots of a 77-year-old tree diversity experiment in Yangambi in the Congo basin were fully inventoried. We calculated above-ground C stocks using allometric equations, and soil C stocks by analyzing soil samples at multiple depths. Linear mixed effects models were used to analyze the effect of taxonomic and functional identity and diversity on the aboveground and soil carbon stocks. Apart from a species identity effect, the proportion of planted species in the total stand basal area (BApl) and effective species richness were identified as compositional parameters with a significant effect on the aboveground carbon (AGC), with BApl being more important. Both AGC and BApl were coupled to the functional identity of the planted species; the planting of short-lived pioneers led to low AGC. We found no clear benefits, but also no drawbacks, for AGC of two species mixture plots over monospecific plots or including nitrogen fixing species in the plantation scheme. However, the latter was the only compositional parameter with a significant positive effect on the soil carbon stock up to 1 m depth. We conclude that the different plantation configurations gave rise to a wide range in carbon stocks. This was predominantly caused by large differences in AGC sequestration over the past 77 years. Altogether, short-lived pioneer species had a low BApl resulting in low carbon sequestration, while partial shade tolerant species achieved the highest AGC stocks. Tolerating spontaneous ingrowth during the plantation development can further increase the AGC stock, given that the appropriate functional type is planted.

Chapter 4 Cover. Yangambi is a research center roughly 100 km downstream of Kisangani. Ir. Bernard Bonyoma (left) and ir. Henri Badjoko (right) know the forest experiments really well. The coloration on the map shows wet tropical forest (darkgreen), moist deciduous tropical forest (light green) and montane tropical forest (red).


29

4.1. Introduction Forest ecosystems contain 45% of the terrestrial carbon (C) stock and are directly interacting with the atmosphere through C sequestration, deforestation and forest degradation (Bonan 2008b). The latter two processes give rise to 95% of the C emissions in the tropics, which equals nearly 10% of the global fossil fuel emissions (Achard 2004, Houghton et al. 2012, Le Quéré et al. 2015). This stresses the importance of carbon offset projects in the tropics, such as afforestation or reforestation, as currently traded in the voluntary carbon market or, to a lesser extent, financed by the Clean Development Mechanism (Canadell & Raupach 2008; Jindal, Swallow & Kerr 2008; Cerbu, Swallow & Thompson 2011). However, implementation of such programs frequently results in plantations of fast growing monocultures of non-native species (Cossalter & Pye-smith 2003). While this approach guarantees high carbon sequestration rates, it ignores other important co-benefits of reforestation in the tropics, such as natural forest restoration and biodiversity recovery, for which native species are better suited (Lamb, Erskine & Parrotta 2005; Greve et al. 2013). But the combination of a bigger species pool, along with the lack of knowledge on the species in the tropics, makes the selection of appropriate species difficult and hazardous for the success of a carbon-offset project (van Breugel et al. 2011). There is a clear need for results of long-term experiments to create a set of management guidelines for selection of suited native species. Research efforts in biodiversity and ecosystem functioning have pointed out that both biodiversity and species identity have an important effect on a variety of ecosystem processes, such as ecosystem productivity and carbon sequestration (Loreau et al. 2001, Hooper et al. 2005, Cardinale et al. 2012). However, most of the published studies focus on short-lived grassland ecosystems (Tilman et al. 2006), from which the results are not directly transferable to structurally more complex ecosystems with longer turnover times such as forests (Hillebrand and Matthiessen 2009). The emphasis of biodiversity and ecosystem functioning research has therefore recently shifted to these long-lived ecosystems, resulting in e.g. tree diversity experiments (Scherer-Lorenzen et al. 2005). These experiments intend to have an orthogonal design, which avoids the risk that differences in management or soil conditions mask the link between tree diversity and ecosystem processes (Scherer-Lorenzen et al. 2005). Yet, most of these manipulative experiments have been installed during the last decade and were setup in the temperate regions (Verheyen et al. 2015), leaving tropical ecosystems understudied. Moreover, results of plantation studies are predominantly derived from experiments with fast growing, non-native tree species (e.g. Forrester et al. 2006). The Yangambi reserve, in the center of the Democratic Republic of Congo (DRC), holds a tree diversity experiment, set up by the Belgian colonials. It provides us with information on 77 years of experimental forest development in the tropics. This study is the first to bring results on this experiment, where we analyze the long-term effects of tree species identity and admixture effects on carbon sequestration, which is highly relevant for the development of carbon plantations in the tropics. 4.2. Methods Study area and experimental set-up This study was performed in a 77-year-old tree diversity experiment, in the surrounding

of the Yangambi Man and Biosphere Reserve (N00’47°; E24’30°), in the DRC.


30

Yangambi is situated approximately 100 km downstream of Kisangani, on the northern

bank of the Congo River. The region has a tropical rainforest climate, Af-type

according to the Köppen climate classification (Peel et al. 2007a), with an annual

rainfall of 1750 mm, one short dry season from January to February, and a

temperature of 24.5°C throughout the year. The site has a gentle topography and soils

are ferralsols (Van Ranst et al. 2010). The experiment was established in 1938 by the

colonial Belgians, which held a large tropical agriculture and forest research center at

this site. Information on the setup and the initial follow-up of the experiment could be

tracked in the Belgian State Archive, until the decolonization in 1960. Since then, the

experimental sites have been under the protection of the INERA (Institut National pour

l’Etude et la Recherche Agronomique), the Congolese national institute for agricultural

research, which means that there have been no logging activities until today. The total

experimental plantation holds more than 50 ha of experimental plots. These were

planted with a wide variety of tree species in different tree species diversity levels

(ranging from monospecific plots up to mixtures of six species). The total planted

species pool consisted of 23 tree species, listed in Table 4.1. However, the majority of

the plots contained only one or two planted tree species. Almost no repetitions of the

different configurations were planted. The experiment contains plots of both 60 by 60

m (0.36 ha) and 40 by 40 m (0.16 ha). All plots were nursed and kept clear of

spontaneous ingrowth, for ten years after planting, except those where Pericopsis

elata [(Harms) Meeuwen] was planted, which were nursed for 20 years. After that, the

plots were deliberately left unmanaged, so spontaneously in-growing species now

accompany the planted species, augmenting the total realized species pool to 143 tree

species in all plots.

For this study, we wanted to assess both the tree species identity and admixture effects, i.e. the effect of adding one other species in the monoculture, on the long-term carbon sequestration. We therefore selected 13 target tree species, i.e. the species which were planted in at least one monospecific and one two-species mixture, as shown in Table 4.1. We grouped the plots in 13 groups, each time consisting of the monoculture(s) and admixture(s) of the target species with the admixed species. Remark that in some cases, the admixed species is also a target tree species, and thus also found as a monoculture in the plantation. We inventoried a total of 29 plots, consisting of 14 monocultures and 15 two-species mixtures. Each plot was subdivided in either 4 (for the 0.16 ha plots) or 9 subplots of 20 by 20 m (for the 0.36 ha plots), resulting in a total dataset of 201 subplots.

Data Collection

An international standardized protocol for tropical forest inventories (RAINFOR, Malhi et al. 2002) was used. We considered both the planted tree species (one or two) and the species that spontaneously established in the subplots. In each subplot, the diameter of all living trees with a diameter larger than 10 cm was measured at 1.3 m height and the trees were identified to species level. Tree height was measured on 20% of all individuals in each plot, selected across all the diameter classes, using a hypsometer (Vertex III, Haglöf, Sweden). We use abbreviations of the planted species, of which the full scientific name can be found in Appendix C1. Extra parameters were assigned to each plot to indicate the nursing treatment (group of P.elata was kept clear of spontaneous ingrowth for a longer time) and the inclusion of nitrogen fixing tree


31

species in the plantation scheme. We treated all Fabaceae members as potential nitrogen fixers (see Appendix C1). At five different places within every plot, soil samples were taken at five depth increments (0-10, 10-20, 20-30, 30-50 and 50-100 cm) and these samples were pooled per depth increment. All samples were dried for 48 hours at 60°C. Carbon and nitrogen content were analyzed using an elemental analyzer (Automated Nitrogen Carbon Analyser, interfaced with a Sercon 20-20 IRMS; SerCon, Cheshire, UK). In addition, spread over the whole study area, seven soil pits to 1 m depth were dug, and bulk density was measured in the wall of these pits at different depths (20, 40 and 80 cm) using container rings of known volume (Eijkelkamp Agrisearch Equipment, Giesbeek, The Netherlands). Furthermore, we determined the bulk density of the upper soil layer (0 -10 cm) at three locations within each plot using container rings. Table 4.1. Experimental design table showing the different configurations that were assessed in this study. We separated the target species (upper part of the table) and the admixed species (lower part). Full names for the abbreviations in the column header can be found in the first column. The diagonal in the upper part of the table shows the monocultures of the target species that are present.

Target species A.c. E.a. E.c. G.c. L.t. M.a. M.e. P.t. P.o. P.m. P.e. P.s. S.t.

Autranella congolensis x

Entandrophragma angolense

x x

Entandrophragma cylindricum

x x

Guarea cedrata x x x

Lovoa trichilioides x x

Mammea africana x

Milicia excels x

Pachyelasma tessmannii x

Panda oleosa x x

Pentaclethra macrophylla x

Pericopsis elata x x x x

Pterocarpus soyauxii x

Strombosiopsis tetrandra x x

Antrocaryon nannanii x

Blighia welwitschii x

Carapa procera x

Chrysophyllum africanum x

Drypetes likwa x

Khaya anthotheca x

Phyllanthus spec. x

Strombosia grandifolia x

Treculia africana x

Zanthoxylum gilletii x

Data analysis For the calculation of the aboveground carbon stock (AGC) in the trees, we used the formula of Chave et al. (2005), including wood density and tree height. First, we fitted different diameter-height relationships from literature (see Appendix C2) by non-linear least-squares estimations for every plot, using the actual tree height measurements from the field. The best fit for every plot was selected based on the Akaike Information Criterion (AIC) and the residual variation. Broken and strongly leaning trees were measured separately and were not included in the fitting process. The best fit was then used to estimate the unknown tree heights. We used wood density data that were


32

collected in the surrounding natural forest by taking 5 x 5 x 5cm3 wood samples from under the bark (Kearsley et al. 2013). If no species data was available from this dataset, we used species averages from the DRYAD global wood density database (Zanne et al. 2009, Chave et al. 2009) or genus averages from both databases to assign wood densities to the individual trees. We assumed a carbon content of 50% in the woody biomass for the final carbon stock calculations. For the soil organic carbon stock (SOC) down to one meter depth, the averages of the bulk density measurements from the soil pits were used in combination with the specific carbon content from the plot-level composite samples for the bottom four increment layers. Because of the higher variation of the bulk density of the topsoil and the high contribution of this soil layer to the total SOC stock to 1 m depth, we used the plot-specific bulk density and C measurements for the topsoil. Based on the results of the inventory, we derived subplot-level planted tree species richness (monospecific or two species mixture), effective tree species richness (planted tree species as well as spontaneous ingrowth), effective Simpson diversity, and the proportion of planted species in the stand basal area (BApl). In a second stage, we replaced the taxonomic target species group as a fixed effect with functional types that summarized the functional role of the planted species (functional type). For that purpose, we compiled a simple trait matrix with information on wood density and shade tolerance of all the planted tree species. Wood density is an important trait, strongly linked to the functional ecology of tree species (e.g. Chave et al. 2009), while shade tolerance is a crucial life-history trait, associated with a wide range of physical and chemical plant traits (Valladares & Niinemets 2008). Information on shade tolerance was derived from literature (Lebrun & Gilbert 1954; Hawthorne 1995; Hubau et al. 2012), giving priority to the publication of Lebrun & Gilbert (1954), who based their classification on field observations of seedlings in natural forest in Yangambi. They used three classes – light-demanding, shade-tolerant and shade species - with increasing tolerance to shade. Based on this trait matrix, we performed a hierarchical clustering on the different tree species, and divided them in functional types. We choose four functional types, since there were two distinct groups of light-demanding species with differing wood density properties (Table 4.2). We tested the differences in wood densities between the types using a non-parametric Kruskal-Wallis test, and the significant association of the shade tolerance with the types using a non-parametric chi-square test.


33

Table 4.2. Functional traits and included tree species for each type; short-lived pioneers (SLP), long-lived pioneers (LLP), partially shade tolerant (PST) and shade tolerant (ST) trees. The average wood density (WD) was calculated by using the average wood density of the tree species present in each class (averages ± the standard deviations). Shade tolerance was classified based on two literature sources (Lebrun & Gilbert 1954; Hawthorne 1995) using three classes (light-demanding, tolerant to shade, shade).

Class WD (g cm-3) Light strategy

Species

Short-lived pioneers (SLP)

0.50 ±0.02 Light-demanding

Antrocaryon nannanii A.n.

Entandrophragma angolense E.a.

Entandrophragma cylindricum E.c.

Khaya anthotheca K.a.

Lovoa trichilioides L.t.

Milicia excelsa M.e.

Treculia africana T.a.

Long-lived pioneers (LLP)

0.71 ±0.08 Light-demanding

Blighia welwitschii B.w.

Pentaclethra macrophylla P.m.

Pericopsis elata P.e.

Phyllanthus spec. P.sp.

Pterocarpus soyauxii P.s.

Zanthoxylum gilletii Z.g.

Partial shade-tolerant (PST)

0.68 ±0.05 Tolerant

Autranella congolensis A.c.

Chrysophyllum africanum C.a.

Drypetes likwa D.l.

Mammea africana M.a.

Pachyelasma tessmannii P.t.

Shade-tolerant (ST) 0.61 ±0.06 Shade

Carapa procera C.p. Guarea cedrata G.c.

Panda oleosa P.o.

Strombosia grandifolia S.g.

Strombosiopsis tetrandra S.t.

To analyze the influence of tree species diversity and composition on subplot-level AGC (n=201), we applied linear mixed effects models. Plot was set as random intercept and AGC was log-transformed. We started off with a model containing all calculated compositional parameters for the subplots as fixed effects (i.e. the target species group, planted tree species richness, effective species richness, effective Simpson diversity, the BApl, the presence/absence of nitrogen fixers among the planted species and the nursing treatment). The non-categorical variables were standardized to make the parameter estimates of the fixed effects comparable. First, we tested the random structure, keeping the fixed effects structure constant (with parameter estimation via restricted maximum likelihood). Once the optimal random structure was found, the fixed effects were backwards selected, based on the AIC and likelihood ratio tests using maximum likelihood estimations. Finally, the marginal and conditional R2 were calculated for the final model, which indicate the proportion of the variance that is explained by the fixed structure, respectively the fixed and random


34

structures together (Nakagawa & Schielzeth 2013). We additionally checked Spearman correlations in the pooled dataset (n=201) between BApl, effective species richness and AGC, to increase the interpretability of the mixed effects model results.

Subsequently, we used the functional types to replace the taxonomic target species groups as a fixed effect in the model. Both final models were compared using likelihood ratio tests and AIC. To gain insight in how BApl, as an important fixed effect, was linked to the functional identity of the planted species, ANOVA was additionally done, with BApl as dependent variable and the functional types as independent variable on the plot level (n=29).Finally, we checked whether the SOC stock (to 1 m depth) was correlated to AGC on the plot level, by calculating Spearman correlation coefficients. Subsequently, we repeated the modelling steps on the plot-level for the SOC with all the compositional parameters in the first model as fixed effects. All statistical analyses were conducted with R version 3.1.1 (R Development Core Team 2015). We used the ‘nmle’ package for the mixed effects modeling (Pinheiro et al. 2013). 4.3. Results The overall AGC and SOC is 212 ± 106 Mg C ha-1, respectively 83 ± 16 Mg C ha-1 (Figure 2.1), with the standard deviations calculated on the plot-level averages, resulting in coefficients of variation of 0.50 (AGC) and 0.19 (SOC). The subplot-level AGC stocks in the complete dataset (n=201) range from 31 Mg C ha-1 to 731 Mg C ha-

1. The average plot-level compositional characteristics per plot are given in Appendix C3.

Figure 4.1. Aboveground carbon (AGC) (Mg C ha-1) and soil organic carbon (SOC) (Mg C ha-1) stocks in the different plots. Plots (x-axis) were grouped based on the target tree species they share (indicated with abbreviations on top), hence some two-species mixtures appear twice in the graph. The AGC are averages based on the AGC of the subplots, and the error bars indicate the standard deviation on the subplot averages. The SOC stock values are calculated using composite samples from the whole plot, and thus have no error bars. The full scientific names for the abbreviations of the target tree species are listed in Appendix C1.

The functional clustering resulted in four distinct types (Table 4.2); two light demanding types with different wood densities (further called short-lived pioneers (SLP) and long-lived pioneers (LLP)), a class with high wood density and mediate shade tolerance (partial shade-tolerant species (PST)) and a high-wood density class with high shade


35

tolerance (shade-tolerant species (ST)). The wood densities and shade tolerance significantly differed between the types (p=0.003, respectively p=0.001). The optimal model structure for the taxonomical approach contained plot as a random intercept. The backward selection of fixed effects resulted in a model with three significant effects: the BApl, the effective species richness and the target species groups (significant, p < 0.1, for four groups), with the parameter estimates and significance shown in Table 4.3. This resulted in a final model with a marginal and conditional R2 of respectively 0.51 and 0.55 (Table 4.3). Simple Spearman-correlations in the dataset (n = 201) show a significant positive correlation between BApl and AGC, a significant negative correlation between effective species richness and BApl, and a weak correlation between effective species richness and AGC (Figure 4.2). The parameter estimates of the functional types in the model (Table 4.3) pointed out that planting species of type PST (Table 4.2) had a positive effect on AGC, in contrast with LLP, ST and especially SLP. Fixed effects selection resulted in the same model as the first model, with functional types replacing the target tree species groups, and the differences in nursing as an additional significant, positive fixed effect. The optimal taxonomical and functional model performed equally well. The ANOVA analysis revealed a significant negative effect of the presence of the SLP class in the plantation scheme on the BApl (p=0.004).The top soil layer (0-10 cm) held on average 33% of the total SOC stock to 1 m depth. The coefficient of variation of the bulk densities decreased with increasing soil depth; 0.16 (5 cm), 0.15 (20 cm), 0.07 (40 cm) to 0.03 (80 cm). There was no significant correlation between AGC and SOC to 1 m depth on plot-level. Including N-fixers in the plantation scheme was the only significant, positive, effect on SOC stocks (r2 =0.11 and p=0.03). 4.4. Discussion Both the average and the coefficient of variation of the AGC stocks were considerably higher than SOC stocks. This shows that the impact of the management choices of tropical plantations have a relatively higher impact on AGC in terms of absolute C sequestration compared to SOC stocks. Although soils are an important carbon stock globally, we target management guidelines for carbon sequestration in tropical plantations, so devote more attention to AGC (Houghton 2005). Remark that we did not assess carbon stocks related to the below-ground biomass. The classification of the functional groups was done using only shade tolerance and wood density, which is consistent with previous work (e.g. Poorter, Bongers & Bongers 2006). We stress that these functional types should also be mainly interpreted in terms of shade tolerance and wood density. Tree species in functional types SLP and LLP have a high light requirement, but differ in their allocation to mechanical stability or growth rate (Selaya and Anten 2008). PST species can establish in low light conditions but need gaps to grow, while the ST species can grow in the low light conditions (Poorter et al. 2006).


36

Table 4.3. Model specifications for the linear mixed effects models of the above-ground carbon (AGC) stocks (n=201). The random structure of each of the models in the table consists of the plot as a random intercept. Model 1 shows the parameters for the taxonomical-approach model, including the target tree species groups as fixed effects (full scientific names for the abbreviations of the target species are given in Table A1). Model 2 shows the results for the functional-approach, with introduction of four functional types (short-lived pioneers (SLP), long-lived pioneers (LLP), partial shade tolerant (PST) and shade tolerant (ST) trees for the target tree species; with specifications in Table 4.2). BApl is the ratio of basal area of the planted species to the total stand basal area in the plot (including spontaneous ingrowth). ESR is effective species richness and equals the number of occurring tree species (planted species and spontaneous ingrowth) in the subplot. AIC stands for Akaike Information Criterion, which was used as selection criterion. R2

marg and R2cond, indicate the proportion of variance

explained by the fixed structure, and the fixed and random structure together, respectively (Nakagawa and Schielzeth 2013).

Fixed effects AIC R2marg R2

cond p-value parameter estimate

Model 1: taxonomical approach 259 0.51 0.55

BApl <0.001 0.26 ±0.04 ESR <0.001 0.18 ±0.04 Groups A.c. 0.001 0.63 ±0.28 E.a. 0.24 -0.24 ±0.23 E.c. 0.38 -0.22 ±0.24 G.c. 0.20 -0.19 ±0.23 L.t. 0.003 -0.51 ±0.26 M.a. 0.76 0.07 ±0.28 M.e. 0.14 -0.47 ±0.27 P.e. 0.07 0.19 ±0.20 P.m. 0.88 -0.21 ±0.26 P.o. 0.48 0.21 ±0.25 P.s. 0.77 -0.22 ±0.28 P.t. 0.06 0.43 ±0.28 S.t. 0.28 -0.14 ±0.25

Model 2: functional approach 257 0.54 0.55

BApl <0.001 0.26 ±0.04 ESR <0.001 0.20 ±0.04 Nursing 0.03 0.30 ±0.12 Functional Class SLP 0.02 -0.35 ±0.14 LLP 0.55 -0.07 ±0.13 PST 0.01 0.40 ±0.15 ST 0.24 -0.12 ±0.10

Tree composition effects on above-ground carbon stocks Setting the plot factor as a random intercept allowed us to quantify unwanted local differences in e.g. soil conditions. The effect of the target species groups is an identity effect that is to be expected from this type of experiment, and has been observed in similar contexts (e.g. Balvanera et al. 2006; Redondo-Brenes 2007; Ruiz-Jaen & Potvin 2011). In terms of AGC, two species mixtures did not outperform the monocultures of the target species (Figure 3.1). This shows that other processes, such as the species identity effect and spontaneous ingrowth of tree species, were more important in the long-term development of these plantations. In the target species


37

group of A.congolensis (A.c. in Figure 3.1), the AGC increase in the mixture with D. likwa compared to the monoculture is remarkable. However, it is important to be cautious with the interpretation of the results; in case there is no monoculture of the admixed species, a positive or negative mixing effect could be falsely reported using only this dataset (Hulvey et al. 2013). BApl and effective species richness both have a significant positive effect on the AGC and SOC of this plantation. In the pooled dataset (n=201), however, these two predictors are negatively correlated to one another and the positive effect of effective species richness is masked (small negative correlation with AGC; Figure 3.2). Based on these combined observations, we conclude that BApl

has a more important effect than species number. We state that, given a high BApl, the species that manage to grow into the stand spontaneously are probably using different niches than the planted species. This way, spontaneous ingrowth can enlarge the resource use of the stand (Loreau and Hector 2001). Considering a low BApl, spontaneous ingrowth is of large importance to compensate for the failure of the planted species to establish. In this case, effective species richness reflects the success of spontaneous species to establish, and augment the carbon storage. Hence within each configuration, both parameters show a positive effect on the AGC (Table 4.3), although the BApl is the dominant driver (Figure 3.2), and should thus be priority in tropical plantation design. The lack of a significant difference in model performance when summarizing the 13 taxonomic groups into 4 functional groups proofs that the species identity effect is strongly linked to the functional life-history of species. From the final model results, we conclude that planting with species of class PST, has a bigger carbon storage potential in the long-term, while using SLP is clearly not a good practice. P. elata was the only target species which was nursed for ten years longer in all plots (two species mixture and monospecific). This is a light demanding species, belonging to class LLP (Table 4.2) and performed in general better than the other, functionally equal species. We cannot safely generalize this observation of a positive effect of extra nursing, since only one functional class and one species had this treatment. In this case, however, keeping the plantation free of spontaneous ingrowth for a longer time, did not only affect the BApl, but also the AGC in the plantations with light-demanding species. Through the significant negative effect of SLP trees in the plantation scheme on the BApl, it is shown that SLP are not successful in carbon sequestration through a failure in BApl. The inclusion of N-fixers in the plantation scheme did not significantly contribute to the AGC, although it has been shown that the effect of N-fixers strongly depends on site conditions and complementarity with the other planted species (Forrester et al. 2006). Additionally, tropical old-growth forests are generally considered N-rich leading to a down-regulation of symbiotic N-fixation (Batterman et al. 2013). In the case of the Yangambi experiment, N-fixers may have been actively fixing in the early stand development, but its effect may have been masked by growth-effects in the long-term.


38

Figure 3.2. Interrelationships between aboveground carbon (AGC), BApl and effective species richness in the different subplots (n=201). AGC stands for above ground carbon in the woody biomass of the subplots. BApl is the ratio of basal area of the planted species to the total stand basal area in the plot (including spontaneous ingrowth). The effective species richness is the number of tree species with a DBH>10 cm that were present in the subplot. The r-values are Spearman correlation coefficients, along with their p-value.

Tree composition effects on soil carbon stocks Including N-fixers in the plantation scheme was the only compositional parameter with a significant, positive effect on the SOC stock, which is known from other reforestation sites in the tropics (e.g. Resh, Binkley & Parrotta 2002). The plots with higher AGC were not associated with high SOC stocks. Although literature reports a positive correlation of aboveground productivity with root allocation (Raich et al. 2014), total SOC stock development after land use change is more complex and less understood (Laganière et al. 2010). Initial carbon content and tree species have been identified as key determinants for the soil carbon processes following afforestation (Laganière et al. 2010, Shi et al. 2013). As we have no baseline data on the initial SOC stocks, we cannot quantify actual changes in soil carbon stock. However, taking into account the relative importance of spontaneous ingrowth, we assume that in this case the effective species composition rather than the planted species determines the changes in this stock. This augments the complexity of disentangling the role of tree species composition in the SOC stock formation, as the effective species pool rises to 143 tree species in the total study area, rather than only considering the 23 planted species. For the long-term development of tropical plantations, we state that changes in the absolute SOC stock, following afforestation, are small compared to the sequestration in AGC. This observation, combined with the far higher variability of the AGC stock and the complexity of predicting the SOC changes when considering processes like spontaneous ingrowth, leads to the suggestion that plantation managers should focus on aboveground C sequestration. Management implications BApl is an important driver of carbon stocks in tropical plantations, and is linked to the functional identity of the planted species. This identity should thus be a primordial consideration for reforestation projects in the tropics, which aim at both the re-establishment of native forest and successful carbon sequestration. The species identity can be translated and generalized into very basic functional types, where low wood-density short-lived pioneers had a clear negative impact on both the BApl and carbon accrual in the long run. Given a high BApl of planted species, spontaneous ingrowth positively affects the carbon sequestration. In this study, we found no clear


39

positive effect of admixing species (but not a negative one either). Including nitrogen fixers in the plantation did not lead to higher above ground carbon stocks, although these plots showed a higher soil carbon stock. Acknowledgements This research is part of the COBIMFO project, funded by the Belgian Science Policy Office (BELSPO). We thank the people from the INERA, which have helped intensively with the inventory and the species identification, as well as all the people from the University of Kisangani (UNIKIS) and the Institut Facultaire des sciences Agronomiques (IFA) that also contributed to this publication. We thank Katja van Nieuland and Stijn Vandevoorde for their help with the sample processing and chemical analyses. Supporting Information Additional Supporting Information may be found in Appendix C. Appendix C1 A list of the abbreviations of the full scientific names of planted species Appendix C2 List of H:DBH relations Appendix C3 Compositional characteristics of the different plots

Chapter 5: Functional composition of tree communities changed topsoil properties in an old experimental tropical plantation

40

Chapter 5: Functional composition of tree communities changed topsoil properties in an old experimental tropical plantation After: Bauters, M., Verbeeck, H., Doetterl, S., Ampoorter, E., Baert, G., Vermeir, P.,

Verheyen, K. and Boeckx, P.: Functional Composition of Tree Communities Changed Topsoil Properties in an Old Experimental Tropical Plantation, Ecosystems, 20, 861–871, 2017.

Abstract Forest biogeochemistry is strongly determined by the interaction between the tree community and the topsoil. Functional strategies of tree species are coupled to specific chemical leaf traits, and thus also to litter composition, which affects mineral soil characteristics. The limited understanding on this interaction is mainly based on shorter-term common garden experiments in temperate forest and needs to be extended to other forest types and climates if we want to understand the universality of this linkage. Especially for highly diverse tropical forests our understanding of this interaction remains limited. Using an old experimental plantation within the central Congo basin, we examined the relationship between leaf and litter chemical composition and topsoil properties. Canopy, litter and topsoil characteristics were measured and we determined how the community-level leaf and litter chemical composition altered the topsoil carbon, major plant nutrients and exchangeable cation concentration, acidity, and pH over the last eight decades. We found that functional composition strongly affected topsoil pH. In turn, topsoil pH strongly determined the soil total carbon and available phosphorus, total nitrogen, and exchangeable potassium. Our results indicate that, as observed in temperate common garden experiments, trees alter chemical topsoil properties primarily through soil acidification, differently induced by functional composition of the tree community. The strong link between this community-level composition and topsoil characteristics on a highly representative soil type for the tropics, improves our understanding of tropical forests biogeochemistry.

Chapter 5 Cover. The views from Yangambi on the Congo river and beyond can be breathtaking. Team awesome from Yangambi after coming back from a long field campaign, from left to right: myself, Héritier, Bernard, Augustin, Henri and Kibinda. The coloration on the map shows wet tropical forest (darkgreen), moist deciduous tropical forest (light green) and montane tropical forest (red).


41

5.1. Introduction

The link between soil properties and tree communities has fascinated ecologists and soil scientists for decades (Zinke 1962, Binkley and Giardina 1998, Finzi et al. 1998c, Reich et al. 2005, Hobbie et al. 2007). To which extent this link is caused by tree species effects, i.e. the presence of certain tree species altering local soil characteristics or by environmental filtering, remains poorly understood. Tree communities are composed of different tree species with different plant functions and strategies (Díaz et al. 2015), with inherent implications for the canopy leaf chemistry (Wright et al. 2004). It is, hence, evident that the long-term presence of species with different canopy leaf chemistry, as commonly found in natural forests, affects at least the topsoil. Recently, research started to focus on the afterlife effects of the variation in this functional leaf chemistry on its abiotic environment (Freschet et al. 2012). The current understanding of mechanisms through which trees can alter their environment and the soil in particular is primarily based on studies of monocultures in a ‘common garden‘ setting (Reich et al. 2005, Vesterdal et al. 2008, De Schrijver et al. 2012). This approach has the advantage that a variety of species are planted in adjacent blocks, so that climate, geology, previous land use, topography and hydrology are relatively homogeneous, and that the observed differences in soil properties in function of time can be inferred to a tree species effect (Binkley and Giardina 1998). In turn, the essential long-term commitment linked to experiments with woody species limits the ease of this type of research and therefore our knowledge of tree species effects is mainly based on a relatively limited set of existing experiments. Additional to the common garden approach, a large amount of studies observed the covariation of soil properties, tree species traits and communities on a larger scale, in more complex, natural forests (Zinke 1962, Finzi et al. 1998b). Such correlative studies have also been done on a large scale in the tropics (John et al. 2007, Toledo et al. 2012, Quesada et al. 2012), but they fail to infer the causality of observed correlations between soil and tree communities. Indeed, as Sollins (1998) pointed out, such studies are a first step, and we need experimental approaches to increase our understanding of the underlying mechanisms in the tropics. For temperate forests, considerable research efforts in planted monocultures have pointed out that trees affect soil properties through fresh leaf and litter chemical composition by direct (e.g. by altering the acidity of the topsoil) and indirect (e.g. through earthworm abundance) effects (Reich et al. 2005, De Schrijver et al. 2012, Mueller et al. 2012). Hence, in temperate forest ecosystems, poor-quality foliage with low non-hydrolyzing cation (as Ca2+) concentration induces soil acidification, lowered microbial activity and increased carbon (C) stocks in the forest floor. Vesterdal et al. (2013) conclude from their meta-analysis that the effect of tree species on mineral C stocks is weak and inconsistent. However, they are rightfully cautious with their statement, by noting that there are only very few common gardens older than 50 years and that tree species effects on mineral soils are likely to become apparent only in the longer run. In any case, additional studies on the effect of trees on soil properties are needed in multiple environments with different geology and climate. To our knowledge, there are only few studies on tree species effects in the Neotropics, ranging in age from 4 to 15 years (Fisher 1995, Stanley and Montagnini 1999, Russell et al. 2007). Nevertheless, lower latitudes offer a very different pedologic and climatic setting to extend our knowledge from the temperate region. Deeply weathered, old tropical soils like Oxisols and Ultisols cover ca. 57% of the surface area (Kauffman et al. 1998) and are characterized by low primary mineral concentration, low pH, low non-hydrolyzing


42

cation concentration, low phosphorus (P) concentration and high amount of exchangeable acidity (USDA 1999). The younger, temperate soils are in general in a higher buffer range compared to strongly weathered, tropical soils. Additionally, the low pH greatly affects both P availability, the major limiting nutrient in lowland tropical ecosystems, and aluminum mobility, which affects the toxicity of the abiotic environment for plants and the microbial community (Matson et al. 1999). Because of these factors, we hypothesize that tree species effects in these poorly buffered tropical systems to be different from the younger soils from the temperate region. In this study, we characterized the chemical links within and between canopy, litter and mineral topsoil of 29 different plots in a unique experimental forest plantation in the central Congo Basin, established nearly 80 years ago (Chapter 4). After a nursing period of ten years, the plots were deliberately left unmanaged, so spontaneously in-growing species now accompany the planted species, augmenting the total realized species pool to 143 tree species in all plots. Additionally, the current diversity at this site bridges the gap between planted monocultures in common gardens, and correlative studies in natural systems. As such, it is a unique site allowing us to test whether our current knowledge of less complex experiments conducted in temperate forest is generalizable to near-natural situations in a different biome in the Tropics. The complexity and diversity of the current experiment, as well as the differing geology and climate, make this nearly 80-year-old tropical experimental plantation an excellent setup to extend the current knowledge on tree-soil relationships from previous experiments. Although this study is sheer correlational, the ‘common garden’ setup on itself allows us to link causality to these observations. Hence, our objectives for this study were (1) to assess the effect of different community-level functional traits on important soil properties, (2) to assess whether the observed mechanisms of tree species effects from temperate common gardens are transferable to the community-level in this complex setup from the Tropics, and (3) to test whether the effects on a longer timescale extend to deeper soil layers. This can take us one step closer to generalize a tree species effect from local experiments to a natural context. 5.2. Material and methods Setup and sampling The nearly 80-year-old experimental Yangambi plantation lies in the center of the Democratic Republic of the Congo. The region has a tropical rainforest Af climate according to the Köppen climate classification (Peel et al. 2007a) with an annual rainfall of 1750 mm and two short dry seasons (January to February and July to September) at an average temperature of 24.5°C throughout the year. The site has a gentle, nearly levelled topography and soils are classified as Oxisols (Van Ranst et al. 2010). The site was maintained from the plantation date in the 1930’s until 1960 by Belgian colonial researchers, and has since been under the protection of INERA (Institut National pour l’Etude et la Recherche Agronomique), the Congolese national institute for agricultural research. All the plots were initially planted in a randomized block design in within the approximate area of 1 km2, hence with homogeneous climate, relief and parent material. For this study we used 29 plots of either 60 m by 60 m (0.36 ha) or 40 m by 40 m (0.16 ha), which were planted with tree monocultures or two species mixtures (for specifications see chapter 4 and Appendix D1), summing up to a total planted species pool of 23 tree species.


43

An international standardized protocol for tropical forest inventories (RAINFOR, Malhi et al. 2002) was used to inventory the plots, and hence all aboveground living stems with a diameter at breast height greater than 10 cm were measured and determined to species level. Each plot was subdivided in either 4 (for the 0.16 ha plots) or 9 subplots (for the 0.36 ha plots) of 20 m by 20 m, resulting in a total of 201 subplots. We characterized the nutrient concentration of the topsoil, litter layer and tree canopy in three randomly chosen subplots per plot (87 subplots in total; 3 replications per plot). In each subplot, we took a composite sample of the top 0-5 cm and 5-10 cm layer of mineral soil at 5 random locations (4 or 9 replications per plot). Additionally, we collected plot-level composite samples of deeper soil layers from five different places within every plot (10-20, 20-30, 30-50 and 50-100 cm; no replications). The litter layer was collected from one randomly located 0.25 m by 0.25 m square per subplot (4 or 9 replications per plot). The canopy of every subplot was characterized by selecting the most abundant tree species, constituting 95% of the basal area of the selected subplots, and by sampling mature leaves of a minimum of two individuals per species using tree climbers. This sampling was carried out at the level of the total experiment. Species occurring in multiple subplots were not sampled in every subplot, resulting in a total of 65 tree species and 354 individuals sampled. Sample analysis Topsoil (0 - 5 cm), litter and leaf samples were dried for 48 hours at 60°C. Roots were picked out of the soil samples before grinding and subsequently C and nitrogen (N) concentration of tree and soil samples were analyzed using an elemental analyzer (Automated Nitrogen Carbon Analyzer), interfaced with an Isotope Ratios Mass Spectrometer (IRMS; 20-20, SerCon, UK). The soil pH (pHKCl) of each sample was analyzed using a glass electrode (Model 920A, Orion, England) after suspension of 14 ml soil in a 70 mL KCl (1 M) solution (ISO10390). Exchangeable Ca, Mg and K were quantified by saturating cation exchange sites with ammonium acetate buffered at pH 7.0 and by subsequently measuring the cation concentrations in the filtrated extracts with Atomic Absorption Spectroscopy (AAS) (AA240FS, Agilent Technologies, US). Exchangeable Al was extracted by 1 M KCl solution and determined by inductively coupled plasma (ICP; Iris intrepid II XSP, Thermo Scientific, US) and optical emission spectrometry. Additionally, we titrated this extract with sodium hydroxide to determine the exchangeable acidity (H+ and Al3+) (Van Ranst et al. 1999). For total P, soil samples were treated with HClO4, HNO3 and H2SO4 for 4 hours at 150°C, after which phosphate concentrations were measured colorimetrically with malachite green (Varian Cary50, Agilent Technologies, US; Lajtha et al. 1999). Resin P was analyzed as a proxy for plant-available P (Saggar et al. 1990, Johnson et al. 2003). This was done by shaking 1 g of soil for 16 hours with distilled water and activated resin membranes, and subsequently shaking the membranes for 16 hours in a 0.5 M HCl solution. P analysis of the extracts was done using standard colorimetry kits. Fresh leaf and litter samples were dry-ashed at 550°C for 5.5 hours; the ash was dissolved in 2M HCl solution and subsequently filtered through a P-free filter. The aliquots where then analyzed for total P by colorimetry and for calcium (Ca), magnesium (Mg) and potassium (K) by AAS (Ryan et al. 2001). Additionally, to have a general idea of the topsoil texture, sand vs silt and clay concentration of 11 randomly selected, pre-treated (organics destroyed with 20% H2O2; aggregates disaggregated with 0.1M Na4P2O7) soil samples of the 0 - 5 cm layer was determined. The procedure for C and pH was repeated for the soil samples of the deeper soil layers.


44

Data analysis Average leaf chemical trait values (C, N, P, C:N, N:P, Ca, Mg, K) were calculated for every selected species. Subsequently, we assigned a basal area weighted-average canopy value to every subplot using the subplot’s species composition, to calculate an average nutrient concentration for the canopy of every selected subplot (n=87). For litter and soil, we had one (composite) sample per subplot. We then averaged all subplot values for canopy, litter and soil variables to plot-level variables (n=29), and further used this plot-level variables in the statistical analyses. Principal component analyses (PCA’s) were done on the standardized variables for every compartment (soil, litter, and canopy) separately to check variability and correlations within each of these compartments. For interpretability we rotated the first two principal components of the soil, litter and canopy compartment, using a Varimax rotation. For the links between canopy, litter and topsoil, we calculated bivariate Pearson correlations between all the variables and subsequently used multiple linear regression to link canopy and litter characteristics to multiple key topsoil variables: C concentration, N concentration, available P, C:N ratio, N:P ratio (with available P), pH and exchangeable acidity (H+ and Al3+). We considered both canopy and litter chemical variables as plant-related traits to relate the biotic aboveground part to the topsoil for two reasons: (1) each species has specific nutrient resorption strategies, meaning that the canopy nutrient values and ratios can deviate from the actual litterfall input, and (2) each species and individual has a different leaf productivity, which is not accounted for in our basal-area weighted mean. As such, litter layer sampling includes freshly fallen leaves as well as partly decomposed leaves. To avoid problems with collinearity, because of the strong correlations of some of our predictor variables, we (1) first did preliminary stepwise multiple regression with the canopy and litter variables separately (using a starting set of 8 variables for both compartments, as shown in Appendix D2) for the different response variables, and (2) used only the retained variables from this first step for a stepwise regression including both canopy and litter variables. In stepwise multiple regressions, Akaike’s information criterion (AIC) was used as stopping criterion, which penalizes for the number of predictor variables that are retained in the model. After every parameter selection step, the variance inflation factors (VIF) were calculated to verify if the retained variables did not cause multicollinearity problems. If the VIF of two or more variables was greater than 3, we retained the single variable with the strongest Pearson correlation with the response variable (Zuur et al. 2010). Finally, we estimated the extent of the tree-community effects on soil carbon and soil pH. For this, canopy and litter layer characteristics of the different plots were related to the pH and the carbon concentrations of the different soil layers, using the same multiple linear regression model. All statistical analyses were conducted with R version 3.2.3 (R Development Core Team 2015). 5.3. Results Patterns within soil, litter and canopy Topsoil texture was considered fairly constant on the experimental area with average sand concentration of 85.3 ± 2%. Furthermore, the sandy and acidic topsoils were characterized by very low concentrations of exchangeable non-hydrolyzing cations, with pHKCL ranging from 3.09 to 3.72. The relative ranges of element concentration in the three compartments were comparable, except for C, C:N and Ca (Table 5.1).


45

Carbon concentration in the soil, on average 22.5 g kg-1, ranged three-fold, while only 10 to 25% in litter and canopy. On the contrary, the C:N ratio and N concentration ranged three-fold in the canopy and two-fold in the litter compartment, while only 20% in the soil, with an average of 13.4. The Ca concentration ranged four- and two-fold in the canopy and litter layer, while ten-fold in the soil compartment. Total soil P ranged two-fold while bio-available P ranged more than five-fold, with averages of respectively 240 mg kg-1 and 12 mg kg-1. On average, the litter N:P concentration was higher than the canopy N:P, while the C:N ratios were similar. Soil exchangeable aluminum and exchangeable acidity ranged two-fold and were two orders of magnitude higher than the non-hydrolyzing cation concentrations.

Table 5.1. Average element concentration for the three compartments. Averages ± SD (and ranges) for the 29 plots are shown. Ϯ Expressed in mg g-1 for canopy and litter, and in meq 100 g-1 for soil.

Variable Canopy Litter Soil (0-5 cm)

C (g kg-1) 471.7±12.5 (448.0 – 497.0) 426.7±28.3 (359.2 – 475.6) 22.5±7.3 (13.1 – 41.9)

N (g kg-1) 32.2±6.7 (17.3 – 41.9) 26.4±3.7 (18.6 – 33.1) 1.7±0.5 (1.1 – 2.9)

C:N 16.27±4.58 (11.6 - 29.51) 16.62±3.00 (12.63 - 25.42) 13.39±0.88 (11.91 - 15.79)

P (g kg-1) 1.4±0.3 (0.7 – 2.2) 0.8±0.2 (0.4 – 1.2) 0.24±0.04 (0.16 - 0.30)

N:P 24.19±4.05 (16.95 - 32.88) 32.24±5.01 (24.55 - 46.22) 7.04±2.39 (3.96 - 13.92)

KϮ 10.72±2.94 (5.94 - 18.44) 2.01±0.58 (0.65 - 3.42) 0.07±0.01 (0.05 - 0.11)

MgϮ 2.57±0.92 (1.03 - 4.42) 2.66±0.82 (0.94 - 5.60) 0.07±0.02 (0.04 - 0.10)

CaϮ 5.75±1.96 (2.40 - 9.39) 9.30±2.77 (3.54 - 20.15) 0.06±0.04 (0.02 - 0.21)

P available (mg kg-1)

12.23±3.69 (4.70 - 26.15)

Al (meq 100 g-1)

3.98±0.77 (2.66 - 6.33)

pH (KCl) 3.38±0.18 (3.09 - 3.72)

Acidity (meq 100 g-1)

7.06±1.45 (5.01 - 10.67)

The rotated PCAs visualize the correlations of the nutrient chemistry within the soil, litter and canopy (Appendix D3 and D4). Both canopy and litter show a similar structure of variation, with N:P ratio and Ca and Mg concentration along the first axis, and C:N ratio and N concentration along the second axis. In the canopy compartment, PCA axis 1 and 2 together explained 84% of the variation, while for the litter layer only 65% of the variation was explained by the first two axes. For soil, we included more variables in the analysis and 64% of the variation was explained. A total of 43% of the variation was explained by the first axis, and visualizes a clear relation between pH on the one hand and acidity, available P, C, N, K and C:N on the other. The N:P ratio, together with Ca and Mg concentration were strongly correlated with the second axis. Patterns between soil, litter and canopy compartments In general, canopy and litter chemical variables were well correlated (Appendix D5). The bivariate statistics are shown in Appendix D, will not be discussed in full detail here. The link between topsoil and both canopy and litter, is expressed via multivariate regression for topsoil C, N, available P, C:N, pH and acidity (Table 5.2). This revealed that canopy C concentration explained as much variation in topsoil C concentration as three litter variables; C concentration, C:N ratio and Mg concentration, with the latter


46

having negative predictor coefficients. With canopy and litter both in the model, only canopy C concentration and C:N ratio of the litter were retained, explaining 30% of the variation in topsoil C. For N concentration, the sets of significant predictors were very similar to those for the topsoil C concentration, but K concentration was also retained as significant predictor for litter, which now explained more variation than the canopy. For the canopy variables, again only canopy C concentration was retained as a significant parameter and a selection on the combined predictor sets resulted in a model explaining 44% of the variation in topsoil N concentration. For soil bio-available P, Ca concentration was the only significant, and negative, predictor of the canopy, while for litter the P concentration, N:P ratio, K concentration and Ca concentration were retained. Together the N:P ratio of the litter, the litter K concentration and the (negative) Ca concentration significantly explained 48% of the variation of bio-available-P. Overall, pHKCl and C:N showed the strongest links to the plant-related predictors (respectively 58 and 60% of the variation explained by canopy and litter; Table 5.2, but also see Appendix D5). The stepwise regression parameter selection for the C:N ratio resulted in respectively canopy C concentration and C:N ratio, and litter C:N ratio and Ca concentration to predict the topsoil C:N ratio. By a selection on both retained sets of predictors, only canopy C:N ratio and litter Ca concentration were retained. Similar, bivariate correlations were found between topsoil pH and canopy C, P, K, Mg and Ca concentration and N:P ratio. For litter, only Mg and Ca concentration were correlated with topsoil pH. In the regression models, respectively the canopy C and P concentration were retained, with the N:P ratio, while for litter the C concentration and (negative) N concentration were the significant predictors for topsoil pHKCl.


47

Table 5.2. Retained significant predictors of multiple regressions on different soil (0-5 cm) response variables. The stepwise regression was done with separate predictor sets for canopy (subscript ‘can’) and litter (subscript ‘lit’) first, of which the retained predictor variables were combined in a third multiple regression. Variable abbreviations used: carbon (C), nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), C:N and N:P ratio. Significance levels for the retained predictor variables were P > 0.1 (non-significant; ns); P < 0.1 (*), P < 0.05 (**) and P < 0.001 (***).

Negative parameter estimates were indicated with (-).

Soil Response R2adj P Explanatory Significant Predictors

C 0.26 0.003 Canopy Ccan**

0.26 0.016 Litter Clit**, C:Nlit

**(-), Mglit**(-)

0.30 0.007 Canopy + Litter Ccan**, C:Nlit

*(-)

N 0.17 0.032 Canopy Ccan**


**(-), Klit*, Mglit

**(-) 0.44 <0.001 Canopy + Litter Ccan

***, C:Nlit*(-), Klit

**

Available P 0.15 0.021 Canopy Cacan**(-)

0.37 0.003 Litter Plit*, N:P lit

**, Klit**, Calit

**(-) 0.48 <0.001 Canopy + Litter Cacan

***(-), N:P lit**, Klit

***

Total P 0.00 >0.1 Canopy Pcanns, N:Pcan

ns, Cacanns (-)

0.02 >0.1 Litter Clitns(-), Plit

ns(-), Mglitns(-)

0.07 >0.1 Canopy + Litter Cacanns (-)

C:N 0.50 <0.001 Canopy Ccan**, C:Ncan

** 0.36 0.001 Litter C:Nlit

*, Calit**(-)

0.58 <0.001 Canopy + Litter C:Ncan**, Calit

**(-)

pH 0.48 <0.001 Canopy Ccan**(-), N:Pcan

**(-) 0.49 <0.001 Litter Nlit

**(-), Calit***

0.60 <0.001 Canopy + Litter Ccan**(-), Nlit

**(-), Calit**

Acidity 0.15 0.050 Canopy Mgcan*(-)

0.24 0.011 Litter Clit**, Mglit

**(-) 0.24 0.011 Canopy + Litter Clit

**, Mglit**(-)

The regression combining the retained predictors of the canopy and litter, canopy C concentration and litter N and Ca concentration explained 60% of the variation. For acidity, only the Mg concentration was retained as significant predictor for the canopy, while C and Mg concentration were retained as significant predictors for the litter. The latter were the only significant variables in the regression with both predictor sets, which explained only 24% of the variation. Mineral soil C concentration in every soil layer was significantly explained by a variable set of canopy and litter chemical characteristics (Table 5.2). The R2

adj of the models combining canopy and litter predictor sets ranged between 0.21 and 0.34 in the soil layers of the upper 30 cm and the lower 50 cm, while for the 30-50 cm layer it was very low and not significant (R2

adj=0.08 and p=0.08). For soil pH, canopy and litter traits significantly explained the variation in the soil layers until 30 cm depth (Table 5.3). In these layers, the multiple regressions in general better explained the pH variation than the C concentration of the corresponding layer, except for the 10-20 cm depth layer. 5.4. Discussion Range in Topsoil conditions The high sand concentration and the low amount of non-hydrolyzing, exchangeable cations in the soils imply that the cation exchange capacity (CEC) of the mineral soil is mainly controlled by soil organic matter. This also explains the strong positive


48

correlation between C in the topsoil and exchangeable Al (Appendix D3c, D2). Aluminum clearly dominated the cation exchange complex (Table 5.1). The strong link between soil C and total N, available P and exchangeable K also follows from the fact that the mineral matrix of these acid, sandy soils is low in abiotic, mineral derived nutrients. Hence, nutrient availability is more likely determined by the amount of topsoil organic matter. In this study site, the contribution of the top 10 cm soil C stock to the total soil C stock (to 1 m depth) ranged from 11 to 30% (unpublished data), which highlights the importance of this thin layer for element cycling. The soil C concentration in this upper layer showed a more than three-fold in the experiment, which is perhaps the most striking result of this study. This soil C concentration was highly correlated with soil pH (Figure 5.1c), which, in turn, enhanced Al availability (Chorover and Sposito 1995; Gruba and others 2013). In acidic conditions as in Oxisols, ligand exchange between OH groups of Fe- or Al-oxides and carboxyl or phenolic OH-groups on the OM is likely to play an important role in the OM stabilization (Gu et al. 1994, Nabnera et al. 2006). Additionally, there are important toxicity effects of acid pH and available Al on soil fauna and the microbial community, as well as increased and effective cation bridging between the trivalent Al and soil C, stabilizing the soil C and both lowering the decomposition processes (Mulder et al. 2001, Hobbie et al. 2007, Mueller et al. 2012). No correlation between soil exchangeable Ca and pH, exchangeable acidity and Al concentration was found (SI 2). Very acidic soils with low exchangeable cation concentrations are in the aluminum buffer range as shown by the dominance of exchangeable Al on the exchange complex (Table 5.1). The low concentrations of Ca or Mg on the cation exchange complex will thus have little effect on the soil pH.


49

Depth (cm) C concentration pH R2

adj P Explanatory Predictors R2adj P Explanatory Predictors

0-5 0.26 0.003 Canopy Ccan** 0.49 <0.001 Canopy Ccan

**(-), N:Pcan**(-)


**(-), Mglit**(-) 0.46 <0.001 Litter Nlit

**(-), Calit***

0.26 0.003 Canopy + Litter Ccan**, C:Nlit

*(-) 0.6 <0.001 Canopy + Litter Ccan**(-), Nlit

**(-), Calit**

5-10 0.2 0.02 Canopy C:Ncan**(-), Kcan

**(-) 0.42 <0.001 Canopy Cacan***

0.32 0.003 Litter Nlit**, Plit

*(-) 0.39 <0.001 Litter Nlit***(-), Mglit

** 0.34 0.002 Canopy + Litter Kcan

**(-), Nlit*** 0.42 <0.001 Canopy + Litter Cacan

***

10-20 0.17 0.04 Canopy Ccan**, C:Ncan

**(-) 0.09 0.06 Canopy Ncan*(-)

0.21 0.02 Litter Plit**, Calit

**(-) 0.13 0.06 Litter Plit**(-), Calit

* 0.21 0.02 Canopy + Litter Plit

**, Calit**(-) 0.13 0.06 Canopy + Litter Plit

**(-), Calit*

20-30 0.1 0.05 Canopy Ncan* 0.5 <0.001 Canopy Ccan

**(-), Ncan***(-), Kcan

** 0.32 0.005 Litter Clit

**(-), Plit**, Calit

**(-) 0.43 <0.001 Litter Clit**, Plit

***(-), Calit**

0.32 0.005 Canopy + Litter Clit**(-), Plit

**, Calit**(-) 0.5 <0.001 Canopy + Litter Ccan

**(-), Ncan***(-), Kcan

**

30-50 0 ns Canopy - 0 Ns Canopy - 0.08 0.07 Litter Klit

*(-) 0 Ns Litter - 0.08 0.07 Canopy + Litter Klit

*(-) 0 Ns Canopy + Litter -

50-100 0 ns Canopy - 0 Ns Canopy - 0.23 0.004 Litter N:Plit

**(-) 0 Ns Litter - 0.23 0.004 Canopy + Litter N:Plit

**(-) 0 Ns Canopy + Litter -

Table 5.3. Multiple regression for soil C and pH in different soil layers with the plant-related canopy (subscript ‘can’) and litter (subscript ‘lit) chemistry as explanatory variables and the soil organic carbon concentration of the different depth increment layers as response variables. Variable abbreviations used: carbon (C), nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), C:N and N:P ratio. Significance levels for the retained predictor variables were P > 0.1 (non-significant; ns); P < 0.1 (*), P < 0.05 (**) and P < 0.001 (***). Negative parameter estimates were indicated with (-).


50

The C and N stoichiometry of the canopy, litter and topsoil invoke the idea of a strong microbial control on the soil C in the upper 5 cm, as suggested by previous research (Russell et al. 2007, Hobbie et al. 2007). At least two observations from our study confirm this: (1) reduced variability in C:N ratios of the topsoil compared to canopy and litter (Table 5.1), caused possibly by microbial processes narrowing the divergence of resource stoichiometry during the build-up of soil C, and (2) the lack of correlations between C and N concentration within the canopy and litter, along with the strong correlation in the topsoil (Appendix D2). However, further research on the soil microbial community in the different plots is needed to corroborate this hypothesis.

Figure 5.1. Correlation plots illustrating the links between canopy chemistry and topsoil properties (0-5 cm) in this experiment. Canopy chemistry was correlated with soil pH (shown for canopy carbon (a) and canopy calcium concentration (b)). This range in soil pH was linked to a strong range in soil carbon (c), available phosphorus (d), soil total nitrogen (e) and exchangeable potassium (f) concentration. The points represent the different plots in the setup (n=29), plots with high above-ground carbon (AGC) are black, mediocre dark grey and low AGC light grey.

How does tree functional composition affect top soil properties of a tropical Oxisol? The randomized planting design of the experiment allows us to safely exclude that the species pool in the different plots was filtered by soil properties (also see Appendix D6). Strongest multivariate relations between canopy and litter traits and soil variables were found for soil pH, explained by leaf C concentration, litter N and Ca. This suggests that tree species are primarily affecting soil conditions by changing the soil pH, in turn inducing the build-up of a higher soil C concentration via mechanisms discussed above. In the framework of previous work on tree-species effects, three


51

mechanisms have been hypothesized to contribute to this link between functional composition and soil acidification: (1) increased input of organic acids, (2) increased soil respiration and (3) sequestration and redistribution of non-hydrolyzing cations by the vegetation, resulting in a net loss of cations from the soil, which is compensated by protons (Jobbágy and Jackson 2003). The latter mechanism (3) invokes a species-specific strategy for non-hydrolyzing cation cycling. Although the mechanisms underlying species-specific Ca cycling are still not fully understood, it has been made clear that species-specific physiologic differences are indeed the main drivers for this chemical leaf trait (Dauer et al. 2007). In our case, the canopy Ca concentration correlated with both litter and soil Ca concentration, but no link was present between litter and soil Ca. The positive link between soil pH and canopy Ca (Figure 5.1) and Mg, corroborates with observations in temperate forest (Finzi et al. 1998a, Reich et al. 2005). As no direct correlation between pH and Ca or Mg on the exchange complex in the mineral soil was found, the input and fluxes of these cations through litter decomposition is probably not the main mechanism for decreases in soil pH values. Instead, both mechanism (1) and (2) are more likely to cause acidification; (1) high C concentration and low non-hydrolyzing cation concentration of foliage and litter of the species on more acid topsoils are linked to low leaf and litter pH and high organic acid concentrations (Cornelissen et al. 2006), and (2) the high cation concentration in litter is predominantly balanced by organic anions, and the decarboxylation of these organic anions during microbial decomposition is a proton-consuming process, hence causing a pH increase (Xu et al. 2006). Additionally, N fixation is known to be a soil acidifying process (Nyatsanga and Pierre 1973, Bolan et al. 1991). Bonnier (1957) reported that he only found nodules on potential fixers in very recently disturbed forest soils during his two years of research at Yangambi, and that nodules were absent in all other forest sites, as also observed in a nearby site (Chapter 2). This suggests that symbiotic N-fixation would no longer be actively influencing soil pH in this nearly 80-year-old experiment, although we do not have data on this. However, in the original experiment three potential N-fixing legumes were planted and we cannot exclude that this had no significant effects on soil in the early development stages of the experiments. Our results suggest that the community-level leaf and litter chemistry have altered important topsoil characteristics (C, total N, available P, exchangeable K and Al; Figure 5.1, c to f) predominantly via an effect on soil pH (Figure 5.1, a to b), which corroborates with what we know from temperate forests (Hobbie et al. 2007, Mueller et al. 2012), and from other shorter-term tropical experiments (Fisher 1995, Russell et al. 2007). However, our data does not allow to fully determining the underlying mechanism. We’ve provided information on the above ground carbon (AGC) stocks in Figure 5.1 (Chapter 4). From this graph it is clear that the effects of the community-level leaf chemistry on mineral topsoil are also strongly linked to our previous results from this setup. The more soil-acidifying communities, with clear effects on other topsoil chemical characteristics, gave rise to higher aboveground carbon stocks, compared to the less acidifying communities in Figure 5.1a. The functional identity of tree species is thus of vital importance for biogeochemistry in our study sites. Depth of the effect In our analysis of pH and C concentration in different soil layers, we found evidence that the effects of the tree communities extend to deeper soil layers on the timescale


52

of this experiment. For both C concentration and pH, we show that links between the functional composition of the tree community and the soil characteristics are to be found until at least 30 cm depth. Overall, pH variability was better explained than the soil C concentration. This confirms our findings from the topsoil layer, i.e. that trees control soil C and nutrient concentrations primarily via effects on soil pH. Furthermore, the tree communities’ effects on pH in deeper soil layers seem also more consistent than on soil carbon (Table 5.3). As Dawud et al. (2016) have shown, both tree diversity and functional identity affect the distribution of C in deeper soil layers. It is likely that these diversity effects are also important in this experiment. Additionally, we did not sample roots in this experiment, although it is known that they contribute considerably to soil C stocks (e.g. Schmidt and others 2011). The limited and varying R2

adj values might thus (partly) be caused by these important root contributions. Recent work, however, has shown that root chemistry is closely linked to the species-level leaf chemistry and (Valverde-barrantes et al. 2015), in which case the discussed effects of nutrient concentrations and stoichiometry of the input material remain valid. Acknowledgements This research is part of the COBIMFO project, funded by the Belgian Science Policy Office (BELSPO). Marijn Bauters holds a VLADOC grant from the Flemish Interuniversity Cooperation (VLIR-UOS). We thank the people from the INERA (Institut National pour l’Etude et la Recherche Agronomique), which have helped intensively with the inventory and the species identification, as well as all the people from the University of Kisangani (UNIKIS) that also contributed to this publication. Supporting Information Additional Supporting Information may be found in Appendix D. Appendix D1 Design table of the experiment Appendix D2 Correlation between chemical characteristics within the canopy (a), the litter (b) and the soil (c) compartment Appendix D3 Varimax-rotated PCA bi-plots for the canopy (a), litter (b) and soil (0-5 cm) (c) compartment. Appendix D4 Scores of the Varimax rotated-principal components. Appendix D5 Pearson correlations between canopy and litter (a), soil and litter (b) and canopy and soil (c) compartment Appendix D6 Correlation plots of the actual (y axis) versus the planted canopy characteristics (x axis).

Chapter 6: Parallel functional and stoichiometric trait shifts in South-American and African forest communities with elevation

53


After: Bauters, M., H. Verbeeck, M. Demol, S. Bruneel, C. Taveirne, D. Van der Heyden, L.

Cizungu, and P. Boeckx. 2017. Parallel functional and stoichiometric trait shifts in South American and African forest communities with elevation. Biogeosciences 14:5313–5321.

Abstract The Amazon and Congo basin are the two largest continuous blocks of tropical forest with a central role for global biogeochemical cycles and ecology. However, both biomes differ in structure and species richness and composition. Understanding future directions of the response of both biomes to environmental change is paramount. We used one elevational gradient on both continents to investigate functional and stoichiometric trait shifts of tropical forest in South America and Africa. We measured community-weighted functional canopy traits and canopy and topsoil δ15N signatures. We found that the functional forest composition response along both transects was parallel, with a shift towards more nitrogen conservative species at higher elevations. Moreover, canopy and topsoil δ15N signals decreased with increasing altitude, suggesting a more conservative N cycle at higher elevations. This cross-continental study provides empirical indications that both South-American and African tropical forest show a parallel response with altitude, driven by nitrogen availability along the elevational gradients, which in turn induces a shift in the functional forest composition. More standardized research, and more research on other elevational gradients is needed to confirm our observations.

Chapter 6 Cover. The tropical montane forests of both Ecuador and Rwanda are amazingly beautiful. Measuring big trees is a team-effort, left to right: Fidèle, myself, Cys Taveirne (on top) and Dries Van der Heyden in Nyungwe national park. The coloration on the map shows wet tropical forest (darkgreen), moist deciduous tropical forest (light green) and montane tropical forest (red).


54

6.1. Introduction A good understanding of the future response of tropical forest ecosystems to global change is required because of their vital role in global biogeochemical cycles and ecology. However, due to the long turnover times in forest ecosystems, it is hard to acquire insight in these future responses. As a result, empirical research has since long turned to studying ecosystems along natural gradients, which can greatly advance our understanding of ecosystem ecology and function in response to environmental shifts. Elevational gradients in particular offer open-air space-for-time experiments. Contrary to latitudinal gradients or elevational gradients in the higher latitude zones, they are not complicated by changes in seasonality, and with careful interpretation can offer great insights in tropical forest functioning (Körner 2007, Malhi et al. 2010, Sundqvist et al. 2013). Hence, elevational transects have been postulated as a viable and useful setup to assess long-term ecosystem responses to environmental changes, and serve as an empirical tool to assess future trajectories of forest ecosystems under global change (Malhi et al. 2010, Sundqvist et al. 2013). This has invoked research efforts on transects in South America, but no such studies have been carried out in central African forests, leaving the second-largest continuous block of tropical forest understudied. Nevertheless, recent work has shown that African and South-American tropical forest currently show important differences in structure (Banin et al. 2012) and species richness and composition (Slik et al. 2015b). These differences call for cross-continental empirical research in both the Amazon and the Congo basin (Corlett and Primack 2006), and in this context we can raise questions about the universality of tropical forest biogeochemistry and functioning across both continents, and subsequently their response to future global change scenarios. Additionally, due to the central role of nutrient availability that drives both net ecosystem productivity (NEP) and ecosystem carbon use efficiency (CUEe) (Fernandez-Martinez et al. 2014), the effect of climatic gradients on nutrient availability should be better understood. Indeed, recent efforts have shown that biosphere-atmosphere carbon exchange in forests is regulated by nutrient availability (Fernandez-Martinez et al. 2014) and therefore, changes in nutrient bio-availability induced by global change need to be accounted for. Canopy chemical traits are proxies that are relatively easy to assess, and from which ecosystem functioning and biogeochemistry can be inferred (Wright et al. 2004, Asner et al. 2015). Nutrient ratios and concentrations in leafs, along with specific leaf area (SLA), are traits that are known to cluster around the leaf economic spectrum, which expresses a trade-off in photosynthetic efficiency and leaf turnover. Indeed, canopy nutrients play key roles in photosynthesis, and are hence vital for carbon exchange processes at the leaf level (Evans 1989, Reich et al. 2009). Consequently, species with high SLA, N and P are associated with high photosynthesis rates (Reich et al. 1997, Wright et al. 2004, Poorter et al. 2009), but have an ‘expensive’ nutrient economy (fast leaf turnover). Previous work has shown that these traits vary systematically with landscape biogeochemistry (Asner et al. 2014, 2015) and hence the functional canopy signature of forests across gradients express the ecological response to changes in nutrient availability. Canopy chemistry has received increasingly more attention because of its inherent link to the plant strategy. Nevertheless, and as rightfully noted by Asner and Martin (2016), there are only limited surveys on canopy functional signatures in the tropics, while this information is vital for a landscape-scale understanding of tropical forest assembly. In addition to leaf traits, both leaf and soil δ15N are known integrators of the local N cycle


55

and analysis of natural abundance of stable N isotope ratios is a powerful and extensively studied proxy for N cycling in ecosystems (Högberg 1997). Previous efforts have shown that shifts towards lower δ15N values indicate a more closed N cycle with lower N availability, and vice versa (Houlton et al. 2006, Brookshire et al. 2012a, Craine et al. 2015). This shift in isotopic ratios is caused by increased rates in fractioning processes such as denitrification, where 14N is preferentially consumed, leaving the source pool enriched with 15N (Hobbie and Ouimette 2009). Hence δ15N values have been used to infer shifts in N openness across natural gradients (Vitousek et al. 1989, Martinelli et al. 1999, Menge et al. 2011), and subsequently, combining both leaf traits and δ15N values is an interesting approach to assess ecosystem responses to environmental gradients. In this study we address the existing lack of standardized cross-continental research and assessed shifts in nutrient availability and forest functional composition along two similar transects in Ecuador and Rwanda. We assessed these shifts through indicative (I) community-level functional traits and (II) nitrogen isotope ratios in topsoil and canopy. We hypothesized that (I) both these community-level traits and stable isotope signals would indicate a shift in nitrogen availability with altitude, and that (II) these shifts would be similar on both continents in terms of direction and magnitude, given a standardized research protocol and a similar adiabatic lapse rate. 6.2. Materials and Methods

Field inventories, sampling and trait analyses

We selected plots at different altitudes on the West flank of the Andes in Ecuador (ranging from 400-3200 masl) and in the Nyungwe national Park Rwanda (1600-3000 masl), in the Southern Great Rift Valley (Appendix E1 and E4 for location and overview maps). Due to reduced accessibility, the gradient in Rwanda was shorter than the South-American transect. We delineated and inventoried plots following an international standardized protocol for tropical forest inventories (RAINFOR,Malhi et al. 2002), with an adapted plot size of 40 by 40 m. In each plot, the diameter of all live stems with a diameter larger than 10 cm was measured at 1.3 m height and the trees were identified to species or genus level. Besides diameters also tree heights were measured, in order to estimate the aboveground carbon storage (AGC) using pan-tropical allometric relationships (Chave et al. 2014). The canopy of every plot was characterized by selecting the most abundant tree species, aiming at a sampling percentage of 80% of the basal area of the plots. For the selected species of all plots, we sampled mature leaves of a minimum of three individuals per species per plot using tree climbers. For most of the individuals we sampled fully sunlit leaves, but this was not always possible for the safety of the climbers, in which case we sampled partly shaded leaves under the top canopy. Previous work on elevational transects has shown that the vertical profile of leaves within a canopy has little effect on the trait values (Fisher et al. 2013). Additionally, composite samples of the topsoil (0-5 cm) were collected at five different places within each plot, and mixed per plot prior to drying. Soil and leaf samples were dried for 48 hours at 60°C. Roots were picked out of the soil samples before grinding and subsequently carbon (C), nitrogen (N) content and δ15N of plant and soil samples were analyzed using an elemental analyzer (Automated Nitrogen Carbon Analyzer; ANCA-SL, SerCon, UK), interfaced with an Isotope Ratios Mass Spectrometer (IRMS; 20-20, SerCon, UK). Leaf samples were dry-ashed at 550°C for 5.5 hours; the ash was dissolved in 2M HCl solution and subsequently filtered through a P-free filter. The aliquots where then analyzed for total


56

P by AAS method No.G-103-93 Rev.2 (Multitest MT7/MT8;Ryan et al. 2001). SLAs were calculated by dividing the leaf areas of all the sampled leafs per individual by their summed dry mass. Leaf areas were determined by either photographing leafs with on white paper with a reference scale or by drawing leaf contours and scanning the drawings. Both the scans and the pictures were processed using the ImageJ software (Schneider et al. 2012). For one abundant species of the higher altitudes on the Rwandan transect (Podocarpus latifolius (Thunb.) R.Br. ex Mirb.) we could not obtain good area estimates, so we adopted SLA figures from literature (Midgley et al. 1995).

Statistical analysis

Average leaf trait values as specific leaf area (SLA), leaf nitrogen content on mass basis (LNC), leaf phosphorus content on mass basis (LPC), δ15N, C:N and N:P ratio were calculated for every selected species, based on the sample values for the different individuals of the species. Subsequently, to calculate community-level traits and leaf δ15N per plot, we calculated a basal area weighted-average canopy value and standard deviation using the species composition and the species averages, following Asner et al. (2016). Hence:

�̅�𝑤 =∑ 𝑤𝑖. 𝑥𝑖𝑁𝑖=1

∑ 𝑥𝑖𝑁𝑖=1

with xw the weighted value for trait x, xi the mean trait value for species i and wi the basal-area based weight of that species in the specific plot. Subsequently for the weighted standard deviations (𝜎𝑤):

𝜎𝑤 =√∑ 𝑤𝑖. (𝑥𝑖 − �̅�𝑤)²𝑁𝑖=1

(𝑁 − 1)∑ 𝑤𝑖𝑁𝑖=1

𝑁

with N the number of nonzero weights. The structure of the trait datasets was assessed qualitatively using Pearson correlation statistics after log-transforming the trait data for normality. Finally, we studied the relations between the different leaf traits and elevation using mixed effects models for the different traits, with a random error structure. The plots were spatially clustered around four altitudes on both transects, hence we introduced these elevational clusters as a random effect, and treated altitude and transect as fixed effects. Models were then fitted using maximum likelihood methods in the ‘lme4’ package in R (Bates et al. 2007). P-values of the fixed effects –elevation, transect and their interaction - were determined based on the denominator degrees of freedom calculated with the Satterhwaite approximation, in the lmerTest package (Kuznetsova et al. 2014). The P-values for the interaction term, along with the Akaike Information Criterion (AIC) for models with and without this interaction term were used to decide whether or not to exclude the interaction term. For reasons of linearity we used the inverse C:N (hence rather N:C) in these analyses. Models for δ15N were assessed for each transect, using mixed effects models, with elevational cluster as a random effect. To explicitly determine divergence and convergence of plant and soil δ15N with altitude, compartment (i.e. canopy leaves or topsoil) was introduced as a fixed effect and the interaction term was left in the model. For the statistical analysis, the R-software was used (R Development Core Team 2015).


57

6.3. Results The pooled trait datasets from both transects showed a consistent and similar correlation structure (Appendix E2), with both the separate and the pooled data showing significant correlations between all traits, except SLA and N:P. The structural vegetation parameters on both transects showed important differences: for the same altitude range, we found a higher stem density, but less species on the Rwandan transect (Table 6.1). Tree height and basal area were comparable, and the carbon stocks showed high variability along both transects. Climatic conditions were similar, with a highly consistent temperature gradient (Appendix E3, Table 6.1), and similar mean annual precipitation in the concurring elevational ranges. The linear mixed effects models with altitude as fixed effect, were able to explain a significant proportion of variation in all traits. This is reflected by both the marginal (m) and conditional (c) R2

adj, respectively proxies for the variation explained by the fixed effects, and the random and fixed effects together (Nakagawa and Schielzeth 2013) (Table 6.2). The interaction term was not significant in any case, hence the trait responses to altitude were parallel on both continents. LNC, N:C, LPC and N:P significantly decreased with altitude (R2

adj,m of respectively 0.83, 0.87, 0.68 and 0.60), with the Rwanda transect showing higher overall values. SLA also decreased significantly, but with a slightly higher intercept for the Ecuadorian transect (R2

adj,m = 0.83). δ15N decreased on both continents with altitude, with a similar effect on both continents (Table 6.3). There was a significant divergence between slope and soil δ15N along the Ecuadorian transect, while Rwanda showed a significant convergence (R2

adj,m = 0.93 and 0.55 for respectively Ecuador and Rwanda).


58

Table 6.1. General characteristics, vegetation structure, climate (mean annual temperature (MAT) and mean annual precipitation (MAT), WorldClim - Fick, 2017) and soil characteristics of the elevational clusters on both transects. Number of trees and species (in the 40 by 40 m plots), basal area (BA), mean tree height (MTH) and above-ground carbon (AGC) are averages per plot ± the standard deviation on the plot-level results, based on the inventories.

Cluster Altitude (masl)

Number of trees per plot

Number of species per plot

BA (m2 ha-1)

MTH (m)

AGC (Ton C ha-1)

MAT (°C)

MAP (mm)

Soil parent material Soil classification

Ecuador

1 406 ± 10 86 ± 13 30 ± 2 25 ± 3.1 18.3 ± 1 96 ± 19 23.7 3720 Lahars Andisol

2 1068 ± 25 84 ± 38 39 ± 15 33 ± 9.4 16.9 ± 0.7 140 ± 25 20.0 3227 Lahars Andisol

3 1871 ± 79 69 ± 11 31 ± 2 33 ± 11 13.5 ± 1.3 112 ± 57 17.5 1619 Redbed volcaniclastics Andisol

4 3217 ± 21 90 ± 25 18 ± 2 49 ± 11 13 ± 1.4 161 ± 34 10.9 1241 Granitic/acid Andisol

Rwanda

1 1760 ± 66 70 ± 18 21 ± 4 34 ± 4 13.9 ± 0.7 121 ± 11 17.6 1518 Shale and Quartzite Inceptisol/Ultisol

2 2200 ± 64 71 ± 18 18 :± 3 45 ± 9 14.3 ± 0.3 179 ± 27 15.9 1628 Shale and Quartzite Inceptisol/Ultisol

3 2512 ± 37 122 ± 60 11 ± 1 31 ± 9 12 ± 1 99 ± 36 14.7 1716 Shale and Quartzite Inceptisol/Ultisol

4 2844 ± 77 109 ± 56 8 ± 2 34 ± 4 11.5 ± 0.5 89 ± 10 12.9 1835 Shale and Quartzite Inceptisol/Entisol


59

Table 6.2. Fixed effects estimates (altitude in km asl) for the different canopy-level response variables; leaf nitrogen content (LNC), inverse C:N ratio, specific leaf area (SLA), leaf phosphorus content (LPC) and N:P ratio, along with the estimated marginal (m) and conditional (c) R2

adj (sensu Nakagawa and Schielzeth (5)). The interaction term for altitude x transect was not significant in any case, and was hence not retained in any model.

Table 6.3 Fixed effects estimates (altitude in km asl) for δ15N in both canopy and topsoil (compartment) on both transects, along with the estimated marginal (m) and conditional (c) R2

adj (sensu Nakagawa and Schielzeth (5)).

Response Effect Estimate SE P-value R2adj,m R2

adj,c

LNC (%) Ecuador intercept 3.04 0.131 <0.001 0.80 0.85

Rwanda intercept 3.65 0.115 0.003 Altitude -0.59 0.000 <0.001 SLA Ecuador intercept 175.28 14.107 <0.001 0.77 0.95

Rwanda intercept 172.55 12.504 0.835 Altitude -36.24 0.007 0.002 N:C Ecuador intercept 0.07 0.003 <0.001 0.83 0.91

Rwanda intercept 0.07 0.003 0.033 Altitude -0.01 0.000 <0.001 LPC (%) Ecuador intercept 0.16 0.012 <0.001 0.60 0.88

Rwanda intercept 0.17 0.011 0.247 Altitude -0.02 0.000 0.009 N:P Ecuador intercept 20.66 1.055 <0.001 0.54 0.74

Rwanda intercept 24.06 0.926 0.014 Altitude -1.82 0.001 0.018

Response Effect Estimat

e SE

P-value

R2adj,m R2

adj,c

δ15N Ecuador(‰)

Canopy intercept 3.96 0.40

6 <0.001

0.93 0.94

Soil intercept

5.19 0.440

0.009

Altitude

-2.59 0.000

<0.001

Altitude x compartment

1.53 0.000

<0.001

δ15N Rwanda (‰)

Canopy intercept 6.83 3.19

3 0.211

0.55 0.82

Soil intercept

12.37 1.749

0.004

Altitude

-2.02 0.001

0.310

Altitude x compartment

-1.37 0.001

0.074


60

Figure 6.1. Trends in community-level functional traits and leaf (full line, closed circles) and topsoil (dashed line, open circles) δ15N of the elevation transects in Ecuador (red) and Rwanda (blue). Leaf nitrogen content (LNC), leaf phosphorus content (LPC), specific leaf area (SLA), and leaf N:C, P:C and N:P ratio decrease with increasing altitude on both transects. Both transect showed decreasing values of δ15N, providing additional evidence for a more closed N-cycle with increasing altitude. Lines represent the fixed altitude effects in the respective statistical models for both Ecuador (red, 400-3200 masl) and Rwanda (blue, 1600-3000 masl).


61

6.4. Discussion

Elevational transects are viable setups to assess long-term ecosystem responses to environmental gradients. We assessed canopy chemistry, functional composition and δ15N signals along one elevational gradient in South-America and one in Central Africa. The measured traits are indicative proxies for the underlying biogeochemistry of the forest ecosystems. The shifts of these proxies along both transects were parallel in both setups. They indicated a lowering N availability with increasing altitude, with a subsequent parallel shift in functional forest composition on both continents. The vegetation structure was varying differently along both transects. The high variability in the stem number, basal area and carbon stocks is potentially caused by the relatively small plot size. Other research efforts, targeting these variables in specific, use plot sizes of 1 hectare, as set forward by the original RAINFOR protocol, in tropical forests worldwide (Phillips et al. 2009). As such, the differing carbon stocks probably do not integrate important stochastic events (e.g. tree fall) from the forest along both slopes. However, interestingly enough we found a lower average carbon stocks and higher number of trees in the upper two Rwandan clusters in comparison to the Ecuadorian forests. This contrasts to what has been reported from large scale forest monitoring networks across the lowland forests of Amazon and the Congo basin (Lewis et al. 2013). More research in larger plots, including dynamics and productivity should validate if this is a consistent observation in highland forest on both continents. On the other hand, the lower species number on the African transect fits well within the recent findings of a pantropical study, reporting a lower tree species diversity in the African tropical forest (Slik et al. 2015b). Different environmental variables are influenced by altitudinal changes, i.e. atmospheric pressure, temperature, cloudiness, moisture, etc. (Körner 2007). Accordingly, elevation is an indirect proxy for the related changes in these variables. In this view, the air temperature decrease with elevation was highly similar on both transects, which means that we can validly assess similar temperature-driven responses of both forest functional composition and the underlying nutrient dynamics. The high collinearity in the trait datasets corresponds well to known trade-offs described as the “leaf economics spectrum” (LES); basically a leaf-level trade-off between leaf construction cost, i.e. low specific leaf area (SLA), leaf nitrogen content (LNC) and leaf phosphorus content (LPC); and photosynthetic efficiency, i.e. high SLA, LNC and LPC (Wright et al. 2004). LNC, LPC and SLA showed a highly significant decrease with altitude (Figure 6.1 and Table 6.2), indicating a functional shift towards more nutrient conservative species communities at higher altitudes on both transects. Indeed, leaves at lower altitudes with high LNC, LPC and SLA and hence a more efficient photosynthetic apparatus and rapid turnover, are replaced by leaves with low LNC, LPC and SLA values at higher altitudes. We’ve added previous published work of South America and South-East Asia, with similar temperature gradients by Asner et al, Kitayama and Aiba and Van de Weg et al. to our transects (Kitayama and Aiba 2002, Van de Weg et al. 2009, Asner et al. 2016; Appendix E3) to assess the consistency of our observed trends. We added the limited amount of studies where community-weighted means were reported along one ‘single mountain range system’, hence neglecting a recent and relevant contribution from Asner et al. (Asner and Martin, 2016). Our comparison showed that the decreasing trend in LNC was consistent with the other studies from South-America (Asner et al., 2016b; Van de


62

Weg et al., 2009), but not with South-East Asia where no significant trend was found (Kitayama and Aiba, 2002). However, leaf mass area (LMA; the inverse of SLA) of all studies showed a similar, increasing trend with elevation. LPC shows a strong and significant trend along both transects in this study, while the other studies report no significant trend. This is consistent with the meta-analysis presented by Tanner et al., which shows consistent negative LNC trends on ‘same mountain’ studies and inconsistent LPC trends (Tanner et al. 1998). A recent effort on a larger scale in Peru has shown that LES trade-off between LNC-LPC or SLA-LPC is indeed decoupled by climatic and geophysical filters, while the leaf SLA-LNC trade-off is more robust . Regarding the studies we included for comparison (Asner et al. 2016) (Appendix E3), only Van de Weg et al. assessed N:P ratio. Although no significant trend was found, they reported that N:P ratio was lowest in the highest sites (Van de Weg et al., 2009). Additionally, decreasing N:P ratios have also been reported on other transects on the Andes (Soethe et al. 2008, Fisher et al. 2013), and recently in Peru using airborne imaging spectroscopy (Asner et al., 2016). In addition to the above community-level functional traits, the decreasing δ15N values on both continents (Figure 6.1) are another strong indication of the decreasing N availability in the upper forests. Along the transects, both topsoil and canopy leaves showed decreasing δ15N values with increasing altitude (Figure 6.1), indicating a more closed N cycle with lower N availability at the higher altitudes of both transects. It has been shown that lowland tropical rainforests exhibit high values of δ15N mainly caused by the high gaseous nitrogen losses via denitrification, a strongly fractionating process (Houlton et al. 2006). The decreasing trends with altitude are interesting and seem to support the existing paradigm that tropical forests shift from P to N limitation in transition from lowland to montane tropical forest (Townsend et al. 2008). This is also reflected in the stoichiometric shifts, as canopy N:P is decreasing with increasing elevation (Figure 6.1). Hence plants incorporate relatively less N compared to P in canopies at higher altitudes. The higher soil δ15N values along the lower part of the Rwanda transect suggests a more open N-cycle compared to the lower part of the Ecuadorian transect. This corroborates with a recent finding of very high N losses at 1900 masl at the Rwanda site (Rütting et al. 2014), and the observation of high retention potential of bio-available N in Chilean Andisols (Huygens et al. 2008). Further research is needed to explain the notable divergence in soil and foliage δ15N along the Ecuadorian transect, mainly driven by the highest elevational cluster. As previously reported this can be due to different degrees of dependence upon ectomycorrhizal fungi (EcM) (Hobbie et al. 2005), different mycorrhizal association types (Craine et al. 2009) or shifts in the uptake of different forms of nitrogen (Kahmen et al. 2008, Averill and Finzi 2011). EcM-associated plant species are expected to show more depleted isotopic ratios, due to isotope fractionation during N transfer to the host plant. This effect is obscured in lowland N rich tropical forests and might just not be detectable at lower altitudes, but might become apparent in N poorer environments such as the higher altitude forests (Mayor et al. 2014). Secondly, a study from a temperate elevational transect has shown that plants increasingly switch to organic N sources with decreasing temperature, without fractionation upon N transfer from EcM to plants (Averill and Finzi, 2011). Resulting from that, they found a convergence rather than divergence of δ15N soil – canopy along altitude, because plants draw N increasingly from a source pool close to the bulk isotopic signature. We have no data on EcM colonization or δ15N of sporocarps in the study plots, so we are not able to disentangle both mechanisms. However, by characterizing both community functional traits and


63

canopy and soil δ15N, the data of these transects is consistent with a decreasing availability of soil N as elevation increases. We suggest the reduced N availability to be caused by an indirect temperature effect on the N-cycle, consistent with observations from a direct fertilization experiment (Fisher et al., 2013). Lower temperatures slow down depolymerization and N mineralization processes, hence also N bio-availability, thereby invoking changes in the functional plant communities along the transects (Marrs et al. 1988, Coûteaux et al. 2002). Future global change will most likely distort N availability both directly via increased reactive N deposition (Galloway et al. 2008, Hietz et al. 2011) and indirectly via a temperature effect on N mineralization in forest soils. This raises questions on the future of plant species within the already threatened montane tropical forest biome, where higher N availability and temperature increase might distort the existing ecological niches and in turn also increase N losses. Further research should therefore focus on process-based knowledge of N and P cycle dynamics along such transects to further assess if the availability is actually limiting the ecosystems. These observations also have repercussions for carbon fluxes: since nutrient availability exerts a stronger control on NEP than on gross primary production (GPP) (Fernandez-Martinez et al., 2014), it is likely that the CUEe will be lower at higher altitudes. It has been hypothesized that this decrease in CUEe is due to an increased investment of photosynthates to non-biomass components, such as root symbionts for nutrient mining and root exudates, in expense of net primary production (NPP) (Vicca et al. 2012). However, recent empirical evidence has shown for one transect in the Andes, that a decrease in GPP with increasing altitude is not accompanied by a trend in CUE (Malhi et al. 2016). More work on carbon budgets along elevational transects is needed to fully understand the role of N and P availability and its interaction with climate gradients for the tropical forest carbon cycle. 6.5. Conclusions Altogether, this study evidences parallel functional shifts with a similar direction and magnitude along two comparable elevation gradients, in tropical forests on two different continents. The data suggests, in two different ways, that this shift is caused by temperature-driven response of nutrient availability. With the first data on an elevational transect in Central Africa, this work adds to the existing set of elevational transects in the tropics. However, more transects are needed, especially in Africa, to validate a universal response of tropical forests to environmental change. Furthermore, work on process-based nutrient dynamics is important to unravel the importance of different global change factors for both forest basins. Acknowledgments This research has been supported by the Belgian Development Cooperation through VLIR-UOS. VLIR-UOS supports partnerships between universities and university colleges in Flanders (Belgium) and the South looking for innovative responses to global and local challenges. Visit www.vliruos.be for more information. We also thank BOS+ Tropen, Mindo Cloud Forest Foundation and the Rwanda Development Board for the logistical support; Fidel Nyirimanzi and Nicanor Mejía for their botanical expertise. Supporting Information Additional Supporting Information may be found in Appendix E. Appendix E1 Overview map


64

Appendix E2 Structure and correlations of the trait data Appendix E3 Trends in community-level traits of previously reported studies Appendix E4 Coordinates, elevation and cluster membership of the different plots on both transects Appendix E5 Summary of the plot-level characteristics Appendix E6 Fixed effects estimates for the different canopy-level response variables for the full model including interaction term

65

III Nitrogen budgets of contrasting forests in the

Congo Basin

Chapter 7: The nitrogen cycle of African tropical forests: a shift in paradigms

66

Chapter 7: The nitrogen cycle of African tropical forests: a shift

in paradigms After:

Bauters, M., Verbeeck, H., Ntaboba, L. C., Barthel, M., Six, J., Boyemba, F., Bodé, S., Makelele, I., Rütting, T. and Boeckx, P.: Nitrogen cycle in African tropical forests: a shift in paradigms (due for submission)

Abstract

The observation of high losses of bio-available nitrogen (N) and N richness in tropical forest is paradoxical with apparent lack of an N input. Hence the current paradigm asserts that biological nitrogen fixation (BNF) and spatial heterogeneity are a major N input for tropical forests. However, the ensemble of data that is at the basis of the paradigm is confounded by several issues, which challenge the generality of the BNF paradigm: full N budgets are rare and geographically biased, organic N compounds are often neglected and soil gross N cycling is still not well characterized in different forest types. We conducted intensive N flux monitoring in tropical forest types in the Congo Basin in contrasting biotic (mycorrhizal association) and abiotic (lowland – highland) environments. In four different setups we monitored N deposition, throughfall, litterfall, leaching and export during one hydrological year and completed this empirical N budget with gross soil N dynamics using 15N-tracing and numerical modeling. We found that lowland forests showed a very tight soil N cycle, with mineralization to immobilization (M/I) rates close to 1 and low nitrification to mineralization rates (N/M), which was in line with the observation of dissolved organic nitrogen (DON) dominated N losses for the most abundant forest type. The ectomycorrhizal (EcM) dominated forest in the lowland, however, showed NO3

- dominated N losses but the same M/I and N/M ratios, suggesting an active role of the EcM community in the N cycling of those forests. Contrastingly, the montane forests showed a more open soil N cycle with higher nitrification, which indeed led to NO3

- dominated N losses in the montane forests. Our observations show that N losses and N budgets can dramatically change by biotic factors, with the variability within location being as high as between locations (lowland - highland). Contrary to Neotropical forests, the budgets of central African forests were imbalanced by a higher input than output. This challenges the role of BNF in mature central African tropical forests, and suggests that unmeasured outputs such as N2 emission play a major role in the N balance.

Chapter 7 Cover. Isaac Makelele, the future of Congo’s biogeochemical work, performing 15N labeling in Yoko (left). In the bamboo forest on the slopes of Mt. Kahuzi with Simon Baumgartner, Matti Barthel (N2O fluxes) and the local ecogardes (right). The coloration on the map shows wet tropical forest (darkgreen), moist deciduous tropical forest (light green) and montane tropical forest (red).


67

7.1. Introduction

Tropical forests dominate the terrestrial carbon (C) cycle, accounting for about one-third of global terrestrial gross primary productivity (Beer et al. 2010) and intact tropical forests have been reported to sequester the equivalent of about half the total terrestrial C sink (Pan et al. 2011). Recent work has put forward an increasing role for plant nutrients in limiting the productivity response of ecosystems, and both data-analyses and modeling efforts have shown that CO2 uptake by terrestrial ecosystems strongly depends on nutrient availability (Bonan 2008a, Peñuelas et al. 2013, Fernandez-Martinez et al. 2014, Wieder et al. 2015). Indeed, this has urged land surface modelers to implement nutrient cycles in this generation’s models (Zaehle and Friend 2010, Wang et al. 2010, Goll et al. 2012, 2017) and stresses the need for a profound and mechanistic understanding of nutrient cycles. Nitrogen (N) and phosphorus (P) are of particular interest, being the two most important macronutrients to sustain plant life. However, the cycling of nitrogen (N), being a highly mobile element, in different forest types is still not fully understood in tropical forests.

The nitrogen paradox of tropical forests

Tropical forests are in general regarded to be N rich ecosystems, where N is cycled in excess. Much like discussed by Taylor et al. (2015), this is supported by multiple lines of evidence such as the high N:P ratio in canopy leaves (McGroddy et al. 2004, Fyllas et al. 2009), high dissolved inorganic N (DIN) losses from streams (Brookshire et al. 2012a), the high natural abundance of 15N in tropical forest soils (Bai and Houlton 2009) and thoughtful extensive reviews on N cycling in tropical forest (Vitousek 1984, Vitousek and Sanford 1986, Bruijnzeel 1991). However, the generally high losses of bio-available N from forested catchments suggest the existence of an N input sustaining these losses (Hedin et al. 2009). Although symbiotic biological nitrogen fixation (BNF) was traditionally put forward as major source of N, research efforts have now shown that this process is down-regulated in N-rich environments, and hence also in mature tropical forests (Barron et al. 2011, Chapter 2). This paradoxical sustained BNF has subsequently been explained by spatial heterogeneity where N poor niches are decoupled from the N richness, and where symbiotic or asymbiotic BNF can be maintained (Reed et al. 2011, Menge, Duncan and Levin 2017). Hence, tropical forests are regarded as a ‘leaky nitrostat’ (Hedin et al. 2009), with an open N cycle, where inputs are regulated following the losses, and a tightly closed P cycle. However, the data at the very basis of the nitrogen paradox; i.e. the observations of sustained N loss in tropical rainforests, are actually few and of varying quality. Firstly, some of the export budgets have reported DIN losses only (Appendix F1), leaving both dissolved organic and particulate organic N (DON and PON) unquantified, although these have proven to be of extreme importance in the overall N balance of some tropical forest ecosystems (Taylor et al. 2015) and some temperate forest ecosystems (Perakis and Hedin 2002). Secondly, research results from both montane and lowland tropical forest catchments have been generalized. In parallel to the vision of N-rich and P-poor lowland rainforests, ecologists have a longstanding paradigm of tropical montane forests being limited by N rather than P (Tanner et al. 1998, Santiago 2015). Hence lumping all ‘tropical forest’ catchments should be cautioned. A final important assumption of the N paradox is that there is low N deposition in most remote sites. Again here, recent studies have shown that DON deposition can contribute largely to total N deposition, although it has been ignored in most studies (Chapter 3, Mace et al. 2003, Cape et al. 2011, Cornell 2011). The few studies that include DON deposition


68

for N balances usually come up with N inputs that could outbalance reported N export ranges (Appendix F2), while in modelling efforts, most often only ‘reactive’, inorganic, N deposition is analyzed or simulated. Finally, the combination of empirical field measurements of N input and N output, to make up a full balance, are extremely rare for tropical forests (Bruijnzeel 1991). This can bias our idea of the N cycle, since N cycling depends on geographic location (deposition; Dentener et al. 2006), forest type (Staelens et al. 2011), climate (Weintraub et al. 2016) and topography (Weintraub et al. 2014). The availability of studies that map the complete N cycle in tropical forests is also geographically biased, being located in Costa Rica (Taylor et al. 2015), Hawaii (Vitousek 1984, Hedin et al. 2003) and some other sites in South America and South-East Asia (Bruijnzeel 1991), with very few data on African tropical forest (Galy-Lacaux et al. 2014). Rather than contradicting the current paradigm of N cycling in tropical forests, we advocate that more integrative observations in contrasting conditions are necessary. This would both benefit the fundamental understanding as well as answer the need for parameterization of land surface models.

Tropical forest nitrogen cycling in contrasting abiotic environments

Soil types and/or climates affect N cycling of natural ecosystems and consequently much can be inferred from the variation in N cycling across contrasting environments (Vitousek 1984, Vitousek and Matson 1988, Vitousek et al. 1995). Simple proxies integrating the local N cycle have been widely assessed in tropical forest environments, such as litterfall (Vitousek 1984), δ15N (Craine et al. 2009, Mayor et al. 2014), canopy stoichiometry (Vitousek et al. 1995, Asner et al. 2015), N losses (Hedin et al. 2003, Brookshire et al. 2012b, Gücker et al. 2016) and net soil transformations (Vitousek and Matson 1988). Although most of these proxies are considered integrative, they only offer limited insight in the mechanisms behind altered N cycle patterns. For example, net soil N rates usually are not correlated with the gross soil N rates, and hence only offer a limited view on the soil N cycle (Davidson et al. 1992). Additionally, gross N transformation rates need to be assessed in situ in intact soil cores, since lab-based or disturbed soil assessments render non-representative rates (Booth et al. 2006, Arnold et al. 2008, Gütlein et al. 2016). However, full N cycle assessments, combining patterns in N inputs, N losses and gross soil N rates are very labor intensive and logistically challenging in remote places, and therefor very rare in tropical forests (see e.g. Gerschlauer et al. 2016 for a review of in situ gross N studies of tropical forests). Nevertheless, the contradictions in existing observations of N loss patterns across tropical forest catchments proof that there is a lot that could be learned from more ‘holistic’ and integrative studies in contrasting environments (Brookshire et al. 2012b, Taylor et al. 2015, Gücker et al. 2016).

Tropical forest nitrogen cycling in contrasting biotic environments

A vital component and actor in the N cycle are the organisms that take part in it. Due to the inherent metabolic need of all organisms ranging from microbes over soil fauna to plants to build in both C and N, organismal growth is inherently linked to N cycle rates. Different studies have pointed to the impacts and effects of plant species on local (Reed et al. 2008, Menge, Duncan and Levin 2017) or landscape-scale biogeochemistry (Knops et al. 2002). However, few studies have directly compared a ‘full N cycle’ assessment across forest types, where geography, and thereby climate and geology, were kept constant. The intimate mixing of monodominant forests


69

consisting of Gilbertiodendron dewevrei (De Wild.) J. Léonard with the ‘mixed’ lowland forest, or monodominant bamboo forests with the mixed montane forest, on the African continent offers a great setup to assess species’ influence on N cycles. On both locations, both forest types are climax ecosystems but represent a distinct set of functional traits (Peh et al. 2011a). These traits are only proxies for the underlying life-history of plants (Díaz et al. 2015), hence the real mechanisms of monodominance are to be found in the underlying processes that are expressions of these traits. For Gilbertiodendron forest for example, authors have suggested that the association with ectomycorrhizal fungi (EcM) is underlying the phenomenon (Peh et al. 2011a, 2011b, Corrales et al. 2016, Kearsley et al. 2017). One study has reported clear changes in the local N cycle from a monodominant forest in South-America and concluded that monodominant forest systems are promising model systems to explore the organization of the tropical N cycle, and the consequences of ecological trait assembly on ecosystem functioning in general (Brookshire and Thomas 2013). Consequently, a more elaborate study aiming at a full N budget in both mixed and monodominant forests would further unravel both 1) the potential variability of N cycling due to biotic drivers and 2) the underlying question on how monodominance can establish in a highly diverse tropical forest.

Aim and hypotheses

With this study, we want to gain insight in the nitrogen cycle of tropical forests by making full N budgets for different forest types in a poorly documented region. We aim to elucidate three main questions: First, how does the montane forest N cycle compare to lowland forest N cycle, i.e. how does environment affect N cycling in tropical forests? Secondly, how does a different forest type within each geographical location affect the N cycle, i.e. what is the variability that can be expected from a change in biotic forest composition? And finally, how does the nitrogen cycle of tropical forests in the Congo basin compare to the better-documented South-American and South-East Asian tropical forests? To answer these questions, we quantified components of the N cycle in four contrasting forest types in the remote forest of the Congo basin, in the Democratic Republic of Congo. We monitored forest N fluxes fortnightly in triplicate in lowland mixed forest, lowland monodominant forest, montane mixed forest and montane monodominant forest during one hydrological year and complemented these measurements with 15N tracing and quantification of gross soil N dynamics.

7.2. Material and Methods

Study sites

The study was carried out in lowland and montane pristine forests in the Congo Basin (DR Congo). At both locations we assessed two forest types with three repetitions of the experimental setup per forest type (totaling n=12). The lowland sites are situated in the tropical forest near Yoko, roughly 30 km south of Kisangani, Tshopo in DR Congo (N00°17’; E25°18’), with mean annual rainfall of 1800 mm and average temperature of 24.2°C. Vegetation at the lowland location is classified as semi-deciduous rainforest, and the climate falls within the Af-type (tropical rainforest climate), following the Köppen-Geiger classification. Soils in the region are typical deeply weathered and nutrient-poor Ferralsols (Van Ranst et al. 2010), with very limited elevational differences and gentle slopes. The site has two dominant forest types being lowland mixed forest (LMF) and lowland monodominant forest (LMoF), where >60% of the basal area consists of one species Gilbertiodendron dewevrei (De


70

Wild.) J. Léonard. The montane forest is situated in the Kahuzi-Biéga national park, roughly 30 km north-west of Bukavu, South-Kivu in DR Congo (S02°18’; E28°43’). The national park is part of the Albertine Rift region, with an altitude ranging from 650 to 3320 m asl. Annual rainfall is between 1500 and 2000 mm, with a mean annual temperature of 20˚C. Most of the accessible part of the park is located around 2200 m asl, where two main forest types can be found: montane mixed forest (MMF) and monodominant bamboo forest (MBF). The latter is characterized by the dominance of the bamboo Yushania alpine (K.Schum.) W.C.Lin. We have established three study plots per forest type, where throughfall and soil solution where sampled. This resulted over the two locations (lowland and highland) in four sets of three study plots in LMF, LMoF, MMF and MBF. On both locations, the mixed forest type represents the dominant vegetation type, while the monodominant and bamboo forest are far less abundant.

Sampling and analysis

Throughfall and bulk precipitation was collected fortnightly using polyethylene (PE) funnels with a diameter of 15 cm supported by a wooden pole of 1.5 m height on which a PE tube was attached and draining into 5 L PE container. A nylon mesh was placed in the neck of the funnel to avoid contamination by large particles. The container was buried in the soil and covered by leaves to avoid the growth of algae and to keep the samples cool. In every experimental setup, we installed eight throughfall collectors as two rows of four collectors, with approximately 8 m distance between all collectors. The soil solution was sampled per study plot by 4 suction-lysimeters at 20 cm depth, 4 lysimeters at 40 cm depth and 3 lysimeters at 80 cm depth. Suction cup lysimeters consisted of a PVC tube fitted with a porous ceramic cup and connected to a buried opaque 2-L glass bottle by a PE tube. A pressure of -500 hPa was applied on each sampling occasion, using a portable vacuum pump (PRENART Equipment, Denmark). On every sampling occasion (fortnightly), the water volume in each collector was measured in the field, and recipients, funnels and mesh were replaced, rinsed and/or washed with distilled water. A volume-weighted composite sample of the devices per plot was made. All samples were stored in a freezer immediately and sent in batch to Belgium for chemical analysis. To assess the chemical stability of samples left out in the field during maximum two weeks, we selected three sites where four out of eight throughfall collectors were sampled daily, and the other four after two weeks. The volume-weighted fortnightly composite samples were first filtered using a nylon membrane filter of 0.45 µm before freezing. NH4

+ was determined colorimetrically by the salycilate–nitroprusside method (Mulvaney 1996) on an auto-analyzer (AA3, Bran and Luebbe, Germany). Nitrate was determined colorimetrically using the same auto-analyzer in form of NO2 after reduction of NO3

- in a Cd–Cu column followed by the reaction of the NO2

- with N-1-napthylethylenediamine to produce a chromophore. Additionally, the total dissolved nitrogen (TDN) in the water sample was determined by adding 1:1 oxidizing solution of NaOH, H3BO3 and K2S2O8, and putting it 1 hour in an autoclave at 121°C, in order to convert NH4

+ and dissolved organic N (DON) into NO3

- (Lachouani et al. 2010). Exchangeable cations (K, Ca, Mg, and Na) were determined by atomic absorption spectrophotometry (Eppendorf, Netheler & Hinz GmbH Hamburg, Germany). When the concentration was below the detection limit we assumed the concentration to be half of the detection limit. Chloride (Cl-) and phosphate (PO4

3-) was measured in the samples by Ion Chromotography (Dionex ICS90, Thermo-Scientific, Synnyvale, CA, USA).


71

Additionally, stream water was sampled in a nearby stream at both locations (lowland and mountains), where a V-notch (90°) weir was installed to survey the river water flux and stream water composition. The flow rate was estimated using a bucket and a stopwatch at every sampling occasion. Additionally, a water level height data logger (WT-HR 1500, Trutrack, New Zealand) was installed approximately 2 meter upstream of the V-notch, logging the water level every two hours. Overall, the data presented here comprises sampling from October 2015 to October 2016 in the lowland site and from December 2015 to December 2016 in the montane site. The catchment area corresponding to the drainage at the outlet point was determined using SRTM data (reference and resolution here), and a D8 flow accumulation algorithm using Whitebox and QGis. Subsequently, relations were fitted for the logger height and the stream flow at the sampling dates, and the accumulated drainage flow was calculated using the logged water heights. Litterfall traps were set up parallel to the throughfall collectors, in the same setup scheme; i.e. two rows of 4 litterfall traps at an approximate 8 m distance between each other. The traps were sampled fortnightly and the collected sampled were dried immediately after sampling. Branches with diameter > 2 cm were discarded, since we were mainly interested in fine litterfall. After drying for 24 hours at 70°C, the samples were transported to Belgium for analysis. For the lowland location all samples were ground for homogenization and subsequently analyzed on an elemental analyzer (ANCA-SL, SerCon, UK), coupled to an IRMS (20-20, SerCon UK). Total N and C turnover were determined by multiplying the N and C content of every sample with the subsequent dry weight of the total sample. For the montane forest types, the samples were pooled per setup and per month, a subset of 3 sampling dates per study site (n=6) was analyzed, and the average N content was multiplied with the study site’s respective litterfall weight.

Data processing and analysis

We used plot-averaged values for the volume of the bulk rainfall, throughfall and lysimeter collectors. The water flux for bulk rainfall and throughfall was calculated by dividing the average water volume by the surface area of the collector. Element deposition was consecutively calculated by multiplying the water volume with the element concentration in that volume. The leaching flux at the level of the suction cup lysimeters was calculated using the Chloride Mass Balance (CMB) method for the lowland cluster (De Schrijver et al. 2004). This method is based on the assumption of conservation of mass between the input of atmospheric chloride and the chloride flux in the subsurface (Eriksson 1969). The nutrient concentration in the collected water was used to calculate the total nutrient output per surface area of the catchment. For the montane sites, we used sodium as an inert tracer instead of the CMB method because we found very high chloride depositions, supposedly from the Nyiragongo volcano outgassing in the Goma region, roughly 100 km north of the montane site. This suggests that substantial retention mechanisms in the soil are impeding the use of CBM for these sites. Due to the lack of relation between N concentration and flow rate, we calculated catchment-scale export by multiplying the accumulated discharge with the concentrations (arithmetic mean, minimum and maximum) of N measured across all stream samples to estimate the catchment scale losses. The catchment scale losses are assumed to be integrative for the losses under the most dominant forest types, being respectively LMF and MMF in the lowland and the highland site. The overall water balance was evaluated at both locations by comparing the resulting


72

evapotranspiration per area with the simulated evapotranspiration from Global Land Evaporation Amsterdam Model (GLEAM; Miralles et al. 2011, Martens et al. 2017). Table 7.1. Summary of the field methods and setup per forest type to measure the nitrogen fluxes at both geographic locations.

Flux Setup per forest type Source

Open field rainfall 8 x 0.15 m PE funnels This study

Throughfall 3 x 8 x 0.15 m Ø PE funnels This study

Litterfall 3 x 8 x 0.65 m Ø litterfall traps This study

Leaching (20 cm depth) 3 x 4 x suction lysimeter This study



N20 emissions 3 x 3 x 0.4 m Ø PVC chambers with closeable led Matti Barthel unpublished

Soil N cycle 3 x 3 x 15N labeling experiment This study

15N tracing experiment for gross soil N dynamics

In parallel to the monitoring of the 12 plots as described above, a specific in situ 15N labelling experiment using the virtual soil core approach (Rütting et al. 2011a) was conducted in August 2016 in the three LMF plots and one LMoF plot (onset of the wet season). Unfortunately, we have no data yet from a tracing experiment in the MMF and MBF, hence we took data from Rütting et al. (2014) for one MMF site in Rwanda, were the exact same methodology was applied. Hence for the gross N rates, the within location differences inflected from biotic contrasts are only assessed at the lowland location. The labelling experiment allows an assessment of in situ gross N dynamics in an undisturbed system. Within each plot, we replicated the experiment three times. As such, per replication, we selected two rows of 4 labelling spots parallel to each other. Subsequently, we simultaneously labelled one row with a 14NH4

15NO3 and the other with 15NH4

14NO3 solution, both with 98 % 15N enrichment. Both solutions contained the same concentrations of NH4

+ and NO3-, and were added in the same

amount. This was done by using specifically designed, handmade injection devices of 19 one-mL injections and a spatially homogenous pattern, in the top 7 cm of the soil. The labelling spots were then subsequently sampled at different time steps after labelling (8h, 24h, 48h and 72h). Sampling was done by taking the ‘inner soil core’ in the center of the labeling area, to avoid border effects of the labeling. Immediately after sampling the soil samples were transferred to the field lab, and extracted by shaking 60 g of the sample with 120 ml of 1M KCl. After exactly 1 hour, the extract was filtered through MN615 filter paper (Macherey-Nagel), and the extracts were exported to Belgium for analysis. As described above, NO3

- and NH4+ concentrations

were determined colorimetrically using an Auto analyzer. Finally the 15N contents of both nitrogen species were analyzed after conversion to N2O (Hauck 1982, Stevens and Laughlin 1994), using a coupled trace gas preparation unit (ANCA-TGII, PDZ Europa, UK) and an Isotope Ratio Mass Spectrometer (IRMS) (20-20, Sercon, UK). The results of the analyses were used in the numerical 15N tracing model Ntrace, running in MatLab (Version 7.13, The MathWorks, Inc.), which has several advantages over the classic pool dilution techniques (Rütting et al. 2011a). This model uses Monte Carlo sampling techniques for parameter estimation, and as such estimates the fluxes of a prior defined set of N pools and transformations. For the latter, we assumed the organic, ammonium and nitrate pool and the major known fluxes between those, i.e. mineralization, NH4

+ immobilization, nitrification, dissimilatory reduction of nitrate to


73

ammonium (DNRA) and NO3- immobilization. Model fits were evaluated using the

Akaike Information Criterion, and the resulting shape of parameter distribution functions (Rütting et al. 2011a).

7.3. Results

Hydrology

Over the course of a year, the lowland location showed throughfall volumes of respectively 1765 ± 28 and 1993 ± 87 mm water in the LMF and LMoF, respectively, and 2131 mm of rainfall in the open field. This concurs with on average 17% canopy interception evaporation in the mixed forest sites, while only 6% in the monodominant sites. The water export from the catchment through river discharge was calculated for the period from 01/04/2015 to 23/01/2016, because of data logger failure after 23/01/2016. For this period, in the catchment of 30.1 ha, 545 mm water left the catchment at the outlet point, being 41% of the incident rainfall (open field) during that period. Extrapolating this value for the rest of the hydrological year resulted in a total evapotranspiration of 1685 mm. For the montane location, throughfall amounted to 1627 ± 127 and 1767 ± 55 mm for the MMF and MBF, respectively, with open field rainfall of 1802 mm. Catchment scale export through river discharge was monitored from 5/9/2015 to 9/9/2016, and 497 mm left the 11.5 ha catchment, which was roughly 27% of the incident rainfall (Appendix F3).

N input, output and litterfall

Wet deposition of TDN was similar at both geographic locations, being respectively 18.2 and 21.2 kg ha-1 yr-1. Canopy passage added a variable amount of N, hence total throughfall N input varied among the different systems being 53.1 ± 3.2, 37.5 ± 10, 37.7 ± 0.7 and 27.2 ± 6.7 kg ha-1 yr-1 in respectively LMF, LMoF, MMF and MBF. The calculated losses of TDN via the lysimeters in the dominating forest type at each location (LMF and MMF) and the river export were in the same order of magnitude, respectively 11.5 ± 2.8 and 7.3 kg ha-1 yr-1 for LMF and 15.5 ± 9.7 and 7.2 kg ha-1 yr-1

for MMF (Table 7.2). Phosphorus wet deposition was low at both the lowland and montane location, respectively 0.4 and 2.3 kg ha-1 yr-1, and only reached considerable levels after canopy passage in throughfall samples. For most of the leaching and export samples the dissolved P content was below detection limit (0.1 mg L-1). The compositional changes of the TDN through the ecosystems, going from throughfall to catchment export, varied across the different forest types (Table 7.3, Figure 7.1). The lowland in general showed high export of N in the form of DON, which was confirmed by the leaching data from the most abundant forest type, LMF. However, LMoF showed a distinct leaching pattern, with NO3

- being by far the most abundant form of N output. Likewise, the montane forest showed a very high output in the form of stream NO3

- (Table 7.3, Figure 7.1). In the montane forests, we found that MBF showed different leaching patterns with less pronounced NO3

- loss dominance. Moreover, NH4

+ was more prevalent in the throughfall and leaching compared to MMF. Additionally, both throughfall inputs and hydrological losses of N were lower in the bamboo soil.


74

Figure 7.1. Rainfall (a, f) and composition of nitrogen species of throughfall (b, g), soil solution at 20 (c, h) and 80 cm depth (d, i) and the export from the catchment (e, j), respectively, for the lowland mixed (LMF; a-e) and the montane mixed forest (MMF; f-j). Blue triangles are nitrate, red squares are ammonium and green circles dissolved organic N (DON). Error bars show the standard deviation on the arrhythmic mean of the three monitoring sites per forest type.

Gross N soil transformations

The 15N labeling experiment showed a limited variability within each site of LMF, but a high variability of gross rates across plots within the same forest type (Appendix F5). However, the mineralization to immobilization (M/I) ratios that were calculated were more consistent within LMF (Table 7.4). Overall, the highest average mineralization rates were found in LMoF and one LMF site (respectively 8.00 and 9.72 µg N g-1 d-1). The monodominant site showed similar M/I and N/M ratios as the LMF. Gross mineralization in MMF was low (2 µg N g-1 d-1), and the site also showed higher N/M and M/I ratios (Table 7.4).


75

Table 7.2. Calculated yearly nutrient budgets for lowland mixed forest (LMF), lowland monodominant forest (LMoF), montane mixed forest (MMF) and montane bamboo forest (MBF). All numbers are in kg ha-1 yr-1 (in kg N for ammonium, nitrate and dissolved organic nitrogen (DON) and kg phosphorus (P)). The catchment output shows the arrhythmic mean with minimum and maximum in parentheses.

Forest type NH4+ NO3

- DON P

LMF

Throughfall 12.2 ± 2.3 14.3 ± 2.0 26.6 ± 1.6 15.3 ± 3.2 Wet Deposition 2.4 2.8 13.0 0.4 Leaching at 20 cm 2.9 ± 1.9 6.6 ± 3.1 16.0 ± 9.7 0 ± 0 Leaching at 40 cm 8 .0± 7.0 6.3 ± 3.9 9.4 ± 2.8 0 ± 0 Leaching at 80 cm 2.4 ± 1.0 2.2 ± 0.7 6.9 ± 1.7 0.1 ± 0.2

LMoF

Throughfall 4.8 ± 0.1 12.0 ± 0.5 20.7 ± 2.3 10.7 ± 3.4 Wet Deposition 2.4 2.8 13.0 0.4 Leaching at 20 cm 3.0 ± 0.5 21.9 ± 3.5 12.6 ± 3.4 0.1 ± 0.1 Leaching at 40 cm 3.8 ± 4.4 18.9 ± 16.7 9.6 ± 8.7 0 ± 0 Leaching at 80 cm 2.0 ± 1.1 9.7 ± 6 3.3 ± 1.2 0 ± 0

Catchment output 0.7 (0.0-1.7) 0.4 (0.0-0.9) 6.2 (0.5-18.5) 0 (0-0.5)

MMF

Throughfall 9.4 ± 1.8 13.9 ± 2.2 14.4 ± 0.3 16.5 ± 4.6 Wet Deposition 9.6 5.8 5.8 2.3 Leaching at 20 cm 2.0 ± 1.1 19.2 ± 12.6 6.5 ± 4.2 0 ± 0 Leaching at 40 cm 1.8 ± 1.0 12.8 ± 14.7 2.7 ± 2.54 0 ± 0 Leaching at 80 cm 0.9 ± 0.5 12.7 ± 8.2 1.9 ± 1.19 0.1 ± 0.2

MBF

Throughfall 12.5 ± 4.6 5.9 ± 1.6 8.8 ± 2.2 8.6 ± 2.7 Wet Deposition 9.6 5.8 5.8 2.3 Leaching at 20 cm 1.1 ± 0.2 1.5 ± 1.4 2.7 ± 1.0 0 ± 0 Leaching at 40 cm 0.7 ± 0.2 1.3 ± 0.8 1.2 ± 0.6 0.2 ± 0.2 Leaching at 80 cm 1.9 ± 2.1 2.4 ± 2.7 0.8 ± 0.5 0.3 ± 0.6

Catchment output 1.4 (0.4-8.3) 3.8 (0.2-7.9) 2 (0.4-4.9) 1.5 (0-5.1)

7.4. Discussion

Lowland versus montane forest

We first look at how location, as a proxy for climatic and edaphic factors, changed the N cycle of the most abundant forest types per location, i.e. the mixed forest on both locations. In general, lowland tropical forest is considered to be a N rich and P limited forest type (Hedin et al. 2009), while this N richness decreases with altitude and is thought to become limited at higher altitudes. Given the high deposition rates (Chapter 3) and the downregulation of symbiotic BNF (Chapter 2), we have all the reason to accept that lowland tropical forest in the Congo basin is indeed N rich. However, for upland forest, the data is much scarcer. Recent work has shown that N availability lowers with altitude in African tropical forest, as in Neotropical forest (Chapter 6), but this does not per se relate to limitation. Moreover, both the wet deposition and catchment-scale export found in this study are very similar across both locations. However, the most abundant forest types at both the lowland and the highland site, respectively the LMF and MMF, showed a striking difference in composition of nitrogen losses.


76

Table 7.3. The fraction of the different dissolved nitrogen species at both geographic locations and catchments, with throughfall and leaching composition in both sampled forest types per location. All numbers represent the fraction of the total budget numbers per forest type.

Lowland Montane

NH4+ NO3

- DON NH4+ NO3

- DON

Mixed input 0.23 0.27 0.50

0.25 0.37 0.38

leaching 20 cm 0.11 0.26 0.63

0.07 0.69 0.23

40 cm 0.34 0.27 0.40

0.10 0.74 0.16

80 cm 0.21 0.19 0.60

0.06 0.82 0.12

Monodominant input 0.13 0.32 0.55

0.46 0.22 0.32

leaching 20 cm 0.08 0.58 0.34

0.21 0.28 0.51

40 cm 0.12 0.59 0.30

0.22 0.41 0.38

80 cm 0.13 0.65 0.22

0.37 0.47 0.16

Catchment-scale 0.09 0.05 0.85

0.19 0.53 0.28 For LMF forest, soil solution became gradually dominated by DON with leaching depth, which was confirmed by the catchment-scale export. Although traditionally, DIN has been regarded as the main hydrological N output from tropical forests (Bruijnzeel 1991, Schwendenmann and Veldkamp 2005, Brookshire et al. 2012a), some findings of DON dominated export from micro-catchments in lowland tropical forest have been reported (Neill et al. 2001, Taylor et al. 2015, Gücker et al. 2016). With a closer look to the data in literature, DON losses seem to be almost always at least as high as DIN losses in lowland tropical forest, whenever measured (Appendix F1). This is in contrast to the general paradigm that nitrification and subsequent NO3

- leaching is the primary mechanism for N export in lowland tropical forest, which are generally considered to be N rich ecosystems (Bruijnzeel 1991, Neill et al. 2001, Hedin et al. 2003, Corre et al. 2010). This observation was also supported by the gross dynamics we found in the LMF. Indeed, the absolute rates across the three LMF sites were highly variable, but the mineralization to immobilization (M/I) ratios were very consistent, leading to an almost complete immobilization of mineralized N. This suggests a very tight turnover and control of the microbial community on the soil N cycle and a low relative nitrification. Table 7.4. Gross soil N transformations (µg N g-1 d-1) for lowland mixed forest (LMF), lowland monodominant forest (LMoF) and the montane mixed forest (MMF) and mineralization to immobilization ratios (M/I) and nitrification to mineralization ratios (N/M). The rates show the mean and standard deviations over the three sites for LMF, and over the experimental repetitions for LMoF and MMF.

N fluxes LMF LMoF MMF

Gross NH4+ production Mineralization 5.62 (3.70) 8.00 (1.51) 2.00 (0.20)

DNRA 0.05 (0.03) 0.03 (0.01) 0.00

Gross NH4+ consumption Nitrification 1.47 (0.35) 1.51 (0.50) 1.50 (0.20)

NH4+ immobilization 4.27 (3.99) 6.65 (1.78) 0.50 (0.30)

Gross NO3- production Nitrification 1.47 (0.35) 1.51 (0.50) 1.50 (0.20)

Gross NO3- consumption DNRA 0.05 (0.03) 0.03 (0.01) 0.00

NO3- immobilization 1.36 (0.36) 1.57 (0.55) 0.20 (0.20)

M/I 0.99 0.97 2.86

N/M 0.26 0.19 0.75


77

Additionally, absolute mineralization rates for LMF were on the low side of reported results from pool dilution experiments (see Appendix F4 for an overview), but in the same order of magnitude of some recent work in tropical lowland forest (Silver et al. 2005, Sotta et al. 2008, Wieder et al. 2013, Allen et al. 2015). The rates and ratios are closely resembling reported rates from lowland tropical forest in Costa Rica (Wieder et al. 2013), were a very similar dominance of organic N losses was found (Taylor et al. 2015). Hence, the M/I ratio which is nearly 1 in LMF indirectly shows that, contrary to DON losses from some pristine temperate forest catchments (Hedin et al. 1995, Perakis and Hedin 2002, Brookshire et al. 2007), the DON in tropical forest does seem to be microbial processed. Contrastingly, the leaching and catchment-export showed strongly NO3

--dominated N losses at the montane location. The gross soil N rates under MMF were in general lower than the lowland, which in in accordance with previous studies (Corre et al. 2010). This manifests in the lower organic matter turnover in montane forest, as evidenced in previous studies, which were also at the basis of the idea of increasing N limitation with altitude in the tropics (Marrs et al. 1988, Vitousek and Matson 1988, Tanner et al. 1998). However, the rates exhibited a substantially higher nitrification relative to the mineralization, with lower immobilization rates, leading to a build-up of the NO3

- pool and subsequently NO3- dominated losses. Previous reports have indeed

also shown high NO3- losses from some Neotropical montane forests, which has led

to the belief that many tropical montane forests are not N-limited (Mcdowell and Asbury 1994, Brookshire et al. 2012b, Rütting et al. 2014). Altogether, the observation of a highly efficient DIN retention in our LMF and DIN leak in montane forests seems at odds with the standing paradigm of high NO3

- losses in N saturated tropical forest. One would expect the N rich lowland forest to be saturated and hence show substantial DIN losses. Furthermore, if we look at the microbial community as a mere aid to efficiently retain N in the upper soil layers, then tight N cycling would indeed translate in N richness, while open soil gross N cycles would cause leaching. In the latter case, the leaching is also an actual loss due to failure to retain N in the upper soil, while the losses in the N rich lowland are rather originating from N saturation. In addition to the microbial cycling, N addition experiments in old-growth N rich forests have shown that abiotic sorption of NO3

- plays in an important role in dampening the expected NO3-

hydrological losses (Lohse and Matson 2005). Additionally, there is an increasing awareness on abiotic processing of nitrification intermediates in conditions of low pH and high Mn content (Heil et al. 2016, Liu et al. 2017). Hence, physical and chemical abiotic parameters might play an important role in retaining NO3

- in the upper soil layer, and decouple the hydrological losses from the biological demands of the microbial or plant community. Our in situ 15N tracing method is not able to differentiate between biotic or abiotic NO3

- immobilization, hence abiotic NO3- immobilization via the ferrous

wheel hypothesis, or NO3- adsorption to variable-charge soils might also contribute to

the reported NO3- immobilization rates (Davidson et al. 2003, Lohse and Matson 2005,

Jiang et al. 2015). However, the sharp contrast with the NO3- dominated losses in

LMoF seems at odds with this, since abiotic conditions (pH, Fe-concentrations) in both forest types are similar (Cassart et al. 2016). Overall, both sites are characterized by very low rates of dissimilatory reduction of nitrate to ammonium (DNRA; Table 7.4). Although rates from other tropical forest sites range widely (Rütting et al. 2011b, Templer et al. 2008), NO3

- availability has been identified to exert a strong control on these rates (Silver et al. 2012). Others have


78

stated that soil C, in combination with humidity, are positively correlated with DNRA rates (Rütting et al. 2011b). Since the rates are low in the MMF (where soil C is higher, and NO3

- availability is high), our results seem to contradict both hypothesis, and we can only conclude that more research on DNRA as a retention mechanism is needed.

Local differences in N cycling linked to biotic differences

The intriguing monodominance of Gilbertiodendron dewevrei in the LMoF forest type has been discussed broadly during the last decade, particularly the mechanism behind its vast monodominance in Afrotropical lowland forest (Torti et al. 2001, Peh et al. 2011b, Kearsley et al. 2017). One of the hypotheses is that a specific combination of plant traits and strategy ensures a low soil nitrogen availability and hence acts as an environmental filter for other species to establish under mature stands, as also observed in monodominant forests in South America (Torti et al. 2001, Brookshire and Thomas 2013). Furthermore, Gilbertiodendron associates with EcM, which in turn are known facilitators of plant growth via organic N uptake. This is in contrast with LMF, which is predominantly associated with arbuscular mycorrhizal (AM) fungi. As such, low soil DIN levels have been reported from EcM monodominant forest stands, presumably driven by the EcM-mediated drawdown of the readily mineralizable DON pool (Corrales et al. 2016). Direct evidence is lacking, in part because the typical δ15N-depletion in EcM plants as observed in the temperate forest is absent in the N rich conditions of lowland tropical forest. Nevertheless structural equation modeling, the high degree of root colonization, and the observations of δ15N-enrichment of EcM sporocarps in the same sites suggest that EcM mediate N-acquisition in EcM forests in the lowland tropics (Tedersoo et al. 2012, Mayor et al. 2015). Building on this, the differing pattern of N losses between lowland AM and EcM forest sites indicates that N-acquisition strategy is different. The switch from DON to NO3

- dominated losses has two possible explanations: 1) the restriction of monodominant forest to EcM-mediated scavenging on DON, hence leaving the DIN pool to leach, or 2) higher net mineralization and hence less efficient immobilization under EcM dominated forests. The results from our 15N tracing experiment revealed that there are no significant differences in gross N cycling in LMoF forest versus LMF. Accordingly, M/I and N/M ratio both showed tight N cycling as in LMF, with relatively low nitrification. Consequently, our data suggest that the dominance of NO3

- in the losses under LMoF in central African forests is caused by a reduced uptake of NO3

- by plants or EcM, and we conclude that altered N cycling is not the mechanism driving the competitive advantage of these EcM communities in Afrotropical forests. Collectively, we conclude that, although the N cycle is altered under EcM communities, other nutrients than N (e.g. P) are promoting monodominance in lowland forests (Newbery et al. 1997, Chuyong et al. 2000, Kuyper 2012). As stated above, we currently have no data on gross N transformations in MBF. Overall, it seems that at least some important differences in throughfall and leaching patterns confirm that biotic differences substantially influence the N cycle in tropical forests.

Nitrogen budget for African tropical forests

Every good nitrogen budget starts with a closed water balance. We evaluated our water balance and by comparing our calculated evapotranspiration data with simulated evapotranspiration by GLEAM (Appendix F3; Miralles et al. 2011, Martens et al. 2017), revealing that the resulting evapotranspiration values from our empirical data are within the order of magnitude of the GLEAM estimates. It must be noted that stemflow was not assessed in this study due to logistic constraints, but it has been shown to be


79

only of minor importance to the water balance of tropical forests (Chuyong et al. 2004, Manfroi et al. 2004, Park and Cameron 2008). Dry deposition is challenging to assess, and the methods available render highly variable results (Hofhansl et al. 2011), hence we did not assess it directly here. However, there is strong evidence from the lowland sites, that important proportions of dry deposition are exogenous (fire-derived) N-inputs, rendering the total deposition on the Congo basin’s tropical forest substantially higher than expected, with an important organic component (Chapter 3). Likewise, a recent study has shown that BNF in the tropics has been overestimated (Sullivan et al. 2014). Indeed, we assume that symbiotic N fixation is also very low in our LMF. This is also in part confirmed by an earlier study close to the LMF study sites, where a downregulation of canopy legume nodulation has been noted along a successional gradient (Chapter 2). Additionally, N2O emissions were measured daily during two weeks in both the dry and the wet season in one of the sites at LMF, LMoF and MMF, which rendered estimates of respectively 1.69, 2.15 and 3.60 kg N ha-1 yr-1 (Matti Barthel unpublished). Altogether for LMF, this mounted to 31.8 kg ha-1 yr-1 deposition, with 9.8 kg ha-1 yr-1 losses of which 80% were hydrological, resulting in an imbalance of 21 kg ha-1 yr-1. For the other sites, not taking into account dry deposition, we found a minimum imbalance of 10 kg ha-1 yr-1. Hence for our central African forests sites, we found an apparent lack of an N output. This is in sharp contrast with the general idea of tropical forest N budgets, based on observations from mainly Neotropical forests (Hedin et al. 2009), where an apparent lack of input was noted, and hence also resolves the N paradox as postulated based on those sites. Furthermore, from our central African sites, we deduce a shift in the paradigm of a central role for BNF in mature tropical forests. There are two outputs which have proven to be important in other tropical forest catchments and which were not assessed in this study, namely 1) the particulate organic nitrogen (PON) export and 2) other gaseous N species (NO and N2). The importance of PON for budgets has been shown for geomorphically active lowland forests recently (Taylor et al. 2015). However, the soils of our lowland study sites are probably amongst the geologically oldest soils in the world, and hence we assume that PON transport is less important in our study site. Other gaseous N losses than N2O, on the other hand, are more likely to play an important role in the N budget. (Chemo) denitrification (i.e. the reduction of NO3

- to NO, N2O or N2 under anoxic conditions) has been hypothesized to be a major loss pathway for some tropical forests (Houlton et al. 2006). However, N2 is only rarely assessed in the field, because it is notoriously difficult to measure due to the high background concentrations of N2. Templer et al. (2008) likewise found very low leaching of NO3

- during their 15N tracing experiment in Puerto Rican tropical forest, and concluded that other sinks such as denitrification to N2 might decrease the leaching susceptibility. Under wet conditions, denitrification is likely to convert most of the NO3

- to N2 (Potter et al. 1996), hence N2 could be a hugely underestimated loss (Davidson et al. 2000, Holtgrieve et al. 2006). Empirical estimations of N2 production from old tropical forests were roughly four times higher than N2O production (Hedin et al. 2003), and up to five times in process-based models (Bai and Houlton 2009). Moreover, recent isotopic studies and ecosystem models have pointed out that denitrification losses in humid tropical forests make up a large fraction of total ecosystem N losses (Houlton et al. 2006, Brookshire et al. 2017). Moreover, others have shown that terrestrial denitrification fluxes are underestimated by up to 98%, and that N gaseous export exceeds NO3

- by a six fold (Fang et al. 2015). It might be that ammonium oxidation coupled to iron reduction (i.e. Feammox) or other


80

forms of chemodenitrification play an important role in these soils with high Fe-content (Yang et al. 2012, 2015) and low pH (Heil et al. 2016, Liu et al. 2017), in which case denitrification would be partly decoupled from biological activity, but further observations are needed to assess this. Altogether, developing methods to measure N2 in situ is vital to understand both the N balance of tropical forests, and to gain insight in the driving factors behind N2O:N2 partitioning during nitrification/(chemo)denitrification.

Figure 7.2. The nitrogen cycle in mixed lowland forest (left) and mixed montane forest (right) in the Congo basin with (a) bulk deposition, (b) dry deposition on canopy, (c) canopy leaching, (d) litterfall, (e) lysimeter leaching at 20, 40 and 80 cm depth, (f) the soil gross N dynamics in detail shown in the lower part of the figure, (g) N2O emissions and (h) catchment scale hydrological export. All numbers, except those in (f) show total dissolved nitrogen (TDN), and all numbers expressed kg N ha-1 yr-1. Red arrows are inputs, blue outputs and green arrows are internal fluxes.

7.5. Conclusions

Altogether, the observations from an in-situ 15N tracing study matched well with the composition of the N losses in the different forest types. We found no elevated NO3

- losses, but DON losses and tight N soil cycling under a lowland tropical forest, which is almost certainly very N rich. This implies that the microbial community in tropical forest soils play a key role in the retention of N for plants, and facilitate for rather than compete with the plant community (Schimel and Bennett 2004). Instead we found a


81

more open soil N cycle combined with high NO3- losses under a montane forest, which

is traditionally considered N limited. We showed that this is due to a higher M/I ratio, in combination with a low N/M ratio. This contrasted with the observations under a lowland EcM-dominated forest, where high NO3

- losses were observed without the latter changes in gross transformation ratios, suggesting that EcM-activity actively alters the soil N cycle, but not the microbial activity. This also shows that both biotic and abiotic drivers greatly affect the local N cycle. Contrary to studies from the Neotropics, the hydrological and N2O losses in all forest types were substantially lower than the minimum input levels. This reiterates the importance of making up full N budgets, and including organic nitrogen in the assessments. Furthermore, we hypothesize that denitrification might be more important than previously thought, but that this was unaccounted for because of difficulty to measure N2.

Acknowledgments

This research has been supported by the Belgian Development Cooperation through VLIR-UOS both through a personal scholarship of M.B. and project funding. VLIR-UOS supports partnerships between universities and university colleges in Flanders (Belgium) and the South looking for innovative responses to global and local challenges. Visit www.vliruos.be for more information. We would like to thank Luc Willems, Greet De bruyn, Katja Vannieuland, Stijn Vandevoorde, Eric Gillis and Wout Colman for help with the sample analysis. Supporting Information Additional Supporting Information may be found in Appendix F. Appendix F1 Overview of studies reporting nitrogen (N) export in tropical forest basins Appendix F2 Overview of studies reporting nitrogen (N) deposition in tropical forest basins Appendix F3 Water budgets of both geographical location and the validation with the GLEAM model Appendix F4 Yearly litterfall mass flux and nitrogen flux per forest type Appendix F5 Gross soil N transformations

83

Conclusions and future prospects

Chapter 8: General conclusions and future prospects

84


Tropical forests worldwide play a major role in global biogeochemical cycles and global ecology (Pan et al. 2011). Understanding the functioning of this biome is thus of vital importance to predict future trajectories of global change. Recent data-analysis and modeling activities have shown that nitrogen (N) and phosphorus (P) availability control ecosystem productivity and CO2-uptake responses (Peñuelas et al. 2013, Fernandez-Martinez et al. 2014, Wieder et al. 2015). In this view, land surface models, which are traditionally very carbon-oriented and photosynthesis-driven, have started implementing nutrient cycles in their models (Zaehle and Friend 2010, Wang et al. 2010, Goll et al. 2012, Smith et al. 2014, Yang et al. 2014). The initial planning for this PhD was to use an existing modeling framework with implemented nutrient cycles and to parametrize for Central Africa using empirical data. Over the course of these years, however, the empirical body of knowledge on tropical forest biogeochemistry in this region of the world has proven to be surprisingly small and getting new empirical data is surprisingly hard. The result is that four years were soon over, only by doing fieldwork and lab and data analysis. In retrospect there are three reasons why this thesis only confines to empirical work: the complexity of forest biogeochemistry in tropical forests, the prevalent knowledge gap of central African forests and the practice of doing fieldwork in central Africa. 8.1. What have we learnt? A dazzling interconnected complexity of ecology and biogeochemistry across scales

N cycling in tropical forests

The downregulation of BNF along forest succession has been shown for Neotropical forests (Barron et al. 2011, Batterman et al. 2013), and chapter 2 nicely confirms that BNF indeed plays a central role in the recuperation of the N cycle of tropical forests (Davidson et al. 2007). However, high fire-derived N deposition (Chapter 3) was only reported for one site in the Andes (Boy et al. 2008), and is assumed to be low in most of the Amazon forest sites (Hedin et al. 2009, Appendix F2). The combination of both findings from this thesis, i.e. the role of BNF in regrowth forest and high exogenous N deposition, is interesting. This questions the role that BNF plays on an ecosystem-level; does it provide a necessary input to promote forest growth (Batterman et al. 2013)? Instead, it might be an individual ecological strategy to suppress non-fixing species, and hence be decoupled from ecosystem demand, as recently suggested by Taylor et al. (2017). Surely enough, the underlying mechanisms are fairly different, as is the way BNF would be implemented in Earth system models; using the supply-demand approach or through the control of foliar N on photosynthesis (Thornton et al. 2007, Zaehle and Friend 2010, Reed et al. 2015). This is also in essence the key question that arises from chapter 6; does the shift in functional community assembly follow the nitrogen supply or the nitrogen demand? Chapter 6 reports on observations of a strong shift to nitrogen conservative species towards higher altitudes – colder environments. Recent work suggests that reduced light availability at higher altitudes, via cloud immersion effects, is at the very basis of the control of productivity in tropical forest ecosystems (Malhi et al. 2016). This means that a shift towards lower canopy N content – and N conservative communities - might be induced by a lower allocation of N to the leafs, because a high photosynthetic capacity is useless with reduced light availability. Hence advocating that a lower


85

demand is the mechanism underlying the observed shifts and not necessarily lower supply. The δ15N trends presented in chapter 6, however, provide an additional line of evidence that N supply shifts with elevation, from an open to a more conservative N cycle. But the underlying question whether or not this is limiting productivity is still vexing. Secondly, quantification of an almost complete N cycle (Chapter 7) has been reported by means of fragmented studies for sites in Costa Rica, Panama and Hawaii (Vitousek 1984, Hedin et al. 2003, Weintraub et al. 2014, Taylor et al. 2015). This has led to the idea of N richness of tropical forests: i.e. tropical forests cycle and recycle large quantities of N. Additionally, several studies report on ‘high’ outputs of bio-available N, outbalancing the inputs (Brookshire et al. 2012a, 2012b, Bai et al. 2012), which subsequently lies at the basis of ‘the N paradox’ of tropical ecosystems (Hedin et al. 2009). Hence this thesis started originally with an assessment of the so-called free-living BNF, which is believed to play a central role in resolving this paradox (Reed et al. 2007, 2011, 2013, Matson et al. 2015). Since no research chapter is to be found in this thesis reporting on those experiments, you can safely judge that the obtained results were not reliable/not reportable. In retrospect, it makes sense that no free-living fixation takes place (or at least not in high rates), since chapter 7 and chapter 3 jointly show that mature forests are subjected to inputs that outbalance the measured outputs in central African forests. This is in sharp contrast with the South-American documented sites and with the current implementation of tropical forests in Earth system models (Stocker et al. 2016). However, is the view underlying the nitrogen paradox correct? Chapter 7 rises at the very least two important points which challenge this line of thinking: 1) that the imbalance for central African forests is the inverse of what has been assumed from South-American reports, and 2) that the quantification of the N cycle in tropical forests changes severely when considering organic N compounds. So perhaps the arguments that have been formulated so far, using only bio-available N in in- and output assessments of tropical forests, are simply naïve? Judging from the high gross soil N dynamics (Figure 7.2) the bio-availability of organic N compounds might not be an issue, and hence research must go beyond NH4

+ or NO3- assessments. Include organic N in biogeochemical and ecological

contexts (Cornell et al. 2003, Jickells et al. 2013, Stocker et al. 2016)!

P cycling in tropical forests

Although tropical forests are mainly considered P-limited (Vitousek et al. 2010), only chapters 5, 6 and 7 very briefly discuss the P cycle. One of the main reasons for this is the difficulty to measure P fluxes and their role in ecosystems, as also discussed by Reed et al. (2015). These difficulties have limited the discussion of P in this thesis in three ways: Firstly, unlike N, P is mainly a rock-derived nutrient and is known to strongly adsorb to mineral surfaces or to be occluded by organic complexes in the soil (Walker and Syers 1976, Vitousek et al. 2010). Hence the availability (sensu Chapter 5) is only one measure for P availability, and not necessarily the real plant availability. Additionally, only 33P or 32P can be used to look at gross soil P fluxes by means of pool dilution, and none of both is a stable isotope. The radioactive decay of both compounds makes the application of developed lab techniques in specialized labs challenging for a tropical field site. Furthermore, it has been shown repeatedly for gross soil N fluxes that measurements on lab-based incubated samples from the field are problematic


86

due important sampling/disturbance effects (Booth et al. 2006, Arnold et al. 2008, Gütlein et al. 2016). Secondly, a surprising finding from chapter 7 showed exceptionally high P throughfall loads. Unfortunately, the direct exchange of orthophosphate of atmosphere and canopy (i.e. dry deposition) is something that is very challenging to quantify (Hofhansl et al. 2011), and with the sampling methods used in chapter 3 and chapter 7, there is no solid conclusion to be drawn. Nevertheless, it seems very likely that P-trapping by canopies (Lawrence et al. 2007, De longe et al. 2008) plays a central role for the P input of central African forests. Finally, the intraspecific variability of P in foliar concentrations is much higher than for N, and we have only a poor understanding of how leaf P concentrations controls photosynthesis (Reed et al. 2015). This in combination with the lack of a direct plant-availability metric for P, limits the conclusions that can be drawn from setups like the ones in chapter 6. Altogether, it is of paramount importance that empirical work on the P cycle continues, and that new techniques facilitate field-based research. The need is high now because of the recent implementation of the P cycle into Earth system models (Reed et al. 2015, Goll et al. 2017).

Ecology and biogeochemistry across spatial scales

The thesis crosscuts the spatial scales from tree-level effects to ecosystem-effects. As such, for chapter 4 and 5, a unique 80-year-old experimental plantation was used to assess how both aboveground and belowground carbon and nutrient cycles differ in contrasting biotic environments. Such local species effects are otherwise notoriously challenging to assess in orthogonal setups due to the time it takes for forest ecosystems to grow. The main conclusions from this setup were that both the aboveground and the belowground biogeochemistry are linked to the ecology of the species. As such, we noticed that, in the lowland tropics, nutrient conservative species are more likely to have high carbon accruals over the longer term (Chapter 4). Moreover, we found that the strategy inflected by the species in turn also affected the local soil conditions. The same nutrient-conservative species were found on more acid soils, with higher SOC contents (Chapter 5). Hence, tree species adopt an ecological strategy by adapting the biogeochemical configuration in their tissues, which in turn affects their growth. Tropical forests are assemblies of thousands of species of which the composition influences the environment on a local scale (Chapter 4, Chapter 5), but inversely of which the composition is determined by the environmental conditions on a larger scale (Chapter 6). This interaction between biotic and abiotic factors and the respective outcome for biogeochemical cycles is a key part in the understanding of these ecosystems for forest management, ecology and for understanding global biogeochemical cycles. Furthermore, it is evident that the we need new techniques such as remote sensing to capture the spatial heterogeneity and bridge the gap from tree-level to regional ecosystem studies (Townsend et al. 2008). The same complex interaction is also evidenced by the fact that nitrogen fluxes in different forest types were as different within the location as between geographic locations in some cases (Chapter 7). This amazing complexity adds up to a challenging matrix for quantitative research, for which it is highly labor-intensive to put


87

numbers next to the arrows in Figure 7.2. Using those numbers in models is a possible sequel to the empirical work, however, we cannot forget that empirical understanding is vital to build process-based models.

8.2. The thesis in an international context

Overall, the results from the legacy experimental plantation (Chapter 4 and 5) are highly relevant for policy-makers and local stakeholders involved in reforestation projects, restoration ecology and carbon offset markets. One might argue, judging from the ubiquitous Eucalypt plantations in eastern DRC, Rwanda and Uganda, that this is the most valuable outcome of this thesis. These non-native plantations are often the preferred choice for the guaranteed success for fast growth and high carbon uptake, but they are far less interesting in terms of ecological restoration of the landscape (Cossalter and Pye-smith 2003, Lamb et al. 2005). On the other hand, as evidenced by other work that was done parallel to this thesis on a reforestation project in Northern Ecuador, NGOs don’t know what native species to plant to attain acceptable results (unpublished). Hence, there is the need for a clear set of guidelines for reforestation with native species in the tropics. A freshly kicked-off collaboration with Swedish colleagues from Göthenborg University on a similar old legacy plantation in Rwanda is an interesting addition the Yangambi arboretum, and will further allow us to build toward this set of guidelines in the tropics. On the other hand, the fundamental work on N cycles (Chapter 7) is in its own way a valuable scientific contribution. It is remarkable how the current paradigms are based on only a few well-documented tropical forests sites. The sites that were setup for N flux monitoring in the DRC, if continued in the future, could one day grow out to be the African counterpart of the renowned sites in Hawaii and La Selva. Maybe most remarkable is the finding of a high fire-derived N deposition in central African forests (Chapter 3). The fact that this was both novel and so simple to measure shows the prevalence of the African knowledge gap. And obviously there is much more work to be done following the finding from this high N deposition: what are the consequences for the ecology of central African forests? Is the forest productivity affected by this chronic fertilization? How will the population growth and increased land-use changes affect these N deposition regimes? What are the feedbacks from atmospheric black carbon caused by the annual biomass burnings? Furthermore, although we found strong (anecdotic) parallels with the Amazon basin in the downregulation of the N-fixing community and the functional responses of the forest on both continents (Chapter 2 and 6), the high N deposition for the lowland forest (Chapter 3) might provide an interesting contrast between both continents. Some simulation papers also show contrasting patterns between the Amazon and Congo basin in terms of deposition (Wang et al. 2017). Hence, this cross-continental or even pantropical empirical work has to be elaborated to stress that both biomes are potentially very different from a biogeochemical and ecological point of view (Corlett and Primack 2006). 8.3. Working in Congo By this point in the thesis, it has been stated several times that the Congo basin is one of the most prominent knowledge gaps in global (terrestrial) biogeochemistry and ecology (Verbeeck et al. 2011). Judging from this and from the reaction of family, friends or co-workers upon mentioning ‘going to the DRC’, I think it is safe to state that


88

the DRC has a bad reputation. Political instability, civil wars and reduced accessibility have fed this conception of the dangerous dark Congo, and have kept away the international academic community. It must be said that working in the DRC is indeed not a walk in the park. It is in no way comparable to doing empirical research in temperate forests. Additionally, the research stations I visited in Ecuador and French Guiana, but even in Uganda and Rwanda, offer a luxury, logistic support and ease of organization that no site in the DRC can reach (to my knowledge). One needs to combine creativity, flexibility, patience and motivation for successful field campaigns. However, having spent myself a considerable amount of time at different locations in the DRC and having taken several colleagues and Belgian MSc students to the field, I can say that for the right type of people it is one of the best and most fun places to work in. The level of adventure is strongly addictive, and I have never felt unsafe in the field. Once one learns how to do things and how to get around, there is only little that can go wrong. On the contrary, with the lack of a working bureaucracy you can turn the country’s handicap into an advantage. There are a number of good research sites which are becoming increasingly easy to reach, and where people, thanks to some pioneers, are also more and more used to work with foreign researchers. There are qualitative young Congolese researchers with a lot of potential waiting in line to do research. Hence I can only hope that the country will welcome more and more researchers in the near future.

8.4 Future prospects

The results of this thesis are in essence only derived from a few core sites in the Congo Basin. Altogether, they show that both ecological and biogeochemical ecosystem-level variables vary widely over the sites, and vary with both biotic and abiotic changes. The sites of this thesis are almost the only sites in the DRC where long-term research projects on biogeochemistry are implemented. This stresses that the set of research sites must be elaborated to unexplored regions. Judging from the map in Appendix B3, for example, it seems that the finding of high fire-derived N deposition can exceed the deposition loads reported here in areas as Salonga National Park, where only very few researchers have been. Nevertheless, I think one of the major contributions of this thesis is the fortification of the network of partners in the DRC who are involved in several projects in the meantime. There is a strong basis now in several locations to continue the development of new research projects. Secondly, several of the Chapters (Chapter 2, 3 and Chapter 7) in the end suggest

that there is major unaccounted pathway for N loss of tropical forests (for at least our

region) and several high-impact publications have already suggested that

denitrification to N2 might be hugely underestimated (Houlton et al. 2006, Brookshire

et al. 2017). The partitioning between N2 and N2O is an important knowledge gap, with

N2O being one of the strongest greenhouse gasses in the atmosphere. The results of

this might thus directly influence our state-of-the-art knowledge on global change, and

future projections of global warming. Hence, it would be highly interesting to look into

the fate of this nitrogen deposition loads as reported in chapter 3, and see if N2

outgassing can account for part of this loss. The development of new techniques to

measure N2 outgassing (which is very hard because of the high background

concentration) is promising, and hopefully we will soon see this happening in the field.

Additionally, a recent paper has pointed out that roughly 75% of the total N loss in a

lowland forest (of a geomorphic active site though) was attributed to particulate organic


89

N export (Taylor et al. 2015). During the fieldwork for chapter 7, the V-weir had to be

cleaned out biweekly in the lowland site to get rid of all the sedimentation. Hence, this

might be a severely underestimated loss of N on the ecosystem-scale. Furthermore,

it is likely that particulate input and output, which are almost never taken into account,

have a strong effect on the P budget of central-African forests. As such, NASA recently

conducted a multiyear assessment on cloud-aerosol lidar and infrared satellite

observations to look at transatlantic transport of Sahara dust (Yu et al. 2015). They

concluded that Sahara dust deposition could be a major P input for the Amazon basin’s

forest. This makes one wonder what the forests of the Congo basin are subjected to

in terms of P deposition. Hence, both the biomass burning activities on the continent

and the potential dust transport, and their respective links with P deposition on the

forests is something that definitely cries out for attention.

Thirdly, related to chapter 6, many more elevational gradients must be set up in Africa. The past has proven that these are extraordinary natural gradients, of which much can be inferred from an ecological perspective (Malhi et al. 2010, 2016, Asner and Martin 2016, Asner et al. 2016, Mayor et al. 2017). However, the transect in Nyungwe (chapter 6) is one of the very first transects which has been assessed intensively in Africa. The recent implementation of a new transect on Mt. Rwenzuri in Uganda, with far more intensive monitoring on biogeochemistry and even a soil transplant experiment is a promising addition. No doubt this will add substantially to the state-of-the-art. This might also help to elucidate whether the observed patterns in N cycling and canopy traits are either supply (hence nutrient supply- induced shift) or demand (due to lower photosynthetic activity and cloud cover; Malhi et al. 2016), which is still a vexing question in this research area. Finally, more work needs to be done on the soil N cycle. The soil cycling rates that were measured for this thesis are one order of magnitude larger than the other recorded ecosystem fluxes (Figure 7.2), which stresses the importance of understanding this cycle. A limited meta-analyses reported that there are only a limited amount of studies using the in situ and intact soil core approach from the tropics (only 9, excluding this study; Appendix F4). The variability of soil N rates in chapter 7 is amazingly high, and to my knowledge there is almost no mechanistic understanding of the drivers of this variability. Additionally, new methods that add depolymerization to the 15N tracing are a promising tool to elaborate the existing set of tools (Wanek et al. 2010), especially since depolymerization is hypothesized to be the rate-limiting step of N cycling in a lot of ecosystems (Schimel and Bennett 2004).

Afterword

90

Afterword

The world population grows, technology advances at high pace and there is no global policy in a world with an increasingly globalized economic system. The consequence is that living standards are increasing for part of the population, while others remain poor. Furthermore, there seems to be an inverse relation between monetary and natural capital on a global scale. This contrast could not be more striking than when travelling from Ghent to a small town in the Democratic Republic of Congo called Djolu. Much of the technological advances have come at a severe cost for natural capital. Moreover, technology advances at such a rate we can never fully understand the consequences of policy choices/new technologies to the fullest. As the pressure on the remaining natural capital continues to build up, the need for an improved understanding of what’s left grows. The complexity we’re dealing with in natural systems is far more overwhelming than anything human-made. The current knowledge on climate and global change tells us that; ice cap melting leads to a feedback warming effect via reduced albedo, permafrost thawing drives us towards tipping points of atmospheric CO2, El-Nino Southern Oscillation triggering causes dieback in the Amazon forest and hence feeds back atmospheric CO2 levels… Therefore, let there be time and money for fundamental science as a fundament for the future. Non-researchers often question the use of fundamental research: how is it helping us moving forward? But is the world not seeing that growing increasingly faster on a finite world is problematic? Is it thus not wiser to invest in an improved understanding of the pressurized system all of us are dependent on? Thus, we need a global effort to increase knowledge on the processes that are undergoing shifts due to anthropogenic activities. Universities are the first-class knowledge centers and partnerships across boundaries are vital to map and understand the regional effects of a global problem. As the title states, I tried to disentangle the biogeochemical cycles of central African forests. Why? African forests are a blank spot in almost every knowledge-aspect. We know very little about its ecology and maybe even less about its biogeochemistry. By consequence, it introduces a huge uncertainty in every modeling effort trying to figure out what exactly is happening/will happen with system Earth in the future. Apart from that, it is probably also the most adventurous and most fun part of the world to work in. This thesis falls from the classic PhD thesis in the sense that it does not focus on one single aspect. This is partly due to the reality of funding. I started working on a leftover budget from COBIMFO, a BELSPO funded project trying to link biodiversity and carbon stocks in Yangambi, DRC. After 1.5 years, I got my own funding to continue my research, thanks to VLIR-UOS who decided to select my application for VLADOC funding. My study area suddenly elaborated to Eastern DRC, and my research shifted from ecology towards biogeochemistry. As you might or might not have read, the research presented here always stays at the interface between these two fields. Hope you enjoyed the reading. This final document is the result of about 3 months of writing, about 12 months in the field and 33 months of preparation and lab- and data-analysis.

Literature cited

91

Literature cited Achard, F. 2004. Improved estimates of net carbon emissions from land cover change in the

tropics for the 1990s. Global Biogeochemical Cycles 18:1–12.

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

Ache, and M. Franco. 2014. Functional traits explain variation in plant life history strategies.

Proceedings of the National Academy of Sciences of the United States of America 111:740–

745.

Allen, K., M. D. Corre, A. Tjoa, and E. Veldkamp. 2015. Soil nitrogen-cycling responses to

conversion of lowland forests to oil palm and rubber plantations in Sumatra, Indonesia. Plos

One 17:5168.

Allen, O. N., and E. K. Allen. 1981. The Leguminosae: a source book of characteristics, uses

and nodulation. The University of Wisconsin Press, Madison.

Alvarez-Clare, S., M. C. Mack, and M. Brooks. 2013. A direct test of nitrogen and phosphorus

limitation to net primary productivity in a lowland tropical wet forest. Ecology 94:1540–1551.

Archibald, S., A. C. Staver, and S. A. Levin. 2012. Evolution of human-driven fire regimes in

Africa. Proceedings of the National Academy of Sciences 109:847–852.

Arnold, J., M. D. Corre, and E. Veldkamp. 2008. Cold storage and laboratory incubation of

intact soil cores do not reflect in-situ nitrogen cycling rates of tropical forest soils. Soil Biology

and Biochemistry 40:2480–2483.

Arnold, J. E. 2008. Internal Nitrogen Cycling in Tropical Forest Soils. Georg-August-

Universität, Göttingen.

Asner, G. P., C. B. Anderson, R. E. Martin, R. Tupayachi, D. E. Knapp, and F. Sinca. 2015.

Landscape biogeochemistry reflected in shifting distributions of chemical traits in the Amazon

forest canopy. Nature Geoscience 8:567–573.

Asner, G. P., and R. E. Martin. 2016. Convergent elevation trends in canopy chemical traits of

tropical forests. Global Change Biology 22:2216–2227.

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

A. Shenkin, N. Salinas, F. Sinca, R. Tupayachi, D. Sandra, K. Q. Huaypar, M. M. Pillco, F.

Delis, C. Alvarez, Y. Malhi, K. Quispe Huaypar, M. Montoya Pillco, F. D. Ccori Álvarez, S.

Díaz, B. Enquist, Y. Malhi, D. Sandra, K. Q. Huaypar, M. M. Pillco, F. Delis, C. Alvarez, and

Y. Malhi. 2016. Scale dependence of canopy trait distributions along a tropical forest elevation

gradient. New Phytologist 214:973–988.

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

Martinez, L. Carranza-jiménez, and P. Martinez. 2014. Amazonian functional diversity from

forest canopy chemical assembly. Proceedings of the National Academy of Sciences

111:5604–5609.

Averill, C., and A. Finzi. 2011. Increasing plant use of organic nitrogen with elevation is

reflected in nitrogen uptake rates and ecosystem δ15N. Ecology 92:883–91.

Bai, E., and B. Z. Houlton. 2009. Coupled isotopic and process-based modeling of gaseous

nitrogen losses from tropical rain forests. Global Biogeochemical Cycles 23:1–10.

Literature cited

92

Bai, E., B. Z. Houlton, and Y. P. Wang. 2012. Isotopic identification of nitrogen hotspots across

natural terrestrial ecosystems. Biogeosciences 9:3287–3304.

Baldocchi, D., J. Finnigan, K. Wilson, and E. V. A. Falge. 2000. On measuring net ecosystem

carbon exchange over tall vegetation on complex terrain. Boundary-Layer Meteorology

96:257–291.

Balvanera, P., A. B. Pfisterer, N. Buchmann, J.-S. He, T. Nakashizuka, D. Raffaelli, and B.

Schmid. 2006. Quantifying the evidence for biodiversity effects on ecosystem functioning and

services. Ecology letters 9:1146–1156.

Banin, L., T. R. Feldpausch, O. L. Phillips, T. R. Baker, J. Lloyd, K. Affum-Baffoe, E. J. M. M.

Arets, N. J. Berry, M. Bradford, R. J. W. Brienen, S. Davies, M. Drescher, N. Higuchi, D. W.

Hilbert, A. Hladik, Y. Iida, K. A. Salim, a. R. Kassim, D. a. King, G. Lopez-Gonzalez, D.

Metcalfe, R. Nilus, K. S.-H. Peh, J. M. Reitsma, B. Sonké, H. Taedoumg, S. Tan, L. White, H.

Wöll, and S. L. Lewis. 2012. What controls tropical forest architecture? Testing environmental,

structural and floristic drivers. Global Ecology and Biogeography 21:1179–1190.

Barron, A. R., D. W. Purves, and L. O. Hedin. 2011. Facultative nitrogen fixation by canopy

legumes in a lowland tropical forest. Oecologia 165:511–20.

Bates, D., D. Sarkar, M. D. Bates, and L. Matrix. 2007. The lme4 Package. October 2:1–6.

Batterman, S. A., L. O. Hedin, M. van Breugel, J. Ransijn, D. J. Craven, and J. S. Hall. 2013.

Key role of symbiotic dinitrogen fixation in tropical forest secondary succession. Nature

502:224–227.

Bauters, M., E. Ampoorter, D. Huygens, E. Kearsley, T. De Haulleville, G. Sellan, H. Verbeeck,

P. Boeckx, and K. Verheyen. 2015. Functional identity explains carbon sequestration in a 77-

year-old experimental tropical plantation. Ecosphere 6:1–11.

Bauters, M., T. W. Drake, H. Verbeeck, S. Bodé, P. Hervé-Fernandez, P. Zito, D. C. Podgorski,

F. Boyemba, I. Makelele, L. C. Ntaboba, R. G. M. Spencer, and P. Boeckx. (n.d.). High fire-

derived nitrogen deposition on central African forests (in review). Proceedings of the National

Academy of Sciences.

Bauters, M., N. Mapenzi, E. Kearsley, B. Vanlauwe, and P. Boeckx. 2016. Facultative nitrogen

fixation by legumes in the central Congo basin is downregulated during late successional

stages. Biotropica 48:281–284.

Bauters, M., H. Verbeeck, M. Demol, S. Bruneel, C. Taveirne, D. Van der Heyden, L. Cizungu,

and P. Boeckx. 2017. Parallel functional and stoichiometric trait shifts in South American and

African forest communities with elevation. Biogeosciences 14:5313–5321.

Bauters, M., H. Verbeeck, S. Doetterl, E. Ampoorter, G. Baert, P. Vermeir, K. Verheyen, and

P. Boeckx. 2017b. Functional Composition of Tree Communities Changed Topsoil Properties

in an Old Experimental Tropical Plantation. Ecosystems 20:861–871.

Beer, C., M. Reichstein, E. Tomelleri, P. Ciais, M. Jung, N. Carvalhais, C. Rödenbeck, M. A.

Arain, D. Baldocchi, G. B. Bonan, A. Bondeau, A. Cescatti, G. Lasslop, A. Lindroth, M. Lomas,

S. Luyssaert, H. Margolis, K. W. Oleson, O. Roupsard, E. Veenendaal, N. Viovy, C. Williams,

F. I. Woodward, and D. Papale. 2010. Terrestrial Gross Carbon dioxide uptake: Global

distribution and covariation with Climate. Science 329:834–839.

Binkley, D., and C. Giardina. 1998. Why do tree species affect soils? The Warp and Woof of

tree-soil interactions. Biogeochemistry 42:89–106.

Literature cited

93

Bolan, N. S., M. J. Hedley, and R. E. White. 1991. Processes of soil acidification during

nitrogen cycling with emphasis on legume based pastures. Plant and Soil 134:53–63.

Bonan, G. B. 2008a. Carbon cycle: Fertilizing change. Nature Geoscience 1:645–646.

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

of forests. Science (New York, N.Y.) 320:1444–1449.

Bonnier, C. 1958. Symbiose Rhizobium-Légumineuses en région équatoriale. Publications de

l’Institut National pour l’étude Agronomique du Congo Belge 72.

Booth, M. S., J. M. Stark, and S. C. Hart. 2006. Soil-mixing effects on inorganic nitrogen

production and consumption in forest and shrubland soils. Plant and Soil 289:5–15.

Boy, J., C. Valarezo, W. Wilcke, R. Rollenbeck, C. Valarezo, and W. Wilcke. 2008. Amazonian

biomass burning-derived acid and nutrient deposition in the north Andean montane forest of

Ecuador. Global Biogeochemical Cycles 22:1–16.

Van Breugel, M., F. Bongers, and M. Martínez-Ramos. 2007. Species dynamics during early

secondary forest succession: Recruitment, mortality and species turnover. Biotropica 39:610–

619.

van Breugel, M., J. S. Hall, D. J. Craven, T. G. Gregoire, A. Park, D. H. Dent, M. H. Wishnie,

E. Mariscal, J. Deago, D. Ibarra, N. Cedeño, and M. S. Ashton. 2011. Early growth and survival

of 49 tropical tree species across sites differing in soil fertility and rainfall in Panama. Forest

Ecology and Management 261:1580–1589.

Brienen, R. J. W., O. L. Phillips, T. R. Feldpausch, and Et Al. 2015. Long-term decline of the

Amazon carbon sink. Nature 519:344–348.

Bronk, D. A., J. H. See, P. Bradley, and L. Killberg. 2007. DON as a source of bioavailable

nitrogen for phytoplankton. Biogeosciences:283–296.

Brookshire, E. N. J., S. Gerber, W. Greene, R. A. Jones, and S. A. Thomas. 2017. Plant uptake

and stream chemistry set global bounds on nitrogen gas emissions from humid tropical forests.

Geophysical Research Letters 44:2502–2510.

Brookshire, E. N. J., S. Gerber, D. N. L. Menge, and L. O. Hedin. 2012a. Large losses of

inorganic nitrogen from tropical rainforests suggest a lack of nitrogen limitation. Ecology letters

15:9–16.

Brookshire, E. N. J., L. O. Hedin, J. D. Newbold, D. M. Sigman, and J. K. Jackson. 2012b.

Sustained losses of bioavailable nitrogen from montane tropical forests. Nature Geoscience

5:123–126.

Brookshire, E. N. J., and S. a Thomas. 2013. Ecosystem consequences of tree

monodominance for nitrogen cycling in lowland tropical forest. PloS one 8:e70491.

Brookshire, E. N. J., H. M. Valett, S. A. Thomas, and J. R. Webster. 2007. Atmospheric N

deposition increases organic N loss from temperate forests. Ecosystems 10:252–262.

Brown, S., A. J. R. Gillespie, and A. E. Lugo. 1989. Biomass Estimation Methods for Tropical

Forests with Applications to Forest Inventory Data. Forest Science 35:881–902.

Bruijnzeel, L. a. 1991. Nutrient input–output budgets of tropical forest ecosystems: a review.

Journal of Tropical Ecology 7:1.

Literature cited

94

Cachier, H., and J. Ducret. 1991. Influence of biomass burning on equatorial African rains.

Nature 352:228–230.

Canadell, J. G., and M. R. Raupach. 2008. Managing Forests for Climate Change Mitigation.

Science 320:1456–1458.

Cape, J. N., S. E. Cornell, T. D. Jickells, and E. Nemitz. 2011. Organic nitrogen in the

atmosphere - Where does it come from? A review of sources and methods. Atmospheric

Research 102:30–48.

Cardinale, B. J., J. E. Duffy, A. Gonzalez, D. U. Hooper, C. Perrings, P. Venail, A. Narwani,

G. M. Mace, D. Tilman, D. a Wardle, A. P. Kinzig, G. C. Daily, M. Loreau, J. B. Grace, A.

Larigauderie, D. S. Srivastava, and S. Naeem. 2012. Biodiversity loss and its impact on

humanity. Nature 486:59–67.

Cassart, B., A. Angbonga Basia, H. Titeux, E. Andivia, and Q. Ponette. 2016. Contrasting

patterns of carbon sequestration between Gilbertiodendron dewevrei monodominant forests

and Scorodophloeus zenkeri mixed forests in the Central Congo basin. Plant and Soil

414:309-326.

Cerbu, G. a., B. M. Swallow, and D. Y. Thompson. 2011. Locating REDD: A global survey and

analysis of REDD readiness and demonstration activities. Environmental Science & Policy

14:168–180.

Chave, J., C. Andalo, S. Brown, M. a Cairns, J. Q. Chambers, D. Eamus, H. Fölster, F.

Fromard, N. Higuchi, T. Kira, J.-P. Lescure, B. W. Nelson, H. Ogawa, H. Puig, B. Riéra, and

T. Yamakura. 2005. Tree allometry and improved estimation of carbon stocks and balance in

tropical forests. Oecologia 145:87–99.

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

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

Chave, J., M. Réjou-Méchain, A. Burquez, E. Chidumayo, M. S. Colgan, W. B. C. Delitti, A.

Duque, T. Eid, P. M. Fearnside, R. C. Goodman, M. Henry, A. Martinez-Yrizar, W. A.

Mugasha, H. C. Muller-Landau, M. Mencuccini, B. W. Nelson, A. Ngomanda, E. M. Nogueira,

E. Ortiz-Malavassi, R. Pélissier, P. Ploton, C. M. Ryan, J. G. Saldarriaga, and G. Vieilledent.

2014. Improved allometric models to estimate the aboveground biomass of tropical trees.

Global Change Biology 20:3177–3190.

Chen, L., C. Liu, L. Zhang, R. Zou, and Z. Zhang. 2017. Variation in Tree Species Ability to

Capture and Retain Airborne Fine Particulate Matter (PM2.5). Scientific Reports:1–11.

Chen, Y., J. T. Randerson, G. R. Van Der Werf, D. C. Morton, M. Mu, P. S. Kasibhatla, and

B. S. Branch. 2010. Nitrogen deposition in tropical forests from savanna and deforestation

fires. Global Change Biology 16:2024–2038.

Chorover, J., and G. Sposito. 1995. Dissolution behavior of kaolinitic tropical soils. Geochimica

et Cosmochimica Acta 59:3109–3121.

Chuyong, G. B., D. M. Newbery, and N. C. Songwe. 2000. Litter nutrients and retranslocation

in a central African rain forest dominated by trees ectomycorrhizal. New Phytologist 148:493–

510.

Chuyong, G. B., D. M. Newbery, and N. C. Songwe. 2004. Rainfall input, throughfall and

stemflow of nutrients in a central African rain forest dominated by ectomycorrhizal trees.

biogeochemistry 148:493–510.

Literature cited

95

Ciais, P., A. Bombelli, M. Williams, S. L. Piao, J. Chave, C. M. Ryan, M. Henry, P. Brender, R.

Valentini, B. Y. P. C. Iais, A. B. Ombelli, M. W. Illiams, S. L. P. Iao, J. C. Have, P. Ciais, A.

Bombelli, M. Williams, S. L. Piao, J. Chave, C. M. Ryan, M. Henry, P. Brender, and R.

Valentini. 2011a. The carbon balance of Africa: synthesis of recent research studies.

Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering

Sciences 369:2038–2057.

Ciais, P., a Bombelli, M. Williams, S. L. Piao, J. Chave, C. M. Ryan, M. Henry, P. Brender,

and R. Valentini. 2011b. The carbon balance of Africa: synthesis of recent research studies.

Philosophical transactions. Series A, Mathematical, physical, and engineering sciences

369:2038–2057.

Ciais, P., C. Sabine, G. Bala, L. Bopp, V. Brovkin, J. Canadell, A. Chhabra, R. DeFries, J.

Galloway, M. Heimann, C. Jones, C. Le Quéré, R. B. Myneni, S. Piao, P. Thornton, and

Intergovernmental Panel on Climate Change. 2013. The physical science basis. Contribution

of working group I to the fifth assessment report of the intergovernmental panel on climate

change. Assessment Report 5:659–740.

Clark, D. A. D. A., S. Brown, D. W. D. W. Kicklighter, J. Q. J. Q. Chambers, J. R. J. R.

Thomlinson, J. Ni, J. O. H. N. R. T. Homlinson, and J. N. I. Ian. 2001. Measuring net primary

production in forests: concepts and field methods. Ecological Applications 11:356–370.

Condit, R., B. M. J. Engelbrecht, D. Pino, R. Pérez, and B. L. Turner. 2013. Species

distributions in response to individual soil nutrients and seasonal drought across a community

of tropical trees. Proceedings of the National Academy of Sciences 110:5064-5068.

Corby, H. D. L. 1974. Systematic implications of nodulation among Rhodesian legumes. Kirkia

9:301–309.

Corilo, Y. 2015. EnviroOrg. Florida State University.

Corlett, R. T., and R. B. Primack. 2006. Tropical rainforests and the need for cross-continental

comparisons. Trends in Ecology and Evolution 21:104–110.

Cornelissen, J. H. C. J. H. C., H. M. M. Quested, R. S. P. R. S. P. van Logtestijn, N. Pérez-

Harguindeguy, D. Gwynn-Jones, S. Díaz, T. V. V. Callaghan, M. C. C. Press, R. Aerts, and E.

Ecology. 2006. Foliar pH as a new plant trait: Can it explain variation in foliar chemistry and

carbon cycling processes among subarctic plant species and types? Oecologia 147:315–326.

Cornell, S. E. 2011. Atmospheric nitrogen deposition: Revisiting the question of the importance

of the organic component. Environmental Pollution 159:2214–2222.

Cornell, S. E., T. D. Jickells, J. N. Cape, a. P. Rowland, and R. a. Duce. 2003. Organic

nitrogen deposition on land and coastal environments: A review of methods and data.

Atmospheric Environment 37:2173–2191.

Corrales, A., S. a. Mangan, B. L. Turner, J. W. Dalling, A. Corrales, S. a. Mangan, B. L. Turner,

and J. W. Dalling. 2016. An ectomycorrhizal nitrogen economy facilitates monodominance in

a neotropical forest. Ecology Letters:383–392.

Corre, M. D., E. Veldkamp, J. Arnold, S. J. Wright, M. D. Corre, E. Veldkamp, J. Arnold, and

S. J. Wright. 2010. Impact of elevated N input on soil N cycling and losses in old-growth

lowland and montane forests in Panama. Ecology 91:1715–1729.

Cossalter, C., and C. Pye-smith. 2003. Fast-wood forestry: myths and realities. The Center for

International Forestry Research (CIFOR), Bogor, Indonesia.

Literature cited

96

Coûteaux, M. M., L. Sarmiento, P. Bottner, D. Acevedo, and J. M. Thiéry. 2002. Decomposition

of standard plant material along an altitudinal transect (65 - 3968 m) in the tropical Andes. Soil

Biology and Biochemistry 34:69–78.

Craine, J. M., E. N. J. Brookshire, M. D. Cramer, N. J. Hasselquist, K. Koba, E. Marin-spiotta,

and L. Wang. 2015. Ecological interpretations of nitrogen isotope ratios of terrestrial plants

and soils. Plant and Soil 396:1–26.

Craine, J. M., A. J. Elmore, M. P. M. Aidar, M. Bustamante, T. E. Dawson, E. A. Hobbie, A.

Kahmen, M. C. Mack, K. K. Mclauchlan, A. Michelsen, G. B. Nardoto, and L. H. Pardo. 2009.

Global patterns of foliar nitrogen isotopes and their relationships with climate, mycorrhizal

fungi, foliar nutrient concentrations, and nitrogen availability. New Phytologist 183:980–992.

Dauer, J. M., J. Chorover, O. a. Chadwick, J. Oleksyn, M. G. Tjoelker, S. E. Hobbie, P. B.

Reich, and D. M. Eissenstat. 2007. Controls over leaf and litter calcium concentrations among

temperate trees. Biogeochemistry 86:175–187.

Davidson, E. A., J. Chorover, and D. B. Dail. 2003. A mechanism of abiotic immobilization of

nitrate in forest ecosystems: The ferrous wheel hypothesis. Global Change Biology 9:228–

236.

Davidson, E. A., S. C. Hart, and M. K. Firestone. 1992. Internal cycling of nitrate in soils of a

mature coniferous forest. Ecology 73:1148–1156.

Davidson, E. A., M. Keller, H. E. Erickson, L. V. Verchot, and E. Veldkamp. 2000. Testing a

conceptual model of soil emissions of nitrous and nitric oxides. BioScience 50:783–792.

Davidson, E. a, C. J. R. de Carvalho, A. M. Figueira, F. Y. Ishida, J. P. H. B. Ometto, G. B.

Nardoto, R. T. Sabá, S. N. Hayashi, E. C. Leal, I. C. G. Vieira, and L. a Martinelli. 2007.

Recuperation of nitrogen cycling in Amazonian forests following agricultural abandonment.

Nature 447:995–8.

Dawud, S. M., K. Raulund-Rasmussen, T. Domisch, L. Finér, B. Jaroszewicz, and L.

Vesterdal. 2016. Is Tree Species Diversity or Species Identity the More Important Driver of

Soil Carbon Stocks, C/N Ratio, and pH? Ecosystems 19:645-660.

Dentener, F., J. Drevet, J. F. Lamarque, I. Bey, B. Eickhout, A. M. Fiore, D. Hauglustaine, L.

W. Horowitz, M. Krol, U. C. Kulshrestha, M. Lawrence, C. Galy-Lacaux, S. Rast, D. Shindell,

D. Stevenson, T. Van Noije, C. Atherton, N. Bell, D. Bergman, T. Butler, J. Cofala, B. Collins,

R. Doherty, K. Ellingsen, J. Galloway, M. Gauss, V. Montanaro, J. F. M??ller, G. Pitari, J.

Rodriguez, M. Sanderson, F. Solmon, S. Strahan, M. Schultz, K. Sudo, S. Szopa, O. Wild, J.

F. Müller, G. Pitari, J. Rodriguez, M. Sanderson, F. Solmon, S. Strahan, M. Schultz, K. Sudo,

S. Szopa, and O. Wild. 2006. Nitrogen and sulfur deposition on regional and global scales: A

multimodel evaluation. Global Biogeochemical Cycles 20:GB4003.

Diabate, M., A. Munive, S. M. de Faria, A. Ba, B. Dreyfus, and A. Galiana. 2005. Occurrence

of nodulation in unexplored leguminous trees native to the West African tropical rainforest and

inoculation response of native species useful in reforestation. The New phytologist 166:231–

9.

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

C. Wirth, I. C. Prentice, E. Garnier, G. Bönisch, M. Westoby, H. Poorter, P. B. Reich, A. T.

Moles, J. Dickie, A. N. Gillison, A. E. Zanne, S. Pierce, B. Shipley, D. Kirkup, F. Casanoves,

J. S. Joswig, A. Günther, V. Falczuk, N. Rüger, M. D. Mahecha, and L. D. Gorné. 2015. The

global spectrum of plant form and function. Nature 529:1–17.

Literature cited

97

Dittmar, T., B. Koch, N. Hertkorn, and G. Kattner. 2008. A simple and efficient method for the

solid-phase extraction of dissolved organic matter ( SPE-DOM ) from seawater. Limnology

and Oceanography: Methods 6:230–235.

Dixon, R. K., S. Brown, R. A. Houghton, A. M. Solomon, M. C. Trexler, and J. Wisnieski. 1994.

Carbon Pools and Flux of Global Forest Ecosystems. Science 263:185–190.

Draxler, R. 1997. Description of HYSPLIT_4 Modeling System.

Draxler, R. R. 1999. HYSPLIT_4 User’s Guide.

Draxler, R. R., and G. D. Hess. 1998. An Overview of the HYSPLIT_4 Modelling System for

Trajectories, Dispersion, and Deposition. Australian Meteorological Magazine 47:295–308.

Eriksson, E. R. K. 1969. respects , since there is considerable information available on chloride

in groundwater , the deposition of chloride , and the groundwater flow pattern in its undisturbed

state . Finally , some estimates of recharge rates based on the water balance method 7:178–

197.

Evans, J. R. 1989. Photosynthesis and nitrogen relationships in leaves of C3 plants. Oecologia

78:9–19.

Falkowski, P. G., T. Fenchel, and E. F. Delong. 2008. The Microbial Engines That Drive

Earth’s Biogeochemical Cycles. Science 320:1034–1039.

Falkowski, P., R. J. Scholes, E. Boyle, J. Canadell, D. Canfield, J. Elser, N. Gruber, K. Hibbard,

P. Högberg, S. Linder, F. T. Mackenzie, B. Moore III, T. Pedersen, Y. Rosenthal, S. Seitzinger,

V. Smetacek, and W. Steffen. 2000. The Global Carbon Cycle: A Test of Our Knowledge of

Earth as a System. Science 290:291–296.

Fang, Y., K. Koba, A. Makabe, C. Takahashi, W. Zhu, T. Hayashi, A. A. Hokari, R. Urakawa,

E. Bai, B. Z. Houlton, D. Xi, S. Zhang, K. Matsushita, Y. Tu, D. Liu, F. Zhu, Z. Wang, G. Zhou,

D. Chen, T. Makita, H. Toda, X. Liu, Q. Chen, D. Zhang, Y. Li, and M. Yoh. 2015. Microbial

denitrification dominates nitrate losses from forest ecosystems. Proceedings of the National

Academy of Sciences of the United States of America 112:1470–1474.

Feldpausch, T. R., J. Lloyd, S. L. Lewis, R. J. W. Brienen, M. Gloor, a. Monteagudo Mendoza,

G. Lopez-Gonzalez, L. Banin, K. Abu Salim, K. Affum-Baffoe, M. Alexiades, S. Almeida, I.

Amaral, a. Andrade, L. E. O. C. Aragão, a. Araujo Murakami, E. J. M. M. Arets, L. Arroyo, G.

a. Aymard C., T. R. Baker, O. S. Bánki, N. J. Berry, N. Cardozo, J. Chave, J. a. Comiskey, E.

Alvarez, a. de Oliveira, a. Di Fiore, G. Djagbletey, T. F. Domingues, T. L. Erwin, P. M.

Fearnside, M. B. França, M. a. Freitas, N. Higuchi, Y. Iida, E. Jiménez, a. R. Kassim, T. J.

Killeen, W. F. Laurance, J. C. Lovett, Y. Malhi, B. S. Marimon, B. H. Marimon-Junior, E. Lenza,

a. R. Marshall, C. Mendoza, D. J. Metcalfe, E. T. a. Mitchard, D. a. Neill, B. W. Nelson, R.

Nilus, E. M. Nogueira, a. Parada, K. S.-H. Peh, a. Pena Cruz, M. C. Peñuela, N. C. a. Pitman,

a. Prieto, C. a. Quesada, F. Ramírez, H. Ramírez-Angulo, J. M. Reitsma, a. Rudas, G. Saiz,

R. P. Salomão, M. Schwarz, N. Silva, J. E. Silva-Espejo, M. Silveira, B. Sonké, J. Stropp, H.

E. Taedoumg, S. Tan, H. ter Steege, J. Terborgh, M. Torello-Raventos, G. M. F. van der

Heijden, R. Vásquez, E. Vilanova, V. a. Vos, L. White, S. Willcock, H. Woell, and O. L. Phillips.

2012. Tree height integrated into pantropical forest biomass estimates. Biogeosciences

9:3381–3403.

Fernandez-Martinez, M., S. Vicca, I. a. Janssens, J. Sardans, S. Luyssaert, M. Campioli, F.

S. C. Iii, P. Ciais, Y. Malhi, M. Obersteiner, D. Papale, S. L. Piao, and M. Reichstein. 2014.

Literature cited

98

Nutrient availability as the key regulator of global forest carbon balance. Nature Climate

Change 4:471–476.

Fick, S. E., and R. J. Hijmans. 2017. WorldClim 2 : new 1-km spatial resolution climate

surfaces for global land areas. International Journal of Climatology 37:4302-4315.

Field, C. B., M. J. Behrenfeld, and J. T. Randerson. 1998. Primary Production of the

Biosphere : Integrating Terrestrial and Oceanic Components 281:237–241.

Finzi, A. C., N. Van Breemen, C. D. Canham, S. E. Applications, N. May, N. Van Breemen,

and C. D. Canham. 1998a. Canopy Tree-Soil Interactions within Temperate Forests : Species

Effects on Soil Carbon and Nitrogen. Ecological Applications 8:440–446.

Finzi, A. C., C. D. Canham, and N. Van Breemen. 1998b. Canopy tree soil interactions within

temperate forests: Species effects on pH and cations. Ecological Applications 8:447–454.

Finzi, a C., N. Van Breemen, and C. D. Canham. 1998c. Canopy Tree–Soil Interactions Within

Temperate Forests: Species Effects on Soil Carbon and Nitrogen. Ecological Applications

8:440–446.

Fisher, J. B., Y. Malhi, I. C. Torres, D. B. Metcalfe, M. J. van de Weg, P. Meir, J. E. Silva-

Espejo, W. H. Huasco, and T. Daniel. 2013. Nutrient limitation in rainforests and cloud forests

along a 3,000-m elevation gradient in the Peruvian Andes. Oecologia 172:889–902.

Fisher, R. F. 1995. Amelioration of Degraded Rain Forest Soils by Plantations of Native Trees.

Soil Science Society of America Journal 59:544–549.

Forrester, D. I., J. Bauhus, A. L. Cowie, and J. K. Vanclay. 2006. Mixed-species plantations of

Eucalyptus with nitrogen-fixing trees: A review. Forest Ecology and Management 233:211–

230.

Fortunel, C., C. E. T. Paine, P. V. a. Fine, N. J. B. Kraft, and C. Baraloto. 2014. Environmental

factors predict community functional composition in Amazonian forests. Journal of Ecology

102:145–155.

Freschet, G. T., R. Aerts, and J. H. C. Cornelissen. 2012. A plant economics spectrum of litter

decomposability. Functional Ecology 26:56–65.

Fujikake, H., H. Yashima, T. Sato, N. Ohtake, K. Sueyoshi, and T. Ohyama. 2002. Rapid and

Reversible Nitrate Inhibition of Nodule Growth and N2 fixation activity in Soybean (Glycine

max (L.) Merr.). Soil Science and Plant Nutrition 48:211–217.

Fyllas, N. M., S. Patiño, T. R. Baker, G. B. Nardoto, L. A. Martinelli, C. A. Quesada, R. Paiva,

A. Okologie, N. M. Fyllas, S. Patiño, T. R. Baker, G. B. Nardoto, L. A. Martinelli, C. A. Quesada,

R. Paiva, M. Schwarz, V. Horna, L. M. Mercado, A. Santos, L. Arroyo, E. M. Jimenez, F. J.

Luizao, D. a. Neill, N. Silva, A. Prieto, A. Rudas, M. Silviera, I. C. G. Vieira, G. Lopez-Gonzalez,

Y. Malhi, O. L. Phillips, and J. Lloyd. 2009. Basin-wide variations in foliar properties of

Amazonian forest: phylogeny, soils and climate. Biogeosciences 6:2677–2708.

Galloway, J. N., F. J. Dentener, D. G. Capone, E. W. Boyer, R. W. Howarth, S. P. Seitzinger,

G. P. Asner, C. C. Cleveland, P. A. Green, E. A. Holland, D. M. Karl, A. F. Michaels, J. H.

Porter, A. R. Townsend, and C. J. Vorosmarty. 2004. Nitrogen cycles : past , present , and

future. Biogeochemistry 70:153–226.

Literature cited

99

Galloway, J. N., A. R. Townsend, J. W. Erisman, M. Bekunda, Z. Cai, J. R. Freney, L. A.

Martinelli, S. P. Seitzinger, and M. A. Sutton. 2008. Transformation of the Nitrogen Cycle :

Recent Trends, Questions, and Potential Solutions. Science 320:889–893.

Galy-Lacaux, C., C. Delon, C. Galy-Lacaux, and C. Delon. 2014. Nitrogen emission and

deposition budget in West and Central Africa. Environmental Research Letters 9:125002.

Gaston, K. J. 2000. Global patterns in biodiversity. Nature 405:220-227.

Geider, R. J., E. H. Delucia, P. G. Falkowski, A. C. Finzi, J. P. Grime, J. Grace, T. M. Kana, J.

L. A. Roche, S. P. Long, B. A. Osborne, T. Platt, I. C. Prentice, J. A. Raven, W. H. Schlesinger,

V. Smetacek, V. Stuart, S. Sathyendranath, R. B. Thomas, T. C. Vogelmann, P. Williams, and

F. I. Woodward. 2001. Primary productivity of planet earth : biological determinants and

physical constraints in terrestrial and aquatic habitats. Global Change Biology 7:849–882.

Gérard, P. 1960. Etude écologique de la forêt dense à Gilbertiodendron dewevrei dans la

région de l’Uele. Publications de l’Institut National pour l’étude Agronomique du Congo Belge

87.

Gerschlauer, F., M. Dannenmann, A. Kühnel, R. Meier, A. Kolar, K. Butterbach-Bahl, and R.

Kiese. 2016. Gross Nitrogen Turnover of Natural and Managed Tropical Ecosystems at Mt.

Kilimanjaro, Tanzania. Ecosystems 19:1271–1288.

Global Modeling and Assimilation Office. (n.d.). MERRA-2 tavg1_2d_aer_Nx: 2d,1-

Hourly,Time-averaged,Single-Level,Assimilation,Aerosol Diagnostics V5.12.4. Goddard Earth

Sciences Data and Information Services Center (GES DISC), Greenbelt, MD, USA,.

Golberg, E. . 1985. Black carbon in the environment. Journal of the American Veterinary

Medical Association 244:353–354.

Goll, D. S., V. Brovkin, B. R. Parida, C. H. Reick, J. Kattge, P. B. Reich, P. M. van Bodegom,

and Ü. Niinemets. 2012. Nutrient limitation reduces land carbon uptake in simulations with a

model of combined carbon, nitrogen and phosphorus cycling. Biogeosciences 9:3547–3569.

Goll, D. S., N. Vuichard, F. Maignan, A. Jornet-Puig, J. Sardans, A. Violette, S. Peng, Y. Sun,

M. Kvakic, M. Guimberteau, B. Guenet, S. Zaehle, J. Peñuelas, I. Janssens, and P. Ciais.

2017. A representation of the phosphorus cycle for ORCHIDEE (revision 4520). Geoscientific

Model Development 10:3745–3770.

Greve, M., B. Reyers, A. Mette Lykke, J. Svenning, and A. M. Lykke. 2013. Spatial optimization

of carbon-stocking projects across Africa integrating stocking potential with co-benefits and

feasibility. Nature communications 4:2975.

van Groenigen, J. W., D. Huygens, P. Boeckx, T. W. Kuyper, I. M. Lubbers, T. Rütting, P. M.

Groffman, J. W. Van Groenigen, D. Huygens, P. Boeckx, T. W. Kuyper, I. M. Lubbers, and T.

Rütting. 2015. The soil N cycle: new insights and key challenges. Soil 1:235–256.

Gruba, P., J. Mulder, and S. Brozek. 2013. Modelling the pH dependency of dissolved calcium

and aluminium in O, A and B horizons of acid forest soils. Geoderma 206:85–91.

Gu, B. H., J. Schmitt, Z. H. Chen, L. Y. Liang, and J. F. McCarthy. 1994. Adsorption and

Desorption of Natural Organic-Matter on Iron-Oxide - Mechanisms and Models. Environmental

Science & Technology 28:38–46.

Literature cited

100

Gücker, B., R. C. S. Silva, D. Graeber, J. A. F. Monteiro, E. N. J. Brookshire, R. C. Chaves,

and I. G. Boechat. 2016. Dissolved nutrient exports from natural and human-impacted

Neotropical catchments. Global Ecology and Biogeography 25:378–390.

Gütlein, A., M. Dannenmann, and R. Kiese. 2016. Gross nitrogen turnover rates of a tropical

lower montane forest soil: Impacts of sample preparation and storage. Soil Biology and

Biochemistry 95:8–10.

Hansen, M. C. C., P. V Potapov, R. Moore, M. Hancher, S. A. a Turubanova, A. Tyukavina,

D. Thau, S. V. V Stehman, S. J. J. Goetz, T. R. R. Loveland, A. Kommareddy, A. Egorov, L.

Chini, C. O. O. Justice, J. R. G. R. G. Townshend, P. V. Patapov, R. Moore, M. Hancher, S.

A. a Turubanova, A. Tyukavina, D. Thau, S. V. V Stehman, S. J. J. Goetz, T. R. R. Loveland,

A. Kommaredy, A. Egorov, L. Chini, C. O. O. Justice, and J. R. G. R. G. Townshend. 2013.

High-Resolution Global Maps of 21st-Century Forest Cover Change. Science 342:850–854.

Hauck, R. D. 1982. Nitrogen - Isotope-ratio analysis. Pages 735–779Methods of soil analysis,

Part 2. Chemical and microbiological properties - Agronomy monograph no. 9, 2nd edition.

Hawthorne, W. D. 1995. Ecological profiles of Ghanaian forest trees. Page Tropical Forestry

Papers, No. 29,. Tropical Forestry Papers 29, Oxford Forestry Institute, Oxford, UK.

Hedin, L. O., J. J. Armesto, and A. H. Johnson. 1995. Patterns of Nutrient Loss from

unpolluted, old-growth temperate Forests: evaluation of Biogeochemical Theory. Ecology

76:493–509.

Hedin, L. O., E. N. N. J. Brookshire, D. N. L. L. Menge, and A. R. Barron. 2009. The Nitrogen

Paradox in Tropical Forest Ecosystems. Annual Review of Ecology, Evolution, and

Systematics 40:613–635.

Hedin, L. O., P. M. Vitousek, and P. A. Matson. 2003. Nutrient losses over four million years

of tropical forest development. Ecology 84:2231–2255.

Heil, J., H. Vereecken, and N. Brüggemann. 2016. A review of chemical reactions of

nitrification intermediates and their role in nitrogen cycling and nitrogen trace gas formation in

soil. European Journal of Soil Science 67:23–39.

Hemingway, J. D. 2017. fouriertransform: open-source tools for FT-ICR MS data analysis

(http://github.com/FluvialSeds/fouriertransform).

Hendrickson, C. L., J. P. Quinn, N. K. Kaiser, D. F. Smith, G. T. Blakney, T. Chen, A. G.

Marshall, C. R. Weisbrod, and S. C. Beu. 2015. 21 Tesla Fourier Transform Ion Cyclotron

Resonance Mass Spectrometer : A National Resource for Ultrahigh Resolution Mass

Analysis:1626–1632.

Hietz, P., B. L. Turner, W. Wanek, a. Richter, C. a. Nock, and S. J. Wright. 2011. Long-Term

Change in the Nitrogen Cycle of Tropical Forests. Science 334:664–666.

Hillebrand, H., and B. Matthiessen. 2009. Biodiversity in a complex world: consolidation and

progress in functional biodiversity research. Ecology letters 12:1405–1419.

Hobbie, E. a., A. Jumpponen, J. Trappe, Æ. A. Jumpponen, and Æ. J. Trappe. 2005. Foliar

and fungal 15N:14N ratios reflect development of mycorrhizae and nitrogen supply during

primary succession: Testing analytical models. Oecologia 146:258–268.

Hobbie, E. a., and A. P. Ouimette. 2009. Controls of nitrogen isotope patterns in soil profiles.

Biogeochemistry 95:355–371.

Literature cited

101

Hobbie, S. E., M. Ogdahl, J. Chorover, O. a. Chadwick, J. Oleksyn, R. Zytkowiak, and P. B.

Reich. 2007. Tree Species Effects on Soil Organic Matter Dynamics: The Role of Soil Cation

Composition. Ecosystems 10:999–1018.

Hofhansl, F., W. Wanek, S. Drage, W. Huber, A. Weissenhofer, and A. Richter. 2011.

Topography strongly affects atmospheric deposition and canopy exchange processes in

different types of wet lowland rainforest, Southwest Costa Rica. Biogeochemistry 106:371–

396.

Högberg, P. 1997. 15N natural abundance in soil – plant systems. New Phytologist 137:179–

203.

Högberg, P., and I. J. Alexander. 1995. Roles of Root Symbioses in African Woodland and

Forest - Evidence From N-15 Abundance and Foliar Analysis. Journal of Ecology 83:217–224.

Holtgrieve, G. W., P. K. Jewett, and P. A. Matson. 2006. Variations in soil N cycling and trace

gas emissions in wet tropical forests. Oecologia 146:584–594.

Hooper, D. U. U., F. S. I. Chapin, J. J. Ewel, A. Hector, P. Inchausti, S. Lavorel, J. H. Lawton,

D. M. Lodge, M. Loreau, S. Naeem, B. Schmid, H. Setälä, A. J. Symstad, J. Vandermeer, D.

A. Wardle, and H. Setala. 2005. Effects of biodiversity on ecosystem functioning: a consensus

ofcurrent knowledge. Ecological Monographs 75:3–35.

Houghton, R. a. 2005. Aboveground Forest Biomass and the Global Carbon Balance. Global

Change Biology 11:945–958.

Houghton, R. a., J. I. House, J. Pongratz, G. R. Van Der Werf, R. S. Defries, M. C. Hansen,

C. Le Quéré, and N. Ramankutty. 2012. Carbon emissions from land use and land-cover

change. Biogeosciences 9:5125–5142.

Houlton, B. Z., D. M. Sigman, and L. O. Hedin. 2006. Isotopic evidence for large gaseous

nitrogen losses from tropical rainforests. Proceedings of the National Academy of Sciences of

the United States of America 103:8745–50.

Houlton, B. Z., Y.-P. Wang, P. M. Vitousek, and C. B. Field. 2008. A unifying framework for

dinitrogen fixation in the terrestrial biosphere. Nature 454:327–30.

Huang, S., S. J. Titus, and D. P. Wiens. 1992. Comparison of nonlinear height-diameter

functions for major Alberta tree species. Canadian Journal of Forest Research 22:1297-1304.

Hubau, W., J. Van den Bulcke, P. Kitin, F. Mees, J. Van Acker, and H. Beeckman. 2012.

Charcoal identification in species-rich biomes: A protocol for Central Africa optimised for the

Mayumbe forest. Review of Palaeobotany and Palynology 171:164–178.

Hulvey, K. B., R. J. Hobbs, R. J. Standish, D. B. Lindenmayer, L. Lach, and M. P. Perring.

2013. Benefits of tree mixes in carbon plantings. Nature Climate Change 3:869–874.

Huygens, D., P. Boeckx, P. Templer, L. Paulino, O. Van Cleemput, C. Oyarzún, C. Müller, and

R. Godoy. 2008. Mechanisms for retention of bioavailable nitrogen in volcanic rainforest soils.

Nature Geoscience 1:543–548.

Ito, A., G. Lin, and J. E. Penner. 2015. Global modeling study of soluble organic nitrogen from

open biomass burning. Atmospheric Environment 121:103–112.

Janssens, I. A., W. Dieleman, S. Luyssaert, J.-A. Subke, M. Reichstein, R. Ceulemans, P.

Ciais, A. J. Dolman, J. Grace, G. Matteucci, D. Papale, S. L. Piao, E.-D. Schulze, J. Tang, and

Literature cited

102

B. E. Law. 2010. Reduction of forest soil respiration in response to nitrogen deposition. Nature

Geoscience 3:315–322.

Jia, Y., G. Yu, Y. Gao, N. He, Q. Wang, C. Jiao, and Y. Zuo. 2016. Global inorganic nitrogen

dry deposition inferred from ground- and space-based measurements. Scientific Reports

6:19810.

Jiang, X., X. Xin, S. Li, J. Zhou, T. Zhu, C. Müller, Z. Cai, and A. L. Wright. 2015. Effects of Fe

oxide on N transformations in subtropical acid soils. Scientific Reports 5:8615.

Jickells, T., A. R. Baker, J. N. Cape, S. E. Cornell, and E. Nemitz. 2013. The cycling of organic

nitrogen through the atmosphere. Philosophical transactions of the Royal Society of London.

Series B, Biological sciences 368:20130115.

Jindal, R., B. Swallow, and J. T. Kerr. 2008. Forestry-based carbon sequestration project in

Africa: Potential benefits and challenges. Natural Resources Forum 32:116–130.

Jobbágy, E. G., and R. B. Jackson. 2003. Patterns and mechanisms of soil acidification in the

conversion of grasslands to forests. Biogeochemistry 64:205–229.

John, R., J. W. Dalling, K. E. Harms, J. B. Yavitt, R. F. Stallard, M. Mirabello, S. P. Hubbell,

R. Valencia, H. Navarrete, M. Vallejo, and R. B. Foster. 2007. Soil nutrients influence spatial

distributions of tropical tree species. Proceedings of the National Academy of Sciences

104:864-869.

Johnson, A. H., J. Frizano, and D. R. Vann. 2003. Biogeochemical implications of labile

phosphorus in forest soils determined by the Hedley fractionation procedure. Oecologia

135:487–99.

Johnson, D. W., and P. S. Curtis. 2001. Effects of forest management on soil C and N storage :

meta analysis. Forest Ecology and Management 140:227–238.

Kahmen, A., W. Wanek, and N. Buchmann. 2008. Foliar d15N values characterize soil N

cycling and reflect nitrate or ammonium preference of plants along a temperate grassland

gradient. Oecologia 156:861–870.

Kanakidou, M., S. Myriokefalitakis, N. Daskalakis, G. Fanourgakis, A. Nenes, A. R. Baker, K.

Tsigaridis, and N. Mihalopoulos. 2016. Past, Present, and Future Atmospheric Nitrogen

Deposition. Journal of the Atmospheric Sciences 73:2039–2047.

Kauffman, S., W. Sombroek, and S. Mantel. 1998. Soils of Rainforests: Characterization and

major constraints of dominant forest soils in the humid tropics. Pages 9–20in A. Schulte and

D. (Eds. . Ruhiyat, editors.Soils of Tropical forest ecosystems. Springer Berlin, Heidelberg.

Kearsley, E., T. de Haulleville, K. Hufkens, A. Kidimbu, B. Toirambe, G. Baert, D. Huygens,

Y. Kebede, P. Defourny, J. Bogaert, H. Beeckman, K. Steppe, P. Boeckx, and H. Verbeeck.

2013. Conventional tree height-diameter relationships significantly overestimate aboveground

carbon stocks in the Central Congo Basin. Nature communications 4:2269.

Kearsley, E., H. Verbeeck, K. Hufkens, F. Van de Perre, S. Doetterl, G. Baert, H. Beeckman,

P. Boeckx, and D. Huygens. 2017. Functional community structure of African monodominant

Gilbertiodendron dewevrei forest influenced by local environmental filtering. Ecology and

Evolution 7:295–304.

Kelty, M. J. 2006. The role of species mixtures in plantation forestry. Forest Ecology and

Management 233:195–204.

Literature cited

103

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

along altitudinal gradients with contrasting soil phosphorus pools on Mount Kinabalu , Borneo.

Journal of Ecology 90:37–51.

Knops, J. M. H., K. L. Bradley, and D. A. Wedin. 2002. Mechanisms of plant species impacts

on ecosystem nitrogen cycling. Ecology Letters 5:454–466.

Koch, B. P., and T. Dittmar. 2006. From mass to structure: An aromaticity index for high-

resolution mass data of natural organic matter. Rapid Communications in Mass Spectrometry

20:926–932.

Körner, C. 2007. The use of “altitude” in ecological research. Trends in ecology & evolution

22:569–74.

Kuyper, T. W. 2012. Ectomycorrhiza and the open nitrogen cycle in an afrotropical rainforest.

New Phytologist 195:728–729.

Kuznetsova, A., P. B. Brockhoff, and R. H. B. Christensen. 2014. lmerTest: Tests for random

and fixed effects for linear mixed effect models (lmer objects of lme4 package).

Lachouani, P., A. H. Frank, and W. Wanek. 2010. A suite of sensitive chemical methods to

determine the δ15N of ammonium, nitrate and total dissolved N in soil extracts. Rapid

communications in mass spectrometry : RCM 24:3615–3623.

Laganière, J., D. A. Angers, and D. Paré. 2010. Carbon accumulation in agricultural soils after

afforestation: a meta-analysis. Global Change Biology 16:439–453.

Lajtha, K., C. T. Driscoll, W. M. Jarrell, and E. T. Elliott. 1999. Soil phosphorus:

characterization and total element analysis. Standard Soil Methods for Long Term Ecological

Research:115–143.

Laliberte, E., G. Zemunik, and B. L. Turner. 2014. Environmental filtering explains variation in

plant diversity along resource gradients. Science 345:1602–1605.

Lamb, D., P. D. Erskine, and J. A. Parrotta. 2005. Restoration of degraded tropical forest

landscapes. Science (New York, N.Y.) 310:1628–1632.

Lang’at, J. K. S., B. K. Y. Kirui, M. W. Skov, J. G. Kairo, M. Mencuccini, and M. Huxham. 2013.

Species mixing boosts root yield in mangrove trees. Oecologia 172:271–278.

Laughlin, D. C., S. J. Richardson, E. F. Wright, and P. J. Bellingham. 2015. Environmental

Filtering and Positive Plant Litter Feedback Simultaneously Explain Correlations Between Leaf

Traits and Soil Fertility. Ecosystems 18:1269–1280.

Lawrence, D., P. D’Odorico, L. Diekmann, M. DeLonge, R. Das, and J. Eaton. 2007. Ecological

feedbacks following deforestation create the potential for a catastrophic ecosystem shift in

tropical dry forest. Proceedings of the National Academy of Sciences 104:20696–20701.

Lebrun, J., and G. Gilbert. 1954. Une classification écologique des forêts du Congo. Série

Scientifique INEAC 63:1–88.

Lewis, S. L., B. Sonké, T. Sunderland, S. K. Begne, G. Lopez-gonzalez, G. V. D. M. F. Van

Der Heijden, O. L. Phillips, K. Affum-baffoe, T. R. Baker, L. Banin, J. J.-F. Bastin, H.

Beeckman, P. Boeckx, J. Bogaert, C. De Cannière, C. J. Clark, M. Collins, G. Djagbletey, M.

N. K. Djuikouo, J. J.-L. Doucet, C. E. N. Ewango, S. Fauset, T. R. Feldpausch, G. Ernest, J.

J.-F. J. Gillet, A. C. Hamilton, D. J. Harris, T. B. Hart, T. De Haulleville, A. Hladik, K. Hufkens,

D. Huygens, P. Jeanmart, K. J. Jeffery, E. Kearsley, M. E. Leal, J. Lloyd, J. C. Lovett, J.-R. J.

Literature cited

104

J.-R. Makana, Y. Malhi, R. Andrew, L. Ojo, K. S. Peh, G. Pickavance, J. R. Poulsen, J. M.

Reitsma, D. Sheil, M. Simo, K. Steppe, H. E. Taedoumg, J. Talbot, R. D. James, D. Taylor, S.

C. Thomas, B. Toirambe, H. Verbeeck, J. Vleminckx, J. Lee, L. J. T. White, S. Willcock, H.

Woell, L. Zemagho, T. White, S. Willcock, H. Woell, L. Zemagho, P. T. R. S. B, B. Sonke, R.

Poulsen, C. Thomas, E. Kearsley, M. E. Leal, J. Lloyd, J. C. Lovett, J.-R. J. J.-R. Makana, Y.

Malhi, R. Andrew, L. Ojo, K. S. Peh, G. Pickavance, J. R. Poulsen, J. M. Reitsma, D. Sheil,

M. Simo, K. Steppe, H. E. Taedoumg, J. Talbot, R. D. James, D. Taylor, S. C. Thomas, B.

Toirambe, H. Verbeeck, J. Vleminckx, J. Lee, L. J. T. White, S. Willcock, H. Woell, and L.

Zemagho. 2013. Above-ground biomass and structure of 260 African tropical forests Above-

gr. Philosophical transactions of the Royal Society of London. Series B, Biological sciences

368:20120295.

Lewis, W. M., J. M. Melack, W. H. McDowell, M. Mcclain, and J. E. Richey. 1999. Nitrogen

yields from undisturbed watersheds in the Americas. Biogeochemistry 46:149–162.

Liu, S., A. E. Berns, H. Vereecken, D. Wu, and N. Brüggemann. 2017. Interactive effects of

MnO2, organic matter and pH on abiotic formation of N2O from hydroxylamine in artificial soil

mixtures. Scientific Reports 7:39590.

Lohse, K. A., and P. Matson. 2005. Consequences of nitrogen additions for soil losses from

wet tropical forests. Ecological applications 15:1629–1648.

De longe, M., P. D’odorico, and D. Lawrence. 2008. Feedbacks between phosphorus

deposition and canopy cover: The emergence of multiple stable states in tropical dry forests.


Loreau, M., and A. Hector. 2001. Partitioning selection and complementarity in biodiversity

experiments. Nature 412:72–76.

Loreau, M., S. Naeem, P. Inchausti, J. Bengtsson, J. P. Grime, A. Hector, D. U. Hooper, M. a

Huston, D. Raffaelli, B. Schmid, D. Tilman, and D. a Wardle. 2001. Biodiversity and ecosystem

functioning: current knowledge and future challenges. Science (New York, N.Y.) 294:804–

808.

Losos, E., and E. G. Leigh. 2004. Tropical Forest Diversity and Dynamism. First edition. The

University of Chicago press books. Chicago. 688 pp.

Mace, K. A., P. Artaxo, and R. A. Duce. 2003. Water-soluble organic nitrogen in Amazon Basin

aerosols during the dry (biomass burning) and wet seasons. Journal of Geophysical Research

108:4512.

Malhi, Y., C. A. J. Girardin, G. R. Goldsmith, C. E. Doughty, N. Salinas, D. B. Metcalfe, W.

Huaraca Huasco, J. E. Silva-Espejo, J. del Aguilla-Pasquell, F. Farfan Amizquita, L. E. O. C.

Aragao, R. Guerrieri, F. Y. Ishida, N. H. A. Bahar, W. Farfan-Rios, O. L. Phillips, P. Meir, and

M. Silman. 2016. The variation of productivity and its allocation along a tropical elevation

gradient: A whole carbon budget perspective. New Phytologist 214:1019–1032.

Malhi, Y., O. L. Phillips, J. Lloyd, T. Baker, J. Wright, S. Almeida, L. Arroyo, T. Frederiksen, J.

Grace, N. Higuchi, T. Killeen, W. F. Laurance, C. Leaño, S. Lewis, P. Meir, a. Monteagudo,

D. Neill, P. Núñez Vargas, S. N. Panfil, S. Patiño, N. Pitman, C. a. Quesada, a. Rudas-Ll., R.

Salomão, S. Saleska, N. Silva, M. Silveira, W. G. Sombroek, R. Valencia, R. Vásquez

Martínez, I. C. G. Vieira, and B. Vinceti. 2002. An international network to monitor the

structure, composition and dynamics of Amazonian forests (RAINFOR). Journal of Vegetation

Science 13:439.

Literature cited

105

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

Elevation gradients in the tropics: Laboratories for ecosystem ecology and global change

research. Global Change Biology 16:3171–3175.

Manfroi, O. J., K. Koichiro, T. Nobuaki, S. Masakazu, M. Nakagawa, T. Nakashizuka, and L.

Chong. 2004. The stemflow of trees in a Bornean lowland tropical forest. Hydrological

Processes 18:2455–2474.

Mann, P. J., A. Davydova, N. Zimov, R. G. M. Spencer, S. Davydov, E. Bulygina, S. Zimov,

and R. M. Holmes. 2012. Controls on the composition and lability of dissolved organic matter

in Siberia’s Kolyma River basin. Journal of Geophysical Research 117:1–15.

Marlon, J. R., P. J. Bartlein, C. Carcaillet, D. G. Gavin, S. P. Harrison, P. E. Higuera, F. Joos,

M. J. Power, and I. C. Prentice. 2009. Climate and human influences on global biomass

burning over the past two millennia. Nature Geoscience 2:307–307.

Marrs, R. H. H., J. Proctor, A. Heaney, and M. D. D. Mountford. 1988. Changes in Soil

Nitrogen-Mineralization and Nitrification Along an Altitudinal Transect in Tropical Rain Forest

in Costa Rica. Journal of Ecology 76:466–482.

Martens, B., D. G. Miralles, H. Lievens, R. Van Der Schalie, R. A. M. De Jeu, D. Fernández-

Prieto, H. E. Beck, W. A. Dorigo, and N. E. C. Verhoest. 2017. GLEAM v3: Satellite-based

land evaporation and root-zone soil moisture. Geoscientific Model Development 10:1903–

1925.

Martinelli, L. a., M. C. Piccolo, a. R. Townsend, P. M. Vitousek, E. Cuevas, W. McDowell, G.

P. Robertson, O. C. Santos, and K. Treseder. 1999. Nitrogen stable isotopic composition of

leaves and soil: Tropical versus temperate forests. Biogeochemistry 46:45–65.

Matson, A. L., M. D. Corre, J. I. Burneo, and E. Veldkamp. 2015. Free-living nitrogen fixation

responds to elevated nutrient inputs in tropical montane forest floor and canopy soils of

southern Ecuador. Biogeochemistry 122:281–294.

Matson, P. a., W. H. McDowell, A. R. Townsend, and P. M. Vitousek. 1999. The globalization

of N deposition: ecosystem consequences in tropical environments. Biogeochemistry 46:67–

83.

Mayor, J., M. Bahram, T. Henkel, F. Buegger, K. Pritsch, and L. Tedersoo. 2015.

Ectomycorrhizal impacts on plant nitrogen nutrition : emerging isotopic patterns , latitudinal

variation and hidden mechanisms. Ecology Letters 18:96–107.

Mayor, J. R., N. J. Sanders, A. T. Classen, R. D. Bardgett, J. Clément, A. Fajardo, S. Lavorel,

M. K. Sundqvist, M. Bahn, C. Chisholm, E. Cieraad, Z. Gedalof, K. Grigulis, G. Kudo, D. L.

Oberski, and D. A. Wardle. 2017. Elevation alters ecosystem properties across temperate

treelines globally. Nature Publishing Group 542:91–95.

Mayor, J. R., S. J. Wright, E. A. G. Schuur, M. E. Brooks, and B. L. Turner. 2014. Stable

nitrogen isotope patterns of trees and soils altered by long-term nitrogen and phosphorus

addition to a lowland tropical rainforest. Biogeochemistry 119:293–306.

Mcdowell, W. H., and C. E. Asbury. 1994. Export of carbon, nitrogen, and major ions from

three tropical montane watersheds. Limnology and Oceanography 39:111–125.

McGill, W. B., and C. V Cole. 1981. Comparative Aspects of Cycling of Organic C, N, S and P

through soil organic matter. Geoderma 26:267-286.

Literature cited

106

McGroddy, M. E., T. Daufresne, and O. L. Hedin. 2004. Scaling of C : N : P Stoichiometry in

Forests Worldwide : Implications of Terrestrial Redfield- Type Ratios. Ecology 85:2390–2401.

Menge, Duncan, N. L., and S. A. Levin. 2017. Spatial heterogeneity can resolve the nitrogen

paradox of tropical forests. Ecology 98:1049–1061.

Menge, D. N. L., and L. O. Hedin. 2009. Nitrogen fixation in different biogeochemical niches

along a 120 000-year chronosequence in New Zealand. Ecology 90:2190–201.

Menge, D. N. L., S. a Levin, and L. O. Hedin. 2009. Facultative versus obligate nitrogen fixation

strategies and their ecosystem consequences. The American Naturalist 174:465–77.

Menge, D. N. L., W. Troy Baisden, S. J. Richardson, D. a Peltzer, and M. M. Barbour. 2011.

Declining foliar and litter δ15N diverge from soil, epiphyte and input δ15N along a 120 000 yr

temperate rainforest chronosequence. The New phytologist 190:941–52.

Midgley, J. J., G. R. Van Wyk, and D. A. Everard. 1995. Leaf attributes of South African forest

species. African Journal of Ecology 33:160–168.

Miralles, D. G., T. R. H. Holmes, R. A. M. De Jeu, J. H. Gash, A. G. C. A. Meesters, and A. J.

Dolman. 2011. Global land-surface evaporation estimated from satellite-based observations.

Hydrology and Earth System Sciences 15:453–469.

Mueller, K. E., D. M. Eissenstat, S. E. Hobbie, J. Oleksyn, A. M. Jagodzinski, P. B. Reich, O.

a. Chadwick, J. Chorover, J. Oleksyn, A. M. Jagodzinski, P. B. Reich, O. a. Chadwick, and J.

Chorover. 2012. Tree species effects on coupled cycles of carbon, nitrogen, and acidity in

mineral soils at a common garden experiment. Biogeochemistry 111:601–614.

Mulder, J., H. A. De Wit, H. W. J. Boonen, and L. R. Bakken. 2001. Increased levels of

aluminium in forest soils: effects on the stores of soil organic carbon. Water, Air, and Soil

Pollution 130:160.

Mulvaney, R. L. 1996. Nitrogen - inorganic forms. Page 1123–1184. Methods of Soil Analysis.

Part 3. Chemical Methods.

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

and J. Kent. 2000. Biodiversity hotspots for conservation priorities. Nature 403:853–858.

Nabnera, G. E. L., K. E. Kschmittb, E. M. Atznerc, G. G. Uggenbergerd, M. von Lützow, I.

Kögel-Knabner, K. Ekschmitt, E. Matzner, G. Guggenberger, B. Marschner, and H. Flessa.

2006. Stabilization of organic matter in temperate soils: Mechanisms and their relevance under

different soil conditions - A review. European Journal of Soil Science 57:426–445.

Nakagawa, S., and H. Schielzeth. 2013. A general and simple method for obtaining R2 from

generalized linear mixed-effects models. Methods in Ecology and Evolution 4:133–142.

NASA. (n.d.). MODIS Collection 6 NRT Hotspot / Active Fire Detections MCD14DL. Available

on-line [https://earthdata.nasa.gov/firms].

Neill, C., L. A. Deegan, S. M. Thomas, and C. C. Cerri. 2001. Deforestation for pasture alters

nitrogen and phosphorus in small Amazonian streams. Ecological Applications 11:1817–1828.

Newbery, D. M., I. J. Alexander, and J. A. Rother. 1997. Phosphorus Dynamics in a Lowland

African Rain Forest : The Influence of Ectomycorrhizal Trees. Ecological Monographs 67:367–

409.

NRC. 2012. A Review of the U.S. Global Change Research Program’s Draft Strategic Plan.

Literature cited

107

Nyatsanga, T., and W. H. Pierre. 1973. Effect of Nitrogen Fixation by Legumes on Soil Acidity.

Agronomy journal 65:936-940.

Obbink, R. B., K. H. Icks, J. G. Alloway, T. S. Pranger, R. A. Lkemade, M. A. Shmore, M. B.

Ustamante, R. Bobbink, K. Hicks, J. Galloway, T. Spranger, R. Alkemade, M. Ashmore, M.

Bustamante, S. Cinderby, E. Davidson, F. Dentener, B. Emmett, J. W. Erisman, M. Fenn, F.

Gilliam, A. Norden, L. Pardo, and W. De Vries. 2010. Global assessment of nitrogen deposition

effects on terrestrial plant diversity. Ecological Applications 20:30–59.

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

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

Piao, A. Rautiainen, S. Sitch, D. Hayes, a D. McGuire, S. Piao, A. Rautiainen, S. Sitch, and

D. Hayes. 2011. A large and persistent carbon sink in the world’s forests. Science (New York,

N.Y.) 333:988–993.

Park, A., and J. L. Cameron. 2008. The influence of canopy traits on throughfall and stemflow

in five tropical trees growing in a Panamanian plantation. Forest Ecology and Management

255:1915–1925.

Pearson, H. L., and P. M. Vitousek. 2008. Stand Dynamics , Nitrogen Accumulation , and

Symbiotic Nitrogen Fixation in Regenerating Stands of Acacia koa. Ecological Applications

11:1381–1394.

Peel, B. L., B. L. Finlayson, and T. A. McMahon. 2007a. Updated world map of the Koppen-

Geiger climate classification.pdf. Hydrology and Earth System Sciences 11:1633–1644.

Peel, M. C., B. L. Finlayson, and T. A. Mcmahon. 2007b. Updated world map of the K ¨ oppen-

Geiger climate classification:1633–1644.

Peh, K. S. H., S. L. Lewis, and J. Lloyd. 2011a. Mechanisms of monodominance in diverse

tropical tree-dominated systems. Journal of Ecology 99:891–898.

Peh, K. S.-H., B. Sonke, J. Lloyd, C. a Quesada, and S. L. Lewis. 2011b. Soil Does Not Explain

Monodominance in a Central African Tropical Forest. PLos one 6:e16996.

Peñuelas, J., B. Poulter, J. Sardans, P. Ciais, M. van der Velde, L. Bopp, O. Boucher, Y.

Godderis, P. Hinsinger, J. Llusia, E. Nardin, S. Vicca, M. Obersteiner, and I. a Janssens. 2013.

Human-induced nitrogen-phosphorus imbalances alter natural and managed ecosystems

across the globe. Nature communications 4:2934.

Peñuelas, J., J. Sardans, A. Rivas-ubach, and I. a. Janssens. 2012. The human-induced

imbalance between C, N and P in Earth’s life system. Global Change Biology 18:3–6.

Perakis, S. S., and L. O. Hedin. 2002. Nitrogen loss from unpolluted South American forests

mainly via dissolved organic compounds. Nature 415:416–9.

Phillips, O., T. Baker, T. Feldpausch, and R. Brienen. 2009. RAINFOR field manual for plot

establishment and remeasurement.

Pinheiro, J., D. Bates, S. DebRoy, D. Sarkar, and T. R. C. Team. 2013. nlme: Linear and

Nonlinear Mixed Effects Models. R Core Development Team.

Poorter, H., U. Niinemets, L. Poorter, I. J. Wright, and R. Villar. 2009. Causes and

consequences of variation in leaf mass per area (LMA):a meta-analysis. New Phytologist

182:565–588.

Literature cited

108

Poorter, L., L. Bongers, and F. Bongers. 2006. Architecture of 54 moist-forest tree species:

Traits, trade-offs, and functional groups. Ecology 87:1289–1301.

Potter, C. S., P. a. Matson, P. M. Vitousek, and E. a. Davidson. 1996. Process modeling of

controls on nitrogen trace gas emissions from soils worldwide. Journal of Geophysical

Research-Atmospheres 101:1361–1377.

Potvin, C., Lady Mancilla, N. Buchmann, J. Monteza, T. Moore, M. Murphy, Y. Oelmann, M.

Scherer-Lorenzen, B. L. Turner, W. Wilcke, F. Zeugin, and S. Wolf. 2011. An ecosystem

approach to biodiversity effects: Carbon pools in a tropical tree plantation. Forest Ecology and

Management 261:1614–1624.

Prentice, I. C., R. J. Scholes, D. W. R. Wallace, D. Archer, M. R. Ashmore, O. Aumont, D.

Baker, M. Battle, M. Bender, L. P. Bopp, P. Bousquet, K. Caldeira, P. Ciais, P. M. Cox, W.

Cramer, F. Dentener, I. G. Enting, C. B. Field, P. Friedlingstein, E. A. Holland, R. A. Houghton,

J. I. House, A. Ishida, A. K. Jain, I. A. Janssens, F. Joos, T. Kaminski, C. D. Keeling, R. F.

Keeling, D. W. Kicklighter, K. E. Kohfeld, W. Knorr, R. Law, T. Lenton, K. Lindsay, A. C.

Manning, R. J. Matear, A. D. Mcguire, J. M. Melillo, R. Meyer, M. Mund, J. C. Orr, S. Piper, K.

Plattner, P. J. Rayner, S. Sitch, R. Slater, S. Taguchi, P. P. Tans, H. Q. Tian, M. F. Weirig, T.

Whorf, and A. Yool. 2001. Working Group I: The Scientific Basis. The Carbon Cycle and

Atmospheric Carbon Dioxide.

Le Quéré, C., R. M. Andrew, J. G. Canadell, S. Sitch, J. Ivar Korsbakken, G. P. Peters, A. C.

Manning, T. A. Boden, P. P. Tans, R. A. Houghton, R. F. Keeling, S. Alin, O. D. Andrews, P.

Anthoni, L. Barbero, L. Bopp, F. Chevallier, L. P. Chini, P. Ciais, K. Currie, C. Delire, S. C.

Doney, P. Friedlingstein, T. Gkritzalis, I. Harris, J. Hauck, V. Haverd, M. Hoppema, K. Klein

Goldewijk, A. K. Jain, E. Kato, A. Körtzinger, P. Landschützer, N. Lefevre, A. Lenton, S.

Lienert, D. Lombardozzi, J. R. Melton, N. Metzl, F. Millero, P. M. S. Monteiro, D. R. Munro, J.

E. M. S. Nabel, S. I. Nakaoka, K. O’Brien, A. Olsen, A. M. Omar, T. Ono, D. Pierrot, B. Poulter,

C. Rödenbeck, J. Salisbury, U. Schuster, J. Schwinger, R. Séférian, I. Skjelvan, B. D. Stocker,

A. J. Sutton, T. Takahashi, H. Tian, B. Tilbrook, I. T. Van Der Laan-Luijkx, G. R. Van Der Werf,

N. Viovy, A. P. Walker, A. J. Wiltshire, and S. Zaehle. 2016. Global Carbon Budget 2016.

Earth System Science Data 8:605–649.

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

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

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

Johannessen, E. Kato, R. F. Keeling, V. Kitidis, K. Klein Goldewijk, C. Koven, C. S. Landa, P.

Landschützer, A. Lenton, I. D. Lima, G. Marland, J. t. Mathis, N. Metzl, Y. Nojiri, A. Olsen, T.

Ono, S. Peng, W. Peters, B. Pfeil, B. Poulter, M. R. Raupach, P. Regnier, C. Rödenbeck, S.

Saito, J. E. Salisbury, U. Schuster, J. Schwinger, R. Séférian, J. Segschneider, T. Steinhoff,

B. D. Stocker, A. J. Sutton, T. Takahashi, B. Tilbrook, G. R. van der Werf, N. Viovy, Y. P.

Wang, R. Wanninkhof, A. Wiltshire, and N. Zeng. 2015. Global carbon budget 2014. Earth

System Science Data 7:47–85.

Quesada, C. a., O. L. Phillips, M. Schwarz, C. I. Czimczik, T. R. Baker, S. Patiño, N. M. Fyllas,

M. G. Hodnett, R. Herrera, S. Almeida, E. Alvarez Dávila, a. Arneth, L. Arroyo, K. J. Chao, N.

Dezzeo, T. Erwin, a. di Fiore, N. Higuchi, E. Honorio Coronado, E. M. Jimenez, T. Killeen, a.

T. Lezama, G. Lloyd, G. López-González, F. J. Luizão, Y. Malhi, a. Monteagudo, D. a. Neill,

P. Núñez Vargas, R. Paiva, J. Peacock, M. C. Peñuela, a. Peña Cruz, N. Pitman, N. Priante

Filho, a. Prieto, H. Ramírez, a. Rudas, R. Salomão, a. J. B. Santos, J. Schmerler, N. Silva,

M. Silveira, R. Vásquez, I. Vieira, J. Terborgh, and J. Lloyd. 2012. Basin-wide variations in

Literature cited

109

Amazon forest structure and function are mediated by both soils and climate. Biogeosciences

9:2203–2246.

R Development Core Team. 2015. R: A language and environment for statistical computing,

http://www.r-project.org. Vienna, Austria, http://www.r-project.org, Vienna, Austria,

http://www.r-project.org.

Raich, J. W., D. A. Clark, L. Schwendenmann, and T. E. Wood. 2014. Aboveground tree

growth varies with belowground carbon allocation in a tropical rainforest environment. PloS

one 9:e100275.

Van Ranst, E., G. Baert, M. Ngongo, and P. Mafuka. 2010. Carte pédologique de Yangambi,

planchette 2: Yangambi, échelle 1:50.000. UGent, Hogent, UNILU, UNIKIN.

Van Ranst, E., M. Verloo, A. Demeyer, and J. Pauwels. 1999. Manual for the Soil Chemistry

and Fertility Laboratory. University of Ghent, Ghent, Belgium.

Reay, D. S., F. Dentener, P. Smith, J. Grace, and R. a. Feely. 2008. Global nitrogen deposition

and carbon sinks. Nature Geoscience 1:430–437.

Redondo-Brenes, A. 2007. Growth, carbon sequestration, and management of native tree

plantations in humid regions of Costa Rica. New Forests 34:253–268.

Reed, S. C., C. C. Cleveland, and A. R. Townsend. 2007. Controls Over Leaf Litter and Soil

Nitrogen Fixation in Two Lowland Tropical Rain Forests. Biotropica 39:585–592.

Reed, S. C., C. C. Cleveland, and A. R. Townsend. 2008. Tree species control rates of free-

living nitrogen fixation in a tropical rain forest. Ecology 89:2924–34.

Reed, S. C., C. C. Cleveland, and A. R. Townsend. 2011. Functional Ecology of Free-Living

Nitrogen Fixation: A Contemporary Perspective. Annual Review of Ecology, Evolution, and

Systematics 42:489–512.

Reed, S. C., C. C. Cleveland, and A. R. Townsend. 2013. Relationships among phosphorus,

molybdenum and free-living nitrogen fixation in tropical rain forests: results from observational

and experimental analyses. Biogeochemistry 114:135–147.

Reed, S. C., X. Yang, and P. E. Thornton. 2015. Incorporating phosphorus cycling into global

modeling efforts: A worthwhile, tractable endeavor. New Phytologist 208:324–329.

Reich, P. B., J. Oleksyn, J. Modrzynski, P. Mrozinski, S. E. Hobbie, D. M. Eissenstat, J.

Chorover, O. a. Chadwick, C. M. Hale, and M. G. Tjoelker. 2005. Linking litter calcium,

earthworms and soil properties: A common garden test with 14 tree species. Ecology Letters

8:811–818.

Reich, P. B., J. Oleksyn, and I. J. Wright. 2009. Leaf phosphorus influences the

photosynthesis-nitrogen relation: A cross-biome analysis of 314 species. Oecologia 160:207–

212.

Reich, P. B., M. B. Walters, D. S. Ellsworth, N. B. Columbia, P. George, and B. Columbia.

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.

Resh, S. C., D. Binkley, and J. a. Parrotta. 2002. Greater Soil Carbon Sequestration under

Nitrogen-fixing Trees Compared with Eucalyptus Species. Ecosystems 5:217–231.

Literature cited

110

Ruiz-Jaen, M. C., and C. Potvin. 2011. Can we predict carbon stocks in tropical ecosystems

from tree diversity? Comparing species and functional diversity in a plantation and a natural

forest. New Phytologist 189:978–987.

Russell, a. E., J. W. Raich, O. J. Valverde-Barrantes, and R. F. Fisher. 2007. Tree Species Effects on Soil Properties in Experimental Plantations in Tropical Moist Forest. Soil Science Society of America Journal 71:1389.

Rütting, T., L. N. Cizungu, D. Roobroeck, M. Bauters, D. Huygens, and P. Boeckx. 2014. Leaky nitrogen cycle in pristine African montane Rainforest soil. Global Biogeochemical Cycles 29:962–973.

Rütting, T., D. Huygens, J. Staelens, C. Müller, and P. Boeckx. 2011a. Advances in 15N-tracing experiments: new labelling and data analysis approaches. Biochemical Society transactions 39:279–83.

Rütting, T., P. Boeckx, C. Müller, and L. Klemedtsson. 2011b. Assessment of the importance of dissimilatory nitrate reduction to ammonium for the terrestrial nitrogen cycle. Biogeosciences 8:1779–1791.

Ryan, J., G. Estefan, and A. Rashid. 2001. Soil and Plant Analysis Laboratory Manual. Second Edition. ICARDA, Aleppo, Syria, Syria.

Saggar, S., M. J. Hedley, and R. E. White. 1990. A simplified resin membrane technique for

extracting phosphorus from soils. Fertilizer Research 24:173–180.

Santiago, L. S. 2015. Nutrient limitation of eco-physiological processes in tropical trees. Trees

- Structure and Function 29:1291–1300.

Scaranello, M. a., L. F. Alves, S. A. Vieira, and P. B. De Camargo. 2012. Height-diameter

relationships of tropical Atlantic moist forest trees in southeastern. Scientia Agricola 69:26–

37.

Scherer-Lorenzen, M., C. Potvin, J. Koricheva, B. Schmid, A. Hector, Z. Bornik, G. Reynolds,

and E. D. Schulze. 2005. The design of experimental tree plantations for functional biodiversity

research. Pages 347–376in M. Scherer-Lorenzen, C. Körner, and E. D. Schulze,

editors.Forest Diversity and Function. Springer Berlin, Heidelberg.

Schimel, J. P., and J. Bennett. 2004. Nitrogen Mineralization: Challenges of a changing

Paradigm. Ecology 85:591–602.

Schmidt, M. W. I., M. S. Torn, S. Abiven, T. Dittmar, G. Guggenberger, I. a. Janssens, J.

Lehmann, D. a. C. Manning, P. Nannipieri, D. P. Rasse, S. Weiner, and S. E. Trumbore. 2011.

Persistence of soil organic matter as an ecosystem property. Nature 478:49–56.

Schneider, C. a, W. S. Rasband, and K. W. Eliceiri. 2012. NIH Image to ImageJ: 25 years of

image analysis. Nature Methods 9:671–675.

De Schrijver, A., P. De Frenne, J. Staelens, G. Verstraeten, B. Muys, L. Vesterdal, K. Wuyts,

L. van Nevel, S. Schelfhout, S. De Neve, and K. Verheyen. 2012. Tree species traits cause

divergence in soil acidification during four decades of postagricultural forest development.


De Schrijver, A., L. Nachtergale, J. Staelens, A. De Schrijver, L. Nachtergale, J. Staelens, S.

Luyssaert, and L. De Keersmaeker. 2004. Comparison of throughfall and soil solution

chemistry between a high-density Corsican pine stand and a naturally regenerated silver birch

stand. Environmental Pollution 131:93–105.

Literature cited

111

Schwendenmann, L., and E. Veldkamp. 2005. The Role of Dissolved Organic Carbon,

Dissolved Organic Nitrogen , and Dissolved Inorganic Nitrogen in a Tropical Wet Forest

Ecosystem 8:339–351.

Scripps. 2014. Keeling Curve. Scripps CO2 Program 391.

Selaya, N. G., and N. P. R. Anten. 2008. Differences in biomass allocation, light interception

and mechanical stability between lianas and trees in early secondary tropical forest. Functional

Ecology 22:30–39.

Sharp, J. H., K. R. Rinker, K. B. Savidge, J. Abell, J. Yves Benaim, D. Bronk, D. J. Burdige,

G. Cauwet, W. Chen, M. D. Doval, D. Hansell, C. Hopkinson, G. Kattner, N. Kaumeyer, K. J.

McGlathery, J. Merriam, N. Morley, K. Nagel, H. Ogawa, C. Pollard, M. Pujo-Pay, P.

Raimbault, R. Sambrotto, S. Seitzinger, G. Spyres, F. Tirendi, T. W. Walsh, and C. S. Wong.

2002. A preliminary methods comparison for measurement of dissolved organic nitrogen in

seawater. Marine Chemistry 78:171–184.

Shi, S., W. Zhang, P. Zhang, Y. Yu, and F. Ding. 2013. A synthesis of change in deep soil organic carbon stores with afforestation of agricultural soils. Forest Ecology and Management 296:53–63.

Silver, W. L., A. W. Thompson, A. Reich, J. J. Ewel, and M. K. Firestone. 2005. Nitrogen cycling in tropical plantation forests: potential controls on nitrogen retention. Ecological Applications 15:1604–1614.

Silver, W. L., D. J. Herman, M. K. Firestone, and N. Sep. 2012. Dissimilatory Nitrate Reduction to Ammonium in Upland Tropical Forest Soils. Ecology 82:2410–2416.

Slik, J. W. F., V. Arroyo-rodríguez, S. Aiba, P. Alvarez-loayza, L. F. Alves, P. Ashton, P. Balvanera, M. L. Bastian, P. J. Bellingham, E. Van Den Berg, L. Bernacci, C. Bispo, L. Blanc, K. Böhning-gaese, P. Boeckx, F. Bongers, B. Boyle, M. Bradford, F. Q. Brearley, M. B. Hockemba, S. Bunyavejchewin, D. Calderado, L. Matos, M. Castillo-santiago, E. L. M. Catharino, S. Chai, Y. Chen, R. K. Colwell, C. L. Robin, C. Clark, D. B. Clark, D. a Clark, H. Culmsee, K. Damas, H. S. Dattaraja, G. Dauby, P. Davidar, S. J. Dewalt, J. Doucet, A. Duque, G. Durigan, K. a O. Eichhorn, P. V Eisenlohr, E. Eler, C. Ewango, N. Farwig, K. J. Feeley, L. Ferreira, R. Field, A. T. De Oliveira, C. Fletcher, O. Forshed, G. Franco, G. Fredriksson, T. Gillespie, J. Gillet, G. Amarnath, D. M. Griffith, J. Grogan, N. Gunatilleke, D. Harris, R. Harrison, A. Hector, J. Homeier, N. Imai, A. Itoh, P. a Jansen, C. a Joly, B. H. J. De Jong, K. Kartawinata, E. Kearsley, D. L. Kelly, D. Kenfack, M. Kessler, K. Kitayama, R. Kooyman, E. Larney, Y. Laumonier, S. Laurance, W. F. Laurance, M. J. Lawes, I. Leao, S. G. Letcher, J. Lindsell, X. Lu, A. Mansor, A. Marjokorpi, E. H. Martin, H. Meilby, F. P. L. Melo, D. J. Metcalfe, V. P. Medjibe, J. Paul, J. Millet, D. Mohandass, J. C. Montero, M. De Morisson, B. Mugerwa, H. Nagamasu, R. Nilus, S. Ochoa-gaona, N. Page, P. Parolin, M. Parren, N. Parthasarathy, E. Paudel, A. Permana, M. T. F. Piedade, N. C. a Pitman, L. Poorter, A. D. Poulsen, J. Poulsen, J. Powers, R. C. Prasad, J. Puyravaud, J. Razafimahaimodison, J. Reitsma, J. Roberto, W. Roberto, H. Romero-saltos, F. Rovero, A. Hamuraby, K. Ruokolainen, E. Rutishauser, F. Saiter, P. Saner, B. a Santos, F. Santos, S. K. Sarker, M. Satdichanh, C. B. Schmitt, J. Schöngart, M. Schulze, M. S. Suganuma, D. Sheil, S. Pinheiro, P. Sist, T. Stevart, R. Sukumar, I. Sun, T. Sunderand, H. S. Suresh, E. Suzuki, M. Tabarelli, J. Tang, N. Targhetta, I. Theilade, D. W. Thomas, P. Tchouto, J. Hurtado, R. Valencia, J. L. C. H. Van Valkenburg, T. Van Do, R. Vasquez, H. Verbeeck, V. Adekunle, S. a Vieira, C. O. Webb, T. Whitfeld, S. a Wich, J. Williams, F. Wittmann, H. Wöll, X. Yang, C. Y. Adou, S. L. Yap, T. Yoneda, R. a Zahawi, R. Zakaria, R. Zang, R. Assis, B. G. Luize, and E. M. Venticinque. 2015a. An estimate of the nmber of tropical tree species. Proceedings of the National Academy of Sciences 112:7472–7477.

Literature cited

112

Smith, B., D. Wärlind, A. Arneth, T. Hickler, P. Leadley, J. Siltberg, and S. Zaehle. 2014.

Implications of incorporating N cycling and N limitations on primary production in an individual-

based dynamic vegetation model. Biogeosciences 11:2027–2054.

Soethe, N., J. Lehmann, and C. Engels. 2008. Nutrient availability at different altitudes in a

tropical montane forest in Ecuador. Journal of Tropical Ecology 24:397–406.

Sokolov, A. P., D. W. Kicklighter, J. M. Melillo, B. S. Felzer, C. A. Schlosser, and T. W. Cronin.

2008. Consequences of considering carbon-nitrogen interactions on the feedbacks between

climate and the terrestrial carbon cycle. Journal of Climate 21:3776–3796.

Sollins, P. 1998. Factors influencing species composition in tropical lowland rain forest: Does

soil matter? Ecology 79:23–30.

Sotta, E. D., M. D. Corre, and E. Veldkamp. 2008. Differing N status and N retention processes

of soils under old-growth lowland forest in Eastern Amazonia, Caxiuan??, Brazil. Soil Biology

and Biochemistry 40:740–750.

Spencer, R. G. M., W. Guo, P. A. Raymond, T. Dittmar, E. Hood, J. Fellman, and A. Stubbins.

2014. Source and biolability of ancient dissolved organic matter in glacier and lake ecosystems

on the Tibetan Plateau. Geochimica et Cosmochimica Acta 142:64–74.

Sprent, J. 2005. West African legumes: the role of nodulation and nitrogen fixation. The New

phytologist 167:326–30.

Sprent, J. I. 2001. Nodulation in Legumes. London: Royal Botanic Gardens, Kew.

Sprent, J. I. 2009. Legume Nodulation: A Global Perspective. Page Wiley-Blackwell. Wiley-

Blackwell, Oxford, UK.

Staelens, J., D. Houle, and A. De Schrijver. 2008. Calculating Dry Deposition and Canopy

Exchange with the Canopy Budget Model : Review of Assumptions and Application to Two

Deciduous Forests. Water Air Soil Pollution 191:149–169.

Staelens, J., T. Rütting, D. Huygens, A. De Schrijver, C. Müller, K. Verheyen, and P. Boeckx.

2011. In situ gross nitrogen transformations differ between temperate deciduous and

coniferous forest soils. Biogeochemistry 108:259–277.

Stanley, W. G., and F. Montagnini. 1999. Biomass and nutrient accumulation in pure and

mixed plantations of indigenous tree species grown on poor soils in the humid tropics of Costa

Rica. Forest Ecology and Management 113:91–103.

Stein, A. F., R. R. Draxler, G. D. Rolph, B. J. B. Stunder, M. D. Cohen, and F. Ngan. 2015.

Noaa’s hysplit atmospheric transport and dispersion modeling system. Bulletin of the

American Meteorological Society 96:2059–2077.

Stevens, R. J., and R. J. Laughlin. 1994. Determining Nitrogen-15 in Nitrite or Nitrate by

Producing Nitrous Oxide. Soil Science Society of America Journal 58:1108.

Stickler, C. M., D. C. Nepstad, M. T. Coe, D. G. McGrath, H. O. Rodrigues, W. S. Walker, B.

S. Soares-Filho, and E. a. Davidson. 2009. The potential ecological costs and cobenefits of

REDD: a critical review and case study from the Amazon region. Global Change Biology

15:2803–2824.

Stocker, B. D., I. C. Prentice, S. E. Cornell, T. Davies-Barnard, A. C. Finzi, O. Franklin, I.

Janssens, T. Larmola, S. Manzoni, T. Näsholm, J. A. Raven, K. T. Rebel, S. Reed, S. Vicca,

Literature cited

113

A. Wiltshire, S. Zaehle, and D. Hall. 2016. Terrestrial nitrogen cycling in Earth system models

revisited. New Phytologist 210:1165–1168.

Sullivan, B. W., W. K. Smith, A. R. Townsend, M. K. Nasto, S. C. Reed, R. L. Chazdon, and

C. C. Cleveland. 2014. Spatially robust estimates of biological nitrogen (N) fixation imply

substantial human alteration of the tropical N cycle. Proceedings of the National Academy of

Sciences of the United States of America 111:8101–6.

Sundqvist, M. K., N. J. Sanders, and D. a. Wardle. 2013. Community and Ecosystem

Responses to Elevational Gradients: Processes, Mechanisms, and Insights for Global

Change. Annual Review of Ecology, Evolution, and Systematics 44:261–280.

Tanner, A. E. V. J., P. M. Vitousek, and E. Cuevas. 1998. Experimental Investigation of

Nutrient Limitation of Forest Growth on Wet Tropical Mountains. Ecology 79:10–22.

Taylor, B. N., R. L. Chazdon, B. Bachelot, and D. N. L. Menge. 2017. Nitrogen-fixing trees

inhibit growth of regenerating Costa Rican rainforests. Proceedings of the National Academy

of Sciences 114:201707094.

Taylor, P. G., W. R. Wieder, S. Weintraub, S. Cohen, C. C. Cleveland, and A. R. Townsend.

2015. Organic forms dominate hydrologic nitrogen export from a lowland tropical watershed.

Ecology 96:1229–1241.

Tedersoo, L., T. Naadel, M. Bahram, K. Pritsch, F. Buegger, M. Leal, U. Kõljalg, and K.

Põldmaa. 2012. Enzymatic activities and stable isotope patterns of ectomycorrhizal fungi in

relation to phylogeny and exploration types in an afrotropical rain forest. New Phytologist

195:832–843.

Templer, P. H., W. L. Silver, J. Pett-Ridge, K. M. DeAngelis, and M. K. Firestone. 2008. Plant

and microbial controls on nitrogen retention and loss in a humid tropical forest. Ecology

89:3030–3040.

Thomas, R. Q., E. N. J. Brookshire, and S. Gerber. 2015. Nitrogen limitation on land: How can

it occur in Earth system models? Global Change Biology 21:1777–1793.

Thornton, P. E., S. C. Doney, K. Lindsay, J. K. Moore, N. Mahowald, J. T. Randerson, I. Fung,

J.-F. Lamarque, J. J. Feddema, and Y.-H. Lee. 2009. Carbon-nitrogen interactions regulate

climate-carbon cycle feedbacks: results from an atmosphere-ocean general circulation model.

Biogeosciences 6:2099–2120.

Thornton, P. E., J. F. Lamarque, N. A. Rosenbloom, and N. M. Mahowald. 2007. Influence of

carbon-nitrogen cycle coupling on land model response to CO2 fertilization and climate

variability. Global Biogeochemical Cycles 21:1–15.

Tilman, D., P. B. Reich, and J. M. H. Knops. 2006. Biodiversity and ecosystem stability in a

decade-long grassland experiment. Nature 441:629–632.

Tobón, C., J. Sevink, and J. M. Verstraten. 2004. Solute fluxes in throughfall and stemflow in

four forest ecosystems in northwest Amazonia. Biogeochemistry 70:1–25.

Toledo, M., M. Peña-Claros, F. Bongers, A. Alarcón, J. Balcázar, J. Chuviña, C. Leaño, J. C.

Licona, and L. Poorter. 2012. Distribution patterns of tropical woody species in response to

climatic and edaphic gradients. Journal of Ecology 100:253–263.

Literature cited

114

Torti, S. D., P. D. Coley, T. A. Kursar, S. D. Torti, P. D. Coley, and T. A. Kursar. 2001. Causes

and Consequences of Monodominance in Tropical Lowland Forests. The American Naturalist

157:141–153.

Townsend, A. R., G. P. Asner, and C. C. Cleveland. 2008. The biogeochemical heterogeneity

of tropical forests. Trends in Ecology and Evolution 23:424–431.

Trimblet, M. J., and R. J. van Aarde. 2012. Geographical and taxonomic biases in research

on biodiversity in human-modified landscapes. Ecosphere 3:1–16.

USDA. 1999. Soil Taxonomy, A basic system of soil classification for making and interpreting

soil surveys.

Valladares, F., and Ü. Niinemets. 2008. Shade tolerance, a key plant feature of complex

nature and consequences. Annual Review of Ecology, Evolution, and Systematics 39:237–

257.

Valverde-barrantes, O. J., K. a Smemo, and C. B. Blackwood. 2015. Fine root morphology is

phylogenetically structured, but nitrogen is related to the plant economics spectrum in

temperate trees. Functional Ecology 29:796–807.

Verbeeck, H., P. Boeckx, and K. Steppe. 2011. Tropical forests : include Congo basin. Nature

479:2011.

Verheyen, K., M. Vanhellemont, H. Auge, L. Baeten, C. Baraloto, N. Barsoum, S. Bilodeau-

Gauthier, H. Bruelheide, B. Castagneyrol, D. Godbold, H. Josephine, A. Hector, H. Jactel, J.

Koricheva, M. Loreau, S. Mereu, C. Messier, B. Muys, P. Nolet, A. Paquette, J. Parker, M.

Perring, Q. Ponette, C. Potvin, P. Reich, A. Smith, and M. Scherer-Lorenzen. 2015.

Contributions of a global network of tree diversity experiments to sustainable forest

plantations. Ambio 45:29–41.

Vesterdal, L., N. Clarke, B. D. Sigurdsson, and P. Gundersen. 2013. Do tree species influence

soil carbon stocks in temperate and boreal forests? Forest Ecology and Management 309:4–

18.

Vesterdal, L., I. K. Schmidt, I. Callesen, L. O. Nilsson, and P. Gundersen. 2008. Carbon and

nitrogen in forest floor and mineral soil under six common European tree species. Forest

Ecology and Management 255:35–48.

Vet, R., R. S. Artz, S. Carou, M. Shaw, C. U. Ro, W. Aas, A. Baker, V. C. Bowersox, F.

Dentener, C. Galy-Lacaux, A. Hou, J. J. Pienaar, R. Gillett, M. C. Forti, S. Gromov, H. Hara,

T. Khodzher, N. M. Mahowald, S. Nickovic, P. S. P. Rao, and N. W. Reid. 2014. A global

assessment of precipitation chemistry and deposition of sulfur, nitrogen, sea salt, base

cations, organic acids, acidity and pH, and phosphorus. Atmospheric Environment 93:3–100.

Vicca, S., S. Luyssaert, J. Peñuelas, M. Campioli, F. S. Chapin, P. Ciais, a. Heinemeyer, P.

Högberg, W. L. Kutsch, B. E. Law, Y. Malhi, D. Papale, S. L. Piao, M. Reichstein, E. D.

Schulze, and I. A. Janssens. 2012. Fertile forests produce biomass more efficiently. Ecology

Letters 15:520–526.

Vitousek, P. M. 1984. Litterfall , Nutrient Cycling , and Nutrient Limitation in Tropical Forests.

Ecology 65:285–298.

Vitousek, P. M., and P. Matson. 1988. Nitrogen Transformations in a Range of Tropical Forest

Soils. Soil Biology and Biochemistry 20:361–367.

Literature cited

115

Vitousek, P. M., S. Porder, B. Z. Houlton, and O. a Chadwick. 2010. Terrestrial phosphorus

limitation : mechanisms , implications , and nitrogen – phosphorus interactions. Ecological

Applications, 20:5–15.

Vitousek, P. M., and R. L. Sanford. 1986. Nutrient Cycling in Moist Tropical Forest. Annual

Review of Ecology and Systematics 17:137–167.

Vitousek, P. M., G. Shearer, and D. H. Kohl. 1989. Foliar 15N natural abundance in Hawaiian

rainforest: patterns and possible mechanisms. Oecologia 78:383–388.

Vitousek, P. M., D. R. Turner, and K. Kitayama. 1995. Foliar Nutrients During Long-Term Soil

Development in Hawaiian Montane Rain Forest. Ecology 76:712–720.

Walker, T. W., and J. K. Syers. 1976. The fate of phosphorus during pedogenesis. Geoderma

15:1–19.

Wanek, W., M. Mooshammer, A. Blöchl, A. Hanreich, and A. Richter. 2010. Determination of

gross rates of amino acid production and immobilization in decomposing leaf litter by a novel 15N isotope pool dilution technique. Soil Biology and Biochemistry 42:1293–1302.

Wang, R., D. Goll, Y. Balkanski, D. Hauglustaine, O. Boucher, P. Ciais, I. Janssens, J.

Penuelas, B. Guenet, J. Sardans, L. Bopp, N. Vuichard, F. Zhou, B. Li, S. Piao, S. Peng, Y.

Huang, and S. Tao. 2017. Global forest carbon uptake due to nitrogen and phosphorus

deposition from 1850 to 2100. Global change biology 38:42–49.

Wang, Y. P., R. M. Law, and B. Pak. 2010. A global model of carbon, nitrogen and phosphorus

cycles for the terrestrial biosphere. Biogeosciences 7:2261–2282.

Waring, A. B. G., L. Álvarez-Cansino, K. Barry, K. K. Becklund, M. G. Gei, O. Lopez, L.

Markesteijn, S. Mangan, M. E. Rodriguez, R. M. Segnitz, S. a Schnitzer, and J. S. Powers.

2015. Viva Zinke! Pervasive and strong effects of plant individuals and species on soil

chemistry: a meta-analysis. Proceedings of the Royal Society B 282.

Van de Weg, M. J., P. Meir, J. Grace, O. K. Atkin, M. J. Van de Weg, P. Meir, J. Grace, and

O. K. Atkin. 2009. Altitudinal variation in leaf mass per unit area, leaf tissue density and foliar

nitrogen and phosphorus content along an Amazon-Andes gradient in Peru. Plant Ecology &

Diversity 2:243–254.

Weintraub, S. R., R. J. Cole, C. G. Schmitt, and J. D. All. 2016. Climatic controls on the isotopic

composition and availability of soil nitrogen across mountainous tropical forest. Ecosphere

7:1–13.

Weintraub, S. R., P. G. Taylor, S. P. Porder, C. C. Cleveland, G. P. Asner, A. R. Townsend,

and P. Asner, Gregory. 2014. Topographic controls on soil nitrogen availability in a lowland

tropical forest. Ecology 96:1561–1574.

van der Werf, G. R., J. T. Randerson, L. Giglio, G. J. Collatz, P. S. Kasibhatla, and J. Arellano,

A. F. 2006. Interannual variability in global biomass burning emissions from 1997 to 2004.

Atmospheric Chemistry and Physics 6:3423–3441.

Wieder, W. R., C. C. Cleveland, W. K. Smith, and K. Todd-Brown. 2015. Future productivity

and carbon storage limited by terrestrial nutrient availability. Nature Geoscience 8:441-445.

Wieder, W. R., C. C. Cleveland, P. G. Taylor, D. R. Nemergut, E. L. Hinckley, L. Philippot, D.

Bru, S. R. Weintraub, M. Martin, and A. R. Townsend. 2013. Experimental removal and

Literature cited

116

addition of leaf litter inputs reduces nitrate production and loss in a lowland tropical forest.

Biogeochemistry 113:629–642.

Wright, I. J., P. B. Reich, M. Westoby, D. D. Ackerly, Z. Baruch, F. Bongers, J. Cavender-

bares, T. Chapin, J. H. C. Cornelissen, M. Diemer, J. Flexas, E. Garnier, P. K. Groom, and J.

Gulias. 2004. The worldwide leaf economics spectrum. Nature 12:821–827.

Xu, J. M., C. Tang, and Z. L. Chen. 2006. The role of plant residues in pH change of acid soils

differing in initial pH. Soil Biology and Biochemistry 38:709–719.

Yang, W. H., D. Liptzin, and J. B. Yavitt. 2015. High potential for iron reduction in upland soils.

Ecology 96:2015–2020.

Yang, W. H., K. A. Weber, and W. L. Silver. 2012. Nitrogen loss from soil through anaerobic

ammonium oxidation coupled to iron reduction. Nature Geoscience 5:538–541.

Yang, X., P. E. Thornton, D. M. Ricciuto, and W. M. Post. 2014. The role of phosphorus

dynamics in tropical forests – a modeling study using CLM-CNP. Biogeosciences 11:1667–

1681.

Yu, H., M. Chin, T. Yuan, H. Bian, L. a. Remer, J. M. Prospero, A. Omar, D. Winker, Y. Yang,

Y. Zhang, Z. Zhang, and C. Zhao. 2015. The Fertilizing Role of African Dust in the Amazon

Rainforest: A First Multiyear Assessment Based on CALIPSO Lidar Observations.

Geophysical Research Letters 42:1984-1991.

Zaehle, S., and A. D. Friend. 2010. Carbon and nitrogen cycle dynamics in the O-CN land

surface model: 1. Model description, site-scale evaluation, and sensitivity to parameter

estimates. Global Biogeochemical Cycles 24:1–13.

Zanne, A. E., G. Lopez-Gonzalez, D. A. A. Coomes, J. Ilic, S. Jansen, S. L. S. L. Lewis, R. B.

B. Miller, N. G. G. Swenson, M. C. C. Wiemann, and J. Chave. 2009. Data from: Towards a

worldwide wood economics spectrum. Dryad Digital Repository. Page Dryad.

Zhang, Y., M. Xu, H. Chen, and J. Adams. 2009. Global pattern of NPP to GPP ratio derived

from MODIS data: Effects of ecosystem type, geographical location and climate. Global

Ecology and Biogeography 18:280–290.

Zinke, P. J. 1962. The Pattern of Influence of Individual Forest Trees on Soil Properties.

Ecology 43:130–133.

Zuur, A. F., E. N. Ieno, and C. S. Elphick. 2010. A protocol for data exploration to avoid

common statistical problems. Methods in Ecology and Evolution 1:3–14.

117

Appendices

Appendices

118

Appendix A Appendix A1 List of species occurrence and nodulation in the different forest succession stages in the study area. Appendix A2 Legume composition of the different successional stages. Appendix A1. List of species occurrence and nodulation in the different forest succession stages in the study area. † 1 = 0-5 yr. Fallow; 2 = 5-10 yr. Fallow; 3 = Secondary forest; 4 = Primary forest; n.p., not present; ‡ +, known to nodulate; -, thought to be non-nodulating; *, no literature records found; **, literature check not possible (not identified to species level); a, Allen & Allen 1981;b, Corby 1974; c, Diabate et al. 2005; d, Högberg & Alexander 1995;e, Sprent 2001; f, Bonnier 1958.

Species Successional Stage† Literature‡

1 2 3 4

Nodulated fraction

Abrus precatorius L. 0.76 0.32 n.p. n.p. + a Albizia spp. 0.48 0.11 n.p. n.p. ** ** Albizia adianthifolia (Schum.) W.Wight n.p. 0.14 0.01 0.00 + b Albizia ferruginea (Guill. & Perr.) Benth. n.p. 0.08 0.00 n.p. + c Albizia glaberrima (Schum. & Thonn.) Benth.. n.p. n.p. 0.00 n.p. + b Albizia gummifera (J.F.Gmel.) C.A.Sm. n.p. 0.10 0.02 0.00 + b Albizia laurentii De Wild n.p. 0.11 n.p. n.p. * * Alysicarpus rugosus (Willd.) DC 0.64 n.p. n.p. n.p. + b Amphimas pterocarpoides Harms n.p. n.p. 0.06 0.00 + c Anthonotha fragrans (Baker f.) Exell & Hillc. n.p. n.p. n.p. 0.00 - c Baikiaea spp. n.p. 0.05 n.p. n.p. ** ** Baikiaea robynsii Ghesq. 0.93 0.00 0.00 0.00 * * Baphia spp. n.p. 0.00 n.p. n.p. ** ** Baphia laurifolia Baill. n.p. n.p. 0.00 0.00 + d Canavalia ensiformis (L.) DC. 0.85 n.p. n.p. n.p. + a Cynometra alexandri C.H.Wright n.p. n.p. 0.00 0.00 - e Cynometra lujae De Wild. n.p. n.p. 0.00 0.00 - e Dalbergia spp. 0.15 0.00 n.p. n.p. ** ** Dalbergia lactea Vatke 0.28 0.03 n.p. n.p. + b Desmodium adscendens (Sw.) DC. 0.92 n.p. n.p. n.p. + a Dialium pachyphyllum Harms n.p. n.p. 0.00 0.00 - f Gilbertiodendron dewevrei (De Wild.) J.Leonard n.p. n.p. n.p. 0.00 - f Neonotonia wightii (Wight & Arn.) J.A. Lackey 0.81 n.p. n.p. n.p. + b Guibourtia demeusei (Harms) J.Leonard n.p. n.p. 0.00 0.00 * * Hylodesmum repandum (Vahl) H.Ohashi & R.R.Mill 0.75 n.p. n.p. n.p. + b Michelsonia microphylla (Troupin) Hauman n.p. n.p. n.p. 0.00 * * Millettia spp. n.p. 0.14 n.p. n.p. ** ** Millettia drastica Baker n.p. 0.05 0.01 0.00 * * Millettia dubia De Wild. n.p. 0.00 0.00 0.00 + f Mimosa pigra L. 0.82 n.p. n.p. n.p. + b Mimosa pudica L 0.91 n.p. n.p. n.p. + a Mucuna flagellipes Hook.f. 0.86 n.p. n.p. n.p. + c Neonotonia wightii (Wight & Arn.) J.A.Lackey 0.78 n.p. n.p. n.p. + b Paramacrolobium coeruleum (Taub.) Leonard n.p. n.p. n.p. 0.00 - a Rhyncosia spp. 0.47 n.p. n.p. n.p. ** ** Scorodophloeus zenkeri Harms n.p. n.p. n.p. 0.00 - f Senna alata (L.) Roxb. 0.00 0.00 n.p. n.p. - c Senna occidentalis (L.) Link 0.00 0.00 n.p. n.p. - c Senna spectabilis (DC.) H.S.Irwin & Barneby 0.00 0.00 n.p. n.p. + a Tephrosia vogelii Hook.f. 0.65 0.41 n.p. n.p. + a Teramnus labialis (L.f.) Spreng. 0.93 n.p. n.p. n.p. + a Tetrapleura tetraptera (Schum. & Thonn.) Taub. n.p. 0.37 0.03 n.p. + c Trifolium spp. 0.80 n.p. n.p. n.p. ** **

Appendices

119

Appendix A2. Legume composition of the different successional stages. For each subfamily following parameters are shown: the mean absolute abundances (A) (in individuals per hectare) ± standard deviation; relative abundance (RA) of the subfamilies within the successional stage; percentage of the individuals that were nodulated (N) ± standard deviation.

Caesalpinioideae Mimosoideae Papilionoideae

Primary Forest

A 6657 ± 625 258 ± 168 1908 ± 339 RA 75% 3% 22% N 0.03% ± 0.06% 0% 0%

Secondary Forest

A 4588 ± 298 3031 ± 361 1855 ± 368 RA 48% 32% 20% N 1.50% ± 1.49% 0.82% ± 1.26% 0%

5-10 Fallow

A 3682 ± 328 2651 ± 450 4110 ± 411 RA 35% 25% 39% N 1.59% ± 1.21% 22.95% ± 11.11% 17.90% ± 7.20%

0-5 Fallow

A 5078 ± 697 3157 ± 493 22218 ± 605 RA 17% 10% 73% N 1.39% ± 2.35% 85.94% ± 4.98% 80.85% ± 6.23%

References Appendix A Allen, K., M. D. Corre, A. Tjoa, and E. Veldkamp. 2015. Soil nitrogen-cycling responses to conversion of lowland forests to oil palm and rubber plantations in Sumatra, Indonesia. Plos One 17:5168. Bonnier, C. 1958. Symbiose Rhizobium-Légumineuses en région équatoriale. Publications de l’Institut National pour l’étude Agronomique du Congo Belge 72. Corby, H. D. L. 1974. Systematic implications of nodulation among Rhodesian legumes. Kirkia 9:301–309. Diabate, M., A. Munive, S. M. de Faria, A. Ba, B. Dreyfus, and A. Galiana. 2005. Occurrence of nodulation in unexplored leguminous trees native to the West African tropical rainforest and inoculation response of native species useful in reforestation. The New phytologist 166:231–9. Högberg, P., and I. J. Alexander. 1995. Roles of Root Symbioses in African Woodland and Forest - Evidence From N-15 Abundance and Foliar Analysis. Journal of Ecology 83:217–224. Sprent, J. I. 2001. Nodulation in Legumes. London: Royal Botanic Gardens, Kew.

Appendices

120

Appendix B Appendix B1 Molecular characterization of rainfall and throughfall samples in central Congo Basin. Appendix B2 Linear regression of mean annual relative abundance of condensed aromatics and annual organic N deposition Appendix B3 Annual-averaged black carbon column mass density over the African continent. Appendix B4 Total dissolved nitrogen (TDN), dissolved inorganic nitrogen (DIN), dissolved organic nitrogen (DON) and average molecular characteristics (mean ± SD) Appendix B5 Backwards wind trajectories (white dot lines) and fire pixels (red dots) per day, looped through the study period. Movie file available via the online version of the publication at pnas.org. Appendix B6 Monthly averaged Black Carbon Column Mass Density over the last 5 years over the African continent, with increasing intensity from yellow to brown. Movie file available via the online version of the publication at pnas.org.

Appendix B1. Molecular characterization of rainfall and throughfall samples in central Congo Basin. a-c, van Krevelen diagrams of the dissolved organic molecular formulae of open field rainfall, monodominant and mixed forest throughfall, and d-f, a subset with only the nitrogen containing formulae. g-i, the original van Krevelen for the open field rainfall and difference plots of monodominant and mixed forest, respectively. The difference plots show the mean relative abundance of the unique molecular formulae in the samples.

Appendices

121

Appendix B2. Linear regression of mean annual relative abundance of condensed aromatics and annual organic N deposition in the open field and at one monodominant and one mixed forest site.

Appendix B3. Annual-averaged black carbon column mass density over the African

continent during the period September 2015 to September 2016.

Appendices

122

Appendix B4. Total dissolved nitrogen (TDN), dissolved inorganic nitrogen (DIN), dissolved organic nitrogen (DON) and average molecular characteristics (mean ± SD) of open field rainfall, mixed forest and monodominant throughfall, sampled during one hydrological year in the remote forest of the central Congo Basin. The compound classification is shown in relative abundance (%), as well as in the number of assigned formulae (#).

Open field Mixed forest Monodominant forest

TDN in rainfall/throughfall (kg N ha-1 yr-1) 18.2 53.1 ± 3.2 37.5 ± 4.2

DIN (kg N ha-1 yr-1) 5.2 26.5 ± 2.2 16.8 ± 0.5

DON (kg N ha-1 yr-1) 13 26.6 ± 2.1 20.7 ± 2.2

Condensed Aromatics (%RA) 2.5 ± 0.8 4.3 ± 1 3.6 ± 1

Polyphenolic (%RA) 8.8 ± 1.1 12.4 ± 2.1 12.4 ± 1.7

Unsaturated (%RA) 74.1 ± 4 71.5 ± 2.5 74.8 ± 2.7

Aliphatic (%RA) 11.1 ± 3.8 9.1 ± 1.8 7 ± 2

N-containing (%RA) 38.6 33.4 29.6

Assigned Fomula (#) 17096 ± 4387 16955 ± 2213 15627 ± 2412

Unique formulae (#) 4817 1130 978

Unique versus open field (#) 2811 2459

Unique formulae versus open field (% RA) 0.73 0.97

N containing Formula (#) 9520 ± 2683 8765 ± 1457 7664 ± 1584

Unique formulae (#) 3266 744 403

Unique versus open field (#) 1380 1039

Unique formulae versus open field (% RA) 0.94 0.86

Appendix B5. Backwards wind trajectories (white dot lines) and fire pixels (red dots) per day, looped through the study period (available through SI of the paper online). Appendix B6. Monthly averaged Black Carbon Column Mass Density over the last 5 years over the African continent, with increasing intensity from yellow to brown (available through SI of the paper online).

Appendices

123

Appendix C Appendix C1 A list of the abbreviations of the full scientific names of planted species Appendix C2 List of H:DBH relations Appendix C3 Compositional characteristics of the different plots

Appendix C1. A list of the abbreviations of the full scientific names of all planted species in the plantation, along with the full and correct scientific name, family and subfamily. Species were identified by the local botanists of the INERA (Institut National pour l’Etude et la Recherche Agronomique Yangambi). There was no consensus about the species-level identification of one planted species of the Phyllanthus genus.

Abbreviation Full scientific name Family Subfamily

A.c. Autranella congolensis (De Wild.) A. Chev. Sapotaceae A.n. Antrocaryon nannanii De Wild. Anacardiaceae B.w. Blighia welwitschii (Hiern) Radlk. Sapindaceae C.a. Chrysophyllum africanum A. DC. Sapotaceae C.p. Carapa procera DC. Meliaceae D.l. Drypetes likwa J. Leonard Euphorbiaceae E.a. Entandrophragma angolense (Welw. ex C. DC.) C. DC. Meliaceae E.c. Entandrophragma cylindricum (Sprague) Sprague Meliaceae G.c. Guarea cedrata (A. Chev.) Pellegr. Meliaceae K.a. Khaya anthotheca (Welw.) C.DC. Meliaceae L.t. Lovoa trichilioides Harms Meliaceae M.a. Mammea africana Sabine Clusiaceae M.e. Milicia excelsa (Welw.) C.C. Berg Moraceae P.e. Pericopsis elata (Harms) Meeuwen Fabaceae Papilionoideae P.m. Pentaclethra macrophylla Benth. Fabaceae Mimosoideae P.o. Panda oleosa Pierre Pandaceae P.s. Pterocarpus soyauxii Taub. Fabaceae Papilionoideae P.sp. Phyllanthus species Phyllanthaceae P.t. Pachyelasma tessmannii (Harms) Harms Fabaceae Caesalpinioideae S.g. Strombosia grandifolia Hook. f. Olacaceae S.t. Strombosiopsis tetrandra Engl. Olacaceae T.a. Treculia africana Decne. Moraceae Z.g. Zanthoxylum gilletii (De Wild.) P.G. Waterman Rutaceae

Appendices

124

Appendix C2. List of H:DBH relations, and additionally their types and references, that were fitted to the tree heights and diameters, measured in the field. Per plot, the best fit was used to estimate the other tree heights, based on their diameter. H is tree height, DBH stands for diameter breast height and a, b and c are the equation’s variable parameters. The last column indicates which formula was selected for each plot, based on the best fit.

H:D relation Type Reference Plot

H=1.3 + a x (1 - exp(-b x DBHc) Weibull Huang, Titus & Wiens 1992 22

H=a x (1 - exp(-b x DBHc) Weibull Scaranello et al. 2012 H=1.3 + a x (1 - exp(-b x DBH) Chapman-Richards Huang, Titus & Wiens 1992 H=a x (1 - exp(-b x DBH))c Chapman-Richards Scaranello et al. 2012 H=exp(a + b x log(DBH)) / Brown, Gillespie & Lugo 1989 H=1.3 + a x (1 + b x exp(-c x DBH)) -1 Logistic Scaranello et al. 2012 2, 7, 9, 13, 25 H=1.3 + a x (1 + b-1 x DBH-c) -1 Modified logistic Huang, Titus & Wiens 1992 3 H=1.3 + exp(a + b x (DBH + 1) -1) Exponential Scaranello et al. 2012 5, 12, 14, 21, 24 H=1.3 + a x exp(b x (DBH + c) -1) Exponential Huang, Titus & Wiens 1992 H=a - b x exp(-c x DBH) Exponential Feldpausch et al. 2012 H=a x (1 - exp(-b x DBH)) Exponential Banin et al. 2012 1, 18 H=1.3 + a x DBH x (b + DBH) -1 Hyperbolic Scaranello et al. 2012 4, 10, 16, 19, 23,

26, 27, 28, 29 H=a x exp(-b x exp(-c x DBH) Gompertz Scaranello et al. 2012 H=1.3 + a x DBHb Power Scaranello et al. 2012 6, 8, 15, 17, 20 H=a x DBHb Power Scaranello et al. 2012 11

Appendices

125

Appendix C3. Compositional parameters and planted species of the different plots, calculated from the different subplots. AGC is above ground carbon in the woody biomass of the trees, BA is basal area and WD wood density. All values represent plot-level averages and standard deviations are calculated with the averages of the subplots within the plot, without taking into account the within-subplot variation. For wood density, we used data from the same species that were measured in the surrounding natural forest (Kearsley et al. 2013). The effective species richness represents the number of occurring tree species (including the spontaneous ingrowth). ). BApl is the ratio of basal area of the planted species to the total stand basal area in the plot (including spontaneous ingrowth). Effective Simpson’s diversity is calculated on the present trees. All species’ abbreviations are explained in Table A1.

Plot AGC

(Mg C ha-1

)

Stem Density

(ha-1

)

BA

(m2 ha

-1)

WD

(g cm-3

)

Effective species

richness (#)

Effective Simpson diversity

BApl Planted Species

1

Planted Species

2

1 294.49 ±63.01 702.8 ±113.5 45.06 ±8.28 0.65 ±.01 7.22 ±0.83 0.75 ±.04 0.78 ±.09 P.e. B.w.

2 200.08 ±61.91 336.1 ±96.1 32.02 ±10.16 0.63 ±.02 5.89 ±2.42 0.70 ±.09 0.61 ±.20 P.e.

3 218.01 ±72.70 544.4 ±87.3 37.34 ±9.74 0.57 ±.02 10.33 ±1.80 0.81 ±.05 0.74 ±.14 P.e. G.c.

4 277.53 ±116.10 397.2 ±61.8 41.56 ±14.77 0.69 ±.02 7.67 ±1.87 0.80 ±.06 0.79 ±.10 P.m. Z.g.

5 319.11 ±83.03 531.3 ±37.5 47.33 ±10.76 0.65 ±.01 6.25 ±0.50 0.74 ±.07 0.79 ±.18 P.e. P.o.

6 338.77 ±126.70 619.4 ±137.4 49.81 ±15.97 0.69 ±.03 8.67 ±2.83 0.70 ±.13 0.71 ±.15 A.c.

7 173.51 ±66.20 586.1 ±98.5 36.69 ±10.01 0.55 ±.03 12.44 ±2.51 0.87 ±.03 0.42 ±.14 P.s. T.a.

8 152.09 ±54.32 433.3 ±91.0 32.42 ±8.57 0.56 ±.03 10.89 ±2.26 0.87 ±.04 0.28 ±.22 P.s.

9 171.78 ±43.42 616.7 ±101.6 34.63 ±5.61 0.54 ±.01 10.67 ±2.45 0.80 ±.07 0.68 ±.18 E.c. A.n.

10 121.85 ±62.75 411.1 ±79.2 21.78 ±7.90 0.63 ±.04 8.56 ±0.73 0.80 ±.06 0.36 ±.15 P.m. C.p.

11 100.99 ±54.43 375.0 ±136.4 22.60 ±8.79 0.54 ±.03 8.56 ±2.19 0.84 ±.04 0.32 ±.16 M.e.

12 239.25 ±59.47 662.5 ±59.5 42.22 ±8.48 0.56 ±.01 7.00 ±2.45 0.46 ±.17 0.78 ±.15 G.c.

13 243.37 ±42.73 362.5 ±92.4 34.39 ±4.49 0.63 ±.04 6.75 ±2.06 0.72 ±.17 0.62 ±.42 P.e.

14 602.46 ±84.29 731.3 ±82.6 70.23 ±9.74 0.74 ±.00 3.50 ±0.58 0.54 ±.02 0.99 ±.00 A.c. D.l.

15 228.06 ±143.85 350.0 ±35.4 36.63 ±17.06 0.54 ±.06 4.25 ±2.87 0.38 ±.30 0.60 ±.44 P.o.

16 167.61 ±81.97 743.8 ±428.8 33.03 ±14.64 0.51 ±.05 12.50 ±3.87 0.80 ±.06 0.16 ±.10 E.c.

17 114.53 ±55.48 237.5 ±85.4 18.20 ±6.43 0.72 ±.06 4.25 ±1.26 0.58 ±.09 0.66 ±.21 P.m.

18 102.32 ±80.85 280.6 ±168.1 21.38 ±14.10 0.57 ±.03 7.56 ±3.43 0.80 ±.11 0.18 ±.14 M.e. P.sp.

19 150.25 ±47.82 306.3 ±55.4 26.67 ±7.15 0.61 ±.00 2.50 ±1.29 0.20 ±.17 0.98 ±.02 S.t.

20 258.40 ±20.50 418.8 ±62.5 40.71 ±3.17 0.65 ±.01 3.00 ±1.41 0.55 ±.07 0.98 ±.03 P.e. S.t.

21 95.85 ±70.07 494.4 ±157.5 23.43 ±12.26 0.52 ±.03 9.89 ±4.01 0.79 ±.11 0.41 ±.27 L.t.

22 106.94 ±38.10 394.4 ±91.7 23.68 ±7.99 0.52 ±.02 8.89 ±1.96 0.82 ±.05 0.29 ±.18 L.t. K.a.

23 144.45 ±64.91 441.7 ±106.8 31.20 ±9.23 0.50 ±.03 8.89 ±2.47 0.82 ±.06 0.45 ±.23 G.c. L.t.

24 299.49 ±138.66 377.8 ±108.6 48.30 ±20.45 0.62 ±.01 6.00 ±1.66 0.73 ±.05 0.78 ±.10 P.t.

25 141.24 ±52.70 558.3 ±114.6 31.66 ±10.86 0.51 ±.03 12.67 ±2.50 0.89 ±.02 0.30 ±.19 E.c. E.a.

26 130.26 ±105.26 587.5 ±94.6 28.94 ±12.62 0.53 ±.07 9.75 ±0.96 0.79 ±.03 0.22 ±.08 E.a.

27 269.35 ±84.00 637.5 ±425.5 43.13 ±11.51 0.62 ±.04 7.50 ±3.11 0.70 ±.08 0.31 ±.36 P.t. C.a.

28 255.84 ±46.01 356.3 ±94.4 38.74 ±6.05 0.60 ±.00 5.00 ±0.82 0.51 ±.07 0.95 ±.02 M.a.

29 255.43 ±166.71 500.0 ±143.6 44.05 ±24.64 0.57 ±.02 10.22 ±3.15 0.83 ±.06 0.52 ±.33 M.a. S.g.

Appendices

126

References Appendix C Banin, L. et al. (2012). What controls tropical forest architecture? Testing environmental, structural and floristic drivers. Global Ecology and Biogeography 21:1179-1190 Brown, S., Gillespie, A. J. R., & Lugo, A. E. (1989). Biomass Estimation Methods for Tropical Forests with Applications to Forest Inventory Data. Forest Science 35:881–902. Feldpausch, T. R. et al. (2012). Tree height integrated into pantropical forest biomass estimates. Biogeosciences 9:3381–3403. Huang, S., Titus, S. J., & Wiens, D. P. (1992). Comparison of nonlinear height-diameter functions for major Alberta tree species. Canadian Journal of Forest Research 22:1297-1304. Kearsley, E. et al. (2013). Conventional tree height-diameter relationships significantly overestimate aboveground carbon stocks in the Central Congo Basin. Nature Communications 4:2269. Scaranello, M. a., Alves, L. F., Vieira, S. A., & Camargo, P. B. De. (2012). Height-diameter relationships of tropical Atlantic moist forest trees in southeastern. Scientia Agricola 69:26–37.

Appendices

127

Appendix D Appendix D1 Design table of the experiment

Appendix D2 Correlation between chemical characteristics within the canopy (a), the litter

(b) and the soil (c) compartment

Appendix D3 Varimax-rotated PCA bi-plots for the canopy (a), litter (b) and soil (0-5 cm) (c)

compartment.

Appendix D4 Scores of the Varimax rotated-principal components.

Appendix D5 Pearson correlations between canopy and litter (a), soil and litter (b) and

canopy and soil (c) compartment

Appendix D6 Correlation plots of the actual (y axis) versus the planted canopy

characteristics (x axis).

Appendices

128

Appendix D1. Design table of the experiment, with the planted species per plot and the community-weighted means of the canopy nutrient values.

Plot Planted Species 1 Planted Species 2 C N

C:N P

N:P K Mg Ca

(g kg-1) (g kg-1) (g kg-1) (g kg-1) (g kg-1) (g kg-1)

14 Autranella congolensis (De Wild.) A. Chev. Drypetes likwa J. Leonard 468.7 20.1 24.4 1.05 19.2 6.2 2.3 6 6 Autranella congolensis (De Wild.) A. Chev. 478.8 19.4 25.8 1.15 16.9 5.9 2.5 7.2 26 Entandrophragma angolense (Welw. ex C. DC.) C. DC. 465.9 33.5 15.6 1.55 21.7 10.7 3.3 7.1 9 Entandrophragma cylindricum (Sprague) Sprague Antrocaryon nannanii De Wild. 455.5 36.6 14.2 1.7 21.5 13.2 2.9 6.3 25 Entandrophragma cylindricum (Sprague) Sprague Entandrophragma angolense (Welw. ex C. DC.) C. DC. 459.7 36.3 13.9 1.56 23.2 12.1 2.9 6.7 16 Entandrophragma cylindricum (Sprague) Sprague 463 41.6 12 2.18 19.6 13.8 3.4 7.9 23 Guarea cedrata (A. Chev.) Pellegr. Lovoa trichilioides Harms 448 34.2 14 1.93 17.8 15.2 4.4 9.4 12 Guarea cedrata (A. Chev.) Pellegr. 478.2 33.8 14.4 1.38 24.4 13.6 2.3 6.2 22 Lovoa trichilioides Harms Khaya anthotheca (Welw.) C.DC. 464.3 27.3 18 1.49 18.9 10.7 2.8 6.6 21 Lovoa trichilioides Harms 463.9 30.4 16 1.25 24.7 14 2.9 6.8 29 Mammea africana Sabine Strombosia grandifolia Hook. f. 486.2 22.3 24.6 0.99 22.6 8.2 2.3 5.5 28 Mammea africana Sabine 497 17.3 29.5 0.75 23.2 7 1.9 4.3 18 Milicia excelsa (Welw.) C.C. Berg Phyllanthus species 455 31.4 15.2 1.42 22.2 11.5 3.8 9 11 Milicia excelsa (Welw.) C.C. Berg 449 29.3 16.3 1.27 23.1 12.4 4.2 8.8 27 Pachyelasma tessmannii (Harms) Harms Chrysophyllum africanum A. DC. 490.2 27 19.4 1.11 25.9 9.3 2 4.2 24 Pachyelasma tessmannii (Harms) Harms 489.7 24.8 20.5 0.75 32.9 8.2 1.7 3.9 15 Panda oleosa Pierre 477.2 30.9 15.6 1.25 24.8 11.5 1.6 3.1 10 Pentaclethra macrophylla Benth. Carapa procera DC. 465.9 30.4 16.5 1.15 26.4 8.3 2.6 7.2 4 Pentaclethra macrophylla Benth. Zanthoxylum gilletii (De Wild.) P.G. Waterman 463.2 35.5 13.2 1.72 20.7 12.3 3.6 6.7 17 Pentaclethra macrophylla Benth. 460.8 35.7 13.1 1.22 29.7 10.1 2.9 5.7 1 Pericopsis elata (Harms) Meeuwen Blighia welwitschii (Hiern) Radlk. 478.9 38.9 12.8 1.47 26.6 8.5 1.8 4.3 3 Pericopsis elata (Harms) Meeuwen Guarea cedrata (A. Chev.) Pellegr. 481.1 35.9 13.8 1.38 26.1 12.2 1.8 4.7 5 Pericopsis elata (Harms) Meeuwen Panda oleosa Pierre 479.3 40.6 12.1 1.33 30.6 7.8 1 2.5 20 Pericopsis elata (Harms) Meeuwen Strombosiopsis tetrandra Engl. 481.3 40.5 12.1 1.36 29.9 10.7 1.3 2.4 13 Pericopsis elata (Harms) Meeuwen 479.2 39.8 12.3 1.41 28.4 7.3 1.4 3.8 2 Pericopsis elata (Harms) Meeuwen 477.8 41.9 11.6 1.37 30.5 7.7 1.2 2.9 7 Pterocarpus soyauxii Taub. Treculia africana Decne. 466 32.1 14.9 1.41 22.8 12.2 3.3 6.5 8 Pterocarpus soyauxii Taub. 472.3 31 15.6 1.34 23.2 12.2 3.7 7.6 19 Strombosiopsis tetrandra Engl. 482.2 34.1 14.3 1.4 24.4 18.4 2.4 3.8

Appendices

129

Appendix D2. Correlation between chemical characteristics within the canopy (a), the litter (b) and the soil (c) compartment. Significance levels were P < 0.1 (*), P < 0.05 (**) and P < 0.001 (***) (ns: non-significant). Variables shown in the tables are carbon (C), nitrogen (N), phosphorus (P), Total phosphorus (Ptot), available phosphorus (Pav), potassium (K), calcium (Ca), magnesium(Mg), C:N and N:P ratio, pH, exchangeable acidity (Ac) and exchangeable aluminium (Al).

(a) C N P C:N N:P K Mg Ca

C 1

N ns 1

P -0.61*** 0.69*** 1

C:N 0.40* -0.96*** -0.69*** 1

N:P 0.44* 0.38* -0.38* ns 1

K -0.43* ns 0.57** -0.49** Ns 1

Mg -0.78*** ns 0.45* ns -0.66*** 0.51** 1

Ca -0.76*** ns 0.39* ns -0.71*** Ns 0.93*** 1

(b) C N P C:N N:P K Mg Ca

C 1

N ns 1

P ns 0.80*** 1

C:N 0.53** -0.90*** -0.63*** 1

N:P ns ns -0.75*** ns 1

K ns ns 0.39* ns ns 1

Mg ns ns Ns ns ns Ns 1

Ca ns ns 0.45* ns -0.41* Ns 0.83*** 1

(c) C N Ptot Pav C:N N:P K Mg Ca Al Ac pH

C 1

N 0.97*** 1

Ptot 0.33* ns 1

Pav 0.54** 0.55** 0.47** 1

C:N 0.38** ns ns ns 1

N:P 0.81*** 0.87*** ns 0.48** ns 1

K 0.74*** 0.78*** ns 0.64*** ns 0.70*** 1

Mg 0.37* 0.44** ns 0.41** ns 0.42** 0.73*** 1

Ca ns ns ns ns ns ns Ns 0.61*** 1

Al 0.57*** 0.66*** ns 0.42** ns 0.76*** 0.60*** 0.32* Ns 1

Ac 0.80*** 0.74*** ns 0.50** 0.46* 0.60*** 0.56** ns Ns 0.53** 1

pH -0.86*** -0.85*** -0.32* -0.63*** -0.37** -0.76*** -0.66*** ns Ns -0.49** -0.66*** 1

Appendices

130

Appendix D3. Varimax-rotated PCA bi-plots for the canopy (a), litter (b) and soil (0-5 cm) (c) compartment. Variables included in the analyses are carbon (C), nitrogen (N), phosphorus (P), total phosphorus (Ptot), available phosphorus (Pav), potassium (K), calcium (Ca), magnesium (Mg), C:N and N:P ratio, exchangeable acidity (Exch_Ac) and exchangeable aluminium (Al). The percentage of variance explained by PC1 and PC2 is indicated on the respective axes. The points represent the different plots in the experiment (n=29). Appendix D4. Scores of the Varimax rotated-principal components.Ϯ Expressed in mg g-1 for canopy and litter, and meq 100 g-1 soil. Variables shown in the tables are carbon (C), nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), available phosphorus (Pav), magnesium(Mg), C:N and N:P ratio, pH, exchangeable acidity (Ac).

Variable Canopy Litter Soil

PC1(43%) PC2(41%) PC1(35%) PC2(30%) PC1 (43%)

PC2(21%)

C (g kg-1) -0.71 -0.54 0.47 -0.65 0.92 0.25 N (g kg-1) -0.24 0.95 0.36 0.85 0.85 0.38 P or Ptot (g kg-1) 0.43 0.81 0.77 0.49 0.47 -0.01 C:N 0.12 -0.98 -0.11 -0.99 0.60 -0.46 N:P -0.85 0.24 -0.79 0.08 0.06 0.30 KϮ 0.40 0.60 0.53 -0.02 0.69 0.65 MgϮ 0.94 0.17 0.69 0.15 0.32 0.87 CaϮ 0.95 0.04 0.69 0.12 -0.13 0.78 Pav (µg g-1) 0.73 0.17 pH (KCl) -0.87 -0.15 Ac (meq 100 g-1 soil) 0.86 -0.10

Appendices

131

Appendix D5. Pearson correlations between canopy and litter (a), soil and litter (b) and canopy and soil (c) compartment. Significance levels were P > 0.1 (non-significant; ns); P < 0.1 (*), P < 0.05 (**) and P < 0.001 (***). Variables shown in the tables are carbon (C), nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium(Mg), available phosphorus (Pav), C:N and N:P ratio, pH, exchangeable acidity (Ac) and exchangeable aluminium (Al), in the canopy (can), litter (lit) and soil (s).

(a) Ccan Ncan C:Ncan Pcan N:Pcan Kcan Mgcan Cacan

Clit ns ns 0.32** ns -0.38** ns Ns Ns

Nlit ns 0.66*** -0.67*** ns 0.53** ns Ns Ns

C:Nlit ns -0.71*** 0.75*** ns -0.61*** ns Ns Ns

Plit ns 0.61*** -0.59*** 0.46** ns 0.31** Ns Ns

N:Plit ns ns ns -0.45** ns ns Ns Ns

Klit ns 0.40** -0.39** 0.46** ns 0.34** Ns Ns

Mglit -0.63*** ns ns 0.47** -0.40** 0.51** 0.62*** 0.51**

Calit -0.52** ns ns 0.41** -0.39** 0.47** 0.51** 0.46**

(b) Clit Nlit C:Nlit Plit N:Plit Klit Mglit Calit

Cs ns ns ns ns ns ns -0.32* Ns

Ns ns ns ns ns ns 0.37** ns Ns

C:Ns ns -0.32* 0.37** -0.34* ns ns -0.46** -0.47**

Ptot ns ns ns ns ns ns ns Ns

Pav ns ns ns ns ns ns -0.38** -0.47**

N:Ptot ns 0.36* ns ns ns 0.41** ns Ns

Ks ns 0.34* -0.32* ns ns ns ns Ns

Mgs ns ns ns ns ns ns ns Ns

Cas -0.34* ns ns ns 0.48*** ns ns Ns

Al ns ns ns ns ns ns ns Ns

pH ns ns ns ns ns ns 0.42** 0.46**

Ac 0.31* ns ns ns ns ns -0.35* Ns

(c) Ccan Ncan C:Ncan Pcan N:Pcan Kcan Mgcan Cacan

Cs 0.55** ns ns ns 0.34** ns -0.37** -0.40**

Ns 0.43** ns ns ns 0.35** ns -0.32* -0.36*

C:Ns 0.64*** -0.43** 0.61*** -0.49** ns -0.48** -0.42** -0.32*

Ptot ns ns ns ns ns ns ns -0.39**

Pav 0.33* ns ns ns 0.40** ns -0.44** -0.47**

N:Ps ns ns ns ns ns ns ns Ns

Ks ns ns ns ns 0.34* ns ns Ns

Mgs ns ns ns ns ns ns ns Ns

Cas ns ns ns ns ns ns 0.41** 0.36**

Exch 0.40** ns -0.35* 0.33* ns ns ns Ns

pH -0.67*** ns ns 0.40** -0.58*** 0.34** 0.57*** 0.61***

Ac 0.52** ns -0.37** ns ns ns -0.44** -0.42**

Appendices

132

Appendix D6. Correlation plots of the actual (x axis) versus the planted canopy characteristics (y axis). The experimental plantation was not managed after a ten years of nursing period, so the original monocultures and 2-species-mixtures have been invaded by spontaneous in-growing species, which increased the complexity of this experiment. These correlation plots show that, in contrast with the obvious taxonomic implications of this invasion, this had only limited implications for the functional identity of the plots. Hence the conclusions made in the paper, linking leaf chemistry to soil characteristics, are valid, despite the fact that the species composition was not fully controlled through time.

Appendices

133

Appendix E

Appendix E1 Overview map Appendix E2 Structure and correlations of the trait data Appendix E3 Trends in community-level traits of previously reported studies Appendix E4 Coordinates, elevation and cluster membership of the different plots on both transects Appendix E5 Summary of the plot-level characteristics Appendix E6 Fixed effects estimates (altitude in km asl) for the different canopy-level response variables for the full model including interaction term

Appendix E1. Overview map with the locations of the Ecuador transect (upper; 400-3200 masl, 4-5 plots per cluster) and the Rwanda transect (lower; 1600 - 3000 masl, 5 plots per cluster) plot locations projected on a DEM (based on the Aster GDEM product (1). For clarity different scale legends are used in both maps.

Appendices

134

Appendix E2. Structure and correlations of the trait data in the Rwanda (upper), Ecuador (middle) and both (lower) transects with Spearman correlation statistics and their significances (p-value < 0.001 ***, < 0.01 **, < 0.05*,< 0.1 .). Diagonal shows the probability density function as fitted with fitted kernel density plots. Shown for specific leaf area (SLA), leaf nitrogen content (LNC), and leaf phosphorus content (LPC).

Appendices

135

Appendix E3. Trends in community-level leaf nitrogen content (LNC), leaf mass area (LMA) and mean annual temperature (MAT) on different transects (as described in Asner et al., 2016; Kitayama & Aiba, 2002; Van de Weg et al., 2009) compared to our transects in Ecuador and Rwanda. Figure was obtained by using reported linear regression parameters, and by applying simple linear regression lines to this study’s data.

Appendices

136

Appendix E4. Coordinates, elevation and cluster membership of the different plots on both transects.

Ecuador Rwanda

Plot Cluster Latitude Longitude Elevation Plot Cluster Latitude Longitude Elevation

1 1 00° 08' 50.58" N 079° 08' 37.03" W 406 1 4 02° 26' 47.28" S 029° 15' 09.96" E 2879

2 1 00° 08' 45.79" N 079° 08' 34.45" W 420 2 4 02° 26' 50.93" S 029° 15' 07.55" E 2875

3 1 00° 08' 50.31" N 079° 08' 31.87" W 404 3 4 02° 26' 25.68" S 029° 15' 00.96" E 2937

4 1 00° 08' 50.49" N 079° 08' 33.30" W 410 4 4 02° 27' 09.89" S 029° 14' 57.36" E 2767

5 1 00° 08' 50.27" N 079° 08' 35.04" W 394 5 4 02° 27' 12.42" S 029° 14' 57.29" E 2761

6 2 00° 02' 10.16" N 078° 52' 00.04" W 1098 6 2 02° 28' 52.79" S 029° 11' 04.08" E 2293

7 2 00° 02' 08.27" N 078° 51' 59.52" W 1055 7 2 02° 28' 41.45" S 029° 10' 53.51" E 2240

8 2 00° 02' 06.95" N 078° 51' 59.51" W 1077 8 1 02° 27' 33.59" S 029° 12' 02.52" E 1745

9 2 00° 02' 11.07" N 078° 52' 02.51" W 1041 9 1 02° 34' 16.07" S 029° 13' 14.40" E 1835

10 3 00° 05' 36.80" N 078° 37' 17.64" W 1953 10 1 02° 34' 15.35" S 029° 13' 04.25" E 1799

11 3 00° 05' 45.48" N 078° 37' 19.25" W 1893 11 1 02° 34' 15.12" S 029° 12' 29.88" E 1760

12 3 00° 05' 48.54" N 078° 37' 22.21" W 1873 12 1 02° 34' 21.47" S 029° 12' 01.91" E 1659

13 3 00° 06' 06.87" N 078° 37' 40.73" W 1764 13 2 02° 28' 23.51" S 029° 12' 21.42" E 2141

14 4 00° 26' 58.63" N 078° 28' 13.58" W 3214 14 2 02° 28' 13.91" S 029° 12' 29.69" E 2158

15 4 00° 26' 59.20" N 078° 28' 14.94" W 3222 15 3 02° 28' 10.44" S 029° 14' 25.68" E 2557

16 4 00° 26' 58.11" N 078° 28' 12.35" W 3191 16 3 02° 28' 11.09" S 029° 14' 18.17" E 2523

17 4 00° 27' 04.12" N 078° 28' 18.17" W 3241 17 2 02° 29' 10.61" S 029° 09' 33.90" E 2167

18 3 02° 28' 43.97" S 029° 12' 07.85" E 2456

19 3 02° 29' 10.61" S 029° 11' 55.74" E 2500

20 3 02° 29' 14.16" S 029° 11' 42.83" E 2522

Appendices

137

Appendix E5a. Summary of the plot-level (mean and standard deviation) leaf nitrogen content (LNC), leaf phosphorus content (LPC), specific leaf area (SLA) and leaf C:N, C:P, N:P and δ15N, along with the total percentage of the basal area (BA) sampled per plot of the Ecuador transect.

Transect Plot Altitude SLA (cm2 g-1) LNC (%) LPC (%) C:N C:P N:P δ15N (‰) Percenge of BA sampled

Ecuador 1 404 149.74 ± 66.55 2.61 ± 0.48 0.13 ± 0.03 17.43 ± 2.88 359.7 ± 65.25 20.89 ± 3.06 2.7 ± 1.5 77%

Ecuador 2 420 136.43 ± 42.31 3.02 ± 0.83 0.15 ± 0.05 15.45 ± 3.95 320.39 ± 107.88 21.06 ± 2.52 3.24 ± 1.4 73%

Ecuador 3 404 148.83 ± 54.33 2.62 ± 0.57 0.13 ± 0.03 17.69 ± 3.11 356.46 ± 76.83 20.62 ± 3.72 2.66 ± 0.79 83%

Ecuador 4 410 146.62 ± 57.34 2.61 ± 0.59 0.13 ± 0.03 17.64 ± 3.26 364.41 ± 76.66 20.92 ± 3.03 2.74 ± 0.96 66%

Ecuador 5 394 146.04 ± 68.13 2.57 ± 0.58 0.12 ± 0.03 17.73 ± 3.15 371.72 ± 67.92 21.32 ± 3.26 2.54 ± 0.86 61%

Ecuador 6 1098 142.58 ± 35.07 2.31 ± 0.52 0.14 ± 0.04 20.6 ± 4.23 369.76 ± 137.5 18.57 ± 6.64 -0.25 ± 1.4 30%

Ecuador 7 1055 164.48 ± 46.3 2.68 ± 0.7 0.15 ± 0.06 17.68 ± 4.64 320.68 ± 120.44 18.52 ± 5.2 1.12 ± 1.87 55%

Ecuador 8 1077 162.65 ± 34.15 2.73 ± 0.57 0.15 ± 0.03 17.22 ± 3.61 304.51 ± 55.54 18.12 ± 2.35 1.5 ± 1.09 55%

Ecuador 9 1041 135.56 ± 45.96 2.15 ± 0.54 0.14 ± 0.05 20.36 ± 3.17 332.95 ± 72.91 16.41 ± 2.91 0.77 ± 1.67 44%

Ecuador 10 1953 115.76 ± 25.38 2.08 ± 0.43 0.13 ± 0.01 22.48 ± 5.34 341.55 ± 43.48 15.78 ± 2.54 0.22 ± 0.81 64%

Ecuador 11 1893 123.92 ± 26.67 1.98 ± 0.57 0.12 ± 0.03 23.74 ± 7.24 365.68 ± 95 16.02 ± 2.5 -0.04 ± 1.1 83%

Ecuador 12 1873 113.68 ± 29.9 2.28 ± 0.38 0.15 ± 0.04 20.51 ± 3.75 323.25 ± 80.78 16.44 ± 4.47 0.18 ± 0.73 76%

Ecuador 13 1764 113.57 ± 18.51 1.99 ± 0.38 0.13 ± 0.01 23.28 ± 4.46 360.03 ± 37.76 15.67 ± 2.39 0.12 ± 0.52 72%

Ecuador 14 3214 47.95 ± 26.12 0.97 ± 0.36 0.06 ± 0.01 50.96 ± 13.85 772.64 ± 115.72 15.56 ± 3.55 -5.07 ± 1.94 85%

Ecuador 15 3222 49.29 ± 24.14 1.02 ± 0.47 0.06 ± 0.01 49.43 ± 15.54 756.2 ± 138.32 15.86 ± 3.99 -4.83 ± 2.3 95%

Ecuador 16 3191 46.88 ± 18.44 0.98 ± 0.43 0.06 ± 0.01 51.07 ± 14.52 770.36 ± 141.15 15.46 ± 3.23 -5.05 ± 2.32 91%

Ecuador 17 3241 59.27 ± 30.31 1.16 ± 0.38 0.07 ± 0.01 44.08 ± 14.2 697.04 ± 151.06 16.47 ± 2.8 -4.23 ± 1.84 95%

Appendices

138

Appendix E5b.Summary of the plot-level (mean and standard deviation) leaf nitrogen content (LNC), leaf phosphorus content (LPC), specific leaf area (SLA) and leaf C:N, C:P, N:P and δ15N, along with the total percentage of the basal area (BA) sampled per plot of the Rwanda transect.

Transect Plot Altitude SLA (cm2 g-1) LNC (%) LPC (%) CN CP NP δ15N (‰) Percentage

of BA sampled

Rwanda 1 2879 42.04 ± 28.52 1.53 ± 0.94 0.09 ± 0.04 34.58 ± 13.74 564.61 ± 171.59 16.81 ± 2.85 -0.12 ± 1.46 90% Rwanda 2 2875 45.83 ± 23.33 1.83 ± 1.03 0.1 ± 0.05 30.78 ± 14.2 508.9 ± 191.33 17.16 ± 2.3 0.52 ± 1.89 95% Rwanda 3 2937 35.53 ± 26.65 1.33 ± 1.09 0.08 ± 0.05 37.72 ± 15.34 598.75 ± 202.9 16.09 ± 2.42 -0.2 ± 1.6 99% Rwanda 4 2767 60.76 ± 31.92 2.09 ± 0.93 0.11 ± 0.04 27.67 ± 11.72 483.82 ± 147.81 19.26 ± 5.05 0.8 ± 1.73 91% Rwanda 5 2761 60.19 ± 32.62 2.09 ± 0.98 0.11 ± 0.04 27.95 ± 12.02 488.41 ± 149.64 19.55 ± 5.66 0.68 ± 1.59 97% Rwanda 6 2293 99.03 ± 21.26 2.54 ± 0.49 0.12 ± 0.03 19.85 ± 3.8 398.62 ± 86.35 21.01 ± 3.94 3.2 ± 1.85 86% Rwanda 7 2240 112.49 ± 25.8 2.49 ± 0.53 0.13 ± 0.02 20.88 ± 4.5 380.75 ± 78.3 19.57 ± 3.17 4.07 ± 1.9 94% Rwanda 8 1745 110.11 ± 25.21 2.59 ± 0.66 0.13 ± 0.02 20.59 ± 4.03 385.18 ± 73.06 20.44 ± 3.37 3.52 ± 1.94 99% Rwanda 9 1835 106.3 ± 32.8 2.68 ± 0.74 0.13 ± 0.04 19.81 ± 4.91 384.09 ± 96.91 20.83 ± 3.7 3.8 ± 1.83 97% Rwanda 10 1799 106.11 ± 30.81 2.62 ± 0.72 0.13 ± 0.05 20.19 ± 5.37 388.53 ± 115.94 20.4 ± 5.11 4.23 ± 1.98 92% Rwanda 11 1760 111.25 ± 21.64 2.6 ± 0.63 0.12 ± 0.02 20.4 ± 4.39 388.3 ± 62.55 20.89 ± 3.51 3.97 ± 1.48 100% Rwanda 12 1659 104.19 ± 39.07 2.39 ± 0.71 0.12 ± 0.04 21.88 ± 4.92 404.88 ± 97.93 19.63 ± 3.97 3.51 ± 2.27 100% Rwanda 13 2141 102.67 ± 21.38 2.61 ± 0.56 0.12 ± 0.03 19.39 ± 4.36 413.08 ± 102.42 22.01 ± 4.45 4.01 ± 1.99 97% Rwanda 14 2158 105.99 ± 22.76 2.55 ± 0.39 0.11 ± 0.02 20.85 ± 3.29 439.03 ± 66.3 23.06 ± 3.52 3.31 ± 1.42 100% Rwanda 15 2557 98.84 ± 24.87 2.5 ± 0.73 0.11 ± 0.02 22.43 ± 4.52 433.44 ± 65.02 22.89 ± 6.92 1.01 ± 0.88 99% Rwanda 16 2523 100.12 ± 27.99 2.38 ± 0.73 0.11 ± 0.02 23.16 ± 5.33 440.96 ± 61.48 22.02 ± 6.44 0.94 ± 0.8 100% Rwanda 17 2167 89.7 ± 30.72 2.18 ± 0.63 0.11 ± 0.03 24.28 ± 7.86 448.71 ± 101.96 19.45 ± 2.82 1.93 ± 1.83 97% Rwanda 18 2456 94.81 ± 32.58 2.04 ± 0.46 0.11 ± 0.02 25.36 ± 5.17 452.75 ± 75.93 19.01 ± 2.84 0.94 ± 0.9 99% Rwanda 19 2500 89.29 ± 32.47 2.24 ± 0.43 0.12 ± 0.02 21.6 ± 3.85 386.74 ± 73.89 18.15 ± 1.34 0.98 ± 1.27 93% Rwanda 20 2522 86.38 ± 23.57 2.2 ± 0.55 0.12 ± 0.03 23.02 ± 5.39 410.66 ± 91.2 18.45 ± 0.81 1.36 ± 1.13 98%

Appendices

139

Appendix E6. Fixed effects estimates (altitude in km asl) for the different canopy-level response variables for the full model including interaction term; leaf nitrogen content (LNC), inverse C:N ratio, specific leaf area (SLA), leaf phosphorus content (LPC), N:P ratio, canopy and topsoil δ15N. The ΔAIC shows the differences in Akaike information criterion between the full model and model without interaction term, with positive values for a reduction in AIC when discarding the interaction term.

Response Effect Estimate Standard Error P-value ΔAIC

LNC (%) Ecuador intercept 3.010 0.145 <0.001 1.10

Rwanda intercept 3.930 0.466 0.114 Altitude -0.572 0.000 0.001 Altitude:Transect -0.140 0.000 0.525 SLA (cm2 g-1) Ecuador intercept 171.377 15.534 <0.001 1.16

Rwanda intercept 199.016 44.535 0.552 Altitude -33.849 0.008 0.009 Altitude:Transect -13.755 0.019 0.496 N:C Ecuador intercept 0.066 0.004 <0.001 1.94

Rwanda intercept 0.075 0.011 0.472 Altitude -0.013 0.000 0.002 Altitude:Transect -0.000 0.000 0.930 LPC (%) Ecuador intercept 0.157 0.014 <0.001 2.00

Rwanda intercept 0.170 0.042 0.763 Altitude -0.024 0.000 0.024 Altitude:Transect 0.000 0.000 0.985 N:P Ecuador intercept 20.440 1.215 <0.001 1.60

Rwanda intercept 25.885 3.842 0.218 Altitude -1.686 0.001 0.051 Altitude:Transect -0.913 0.002 0.607

Appendices

140

References Appendix E NASA/METI (2011) The Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Global Digital Elevation Model (GDEM) version 2. Available at: http://earthexplorer.usgs.gov/.

Asner, G. P., R. E. Martin, C. B. Anderson, K. Kryston, N. Vaughn, D. E. Knapp, L. P. Bentley, A. Shenkin, N. Salinas, F. Sinca, R. Tupayachi, D. Sandra, K. Q. Huaypar, M. M. Pillco, F. Delis, C. Alvarez, Y. Malhi, K. Quispe Huaypar, M. Montoya Pillco, F. D. Ccori Álvarez, S. Díaz, B. Enquist, Y. Malhi, D. Sandra, K. Q. Huaypar, M. M. Pillco, F. Delis, C. Alvarez, and Y. Malhi. 2016. Scale dependence of canopy trait distributions along a tropical forest elevation gradient. New Phytologist 214:973–988.

Kitayama, K., and S. Aiba. 2002. Ecosystem structure and productivity of tropical rain forests along altitudinal gradients with contrasting soil phosphorus pools on Mount Kinabalu , Borneo. Journal of Ecology 90:37–51.

Nakagawa, S., and H. Schielzeth. 2013. A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods in Ecology and Evolution 4:133–142.

Van de Weg, M. J., P. Meir, J. Grace, O. K. Atkin, M. J. Van de Weg, P. Meir, J. Grace, and O. K. Atkin. 2009. Altitudinal variation in leaf mass per unit area, leaf tissue density and foliar nitrogen and phosphorus content along an Amazon-Andes gradient in Peru. Plant Ecology & Diversity 2:243–254.

http://earthexplorer.usgs.gov/

Appendices

141

Appendix F Appendix F1 Overview of studies reporting nitrogen (N) export in tropical forest basins Appendix F2 Overview of studies reporting nitrogen (N) deposition in tropical forest basins Appendix F3 Water budgets of both geographical location and the validation with the GLEAM model Appendix F4 Yearly litterfall mass flux and nitrogen flux per forest typ Appendix F5 Gross soil N transformation Appendix F1. Overview of studies reporting nitrogen (N) export in tropical forest basins. In addition to this table, an extensive overview can be found for N export in Lewis et al. 1999 (from which part of this table was adopted). All reported values in the table are expressed in kg N ha-1 yr-1; Total dissolved nitrogen (TDN), dissolved organic nitrogen (DON), dissolved inorganic nitrogen (DIN), and particulate organic nitrogen (PON). Where multiple years of data was available, we just show the most recent yearly data here. Only papers where numbers in kg n ha-1 yr-1 could be found were included, for recent concentration-expressed export see Gücker et al. (2016)

Ecosystem Publication Location Export

NH4+ NO3

- DIN DON TDN PON TN

Lowland tropical Germer et al. 2009 Amazon 4.35 0.45

Lesack 1993 Braco do mota 0.14 0.6 0.74 2.87 3.61 0.7 4.31

Lewis 1995 Jurua 0 1.82 1.82 1.21 3.03 1.44 4.45

Lewis 1995 Japura 0 2.04 2.04 1.84 3.88 2.44 8.6

Lewis 1995 Negro 0 0.67 0.67 2.48 3.15 0.02 1.47

Lewis 1995 Amazon, Obidos 0.24 1.68 1.92 1.94 3.86 2.42 6.06

Lewis et al. 1995 Trombetas 0.53 3.01

Taylor et al. 2015 Costa Rica 0.06 0.26 0.32 1.37 1.69 14.6 16.29

Yusop et al. 2006 Malaysia 1.61 5.80 7.40

Montane tropical Brookshire et al. 2012 Costa Rica 0 5 5 3 8

Bruijnzeel et al. 1993 Sabah 1 14 15

Bücker et al. 2011 Ecuador 4.63

Lewis 1995 Solimoes at ica 0 2.4 2.4 1.36 3.76 3.9 7.66

Lewis 1995 Madeira 0 1.3 1.3 1.3 2.6 2.03 4.63

Lewis and Saunders 1989 Caroni 0.61 1.46 2.07 2.35 4.42 0.75 5.17

Lewis et al 1986 Caura 0.89 1.49 2.38 4 6.38 3.64 9.98

Mcdowell & Asbury 1994 Puerto Rico 0.49 1.83 2.32 3.78 6.1 1.26 7.36

McDowell et al. 1995 Carribean 1.48 3.33 4.81

Newbold et al. 1995 Costa Rica 5.05 2.2 7.25

Rütting et al. 2014 Rwanda 1.1 19.7 20.8

Appendices

142

Appendix F2. Overview of studies reporting nitrogen (N) deposition in tropical forest basins. All values shown are throughfall values, unless stated otherwise. For an overview of bulk deposition values see Boy et al. (2008). All reported values for total dissolved nitrogen (TDN), dissolved organic nitrogen (DON) and dissolved inorganic nitrogen (DIN) in the table are expressed in kg N ha-1 yr-1.

Ecosystem Publication Location Deposition NH4

+ NO3- DIN DON TDN

Lowland tropical Andreae et al 1990 Amazon 1.7 3.1 4.8

Eklund et al 1997 Costa rica 3.8 2.6 6.4 3.8 10.2

Galy-Lacaux et al. 2014 Congo-Brazzaville 15.1

Hofhansl 2011 Costa Rica 5.1 2 7.1 0.6 7.7

Lacaux et al. 1992 Brazzaville (bulk) 3.65 16.05 19.7

Muoghalu 2006 Nigeria 2.11 32 34.11

Schwendenmann & Veldkamp 2005 Costa Rica 17 9 26

Sigha-Nkamdjou et al. 2003 Cameroon (bulk) 3.9 8.6 12.5

Williams et al 1997 Amazon 1.5 7.1 8.6

Montane tropical Asbury et al. 1994 Puerto Rico 6.9 5.3 12.2

Boy et al. 2008 Ecuador (figures) 5.7 7.25 12.95 10.45 23.4

Clark et al. 1998 Costa Rica 1.7 1.7

Hölscher et al. 2003 Costa Rica 3.9 0.8 4.7

Liu et al. 2003 China 2.7 0.91 3.61

Makowski et al. 2013 Peru 6.125

Mayer et al. 2000 SE Brazil 24.1 8.3 32.4

McDowell et al. 1990 Puerto Rico 1.78 2.03 3.81

Schrumpf et al. 2005 Tanzania (Kili) 3.3 0.85 4.15 8.1 12.25

Veneklaas 1990 Colombia 16.53

Appendices

143

Appendix F3. Water budgets of both geographical locations and validation with the GLEAM model. The reported numbers are empirical and model validations integrated over the period we were able to measure the outflow of the catchment in both geographical locations (CE= canopy evaporation, ET= evapotranspiration).

Location Waterbalance Amount (mm) fraction of the budget

Lowland Open field Rainfall 1336 CE LMF 150 11%

CE LMoF 213 16%

Catchment outlet 545 41%

Total ET 791 59%

GLEAM ET 918

Montane Open field Rainfall 1851 CE MMF 334 18%

CE MBF 160 9%

Catchment outlet 497 27%

Total ET 1354 73%

GLEAM ET 1136

Appendices

144

Appendix F4. Reported values for gross nitrogen transformation rates in the tropics, based on in situ intact soil core pool dilution experiments (with exception of this study and Rütting et al. 2015 which used tracing experiments with numerical modeling). Table based on a mini-review by Gerschlauer et al. (2016), with mineralization (M), immobilization (I), consumption (C), calculated immobilization (IC) for ammonium (am) and nitrate (nit) and dissimilatory nitrate reduction to ammonium (DNRA).

Forest type Location M Iam Cam IC,am N Init Cnit IC,nit DNRA M/I N/M Source

Lowland Indonesia 11.5 16.8 0.9 2 0.4 0.61 0.08 Allen et al. 2015

Costa Rica 7.9 11.45 6.35 5.1 5.2 5 0.2 0.68 0.65 Silver et al. 2005

Brazil 13.9 13.7 3.8 5.8 0.8 0.71 0.27 Sotta et al. 2009

Ecuador 52 66 64.6 1.4 4.5 0.80 0.03 Arnold et al. 2008, 2009

Panama 77 84 82.7 1.3 0.93 0.02 Arnold et al. 2008, 2009

Brazil 7.2 6.8 0.3 0.9 0.3 0.94 0.04 Sotta et al. 2009

Costa Rica 4.8 5.1 0.9 0.1 0.94 0.19 Wieder et al. 2013

DRC 5.65 4.27 1.47 1.36 0.05 1.00 0.26 This study

Panama 35.7 27 32.9 32.7 0.2 0.1 1.32 0.01 Corre et al. 2010

Indonesia 5.4 2.7 1.9 0.9 0.2 1.50 0.35 Allen et al. 2015

Lower montane Puerto Rico 1.1 2.63 0.37 2.26 3.2 3.1 0.1 0.31 2.05 Templer et al. 2008

Puerto Rico 3.5 11.05 9.9 1.15 1 -0.3 1.3 0.32 0.33 Templer et al. 2008

Puerto Rico 1.4 3.08 2.46 0.62 1.3 0 0.37 0.44 Templer et al. 2008

Montane Ecuador 20 26 23.8 2.2 1.3 1.3 1.3 0.80 0.11 Arnold et al. 2008, 2009

Ecuador 70 84 83.3 0.7 4.5 4.5 4.5 0.80 0.01 Arnold et al. 2008, 2009

Mexico 14 16 9.3 6.7 5.8 0.93 0.48 Davidson et al. 1993

Panama 22.4 22.4 22.1 0.3 0.2 1.00 0.01 Corre et al. 2010

Rwanda 2 0.5 1.5 0.2 0 2.86 0.75 Rütting et al. 2015

Mexico 19 20 -1 21 3.1 9.05 1.11 Davidson et al. 1993

Plantation Costa Rica 3.9 3.24 -1.16 4.4 2.3 2.1 0.2 3.42 1.13 Silver et al. 2005

Appendices

145

Appendix F5. Gross soil N transformations (µg N g-1 d-1) for the three lowland mixed forest (LMF) sites and mineralization to immobilization ratios (M/I) and nitrification to mineralization ratios (N/M). The rates show the mean and standard deviations over the experimental repetitions in LMF.

Literature Appendix F

Allen, K., M. D. Corre, A. Tjoa, and E. Veldkamp. 2015. Soil nitrogen-cycling responses to conversion of lowland forests to oil palm and rubber plantations in Sumatra, Indonesia. Plos One 17:5168.

Andreae, M. O., R. W. Talbot, H. Berresheim, and K. M. Beecher. 1990. Precipitation chemistry in central Amazonia. Journal of Geophysical Research 95:16,916-987,999.

Arnold, J., M. D. Corre, and E. Veldkamp. 2008. Cold storage and laboratory incubation of intact soil cores do not reflect in-situ nitrogen cycling rates of tropical forest soils. Soil Biology and Biochemistry 40:2480–2483.

Arnold, J., M. D. Corre, and E. Veldkamp. 2009. Soil N cycling in old-growth forests across an Andosol toposequence in Ecuador. Forest Ecology and Management 257:2079–2087.

Asbury, C. E., W. H. McDowell, R. Trinidad-Pizarro, and S. Berrios. 1994. Solute deposition from cloud water to the canopy of a puerto rican montane forest. Atmospheric Environment 28:1773–1780.

Boy, J., C. Valarezo, W. Wilcke, R. Rollenbeck, C. Valarezo, and W. Wilcke. 2008. Amazonian biomass burning-derived acid and nutrient deposition in the north Andean montane forest of Ecuador. Global Biogeochemical Cycles 22:1–16.

Brookshire, E. N. J., S. Gerber, D. N. L. Menge, and L. O. Hedin. 2012. Large losses of inorganic nitrogen from tropical rainforests suggest a lack of nitrogen limitation. Ecology letters 15:9–16.

Bruijnzeel, L. A., M. J. J. Waterloo, J. Proctor, A. T. T. Kuiters, B. Kotterink, L. A. Bruijnzeel, M. J. J. Waterloo, J. Proctor, A. T. T. Kuiters, and B. Kotterink. 1993. Hydrological Observations in Montane Rain Forests on Gunung Silam, Sabah , Malaysia, with special reference effect to the ’ Massenerhebung ’. Journal of Ecology 81:145–167.

Bücker, A., P. Crespo, H.-G. Frede, and L. Breuer. 2011. Solute behaviour and export rates in neotropical montane catchments under different land-uses. Journal of Tropical Ecology 27:305–317.

N fluxes LMF1 LMF2 LMF3

Gross NH4+ production Mineralization 9.72 (2.20) 4.62 (0.36) 2.52 (0.41)

DNRA 0.09 (0.03) 0.03 (0.00) 0.04 (0.01)

Gross NH4+ consumption Nitrification 1.07 (0.10) 1.74 (0.17) 1.6 (0.17)

NH4+ immobilization 8.73 (2.21) 3.01 (0.41) 1.06 (0.48)

Gross NO3- production Nitrification 1.07 (0.10) 1.74 (0.17) 1.60 (0.17)

Gross NO3- consumption DNRA 0.09 (0.03) 0.03 (0.00) 0.04 (0.01)

NO3- immobilization 0.96 (0.09) 1.65 (0.17) 1.46 (0.14)

M/I 1.00 0.99 1.00

N/M 0.11 0.38 0.63

Appendices

146

Clark, K. L., N. M. Nadkarni, D. Schaefer, and H. L. Gholz. 1998. Atmospheric deposition and net retention of ions by the canopy in a tropical montane forest, Monteverde, Costa Rica. Journal of Tropical Ecology 14:27–45.

Corre, M. D., E. Veldkamp, J. Arnold, S. J. Wright, M. D. Corre, E. Veldkamp, J. Arnold, and S. J. Wright. 2010. Impact of elevated N input on soil N cycling and losses in old-growth lowland and montane forests in Panama. Ecology 91:1715–1729.

Davidson, A. E. a, P. A. a Matson, P. M. M. Vitousek, R. Riley, K. Dunkin, J. M. M. Maass, E. A. Davidson, P. A. a Matson, P. M. M. Vitousek, R. Riley, K. Dunkin, G. Garcia-Mendez, and J. M. M. Maass. 1993. Processes regulating soil emissions of NO and N2O in a seasonally dry tropical forest. Ecology 74:130–139.

Eklund, T. J., and W. H. McDowell. 1997. Seasonal variation of tropical precipitation chemistry: La Selva, Costa Rica. Atmospheric Environment 31:3903–3910.

Galy-Lacaux, C., C. Delon, C. Galy-Lacaux, and C. Delon. 2014. Nitrogen emission and deposition budget in West and Central Africa. Environmental Research Letters 9:125002.

Germer, S., C. Neill, T. Vetter, J. Chaves, A. V. Krusche, and H. Elsenbeer. 2009. Implications of long-term land-use change for the hydrology and solute budgets of small catchments in Amazonia. Journal of Hydrology 364:349–363.

Hofhansl, F., W. Wanek, S. Drage, W. Huber, A. Weissenhofer, and A. Richter. 2011. Topography strongly affects atmospheric deposition and canopy exchange processes in different types of wet lowland rainforest, Southwest Costa Rica. Biogeochemistry 106:371–396.

Hölscher, D., L. Köhler, C. Leuschner, and M. Kappelle. 2003. Nutrient fluxes in stemflow and throughfall in three successional stages of an upper montane rain forest in Costa Rica. Journal of Tropical Ecology 19:557–565.

Lacaux, J. P., R. Delmas, G. Kouadio, B. Cros, and M. O. Andreae. 1992. Precipitation Chemistry in the Mayombé Forest of Equatorial Africa. Journal of Geophysical 97:6195–6206.

Lesack, L. F. W. 1993. Export of nutrients and major ionic solutes from a rain forest catchment in the Central Amazon Basin. Water Resources Research 29:743–758.

Lesack, L. W., and J. Melack. 1996. Mass balance of major solutes in a rainforest catchment in the Central Amazon: Implications for nutrient budgets in tropical rainforests. Biogeochemistry 32:115–142.

Lewis Jr, W. M., J. F. Saunders III, S. N. Levine, and F. H. Weibezahn. 1986. Organic carbon in the Caura River, Venezuela. Limnology and Oceanography 31:653–656.

Lewis, W. M., S. K. Hamilton, and J. F. Saunders. 1995. Rivers of Northern South America. Pages 219–256 in C. E. Cushing, K. W. Cummins, and G. W. Minshall, editors.River and stream ecosystems. Elsevier, Amsterdam.

Lewis, W. M., J. M. Melack, W. H. McDowell, M. Mcclain, and J. E. Richey. 1999. Nitrogen yields from undisturbed watersheds in the Americas. Biogeochemistry 46:149–162.

Appendices

147

Lewis, W. M., and J. F. Saunders. 1989. Concentration and Transport of Dissolved and Suspended Substances in the Orinoco River. Biogeochemistry 7:203–240.

Liu, W., J. E. D. Fox, and Z. Xu. 2003. Nutrient budget of a montane evergreen broad-leaved forest at Ailao Mountain National Nature Reserve, Yunnan, southwest China. Hydrological Processes 17:1119–1134.

Makowski Giannoni, S., K. Trachte, R. Rollenbeck, L. Lehnert, J. Fuchs, and J. Bendix. 2016. Atmospheric salt deposition in a tropical mountain rainforest at the eastern Andean slopes of south Ecuador-Pacific or Atlantic origin? Atmospheric Chemistry and Physics 16:10241–10261.

Mayer, R., S. Liess, M. I. M. S. Lopes, and K. Kreutzer. 2000. Atmospheric pollution in a tropical rain forest: Effects of deposition upon biosphere and hydrosphere II. Fluxes of chemicals and element budgets. Water, Air, and Soil Pollution 121:79–92.

McDowell, W. H., a. E. Lugo, and a. James. 1995. Exports of nutrients and major ions from Caribbean catchments. J. N. Am. Bonthol. Soc. 14:12–20.

Mcdowell, W. H., and C. E. Asbury. 1994. Export of carbon, nitrogen, and major ions from three tropical montane watersheds. Limnology and Oceanography 39:111–125.

McDowell, W. H., C. G. Sánchez, C. E. Asbury, and C. R. Ramos Pérez. 1990. Influence of sea salt aerosols and long range transport on precipitation chemistry at El Verde, Puerto Rico. Atmospheric Environment Part A, General Topics 24:2813–2821.

Muoghalu, J. I., and A. Oakhumen. 2000. Nutrient content of incident rainfall, throughfall and stemflow in a Nigerian secondary lowland rainforest. Applied Vegetation Science 3:181–188.

Newbold, J. D., B. W. Sweeney, J. K. Jackson, and L. A. Kaplan. 1995. Concentrations and export of solutes from six mountain streams in northwestern Costa Rica. Journal of North American Benthological Society 14:21-37.

Rütting, T., L. N. Cizungu, D. Roobroeck, M. Bauters, D. Huygens, and P. Boeckx. 2014. Leaky nitrogen cycle in pristine African montane Rainforest soil. Global Biogeochemical Cycles 29:962–973.

Schrumpf, M., W. Zech, J. C. Axmacher, and H. V. M. Lyaruu. 2005. Biogeochemistry of an afrotropical montane rain forest on Mt. Kilimanjaro, Tanzania. Journal of Tropical Ecology 22:77.

Schwendenmann, L., and E. Veldkamp. 2005. The Role of Dissolved Organic Carbon , Dissolved Organic Nitrogen , and Dissolved Inorganic Nitrogen in a Tropical Wet Forest Ecosystem:339–351.

Sigha-Nkamdjou, L., C. Galy-Lacaux, V. Pont, S. Richard, D. Sighomnou, and J. P. Lacaux. 2003. Rainwater chemistry and wet deposition over the equatorial forested ecosystem of Zoétélé (Cameroon). Journal of Atmospheric Chemistry 46:173–198.

Silver, W. L., A. W. Thompson, A. Reich, J. J. Ewel, and M. K. Firestone. 2005. Nitrogen cycling in tropical plantation forests: potential controls on nitrogen retention. Ecological Applications 15:1604–1614.

Appendices

148

Sotta, E. D., M. D. Corre, and E. Veldkamp. 2008. Differing N status and N retention processes of soils under old-growth lowland forest in Eastern Amazonia, Caxiuan??, Brazil. Soil Biology and Biochemistry 40:740–750.

Taylor, P. G., W. R. Wieder, S. Weintraub, S. Cohen, C. C. Cleveland, and A. R. Townsend. 2015. Organic forms dominate hydrologic nitrogen export from a lowland tropical watershed. Ecology 96:1229–1241.

Templer, P. H., W. L. Silver, J. Pett-Ridge, K. M. DeAngelis, and M. K. Firestone. 2008. Plant and microbial controls on nitrogen retention and loss in a humid tropical forest. Ecology 89:3030–3040.

Veneklaas, E. J. 1990. Nutrient Fluxes in Bulk Precipitation and Throughfall in Two Montane Tropical Rain Forests , Colombia Author ( s ): E . J . Veneklaas Source : Journal of Ecology , Vol . 78 , No . 4 ( Dec ., 1990 ), pp . 974-992 Published by : British Ecological Society S. Journal of Ecology 78:974–992.

Wieder, W. R., C. C. Cleveland, P. G. Taylor, D. R. Nemergut, E. L. Hinckley, L. Philippot, D. Bru, S. R. Weintraub, M. Martin, and A. R. Townsend. 2013. Experimental removal and addition of leaf litter inputs reduces nitrate production and loss in a lowland tropical forest. Biogeochemistry 113:629–642.

Williams, M. R., T. R. Fisher, and J. M. Melack. 1997. Solute dynamics in soil water and groundwater in a central Amazon catchment undergoing deforestation. Biogeochemistry:303–335.

Wl, S., T. Aw, R. a, E. Jj, F. Mk, S. Division, L. Selva, B. Station, P. Southwest, W. L. Silver, A. W. Thompson, A. Reich, J. J. Ewel, and M. K. Firestone. 2005. Nitrogen cycling in tropical plantation forests: potential controls on nitrogen retention. Ecological Applications 15:1604–1614.

Yusop, Z., I. Douglas, and A. R. Nik. 2006. Export of dissolved and undissolved nutrients from forested catchments in Peninsular Malaysia. Forest Ecology and Management 224:26–44.

Acknowledgements

149

Acknowledgements

The acknowledgements, a very popular part of the doctoral thesis. I better make it worth it!

This journey has started on the reception of the graduation of my Master degree at Ghent

University. Hans, you were my MsC thesis promoter, and I still remember how things got

started (the conversation might deviate from the actual one, but who cares):

H: ‘Ah Marijn, you’re looking for something temporarily? I think Pascal (ed. who I never

met because I skipped all his courses in the bachelor degree. Sorry for that, I was in ‘Land

and bos’, aka ‘Laat en Bros’) has some money to work in the Congo.

*Grabbing three beers*

H: ‘Pascal, this is Marijn, he did his thesis with me on Ecuadorean cloudforest. Do you still

have that spare money for a position on the BELSPO project?’

P: ‘Yes, I do. But I wanted a post-doc actually. (Then to me, quite direct and blunt) And

you, will you be able to work in the DRC? You realize it’s not like working in the forest here

in Belgium?’

I leave a silence and think to myself: I need to show this guy some attitude, some

backbone. I first drink half of my beer in one sip, and then answer;

M: ‘Let’s say I travelled a lot, and the less time I spent in civilized environments, the more

I tend to like the adventure. Yes I would like it. How’s the Congolese beer?’

Pascal drinks his bottle in one sip. Hans downs his one as well, and orders three more.

P: ‘You don’t mind the taste with temperatures like that during the day!’ Leaves silence.

‘You know it’s only for 1 year or so?’

M: ‘I don’t mind for now... I’m leaving for Mexico for a couple of months. Can we kick-off

in January?’

P and H: ‘Ok, we’ll send the mail for the contract next week, than you sign before leaving.’

Great memory. I ended up staying longer, applying for extra money, and doing more than

anticipated. To my two promoters, Pascal and Hans, who are the first and foremost reason this

booklet even exists. It has been fantastic, and I feel very lucky, to have worked with the both

of you. Two different characters, making up to a nice complementary combo. Thank you for

giving me the freedom (!!!) to do all this stuff, but also for the great ideas and guidance when

needed! I think together we turned this into a nice PhD. I would also like to thank the jury

members and the chairman for the qualitative and timely review of my work, it has made the

final document much better.

Secondly, I would like to thank both of the groups I worked in, ISOFYS and CAVElab; a nice

working environment has been a key and important driver for me. I cannot begin to list all the

things I appreciate so much, or my acknowledgements would get longer than any of the

chapters.

Isofys: Samuel, Rataplan, my russian dancing queen with an amazing opera voice, I am sorry

for stealing your bike 10 years ago. If I would have known you back then, I would not have

done it. Thanks for the many moments of thinking together, challenging me to understand

things better. Mama Katja, you’re the best. Thanks for the ‘vogelnestjes’, thanks for the

Acknowledgements

150

support, thanks for listening... Stijntje, ‘beter aon je messink dan aon je toafel!’, the big friendly

giant, how many samples together? How many samples have you pushed through for me?

Analytisch! Lientje, my partner in crime at ISOFYS, the end is near hang on! Hari Ram –the

guru- Upadhayay, I will always be the foremost follower on your Youtube channel! Pedrito –

Vicuñaslayer, you have taught me the real meaning of cryptic; ‘Water isotopes... It’s like pulling

a dead cat out of the sewer!’. Saskia, the godess of administration, I cannot express my

gratitude for helping me to export yet another 200 kg of material to the DRC. Thanks for all the

help, and for the nice chats we have! Eric, thank you for all the help in the lab, for being always

friendly and nice. Joseph and Benjamin, there is a god, let’s have faith that our paths will cross

many more times. Last but not least; I really really enjoyed the discussion with the post-docs

in the early years of my PhD; Dries, Sebastian, Marco and Samuel. I think the atmosphere in

the office with you guys around in the beginning has really kickstarted my PhD. It was nice to

turn around on my chair and throw a question at you, often resulting in long discussions on

something else. I have learned a lot from your attitude in science.

Cavelab: crazy how this very young and new group has worked out to be such a nice and

warm aggregation of people. Hannes – the legend- aka the sick mind. You truely are a legend,

and I hope that I can enjoy you’re company for much longer. Marie – crazy cat lady – thanks

for taking care of Bertie, my wookie friend, I hope you stick around in Ghent much longer!

Sruthi – AA- thanks for being so cute and funny when drunk. And thanks for again and again

trying to drink too much Karmeliet, stating you can take it. Manfie, my prince, thanks for

crashing at our place (at everyones place for that matter), your company is always much

appreciated. Francis, we go way back, and I truely hope we will continue our friendship accross

borders one day! Elizabeth, thanks for smoothing the path in Yangambi, thanks for the

discussions! Pui Yi and Ann, thanks for always being nice, and making sure all administration

is in order! To all the great newer members; Wim, Kim, Felicien, Miro; rest assure that you

have happen to work in the craziest group UGent has seen. I hope we too can collaborate

much more in the future!

On the Congolese side of this world, in the field, there are also many acknowledgements in

place. At first, Prof. Landry Cizungu Ntaboba, my local guidance: Landry, you are the one that

has shown me how to do things in the DRC. I think being around you in the field has taught

me invaluable lessons which have made this PhD possible. I hope we will continue to

collaborate, have beers, and discuss much more. Franche collaboration! To the dreamteam of

my first fieldcampaign; Thales – the captain – de Haulleville, Giacomo – SIF – Sellan and

Emmanuel – mon conseiller – Kasongo, you guys rock and it was amazing being in the DRC

for the first time with you. To you, Isaac Makelele, who has done so much efforts for research!

I hope I can ever meet another Congolese with your motivation for science. I am thrilled that

your efforts have paid off, and that you will start your own career. You are the future of Congo

Biogeochemistry research. To the current Congo biogeochemistry team, I was super excited

to meet young guys doing biogeochemical research in the DRC (after having had the feeling

of being alone for two years); Matti and Travis, I am very very certain that our skypes,

whatsapps, fieldcampaigns and late night drinking in Belgium/US/Switzerland will continue, R-

O-C-K in the D-R-C-A!! To all the other local heroes Héritier Ololo Fundji, Prof. Faustin

Boyemba, Bernard Bonyoma, Henri Badjoko, Kibinda, Para, Dikiss, Michel, Bernard Grimpeur

(Limama), Nolema, Emmanuel (bukavu), Moise, Lambert, Kizito... Thank you for all the

support, nice chats, endless discussions about money, all the beers together, the legendary

Yangambi party...

I also want to thank my funders, Belgian Science Policy (BELSPO) and VLIR-UOS. I hope this

thesis has been worth funding in your eyes. Thank you for the opportunity! Thanks also to

Plantentuin Meise for the partial funding of a fieldcampaign, to Parc National Kahuzi-Biéga,

Acknowledgements

151

Nyungwe park and INERA Yangambi, IITA, etc. Of course, I also thank all the other

collaboraters with whom I shared good moments: Tobias Rütting, Geert Baert, Luc Willems,

Greet de Bruyn, Evy, Kristof van Oost, etc. And all the MsC students I worked with, it was a

real pleasure: Miro, Stijn, Dries, Cys, Oscarito, Kasper, Evelien, Francis, etc.

The success of a PhD is also, especially with field work, owing to a part of personality, the way

of dealing with people, the way of thinking... This I owe to all my family and friends, everyone

really; Jeroen, Fien, Lies, Vojder, the Verdoncks, Hannie, all the boyzz and girlz in the hood!

Having beers from time to time is important for me. In special, I want to thank my mother Iris,

who has done several PhD’s in life, and is one of the wisest persons I’ve met. Thanks to

Andreas, a fallen hero, who is always close, no matter where I travel.

Last but not least, I want to thank my companion in life, Marie-leen Verdonck. You are now

allowed to throw my laptop out of the window! The fact that you still want to stick around after

my endless times abroad or behind my laptop, my tuning out moments, my lack of affinity with

housekeeping stuff etc. must mean you really like me . Words cannot express my gratitude

for having you by my side every day of the year, so let’s go for something simple: thank you

for all your support and love! You’re the best!

Academic Curriculum vitae

152


Personal information

Name Marijn Bauters Date of birth 23 June 1990 Place of birth Kortrijk Nationality Belgian Marital status Voluntary prisoner of Marie-Leen Verdonck Home address Albrecht Rodenbachstraat 129

B-9040 St Amandsberg +32 483 45 29 72 [email protected]

Work address Coupure Links 653 B-9000 Gent +32 9 264 60 06 [email protected]

Education

2013-2017 PhD in Bioscience Engineering – ISOFYS and CAVE lab, Ghent University

2011-2013 Master in Bioscience Engineering: Soil and Water management (Great Distinction)

2008-2011 Bachelor in Bioscience Engineering (Distinction) 2002-2008 Latin-Mathematics (4 yr) + Sciences-

Mathematics (2 yr), Spes Nostra Heule

Experience (fieldwork) abroad

DRC – Kisangani, Yoko, Yangambi, Djolu, Bukavu, Kahuzi-Biéga

Rwanda – Nyungwe

Uganda – Kibale, Rwenzuri

Ecuador – Pichinchia and Imbabura

French Guiana – Paracou

Scientific contributions

Artikels / Articles - A1

Bauters, M., T. W. Drake, H. Verbeeck, S. Bodé, P. Hervé-Fernandez, P. Zito, D. C. Podgorski, F. Boyemba, I. Makelele, L. C. Ntaboba, R. G. M. Spencer, and P. Boeckx. 2018. High fire-derived nitrogen deposition on central African forests. Proceedings of the National Academy of Sciences. Early view.

Bauters, M., H. Verbeeck, M. Demol, S. Bruneel, C. Taveirne, D. Van der Heyden, L. Cizungu, and P. Boeckx. 2017. Parallel functional and stoichiometric trait shifts in South American and African forest communities with elevation. Biogeosciences 14:5313–5321.


153

Bauters, M., H. Verbeeck, S. Doetterl, E. Ampoorter, G. Baert, P. Vermeir, K. Verheyen, and P. Boeckx. 2017. Functional Composition of Tree Communities Changed Topsoil Properties in an Old Experimental Tropical Plantation. Ecosystems 20:861–871.

Bauters, M., N. Mapenzi, E. Kearsley, B. Vanlauwe, and P. Boeckx. 2016. Facultative nitrogen fixation by legumes in the central Congo basin is downregulated during late successional stages. Biotropica 48:281–284.

Bauters, M., E. Ampoorter, D. Huygens, E. Kearsley, T. De Haulleville, G. Sellan, H. Verbeeck, P. Boeckx, and K. Verheyen. 2015. Functional identity explains carbon sequestration in a 77-year-old experimental tropical plantation. Ecosphere 6:1–11.

Doetterl, S., Kearsley, E., Bauters, M., Hufkens, K., Lisingo, J., Baert, G., Verbeeck, H. & Boeckx, P. (2015) Aboveground vs. Belowground Carbon Stocks in African Tropical Lowland Rainforest: Drivers and Implications. Plos One, 10, e0143209.

Rütting, T., Cizungu, L.N., Roobroeck, D., Bauters, M., Huygens, D. & Boeckx, P. (2014) Leaky nitrogen cycle in pristine African montane Rainforest soil. Global Biogeochemical Cycles, 29, 962–973.

Sellan, G., Simini, F., Maritan, A., de Haulleville, T., Bauters, M., Beeckman, H. & Anfodillo, T. (2017) Testing a general approach to assess the degree of disturbance in tropical forests. Journal of Vegetation Science; 1-10

Artikels / Articles – A3 Bauters, M.; Strubbe, M.; Kearsley, E.; Steppe, K. & Verbeeck, H. (2013) Herbebossen in de Andes: op zoek naar koolstofvoorraden in Noord-Ecuador. Bosrevue, 45, 1–5. Bruneel, S., Demol, M., Verbeeck, H. & Bauters, M. In de mist van het nevelwoud. Bosrevue 57, (2016).

Artikels / Articles – A4

Bauters, M.; Spiessens, B.; Van Den Berge, S.; Verdonck, M.-L.; Volckaert, M. & Verheyen, K. (2011) Ecosysteemdiensten in de praktijk : een studie voor de bossen van Geel-Bel. ANTenne, 5, 13–17.

Posters/ Posters

Bauters, M.; Verbeeck, H.; Kearsley, E.; Steppe, K.; Kwantificeren van koolstofopslag in

tropisch regenwoud in Ecuador, Starters in het bosonderzoek 2013, Brussels.

Bauters, M.; Bodé, S.; Boeckx, P.; The use of isotopic techniques for nitrogen fixation

assessment, BASIS-conference 2015, Utrecht.

Verbeeck, H.; Kearsley, E.; Bauters, M.; Beeckman, H.; Huygens, D.; Steppe, K.; Boeckx, P.; Linking carbon storage with functional diversity in tropical rainforest in the central Congo Basin, EGU General Assembly 2015, Vienna. Bauters, M.; Bruneel, S.; Demol, M.; Taveirne, C.; Van der Heyden, D.; Boeckx, P.; Kearsley, E.; Cizungu, L.; Verbeeck, H.; Cross-continental comparison of the functional composition and carbon allocation of two altitudinal forest transects in Ecuador and Rwanda. , EGU General Assembly 2016, Vienna. Bauters, M.; Verbeeck, H.; Cizungu, L.; Mujunya, B.B.; Boyemba, F.; Baert, G.; Verheyen, K.; Boeckx, P.; A permanent monitoring network for N-fluxes in different forest types in the central Congo Basin; ATBC Annual Meeting 2016


154

Bauters, M.; Verbeeck, H.; Boeckx, P.; The nitrogen budget for different forest types in the central Congo Basin, EGU General Assembly 2017, Vienna. Bauters, M.; Verbeeck, H.; Bruneel, S.; Demol, M.; Taveirne, C.; Van der Heyden, D.; Boeckx, P.; Kearsley, E.; Cizungu, L.; Verbeeck, H.; Cross-continental comparison of the functional composition and carbon allocation of two altitudinal forest transects in Ecuador and Rwanda. EGU General Assembly 2017, Vienna Bauters, M.; Verbeeck, H.; Bruneel, S.; Demol, M.; Taveirne, C.; Van der Heyden, D.; Boeckx, P.; Kearsley, E.; Cizungu, L.; Verbeeck, H.; Cross-continental comparison of the functional composition and carbon allocation of two altitudinal forest transects in Ecuador and Rwanda. ATBC Annual Meeting 2017

Presentaties/Orals

Bauters, M.; Verbeeck, H.; Boeckx, P.; The nitrogen budget for different forest types in the central Congo Basin, EGU General Assembly 2016, Vienna. Bauters, M.; Verbeeck, H., Ampoorter, E., Döetterl, S., Baert, G., Verheyen, K. , Boeckx, P. ; Tree species effects on topsoil properties in an old tropical plantation, EGU General Assembly 2016, Vienna. Bauters, M.; Verbeeck, H.; Cizungu, L.; Boeckx, P.; Towards a nitrogen budget for different forest types of the central Congo Basin, International Nitrogen Initiative 2016, Melbourne Verbeeck, H., Bauters, M., Bruneel, S., Demol, M., Taveirne, C., Van der Heyden, D., Kearsley, E., Cizungu, L. and Boeckx, P.; Cross-continental comparison of the functional composition and carbon allocation of two altitudinal forest transects in Ecuador and Rwanda, GTOE meeting 2016, Brussels Kearsley, K., Hufkens, K., Verbeeck, H., Bauters, M., Beeckman, H., Boeckx, P. and Huygens, D.; Rare tree species support functional diversity in resource-acquisition in central African forests, ATBC annual meeting 2016, Montpellier Verbeeck, H., Bauters, M., Bruneel, S., Demol, M., Taveirne, C., Van der Heyden, D., Kearsley, E., Cizungu, L. and Boeckx, P.; Cross-continental comparison of the functional composition and carbon allocation of two altitudinal forest transects in Ecuador and Rwanda, ATBC annual meeting 2016, Montpellier Bauters, M. Nitrogen cycle in central African Forests, Götenburg University, invited seminar, 2017 Bauters, M., Bodé, S., Rütting, T., Verbeeck, H., Cizungu Ntaboba, L., Boeckx, P.; A 15N tracing technique to disentangle gross N dynamics in tropical forest soils, BASIS meeting, Utrecht, 2017 Bauters, M., Verbeeck, H., Makelele I., Cizungu Ntaboba, L., Boeckx, P.; High fire-derived N deposition in central African forests, iLEAPS conference 2017, Oxford

Contributions to Internal reports

BELSPO – COBIMFO final project report

VLIR-UOS – FORMONCO


155

Reviewing

European journal of Soil Science (6), Biogeosciences (1), Plant Ecology and Evolution (1),

Agriculture Ecosystems and Environment (4), Global Change Biology (1), Global Ecology

and Conservation (1).

Tutoring work

Browaeys, K. Using drone-based imagery as a tool for reforestation success assessment in Northern Ecuador, Master thesis, Ghent Univ. 2017-2018.

Van Gansbeke S. Koolstof opslag in natuurlijke bossen langsheen een hoogtegradient in de Rwenzuri, Oeganda. Bachelorproef, HoGent. 2017.

Desloover A. Functionele diversiteit in natuurlijke bossen langsheen een hoogtegradient in de Rwenzuri, Oeganda. Bachelorproef, HoGent. 2017.

Colman W. Kwantificeren van stikstofdepositie, uitspoeling en emissie in verschillende ecosystemen. Bachelorproef, Odysee. 2017.

Vercleyen O. Carbon accrual, community assembly and diversity recovery along a successional gradient in the tropical rainforest of the central Congo Basin. Master thesis, Ghent Univ. 2017.

Thierens E. Functional Identity and Reforestation Success of Different Tree Species in the Northern Andes Range of Ecuador. Master thesis, Ghent Univ. 2017.

Bulonza ME. Evaluation des dépositions d’azote dans les forêts humides de haute altitude ‘cas de la forêt mixte du parc national de Kahuzi-Biéga). Bachelor thesis, Univ Cath Bukavu. 2016

Makelele IA. Le cycle des nutriments dans les forêts tropicales du bassin du Congo: Contribution du pluvioléssivage et évaluation des pertes par écoulement de rivière au sein d’un bassin versant de la réserve de Yoko. Master thesis, Univ Kisangani. 2016.

Van der Heyden D. Carbon storage and nutrient shifts along an altitudinal gradient in Nyungwe forest , Rwanda. Master thesis, Ghent Univ. 2016.

Demol M. Functional diversity in natural forests along an altitudinal gradient in Northern Ecuador. Master thesis, Ghent Univ. 2016.

Taveirne C. Functional diversity study on an altitudinal forest transect in Central Africa, Nyungwe National park, Rwanda. Master thesis, Ghent Univ. 2016.

Bruneel S. Carbon sequestration in natural forests and reforestations along an altitudinal gradient in the Andes of Ecuador. Master thesis, Ghent Univ. 2016.

Mumbanza FM. Inter-and intra-species leaf trait variability in a planted rainforest in

Yangambi (D.R. Congo). Master thesis, Ghent Univ. 2014.

biogeochemical cycles in contrastingéén na grootste tropisch regenwoud ter wereld. in dit...

Documents