Evolution of Carbon Stabilization Mechanisms in a Podzolic Soil Chronosequence

Faculté d'Ingénierie Biologique, Agronomique et Environnementale

Evolution of carbon stabilization mechanisms in a podzolic soil chronosequence

Marie-Liesse Vermeire Louvain-la-Neuve, mars 2017

Thèse présentée en vue de l’obtention du grade de docteur en sciences agronomiques et ingénierie biologique

Membres du Jury :

Président Prof. Marnik Vanclooster Promoteurs Prof. Jean-Thomas Cornélis Prof. Bruno Delvaux

Lecteurs Prof. Steeve Bonneville Dr. Sophie Cornu Dr. Delphine Derrien Prof. Kristof Van Oost

« N’allez pas là où le chemin peut mener. Allez là où il n’y a pas de chemin et laissez une trace. » Ralph Wardo Emerson

« Commence par faire le nécessaire, Puis fais ce qu’il est possible de faire Et tu réaliseras l’impossible sans t’en apercevoir » Saint François d’Assise

Table of contents REMERCIEMENTS ...... I RÉSUMÉ ...... VII SUMMARY ...... IX PART 1: INTRODUCTION AND OBJECTIVES ...... 1

I. GLOBAL CONTEXT ...... 1 II. OBJECTIVES...... 6 III. THESIS OUTLINE ...... 10 PART 2: GENERAL OVERVIEW ...... 13

CHAPTER 1 - SOIL ORGANIC MATTER ...... 13 1.1 Definition ...... 13 1.2 Origin and composition ...... 13 1.2.1 Plant residues ...... 15 1.2.2 Soil microbial biomass ...... 16 1.2.3 Plant and microbe exudates ...... 18 1.3 Decomposition and transformation ...... 19 CHAPTER 2 - SOM PROTECTION MECHANISMS ...... 25 2.1 Chemical structure and recalcitrance ...... 26 2.1.1 Chemical recalcitrance of organic constituents ...... 26 2.1.2 A concept not sufficient alone ...... 28 2.1.3 Persistence as an ecosystem property ...... 29 2.1.4 Recalcitrance role in the degradation continuum ...... 30 2.2 Inhibition of decomposers activity ...... 34 2.2.1 Presence of the adequate microorganisms ...... 34 2.2.2 Habitat constraints ...... 35 2.2.2.1 Environmental constraints ...... 35 2.2.2.2 Accessibility of the substrate ...... 38 2.2.3 Substrate constraint ...... 40 2.3 Adsorption onto minerals and coprecipitation ...... 43 2.3.1 Mineral transformations during pedogenesis ...... 43 2.3.1.1 Weathering of primary minerals ...... 43 2.3.1.2 Formation of secondary minerals ...... 45 2.3.2 Minerals involved in OMA ...... 48 2.3.3 Properties of OM ...... 51 2.3.4 Bindings types ...... 53 2.3.4.1 Ligand exchange ...... 53 2.3.4.2 Outersphere complexation ...... 54 2.3.4.3 Polyvalent cation bridges ...... 54 2.3.4.4 Weak interactions ...... 54 2.3.5 Conceptual model of OMA structure ...... 55 2.3.6 Controls on adsorption/coprecipitation processes ...... 58 2.3.7 Role of OMA in soil carbon storage and preservation ...... 61 2.3.7.1 Evidences ...... 61 2.3.7.2 Mechanisms ...... 63 2.3.8 Interaction OMA - microorganisms ...... 63 2.3.8.1 Via the organic phase ...... 64 2.3.8.2 Via the mineral phase ...... 65 2.3.9 Difficulty to isolate OMA and methodological choices ...... 69 2.4 Shematic synthesis ...... 72 CHAPTER 3 - PEDOGENIC MODEL: PODZOL...... 75 3.1 Podzol description and repartition ...... 75 3.2 Podzol formation ...... 76 PART 3. EVOLUTION OF C STABILIZATION MECHANISMS IN A PODZOLIC CHRONOSEQUENCE ...... 79

CHAPTER 4 - DESCRIPTION OF THE STUDY SITE ...... 79 4.1 Geography and climate ...... 79 4.2 Geological context ...... 80 4.3 Soil chronosequence sampling ...... 83 4.4 Analytical methods ...... 83 4.4.1 Major soil properties ...... 85 4.4.2 Mineralogical characterizations ...... 85 4.4.3 Rare Earth Elements analysis and data treatment ...... 87 4.4.4 Soil organic matter distribution and composition ...... 90 4.4.5 Soil microorganisms ...... 91 4.4.5.1 Compound specific analysis of amino-sugars ...... 91 4.4.5.2 PLFA extraction and quantification ...... 93 4.4.6 Soil respiration measurements ...... 94 4.4.7 Fe and Si isotopes ...... 95 4.5 Pedological context...... 95 CHAPTER 5 - EVOLUTION OF THE SOIL ENVIRONMENT ...... 99 5.1 Field observations ...... 99 5.2 Micromorphological observations ...... 100 CHAPTER 6 - MINERALOGICAL EVOLUTIONS ...... 103 6.1 Description of the secondary mineral phases ...... 104 6.2 Rare Earth Elements dynamics along pedogenesis in a chronosequence of podzolic soils ...... 108 6.2.1 Introduction ...... 109 6.2.2 Material and methods ...... 111 6.2.3 Results and discussion ...... 111 6.2.4 Conclusion ...... 130 6.3 Synthesis...... 132 6.3.1 Validation of the chronosequence...... 132 6.3.2 Characterization of the parent material ...... 132 6.3.3 Processes involved in podzolization ...... 133 6.3.3.1 Acidification and weathering ...... 133 6.3.3.2 Fe, Al and Si mobilization and accumulation ...... 135 CHAPTER 7 - EVOLUTION OF SOM, MICROBIAL POPULATIONS AND ORGANO-MINERAL ASSOCIATIONS ...... 137 7.1 Development of a C accumulation horizon ...... 138 7.2 Evolution of OM composition and quality ...... 138 7.2.1 Bulk SOM composition ...... 138 7.2.2 Evolution of the repartition of C within fractions ...... 142 7.3 Evolution of the microbial populations ...... 145 7.4 Evolution of total respiration and C decomposability ...... 149 CHAPTER 8 - SYNTHESIS: CO-EVOLUTION OF MINERAL PHASES, MICROBIAL POPULATIONS AND SOM ...... 155 8.1 P1-120 yrs and P2-175 yrs – colonization ...... 156 8.2 P3-270 yrs – transition ...... 157 8.3 P4-330 yrs and P5-530 yrs – developed podzols ...... 159 8.3.1 Litter degradation and enhanced mineral weathering in the eluvial (E) horizon ...... 159 8.3.2 Translocation of organic and inorganic elements ...... 162 8.3.3 OMA accumulation in the illuvial B horizon ...... 162 8.3.3.1 Secondary minerals and OM precipitation ...... 162 8.3.3.2 Dynamics in the illuvial B horizon ...... 165 PART 4 - OMA DYNAMICS IN ANOXIC CONDITIONS ...... 171

CHAPTER 9 - MICROBIAL FE(III) REDUCTION OF ORGANO-MINERAL ASSOCIATIONS ALONG A PODZOLIC SOIL CHRONOSEQUENCE: IMPACT ON C RELEASE? ...... 172 9.1 Introduction ...... 173 9.2 Materials and methods...... 175 9.2.1 Soil samples ...... 175 9.2.2 Chemical extractions: Fe and C pools in soil samples ...... 177 9.2.3 Microbial incubations ...... 178 9.2.4 Scanning Electron Microscopy (SEM) ...... 180 9.3 Results and discussion ...... 181 9.3.1 Characterization of Fe and C pools ...... 181 9.3.2 Microbial Fe(III) reduction: rate and extent ...... 185 9.3.3 Influence of microbial Fe(III) reduction on OM release ...... 192 9.3.4 Influence of OM on microbial Fe(III) reduction ...... 196 9.3.5 Implications for podzolization processes ...... 199 PART 5: GENERAL CONCLUSION AND PERSPECTIVES ...... 201

I. GENERAL CONCLUSION...... 201 II. PERSPECTIVES ...... 209 III. FINAL WORDS ...... 214 PART 6: APPENDIX ...... 215

A1. RELATED PUBLICATIONS AND COLLABORATIONS ...... 217 A1.1 Conference proceedings ...... 217 A1.2 Publications ...... 218 A1.3 Participation to conferences and workshops ...... 218 A1.4 Research stays and field campaign ...... 219 A1.5 Collaboration with other Belgian universities ...... 220 A2. CO-AUTHORED PAPERS ...... 221 A2.1 - Silicon isotopes record dissolution and re-precipitation of pedogenic clay minerals in a podzolic soil chronosequence ...... 221 A2.1.1 Introduction ...... 222 A2.1.2 Materials and methods ...... 225 A2.1.2.1 Sample collection and location ...... 225 A2.1.2.2 Physico-chemical characterizations ...... 228 A2.1.2.3 X-ray diffraction patterns ...... 228 A2.1.2.4 Isotopic and geochemical analyses ...... 229 A2.1.3 Results ...... 230 A2.1.3.1 Soil Mineralogy...... 230 A2.1.3.2 Si isotopic modifications in the clay fraction over time ...... 237 A2.1.3.3 Geochemical modifications in the clay fraction over time...... 240 A2.1.4 Discussion ...... 240 A2.1.4.1 Evolution of clay-sized mineralogy ...... 240 A2.1.4.2 Dissolution and re-precipitation of pedogenic clay minerals during podzolization ...... 243 A2.1.4.3 Implications for podzolization theory ...... 249 A2.1.4.4 Implications for tracing the effects of environmental changes on soils ...... 250 A2.1.5 Conclusions ...... 251 A2.2 Can Fe isotope fractionations trace the pedogenetic mechanisms involved in podzolization? ...... 253 A2.2.1 Introduction ...... 254 A2.2.2 Sampling and methods ...... 255 A2.2.2.1 Site and sampling ...... 255 A2.2.2.2 Soil properties ...... 256 A2.2.2.3 Iron isotope analyses ...... 257 A2.2.2.4 Mass balance calculations ...... 258 A2.2.3 Results ...... 259 A2.2.3.1 Depth evolution of soil properties along the chronosequence ...... 259 A2.2.3.2 Fe partial extractions ...... 262 A2.2.3.3 Fe isotopes depth evolution along the soil sequence ...... 262 A2.2.4 Discussion ...... 263 A2.2.4.1 Lateral and vertical variations in the soil parent sediment prior to pedogenesis ...... 263 A2.2.4.2 Dynamics of different processes involved in podsolization ...... 269 A2.2.4.3 Fe isotopic fractionation as a function of the different podzolization mechanisms ...... 274 A2.2.5 Conclusion...... 275 A3. CARBON STABLE ISOTOPES ...... 277 A4. SOMFRAC RING TRIAL ...... 278 REFERENCES...... 281

Remerciements

Une phrase m’est fréquemment revenue à l’esprit, à l’heure d’écrire mes remerciements pour toutes les personnes qui ont rendu cette thèse possible ou qui en ont fait une expérience si riche. Cette phrase de Forrest Gump : "Je ne sais pas si on a chacun un destin, ou si on se laisse porter par le hasard comme sur une brise. Je crois que c'est peut-être un peu des deux".

Destin … hasard … je ne sais pas non plus. Mais il semblerait bien qu’une brise ait soufflé.

Je n’aurais jamais pensé, en poussant la porte du bureau de Monsieur Bruno Delvaux en 2008, à la recherche d’un projet de mémoire, que je me retrouverais à écrire ces lignes, presque 10 ans plus tard. C’est pourtant à la suite de cette réunion passionnante que je me suis lancée à la découverte du silicium et du bananier, dans le superbe cadre qu’est la Guadeloupe. Ce mémoire a initié mon gout pour la science du sol, et permis de rencontrer mon promoteur actuel, Jean Thomas Cornélis, terminant à l’époque sa thèse. Merci à Monsieur Delvaux de m’avoir ouvert les portes du labo SOLS et rendu possible les rencontres qui m’ont menée jusqu’ici. Merci également pour son encadrement en tant que co-promoteur au cours de mes deux dernières années de thèse.

Je tiens à adresser un merci tout particulier à Jean-Thomas Cornélis, qui a su me convaincre de commencer cette thèse alors que j’avais quitté l’université pour d’autres pâturages. Merci de m’avoir permis de vivre ce qui est pour l’instant les quatre plus belles années de ma vie, et révélée à ce qui est maintenant pour moi une passion : la recherche. Depuis ces premières discussions sur matière organique et pédogenèse, son enthousiasme communicatif a toujours su me motiver, ou me remonter le moral, même dans les plus gros creux indissociables du parcours de thèse. Je le remercie pour sa porte toujours ouverte, son soutien, et pour nos riches discussions scientifiques. Ses qualités humaines remarquables font de lui un

créateur de lien, et ses joyeux coups d’éventail ont souvent alimenté la brise qui a soufflé tout au long de ma thèse.

Merci également à Pierre Delmelle, qui a accepté d’être mon co-promoteur lors des deux premières années de thèse, et qui m’a fait bénéficier de sa précieuse expérience en recherche et consacré de son temps.

Merci à Kristof Van Oost pour avoir été, avec Jean-Thomas et Monsieur Delvaux, à l’origine de ce projet FRFC du Fonds National de la Recherche Scientifique (F.R.S.-FNRS, Belgium), financement dont j’ai bénéficié durant ces quatre années et favorisant la recherche collective. Merci également pour sa participation à mon encadrement, et pour les rencontres lors de réunions ou conférences qui ont ponctué ma thèse. Merci à Marnik Vanclooster d’avoir accepté d’être président du Jury.

Merci également au FNRS, pour avoir financé ce projet ainsi que les participations à des congrès scientifiques et séjours à l'étranger, si enrichissants pour un chercheur et nécessaires au partage des savoirs.

Mon encadrement de thèse s’est également enrichi de nouvelles collaborations, et je voudrais remercier plusieurs personnes- clé par volet de recherche, qui m’ont énormément apporté.

Après avoir roulé ma bosse à l’UCL pour réaliser les caractérisations de base de mes précieux échantillons de Vancouver, une bourrasque m’a transportée vers le labo ISOFYS de l’Université de Gand. C’est avec Sebastian Doetterl que j’ai découvert le monde des microorganismes du sol. Merci à lui pour son aide et encadrement pour les analyses AS et PLFA, mais aussi pour m’avoir ouvert sa maison lors des quelques séjours à Gand et pour les nombreux bons moments autour d’un verre ou en conférences. Nos discussions m’ont apporté beaucoup. Merci également à Pascal Boeckx, pour m’avoir ouvert l’accès à son labo, et à Samuel Bodé pour son aide lors des manips et ses analyses de mes échantillons.

C’est ensuite un vent de mistral qui m’a menée jusqu’à Aix-en- Provence et Marseille, pour étudier les terres rares au labo de Géochimie des sols et des eaux de l’INRA avec Sophie Cornu et Zuzana Fekiacova. Merci à Sophie d’avoir accepté de participer à mon encadrement de thèse, et pour son accompagnement lors de la rédaction de notre article. J’ai appris énormément en termes de rigueur et méthode d’écriture et de traitement de données. Ses remarques franches mais toujours bienveillantes, m’ont parfois bousculée mais toujours beaucoup fait avancer. Je la remercie également pour m’avoir accueilli chez elle à plusieurs reprises, partagé son quotidien familial, et fais découvrir cette ville passionnante qu’est Marseille. Merci à Zuzana pour m’avoir transmis son expertise et appris la rigueur de la chimie isotopique. Un tout grand merci à elle également pour m’avoir hébergée pendant un mois. Je n’oublierai pas nos échanges culinaires (fromage fumé slovaque versus boulets liégeois), nos fous rires et le partage de nos univers. Merci également aux autres membres du labo pour la bonne ambiance et leur accueil.

Après les terres rares, une bise m’a menée jusqu’à une certaine bactérie rose et fortement odorante, entretenant une liaison sulfureuse avec les oxydes de fer. C’est alors Steeve Bonneville, de l’ULB, qui m’a encadrée pour la dernière tranche de ma thèse. Je tiens à lui adresser mes plus vifs remerciements pour notre étroite collaboration. Merci à lui pour son investissement, ses nombreux conseils, et pour m’avoir initiée à cette nouvelle branche de recherche. Sans sa grande connaissance du sujet et son précieux soutien, ce projet de recherche n’aurait jamais pu aboutir.

Merci également à Delphine Derrien d'avoir accepté d'évaluer ma thèse et nourri nos discussions de remarques et conseils constructifs, tant lors de la défense préliminaire que lors de conférences ou autres réunions. La brise ne m’a pas encore conduite jusqu’à Champenoux, mais nos agréables échanges m’ont toujours laissé un gout de trop peu ... sait-on jamais où les vents nous mènent ?

iii

Merci aux autres personnes qui ont joué un rôle dans ce travail de recherche : Laurence Ryelandt pour les analyses SEM-EDX, et Marko Bravin pour les analyses carbone. Je voudrais également remercier les membres du labo GEBI, qui m’ont accueilli pour les manips microbio, en particulier Benoît Stenuit et Thomas Nicolay pour leurs précieux conseils. Je remercie tous les membres du projet SOGLO pour la merveilleuse expérience de terrain que nous avons partagée au Brésil.

C’est maintenant une tempête d’autres visages qui me viennent en tête, et qui ont également marqué ces quatre dernières années. J’ai eu la chance de rencontrer beaucoup de personnes formidables dans cette université. Je remercie François Wiaux, que j’ai découvert lors d’une mythique conférence en Islande, et devenu depuis un ami proche. Merci à lui pour les repas baleine et chèvrerie (quel bestiaire), et tous les autres qui ont suivi. Sa rencontre m’a énormément enrichie. Merci également aux membres du labo ECAV, nos voisins et spécialistes des plantes, avec qui nous avons partagé les toilettes. A part une obscure histoire de vol de savon, la cohabitation s’est toujours bien passée. Il semblerait que sol et plante soient complémentaires. Qui l’eut crut ? Sol et pierre ne sont pas non plus étrangers… Hasard ? Destin ? Merci à toi, Pierre.

… Et bien sûr l’unité SOLS

Quel labo ! Le « best lab in the world », comme nous l’avons appelé d’un commun accord. Selon un proverbe arabe “L'amour est un caravansérail : on y trouve que ce qu'on y apporte.” Il semblerait que l’environnement dans lequel j’ai eu la chance d’évoluer tout au long de ma thèse, florilège de cultures et de personnalités, est l’exception qui confirme la règle. Chacun apporte tellement, que même lorsque l’un des membres arrive à bout de force, le groupe continue de le porter. Merci aux collègues pour les excellents moments passés ensemble, pour les amitiés qui ont germé, pour le partage de vos expériences et votre précieux soutien tant dans les bons que dans les mauvais moments. Merci à Anne Iserentant et Claudine Givron, pour leur aide au laboratoire, leur disponibilité et grande expérience. Merci à Joseph Duffey, Philippe Sonnet et Sophie Opferghelt pour les

nombreuses discussions intéressantes. Merci aux membres du « bureau du bonheur » (B273, appelé « le poulailler » par les bureaux concurrents jaloux) : Marie Detienne et Yolanda Ameijeiras-Mariño, à mes côtés depuis le début, Carlos Aragon-Tobar et Gabriella Ruiz. Merci aux habitants du « irresistible office », Brieuc Hardy, Inga Tarasenko, Rawaa Ammar, Mathilde Paque, Zimin Li, Elena Maters, qui ont toujours été de valeureux adversaires dans la course à la meilleure ambiance de bureau. Merci aux occupants du « bureau sans nom » (snobisme de leur part), Aubry Vandeuren, Benoit Pereira, François Gaspard, Koffi Tomety-Mensah, Séverin Nijimbere, Clairia Kankurize. Merci également aux mémorants, qui animent couloirs et labos.

Moins classique mais je me devais de le faire : un grand merci au café, carburant et lubrifiant social du chercheur. En moyenne, si l’on considère que le volume moyen d’une tasse de café est de 125 ml, 547 litres ont été nécessaires pour alimenter les quatre ans de travail cérébral ou de labo, et les formidables rencontres et discussions lors des pauses-café.

Enfin, j’aimerais remercier du fond du cœur ma famille et mes amis, sans qui je ne serais rien. Je pourrais couvrir des pages entières, mais me contenterai de dire : merci pour m’avoir toujours soutenue, pour avoir su me guider dans les périodes de doutes, pour toutes les tranches de rires qui rendent la vie belle et folle, et surtout pour tout l’amour que vous m’apportez depuis des années.

Hasard ou destin, je n’en sais toujours pas plus, mais en tout cas, il semblerait que c’est plus en sachant accueillir ce qui nous est offert qu’en s’acharnant à obtenir ce que l’on veut qu’on vit les plus belles expériences.

Merci encore à tous, pour tout ce que vous m’avez offert.

Résumé

Une compréhension des processus de décomposition et stabilisation de la matière organique du sol (MOS) est nécessaire pour prédire l’évolution de son stock. La stabilisation de la MOS résulte de trois mécanismes biologiques et physico-chimiques : (1) la « récalcitrance » de la MOS, due à ses propriétés moléculaires ; (2) l’inhibition de l’activité des décomposeurs ; (3) l’interaction entre la matière organique (MO), les surfaces minérales et les ions métalliques (formation d’associations organo-minérales, AOM). La compréhension de la dynamique de la MOS requiert par conséquent une approche intégrée et pluridisciplinaire, tenant compte des interactions entre minéraux, micro-organismes, MOS et de l’environnement physico- chimique résultant. Les objectifs de cette thèse sont : (1) de développer un schéma conceptuel multidisciplinaire du sol, intégrant les différentes propriétés impliquées dans la stabilisation du carbone (C) ; (2) d’appliquer ce modèle à une chronoséquence de sols podzoliques (profils âgés de 0 à 530 ans), pour étudier l’évolution à court et long terme des mécanismes de protection de la MOS (de quelques jours à plusieurs siècles). Dans la chronoséquence de sols podzoliques, la pédogenèse a commencé par un développement de la végétation dans les plus jeunes profils (P1-120 ans et P2-175 ans), et un input de C depuis la litière. L’activité biologique résultante a entrainé une diminution du pH du sol dans les horizons de surface, tel que mesurée dans l’horizon E du profil P3 (270 ans). L’acidification nette a ensuite engendré des modifications minéralogiques, observables dans les deux podzols matures : P4 et P5 (330 et 530 ans). Ces modifications sont : une altération dans les horizons de surface, et une précipitation de minéraux secondaires, associés à de la matière organique, dans les horizons illuviaux. L’accumulation d’AOM a causé l’obstruction du système poral de l’horizon illuvial, et donc le développement de conditions physico-chimiques différentes. Les mécanismes de vii

protection de la MOS se sont révélés être spécifiques à chaque site et horizon. En effet, la protection de la MO est limitée dans les profils P1 et P2. Une accumulation de composés récalcitrants a été mesurée dans l’horizon superficiel des profils P3, P4 et P5. Enfin, la protection la plus efficace de la MOS a été observée dans les horizons B des profils P4 et P5, de par la formation d’AOM et l’inhibition de l’activité des décomposeurs en raison de la cimentation. Les populations microbiennes étaient dominées par les champignons dans les trois plus jeunes profils (P1 à P3), et dans les horizons de surface (E) des profils P4 et P5 ; tandis qu’elles étaient dominées par les bactéries dans les horizons indurés (B) des profils P4 et P5. Nos résultats confirment que les minéraux, la MOS et les populations microbiennes évoluent de manière interdépendante, ce qui engendre des mécanismes de protection de la MOS variant sur de courtes périodes de temps au cours de la formation du sol. Notre schéma conceptuel pourrait être appliqué à différents types de sols, de manière à mieux comprendre la dynamique de la MOS dans des conditions pédologiques contrastées.

viii

Summary

Predicting the evolution of the stock of soil organic matter (SOM) requires a process-based understanding of its decomposition and stabilization pathways. The stabilization of SOM results from biological and physico-chemical mechanisms: (1) the recalcitrance of SOM due to its molecular properties; (2) the inhibition of decomposers activity; (3) the interaction between SOM, mineral surfaces and metal ions (organo-mineral associations, OMA). The understanding of SOM dynamics thus requires an integrated and multidisciplinary approach, considering the interactions in soils between minerals, microorganisms, SOM and the resulting physico-chemical environment.

The PhD thesis is aimed at: (1) developing a multidisciplinary soil conceptual diagram integrating soil properties involved in carbon (C) stabilization; (2) applying this model to a podzolic soil chronosequence (0 to 530 years old soil profiles), and evaluating the short and long term (from day to century) evolution of SOM protection mechanisms.

In the podzolic soil chronosequence, pedogenesis started with early vegetation development and C input from the litter, as observed in the youngest profiles P1 and P2 (120, 175 yrs). The biological activity led to a soil pH decrease in the surface horizon, as measured in P3 E (270 yrs). Net acidification, incipient in P1 and P2, induced mineralogical modifications, as clearly observed in the well-developed podzols P4 and P5 (330, 530 yrs). These modifications are: mineral weathering in the topsoil, precipitation of secondary minerals, associated with OM, in the illuvial B horizon. The accumulation of OMA led to a cementation in B, and thereby to distinct physico-chemical conditions. SOM protection mechanisms were both site and horizon- specific. Indeed, SOM protection was limited in P1 and P2; recalcitrant C compounds were selectively preserved in the topsoil of P3, P4 and P5; eventually, the largest protection of SOM occurred in the B horizons of P4 and P5 through the formation of OMA and inhibition of the decomposers activity due to cementation. The microbial populations were mainly fungi-dominated in the youngest P1, P2 and

P3, and in the topsoil of P4 and P5. They were, however, mainly bacteria-dominated in the cemented B horizons of P4 and P5.

Our results confirm that minerals, SOM and microbial populations evolve interdependently, resulting in SOM protection mechanisms changing over short-time scales during soil formation. Our conceptual diagram could best be applied to different soil types in order to better assess SOM dynamics in contrasted soil conditions.

Part 1 Introduction and objectives

Part 1: Introduction and objectives

I. Global context

Soil – a precious but threatened resource

"Upon this handful of soil our survival depends. Husband it and it will grow our food, our fuel, and our shelter and surround us with beauty. Abuse it and the soil will collapse and die, taking humanity with it" From Vedas Sanskrit Scripture – 1500 BC

“Essentially, all life depends upon the soil ... There can be no life without soil and no soil without life; they have evolved together.” Charles E. Kellogg, USDA Yearbook of Agriculture, 1938.

Since thousands of years, soils have been recognized as a major resource for mankind. They are the foundation of food, feed, fuel and fiber production, and the main provider of many critical ecosystem services (Ponge, 2015). They are the substrate for agriculture and forest production, the Earth’s largest water filter and storage tank, representing also a protection against erosion and flooding, and they

Part 1 Introduction and objectives host a tremendous diversity of organisms of key importance to ecosystem processes (about ¼ of the global planet biodiversity, FAO status of soils, 2015) (Blum, 2005). Soil is also closely connected to the culture and civilization of an ethnic group living in a given place, including their religion, thoughts, livelihood and health (Minami, 2009). As stated by Minami (2009): “It is important for people to protect the soil, their agriculture and the environment because the collapse of soil leads to the collapse of human culture, civilization, livelihood and health”.

However, this highly valuable natural resource is threatened. Since the industrial revolution and the development of human population, it is under increasing pressure to satisfy demands for food, fiber, drinking water, energy production and raw materials extraction (Kleber and Johnson, 2010, Lal, 2009). Recent studies suggest that the world will need 70 to 100% more food by 2050, to feed the predicted 9 billion of people, requiring a major sustainable intensification of agricultural production (Godfray et al., 2010). In parallel to this increasing demand and competition for soil uses, healthy soil surfaces are decreasing. As discussed in detail in the status of soils report of the FAO (2015), today, 33 % of land is moderately to highly degraded due to the sealing (as cities develop), erosion, salinization, compaction, acidification and chemical pollution of soils (figure 0.1). The geographical footprint of Europe’s cities has increased by nearly 80% since the 1950s, and continues to expand (Banwart, 2011). Worldwide, 2 ha of soils are sealed every minute. Soil losses in some locations of the world are in excess of 50 T.ha-1 in a year, which is up to 100 times faster than the rate of soil formation (Banwart, 2011). Soil can be considered as a non-renewable resource (within a lifespan), as it can take up to 1000 years to produce 1cm of soil. Furthermore, the natural area of productive soils is limited (figure 0.2), and there is little opportunity for expansion in the agricultural area. The current rate of soil degradation threatens the capacity of future generations to meet their most basic needs. Consequently, sustainable management of the world’s soils have become imperative for reversing the trend of soil degradation and ensuring current and future global food security.

Part 1 Introduction and objectives

Figure 0.1 - Status of soil degradation in the world (from Banwart (2011)).

Figure 0.2 – World surfaces suitable for agriculture (Atlas of the Biosphere, Center for sustainability and the Global Environment, University of Wisconsin, Madison).

Part 1 Introduction and objectives

Soil organic matter – key factor for soil functions

Despite its often minor contribution to the total mass of mineral soils, soil organic matter (SOM) is critical for soil function and soil quality. Through its influence on a variety of soil properties and ecosystem processes, SOM is fundamental for soil fertility. As described by Baldock and Broos (2012), SOM:

 Impacts soil biological properties: SOM is a reservoir of metabolic energy and source of nutrients (N, P, S, etc.) for living organisms.

 Impacts soil physical properties: SOM participates to the stabilization of soil structure through interactions with minerals, which, in turn, have a great impact on root development, erosion resistance, and water retention. SOM also influences soil thermal properties because of its dark color.

 Impacts soil chemical properties: because of the numerous charges and reactive functional groups on its surface, SOM increases the soil cationic exchange capacity, and have a pH buffering effect. Furthermore, its capacity to form complexes with metals, trace elements and pollutants impacts soil weathering, reduce losses of important micronutrients, and reduce metal and pollutants toxicity.

Furthermore, the carbon contained within the soil represents a significant reservoir of C in the global C biogeochemical cycle. The SOC stock is at least three times higher than the atmosphere and terrestrial vegetation C stocks combined (Schmidt et al., 2011). SOC has been estimated to account for 1200-1550 Pg of C (1Pg = 1015 g) to a depth of 1m, and for 2300-2450 Pg of C to a depth of 2-3 m (Lal, 2004, IPCC-WG1, 2013).

Part 1 Introduction and objectives

Figure 0.3 - Simplified schematic of the global carbon cycle (from the IPCC fifth assessment report, WG1, 2013). Numbers represent carbon stocks in PgC (1 PgC = 1015 g C) and arrows annual carbon exchange fluxes (in PgC yr–1). Black: estimated stocks and fluxes prior to the Industrial Era (~ 1750). Red: changes in stocks or fluxes since 1750.

Carbon storage is an important ecosystem function of soils that has gained increasing attention in recent years due to its interactions with the earth’s climate system. The increase in atmospheric CO2 concentrations since the industrial revolution is contributing to recent climate change, which is among the major challenges facing the world (Lorenz and Lal, 2014). That context gave rise to a huge research effort on SOC, and its interactions with climate change (Davidson and Janssens, 2006, Heimann and Reichstein, 2008, von Lützow and Kögel-

Part 1 Introduction and objectives

Knabner, 2009, Reichstein et al., 2005, Smith et al., 2008, Trumbore and Czimczik, 2008, Amundson, 2001). Sequestration of C within soils has been proposed as an effective natural strategy to offset increase of atmospheric CO2-C concentrations (Lal, 2004, Lorenz et al., 2007). Among others, the 4 per 1000 initiative proposes to implement farming practices that maintain or enhance soil carbon stock on as many agricultural soils as possible, and to preserve carbon-rich soils. This initiative, launched on 1st. December 2015 during the climate negotiations (COP21), is based on the theory that a “4‰” annual growth rate of the soil carbon stock would make it possible to stop the present increase in atmospheric CO2.

Considering its major role in soil fertility and within the global C cycle, understanding the mechanisms that control SOM formation and evolution within the soils is fundamental.

II. Objectives

‘Let us not drop at once into the soil and lose ourselves in the darkness of its details, but first let us look about and see how our field is related to the world at large and to the powers that energize in it. Let us begin with sunshine and the work it does.’ (King, 1907)

The C flow to the soil begins with the sun (Janzen, 2015). Once fixed in the vegetation, C atoms reach the soil where they can spend hours to centuries or longer (Torn et al., 2009). The accumulation of C in soils results primarily from the balance between inputs (net primary production) and losses (microbial mineralization, leaching). Stabilization is defined as protection of OM from mineralization, leading to prolonged turnover times in soil (von Lützow et al., 2006). It is the integrated effect of different “protection mechanisms”, operating simultaneously.

Part 1 Introduction and objectives

Three “protection mechanisms” are generally distinguished in soils: (1) the recalcitrance of SOM, due to its molecular properties; (2) the inhibition of decomposers activity, due to inaccessibility of the substrate or inadequate environment; (3) the interaction with mineral surfaces and metal ions. The interactions with the mineral soil matrix, and the quantity and strength of organo-mineral bonds, are considered to be the major control on long-term SOM stabilization (Kögel-Knabner et al., 2008b). Consequently, three compartments must be taken into account in soils to understand SOM dynamics and stabilization processes: the organic compartment (dead and living OM), the mineral compartment and the “physico-chemical” compartment (pore space).

Soils are not static but evolve constantly with time. As the three constitutive compartments of the soil strongly interacts, they will co- evolve with pedogenesis and soil formation, thereby impacting the SOM protection mechanisms. Understanding how these three compartments evolve with time, and the implications for C stabilization mechanisms require an integrated and multidisciplinary approach.

 First objective:

 Based on a literature review, develop a multidisciplinary soil conceptual model, integrating soil properties involved in carbon (C) stabilization.

Podzols represent a promising soil type for studying the co- evolution of SOM, minerals and physico-chemical environment, and implications for SOM protection mechanisms. First, from a quantitative point of view, they are known to store important quantities of OM. Podzols cover approximately 485 million hectares world-wide (World Reference Base for Soil Resources (WRB) soil classification system, FAO 2014), mainly in the temperate and boreal regions of the Northern Hemisphere, but also in humid temperate climates and in the humid tropics. Podzols are characterized by the formation of a sandy, bleached horizon (E horizon) overlying a dark 7

Part 1 Introduction and objectives horizon with accumulating organic matter as well as Fe- and Al- compounds (spodic or Bh horizon). Podzols (“spodosols” in the U.S.D.A. Soil Taxonomy) is the fifth’s soil order storing most OC per Ha, after Histosols, Andosols (Andisols) and Cambisols (Inceptisols) (Table 0.1). The global C stock in podzols represents approximatedy 5% of the total SOC (Table 0.1). Such large amounts of carbon may play a central role in the global carbon balance.

Table 0.1 – Mass of organic carbon in the world’s soils. Values for the upper 1 m represent most of the carbon in the soil profile (Brady and Weil, 2007).

Organic carbona in upper 100 cm Soil order Global area 10³ km² Mg/ha Global Pgb % of global Histosolsc 1,745 2,045 357 23 Andisols 2,552 306 78 5 Inceptisolsc 21,580 163 352 22 Spodosols 4,878 146 71 5 Mollisols 5,480 131 73 5 Oxisols 11,772 101 119 8 Entisols 14,921 99 148 9 Ultisols 11,330 93 105 7 Alfisols 18,283 69 127 8 Vertisols 3,287 58 19 1 Aridisols 31,743 35 110 7 Misc. land 7,644 24 19 1 Total 135,215 1576 100

a Organic matter may be roughly estimated as 1.7 to 2.0 times this value. The value traditionally used is 1.72. b Petagram = 1015 g. c Carbon stored in Gelisols is included with these soils.

Secondly, from a qualitiative point of view, podzols represent a system where SOM plays a key role in pedogenesis. Complexing organic acids, very aggressive towards minerals, are the main weathering agents in the surface horizon. In addition, organic matters are mobile within the podzolic profiles, leading to a physical separation of two main pedogenetic processes: (1) intense weathering in the surface horizon; (2) secondary mineral phases formation and organo- mineral associations accumulation in illuvial horizons.

Part 1 Introduction and objectives

 Objective two:

 Apply the conceptual model to a podzolic soil chronosequence, to study biogeochemical processes controlling the evolutions of SOM protection mechanisms during podzol genesis.

A soil chronosequence consists of a series of soils that have the same parent material, topography and climatic history (Jenny, 1941 ). The only factor that varies among the sites is their age (i.e. the duration of the pedogenesis) and biotic factors (vegetation and organisms evolve with time).

 Research questions:

 Long term (centuries): In a chronosequence of soil profiles, ranging from a beach sand parent material to a mature podzol of 530 years, what are the horizon- and profile-specific C protection mechanisms, and what are the main pedogenetic processes controlling their evolution ?

 Short term (days): Among other soil environmental factors, oxygen limitations are a largely unrecognized control on C dynamics and stabilization (Keiluweit et al., 2016). In soil samples of increasing age (from the chronosequence), what is the impact of temporary anoxic conditions on organo-mineral associations, considered to be the major control on long-term SOM stabilization?

Part 1 Introduction and objectives

III. Thesis outline

A schematic representation of the thesis organization is presented in the figure 0.4.

The first part of this work is the present introduction, giving (1) an overview of the global context and importance of SOM; (2) the research objectives.

The Part 2 is devoted to the conception of a soil conceptual model, integrating soil properties involved in carbon (C) stabilization. It gives a general overview of the current knowledge on SOM origin, composition, decomposition (Chapter 1) and protection mechanisms (Chapter 2). Part 2 also contains a description of podzols - the soil type used to apply the model - and its formation processes (Chapter 3).

The Part 3 is devoted to the characterization of SOM protection mechanisms in a podzolic chronosequence (from 0 to 530 years). First, we describe the selected study site, as well as the sampling and analytical methods applied (Chapter 4). The evolution of the pore system, determining the soil environment, will be characterized in chapter 5. The chapter 6 focuses on the mineralogical modifications occurring in the chronosequence. The objective is to characterize the significant mineral phases for the organo-mineral associations (OMA) formation (secondary mineral phases), and the processes behind their formation and evolution. In the chapter 7, the evolution of SOM, microbial populations and organo-mineral associations will be described, as well as the C dynamics (C decomposition in the different samples). The chapter 8 presents a synthesis of pedogenetic processes and C protection mechanisms in the chronosequence.

In the Part 4, we will study the short term evolutions of C protection mechanisms (days). The objective of chapter 9, is to determine the impact of anoxic (reducing) conditions on OMA in three chronosequence soil horizons of increasing age.

In the Part 5, we discuss the implications of this thesis, and give general conclusions and perspectives.

Part 1 Introduction and objectives

Part 1: Introduction and objectives Part 5: General conclusion and perspectives

Part 3: SOM protection mechanisms in the chronosequence

Chapter 4 – Study site description and analytical methods

Chapter 8 – Synthesis

Chapter 7 Chapter 5 Chapter 6 Evolution of SOM, Microbial Evolution of Evolution populations and organo-mineral the soil of the associations environment minerals

Part 2: General Overview

Chapter 1 SOM

Chapter 2 Protection Mechanisms

Chapter 3 Podzols

Part 4: OMA dynamics in anoxic conditions

Chapter 9 – impact of microbial Fe(III) reduction on OMA

Figure 0.4 - Schematic representation of the thesis outline.

Part 1 Introduction and objectives

Part 2: General overview Chapter 1 – Soil organic matter

Part 2: General overview

Chapter 1 - Soil organic matter

1.1 Definition

Soil organic matter (SOM) can be considered as the sum of all biogenic organic components in soils. It sums up natural and thermally altered biologically derived organic materials in soil or on the soil surface irrespective of its source, whether it is living or dead, or stage of decomposition, but excluding aboveground portion of living plants (Baldock and Broos, 2012).

It follows from this definition that SOM is composed of a high variety of compounds, with respect to their sources, chemical compositions, sizes (from simple monomeric molecules to complex polymeric compounds) and stage of decomposition. This is one definition among many others, as there is still no consensus on SOM definition within the soils community (Kleber and Johnson, 2010, Lehmann and Kleber, 2015).

1.2 Origin and composition

Organic matter in soils is composed of approximatively 85% of non-living matters (dead cells from plants, animals and microorganisms), 10% of roots, 5% of microflora and fauna (Paul and Clark, 1989).

Part 2: General overview Chapter 1 – Soil organic matter

a. Soil constituents [%v] b. SOM composition [%dw] Living Minerals microorganisms 5% 45% Dead matters 85% Roots OM 5% 10% Water 25% Air 25%

Hemicellulose 20% Lignins 20% Cellulose c. Plant residues composition 30% Proteins [%dw] (main OM input) 6% Extractable material 24%

Figure 1.1 – a. Composition of the soil (general volumic proportions, highly variable between different soils); b. Soil organic matter (SOM) composition (dry weight proportions); c. Plant residue composition (dry weight proportions). Extractable material include water- extractable (simple sugars, amino-acids, organic acids) and organic solvent-extractable material (fats, oils, waxes) (from Oades, 1989).

Part 2: General overview Chapter 1 – Soil organic matter

1.2.1 Plant residues

Plant residues are the dominant input and thus the primary source of organic carbon (OC) in the soil (Kuzyakov and Domanski, 2000, Kögel-Knabner, 2002). The plant residues consist of aboveground input (leaves, branches, fruits), but also below-ground inputs (roots and rhizodepositions). Their amount and composition vary greatly depending on the ecosystem and vegetation type (forest vs. grassland, humid tropics vs. arid regions, etc.). Root contribution to soil C pools is hypothesized to increase with soil depth (Rasse et al., 2005). Belowground plant production is often as high as, or higher than, aboveground production in temperate grasslands and forests (Ekschmitt et al., 2008). Some research suggests root inputs to soil represent 5–33% of daily photoassimilate (Jones et al., 2009). Plant material can be divided in three components, with diverse function and composition (Figure 1.1).

The protoplasm constitutes the inside of a living cell. The most relevant group of protoplasmic substances for the SOM are proteins and storage materials. Proteins represent approximately 6% of plant residue (Baldock and Broos, 2012), and is the most active group of substances in plant cells (Kögel-Knabner, 2002). They are composed of chains of amino-acids (polypeptides), and serve various functions, e.g. transport proteins, structure proteins, or enzymes. Storage materials are polymers of monosaccharides, like starch - polysaccharide composed of amylose and amylopectin, two polymers of glucose - and fructans - a polymer of fructose.

The cell membrane is composed of globular proteins embedded in a lipid bilayer (Baldock and Broos, 2012).

The cell walls constitute the majority of the mass of plant residues deposited in soils. They are composed of structural components. Among them, cellulose is the most important, representing approximately 30 % of plant residue (Baldock and Broos, 2012). Cellulose is a linear polymer composed of glucose units (>10,000) linked by glycosidic bounds. Two other important structure

Part 2: General overview Chapter 1 – Soil organic matter polysaccharides, differing from cellulose in their composition in sugar units and lower degree of polymerization, are hemicellulose (20% of plant residue,Baldock and Broos (2012)) and pectin (only represents less than 1% of plant residue, Kögel-Knabner (2002)). With the polysaccharides, lignin fills out the cell walls. It is the second most important biopolymer in nature and represents approximately 20% of plant residues (Thevenot et al., 2010). The lignin structure consists of aromatic rings with side chains and -OH and -OCH3 groups (phenyl- propanoïd units) linked by various strong covalent bonds (Thevenot et al., 2010). Tannins is another group of polyphenols, quantitatively important component in cell walls in higher plants (Wershaw, 2004).

Cuticular and root waxes are surface lipids covering the epidermis; they protect the surface of shoot and root from biotic and abiotic stresses, limiting gas and water exchanges (Kolattukudy, 2001, Nawrath, 2006). Suberin and cutin are ubiquitous extracellular lipid polymers (polyester) found in plants (Li et al., 2007). Cutin composes the cuticle of the aerial parts of plants, while suberin is present in the bark of vascular plants and in the epidermis, endodermis, exodermis, and phellem of roots (Wershaw, 2004). Lipids consist of a heterogeneous group of substances that can have various other functions in cells, like energy reserve (Dinel et al., 1990).

1.2.2 Soil microbial biomass

The soil microbial biomass has been defined as the part of SOM which constitutes living microorganisms smaller than 5–10 mm3, and contributes between 1 and 5% of SOM (Glaser et al., 2004). Soil communities are extremely diverse and complex, with billions of individual organisms and millions of species being found within a single ecosystem (Bardgett and van der Putten, 2014). Densities of up to 1010 bacteria and several kilometers of fungal hyphae per g soil have been measured in a variety of soils (Lavelle et al., 1993). The majority of soil processes involving organic matter (OM) processes is accomplished by soil microbes. At least 80%, and usually over 95%, of CO2 derives from microbial respiration (Lavelle et al., 1993). In addition to be the main agents of transformation of non-living OM, microbial products (dead

Part 2: General overview Chapter 1 – Soil organic matter cells and exudates) are also a major source of SOM (Knicker, 2011). As decomposition proceeds, a transformation of vegetal OM to microbial substances occurs (more details are given in section 2.1.3), and 80% of the SOC may be derived from microbial biomass (Liang and Balser, 2011).

Some bacterial and fungal constituents are similar to the ones present in plant materials: proteins, lipid bilayer membrane and similar heterogeneous non-cellulosic polysaccharides. Lignin does not occur in microorganisms (Higuchi, 1990), and cellulose rarely (De Leeuw and Largeau, 1993). Specific components also occur, especially the polymers of amino-sugars (AS) constituting their cell walls. Fungal cell wall is composed of chitin (acetyl-glucosamine polymer), while bacterial cell walls is composed of peptidoglycan (acetyl-glucosamine and acetyl-muramic acid polymer, (Bodé et al., 2009)). Though the polysaccharides of micobial cell walls degrade relatively easily, their basic units such as glucosamine, galactosamine and muramic acid (three AS) accumulate in soils during litter decomposition (Kögel- Knabner, 2002).

Methodological box 1.

We decided to use two biomarkers to determine the composition of microbial communities in our soil samples (Chapter 7):

 Phospholipid fatty acids (PLFAs), deriving primarily from microbial cell membranes, are widely accepted as biomarkers that indicate viable microbial biomass and provide a microbial community ‘fingerprint’ (Liang et al., 2008b). They are not found in storage products or in dead cells (Zelles, 1999).

 Amino sugars (AS) quantification is a useful tool to study the evolution of the dead microbial cells pool. The large majority of amino sugars found in soil are of microbial origin (typically >99.5 %, (Bodé et al., 2013)), and they persist after lysis of microbial cells (Amelung et al., 2001, Liang et al., 2008b, Kandeler et al., 2000).

Part 2: General overview Chapter 1 – Soil organic matter

1.2.3 Plant and microbe exudates

Rhizodeposits are root exudates released by living roots into the rhizosphere during plant growth (Kuzyakov and Domanski, 2000). Root exudates contain ions (i.e. H+), water, oxygen and inorganic acids, but they mainly consist of carbon-based compounds (Bais et al., 2006). These organic compounds can be separated into (i) low-molecular weight compounds, including organic acids, amino acids, sugars, phenolics and secondary metabolites; (ii) high-molecular weight compounds like mucilage and proteins (as recently reviewed by Badri and Vivanco (2009) and Jones et al. (2009)). These secretions represent a significant carbon cost to the plant, typically 7–8% of the total photosynthetic carbon (Rasse et al., 2005), and up to 30–40% for young seedlings (Badri and Vivanco, 2009). Rhizodeposits have multiple impacts in the rhizosphere: they especially improve soil structure and the nutrition of soil organisms (Broeckling et al., 2008, Pollierer et al., 2007). Consequently, the rhizosphere is a “hot spot” for microbial activity (Kuzyakov and Blagodatskaya, 2015).

Extracellular polymeric substances (EPS) are produced by microbes to increase the adhesion of cells to solid surfaces, to facilitate bridging between cells in biofilms, to act as a protective barrier against desiccation, toxic metals or other antimicrobial agents, and to allow the accumulation of nutrients from soil solution. EPS are composed of polysaccharides and proteins, with minor contributions of nucleic acids and lipids (for a review on composition and functions of EPS, refer to Flemming and Wingender (2010)). The exact composition depends on the specific bacterial or fungal strain, growth stage, and the physico-chemical conditions of the surrounding solution (Flemming and Wingender, 2010).

Part 2: General overview Chapter 1 – Soil organic matter

1.3 Decomposition and transformation

Decomposition is described by Lavelle et al. (1993) as a "cascade" process during which a given resource is progressively transformed into a set of secondary resources which, in turn, are transformed into resources of tertiary or higher order. Approximately 85%–90% of organic-material decomposition in soils are mediated microbially, while abiotic chemical oxidation, in contrast, is likely to account for <5% of OM decomposition (Lavelle et al., 1993).

At the onset of decomposition, the fragments and exudates of plants and soil fauna are first broken up into small pieces. Only molecules typically < 600 Da can be actively transported across cell walls into microorganisms (Hedges et al., 2000). This first substrate alteration is made outside of the cells by extracellular enzymes, in the soil solution. Consequently, the substrates have to pass through an aqueous phase before uptake and utilization by soil microbial decomposers community. Soil fauna consumes a smaller proportion of SOC compared to microorganisms, but faunal activities enhance the ability of microbial decomposers to utilize organic residues (Baldock and Broos, 2012, Wolters, 2000). This enhancement results from the fragmentation of plant debris and the distribution of organic material in the soil matrix (Dungait et al., 2012).

Once inside the cells, simple substrate (soluble molecules able to cross cell membranes) can be mineralized, assimilated or transformed (below definitions from Baldock and Broos (2012)). Mineralization implies the conversion of organic substrates into + soluble or gaseous inorganic forms (OC to CO2 or ON to NH3 or NH4 ), in response to respiration or other metabolic processes. The carbon substrate can either be assimilated to produce biochemical compounds required for for the maintenance and growth of the decomposer community, or transformed into metabolic products which are further excreted back into soil solution. The balance between the assimilated vs mineralized substrate is called “substrate use efficiency” (Cotrufo et al., 2013, Manzoni et al., 2012). Organic molecules produced by soil microorganisms are, after plant input, the

Part 2: General overview Chapter 1 – Soil organic matter secondary source of decomposable OC in soils. Many, if not most, organic molecules in soil from the non-living pool are of microbial origin (Simpson et al., 2007, Miltner et al., 2012, Schmidt et al., 2011).

The molecules that escape mineralization are allocated to different microbial byproducts. The latter are either subject to further degradation or incorporated intimately within the mineral matrix. Given the difficulty to isolate and describe this amorphous decomposing organic materials, it was proposed to call it ‘‘the molecularly-uncharacterized component of nonliving organic matter in natural environments’’ or MUC (Hedges et al., 2000). Currently, two opposite theories coexist on the fate of the decomposition products: “the synthetic concept” vs. “the degradative concept” (Kleber and Johnson, 2010).

The first concept, the classical and oldest view, is called humification or secondary synthesis. This theory is based on an extraction procedure, dating from a report published in 1786. It became widely adopted to separate soil organic compounds from mineral constituents (Kleber and Johnson, 2010). The procedure involves the extraction with an alkaline solution (NaOH-pH13). The insolubilized fraction is called “humin” and is considered as more resistant. The solubilized fraction that precipitates after adding an acid is called “humic acid” and the fraction that remains soluble after reacidification is called “fulvic acid”. The three fractions extracted (humin, fulvic and humic acids) were universally accepted as experimental proxies for soil organic matter, and treated as physically existing entities (Lehmann and Kleber, 2015). The humification theory arose from research efforts to understand the behavior and properties of the above-mentioned fractions (von Lützow et al., 2006). This theory assumes a further transformation of the initial decomposition products into new condensation products with unique chemical structures, different from those of the starting material, through biotic and abiotic secondary synthesis reactions (Stevenson, 1994, Kononova, 1961, Hayes and Swift, 1990, Schulten and Schnitzer, 1997). The resulting macromolecules (humic substances) are thought to be large, complex and high-molecular-weight structures, with a high content of aromatic rings, and joined by strong covalent bonds

Part 2: General overview Chapter 1 – Soil organic matter

(Stevenson, 1994). These substances are considered as inherently resistant to decomposition, and consequently older than the rest of the soil organic matter (Dungait et al., 2012).

This paradigm has been increasingly questioned. Excellent reviews have linked experimental facts to new insights about OM degradation and association with minerals. They refute the secondary synthesis of humic substances (Burdon, 2001, Sutton and Sposito, 2005, Lehmann and Kleber, 2015, Kleber and Johnson, 2010, Kelleher and Simpson, 2006, Baldock and Broos, 2012). Fundamental facts are: (i) organic matter is a ‘‘thermodynamic anomaly” which tents to be converted in lower energy components such as carbon dioxide, water, nitrate, and phosphate as predicted by entropy; a biosynthesis of recalcitrant molecules would be a waste of energy for microorganisms, and the existing abiotic reactions proposed in the literature (e.g. Maillard reaction) are also rejected; (ii) the harsh alkaline treatment induces artefacts; (iii) no study proves that humic substances occur separately in soil environments. Related to the latter argument, Near- edge X-ray fine structure spectroscopy (NEXAFS) coupled to scanning transmission X-ray microscopy (STXM) (Lehmann et al., 2008) or nuclear magnetic resonance (NMR) (Kelleher and Simpson, 2006) has shown that SOM is composed of a complex mixture of highly variable, but identifiable, biopolymers of plant and microbial origin and their decomposition products, but not of distinct chemical categories.

Thus a recent model was developed, supported by an increasing number of scientists (Burdon, 2001, Piccolo, 2001, Sutton and Sposito, 2005, Wershaw, 2004, Sposito, 2008, Cotrufo et al., 2013): “the degradative concept”. This model suggests that biological macromolecules are progressively degraded into fragments of a structure and composition that can be related to their parent molecules (Kleber and Johnson, 2010). The progressive decomposition by microorganisms induces an enzymatic depolymerization and oxidation of the molecules (addition of oxygen or other electronegative element) (Kleber et al., 2015). By doing so, the microorganisms make the molecules moving down a thermodynamic gradient from large energy-rich compounds to smaller energy-poor compounds (Hedges et al., 2000). Microbes thus use the energy

Part 2: General overview Chapter 1 – Soil organic matter released to support their own metabolic processes. The increase in oxidative state (i.e. the increase in carboxyl or hydroxyl groups) with increasing decomposition is a very well-documented fact about SOM (Lützow and Kögel-Knabner, 2010). This decrease in size and increase in polar and ionizable functional groups induces an increasing solubility of organic materials as their decomposition progresses (Kleber et al., 2015). This property facilitates their mobility in soils by diffusion and bulk flow processes, and their uptake and mineralization by microorganisms (Kleber and Johnson, 2010). Furthermore, the addition of polar functional groups transforms originally nonpolar components (like structural plant and animal residues) into amphiphilic molecules (Wershaw, 1994). Consequently, the degradative concept has resulted in the development of models in which SOM is depicted as being composed of molecular aggregates (“supramolecular aggregate model”, Piccolo (2001)) of amphiphilic fragments self-assembling spontaneously within the aqueous soil solution into micelle-like structures, held together by non-covalent bonds (Wershaw, 1994, Wershaw, 2004).

This modern view for describing the fate of organic debris has been called by Lehmann and Kleber (2015) “the soil continuum model”, considering soil organic matter as a continuum of organic fragments of all sizes and at various stages of decomposition - from intact plant material to highly oxidized carbon in carboxylic acids (Figure 1.2). This has implications on the C stabilization mechanisms theories. With decomposition, the opportunity for protection against further mineralization increases through greater reactivity towards mineral surfaces and incorporation into aggregates (Lehmann and Kleber, 2015), as developed in the next section.

Part 2: General overview Chapter 1 – Soil organic matter

Figure 1.2 - The soil continuum model, from Lehmann and Kleber (2015).

Part 2: General overview Chapter 1 – Soil organic matter

Part 2: General overview Chapter 2 – SOM protection mechanisms

Chapter 2 - SOM protection mechanisms

The accumulation of C in soils results from the balance between inputs (net primary production) and losses (microbial mineralization, leaching). Stabilization is defined as protection of OM from mineralization, leading to prolonged turnover time in soil (von Lützow et al., 2006). It is the integrated effect of different “protection mechanisms”, operating simultaneously.

The main authors working on C stabilization (Sollins et al., 1996, Sollins et al., 2007, Baldock and Skjemstad, 2000, Baldock et al., 2004, von Lützow et al., 2006, von Lützow et al., 2008, Kleber and Johnson, 2010, Six et al., 2002) generally recognize three main protective mechanisms: (1) the recalcitrance of SOM, due to its molecular properties; (2) the inhibition of decomposers activity, due to inaccessibility of the substrate or inadequate environment; (3) the interaction with mineral surfaces and metal ions. These authors nuance the description of each mechanism, and their relative importance.

Integrating the SOM stabilization mechanisms in the SOM decomposition model (Lehmann and Kleber, 2015) was developed by Cotrufo et al. (2013), and called the Microbial Efficiency-Matrix Stabilization (MEMS) framework. This framework sees the decomposition, transformation and stabilization of C as a continuum, with two key processes: (1) a ‘microbial filtering’ regulating the flow of C from plant material to SOM. This process is controlled by the two first protective mechanisms (chemical recalcitrance and environmental inhibition); (2) the interaction between microbial products of decomposition and mineral matrix, the quantity and strength of organo-mineral bonds being the major control on long- term SOM stabilization (Kögel-Knabner et al., 2008b). We will now describe each protective mechanism in this context.

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.1 Chemical structure and recalcitrance

The intrinsic recalcitrance depends on properties of organic molecules at molecular level (including elemental composition, presence of functional groups, and molecular conformation) that restrict microbial decomposition (Sollins et al., 1996). Generally, molecular characteristics that render natural OM less susceptible to degradation are molecular size, complexity and degree of polymerization (number of bounds within the molecules), strength of the intra and intermolecular bonds, content of aromatic and aliphatic functional groups, certain N-containing substituents and functional groups, hydrophobicity and increase in electron richness (decreasing oxidation state) of the substrate (Kleber, 2010).

2.1.1 Chemical recalcitrance of organic constituents

Plant constituents consist of the primary input of OM into the soils. They have diverse chemical recalcitrance and are processed differently by soil organisms (Figure 2.1). The different rates of degradation for each group of plant components were reviewed by Dungait et al. (2012). Metabolic compounds (such as simple sugars, organic acids including amino acids, proteins) exhibit low molecular weight, and high lability. Their half-life can be less than 1 h in surface soil horizons. Structural compounds decompose more slowly. However, many aerobic and anaerobic bacteria and fungi use polysaccharides (cellulose, hemicellulose and pectins) as substrate through the hydrolysis of the glycosidic links using exo-enzymes. The mineralization rates are in the order of magnitude of a few month. Of the structural compounds, lignin and other polyphenols are more resistant to decomposition due to their complex structure and the abundance of aromatic structures. Only white rot fungi, through lignin peroxidase, are able to completely decompose lignin into CO2 but they need a degradable substrate as energy source, enough O2 but not too much available N, and adequate pH-moisture conditions (for a review of lignin degradation by white rot fungi - ten Have and Teunissen (2001)). Other fungi (soft rot and brown rot fungi) and few bacteria

Part 2: General overview Chapter 2 – SOM protection mechanisms can induce structural changes and reduce the molecular size of lignin, but are not able to achieve complete mineralization. The literature is contradictory with regard to lignin decomposition times, but most lignins decompose within 5 year (Thevenot et al., 2010). Lipids and waxes are also considered to be resistant to degradation, due to their apolar chemistry, and the occurrence of aliphatic bonds (C–C bonds). The latter are indeed less easily decomposed than hydrolytic bonds (Derenne and Largeau, 2001).

Figure 2.1 – Conceptual scheme of the C flow during degradation of plant debris (from Miltner et al 2012)

Charcoal, or black carbon, is a particular case of SOM, because it has undergone a pyrolytic degradation (Wershaw, 2004). Black carbon is considered highly resistant to degradation, and is able to persist in soil environment during thousands of years (Major et al., 2010, Liang et al., 2008a). All Pyrogenic C compounds are characterized by fused aromatic rings (highly condensed structure), but varying in cluster size, and the occurrence of other elements (N, O)

Part 2: General overview Chapter 2 – SOM protection mechanisms and functional groups (Preston and Schmidt, 2006). They do not represent a well-defined chemical compound or structure (Krull et al., 2009, Derenne and Largeau, 2001). Thus, the degradability of plant components follows the order:

Sugars > proteins > hemicellulose > cellulose > lignin/fats > charcoal

2.1.2 A concept not sufficient alone

The selective preservation of recalcitrant organic compounds, leading to a progressive change of OM composition, was assumed to be one of the main mechanism of SOM stabilization (Lützow and Kögel-Knabner, 2010, Marschner et al., 2008) and was the basis of many theories to explain the turnover of SOM (Kleber, 2010). However, this view is increasingly questioned.

The formation of recalcitrant humic substances has long been considered as the major pathway for SOM formation and stabilization (Stevenson, 1994). However, as discussed in the previous section, increasing evidences show that these humic substances does not exist in soils, and a formation of complex polymeric structure enriched in aromatic rings cannot explain the stabilization of C.

The selective preservation of more resistant plant components, like lignin or lipids, is often regarded as precursor of stable OM formation (von Lützow et al., 2006). However, many authors have shown that, under adequate conditions and in presence of appropriate decomposer organisms and enzymatic activity, even presumably persistent materials will decompose (Lehmann and Kleber, 2015), including lignin (Thevenot et al., 2010), very long chain n-alkanes (Quenea et al., 2004, Wiesenberg et al., 2004), polycondensed aromatics (Gramss et al., 1999), and even fire-derived carbon (Hamer et al., 2004, Wengel et al., 2006). These substances may turn over more rapidly in soil than total SOM (Marschner et al., 2008, Amelung et al., 2008).

Part 2: General overview Chapter 2 – SOM protection mechanisms

In parallel with recalcitrant material turning over quickly, there is increasing evidence that ‘‘old’’ SOM may well contain substantial proportions of thermodynamically and chemically labile carbon” (Sollins et al., 2006, Nunan et al., 2015, Kleber et al., 2011). Microbial polysaccharides, proteins and other easily biodegradable MO can persist for long periods in soils (Amelung et al., 2002, Kiem and Kögel- Knabner, 2003, Kleber, 2010), and may be even older than bulk SOM (Derrien et al., 2006, Gleixner et al., 2002, Kleber and Johnson, 2010).

Finally, several models of SOM genesis overemphasize the contribution of plant residues and their degradation products as the molecular sources for SOM (Miltner et al., 2012). There are increasing evidences that (1) recalcitrant plant components do not preferentially accumulate in SOM (Marschner et al., 2008, Cotrufo et al., 2013), (2) microbial products are the largest contributor to stable SOM (Knicker, 2011). Besides, 80% of the soil organic C may be derived from microbial biomass (Liang and Balser, 2011).

2.1.3 Persistence as an ecosystem property

Consequently, persistence of OM, is now increasingly recognized as an ecosystem property, instead of a molecular property (Schmidt et al., 2011). The chemical composition influences its decomposition rate by determining the complexity of the decomposition operation (Kleber, 2010). Bosatta and Ågren (1999) stated in their “continuous quality model of OM” that OM quality is determined by the number of enzymatic steps required to release a carbon atom from an organic compound into carbon dioxide. With an increasing complexity and number of steps needed, the quality of the carbon decreases, and the activation energy needed and temperature sensitivity increases (Barré et al., 2016). The molecular recalcitrance of natural OM is not absolute, but relative (von Lützow et al., 2006); it depends on the presence of microorganisms (to have the full suite of required enzymes), and the adequate environment for microbial activity (see section 2.2).

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.1.4 Recalcitrance role in the degradation continuum

A selective preservation of some plant compounds due to their chemical nature, while passing through the “microbial filter” (Figure 2.2), will lead to a progressive change of residue composition during the initial stages of decomposition (Marschner et al., 2008). Lignin accumulates, carbohydrates and proteins are be preferentially degraded. The recalcitrance does play a crucial role in the short to medium term C stabilization (within years or decades) (Kögel-Knabner et al., 2008b).

As decomposition proceeds, a transformation occurs to microbial substances that are not easily recognized (Hedges et al., 2000, Hatcher, 2004). According to Grandy and Neff (2008), virtually all soil C is, at some point, used as a substrate by microorganisms. The filtering effect of microorganism reduces the differences in the molecular structure, coming from the differences in plant chemistry. Consequently, the decomposed fraction in different soils generally presents similarities in structure, and bulk soil OM does also not reflect the different plant input (Kögel-Knabner et al., 2008b). The microbial biomass and byproducts will progressively be incorporated into the silt and clay fraction during further decomposition phases, where microbially-derived N-containing compounds, polysaccharides, waxes and lipids, and stabilized aromatic compounds will accumulate (Figure 2.3). In this fraction, interactions with soil minerals do occur and become the dominant stabilization mechanism, controlling SOM persistence over the long term (decadal to millennial time scales) (Kögel-Knabner et al., 2008b). Old SOM with long turnover times is generally either associated with soil minerals, or present as fossil or black C (Marschner et al., 2008). The formation of organo-mineral associations will be detailed in section 2.3.

Part 2: General overview Chapter 2 – SOM protection mechanisms

Figure 2.2 – see next page.

Part 2: General overview Chapter 2 – SOM protection mechanisms

Figure 2.2 - Representation of the effects of plant litter quality on CO2 efflux and soil organic matter stabilization in the Microbial Efficiency- Matrix Stabilization (MEMS) framework. “During decomposition, above- and below-ground plant litters undergo microbial processing which determines the quantity and chemical nature of decomposition products. Proportionally more dissolved organic matter and more carbohydrates and peptides are formed from high-quality (e.g., fine roots and herbaceous) litter than low-quality (e.g., needle and wood) litter, which loses most of the C as CO2. The ultimate fate of the decomposition products depends on their interactions with the soil matrix.” (From Cotrufo et al 2013)

Figure 2.3 – C structure associated with different soil size classes. The black regions indicate plant-derived compounds and the shaded regions indicate microbially-derived compounds (from Grandy and Neff, 2008).

Part 2: General overview Chapter 2 – SOM protection mechanisms

 First compartment of the model:

SOM compartment - a continuum of organic fragments, from coarse plant residues to small, highly oxidized molecules, with an increasing proportion of microbial-derived C.

 Recalcitrance in the model:

The selective preservation of recalcitrant compounds due to their chemical nature, while passing through the “microbial filter”, is a significant protection mechanism during the initial stages of decomposition. Recalcitrance does play a crucial role in the short to medium term C stabilization (within years or decades).

Recalcitrance

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.2 Inhibition of decomposers activity

As developed in the previous sections, the degradation of organic molecules is, for the most important part, mediated by microbial enzymes. Consequently, degradation can occur only when the appropriate microbial community is exposed to the substrate during enough time and in adequate environments. The second « protection mechanism » refer to all environmental and biological factors that prevent the interaction of extracellular enzymes with an accessible substrate (i.e. that is not adsorbed on minerals).

2.2.1 Presence of the adequate microorganisms

Microorganisms are the main decomposers, and mediate approximately 85%–90% of OM decomposition in soils (Lavelle et al., 1993). The degradation of a specific substrate strongly depends on the presence or not of microorganisms with the appropriate DNA to form and release the adapted enzymes (Baldock and Broos, 2012). Soil communities are extremely diverse and complex, with millions of species and billions of individual organisms being found within a single ecosystem (Bardgett and van der Putten, 2014). Consequently, soil microorganisms demonstrate a huge physiological and biochemical capacity due to their combined enzymatic repertoire (Dungait et al., 2012). Historically, soil microbial ecologists considered that ‘everything is everywhere, but the environment selects’, according to the view of Becking in 1934 (De Wit and Bouvier, 2006). Recent evidences readily challenge this paradigm, since they show that most soil organisms are restricted in their global distributions (Bardgett and van der Putten, 2014, Nannipieri et al., 2003). However, according to Dungait et al. (2012), a good enough working assumptions is that (i) the soil microbial community is collectively infallible in terms of the range of organic molecules it can degrade, and (ii) the capacity to degrade any substrate exists in almost any soil.

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.2.2 Habitat constraints

Soil present a complex architecture. The inorganic constituents are distributed in distinct particle size classes (sand, silt and clay) and they differ in mineralogical nature. They mutually organize, but also with the organic and biologic components, in micro-aggregates (20 - 250 µm), themselves associated in macro-aggregates (>250 µm) (Tisdall and Oades, 1982, Six et al., 2004). The spatial arrangement of the solid particles result in a pattern of pore spaces of various shapes and sizes (Chenu and Stotzky, 2002). This 3D organization of interconnected pores, ranging from the (nano-)pores of the primary particles to the macropores formed within and between the macroaggregates, provides an extremely large and heterogeneous interfacial domain (Totsche et al., 2010). The pore space is filled up with air and water. It is the habitat of soil microorganisms, and the medium for physico-chemical and biological reactions (Chenu and Stotzky, 2002). Soil architecture - pore system and aggregates - influences biological decomposition through impacting both the (1) environmental conditions and (2) accessibility of OM. The presence of ‘biologically non-preferred soil spaces’ is now recognized as a major C stabilization mechanisms in soils (Ekschmitt et al., 2008).

2.2.2.1 Environmental constraints

Available water, temperature, pressure, air composition, pH, oxidation–reduction all significantly influence the size, distribution, metabolic activity and composition of soil biotic communities, and therefore C cycling (Nannipieri et al., 2003, Ettema and Wardle, 2002, Beare et al., 1995). If the organisms are in some way suppressed, the low activity of the soil microbes becomes an effective OM protection mechanism (Torn et al., 2009). Water availability is probably the most important factor affecting microbial life in soils and decomposition processes (Chenu and Stotzky, 2002). Moisture is always needed because it is the medium for chemical reactions, exchanges and displacements. Water content is closely linked with oxygen supply. With increasing water content, water becomes more and more

Part 2: General overview Chapter 2 – SOM protection mechanisms available to decomposer organisms; however, oxygen availability may decreases. An optimum air-filled porosity exists at which the processes of decomposition and mineralization of organic C will be optimized for a given soil (Baldock and Skjemstad, 2000).

Table 2.1 – Differences in requirements between fungi and bacteria, and preferential location within a soil profile (combination of the reviews of Chenu and Stotzky, 2002; Baldock and Skjemstad, 2000; Von lutzow et al., 2006; Gadd, 2009; and Kogel Knabner et al., 2008b).

Bacteria Fungi Requirements Oxygen Oxic and anoxic metabolism Obligatory aerobes (most of them) Water Need water. More resistant to dry Optimal water-air ratio for conditions than bacteria. aerobe bacteria: Field Lowest water potential for capacity. activity: -25 000 kPa. Lowest water potential for Only fungal hyphae can activity: -1500 kPa extend through air-filled pores pH Prefer pH close to neutrality More resistant to acidic conditions Substrate Principally heterotrophs. All heterotrophs. The small autotroph fraction Mainly involved in litter exerts crucial roles in the degradation ecosystem (e.g. iron cycle). Fungi are the primary agents Bacteria depends on the of decomposition arrival of labile substrate Displacement Rely on diffusion of organic Ability to translocate and inorganic compounds in nutrients through the mycelial water for their nutrition network. Hyphae can grow in (limited displacements) air-filled pores. Preferential location Fraction Dominate the small-size More abundant in coarse fractions. Can access the (sandy) fractions. Mainly inside of microaggregats. present in the outer regions Bacteria-dominated of macroaggregates. communities in silt- and clay- sized fractions. Depth Located everywhere, but Mostly concentrated in the proportion increases in the upper 20 cm of soil deeper horizons

Part 2: General overview Chapter 2 – SOM protection mechanisms

Relative to the larger pores between aggregates (macropores), the smaller internal pores (micropores, within macro and microaggregates) are likely to remain filled with water and restrict oxygen availability. If oxygen supply (via diffusion) is below O2 consumption, the interior of structural units (micro and macro aggregates, peds) become O2 depleted (Keiluweit et al., 2016). Sexstone et al. (1985) showed that the interior of aggregates could be anaerobic, even under well-aerated conditions, resulting in an abundance of anaerobic microsites and associated metabolic gradients even within seemingly aerobic, well-drained soils. It has long been known that decreasing O2 concentrations generally decreases OM mineralization rates; residence times of OM tend to be larger in soil spaces with limited O2 supply (aggregates or peds) (Keiluweit et al., 2016, Hedges and Keil, 1995). Constraints induced by O2 limitations on microbial metabolism is a largely unrecognized and underestimated control on overall rates of C oxidation in upland soils (Keiluweit et al., 2016). Furthermore, oxygen limitations within the soil matrix prompt microbes to switch to alternative metabolism and terminal electron acceptors, less efficient in C oxidation. Even if nitrate and Mn reduction occur more rapidly, Fe(III) is increasingly recognized as the quantitatively most important alternative electron acceptor in a large range of upland soils, because of its abundance (Keiluweit et al., 2016).

Methodological Box 2.

Constraints induced by oxygen limitations on microbial metabolism (C oxidation, Fe(III) reduction, etc.), and resulting impacts on OMA, are largely unrecognized in soils. We attempted to evaluate the effect of a reducing event on OMA in podzol illuvial horizons in chapter 9, using a Fe-reducing bacteria (Shewanella putrefaciens).

Part 2: General overview Chapter 2 – SOM protection mechanisms

Other factors, such as pH, presence of metal ions (Paul and Clark, 1989, Sollins et al., 1996) and temperature (Conant et al., 2011, von Lützow and Kögel-Knabner, 2009) also influences both the cell and enzyme activities.

The micro-environmental conditions are determined by the local climate (precipitations, temperatures). At the macro environment scale, biotic suppression and subsequent C stabilization may occur in some ecosystems through adverse conditions, such as O2 limitation in flooded zones, desiccation in desert environments, freezing temperatures in boreal and arctic systems or low pH in mine spoils (Torn et al., 2009, Schmidt et al., 2011, Trumbore, 2009).

2.2.2.2 Accessibility of the substrate

Biodegradation requires a contact between the microbial cell or its enzymes and an organic substrate (Chenu and Stotzky, 2002). The soil volume occupied by micro-organisms is less than 1%, and distributed heterogeneously in very small microhabitats (Schmidt et al., 2011, Kuzyakov and Blagodatskaya, 2015). Microbial activity and colonization depend on the pore-size distribution and special configuration of pores, as described in Ekschmitt et al. (2005). The pore space of a soil is a critical property that determines the relative abundance of habitats and their connection, depending on the water content (capillary laws) (Chenu and Stotzky, 2002). Some microenvironments are not accessible, because the pore necks are too small, or, in the case of bacteria, yeasts and protozoa, the water pathways are discontinuous (only filamentous fungi can grow in air- filled pores) (Hassink et al., 1993, Chenu and Stotzky, 2002). In pores smaller than organism size, the decomposition of OM can only occur via diffusion of extracellular enzymes towards the substrate, followed by diffusion of the product of enzyme reaction back to the organism (Baldock and Skjemstad, 2000) (Figure 2.4). Consequently, physical variables associated with soil architecture (distance between microorganisms and substrate; soil pore size, length, connectivity and tortuosity) can limit the accessibility of SOM to decomposers and their enzymes, hence protecting SOM from degradation (Torn et al., 2009).

Part 2: General overview Chapter 2 – SOM protection mechanisms

Occlusion of OM within aggregates is one aspect of this “physical” protection mechanism, leading to a disconnection between decomposers and SOM (Six et al., 2002). Soil aggregation is a transient property since aggregates are constantly formed and perturbed due to chemical, microbial, plant, animal, and physical processes (Sollins et al., 1996, Huang et al., 2005). Examples of natural and anthropogenic soil disturbance include tilling, freeze/thaw and dry/wet cycles, erosion, bioturbation, wind throw, and fire (Six et al., 2004). Many studies have shown an increase of respiration and mineralization when aggregates are disrupted, probably because of the exposition of SOM, previously sequestered, to a physical environment more adequate for microbial and meso-faunal degradation (von Lützow et al., 2006, Chenu and Stotzky, 2002).

Figure 2.4 - Size spectrum of particles of different provenence in soil, biota, and pores (from Totsche et al 2010)

Deep soils are more likely to be colder, waterlogged, anoxic and substrate-limited (C-poor) compared with surface horizons, leading to smaller and less active microbial communities at depth

Part 2: General overview Chapter 2 – SOM protection mechanisms

(Dungait et al., 2012). Consequently, the reduced exposition of SOM to microbes mentioned here above will be more acute in the subsoil, and the degradation of C more limited therein (Ekschmitt et al., 2008, Chabbi et al., 2009, Salome et al., 2010, Rumpel and Kögel-Knabner, 2011).

2.2.3 Substrate constraint

The substrate degradation depends on the availability of nutrients (carbon and energy sources, mineral nutrients, growth factors, ionic composition) for the decomposer community (Nannipieri et al., 2003), and SOM decomposition may be affected by substrate- driven ‘biological rate limitation’ (Dungait et al., 2012). The production of enzyme capable to degrade SOM requires energy, some compounds being energetically more expensive than others (Kuzyakov et al., 2000). If the energy cost of enzyme production is not paid back by decomposition products, a negative feedback loop on microbial activity will be established. In this case, the energetical limitation reduces the capacity of microbes to decompose SOM (Ekschmitt et al., 2005). Consequently, decomposition of one soil carbon pool can be influenced by another OM or nutrient input (Manzoni and Porporato, 2009, Neff et al., 2002). For example, N and P availability may limit decomposition, more strongly in low-P than in low-N environments (Torn et al., 2005, Hobbie and Vitousek, 2000). The availability of easily decomposable carbon sources for microbes have been shown to increase the degradation of lignin (Klotzbücher et al., 2011), and black C (Hamer et al., 2004, Wengel et al., 2006). This temporary extra decomposition of organic C after addition of nutrients or easily- decomposable substances to the soil (like root exudates, DOM, etc.) is well-documented, and called “positive priming effect” (Kuzyakov et al., 2000). The most likely reason for the occurrence of priming effects is an activation/stimulation of the microorganism’s activities through easily available substrates, inducing enzyme production (co-metabolic decomposition of SOM) (Kuzyakov et al., 2000). The lack of fresh OC supply might be one important aspect of C stabilization in deep soil layers (Fontaine et al., 2007). In the bulk, non-rhizosphere soil, the microbial biomass is limited by the availability of C and/or N. Nutrient

Part 2: General overview Chapter 2 – SOM protection mechanisms sources must be derived from in-situ sources (the native soil organic matter, microbial turnover) or from soluble substrates leached out from upper horizons (Kemmitt et al., 2008).

In addition, organisms can be specialized and have a preferential substrate utilization. They are generally classified into K and R strategists in terms of substrate use, both abundant within bacterial and fungal phyla (Paul and Clark, 1989, Fontaine et al., 2003, Kuzyakov and Blagodatskaya, 2015). R-strategists only decompose the easily assimilable OM, and are consequently adapted to grow quickly depending on availability of their specific substrate because of the high competition for it. K-strategists are capable of degrade more recalcitrant substrate, and grow slowly. They colonize only at the late stages of plant residues decomposition. Even if large amounts of easily degradable substrate are supplied, K-strategists may not have enough time to assimilate these because they respond too slowly as compared to R-strategists. The decomposition pathways that require a cascade of degradation steps performed by different micro-organisms can be delayed by the accidental and temporary lack of one member of the succession chain (Ekschmitt et al., 2005).

To conclude, the combined effect of selective habitat exploitation and preferential substrate utilization represent a “partial refuge” - sensu Ekschmitt et al. (2005) - for OM, having as consequence that some proportion of accessible substrate will remain unaffected by the decomposer community.

Part 2: General overview Chapter 2 – SOM protection mechanisms

 Second compartment of the model:

Physico-chemical compartment: the pore space (from the (nano- )pores of the primary particles to the macropores formed within and between the macroaggregates), filled up with air and water. It is the habitat of soil microorganisms, and water within the pores the medium for chemical and biological reactions.

 Inhibition of decomposers activity in the model:

Soil architecture influences biological decomposition through impacting both the (1) environmental conditions and (2) accessibility of OM.

Inhibition of decomposers activity

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.3 Adsorption onto minerals and coprecipitation

As developed in the chapter 1, SOM consists of a complex mixture of molecules presenting different morphologies and chemical structure, and they follow a gradient of biological oxidation. This continuum of decomposability can be altered by interactions with minerals, able to stabilize organic substances and protect them against biological oxidation. This protection is not a complete and permanent removal of OM from decomposition; it rather reduces the decomposition rate as compared to that of similar unprotected materials (Baldock and Skjemstad, 2000). Growing evidence demonstrates that the quantity and strength of organo-mineral bonds are crucial to control the long term stabilization of SOM (decades to millennia) (Kögel-Knabner et al., 2008b).

2.3.1 Mineral transformations during pedogenesis

Minerals constitute the major part of soil volume and mass. In parallel to SOM transformations, starting from coarse plant residues to small and highly oxidized SOM molecules, mineralogical transformations occur during soil genesis, starting from a parent material (primary minerals) to the formation of secondary minerals involved in organo-mineral associations (OMA). The mineralogical modification processes described in this section are based on the reviews of Churchman and Lowe (2012) and Cornell and Schwertmann (2003).

2.3.1.1 Weathering of primary minerals

Mineralogical transformations start with the exposure of the parent material to weathering. Weathering, in most cases, involve the reaction of the minerals with water (hydrolysis). A major soil formation factor is the composition of the parent material in primary minerals (and possibly also in secondary minerals), and their relative stability to

Part 2: General overview Chapter 2 – SOM protection mechanisms weathering (solubility in the weathering solution). Their stability depends on their chemical composition and mineralogical constitution. The relative stability towards alteration of the principal soil primary minerals has been established by Goldich (1938). The order of the different minerals is in the identical (but reverse) order that the Bowen’s reaction series (Bowen, 1922) describing the order in which the minerals crystallized out of a magma on cooling (Figure 2.5). Some components of the primary minerals can greatly influence the rate and nature of their breakdown. Among them, Fe is the most important, and common in many primary minerals. Fe generally occur in reduced Fe(II) form in primary minerals, like less abundant Mn. Fe oxidation results in mineral structure weakening. Consequently, of the large group of rock silicates, those containing Fe(II) are the least stable members.

Figure 2.5 – Bowen’s reaction series (left) and Goldich stability series (right) (SlidePlayer.com)

After the mineral nature, the second factor impacting weathering rates is the composition of the surrounding solution (nature and concentration of the reagents, pH, redox potential, presence of organic ligands). Though soil weathering often occurs in acid conditions, mineral dissolution and neoformation involve H+- consumption, hence increasing pH. The process may also involve a

Part 2: General overview Chapter 2 – SOM protection mechanisms range of organic compounds. The latter can complex some cations released from silicates, enhance their breakdown and affect the products formed. Furthermore, soil living organisms impact directly or indirectly mineral dissolution. Refer to section 2.3.8.2 for the impact of microorganisms and plants on mineral weathering.

The weathering of primary minerals results in the release of their constituting elements in solution. In soils, primary minerals remaining after chemical and physical weathering will occur most commonly in coarser particles (sand- and silt-sized). Phyllosilicates (quartz, micas, chlorite, vermiculite, talc) and some oxides (such as those of titanium) and also apatite, occur in the clay fraction.

2.3.1.2 Formation of secondary minerals

Once solutes are released from the hydrolysis of primary minerals, “secondary” pedogenic phases form. The most important secondary minerals are Al and Fe oxides, and aluminosilicates (refer to section 2.3.2 for their description). Secondary minerals are the most reactive inorganic constituents in soils. Once formed, they can be re- dissolved. Secondary minerals formation and dissolution is impacted by (1) pH; (2) the composition of soil solution (inorganic and organic ions); (3) the action of living organisms (refer to section 2.3.8); (4) the redox potential (for Fe- and Mn- containing minerals).

As described here above, the composition of the soil solution plays a decisive role in the kinetics, course and ultimate product of weathering and secondary synthesis. The soil solution composition vary greatly in space and time, due to constant water and gas exchanges in the porous system. Notably, leaching governs the rate of weathering by removing solutes and displacing dissolution equilibria, as well as the location of secondary mineral formation. Consequently, the type of secondary phases formed and their fate is difficult to predict in soils.

 The pore space is both the habitat for microorganisms, and the reactor for biological and chemical reactions.

Part 2: General overview Chapter 2 – SOM protection mechanisms

Methodological box 3.

In the podzol chronosequence, soil weathering and formation of secondary phases (involved in OMAs) have been followed up using classical soil characterizations (total elemental content, selective extractions*, XRD, etc.) as well as three geochemical tracers (rare earth elements, Si isotopes and Fe isotopes) (chapter 6).

*Diagram 1: Fe- Al- and Si- phases extracted by pyrophosphate (p), oxalate (o) and dithionite/citrate/bicarbonate (DCB).

Oxyhydroxides Silicates Organo-mineral

Complexes Poorly Crystalline ITM Crystalline crystalline Fep Alp Sip Feo Alo Sio FeDCB AlDCB SiDCB

Part 2: General overview Chapter 2 – SOM protection mechanisms

 Third compartment of the model:

Mineral compartment: mineralogical transformations during soil genesis, starting from a parent material to the formation of secondary minerals involved in organo-mineral associations (OMA).

 OMA in the model:

The quantity and strength of organo-mineral bonds are the major control on long-term SOM stabilization. OMA can form significantly only after modifications have occurred in the SOM and mineral compartments: - Organic substrates have been transformed into microbial biomass and products of decomposition, and incorporated into the clay fraction (also refer to section 2.3.3). - Mineralogical transformations have led to the formation of secondary pedogenic minerals.

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.3.2 Minerals involved in OMA

The most significant minerals involved in the formation of OMAs are sesquioxides (oxides, hydroxides, oxihydroxides of Al and Fe), phyllosilicates and short range-ordered aluminosilicates, as described by Kleber et al. (2015). Their reactivity depends on three types of surface properties: (1) Variable charges, occurring on the surfaces of metal oxides and short range-ordered aluminosilicates, and at the edges of phyllosilicates; they are caused by pH-dependent processes of protonation-deprotonation of surface hydroxyl groups (M-OH, where M represents a structural metal cation, such as Fe3+ or Al3+). These charges are increasingly negative as pH increases, and are positive for Fe- and Al- oxyhydroxides in most soil pH; (2) Permanent charges, occurring on external and interlayer siloxane surfaces of phyllosilicates (for example, smectite, vermiculite, illite); they are caused by isomorphic substitutions of structural cations by cations of lower charge within the crystal structure of clay minerals (Al3+ by Mg2+, Si4+ by Fe3+ or Al3+), resulting in a negative charge unaffected by pH; (3) Permanently non-charged surfaces, constituted by neutral siloxane surfaces of 1:1 phyllosilicates of kaolin group minerals such as kaolinite.

The properties controlling the ability of mineral particles to interact with inorganic and organic solutes are the surface charge characteristics (type and density of charges); and the size and specific surface area (SSA) of the mineral particles (von Lützow et al., 2006, Kögel-Knabner et al., 2008b). The amount of reactive surface groups per unit mass of mineral generally increases with decreasing particle size due to an increase in SSA (Kleber et al., 2015). Minerals in OMAs typically belong to the clay-sized fraction of (<2 µm), but the most significant and reactive components are nanoparticles, because of their small size and high SSA (Theng and Yuan, 2008). Nanoparticles are frequently observed in soils, where they are present as surface coatings, associated with microorganisms and OM (Theng and Yuan, 2008).

Part 2: General overview Chapter 2 – SOM protection mechanisms

Iron oxides, hydroxides and oxihydroxides (hereafter referred as « iron oxides » or Fe oxides) in soil are of small crystal size and/or low crystal order. They are thus particularly important for adsorption processes. The following description of the main forms of Fe oxides is based on the review of Cornell and Schwertmann (2003). Fe oxides consist of arrays of Fe(II) or Fe(III) ions and O2- or OH- ions, arranged in octahedra and/or tetrahedra, and are characterized by a very low solubility. Among the 16 existing Fe oxides, the most common in soils are:

- Goethite (α-FeOOH, 8 – 200 m² g-1): the most common Fe oxide in soils by far, and one of the most stable thermodynamically. It is consequently either the first to form or the end product of many transformations. -1 - Hematite, (α-Fe2O3, <40 m² g ): the second most frequent Fe oxide in soils. In contrast to goethite, it is more widespread in soils developed in warmer, predominantly subtropical and tropical climates. It has a similar thermodynamic stability than goethite, and is also often the end member of transformation of other iron oxides. - Lepidocrocite (γ-FeOOH, 15 – 260 m² g-1): is less common than the two previous oxides, but not rare; it mostly occurs in redoxomorphic environments. It is metastable relative to goethite. -1 - Ferrihydrite (Fe5O8H.H2O, 100-700 m² g ): is widespread in soil environments. Unlike the other iron oxides, it exists exclusively as nano-crystals. Due to its metastable nature, it transforms into more stable iron oxides, unless stabilized in some way. Ferrihydrite thus occurs in relatively young soils, or in those in which its transformation to more stable oxides is inhibited or retarded. The conditions that hinder crystallization and lead to ferrihydrite formation instead of goethite are a high rate of Fe(II) oxidation, and the presence of a fair supply of OM and dissolved Si (like in the B horizon of podzols). Ferrihydrite is however often found associated with goethite. -1 - Magnetite (Fe3O4, 4-100 m² g ), and green rust are mixed valence oxides (containing both Fe(II) and Fe(III)).

Part 2: General overview Chapter 2 – SOM protection mechanisms

Almost all the iron oxides are crystalline, except ferrihydrite which is poorly crystalline. However, the degree of structural order and crystal size (and consequently the stability) vary depending on the conditions of formation (pH, organic ions, cation, etc.). Crystals of soil Fe oxides are usually less well developed than those of synthetic ones. The “poisoning” effect of OM on oxide crystallinity and growth is well documented (Zachara et al., 2002, Schwertmann, 1966) (Figure 2.6). Organics substantially enhance the SSA and alter surface charge characteristics (Huang et al., 2005). At high OM content, all Fe is organically complexed, and no oxides form.

Figure 2.6 – Schematic representation of the effect of OM content and rate of Fe supply on the formation of various Fe forms in soils (Cornell and Schwertmann, 2003).

Fe occurs in two main redox states in soils: oxidized ferric Fe (Fe(III)), which is poorly soluble at circumneutral pH; and reduced ferrous Fe (Fe(II)), which is easily soluble and therefore more bioavailable (Melton et al., 2014). Once formed, Fe(III) oxides can persist for long periods of time due to their high thermodynamic stability. However, according to Cornell and Schwertmann, they can be dissolved either by complexation with organic compounds or by

Part 2: General overview Chapter 2 – SOM protection mechanisms reduction, the latter taking place only in anaerobic conditions, essentially via microbial activity. The Fe(II) formed, more soluble and mobile, can move in (by diffusion) or with (by convection) the soil water until it reaches aerobic environments, where it is re-oxidized and re-precipitated, often as Fe(III) oxides.

In contrast with Fe oxides, there are few insoluble, crystalline Al oxides, excepted in certain soils (andosols, ferralitic soils). Gibbsite - Al(OH)3 – is by far the most frequent type. It forms in hydrolytic, desilicated soil environments in which silica loss prevents the formation of kaolinite i.e. in humid tropical conditions (Duchaufour, 1997). As for Fe oxides, the presence of OM also impairs the crystallization of Al oxides (Xu et al., 2010). The most significant forms of Al in the clay-size fraction, are the pedogenic clay minerals. Among them, phyllosilicates, like illite, vermiculite and smectite, are the most important for the formation of OMA because of their important surface charge. Short range-ordered aluminosilicates (allophane and imogolites) are less abundant, excepted in andosols or in podzols (Gustafsson et al., 1995, Gustafsson, 2001). They present very high surface areas, ranging from 400 to 1500 m2 g-1 (Gustafsson, 2001, Parfitt, 2009). Allophane (Si3Al4O12.nH2O), is made up of hollow spherules with a diameter of 3.5-5 nm, while imogolite (Si2Al4O10.5H2O) presents more ordered hollow tubular structure with an outer diameter of ~2.1 nm and a wall thickness of ~0.7 nm (Parfitt, 2009).

2.3.3 Properties of OM

Some properties of OM promote the formation of OMAs. First, the physical nature of the organic phase (the size) is quantitatively relevant (Kleber et al., 2015). To allow dead plant and (micro-)faunal debris to be in contact to mineral surfaces and interact, OM has to undergo a reduction in size and a chemical alteration. The microbial “oxidative depolymerization”, and progressive transformation of plant residues to microbial components and byproducts, as described in the section 1.3, increase the solubility and thus the chemical reactivity of organic substances towards metal cations and mineral surfaces (Kleber

Part 2: General overview Chapter 2 – SOM protection mechanisms et al., 2015). DOC (<0.45 µm) represents a pool of mobile, small size molecular fragments, potentially highly reactive towards mineral surfaces (Kalbitz et al., 2000).

While the surface characteristics of minerals are well known, the chemical properties of SOM are poorly documented since they are extremely complex (Kleber et al., 2007). There is still no clear picture in the literature on the organization of organic molecules on the mineral surfaces, and dynamics of OMA. As proposed by Kleber et al. (2007), it is appropriate to view the components of decomposing SOM as consisting of a mixture of diverse but typically amphiphilic molecules, along a continuum of amphiphilicity that ranges from molecules that are solely nonpolar and hydrophobic, to those that are predominantly amphiphilic.

Nonpolar, hydrophobic compounds are composed predominantly of alkyl (C-C) and aromatic functional groups. They cannot be ionized and are electrostatically neutral at most soil pH values. Alkyl is part of the long chain alkanes forming lipidic molecules, like plant wax. Aromatic functional groups are found for example in lignin, tannins, cutin and suberin (Kögel-Knabner, 2002).

Aromatic functional group:

Mildly polar materials are for example many carbohydrates and their derivatives. They bear functional groups that do not ionize under typical soil and water pH conditions, like alcohols (C-OH) or ethers (C-O-C), also referred as O-Alkyl compounds (Baldock and Broos, 2012).

At the other end of the continuum are the hydrophilic, polar or charged functional groups, like carboxyls (R-COOH  R-COO-), - phosphoester, hydroxamate (–OH  –O ) and amines (R-NH2  R- + NH3 ), which can develop charge under typical soil and water pH conditions.

Part 2: General overview Chapter 2 – SOM protection mechanisms

Carboxyl: Hydroxamate: Phosphoester:

Amines: Quaternary Pyridinium: ammonium:

The hydrophilic and hydrophobic components of OM play different roles in the formation of OMA. While polar functional groups are able to interact with the charged mineral surfaces, hydrophobic organic solutes will be responsible of hydrophobic interactions, and tends to assemble into micelle-like structures in order to decrease total entropy (von Wandruszka, 1998, Wershaw, 1994, Kleber et al., 2007).

2.3.4 Bindings types

The description of the different binding types is adapted from von Lützow et al. (2006) and Kleber et al. (2015).

2.3.4.1 Ligand exchange

Inorganic hydroxyl groups present on variable charge surfaces can undergo a rapid ligand exchange reaction with negatively charged carboxyl or hydroxamate groups of the OM, resulting in the formation of a polar covalent bond (metal-O-C) and the release of OH- ion. This bounding mechanism, also called “chemisorption”, “specific adsorption”, or “innersphere complexation”, is the strongest and present very high bound energies (several hundred of kJ mol-1).

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.3.4.2 Outersphere complexation

Outersphere complexation, also called “physisorption” or “nonspecific adsorption”, occurs when an adsorbate is electrostatically held at the surface by long-range Coulomb forces between ions of opposite charges. The energies of this bounding mechanism are much lower (<20 kJ mol-1). This bounding can occur (i) between positive charges of variable charge mineral surfaces and negative charges of organic functional groups, or (ii) between negative charges of permanent charge mineral surfaces and positive charges of organic functional groups (cation exchange).

2.3.4.3 Polyvalent cation bridges

Organic anions are normally repelled from negatively charged surfaces in soils (for example, permanent charges on siloxane surfaces), but they can bind to polyvalent cations present on the exchange complex. The major polyvalent cations present in soil are Ca2+ and Mg2+ in neutral and alkaline soils, and hydroxyl-polycations of Fe3+ and Al3+ in acid soils. The strength of cation bridge follows the order: Fe3+ < Al3+ < Pb2+ < Ca2+ < Mn2+ < Mg2+. The mechanisms involved in cation bridging are diverse and may include innersphere and outersphere complexation, as well as weak interactions.

2.3.4.4 Weak interactions

Organic ligands can also be stabilized at mineral surfaces by three weak interactions. (i) H-bondings occur between a hydrogen atom with a positive partial charge and a partially negatively charged O or N atom (4-13 kJ mol-1). (ii) Van der Waals interactions occur between two permanent or temportary dipoles, arising from shift of orbital electrons to one side of one atom or molecule (2-4 kJ mol-1). (iii) Hydrophobic interactions are driven by the exclusion of non-polar residues (e.g. aromatic or alkyl C, or hydrophobic surfaces) from water to force the non-polar groups together (~2 kJ mol-1 at 25°C). Weak

Part 2: General overview Chapter 2 – SOM protection mechanisms interactions might occur in complement with the previous ones. Furthermore, Van der Waals and hydrophobic interactions might be responsible of bindings between permanently non charged surfaces and nonpolar organic molecules (alkyl or aromatic C).

2.3.5 Conceptual model of OMA structure

Many authors increasingly consider the OM not distributed evenly over the available surface, but organized on minerals in a patchy, multiple layered fashion (Kaiser and Guggenberger, 2003, Eusterhues et al., 2005, Hedges and Keil, 1995). Since OM is divers and multifunctional, several adsorption mechanisms probably operate simultaneously (multimode adsorption). Based on recent evidences, Kleber et al. (2007) propose a zonal structure of SOM on mineral surfaces (Figure 2.7).

The first “basal” layer is called the “contact zone”. The polar groups of the OM forms strong ligand exchange bounds with the charged surface groups present at mineral surface. Organic compounds directly attached to mineral surfaces do not participate in significant exchanges with the soil solution, and are protected against decomposition. The hydrophobic end of the amphiphilic molecules adsorbed to the surfaces form a “hydrophobic zone”. Hydrophobic interactions will be possible in that zone with hydrophobic moieties present in the soil solution. As stated by Kleber et al., (2007), the hydrophobic zone “must not be considered an essential and continuous layer present within all OMA, but rather a zone likely to be found to varying degrees in many of these associations”. Finally, with increasing distance from the surface, a variety of modes of attachment of further organic molecules is possible. In that “kinetic zone”, molecules are presumably more loosely bound and have faster exchanges with the solution.

Part 2: General overview Chapter 2 – SOM protection mechanisms

Figure 2.7 - The zonal model of organo-mineral interactions, from Kleber et al. (2007).

A result of this model is the presence of at least two different C pools in the OMA: (i) one is directly adsorbed to the minerals by strong bounds, and is more stable and possibly older; (ii) others are adsorbed in outer regions; they are the most actively cycling and younger C part of the mineral-stabilized C (Sollins et al., 2006, Rillig et al., 2007, Swanston et al., 2005, Schmidt et al., 1990).

Consequently, each soil exhibits a finite capacity to protect OM, depending on the total charges and surface available (Grandy and Neff, 2008). When that capacity is saturated, further addition of OM will remain bioavailable (Baldock and Skjemstad, 2000). The protection depends on surface loading. When the surface loading increases, the proportion of OM not directly attached to the mineral surface, and weakly protected, increases (Kaiser and Guggenberger, 2003, Kaiser

Part 2: General overview Chapter 2 – SOM protection mechanisms and Guggenberger, 2007, Ellerbrock and Kaiser, 2005). Organic matter loadings in sediments and soil are often near those of the called “mono-layer-equivalent coating”, assumed to be in the range of 0.5 - 1 mg C per m2 of surface area (Mayer, 1994b, Kögel-Knabner et al., 2008b, Mayer, 1994a). The term monolayer-equivalent is stressed, because organic matter is not dispersed over all mineral surfaces, but present a patchy distribution. However, the loading can largely vary between soil types and horizons within the same soil (Kögel-Knabner et al., 2008b), depending on (1) the pH, (2) competition with inorganic ions for sorption sites (see next section), (3) the OM input (Kaiser and Guggenberger, 2003), and (4) the type of mineral phases present. For example, Fe oxides might be the most important sorbents for the formation of OMA in the subsoil, with OM loadings in the range of 1 - 2 mg C per m–2 oxide surface (these loadings must be regarded as maximum values) (Eusterhues et al., 2005, Kögel-Knabner et al., 2008b).

In addition to adsorption processes, OM can also form coprecipitates with Al and Fe oxides of decreasing sizes. A stated by Kleber et al. (2015), “in natural systems with dissolved OM being ubiquitously present, Fe and Al oxides always form as complex mixtures, encompassing a continuum from low-polymeric metal– organic complexes to well crystalline phases with surface-attached OM”. In contrast to OM adsorption on pre-existing surfaces, the process of co-precipitation leads to adsorption and occlusion (physical entrapment) of organic molecules in the interstices between the oxide crystal particles (Eusterhues et al., 2014a, Cismasu et al., 2011). Different processes might be involved: (i) complexation of hydrolyzed Fe and Al species by OM; (ii) precipitation of the metal–organic complexes, eventually followed by nucleation and crystal growth; (iii) adsorption of OM to neoformed Fe and Al oxides; and (iv) occlusion of pure OM into precipitated Fe and Al oxide aggregates (Kleber et al., 2015). The impact of the formation of coprecipitates on C stabilization and the factors governing their reactivity and physicochemical properties (e.g. the impact of ligand type) remain poorly known. Furthermore, a lot of studies concern synthetic coprecipitates. Eusterhues et al. (2014b) found adsorbed OM loadings in accordance with the monolayer equivalent loading (0.52 mg m-2), but considerably

Part 2: General overview Chapter 2 – SOM protection mechanisms larger organic matter loadings in coprecipitates (~ 1.1 mg m−2). The authors explain these differences by a larger surface area of coprecipitated ferrihydrites or by the presence of occluded OM in addition to adsorbed organic matter in coprecipitates (Eusterhues et al., 2014b). In nature, however, secondary Fe oxides precipitate from solutions that contain a multitude of inorganic and organic compounds that coexist with each other, for example after oxidation of Fe2+ at anoxic-oxic interfaces, and in transient conditions (permanently changing soil solution composition, redox state, etc.) (Fritzsche et al., 2015). These factors will considerably influence the properties of the obtained Fe oxides.

Complexation of Al and Fe ions by organic substances increases considerably the substrate stability (Sollins et al., 1996). The two ions behave differently in soils (Duchaufour, 1997): Fe2+ is the only soluble form of ionic Fe, but is quickly oxidized in Fe3+ in oxic conditions (decreasing reaction rate with decreasing pH, and very slow if pH<3) (Cornell and Schwertmann, 2003). Aqueous Fe3+ thus rapidly precipitates as Fe oxide. The Fe2+ ionic or complexed forms of Fe are typically much less efficient in complexation with OM than Fe(III). Besides, they are much less abundant, except in anoxic or very acidic conditions (podzols or hydromorphous soils) (Duchaufour, 1997). In contrast, the ionic and exchangeable form of Al is more represented, even in oxic environments under acidic (pH <6.0) or alkaline (pH >8.0) conditions, and/or in the presence of complexing ligands. Al3+ is the only stable ion, and is strongly selectively adsorbed on the exchange complex because of its high valence. Aqueous Al3+ is the most chemically and biologically available form, and can be toxic for plants and microorganisms (Sollins et al., 1996).

2.3.6 Controls on adsorption/coprecipitation processes

The adsorption-desorption processes are controlled by different characteristics of the soil solution. First, the pH affects the surface charge on variable-charge mineral surfaces and the extent of dissociation of functional groups of organic compounds. Therefore it

Part 2: General overview Chapter 2 – SOM protection mechanisms affects sorption processes (Weng et al., 2005, Sollins et al., 1996). Higher OM adsorption is measured at low pH (Weng et al., 2007).

The solution composition and concentration also impacts adsorption processes. The presence of inorganic cations can have multiple impacts. They are assumed to establish bridges between the negatively charged clay surface and the negatively charged organic functional group (Sollins et al., 1996). However, in some soils they can also induce a clay flocculation, and therefore decrease the surface available for adsorption (Baldock and Broos, 2012). The presence of inorganic anions such as phosphate or chloride can markedly decrease the amount of organic ligand adsorbed due to a competition for the adsorption sites (Hiemstra et al., 2010b).

A release of the adsorbed or coprecipitated OM can occur following the dissolution/transformation of the mineral counterpart. Most of the dissolution of clay minerals in soils is biologically (root, bacteria and fungi) mediated (Bonneville et al., 2011), and will be discussed in section 2.3.8.

Ligand structure and concentration greatly affect the adsorption processes. At low ligand concentration in the soil solution, the formation of multiple, strong innersphere complexation per molecule is possible; it allows a better stabilization (Kaiser and Guggenberger, 2007). However, low OM concentrations might also cause a “local disequilibrium” between the solid and liquid phase, and induce desorption of adsorbed OM (Kahle et al., 2003, Hedges and Keil, 1999, Feng et al., 2014). Here the sorption equilibria depend on the strength of the bounding: very slow or no desorption is observed for innersphere complexes, whereas larger desorption is measured for weaker electrostatic forces (Mikutta et al., 2007, Kaiser and Guggenberger, 2000). Furthermore, different organic ligands also compete with each other for adsorption sites. A preferential adsorption of more hydrophobic (aromatic, alkyl), less decomposed, high-molecular weight, globular OM over that with contrary properties, is frequently reported in the literature (von Lützow et al., 2006, Kleber et al., 2015, Cornell and Schwertmann, 2003, Kleber et

Part 2: General overview Chapter 2 – SOM protection mechanisms al., 2007, Kalbitz et al., 2005, Specht et al., 2000, Guggenberger and Kaiser, 2003, Kaiser and Guggenberger, 2007).

As far as co-precipitation processes are concerned, the key factors influencing the precipitation of OM by hydrolyzing metals are (1) the soil solution molar metal-to-carbon (M/C) ratio (with polymeric Fe species forming at lower M/C ratios than respective Al polymers); (2) pH and other factors impacting the type of mineral that will be neoformed (cations, redox state, etc.); (3) the specific affinity of a given metal cation toward a given organic complexant (Kleber et al., 2015, Cornell and Schwertmann, 2003). Preferential coprecipitation of N-poor, aromatic, carboxyl rich compounds has been documented (Scheel et al., 2008). Co-precipitated OM may differ in amount and composition from adsorbed organic matter (Eusterhues et al., 2010). However, more research is needed to understand the properties of these co-precipitates.

Due to the preferential adsorption/coprecipitation with minerals of some organic components, the soil can be considered as a chromatographic column with the more surface-reactive compounds being retained and the more mobile ones being leached downward the profile in the aqueous phase (Marschner et al., 2008, Neff and Asner, 2001, Kaiser and Kalbitz, 2012). Furthermore, it is well known that “fresh” DOM (high-affinity plant-derived compounds for example) may displace “old,” low-affinity adsorbed OM compounds from mineral surfaces which are subsequently released back into the solution by competitive sorbate displacement (Kaiser and Kalbitz, 2012, Gu et al., 1996a, Gu et al., 1996b). Depending on the porosity, and therefore the soil texture and structure, OM can also bypass the chromatographic soil column if it is in conditions of preferential flow (Kaiser and Kalbitz, 2012). Consequently, the OM composition of OMA at different depth will depend on (i) the C input reaching the concerned horizon after being transferred from the surface with soil solution (its composition is impacted or not by competitive sorbate displacement, depending on soil texture); (ii) the microbial transformation in situ; (iii) the root inputs. This complex combination of factors explains the differences in OM compositions as measured for example in OMA at different soil depths (Rumpel and Kögel-

Part 2: General overview Chapter 2 – SOM protection mechanisms

Knabner, 2011, Rumpel et al., 2004), or between sandy or loamy soils (Kögel-Knabner et al., 2008b). For example, an accumulation of small, highly oxidized organic acid can be observed in the subsoil fine fraction of a sandy soil, even if these components are not preferential sorbents for minerals (Rumpel et al., 2004).

2.3.7 Role of OMA in soil carbon storage and preservation

2.3.7.1 Evidences

The stabilizing effect against degradation resulting from the binding of OM to mineral surfaces have been deducted from three different ways in the literature.

Correlation between the clay/oxide content and accumulation of C. As reviewed by Baldock and Skjemstad (2000), Kleber et al. (2015) and Sollins et al, (1996), many studies observe a positive correlation, in a range of soil types, between soil C concentrations and clay or Fe/Al oxides contents, suggesting their influence on SOM stabilization (indirect evidence).

Increase in age and turnover times. As defined by Torn et al. (2009), the turnover time of a reservoir is the time it would take for the reservoir to be completely emptied if there were no further inputs (its refresh rate); while the average age of C atoms in the reservoir is the average time spent by all the atoms currently in the reservoir. The calculation of the age and turnover time of C is usually made by isotopic measurements (description of this section comes from the reviews of Torn et al. (2009), and Trumbore (2009)). Carbon has three isotopes: 98.9% of earth’s C is 12C, ~1.1% is 13C (a stable isotope) and about 1 in 1012 carbon atoms is 14C. Due to differences in their photosynthetic pathways, C3 and C4 plant biomass have different 13C/12C ratios (C3: δ13C ≈ -27‰; C4: δ13C ≈ -13‰; Still et al. 2003). Where a vegetation changes from C3 to C4 plants (or vice-versa), the rate of change of 13C/12C ratios in SOM gives an estimation of the turnover time of SOM, and provides a tool for studying C dynamics on decadal timescales. Radiocarbon (14C) is a useful tool for studying C

Part 2: General overview Chapter 2 – SOM protection mechanisms exchange between terrestrial ecosystems and the atmosphere on two timescales: (1) Radiocarbon is an unstable cosmogenic radionuclide, continuously produced in the upper atmosphere, and decaying with a half-life of 5,730 y. After a C reservoir ceases to actively exchange 14C with the atmosphere, its 14C content begins to decrease because of radioactive decay. For C pools like SOM, constantly receiving new C inputs from plants, and loosing C through decomposition, the 14C/12C ratio in a given organic matter pool reflects both the radioactive decay and the rate of decomposition. It is useful for studying very stable C pools, and on long time scales (until the order of 60,000 years). (2) During the early 1960s, atmospheric thermonuclear weapons testing nearly doubled the amount of 14C in the atmosphere. The most straightforward application of this “bomb 14C” is to compare the 14C content of SOM sampled prior to 1960 with the one of contemporary samples from the same location. Many studies showed a correlation between the turnover times of bulk SOM and the abundance of specific reactive secondary minerals (for example, Fed, Feo and Alo concentrations) (refer to references in Kögel-Knabner et al., 2008b or Kleber et al, 2015). Furthermore, SOM in clay fractions, or in the OMA is generally older or has a longer turnover time than OM in other soil OM fractions (light, free particulate OM), or in bulk SOM (references in von Lützow et al. (2006)).

Decrease of the degradation rates. The stabilization of OM by OMA can be traced by the measurements of C mineralization in incubation experiments. Laboratory incubations provide a controlled environment for characterizing and comparing C and nutrient dynamics in isolated soils (Torn et al., 2009). A diminution of degradation rates has been observed with addition of Al and Fe oxides or aluminosilicates (see references in Sollins et al, 1996 and Baldock and Skjemstad, 2000). Incubations provide one way to quantify the amount of fast-turnover C in soils, and they are useful for comparative and process-level investigations (Torn et al., 2009). However, they are subject to artifacts, and it is generally inadvisable to extrapolate rates from the laboratory to ecological settings.

Part 2: General overview Chapter 2 – SOM protection mechanisms

Methodological box 4.

We decided to use incubation experiment to evidence C stabilization in our samples. We compare degradation rates in the different horizons of the chronosequence in chapter 7.

2.3.7.2 Mechanisms

Various mechanisms might be involved to explain the protective effect on degradation of the adsorption of OM on minerals (Von Lützow et al., 2006). First, the adsorption of the extracellular enzymes relevant for OM degradation onto clay minerals rather than to the substrate reduces its activity. Secondly, OM adsorption can result in conformational changes, like changes in molecular size, charge and steric properties, that will decrease their accessibility to the action of extracellular enzymes, and thereby their availability as microbial substrates. Finally, the adsorption affinity of a reactive group to the mineral might exceeds that of the enzyme active site. Consequently, the cleaving of linkages between SOM surfaces and metal ions requires additional energy that must be overcome before decomposition, protecting SOM compared to non-adsorbed substances.

2.3.8 Interaction OMA - microorganisms

In addition to their role in OM degradation, and their contribution to SOM after microbe cell death, microorganisms also impacts actively OMAs in different ways.

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.3.8.1 Via the organic phase

Microorganisms can interact with the mineral phase through direct attachment of cells (via the production of outer membrane proteins) and deposition of mineral products (Extra-cellular polymeric substances - EPS) (Huang et al., 2005). Almost 80–90% of the soil microorganisms strongly attach themselves to mineral surfaces (Nannipieri et al., 2003). Microorganisms do not live as pure cultures of dispersed single cells, but instead they accumulate at interfaces to form poly-microbial aggregates, called ‘biofilms’ (a review on biofilms, is provided by (Flemming and Wingender, 2010). In most biofilms, the microorganisms account for less than 10% of the dry mass, whereas the matrix - consisting of EPS - can account for over 90%. EPS consists of a conglomeration of a large range of different types of biopolymers (mainly polysaccharides and proteins, followed by nucleic acids, lipids and other). They largely vary depending on microorganisms and physiochemical conditions. EPS are notably responsible for the strong adhesion to mineral surfaces and for the cohesion between cells in the biofilm. They constitute a protection against desiccation or toxic substances, permit redox reactions, and the storage of nutriments. According to Flemming and Wingender (2010), “some components of EPS are only slowly biodegradable and, owing to the complexity of EPS, complete degradation of all components requires a wide range of enzymes”. EPS are not unique to bacteria. Fungi (yeasts and moulds) and microalgae readily produce EPS. Fungi exude polysaccharide and glycoprotein mucilages which form a protective and lubricating matrix around hyphae (“fungal glue”), and which often have adhesive properties to stick soil components together (Ritz and Young, 2004). Due to the large ratio [surface area:volume] of the fungal mycelium, fungi produce large quantities of hydrophobic compounds in the outer wall of hyphae to be well insulated from the external environment (Ritz and Young, 2004). Such insulating compounds, like melanins and hydrophobins, are generally resistant to decomposition, and carry out a range of functional roles by mediating the hydrophobicity (water repellency) of hyphae, attachment to surfaces, and acting as elicitors in biotic interactions (Wösten, 2001).

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.3.8.2 Via the mineral phase

Microorganisms play a fundamental role in mineral transformations in the natural environment (Gadd, 2007). Microorganisms can induce both the formation and dissolution of minerals, including metal oxides and phyllosilicates, the most reactive and relevant minerals for the formation of OMA in soils. The microbial impact can be indirect - through variation of the soil solution composition/chemistry by exudation and uptake processes - or direct – through direct contact and enzymatic oxidation or reduction.

 Mineral dissolution

Plants, fungi and bacteria can colonize mineral surfaces. Yet their relative impacts on mineral weathering are difficult to evaluate and separate from purely abiotic processes. They are thus still debated in the literature (Finlay et al., 2009). Nevertheless, it is now established that mineral weathering can be significantly accelerated, or even initiated, by microorganisms (Uroz et al., 2009). The role of fungi in these processes has long been recognized (especially for mycorrhizal fungi, Taylor et al. (2009)), but the relative importance of bacteria and the molecular mechanisms involved remain poorly understood (Uroz et al., 2009). Filamentous fungi have an advantage over bacteria, because (1) their ratio [surface area:volume] is larger, and (2) their energy supply is facilitated by a transport of C via their mycelia and eventual symbiosis with plant roots (thus it does not depend only on diffusion through bulk solution). Consequently, fungi are one of the prevalent microorganisms in the biotic weathering processes (Smits et al., 2009, Ritz and Young, 2004). Roots, fungi and bacteria use similar strategies for mineral dissolution, with the exception of the ability to exert physical force, restricted to fungi and roots, and the redox processes in anaerobic conditions, restricted to bacteria (Kleber et al., 2015).

(1) Roots and fungal hyphae grow in intimate contact with mineral grains, and they may exert mechanical forces by osmotically applied turgor pressure (Smits et al., 2009). The physical disruption

Part 2: General overview Chapter 2 – SOM protection mechanisms

(biomechanical process) induces an increase of the available surface area for dissolution of rocks and minerals, resulting in biomechanical weathering of the minerals (Bonneville et al., 2011, Gadd, 2007).

(2) Microbial consumption of major nutrients such as P, K, Mg, and Fe, decreases their concentration in soil solution or in the zone in direct contact with the cells, and may lead to an undersaturation of the soil solution. This promotes minerals dissolution by equilibrium displacement (Brantley et al., 2011, Finlay et al., 2009, Taylor et al., 2009, Bonneville et al., 2011).

(3) Roots, bacteria and fungi act on the solution pH by exuding a wide range of acids, and accelerate mineral weathering by proton- promoted dissolution (Uroz et al., 2009, Smits et al., 2009). Inorganic acids include carbonic acids formed from CO2 produced by respiration, as well as sulfuric and nitric acid produced by sulfur- oxidizing and nitrifying bacteria at the border of anaerobic and aerobic environments (Kleber et al., 2015). Organic acids are low- molecular weight substances, such as oxalic, citric, formic, acetic acid, formate, and actetate. Some of these organic acids also have complexing properties that will promote dissolution of minerals (see next point).

(4) Microorganisms can also induce a ligand-promoted dissolution. They secrete a range of organic chelators that can increase mineral weathering by (1) complexing ions in solution and decrease their concentration; or (2) complexing ions from the mineral surface and weaken the links with the rest of the solid (Smits et al., 2009, Uroz et al., 2009). Among the secreted organic ligands, siderophores are low molecular weight organic ligands with high affinity and specificity for iron binding (Kraemer, 2004, Neilands, 1995). Their secretion is an important mechanism to overcome the limited iron bioavailability due to the low solubility and slow dissolution kinetics of iron-bearing minerals, as iron is an essential micronutrient in most organisms (Kraemer, 2004, Neilands 1995).

Part 2: General overview Chapter 2 – SOM protection mechanisms

(5) While fungi only occur in aerobic environments, activities of bacteria in anaerobic environment have a profound geochemical importance (Gadd, 2007). Bacteria have the incredible capacity to use solid minerals for respiration and by doing so, induce a reductive dissolution (Newman, 2001). Respiration is the process of harvesting energy by transferring electrons from an electron donor to an electron acceptor (Newman, 2001). In dissimilatory metal reduction (DMR), microorganisms can use H2 or organic sources as electron donor, and transfer electrons to external redox active metals (instead of soluble terminal electron acceptor like oxygen) reducing it without assimilating the metal (reviewed by Weber et al., 2006, Melton et al., 2014, Lovley, 2013). DMR bacteria have the capacity to transfer electrons to Fe(III)- and Mn(IV)-oxides, or structural metals in clay minerals. The electron transfer processes involved are still a matter of debate. It can involve (1) a direct transfer of electrons from the cell surface to the mineral, through i.e., membrane-bound redox enzymes or bacterial “nanowires”; (2) or the use of “electron shuttles”, such as organic compounds with quinone moieties. The rate of reductive dissolution will be influenced by (1) the properties of the overall system (t°, UV light), (2) the composition of the solution phase (pH, redox potential, concentration of acids, reductants and complexing agents), and (3) the properties of oxide (SSA, stoichiometry, crystal chemistry, crystal habit and presence of defects or guest ions) (Cornell and Schwertmann, 2003). A wide phylogenetic diversity of archaea and bacteria capable of dissimilatory reduction have been identified throughout a range of chemical and physical conditions, demonstrating the ubiquity of this type of microbial metabolism. The two most comprehensively studied mineral-respiring organisms are Geobacter and Shewanella. Dissimilatory Fe(III) and Mn(IV) reduction is one of the most geochemically significant event that naturally takes place in soils and controls iron biogeochemical cycle in most environments.

Part 2: General overview Chapter 2 – SOM protection mechanisms

Methodological box 5.

Microbially-mediated reduction strongly influence the cycling of Fe, but its effect on the OM adsorbed on Fe phases is still poorly known. The objective of chapter 9 is to determine the impact of microbial Fe reduction on the release of mineral-associated C.

 Mineral formation

Biomineralization refers to the processes by which organisms form minerals. These processes have consequences on OMA, by creating new minerals that may bound OM (Kleber et al., 2015). Biomineralization processes are divided into two fundamentally different groups based upon their degree of biological control (reviewed by Weiner and Dove (2003)). In “biologically controlled” mineralization (BCM), the organism uses cellular activities to direct the nucleation, growth, morphology and final location of the mineral. BCP processes can be described as occurring mainly extra- or intracellularly. In extracellular mineralization, the cell produces a macromolecular matrix (composed of proteins, polysaccharides or glycoproteins) outside the cell, genetically programmed to perform essential regulating and/or organizing functions that will result in the formation of composite biominerals. Intracellular mineralization is a widespread strategy, and occurs within specialized vesicles or vacuoles that direct the nucleation of biominerals within the cell. In this situation, the cell exerts a high degree of control upon the resulting biomineral composition and morphology.

“Biologically induced” mineralization (BIM) is the secondary precipitation of minerals that occurs as a result of interactions between biological activity and the environment. The metabolic processes employed by the organism modify the local microenvironment, i.e. changes in pH, pCO2, or secretion products, creating conditions that promote the chemical precipitation of extracellular secondary mineral phases. These chemical conditions

Part 2: General overview Chapter 2 – SOM protection mechanisms favor particular mineral types in an indirect way, and the biological system has little control over the type and habit of minerals deposited. Cell surfaces often act as causal agents for nucleation and subsequent mineral growth: reactive functional groups on the cell surface electrostatically bind cations, which may lead to the local supersaturation needed to induce nucleation and crystal growth.

The iron biominerals are of particular significance because they represent approximately 40% of the minerals formed by organisms (Weiner and Dove, 2003, Cornell and Schwertmann, 2003). In soils, BIM is more important than BCM (Kleber et al., 2015). Microbially-mediated Fe(II) oxidation is known to contribute significantly to the iron biogeochemical cycle in oxic, acidic conditions (Cornell and Schwertmann, 2003), and can also occur, but less significantly, in anoxic environments (in the light: photoautotrophic- and in the dark: nitrate-dependent Fe(II) oxidation) (Weber et al., 2006). The Fe(III) that is produced by Fe(II) oxidizers rapidly precipitates as Fe(III) minerals (Kappler and Straub, 2005). The minerals formed include two-line ferrihydrite, lepidocrocite, goethite and akageneite (Melton et al., 2014). Dissimilatory Fe(III) reduction can lead to sursaturation of aqueous Fe2+ and precipitation of Fe(II)- bearing minerals including siderite, vivianite and geologically significant mixed-valence Fe(II)–Fe(III) minerals, such as magnetite and green rust (Weber et al., 2006, Zachara et al., 2002).

2.3.9 Difficulty to isolate OMA and methodological choices

The objective of SOM fractionation is the separation for analysis of homogeneous C pools that are characterized by distinct properties, such as C turnover rates or protection mechanism. However, most SOM properties present a wide range continuum, and operationally defined SOM fractions remain a mixture of heterogeneous compounds from various sources. Thus, the fractionation of SOM is challenging. No consensus exists in the scientific community on the fractionation schemes and protocols. A

Part 2: General overview Chapter 2 – SOM protection mechanisms review of the fractionation mechanisms have been provided by von Lützow et al. (2007).

The first category of procedures is a physical separation of fractions of different sizes and density, the OMA being the smallest and densest fraction. There is still no accepted standard protocol with respect to, e.g., the input energy to disrupt the aggregates, or the density of heavy liquids to isolate the mineral-associated OM.

The second category of procedures is to fractionate the different C pools by chemical treatments that are assumed to specifically either attack or isolate OM components. Chemical degradation of soil OM by oxidizing reagents is a method to preferentially remove OM structures unprotected by the mineral matrix (Torn et al., 2009). Furthermore, the chemical reactivity of organic matter to oxidation can be used as a proxy for readiness to microbial degradation (Torn et al., 2009). NaOCl have been shown to oxidize preferentially chemically labile compounds (such as polysaccharides and aromatics), and younger OM, leaving behind older compounds, inherently stable against chemical attack (black C and aliphatic compounds such as n-alkanes and n-fatty acids) and protected by association with soil minerals (Mikutta et al., 2006; Zimmermann et al., 2007, Kögel-Knabner et al., 2008b). Furthermore, NaOCl appears to cause less mineral alteration than other chemicals (Zimmermann et al., 2007, Mikutta et al., 2006)). After the oxidation with NaOCl, HF is used to release SOM from organo–mineral associations through dissolution of hydrated silicate minerals and the formation of complexes with Fe and Al (Von lützow et al., 2007). OM left after the NaOCl and HF treatment is hypothesized to have a recalcitrant chemical composition (Mikutta et al., 2006).

Part 2: General overview Chapter 2 – SOM protection mechanisms

Methodological box 6.

Among other techniques, we decided to estimate the amount of OM in OMA using the NaOCl-FH fractionation scheme* (Mikutta et al., 2006) (chapter 7). We are conscious that this fractionation scheme should be considered as approximate due to the potential for incomplete extraction and/or nonselective extraction inherent to all chemical procedures (Baldock and Broos, 2012).

* Diagram 2: C fractions extracted by the NaOCl-HF procedure.

Bulk C C C soil Total Stable Recalcitrant

Flotation NaOCl HF in water Mineral-

Particulate Protected Oxidizable OM C C In the framework of the lack of homogenized fractionation protocols, an attempt to evaluate the different procedures have been launched by Axel Don (Thünen‐Institute, DE), called the SOMFrac Ring Trial. The objective is to make a comparative study on SOM fractionation methods, and to evaluate existing and emerging SOM fractionation schemes in their ability to separate meaningful and as homogeneous as possible SOM fractions. We participated this Ring trial by applying the NaOCl-HF procedure to the soil samples. A detailed description of the ring trial are presented in appendix A4.

Part 2: General overview Chapter 2 – SOM protection mechanisms

2.4 Shematic synthesis

The description of SOM composition, decomposition and stabilization mechanisms presented in the chapter 1 and 2 can be synthetized in a model dividing soil in three compartments: (1) the SOM compartment (living and dead matters); (2) the mineral compartment; and (3) the physico-chemical compartment (the pore space) (Figure 2.8).

The modifications of SOM, starting from intact plant material to small and highly oxidized molecules, with increasing contribution of microbial-derived molecules represents a “SOM compartment” (Chapter 1 and section 2.1).

In parallel, mineralogical transformations occur during soil genesis, starting from a parent material (primary minerals) to the formation of secondary minerals involved in organo-mineral associations (OMA) (section 2.3).

Microorganisms are responsible of most modifications in both SOM and mineral compartments (SOM degradation – section 1.3; Minerals formation/dissolution - section 2.3.8).

The poral system is the habitat of soil organisms, and water within the pores the medium for chemical and biological reactions. The composition of the soil solution (pH, redox potential, concentration in organic and inorganic ions) plays a decisive role in (1) the kinetics, course and ultimate product of weathering and secondary synthesis (section 2.3.1); (2) the adsorption-desorption processes on the mineral surfaces (section 2.3.6); (3) the SOM uptake and utilization by soil microbial community (section 2.2). The soil solution composition vary greatly in space and time, due to constant water and gas exchanges in the porous system. Notably, leaching governs the rate of weathering (by removing solutes and displacing dissolution equilibria), and the location of secondary minerals formation. The soil can be considered as a chromatographic column with the more surface-reactive compounds being retained and the more mobile ones being leached downward the profile in the aqueous phase. Pedogenesis is thus controlled by the system hydrodynamics.

Inhibition of decomposers activity

Recalcitrance

Figure 2.8 - Conceptual model of the three soil compartments, evolving during pedogenesis (SOM, minerals and pore space), and repartition of SOM protection mechanisms

Part 2: General overview Chapter 3 – Podzol

Chapter 3 - Pedogenic model: Podzol

3.1 Podzol description and repartition

Podzol derives from Russian pod – underneath - and zola – ash. Podzols are characterized by an eluvial subsurface horizon (E) ashy, bleached, quartzitic or phytolithic and coarse textured, directly underlain by a dark accumulation horizon (B), enriched in brown or black illuviated SOM and/or reddish iron compounds (WRB, 2015).

Figure 3.1 – Podzol in Belgium (from the Global soil biodiversity atlas, FAO (2016)).

Podzolization is a soil forming process, where the leaching of elements plays a major role and leads to a development of two distinct horizons: (1) the surface horizon (E) where weathering is the main process, and leaching faster than the release of elements; (2) the underlying illuvial horizons (B) where secondary minerals form. Podzolization thus occurs in well drained soils (coarse-grained and silica-rich parent material), and under certain environmental

Part 2: General overview Chapter 3 – Podzol conditions (high precipitation and vegetation releasing strong organic acids from litter decomposition - like forest or heath vegetation). Podzols are common in continental cool summer, and marine west coast climates. Podzols cover more than 3% of the Earth's land surface (485 million ha worldwide), mainly in the temperate and boreal regions of the Northern Hemisphere (WRB, 2015). Tropical podzols occur on less than 10 million ha (WRB, 2015).

Figure 3.2 – Distribution of podzols (from WRB 2014)

3.2 Podzol formation

This part is a synthesis of the recent reviews about podzolization: Lundström et al. (2000a), Buurman and Jongmans (2005), Sauer et al. (2007)and Sanborn et al. (2011). The processes involved in podzolization are still subject to debate. Podzolization always involve the translocation of mobile constituents, but three major groups of processes have been proposed.

Part 2: General overview Chapter 3 – Podzol

The first is the formation and downward transport of soluble unsaturated complexes of organic acids with Al and Fe ions (Fulvate theory, McKeague et al. (1978)). In this view, the immobilization is explained by: (1) the precipitation of organo-metallic complexes due to continuing addition of metals during downward migration until a critical C:metal (C/M) ratio has been reached; this process decrease the solubility of organo-metallic complexes which finally precipitate; (2) the microbial decomposition of the organic part of the complex during downward migration, inducing a decrease of the C/M ratio and the precipitation of the complex. Microbial degradation of the organic complexes and subsequent release and precipitation of Al and Fe has also been proposed to explain formation of “imogolite type material” (ITM) in the B horizon. Formation, transport of Al and Fe through the profile as organic complexes, and precipitation has been generally accepted as the dominant mechanism of weathering, eluviation and immobilization in the illuvial horizon.

A second theory proposes a downward transport of Al and Si as inorganic colloidal sols (Proto-imogolite theory, Farmer et al. (1980)). This view derives from the common observation of inorganic, amorphous or imogolite-type Al–Si precipitates in podzol B horizons. This theory proposes, in a second step, the migration of organic acids and adsorption on the ITM. This theory is controverted, given the high complexing capacity of OM, and because most Podzols have concentrations of dissolved organic acids that are high enough to prevent the formation of proto-imogolite (only observed in the Bs horizon under the horizon saturated with OM).

A third theory, called the Fulvate-bicarbonate theory (Ugolini and Dahlgren, 1987), proposed two stages of profile development which occur sequentially or simultaneously: (a) in situ formation of imogolite/allophane in the Bs horizon by a carbonic acid weathering process, and (b) precipitation of organic acid on the Al-rich precipitates in the Bs horizon.

As described by Sauer et al. (2007), many other factors have been proposed to play a role in immobilization, like (1) the

Part 2: General overview Chapter 3 – Podzol polymerization of organic acids induced by an adsorption of base cations; (2) the reduction of Fe by organic acids and its migration in reduced metal-organic (MO) complexes, followed by a re-oxydation of OM-Fe complexes; (3) the gradual increase of pH with depth, that reduces the solubility of dissolved Fe and Al, or until reaching the point of zero charge (PZC) of the complexes and inducing their flocculation; (4) The physical-mechanical filtering effect of the soil pores; (5) the adsorption of organic substances to soil particles; (6) the accumulation of organic matter not as a result of illuviation, but to roots contribution.

O-horizon Ah-horizon

E-horizon

Bh/Bhs/Bs- horizons

Bw-horizon

BC-horizon

C-horizon

Figure 3.3 – Schematic representation of a podzol (adapted from (Davydchuk et al., 1990)). 78

Part 3 Chapter 4 – Description of the study site

Part 3. Evolution of C stabilization mechanisms in a podzolic chronosequence

Chapter 4 - Description of the study site

4.1 Geography and climate

The study site is a chronosequence located near Cox Bay, on the west coast of Vancouver Island, British Columbia (latitude 49° 6'N, longitude 125° 52'W, Figure 4.1). Climate is characterized by a lack of temperature extremes and abundant precipitation: the mean annual precipitation amounts to 3200 mm and the average temperature 8.9 is °C (Singleton and Lavkulich, 1987). Due to rainfall, moderate temperature and close proximity to the ocean, the humidity remains high throughout the year (between 75 and 95%) (Cordes, 1972). The incoming ocean spray brings significant amount of Na, Mg, Ca and K to soils. Such an input improves the soil nutrient status, without inducing nutrient accumulation, because of rainfall at the site (Cordes, 1972). The vegetation developed on this sequence is a sitka spruce forest with a dense understorey of shrubs. The other species present are notably salal, douglas fir, hemlock, red cedar, and western sword fern. Sitka spruce makes up to 89 % of the volume, Western hemlock makes up 4 %, and Western red cedar 7 % (Cordes, 1972).

Part 3 Chapter 4 – Description of the study site

Figure 4.1 - Location of the Cox Bay soil chronosequence (from Singleton and Lavkulich, 1987)

4.2 Geological context

The parent material of the soil chronosequence is the Cox Bay beach sand (Singleton and Lavkulich, 1987). Headland rocks surrounding Cox Bay and referred to as the Tofino Area Greywacke Unit is the source of the sandy parent materials in the sampled transect (Muller and Carson, 1969). Only slight variations in parent materials suggest a uniform depositional sequence in the study area (Singleton and Lavkulich, 1987). The primary minerals present in the beach sand C material are quartz, sodic feldspars, amphibole (hornblende), pyroxene (augite), kaolinite, micas (illite) and chlorite, as well kaolinite precipitating in the dissolution pits of feldspars (Cornelis et al., 2014a). Quartz is by far the dominant mineral. The

Part 3 Chapter 4 – Description of the study site parent material does not contain inherited clay minerals, except kaolinite present in the weathered feldspar.

Soils in the sequence have developed on sandy beach deposits, which are advancing toward the ocean in a configuration parallel to the existing shoreline at a rate of 0.26 m per year. Increasing soil development, and consequently progressive deepening and differentiation of genetic horizons during podzolization, is observed with increasing distance from the active beach (Cornelis et al., 2014). Five profiles (P) were sampled along a transect (0 - 147 m) perpendicular to the present shoreline (Figure 4.2). The ages of the profiles were determined by dendrochronology (tree-rings dating). As a strip of sand of approximately 13 m wide lies between the active beach and the sand deposits containing tree seedlings, 50 years (time needed for this strip to accumulate) were added to the tree age estimated for each site to determine the site ages (Singleton and Lavkulich, 1987).

Sitka spruce (P. sitchensis) alone is the forest cover at the youngest site (P1-120). It is associated with (1) salal (Gaultheria shallon) at P2-175 and P3-270 sites, and with (2) salal (G. shallon), western red cedar (Thuja plicata), western hemlock (Tsuga heterophylla), douglas fir (Pseudotsuga menziesii) and western sword fern (Polystichum munitum) at the oldest sites (P4-335 and P5-530) (Cornelis et al., 2014).

Figure 4.2 - Cross section of the Cox Bay chronosequence, showing site locations and soil horizons depending on their respective age of soil formation: C-0 year, P1-120 years, P2-175 years, P3-270 years, P4-330 years and P5-530 years.

Part 3 Chapter 4 – Description of the study site

4.3 Soil chronosequence sampling

One representative sample was taken per horizon, in the five pedons (P) (P1-120 years, P2-175 years, P3-270 years, P4-330 years and P5-530 years). After sampling, the soil samples were air-dried and sieved at 2 mm according to NFISO11464.

In total, 5 Kubiena Boxes were collected during the sampling campaign: (1) +4cm in Oh to -4 cm in BC1 of P1; (2) between 0 and -10 cm in the P1 BC1 horizon; (3) between -10 and -20 cm in the P3 Bh horizon; (4) between -13 and -22cm, in the P4 Bh, Bhs and Bs horizons; (5) between -4 and -13cm in the P5 E, Bh, Bhs and Bs horizons. The samples were dried in the oven at 50-60°C and impregnated under vacuum with an unsaturated polyester resin. Subsequently they were cut into slices, polished to a thickness of 25 to 30 µm and fixed on a glass slide. Afterwards they were studied with a polarizing microscope and described according to the terminology of Stoops (2003).

4.4 Analytical methods

A summary of the analysis carried out on the Vancouver chronosequence samples is presented in the Figure 4.3. This section presents the methods used in the Part 3. The methods used in the part 4 will be presented in the chapter 9.

Part 3 Chapter 4 – Description of the study site

Part 3: SOM protection mechanisms in the chronosequence

Chapter 4 - General soil characterizations:

- pH, electrical conductivity - Exchangeable cations and cationic exchange capacity - Particle size analysis - Total elemental content

Chapter 7 Chapter 5 Chapter 6

SOM: - Kubiena boxes - Selective - Total C and N content - Micromorphology extractions - NaOCl-HF Fractionation - XRD analysis - Incubations - REE - Fe isotopes Biota: - Si isotopes - Amino Sugars - PLFA

Part 4 : OMA dynamics in redox environment

Chapter 9 - Microbial incubations (S. putrefaciens) - Scanning Electron Microscopy (SEM)

Figure 4.3 - Summary of the analysis carried out on the Vancouver chronosequence samples.

Part 3 Chapter 4 – Description of the study site

4.4.1 Major soil properties

Classical soil characterizations were performed. Soil pH was measured in 5 g:25 ml soil:water suspension (Page et al., 1982). Cation exchange capacity (CEC) and the content of exchangeable cations were determined, according to Page et al. (1982), in ammonium acetate 1 M at pH 7 and measured by ICP-AES. Soil particle-size analysis was achieved by quantitative recovery of clay (<2 µm), silt (2– 50 µm) and sand (>50 µm) fractions after sonication and dispersion + with Na -saturated resins without any previous H2O2 oxidation of OM, as described in Henriet et al. (2008). Total major elemental contents were measured by inductively coupled plasma/atomic emission spectrometry (ICP–AES) after fusion in Li-metaborate + Li-tetraborate at 1000 °C (Chao and Sanzolone, 1992). This was performed on samples prepared according to NF ISO 11464 and crushed to <250 μm as recommended by NF × 31147. The soil weathering degree in the different soil horizons was estimated by measuring the Total Reserve in Bases (TRB). It is the sum of the total contents of alkaline and alkaline-earth cations (Na+, K+, Ca2+, Mg2+; in cmolc·kg−1). TRB is a proxy for the weathering stage of soil because it estimates the content of weatherable minerals (Herbillon, 1986). The TRB can be used to compare soil horizons to the parent material, evaluating the relative loss of Ca, Na, K and Mg cations during weathering.

4.4.2 Mineralogical characterizations

The soil mineralogy of the fine earth fraction (<2 mm fraction, crushed to a fine powder with an agate mortar and pestle) has been determined using X-Ray diffraction (XRD). All diffractograms were studied using the X-Ray peak matching software EVA (Bruker) and its database of minerals patterns. The proportions of different crystalline minerals were calculated using SIROQUANT© 4.0, one of the most recent software using Rietveld refinement (Rietveld, 1969) for quantitative analysis of XRD patterns. SIROQUANT uses full profile fitting routines to generate a synthetic pattern that can be systematically refined via a least-squares minimization of the

Part 3 Chapter 4 – Description of the study site difference with the diffraction pattern obtained experimentally. The output data also show the error associated with each individual component, the ESD value (“estimated standard deviations” of the weight percentages). These errors are the square roots of the diagonal elements of the least-squares variance covariance matrix. An estimate of the overall goodness of fit for each analysis is also provided in the output, expressed as a global chi-squared value. Errors given by Siroquant for each individual mineral determination has to be multiplied by the square root of the global chi^2 value for the analysis in question to give the total error per mineral.

Secondary mineral phases of Si, Fe and Al, recognized as playing a key role in SOM dynamics in soils, were selectively extracted from the fine earth soil fractions (<2 mm fraction, uncrushed) using sodium pyrophosphate, dark oxalate and dithionite-citrate- bicarbonate (DCB) following the methods of Bascomb (1968); Blakemore et al. (1987), and Mehra and Jackson (1960), respectively, and quantified by ICP–AES (Thermo Scientific, iCAP 6000). The fractions obtained with pyrophosphate (Alp, Fep) indicate principally Al and Fe present in organo-metallic complexes. However, alkaline extractant could also extract Al from Al hydroxide phases (poorly crystalline Al(OH)3) and from poorly crystalline aluminosilicates (Schuppli et al., 1983, Kaiser and Zech, 1996). The fractions obtained with oxalate (Alo, Feo, Sio) comprises principally i) Al and Fe in organic complexes, ii) Fe in poorly-crystalline (hydr)oxides (ferrihydrite and eventually lepidocrocite if present (Poulton and Canfield, 2005)), iii) Al and Si in poorly crystalline aluminosilicates. The Fe in poorly-crystalline (hydr)oxides and the Al in poorly crystalline aluminosilicates are obtained by difference between oxalate and pyrophosphate extract. The DCB-extractable Fe (Fed) provides an estimate of the content of “free iron” in soils, i.e. i) Fe in organic complexes ii) Fe in poorly- crystalline (hydr)oxides, and iii) Fe in crystalline oxyhydroxides such as hematite and goethite. Thus, the Fe from crystalline Fe oxyhydroxides is obtained by difference between Fed and Feo. These selective extractions are not fully quantitative, and must be treated with caution due to uncertainty about the origin of the extracted minerals, but can nevertheless be used as indicators of the relative evolution of the mineral phases (dividing broadly the secondary Fe and Al phases

Part 3 Chapter 4 – Description of the study site into crystalline, poorly crystalline and organically-bound fractions) as a function of the soil development under identical soil parent material. Within the total iron content (Fet) in soils, the Fed/Fet ratio reflects the relative proportion of free iron in the total Fe pool in soil.

4.4.3 Rare Earth Elements analysis and data treatment

Sample preparations for the rare earth elements (REE) analyses were carried out in clean environment (CEREGE, Aix en Provence). Approximately 250 mg of sample powder (<2 mm fraction, crushed to a fine powder with an agate mortar and pestle) was first treated with 30% H2O2 in order to eliminate OM and then dissolved using a mixture of concentrated HF-HNO3, followed by concentrated HCl acids, at ~130 °C. The dissolution was made under laminar flow box in order to minimize sample contamination. The dissolved samples were measured for REE by ICP-MS (NexION 300X, PerkinElmer). To ascertain the accuracy of the REE analysis, two international standards (GSS-2 and GSS-3) were analyzed, using the same technique and during the same analytical session than the “unknown” samples. Those standards were selected because they were estimated to have a REE concentrations close to the ones of the samples, based on the REE concentrations in other podzols in the literature. Furthermore, the repeatability of the analytical technique was verified by analyzing one sample (C horizon) in triplicate. The accuracy of the ICP-MS was estimated by measuring four times the concentrations of each international standards, as an unknown sample. One blank of the sample processing procedure was also included in the analytical session. During the course of the ICP-MS analysis, one blank and one international standard was measured every four samples. The reference and measured values of the international standards, as well as the measured values for the triplicates of the C horizon sample are presented in Table 4.1. All reagents were ultrapure distilled acids and overall procedural blanks contained negligible quantities of REE compared to the sample REE content. REE concentrations measured in the reference samples were within 90% of the reference concentrations for these elements and the errors on the measure were less than 5% for most of the REE.

Table 4.1 - Reference and measured REE concentrations of two international standards (GSS2 REF and GSS3 REF), and measured REE concentrations in three replicates of one C horizon sample from the study area (C-1 to C-3).

LREE (µg g-1) MREE (µg g-1) HREE (µg g-1) I.D. La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu C-1 18,390 34,770 4,750 20,086 4,378 1,366 4,137 0,697 4,345 0,863 2,551 0,348 2,320 0,320 C-2 21,261 39,535 5,243 21,503 4,803 1,426 4,516 0,723 4,557 0,920 2,698 0,373 2,433 0,368 C-3 22,504 39,508 5,285 21,614 4,876 1,363 4,296 0,729 4,665 0,927 2,780 0,382 2,483 0,364 GSS2 REF 164,00 402,00 57,00 210,00 18,00 3,00 7,80 0,97 4,40 0,93 2,10 0,42 2,00 0,32 GSS2-1 171,749 393,008 60,096 220,460 18,591 2,816 7,334 0,757 4,341 0,649 1,910 0,234 1,567 0,234 GSS2-2 164,868 409,497 58,277 212,030 18,249 2,931 7,552 0,937 4,161 0,775 2,198 0,306 1,837 0,306 GSS2-3 162,778 407,441 56,962 210,192 17,150 2,887 7,623 0,901 4,287 0,703 2,072 0,270 1,873 0,234 GSS2-4 166,687 401,918 57,736 215,723 18,357 2,841 7,791 0,919 4,576 0,757 2,180 0,288 1,964 0,270 GSS3 REF 21,00 39,00 4,80 18,40 3,30 0,72 2,90 0,49 2,60 0,53 1,50 0,28 1,70 0,29 GSS3-1 21,684 38,861 4,818 17,875 3,427 0,716 2,585 0,444 2,548 0,511 1,480 0,220 1,431 0,224 GSS3-2 20,818 39,895 4,670 17,238 3,275 0,715 2,662 0,417 2,350 0,475 1,337 0,206 1,305 0,206 GSS3-3 22,339 39,568 5,006 18,763 3,508 0,700 2,681 0,408 3,140 0,493 1,552 0,206 1,368 0,220 GSS3-4 21,518 39,713 4,755 18,149 3,490 0,678 2,629 0,444 2,467 0,520 1,534 0,224 1,494 0,215

Part 3 Chapter 4 – Description of the study site

Measured REE concentrations were normalized to both an external (the Upper Continental Crust-UCC (Rudnick and Gao, 2003, Laveuf and Cornu, 2009)) and an internal reference (the parent material from which the soil profile develops after Braun et al., 1998, Aubert et al., 2001, Dequincey et al., 2006), and presented as distribution patterns with the individual REE listed in the order of their atomic number in the x-axis.

Depletion or enrichment of a group or of an individual REE relative to the others was quantified through the calculation of fractionation ratios and anomalies. The fractionation between light REE (LREE, i.e., from La to Nd), medium REE (MREE, i.e. from Sm to Ho) and heavy REEs (HREEs, i.e., from Er to Lu) was quantified by the ratios La/Gd, Gd/Lu and La/Lu, calculated after normalization of the concentrations to the parent material. The magnitude of Eu and Ce anomalies (Eu/Eu* and Ce/Ce*), the only two REE encountered in two oxidation states under earth surface conditions, and that can consequently display a specific behavior, was calculated by the following ratios:

Equ. 1 푪풆 (푪풆풔풐풊풍/푪풆풓풆풇) 푪풆 − 풂풏풐풎풂풍풚 = [ ∗] = ퟏ/ퟐ ퟏ/ퟐ 푪풆 [(푳풂풔풐풊풍/푳풂풓풆풇) (푷풓풔풐풊풍/푷풓풓풆풇) ]

Equ. 2 푬풖 (푬풖풔풐풊풍/푬풖풓풆풇) 푬풖 − 풂풏풐풎풂풍풚 = [ ∗] = ퟏ/ퟐ ퟏ/ퟐ 푬풖 [(푺풎풔풐풊풍/푺풎풓풆풇) (푮풅풔풐풊풍/푮풅풓풆풇) ]

where the subscript “ref” correspond to the REE concentration in the reference material (UCC or soil parent material) and the subscript “soil” correspond to the REE concentration in the soil sample (Mourier et al., 2008, Ndjigui et al., 2008, Vázquez-Ortega et al., 2015).

Part 3 Chapter 4 – Description of the study site

4.4.4 Soil organic matter distribution and composition

Bulk total organic carbon and nitrogen concentration (CB, NB) were determined on 1 g powdered soil subsamples using a VarioMax CN dry combustion Analyzer (Elementar GmbH, Germany), with a measuring range of 0.2 – 400 mg C g-1 soil (absolute C in sample) and a reproducibility of < 0.5% (relative deviation).

The C and N fractions within total OM were estimated following the protocol of Mikutta et al. (2006), with an additional initial step to remove particulate OM (POM) and dissolved OM (DOM). Soils were first shaken overnight with deionized water, then centrifuged and the supernatant removed (containing POM and DOM). Afterwards, samples were dried at 60°C and a subsample was crushed and analyzed for total organic C and N content (CT and NT). In our soils, the POM and DOM content was low, and the bunk C and N (CB and NB) were not significantly different than the CT and NT.

The amount of NaOCl-resistant C and N (CS and NS) was quantified as the amount of C and N left after treatment of the soil sample with NaOCl (Kleber et al., 2005). Chemical degradation of soil OM by oxidizing reagents is a method to preferentially remove OM structures unprotected by the mineral matrix (Torn et al., 2009). Furthermore, the chemical reactivity of OM to oxidation can be used as a proxy for readiness to microbial degradation (Torn et al., 2009). The oxidizing NaOCl attack is reported to be one of the most efficient and reliable to isolate stable OM (NaOCl-resistant OM), without dissolving pedogenic oxides (Mikutta et al., 2005b, Siregar et al., 2005, von Lützow et al., 2007, Mikutta et al., 2006, Zimmermann et al., 2007). Briefly, 3 g of air-dried soil sample were reacted three times with 30 mL of 6 wt% NaOCl adjusted to pH 8.0 during 6h at 25°C. Samples were then washed twice with 30 mL 1 M NaCl and then with deionized water until the solution was chloride free (i.e. no reaction with AgNO3 occurred). The samples were then dried at 60 °C and homogenized before C and N measurement (CS and NS) on crushed subsamples. The “labile” C and N (CL and NL) was calculated by

Part 3 Chapter 4 – Description of the study site

subtracting the CS and NS fraction from the total C and N content (CT and NT) of the soil sample, respectively.

Within the stable fraction, we distinguished between the amount of mineral-protected (CMP and NMP) and recalcitrant (CR and NR) organic C and N. The NaOCl-treated samples were subsequently extracted with HF in order to dissolve mineral constituents and attached OM. OM left after the NaOCl and HF treatment is here defined as chemically resistant (= recalcitrant, sensu Mikutta et al., 2006). HF treatment is considered to only dissolve soil minerals and mineral-associated organic matter, while the non-associated soil organic matter remains nearly unaffected (Eusterhues et al., 2003). Briefly, 2.25 g of NaOCl-treated dry samples were transferred into pre- weighed centrifuge bottles, and shaken four times with 15 mL 10% HF during 2h, and then washed five times with 15 mL deionized water. Between each shaking step, the samples were centrifuged and the supernatant discarded. After the rinsing, the samples residuum were dried at 60°C, crushed and analyzed for OC and ON (CR and NR). The mineral-protected C and N contents (CMP and NMP) were estimated by the difference between the stable OC and N and the recalcitrant OC and N. At each step of the sequential extraction, the initial and residual sample weights were recorded, and C and N contents were expressed on a bulk soil basis (g kg-1 soil). The OC and N content in the fractions were measured on 10 mg subsample by flash dry combustion with a FLASH 2000 organic elemental Analyzer (ThermoFisher Scientific). The C/N ratio were calculated within each fraction. An analysis of the repeatability between replicates shows deviations of ca. 5, 4, 11 and 33 % between the replicates for total, recalcitrant, labile and mineral- protected C, respectively, and of 5, 7, 18 and 20 % between the replicates for total, recalcitrant, labile and mineral-protected N.

4.4.5 Soil microorganisms

4.4.5.1 Compound specific analysis of amino-sugars

The most abundant amino sugars (AS) in soils, glucosamine (GluN), galactosamine (GalN) and muramic acid (Mur) (Bodé et al.,

Part 3 Chapter 4 – Description of the study site

2013), were used as biomarkers for microbial residues. Amino sugar extraction and analysis was based on the procedure described by Bodé et al. (2009 and 2013), but with some modifications (the cation- exchange resin step was replaced by a precipitation of the impurities by KOH), to adapt to the high Fe and Al content of some soil samples. Briefly, air-dried soil samples (sieved at 2 mm, crushed) amounts corresponding to approximately 0.3 mg of N were hydrolyzed, then dried under reduced pressure, in a Rotavap device. To assess the recovery of the extraction and purification steps, and for calculating and correcting data for losses, 100 µL of a 1mg mL-1 myo-inositol solution were added after the hydrolysis step as an internal recovery standard. The residues were then re-dissolved in MilliQ water, and the suspensions pH adjusted to 6.6-6.8 with KOH(aq) to precipitate impurities (mainly Fe and Al). The solutions were then centrifuged, and the supernatant, containing the AS, freeze dried. In order to reduce the amount of inorganic salts in the final HPLC aliquots, the samples were re-dissolved in methanol, centrifuged and the supernatant (containing the AS fraction) transferred in a 10 mL glass tube. A second internal recovery standard was added at this step: 100 µL of a 1mg mL- 1 N-Methyl-D-glucamine solution. The sample was dried by volatilizing the methanol under a stream of nitrogen, then freeze dried. Afterwards, the AS were transformed into aldononitrile derivatives and re-dissolved in ethylacetate/hexane. The concentration of the basic amino sugar (glucosamine, galactosamine and muramic acid) in the analytical aliquot were determined by liquid chromatography using the method described by Bodé et al. (2009). The liquid chromatographic separation was performed using a LC pump (Surveyor MS-Pump Plus, Thermo Scientific, Bremen, Germany) mounted with a PA20 CarboPac analytical anion-exchange column (3- 150 mm, 6.5 μm) and a PA20 guard column (Thermo Scientific, Bremen, Germany). Basic AS (GluN, GalN, Mur acid) were eluted with 2mM NaOH at a temperature of 15 °C and a flow rate of 300 mL min-1.

Given the high number of samples and the complexity of the AS extraction, not all samples could be analyzed in triplicates, but 1/3 of the samples (selected in order to cover the full range of measured C contents) were used to calculate the repeatability of the measure. An analysis of the repeatability between replicates shows deviations

Part 3 Chapter 4 – Description of the study site of ca. 16 , 13 and 22 % between the replicates for GalN, GluN and Mur acid, respectively. For values < 10 μg g-1 soil, deviations were higher (ca. 30-40 %). Hence, we limit our analysis of this data to observations that exceed the latter mentioned uncertainty and to a description of trends within the dataset.

4.4.5.2 PLFA extraction and quantification

At the end of the incubations (refer to next section), soil samples were freeze dried, and kept frozen until being analyzed for PLFA in triplicate. The objective was to describe living soil microbial biomass responsible of the soil respiration measured. The extraction, quantification and compound-specific analysis of phospholipid fatty acids (PLFAs) were performed following the method described by Denef et al. (2007). Total lipids were extracted from 6 to 10 g of freeze- dried soil (depending on the C content) using phosphate buffer/ chloroform/methanol at a 0.9:1:2 ratio. Total lipids, retrieved in the chloroform phase, were partitioned on silica gel columns by sequential elution with chloroform (neutral lipids), acetone (glycolipids), and methanol (phospholipids). The polar lipid fraction, eluting with methanol, was then subjected to mild alkaline transesterification (using methanolic KOH) to form fatty acid methyl esters (FAMEs) which were subsequently analyzed by capillary gas chromatography- combustion-isotope ratio mass spectrometry (GC-C-IRMS) (GC-C/TC Delta PLUS XP Thermo Scientific) via a GC/C III interface.

In total, 28 PLFA peaks were detected, and their sum used as the total PLFA concentration. 22 specific PLFA peaks were used as biomarker for different microbial communities. The PLFAs i-C15:0, a- C15:0, i-C16:0, a-C16:0, i-C17:0, a-C17:0 were designated as gram- positive bacterial whereas cy-C16:1u7, cy-C17:0, cy-C18:1u11, cy- C18:1u7, cy-C19:0 were designated as gram-negative bacterial. The PLFAs C13:0, C14:0, C15:0, C17:0, C18:0, C20:0 were designated as non-specific bacteria, cy-C18:2u6, cy-C18:1u9, were used as indicators of saprophytic fungi. The PLFAs 10Me-C16:0 and 10Me-C18:0 were used to indicate soil actinomycetes (Zelles, 1997, Frostegård and Bååth, 1996, Denef et al., 2009). The universal PLFA C16:0, occurring

Part 3 Chapter 4 – Description of the study site in the membranes of all organisms, was generally the most abundant PLFA. Standard fatty acid nomenclature was used to describe PLFAs. The number before the colon refers to the total number of C atoms; the numbers following the colon refer to the number of double bonds and their location (after the ‘u’) in the fatty acid molecule. The prefixes ‘‘Me,’’ ‘‘cy,’’ ‘‘i,’’ and ‘‘a’’ refer to the methyl group, cyclopropane groups, and iso- and anteiso-branched fatty acids, respectively. For each sample, the abundance of individual PLFAs was calculated in absolute amounts of C (nmol PLFA-C g-1 soil).

4.4.6 Soil respiration measurements

In order to compare bulk OM “degradability” (susceptibility to decomposition), and the quality of protection mechanisms between soil horizons, an incubation experiment was carried out. The objective is to measure microbial activity and OM degradation based heterotrophic CO2 release. The bulk soil samples for each horizons were first moistened at field capacity with deionized water in a pressure pan (pF 2.4, 0.25 bar), and equilibrated during 15 days. Afterwards, an amount equivalent to 40 g of dry soil was transferred in hermetic incubation flasks, containing a 30-ml vial filled with 25 ml of 0.5 M NaOH solution. The gaseous CO2 released from respiration reacts with NaOH according to the following reaction:

2 NaOH + CO2  Na2CO3 + H2O

The CO2 trapped in the NaOH solution was determined by measuring the changes in electrical conductivity of the NaOH solution (conductivity method of Rodella and Saboya (1999)). The electrical conductivity of the NaOH solution inside incubation chamber (cx), of the standard 0.497M NaOH solution (c1) and of the standard 0.248 M Na2CO3 solution (c2) were used to estimate the mass of absorbed CO2 by the expression:

c1 − cx mg CO2 = 22 × [ ] × V × C c1 − c2

Part 3 Chapter 4 – Description of the study site where V is the volume (ml) of the standard NaOH solution and C is its concentration expressed in mol.l−1 (M). The incubations were conducted in triplicates per soil horizon, and lasted 144 days in a temperature-controlled dark room at 20 °C. Incubation flasks without soil samples served as blanks. The conductivity within the NaOH vials was determined every 3–7 days. Soil-derived CO2 was calculated by correcting the measured values in the soil-containing flasks with values in the blanks (due to atmospheric CO2). Oxygen was regularly supplied in the incubation flasks, by opening them at each measurement time step. The soil samples were kept moisten at field capacity by regularly controlling the weight of the incubation flasks, and adding deionized water to compensate the weight loss.

4.4.7 Fe and Si isotopes

The methods for Si and Fe isotope analysis are presented in Appendix A2.1 and A2.2, respectively.

4.5 Pedological context

According to the IUSS Working Group WRB system (World Reference Base for Soil Resources, 2015) the soils are classified as Dystric Cambisol at the youngest sites (P1-120 and P2-175 yrs), Albic Podzol at the intermediate site (P3-270 yrs), and Placic Podzol at the oldest sites (P4-335 and P5-530 yrs).

The Placic Podzols are characterized by the following sequence of soil horizons from surface to depth (Cornelis et al., 2014): an eluvial albic E horizon, strongly weathered; an illuvial spodic Bh horizon (enriched in OM); a Bhs horizon, enriched in Fe oxyhydroxides and OM; a Bs horizon, enriched in poorly crystalline aluminosilicates and Fe oxyhydroxides; a Bw horizon characterized by its color and structure but exhibiting no illuvial accumulation; and a poorly structured BC horizon (Figure 4.2).

Part 3 Chapter 4 – Description of the study site

The major soil properties are presented in Tables 4.2 and 4.3. The soils are very sandy, with more than 80% of sand (99.9 % of sand in the beach C parent material). The content of weatherable minerals is very high in the parent material, as estimated from their total reserve in bases (TRB = sum of total Ca, Mg, K and Na contents, -1 expressed in cmolc kg . 534.67 in the beach sand), and the pH is close to neutrality (7.7). With time and pedogenesis, a progressive evolution of the pH, minerals and SOM is observed. It will be detailed in the next chapters.

Table 4.2 - Major soil characteristics: WRB classification, pH in water, cationic exchange capacity (CEC), exchangeable cations, base saturation (%BS), particle-size distribution and total organic carbon (TOC) content

Exchangeable cations Particle size Profile Age Horizon Depth WRB* pH CEC -1 %BS (cmolc kg ) distribution (%) TOC 2+ + 2+ + cm (water) cmolc/kg Ca K Mg Na Sand Silt Clay g/kg P.M.** 0 C 7.7 1.55 0.19 0.23 0.33 0.73 95.57 99.9 0.1 0.0 0.32 P1 120 BC1 0-35 DC 5.9 3.72 0.25 0.10 0.35 0.18 23.76 99.2 0.6 0.3 9.12 P1 120 BC2 35-60 DC 5.9 3.32 0.23 0.06 0.33 0.18 24.21 7.79 P2 175 BW 3-44 DC 5.8 4.34 0.23 0.07 0.35 0.25 20.66 99.0 0.6 0.4 3.33 P2 175 BC 44-75 DC 5.9 2.38 0.12 0.08 0.18 0.22 25.30 99.6 0.2 0.1 2.55 P3 270 E 0-7 AP 4.6 6.04 0.48 0.05 0.56 0.11 19.94 90.8 6.1 3.1 12.63 P3 270 Bh 7-23 AP 5.1 5.94 0.59 0.04 0.41 0.13 19.76 97.2 1.7 1.1 16.17 P3 270 BW 23-57 AP 5.3 5.57 0.52 0.04 0.40 0.11 19.14 97.4 1.8 0.9 8.18 P3 270 BC > 57 AP 5.4 5.71 0.39 0.10 0.22 0.11 14.53 98.2 1.1 0.7 6.87 P4 330 E 0-10 PP 4.9 5.66 0.07 0.02 0.05 0.04 3.22 82.3 14.4 3.3 8.48 P4 330 Bh 10-17 PP 5.5 15.72 0.22 0.04 0.10 0.06 2.67 88.0 8.7 3.2 16.18 P4 330 Bhs 17-17.5 PP 15.18 0.13 0.04 0.07 0.06 1.96 90.0 6.8 3.2 22.99 P4 330 Bs 17.5-23 PP 5.3 5.88 0.04 0.03 0.01 0.03 1.80 94.9 4.2 0.9 4.53 P4 330 Bw 23-63 PP 5.4 4.31 0.03 0.03 0.01 0.04 2.58 96.1 2.4 1.5 2.67 P4 330 BC1 63-113 PP 5.3 3.15 0.02 0.06 0.00 0.02 3.21 1.26 P4 330 BC2 113-193 PP 5.3 2.72 0.02 0.01 0.00 0.02 2.33 1.48 P4 330 BC3 > 193 PP 5.2 2.71 0.02 0.04 0.00 0.02 3.14 98.2 1.8 0.0 1.03 P5 530 E 0-8 PP 4.5 5.31 0.22 0.04 0.15 0.06 8.63 14.82 P5 530 Bh 8-9.5 PP 4.5 15.32 0.33 0.03 0.16 0.06 3.78 31.18 P5 530 Bhs 9.5-10 PP 4.5 23.35 0.24 0.08 0.13 0.07 2.23 40.48 P5 530 Bs 10-15 PP 4.8 9.85 0.04 0.04 0.03 0.04 1.60 11.60 P5 530 Bw 15-40 PP 5.0 3.53 0.02 0.01 0.00 0.03 2.02 1.17 P5 530 BC 40-60 PP 5.1 2.45 0.03 0.02 0.00 0.01 2.56 0.64 * WRB classification: DC stands for Dystric Cambisol, AP for Albic Podzol and PP for Placic Podzol **PM: parental material

Table 4.3 – Total elemental contents of the bulk soil, total reserve in bases (TRB) and proportion of ‘free iron’ as measured by the ratio between DCB-extractable iron and total iron (Fed/Fet)

-1 Profile Horizon Depth Total elemental content (g kg ) TRB Fed/Fet

cm Si Al Fe Ca K Mg Na Ba Mn P Sr Zr Ti cmolc/kg P.M. C 285.27 75.87 48.92 55.63 5.93 18.03 21.50 0.29 1.18 0.67 0.44 0.23 7.14 534.67 0.03 P1-120 BC1 0-35 302.71 70.95 34.47 42.72 6.57 13.31 22.94 0.31 0.74 0.55 0.38 0.10 4.37 439.32 0.05 P1-120 BC2 35-60 293.11 73.67 41.94 49.93 6.23 16.01 21.58 0.30 0.96 0.57 0.42 0.13 5.89 490.74 0.05 P2-175 Bw 3-44 309.65 70.01 33.62 39.85 6.89 13.20 24.23 0.32 0.68 0.87 0.38 0.09 4.10 430.49 0.04 P2-175 BC 44-75 328.63 68.07 25.00 31.16 7.98 9.85 25.67 0.36 0.49 0.47 0.35 0.07 2.95 368.66 0.08 P3-270 E 0-7 335.69 62.66 18.99 23.85 7.89 6.99 26.45 0.35 0.37 0.19 0.35 0.08 2.87 311.77 0.13 P3-270 Bh 7-23 324.81 63.76 20.67 22.96 8.30 7.59 26.14 0.39 0.37 0.40 0.33 0.06 2.33 311.96 0.13 P3-270 Bw 23-57 341.35 62.58 18.43 22.14 8.28 6.82 25.69 0.36 0.34 0.34 0.32 0.08 2.20 299.55 0.12 P3-270 BC > 57 323.13 69.07 25.87 29.10 7.68 10.03 26.11 0.35 0.45 0.39 0.34 0.07 3.02 361.04 0.08 P4-330 E 0-10 362.91 54.13 10.81 17.29 8.05 3.48 23.24 0.32 0.27 0.06 0.27 0.20 3.39 236.55 0.10 P4-330 Bh 10-17 317.90 65.90 25.29 24.23 7.88 9.44 23.95 0.36 0.44 0.35 0.30 0.09 3.50 322.84 0.13 P4-330 Bhs 17-17.5 303.58 66.76 37.76 25.69 7.27 10.12 22.91 0.30 0.46 0.34 0.30 0.08 3.67 329.71 0.57 P4-330 Bs 17.5-23 314.20 71.50 31.18 30.53 7.97 11.84 24.85 0.34 0.53 0.37 0.31 0.08 3.59 378.19 0.12 P4-330 Bw 23-63 327.76 68.38 25.27 28.12 7.92 9.80 24.89 0.34 0.49 0.32 0.33 0.10 3.04 349.44 0.10 P4-330 BC1 63-113 330.87 66.73 24.87 28.71 8.11 9.93 24.51 0.34 0.48 0.43 0.32 0.07 3.01 352.30 0.05 P4-330 BC2 113-193 336.78 65.32 22.31 26.90 8.37 8.84 25.08 0.38 0.47 0.37 0.34 0.07 2.59 337.48 0.05 P4-330 BC3 > 193 320.47 70.02 28.95 33.61 7.96 11.44 25.21 0.34 0.59 0.47 0.34 0.07 3.47 391.87 0.04 P5-530 E 0-8 365.59 41.04 10.66 11.85 5.97 2.28 17.63 0.26 0.27 0.11 0.21 0.23 3.11 169.85 0.14 P5-530 Bh 8-9.5 326.16 53.98 19.06 18.48 7.16 4.68 21.20 0.31 0.35 0.19 0.28 0.08 4.11 241.26 0.35 P5-530 Bhs 9.5-10 315.66 58.03 37.94 19.49 7.48 6.44 21.97 0.30 0.30 0.16 0.27 0.10 4.04 264.89 0.60 P5-530 Bs 10-15 321.61 66.62 29.29 21.90 7.55 8.65 22.50 0.33 0.42 0.18 0.29 0.07 3.00 297.65 0.33 P5-530 Bw 15-40 327.80 68.02 26.79 27.30 8.05 10.44 25.15 0.35 0.51 0.25 0.32 0.07 3.21 352.22 0.07 P5-530 BC 40-60 329.07 66.98 26.28 28.45 7.99 10.44 25.02 0.34 0.51 0.26 0.32 0.07 3.09 357.10 0.05

Part 3 Chapter 5 – Evolution of the soil environment

Chapter 5 - Evolution of the soil environment

In the present chapter, we will describe the soil environment of the Vancouver chronosequence samples (the pore space).

5.1 Field observations

During sampling, a progressive cementation/induration of the illuvial B horizons was observed, particularly in P4 and P5 Bhs, due to the accumulation of Fe and OM. Because of the high mean annual precipitation, episodic waterlogging alternating with a rapid return to oxic conditions occurs in these horizons. A picture of the P4 and P5 profiles is presented in the Figure 5.1.

Part 3 Chapter 5 – Evolution of the soil environment

A. B.

Figure 5.1 – Picture of the P4 (A.) and P5 profile (B.)

5.2 Micromorphological observations

The micromorphological observations (thin sections, from kubiena boxes) give additional information on the in-situ organization of OM within the horizons (Figure 5.2).

In P1, the BC1 horizon presents an aerated sandy structure, with the presence of recognizable roots (red arrow), and fecal pellets (green arrow). Black C (pyrogenic material) is also observed.

The P3 Bh horizon presents an aerated sandy structure, but the micromorphology of the OM is different and corresponds to one of the two main SOM types distinguished in podzol-B horizons: polymorphic organic matter. Polymorphic OM presents the aspect of small unconsolidated OM clusters, and has a friable aspect. It is frequently associated with decaying roots and organic debris by mesofauna, and considered to have coprolithic pellets as basic constituents (Buurman and Jongmans, 2005).

In P4 and P5, the micromorphology of OM is very different from the three younger profiles, and corresponds to the second main SOM type present in podzols: monomorphic organic matter coatings. The latter are usually continuous, brown to dark brown, homogeneous or

100

Part 3 Chapter 5 – Evolution of the soil environment layered, and frequently show cracks. These cracks are interpreted as desiccation features of strongly hydrated OM gels (Buurman and Jongmans, 2005). This coating-like structure is observed in the P4 and P5 Bh, Bhs and Bs horizons, and completely fills the pore system in the Bhs horizon (with desiccation cracks observable). The Bhs horizon of P4 and P5 profiles is cemented, and indurated as observed during the sampling campaign. Monomorphic coatings are most commonly found in cemented horizons and it is an evidence of periodic water stagnation, alternating phases of desiccation and saturation (poorly drained podzols) (Legros, 2007, Buurman and Jongmans, 2005). Monomorphic coatings are more typical of non-boreal podzols, and rarely observed in boreal podzols (Buurman and Jongmans, 2005).

Implications for SOM protection mechanisms:

The micromorphological data reveal a very different environment for OM degradation and microorganisms in the chronosequence. In P1, P2 and P3 pedons, as well as in P4 and P5 E horizons, the structure is open, as it exhibits clear pore space. However, the micromorphological features of Bh, Bhs and Bs in P4 and P5 reveal the cemented nature of the illuvial horizon, as well as evidences periodic water saturation. These properties make these horizons periodically anoxic and not accessible to roots. Such properties can impact OM dynamics and further degradation.

101

P1: 0-4 cm – BC1 P3: 10 - 20 cm – Bh P4: 13-22 cm (Bh-Bhs-Bs) P5: 4-13 cm (Bh-Bhs-Bs)

2.6 mm

Figure 5.2 - Micromorphological observations from thin sections of kubiena boxes taken in the P1 BC1, P3 Bh, and P4 and P5 Bh, Bhs and Bs horizons.

102

Part 3 Chapter 6 – Mineralogical evolutions

Chapter 6 - Mineralogical evolutions

This chapter is devoted to the characterization of mineral transformations in the chronosequence, leading to the formation of secondary minerals involved in OMA.

The objectives are (1) to insure that the Cox Bay soils satisfy the conditions to be considered as a chronosequence; (2) to characterize the minerals present in the parent material; (3) to describe the mineralogical evolutions/transformations with time (weathering and secondary mineral phases precipitation), as well as the main pedogenetic processes involved.

103

Part 3 Chapter 6 – Mineralogical evolutions

To do so, we decided to use classical soil characterizations (total elemental content, selective extractions, XRD, etc.) as well as to follow an exploratory approach, using three promising tracors of pedogenic processes: Rare earth elements (REE), Si isotopes and Fe isotope. These tracers have never been applied in a chronosequence of podzols. We selected REE because they are present in most primary minerals, and are known to be scavenged by SOM, Fe and Mn oxides. Fe isotopes were selected because of the importance of the Fe phases and redox processes in podzolic soils, and Si isotopes to gaign a better insight into the origin of Si in pedogenic clay minerals.

This chapter is organized as follow. First, the secondary mineral phases are described in section 6.1. Then, the section 6.2 presents the REE results. Finally, a synthesis of the informations acquired on the Vancouver chronosequence mineralogy (including Si and Fe isotopes results, shown in Appendix 2.1 and 2.2), is presented in the section 6.3.

6.1 Description of the secondary mineral phases

The contents of Al and Fe secondary phases are determined by using selective extractions: DCB (d), oxalate (o), pyrophosphate (p). Here, we restrict our discussion to the main phases since the interpretation of the results of selective extractions is somehow tricky. As inferred from figure 6.1, little contents of secondary Fe and Al mineral phases are identified in P1, P2 and P3 (Fed and Alo < 2.6 and 1.4 g.kg-1, respectively). They are much larger in P4, with a maximum -1 free Fe content in the Bhs horizon (Fed = 21.6 g.kg ), and Alo content similar in Bh, Bhs and Bs horizons (8.8, 9.2 and 9.1 g.kg-1, respectively) (Figure 6.1). In P5, the content of extractable Fe and Al increases -1 compared to P4. Fed content amounts to 26.1 g.kg in the Bhs horizon, -1 and Alo content amounts to 12.6 g.kg in the Bs horizon.

104 b.

a. P1 – 120 yrs P2 – 175 yrs P3 – 270 yrs P4 – 330 yrs P5 – 530 yrs 0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30

E E E BC1 Bh Bh Bh Bw Bhs Bhs

Bs Bs Alo Bw Sio Bw Bw BC BC2 Fed BC BC BC

b. 0 200 400 600 0 200 400 600 0 200 400 600 0 200 400 600 0 200 400 600

E E TRB E BC1 Bh Bh Bw Bh Bhs Bhs

Bw Bs Bs Bw Bw pH BC2 BC BC BC BC 4 5 6 7 4 5 6 7 4 5 6 7 4 5 6 7 4 5 6 7

Figure 6.1 – In the 5 profiles of increasing age (P1 to P5): (a.) Evolution of the secondary mineral phases, quantified by -1 the oxalate-extractable Al and Si (Alo and Sio) and DCB-extractable Fe (Fed) [g kg ]; (b.) Evolution of weathering, + + 2+ 2+ −1 estimated by the total reserve in bases (TRB = sum of the total contents of Na , K , Ca , Mg ; in cmolc·kg ) and evolution of the pH. 105

Part 3 Chapter 6 – Mineralogical evolutions

The nature of the secondary phases in P4 and P5 profiles differ between the Bh, Bhs and Bs horizons (Figure 6.2).

The main Al secondary phases in P4 and P5 Bh and Bhs are organo-Al complexes (Alp). Indeed, Alp represents 68, 73, 93 and 68 % of the total extractable Al (Alo), in P4 Bh, Bhs and P5 Bh, Bhs, respectively. The main Al secondary phase in the Bs horizons are poorly crystalline aluminosilicates, i.e. paracrystalline/short range ordered imogolite and/or its less crystalline form proto-imogolite allophane (both referred here under as “imogolite type material” - ITM). The presence of ITM can be confirmed by the combination of several indications: (1) the ratio Alo-Alp/Sio is ~ 2 (2.1 in P4 Bs and 2.2 in P5 Bs); (2) pH > ~ 4.7 - 5, as ITM are very unstable at lower pH (Gustafsson et al., 1995, Lundström et al., 1995, Mossin et al., 2002) (pH 5.3 in P4 Bs and 4.8 in P5 Bs); (3) the proportion of metals complexed with OM is low (Alp represents 19% of Alo, in both P4 and P5 Bs), as the presence of OM suppresses the formation of imogolite and proto-imogolite (Mossin et al 2002). The occurrence of ITM as the major secondary Al precipitate in the Bs horizon of podzols have been widely documented (Parfitt, 2009, Mossin et al., 2002, Gustafsson et al., 1995). Among aluminosilicates, ITM are the most reactive towards dissolution/precipitation/adsorption reactions because of their very high SSA (Gustafsson et al., 1999). The repartition of the Al phases differ between P4 and P5. The total Al phases content is lower in P5 Bh and Bhs comparatively to the same horizons in P4, but increased in P5 Bs compared to P4 Bs (Figure 6.2).

The main Fe secondary phases in the P4 and P5 illuvial horizons are (1) organo-Fe complexes: Fep, maximum in P4 Bhs and P5 -1 Bhs (10.0 and 12.4 g.kg , 46 and 48 % of the Fed, respectively); (2) poorly crystalline Fe oxides: Feo-Fep, the dominant Fe phase in P4 Bs -1 -1 (3.2 g.kg , 87% of the Fed) and P5 Bhs and Bs (13.9 and 6.1 g.kg , 53 and 63% of the Fed, respectively), and (3) crystalline Fe oxihydroxides: Fed-Feo, highest proportion (47% of the Fed) and total content (10.0 g.kg-1) in P4 Bhs horizon. Ferrihydrite (a poorly cristalline Fe oxyhydroxide) and goethite (crystalline Fe oxyhydroxide) are known to be the most important Fe mineral phases in podzolized soils (Cornell

106

Part 3 Chapter 6 – Mineralogical evolutions

and Schwertmann, 2003, Eusterhues et al., 2003, Eusterhues et al., 2005).

-1 Secondary phases [g.kg soil] 0 10 20 30 0 10 20 30

P4 Bh P5 Bh

P4 BhS P5 Bhs

P4 Bs P5 Bs

Figure 6.2 – Fe, Al and Si fractions in the P4 and P5 illuvial horizons (Bh, Bhs, Bs). Total free Fe (blue bar) is the sum of Fe in organo-metal complexes (Fep), Fe in poorly crystalline Fe oxyhydroxides (Feo-p) and Fe in crystalline oxyhydroxides (Fed-o). Total free Al (red and yellow bar) is the sum of Al in organo-metal complexes (Alp), and Al in poorly crystalline phases (Alo-p). Sio is the Si present in poorly crystalline aluminosilicates.

107

Part 3 Chapter 6 – Mineralogical evolutions

6.2 Rare Earth Elements dynamics along pedogenesis in a chronosequence of podzolic soils1

Abstract:

Rare earth elements (REE) total concentration and signature in soils are known to be impacted by successive soil-forming processes. So it can be used as probe of soil processes. However, few studies focus on their behavior in Podzols. Podzols result from the combination of two main pedogenic processes: (1) the strong weathering in the surface eluvial horizon; (2) the downward transfer of dissolved organic matter (OM) and mobile Al and Fe, and their accumulation in the illuvial horizon beneath. Iron oxides and OM are known to have strong affinities with REE, and to play an important role in transfer and immobilization of REE. In order to decipher the relative importance of Fe oxide and OM in REE fate during podzolization, and to investigate whether REE can trace Podzol formation, we study here the evolution of REE signatures along five pedons, aged from 120 to 530 years, in a Cambisol-Podzol chronosequence located in the Cox Bay of Vancouver Island. Our results show that the REE content is strongly correlated to the general loss of elements and mineral weathering. Furthermore, the accumulation of secondary OM, Al and Fe-bearing phases does not impact the REE signature of the bulk soil. Both our results and the ones available in the literature indicate that the release of REE induced by weathering and subsequent leaching in percolating water are the main pathways determining the REE fate in Podzols. Furthermore, we show that REE can be released and mobilized in very short periods of time during podzolization (330 yrs).

1 Adapted from Vermeire M.L., Cornu, S., Fekiacova, Z., Detienne, M., Delvaux, B., Cornélis, J.T., 2016, Rare Earth Elements dynamics along pedogenesis in a chronosequence of podzolic soils, Chemical Geology 446: 163-174 108

Part 3 Chapter 6 – Mineralogical evolutions

6.2.1 Introduction

The group of rare earth elements (REE) consists of 17 elements; the lanthanides and lanthanum (La) together with scandium (Sc) and yttrium (Y) (Saatz et al., 2015). They form a series that behaves geochemically coherently owing to the small but steady decrease in ionic radius with increase in atomic number (Henderson, 1984, Panahi et al., 2000, Yusoff et al., 2013). They are considered as promising tracers in pedogenesis (Taunton et al., 2000, Aubert et al., 2001, Aide and Smith-Aide, 2003, Chabaux et al., 2003, Compton et al., 2003, Ndjigui et al., 2008, Laveuf and Cornu, 2009, Harlavan et al., 2009, Ma et al., 2011, Gong et al., 2011, Yusoff et al., 2013). The origin of REE in soils lies in the parent material (PM), since anthropogenic sources are restricted (Hu et al., 2006). During pedogenesis, REE signature is affected by a variety of processes (dissolution, oxydo-reduction, precipitation and complexation). These processes induce internal fractionations and/or anomalies related to REE mass or different oxidation states for Ce and Eu. Consequently, REE concentrations normalized to a reference PM and fractionation pattern observed in a soil profile provide a useful tool for elucidating soil-forming processes leading to the formation of a specific soil horizon (Yusoff et al., 2013).

Iron- and Mn- oxides are known to scavenge REE (Rankin and Childs, 1976, Palumbo et al., 2001) through one or a combination of the following mechanisms: coprecipitation, adsorption, surface complex formation, ion exchange and penetration of the lattice (Chao, 1976, Cao et al., 2001), in amounts varying with soil type (Li et al., 2006, Wang et al., 2001, Zhang and Shan, 2001) and depth (Land et al., 1999, Yan et al., 1999). Soil organic matter (SOM) is more efficient than Fe oxides in concentrating REE, given the strong complexing ability of organic molecules (Davranche et al., 2011). Therefore, SOM plays an important role in the transfer and immobilization of REE, controlling inter-horizon REE distribution (Koeppenkastrop and De Carlo, 1992, Tang and Johannesson, 2003, Pourret et al., 2007a, Goyne et al., 2010, Davranche et al., 2011, Aide and Aide, 2012). The differential binding affinity of the REE for SOM across the REE series is still poorly understood. Two general tendencies can be observed in natural

109

Part 3 Chapter 6 – Mineralogical evolutions waters (Tang and Johannesson, 2010): trend M and trend H. The trend M is observed in the liquid phase if the stability constants of REE- Humic substances complexes are greatest for the middle REE (MREE), followed by the heavy REE (HREE), and light REE (LREE). The M trend gives a MREE enrichment signature in the liquid phase (Johannesson et al., 2004, Yamamoto et al., 2005, Pourret et al., 2007b, Pedrot et al., 2008, Tang and Johannesson, 2010, Davranche et al., 2011, Cidu et al., 2013). The trend H is observed if the stability constants of REE-Humic substances increase with increasing atomic number across the REE series. They are thus largest for the HREE; this is called “lanthanide contraction effect” (Sonke and Salters, 2006, Sonke, 2006, Stern et al., 2007, Laveuf and Cornu, 2009, Gangloff et al., 2014, Vázquez-Ortega et al., 2015).

Podzolization combines two main processes: (1) strong mineral weathering in the eluvial surface E horizon, eventually depleted in elements and enriched in resistant minerals as quartz and resistant accessory minerals, and characterized by a light-grey color; (2) eluviation with percolating water of dissolved organic matter (DOM) complexed with Al and Fe which will precipitate and accumulate in soil horizons beneath to form dark reddish/brownish colored illuvial horizons (Bh, Bhs or Bs) (Lundström et al., 2000a). Among the mechanisms that have been proposed to explain mobilization and translocation phenomenon involved in podzolization, the formation and downward transport of unsaturated complexes of organic acids with Al and Fe (the fulvate theory, McKeague et al. (1978)) has been generally accepted as the dominant mechanism of eluviation (Lundström et al., 2000a). Dissolved organic acids are crucial components in Podzol development, both because they demonstrate a large ability to promote dissolution of minerals, and they form complexes with Al and Fe, that will transport those elements deeper in the soil profile (van Hees et al., 2000, Lundström et al., 2000a, Kaiser and Kalbitz, 2012, Gangloff et al., 2014). These reactions can be fast (~100 years) depending on environmental conditions (Sauer et al., 2008, Cornu et al., 2008). As metal cations and DOM play a key role in podzolization process, it is reasonable to hypothesize that the REE signature of the soil will be impacted by a podzolic development.

110

Part 3 Chapter 6 – Mineralogical evolutions

In order to understand the REE’s temporal dynamics along with Fe and OM fate in Podzol, and to assess the ability of REE to trace this pedogenic processes, we study the REE distribution in a podzolic soil chronosequence. The soil members of the chronosequence were already characterized. Besides, they exhibit successive processes in terms of dissolution, synthesis and transfer of clay minerals (Cornelis et al., 2014) as well as a strong redistribution in Fe and OM. A chronosequence involves a sequence of soils developed in the same conditions of soil forming factors except time (Huggett, 1998, Walker et al., 2010). They are valuable tools for investigating temporal dynamics of pedogenic processes.

6.2.2 Material and methods

Refer to chapter 4 for the description of the study site and of methods.

6.2.3 Results and discussion

6.2.3.1 REE pattern normalized to the UCC as tracers of soil parent material

Figure 6.3 shows the REE signature of the beach sand (C) and of the deepest BC horizons of the P1-120, P2-175, P3-270, P4-330 and P5-530 yrs profiles normalized to the UCC, while REE concentrations are reported in Table 6.1. All the considered samples present the same REE pattern depleted in LREE with a strong positive Eu anomaly (Figure 6.3), confirming that the sediment material at the origin of theses horizons comes from the same source. The strong positive Eu anomaly may be related to the large content in plagioclase of the sediment (Figure 6.4), since feldspars are known to be enriched in Eu (Vázquez- Ortega et al., 2015).

The C horizon exhibits the highest total content in REE (107.15 µg REE g-1 soil). This content decreases in the P1-120 BC horizons (82.52 µg REE g-1 soil) and in the deep BC horizons of the P2-175, P3-

111

Part 3 Chapter 6 – Mineralogical evolutions

270, P4-330 and P5-530. These four last horizons show similar total REE concentrations (48.89, 54.03, 57.37 and 51.87 µg REE g-1 soil, respectively). There is no significant evolution of the Eu anomalies, varying from 1.45 to 1.63 when normalized to the UCC.

2,0 1,8 1,6 1,4 1,2 1,0 0,8

0,6 REE normalized to UCC to normalized REE 0,4 0,2 0,0 La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

C P1 BC P2 BC P3 BC P4 BC P5 BC

Figure 6.3 - REE patterns of the C beach parental material, and P1-120, P2-175, P3-270, P4-330 and P5-530 BC normalized to the UCC. Error bars correspond to the result of error propagation calculations, based on the RSD given by the machine for each measure

112

70 70 Quartz b 65 a 65 feldspars Illite (micas) 60 60 Hornblende 55 55 Augite Chlorite 50 50 45 45 40 40 35 35 30 30 25 25

20 20 Abundance Abundance (%) 15 15 10 10 5 5 0 0 C P1 BC P2 BC P3 BC P4 BC P5 BC P3 E P4 E P5 E Figure 6.4 - Bulk mineral composition in (a) the deepest horizons (BC) and the parent material (C), and (b) the surface horizons (E) of the studied podzol profiles. Error bars correspond to the total error associated with each determination. It is calculated by multiplying the error given by SIROQUANT for each individual mineral (the e.s.d. values) by the square root of the global χ² for the analysis in question.

113

Table 6.1 - REE contents in the studied podzol samples and in the Upper Continental Crust (UCC)

Profile Horizon Depth LREE (µg g-1) MREE (µg g-1) HREE (µg g-1) cm La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu P.M. C 20.72 37.94 5.09 21.07 4.69 1.38 4.32 0.72 4.52 0.90 2.68 0.37 2.41 0.35 P1 BC 0-60 14.69 29.80 3.87 16.28 3.70 1.16 3.50 0.57 3.59 0.72 2.15 0.30 1.92 0.27 P2 Bw 3-44 10.45 21.10 2.79 11.93 2.90 0.97 2.89 0.48 3.01 0.60 1.83 0.25 1.75 0.23 P2 BC 44-75 8.54 16.99 2.22 9.49 2.19 0.76 2.26 0.37 2.38 0.47 1.49 0.20 1.32 0.20 P3 E 0-7 7.67 14.25 1.92 7.84 1.77 0.64 1.71 0.26 1.83 0.37 1.16 0.17 1.09 0.16 P3 Bh 7-23 7.18 13.60 1.84 7.47 1.79 0.65 1.69 0.30 1.85 0.38 1.08 0.15 1.01 0.16 P3 Bw 23-57 6.88 13.23 1.77 7.08 1.63 0.61 1.63 0.27 1.68 0.35 1.06 0.16 1.07 0.16 P3 BC > 57 9.76 18.55 2.52 10.97 2.40 0.82 2.37 0.40 2.48 0.50 1.51 0.21 1.32 0.22 P4 E 0-10 7.34 13.33 1.80 7.45 1.60 0.52 1.35 0.23 1.51 0.30 0.92 0.14 0.96 0.13 P4 Bh 10-17 10.13 17.84 2.45 10.07 2.30 0.77 2.10 0.37 2.42 0.47 1.46 0.20 1.28 0.19 P4 Bhs 17-17.5 11.37 20.30 2.90 11.84 2.72 0.84 2.42 0.44 2.69 0.55 1.61 0.22 1.44 0.22 P4 Bs 17.5-23 10.27 19.32 2.74 11.83 2.77 0.91 2.62 0.47 3.09 0.61 1.82 0.25 1.67 0.24 P4 Bw 23-63 9.78 19.71 2.67 11.48 2.70 0.89 2.64 0.45 2.81 0.56 1.67 0.23 1.53 0.24 P4 BC1 63-113 10.76 19.02 2.72 11.37 2.72 0.85 2.63 0.42 2.68 0.54 1.69 0.22 1.52 0.23 P5 E 0-8 7.33 12.18 1.54 6.10 1.22 0.38 0.96 0.18 1.09 0.23 0.69 0.11 0.73 0.13 P5 Bh 8-9.5 8.28 14.42 1.99 7.93 1.71 0.53 1.40 0.24 1.60 0.34 1.02 0.14 1.06 0.16 P5 Bhs 9.5-10 8.67 14.68 1.95 8.29 1.63 0.58 1.49 0.27 1.68 0.35 1.09 0.16 1.09 0.16 P5 Bs 10-15 8.65 15.66 2.20 9.64 2.25 0.74 2.06 0.37 2.37 0.47 1.39 0.21 1.32 0.20 P5 Bw 15-40 10.33 17.41 2.43 10.26 2.51 0.74 2.17 0.40 2.57 0.51 1.52 0.22 1.42 0.21 P5 BC 40-60 9.67 16.72 2.41 10.42 2.57 0.81 2.14 0.39 2.64 0.52 1.65 0.21 1.48 0.22 UCC* 30.00 64.00 7.10 26.00 4.50 0.88 3.80 0.64 3.50 0.80 2.30 0.33 2.20 0.32 * Values for the UCC comes from Laveuf and Cornu (2009)

114

Part 3 Chapter 6 – Mineralogical evolutions

We observe a decrease in REE concentrations from the beach sand to the BC horizon of the P1 and a further decrease in the P2-BC horizons of the other soil profiles. Such decrease may be due to mineral weathering. No further REE decrease is recorded in the BC horizons deeper than 50 cm for older soil profiles (from P2 to P5). In addition, concentrations of Zr and Ti, two poorly mobile elements that are expected to accumulate with an increasing weathering stage, decrease from the sand beach to P1 and P2 BC horizons (0.23, 0.13 and 0.07 g kg-1, respectively, for Zr, and 7.14, 5.89 and 2.95 g kg-1, respectively, for Ti), and then remain constant from P2 BC to P5 BC (Table 4.3). Furthermore, the surface P2 Bw horizon displays higher total concentrations in major elements, Zr, Ti and REE than the deeper P2 BC horizon (Table 4.3, 6.1 and 6.2). All these observations do not support the hypothesis of an early weathering as explanation for the observed decrease of REE signature in the deep horizons from sand beach to BC horizon of the P2. It suggests that the decrease in REE from the sand beach to the BC of the P2 is most probably due to a change in the sedimentation dynamic leading to a variable content of sand, silt and clay in the C beach and BC horizons. This is consistent with the significant increase in quartz content between C and P1 BC (Figure 6.4). Quartz is known to be REE-free and thus acts as a diluting agent of these elements (Compton et al., 2003).

The three podzols, P3-270, P4-330 and P5-530, have similar REE patterns and Zr and Ti concentrations in their BC horizons. They may thus be considered as developed from the same parental BC material. This is confirmed by the similar mineralogical composition of the three BC horizons (Figure 6.4).

115

Table 6.2 – Total REE content (Σ REE), Ce (Ce/Ce*) and Eu (Eu/Eu*) anomalies calculated according to both the Upper Continental Crust (UCC – from Laveuf and Cornu, 2009) and the average of the studied BC horizons (MBC), and REE fractionation ratios (La/Gd, Gd/Lu, La/Lu) normalized to MBC

Profile Horizon Depth cm Σ REE Ce/Ce* Eu/Eu* Ce/Ce* Eu/Eu* La/Gd Gd/Lu La/Lu µg g-1 UCC UCC MBC MBC MBC MBC MBC P.M. C 107.15 0.84 1.45 1.03 0.92 1.14 1.14 1.30 P1 BC 0-60 82.52 0.90 1.51 1.11 0.96 0.99 1.19 1.18 P2 Bw 3-44 61.16 0.89 1.57 1.09 0.99 0.86 1.18 1.01 P2 BC 44-75 48.89 0.89 1.61 1.09 1.02 0.89 1.05 0.94 P3 E 0-7 40.85 0.85 1.74 1.04 1.10 1.06 0.99 1.05 P3 Bh 7-23 39.15 0.85 1.75 1.05 1.11 1.01 1.00 1.01 P3 Bw 23-57 37.57 0.86 1.74 1.06 1.10 1.00 0.94 0.93 P3 BC > 57 54.03 0.85 1.62 1.05 1.03 0.97 1.01 0.99 P4 E 0-10 37.58 0.84 1.65 1.03 1.04 1.29 0.94 1.21 P4 Bh 10-17 52.06 0.82 1.65 1.00 1.04 1.14 1.03 1.18 P4 Bhs 17-17.5 59.57 0.81 1.53 0.99 0.97 1.11 1.03 1.14 P4 Bs 17.5-23 58.60 0.83 1.58 1.02 1.00 0.93 1.03 0.95 P4 Bw 23-63 57.36 0.88 1.57 1.08 0.99 0.88 1.04 0.91 P4 BC1 63-113 57.37 0.80 1.50 0.98 0.95 0.97 1.07 1.03 P5 E 0-8 32.86 0.83 1.67 1.02 1.06 1.80 0.70 1.27 P5 Bh 8-9.5 40.84 0.81 1.62 0.99 1.03 1.40 0.80 1.12 P5 Bhs 9.5-10 42.09 0.81 1.76 1.00 1.11 1.37 0.84 1.16 P5 Bs 10-15 47.54 0.82 1.62 1.00 1.03 0.99 0.97 0.96 P5 Bw 15-40 52.69 0.79 1.48 0.97 0.94 1.13 0.96 1.08 P5 BC 40-60 51.87 0.79 1.63 0.97 1.03 1.07 0.91 0.97

116

Part 3 Chapter 6 – Mineralogical evolutions

In order to identify the impact of podzolization on REE signature, we normalized the REE concentrations of the pedological features to the REE concentrations of the parent material (Braun et al., 1998, Aubert et al., 2001, Dequincey et al., 2006, Laveuf et al., 2008). Therefore, in the following, the mean composition of the 270-330-530 yrs BC horizons (mean BC, MBC) will be used as reference value for the parental material of the three podzols, and the REE composition of the three podzolic profiles horizons will be normalized to this value.

6.2.3.2 REE evolution with depth in podzolic soil chronosequence as a tracer of pedogenesis

Within the podzolic profiles P3-270, P4-330 and P5-530 yrs, the REE contents of the upper E-horizons are lower than the ones of the BC horizons (a loss of total REE content of 24, 34 and 37%, respectively, in the E horizon compared to the deep BC horizon, Figure 6.5 a, e, i, Figure 6.6, and Table 6.2). The E horizon is the most depleted in the P5 profile. In P3-270 yrs, the REE content already decreases in the Bw horizons compared to the BC and then remain stable up to the upper E horizon (Table 6.1), while in P4, REE concentrations start to decrease from the Bh horizon and from the Bhs in P5. Significant positive correlation was observed between the evolution with depth of REE patterns and of TRB (r = 0.896) while a negative correlation was observed with the Si/Al ratio (r = -0.866) (Table 4.3, Figure 6.5 b, f, j). The TRB is classically used as a weathering index in soils (Herbillon, 1986) and Si/Al ratio can be considered as a proxy for quartz accumulation at the surface since quartz is a non weatherable mineral in soils. Therefore REE losses are linked to an increase of the weathering stage. Principal mineralogical evolutions with increasing age in the surface horizon (from the P3 E to the P4 E) of the Vancouver chronosequence is a decrease of the modal abundance of primary minerals, except for quartz (Figure 6.4). Feldspars and quartz are known to contain negligible amounts of REE with the exception of Eu in feldspars (Towell et al., 1969, Condie et al., 1995, Compton et al., 2003).

117

Figure 6.5 - Total REE, La/Gd, TRB, Si/Al, total and pyrophosphate-extracted Al, total and DCB-extracted Fe - in P3, P4 and P5. 118

Part 3 Chapter 6 – Mineralogical evolutions

Preferential dissolution of feldspars relative to other REE-bearing primary minerals should, therefore, result in negative solid Eu- anomalies. However, no significant evolution of the Eu anomaly was observed in E horizons. The observed signature can originate from the dissolution of other mineral phases present in this horizon (illite, hornblende, augite, chlorite), less abundant in the chronosequence. For example, amphiboles can contain high concentrations in REEs and negative anomalies in Eu (Skublov and Drugova, 2003). The weathering of two minerals with opposite Eu anomalies may explain the absence of change in this anomaly through time. Furthermore, accessory minerals (i.e. heavy minerals and phosphates) are known to present high REE concentrations (Bea, 1996), and to control the presence and dynamics of REE in weathering profiles (Braun et al., 1998).

Heavy minerals known to host REEs are Zr- and Ti- bearing phases, such as titanite (sphene), anatase, ilmenite and zircon (Braun et al., 1990, Braun et al., 1998, Aubert et al., 2001, Takahashi et al., 2003). Since most of the heavy minerals are rather stable through weathering (Nickel, 1973), the REEs included in these minerals are expected to accumulate in the weathered horizons. Consequently, those phases cannot explain the loss of REE with weathering observed in the E horizon.

Phosphate-bearing minerals (apatite and monazite, xenotime, rhabdophane, etc.) typically contain thousands of mg kg−1 of REEs (Henderson, 1984, Hughes et al., 1991, Frietsch and Perdahl, 1995, Taunton et al., 2000, Jordens et al., 2013). Consequently, they can largely influence REE content even if small quantity of primary phosphates is present and weathered (Braun et al., 1993, Braun et al., 1998, Aubert et al., 2001, Galan et al., 2007, Stille et al., 2009, Berger et al., 2014, Hissler et al., 2015). Some of these phosphate phases, like apatite, tend to disappear in highly weathered material, with significant organic material (Taunton et al., 2000, Berger et al., 2014). In the Vancouver chronosequence podzols, we observe a correlative behavior between the total REE content and the total P content (r = 0,624), suggesting that P-phases might play a significant role controlling the REE budget in the weathering profile. The total P content in the E horizon of the P3, P4 and P5 profiles is respectively,

119

Part 3 Chapter 6 – Mineralogical evolutions

51, 86 and 56 % lower than the content in the BC horizon (Table 4.3). As those phases, if present, are in small amount, they could not be quantified using the classical XRD method combined with Rietveld analysis.

Furthermore, the signature in the surface horizons is impacted by a preferential decrease of the MREE content, as observed in the profiles P4 and P5 in Figure 6.6. This specific signature could result from “M trend”, a preferential complexation of the MREE by DOM, impoverishing the solid residue (Tang and Johannesson, 2010). Indeed, some authors observed a MREE enrichment into soil solutions, and hypothesized that this could be a fingerprint of the solubilization of REE bound to OM (Davranche et al., 2011). We observe diagnostic horizons of podzolization in the three podzolic profiles, with an accumulation of OM mainly in the Bh and Bhs horizons in P3, P4 and P5 (Table 4.2). In addition, we observe an accumulation of Al and Fe secondary phases (i.e. organo-metallic complexes and short range- ordered Fe- and Al-(hydr)oxides, in the Bh, Bhs and Bs horizons of P4 (Figure 6.5 g and h) and P5 mature podzols (Figure 6.5 k and l). The majority of the Fe secondary phases are measured in the P4-330 and P5-530 Bhs horizons, as shown by the increase of the ratio Fed/Fet in these horizons (Table 4.3). These are also the horizons with the highest C content, Fe oxides being known to provide an important mineral surface for the sorption of OM (Dümig et al., 2012). Surprisingly, no accumulation of REE is observed in the Bh-Bhs-Bs horizons of the podzols, despite these horizons contain a high amount of OM and Fe and Al oxyhydroxides (Figure 6.5 f and j, Figure 6.6), known to scavenge REE (Rankin and Childs, 1976, Palumbo et al., 2001, Davranche et al., 2011, Cidu et al., 2013). In addition, no significant correlation was found between total REE, MREE, LREE and HREE, and the Al, Fe and Si extractable secondary phases, nor positive Ce anomalies, which would be a characteristic feature of REE association with Mn and Fe (Tripathi and Rajamani, 2007, Yusoff et al., 2013). Thus, we suggest that sorption, adsorption, co-precipitation, surface complexes formation, ion exchange and penetration of the lattice of the secondary Al and Fe phases did not affect significantly the total content of REE in the bulk soil samples of Bh, Bhs and Bs horizons.

120

Part 3 Chapter 6 – Mineralogical evolutions

We can conclude that the REE patterns in the E-horizons are impacted by weathering but they do not follow the fate of Fe- and Al- oxyhydroxides and OM in the soil profile. However, to further examine the potential of REE to trace illuviation process, we completed our results with additional REE data analyzed in Podzols, available in the literature (see next section).

6.2.3.3 Pedological processes behind the evolution of the REE signature

We compared the REE signature evolution of our studied Podzols with those published in the literature (Öhlander et al., 1996, Land et al., 1999, Öhlander et al., 2000, Aubert et al., 2004, Tyler, 2004, Mourier et al., 2008, Vodyanitskii et al., 2011). Our compilation contains only studies where the REE contents in the E, Bh/Bhs/Bs horizons and the parent material are available (Figure 6.7). The described podzols developed under a coniferous forest, and in a temperate humid climate. Yet, they are older (between 8700 and 15000 yrs old), they have different parent material and they received less precipitations compared to the Podzols from the Vancouver chronosequence (Table 6.3).

In all Podzols, we observe that the topsoil E and B horizons are depleted in REE compared to the parent material (Figure 6.7). Losses of REE from the parent material due to weathering has been demonstrated not only for Podzols (Vodyanitskii et al., 2010, Vodyanitskii et al., 2011) but also for other different soil types developed on various parent materials such as granodiorite, Archean granite, serpentinite, shale, etc. (Nesbitt, 1979, Marsh, 1991, Prudencio et al., 1993, Mongelli, 1993, Panahi et al., 2000, Aubert et al., 2001, Ndjigui et al., 2008, Ma et al., 2011).

121

Part 3 Chapter 6 – Mineralogical evolutions

1,8 A. 1,6 1,4 1,2 1,0 0,8 0,6 0,4 0,2 0,0 REE ofREE P4 the normalized BC tomean La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu P4 E P4 Bh P4 Bhs P4 Bs P4 Bw P4 BC 1,8 B. 1,6 1,4 1,2 1,0 0,8 0,6 0,4 0,2 0,0

REE ofREE P5 the normalized BC tomean La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu P5 E P5 Bh P5 Bhs P5 Bs P5 Bw P5 BC

Figure 6.6 - REE patterns of the P4 horizons (A.) and of the P5 horizons (B.), normalized to mean BC value of the three podzols. Error bars correspond to the result of error propagation calculations, based on the RSD given by the machine for each.

122

Table 6.3 - Parent material, total REE content in the parent material, vegetation, mean annual rainfall, mean temperature and soil age for the four podzolic profiles in the literature used for comparison.

Parent material and/or Total REE mean annual Reference vegetation mean t° soil age primary minerals in the PM rainfall µg.g-1 mm/yr °C yrs Coniferous forest (Pinus cembra and Larix Mourier et decidua, Picea abies and Abies alba.)Forest al 2008 Quartz, muscovite graphite 255.8 947 ± 184 7.1 ± 0.6 <15000 understorey : V. ulliginosum, V. myrtillus and Loup 2 Rhododendron ferrugineum Coniferous forest (Pinus cembra and Larix Mourier et Quartz, muscovite graphite decidua, with scattered Picea abies and Abies al 2008 traces of k-feldspar, albite and 155.7 947 ± 184 7.1 ± 0.6 <15000 alba.)Forest understorey: Vaccinium vitis-idea Orelle 2 zircon tourmaline apatite and Juniperus sibirica quartz, plagioclase, K-feldspar, biotite accessory amounts of Land et al amphibole, epidote, zircon, 180.3 Coniferous forest (spruce and pine) 500 -0.2 8700 1999 ilmenite, apatite, garnet, and clay minerals At least since the Middle Ages: Heather (Calluna moraine, derived from a vulgaris). After 1950s: spruce (Picea abies). mixture of quartzite Ground vegetation nowadays: Vaccinium 16 (July) 13000- Tyler 2004 (Cambrian sandstone) and 94.0 700 myrtillus, Dicranum scoparium, C. vulgaris, -2 (January) 14000 gneiss (rich in potassium Deschampsia lexuosa, and Pleurozium feldspar) schreberi.

123

Table 6.4 - Total C content; total element content; TRB; oxalate- and pyrophosphate-extracted Al (Alo and Alp respectively); DCB-, oxalate- and pyrophosphate-extracted Fe (Fed, Feo and Fep, respectively) of the 4 podzols from the literature.

depth C wt. % Si Al Fe Ca K Mg Na TRB Alo Alp Fed Feo Fep

g/kg g/kg g/kg g/kg g/kg g/kg g/kg cmolc/kg g/kg g/kg g/kg g/kg g/kg Mourier et E 3-11 6.60 345.07 86.50 20.08 1.86 25.64 4.95 10.98 138.33 0.021 0.015 0.084 0.030 0.029 al 2008 Bhs 11-20 2.10 299.49 94.60 73.69 1.29 25.89 9.95 10.01 153.94 0.041 0.025 0.231 0.128 0.072 Loup 2 Bs 1 20-33 2.30 307.53 97.56 58.64 1.29 24.23 9.77 11.13 153.79 0.067 0.052 0.149 0.075 0.056 Bs 2 33-53 3.20 306.92 97.83 58.57 1.57 23.65 10.01 11.57 155.95 0.079 0.069 0.149 0.087 0.074 C 53-73 1.40 311.92 95.76 49.76 1.50 26.64 11.04 11.80 168.59 0.060 0.039 0.086 0.055 0.028 Mourier et E 7-13 7.80 357.08 70.25 30.30 1.79 18.42 6.57 10.61 124.77 0.024 0.021 0.199 0.069 0.046 al 2008 Bs 1 13-27 3.20 360.92 68.87 29.60 1.29 16.43 6.27 10.68 117.51 0.020 0.016 0.169 0.069 0.062 Orelle 2 Bs 2 27-56 1.40 354.84 67.50 35.55 0.71 18.92 10.25 9.79 134.96 0.063 0.054 0.181 0.070 0.063 C 56-96 1.10 354.09 68.40 31.84 1.14 20.58 11.52 9.87 145.80 0.026 0.023 0.118 0.037 0.033 Land et al E 0-5 0.37 343.22 61.06 16.24 1999 B 10-15 2.32 222.38 87.71 74.95 C 80-85 0.14 292.87 77.94 42.57 Tyler 2004 E 14-37 0.30 29.90 4.54 2.26 21.90 0.20 8.24 98.31 B1 37-50 2.60 31.80 18.40 2.66 20.80 1.17 8.26 100.59 B2 50-67 1.50 32.50 14.20 3.74 20.20 1.44 9.30 107.37 B3 67-87 0.84 B/C 87-100 0.41 C >100 0.46 41.20 11.90 5.30 22.10 1.45 9.46 116.86

124

1,2 1,2 a b 1,0 1,0

0,8 0,8

0,6 0,6

0,4 0,4

0,2 0,2 E E B1 B B2 0,0 0,0 La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu 1,2 1,2 c d 1,0 1,0 0,8 0,8 0,6 0,6

EE normalized C horizon to the EE

R 0,4 0,4 E E Bhs Bs1 0,2 Bs1 0,2 Bs2 Bs2 Bs2 bis 0,0 0,0 La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 6.7 - REE pattern normalized to the C horizon for podzol profiles from the literature: (a) Land et al. (1999); (b) Tyler (2004) ; (c) and (d) Mourier et al. (2008) - Loup 2 and Orelle 2.

125

Part 3 Chapter 6 – Mineralogical evolutions

A second similitude between the podzols is that REE do not accumulate significantly in the B(h, hs, s) horizons compared to the BC parent material (Figure 6.6 and 6.7). Mourier et al (2008) also measured the DCB, oxalate and pyrophosphate-extractable Al and Fe content in the soil samples (Table 6.4), and observed an accumulation of secondary Al and Fe phases. The accumulation of secondary phases is one order of magnitude lower than the one observed in the P4 and P5 profiles of the Vancouver chronosequence, making even more surprising the absence of REE accumulation in the latter. This confirms our observation that the distribution of Fe, Al-OM complexes and coprecipitates in the soil profile do not significantly contribute to the accumulation of REE in the bulk soil during Podzol development. The main pathway followed by REE during podzolization is a release from the minerals during weathering, and a subsequent leaching with percolating water, resulting to a net loss of REE from the soil profile. This trend is different from the one observed in laterites under (sub)tropical conditions, where REE accumulation in B-horizon is quite common, often associated with Fe- and Mn-phases in their oxidized form (Braun et al., 1990, Braun et al., 1993, Sanematsu et al., 2011, Berger et al., 2014, Janots et al., 2015). The difference observed between podzolic pedosystems and lateritic soil profiles in humid tropical regions might come from a difference in soil process involved in the REE redistribution and accumulation. Indeed, the podzolization process is mainly controlled by organo-metallic complexation, but the ones involved in the mentioned studies are due to the ferralitization process, where OM mineralization is fast and where the organo- mineral association phenomenon is less important. The nature of redox process might explain the development or not of an anomaly in Ce, a redox-sensitive element (Nakada et al., 2013). In our chronosequence, we observed no Ce anomaly. Speciation of Ce with organic ligands might be a way to solubilize both Ce(III) and Ce(IV) without fractionating them (e.g., Janot et al. 2015).

Comparing REE patterns analyzed in the Vancouver Podzols with those published in the literature shows that there is no homogeneity in the fractionation signatures observed in the bulk soil horizons of Podzols (Figure 6.7). Land et al. (1999) observe a

126

Part 3 Chapter 6 – Mineralogical evolutions preferential depletion in LREE, Tyler (2004) a similar depletion of all REE, and Mourier et al. (2008) a preferential MREE depletion, as also observed in our study. This variability can be explained by the variety of factor that can impact the REE signature.

First, various compositions and solubility of minerals present in the soil parent material can partly govern the REE released with weathering. However, the lack of quantification of mineral composition in the published studies, including accessory minerals, prevents further comparison among studies.

Secondly, the type of OM present in solution can play a role in the process of REE complexation, and consequently in the REE mobilized with the organic carrier within the soil profile. The LREE are known to be preferentially bound to carboxylic groups, whereas the HREE are preferentially bound to carboxy–phenolic and phenolic groups (Marsac et al., 2011, Gangloff et al., 2014). Some authors (Yamamoto et al., 2005, Pourret et al., 2007a, Tang and Johannesson, 2010, Davranche et al., 2011, Cidu et al., 2013) make the distinction between a “colloidal pool” (humic acids) in soil waters, enriched in MREE, and a “dissolved pool” with a low REE concentration (complexed with fulvic acids), LREE depleted, but HREE enriched. The type of chemical bonds with organic molecules and the resulting size of the organo-REE complexes can therefore impact the fractionation of REE during weathering and as such the signature of the REE leached from the soil profile.

Finally, the physico-chemical conditions of the solution (e.g. the pH) and the other elements present in solution can have an impact on the mobility of REE and interfere with the complexation of REE by OM. Some studies show that for a same OM, the metal loading can affect the preferential affinity for MREE or HREE. Al and Fe may compete with REE in forming organic complexes, and an increase in concentrations of these cations can cause a decrease in the amount of REE bound to dissolved OM (Tang and Johannesson, 2003, Pourret et al., 2007a, Cidu et al., 2013). A MREE enrichment in the liquid phase is shown by patterns at high metal loading, whereas patterns at low metal loading display a regular increase from La to Lu (Marsac et al.,

127

Part 3 Chapter 6 – Mineralogical evolutions

2010), that would imply a preferential loss of HREE from the horizon. The ion activity in soil solution can therefore also substantially govern the preferential leaching of MREE during soil weathering over time.

Further studies, including a characterization of the accessory minerals and the type of DOM present, are needed to determine the relative importance of these factors on the REE fractionation during podzolization.

6.2.3.4 Kinetic of weathering in the E horizon of the Podzol chronosequence

The Figure 6.8 a presents the evolution of parameters in the E horizon relative to parent material over time in the podzolic soil chronosequence: total REE content, La/Gd ratio (tracing the MREE depletion), TRB (index of soil weathering degree) and Al/Si ratio (index of quartz enrichment). Both normalized REE concentrations and TRB decrease in the E horizon over time. However they do not decrease with the same rate, the TRB decreasing more rapidly than the REEs.

In parallel, both normalized Si/Al and La/Gd ratios increase over time with the same rate. This confirms the hypothesis that MREE depletion during soil evolution is partly controlled by the weathering of specific silicate minerals (e.g., feldspars, amphiboles, illite) and the enrichment of quartz and some accessory phases.

Previously published studies, focused on REE content in Podzol, present parent material ages ranging between 8700 and 15000 yrs (Öhlander et al., 1996, Land et al., 1999, Tyler, 2004, Mourier et al., 2008), or more (Aubert et al., 2001, Aubert et al., 2004), which does not allow determination of the early rates of REE losses, nor the early temporal evolution. Here we show a significant depletion in REE content in the surface E horizon relative to the parent material after ~300 years. This proves that REE can be released and mobilized in very short periods of time after the beginning of podzolization.

128

Part 3 Chapter 6 – Mineralogical evolutions

2 A. 1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 TRB

horizon,relative toPM La/Gd 0,2 REE Parameters Parameters evolution the in E Si/Al 0 200 300 400 500 600 Time [yrs]

1 B. 0,9 Mourier et al, 2008 (O2) 0,8 P3-270 yrs Mourier et al, 2008 (L2) 0,7 P4-330 yrs 0,6 P5-530 yrs 0,5 0,4

relaive toPM Land et al, 1999 0,3 Tyler, 2004 0,2

Sum Sum ofREE the in E horizon, 0,1 0 0 5000 10000 15000 20000 Time [yrs]

Figure 6.8 – Kinetics of REE release relative to parent material (PM) over time: (A.) temporal evolution of the total REE concentration, TRB, La/Gd, Al/Si in the E horizon of the studied Vancouver podzolic soil chronosequence – (B.) temporal evolution of the total REE concentration in the E horizon of the studied Vancouver podzolic soil chronosequence and in other podzolic soil profiles from the literature (Land et al., 1999; Tyler, 2004 and Mourier et al., 2008). A logarithmic fit considering all Podzols, except the two of Mourier et al. (2008) was performed.

129

Part 3 Chapter 6 – Mineralogical evolutions

When REE loss in the surface E horizon, compared to the PM, of the Podzols in the literature for which the information is available is plotted in function of time (Figure 6.8 B.), a logarithmic decrease is observed for all Podzols, except the two Podzols of Mourier et al. (2008). Such a trend is consistent with the fact that weathering processes are faster at the beginning of pedogenesis, and decrease with time, as the easily weatherable minerals stock decrease. This significant trend (r²=0.9837) of the weathering kinetics is observed for Podzols developped under different climates (rainfall more than 4 times larger in Vancouver Island than in the sites of Land et al., 1999 and Tyler, 2004; Table 6.3) and from a PM having different total REE content. The mineralogy, even if different between the compared Podzols, has a common point of being rich in quartz and feldspars. The Podzols studied by Mourier et al. (2008) have a very different kinetics, with much lower REE losses compared to the PM. The main difference between the Podzols of Mourier et al. (2008) and the ones of Vancouver Island, Land et al. (1999) and Tyler (2004), is the parent material composed of quartz, muscovite and graphite. Muscovite is known to be more stable than feldspars (Dixon and Weed, 1989), which could explain the weak REE loss after 15000 yrs in the surface horizon, even if the PM is relatively rich in REE. The content and composition of accessory minerals, not quantified in this study, may also explain the differences observed.

6.2.4 Conclusion

Our study shows that a large proportion of REE initially present in the parent material is rapidly lost from the soil profile through mineral weathering and leaching of dissolved elements and/or colloidal particles with percolating water (34 and 37% of loss in the E horizon compared to the BC horizon after 330 and 530 years, respectively). Soil-forming factors, such as vegetation and parent material composition can influence soil physico-chemical properties that in turn play a key role in the fractionation of REE during weathering. Environmental conditions are therefore important drivers controlling the rate of REE leached out of soil profile but also the REE

130

Part 3 Chapter 6 – Mineralogical evolutions signature of soil solution and the preferential leaching of some lanthanides to the hydrosphere.

Furthermore, the accumulation of secondary OM, Al and Fe- bearing phases does not impact the REE signature of the bulk soil. Sorption, adsorption, co-precipitation, surface complexes formation, and ion exchange with the secondary Al and Fe phases are consequently not mechanisms affecting significantly the total content of REE in the Bh, Bhs and Bs horizons.

At last, the demonstration, for the first time, that a large proportion of REE initially present in the parent material can be released and mobilized in very short periods of time during podzolization (330 yrs) has an important implication for the geochemical behavior of REE at the ecosystem scale and in the hydrological system. Further studies are needed to explain the differential evolutions of signatures in different podzols, and to better understand how the OM and physico-chemical factors impact the release, fractionation and leaching of REE outside the soil profile.

131

Part 3 Chapter 6 – Mineralogical evolutions

6.3 Synthesis

6.3.1 Validation of the chronosequence

As developed in the section 6.22 and appendix A2.23, a chronosequence is a series of soils that developed from the same parental material, vegetation, topography and climate, with pedogenesis duration being the only varying factor (Jenny, 1961). As distances between pedons are short and the slope is gentle, both climate and relief can be considered as constant over the soil sequence. Nevertheless, microtopography likely has led to variations in soil water regimes between pedons. Vegetation has developed with time: though plant community similar, their specific composition and quantitative return to soil have evolved with increasing distance from the sea.

6.3.2 Characterization of the parent material

As developed in detail in the section 6.2, the REE results confirmed that the sediment at the origin of the parent material of the sequence (beach sand and BC horizon of all pits) comes from the same source. Furthermore (also refer to Appendix A2.2), we showed a variation of the parent material composition from the C beach sand to the P2 BC horizons. Based on the REE signature, total Fe, Zr and Ti concentrations, content in sand and Si/Al ratio, we suggested that these variations are due to a change in the sedimentation dynamics, leading to a variable content of sand, silt and clay. We also confirmed that the three podzols developed from the same parent material, as P3, P4 and P5 BC horizons have a similar mineralogical composition.

2 Adapted from Vermeire M.L., Cornu, S., Fekiacova, Z., Detienne, M., Delvaux, B., Cornélis, J.T., 2016, Rare Earth Elements dynamics along pedogenesis in a chronosequence of podzolic soils, Chemical Geology 446: 163-174

3 Adapted from Fekiacova, Z., Vermeire, M. L., Bechon, L., Cornelis, J. T., Cornu, S., 2017, Can Fe isotope fractionations trace the pedogenetic mechanisms involved in podzolization? Accepted in Geoderma. DOI: 10.1016/j.geoderma.2017.02.020 132

Part 3 Chapter 6 – Mineralogical evolutions

6.3.3 Processes involved in podzolization

In the chronosequence, the parent material is a sedimentary material, composed of “lithogenic” primary minerals (beach sand). With time and pedogenesis, we described (1) a weathering of the “lithogenic” minerals”, (2) and the accumulation of “pedogenic” secondary minerals: clay minerals, Fe and Al oxides (sections 6.1, 6.2, A2.14 and A2.2).

6.3.3.1 Acidification and weathering

Acidification and weathering are known to be the main processes occurring in the E compartment during podzolization. In the Vancouver chronosequence, a fast decrease of the pH in the surface horizon was measured from the beach sand C horizon (pH 8) to the P3- 270 E horizon (Figure 6.1). From P3-270 yrs to P5-530 yrs surface horizon, the pH remained stable around 4.7. Acidification was a prerequisite condition for the start of podzolization. Once the threshold pH value reached (in P3 E horizon), Al, Fe, alkaline and alkaline-earth cations mobilization in the surface horizon was triggered rapidly, in less than 50 years. The weathering stage is measured by the total reserve in bases (TRB), a weathering index estimating the content of weatherable minerals in soils (Herbillon, 1986). As shown in Figure 6.1, TRB does not vary with depth in P1-120 yrs, P2-175 yrs and P3-230 yrs as it ranges between 300 and 400 -1 -1 cmolc.kg . However, it decreases in the P4-330 yrs (236 cmolc.kg in -1 the E horizon), and reaches 170 cmolc.kg in the P5-530 yrs E horizon. This timing for incipient podzolization is in accordance with most studies (as reviewed by Sauer et al. (2008)), observing the formation of a weak E horizon after about 200 to 500 y.

4 Adapted from Cornélis, J.T., Weis D., Lavkulich, L., Vermeire M.L., Delvaux B., Barling, J. (2014) Silicon isotopes record dissolution and re-precipitation of pedogenic clay minerals in a podzolic soil chronosequence, Geoderma 235–236, 19– 29. 133

Part 3 Chapter 6 – Mineralogical evolutions

As revealed in the section 6.2, REE are a good weathering tracor. Their content is trongly correlated to the general loss of elements and mineral weathering (estimated by the TRB, Si/Al ratio and in quartz content). The signature and release of REE is mainly determined by the composition and solubility of minerals present in the parent material.

Both Fe and Si isotopes revealed a preferential release of light elements during weathering:

Fe isotopes in the bulk fraction revealed a preferential release of light element during weathering in the Vancouver chronosequence. Acidification did not impact the bulk soil Fe isotopic signature, but from P3 270 yrs, the preferential release of light Fe isotopes from the minerals resulted in enrichment of heavy Fe isotopes and positive δ56Fe values in the surface horizon (refer to section A2.2.4). We hypothesize that the intensity of this fractionation may be related to the nature of the silicates weathered.

The first stages of weathering (from beach sand to P2) are only measurable in the clay fraction, representing a small proportion of the bulk soil. The clay minerals modifications (transformation and neoformation) started already in P1 and P2, before podzolization and C/Al/Fe mobilization (as described in the section A2.1.4). At the very beginning of soil formation (in the youngest profiles, P1 and P2), the neoformation of kaolinite, illite and chlorite from dissolution of primary minerals is observed. In developed podzols, we observed (1) the disappearance of kaolinite in the weathered E horizon, (2) an increase of relative abundance of kaolinite in Bh and Bhs horizons compared to E horizon, (3) the accumulation of imogolite-type materials in Bhs and Bs horizons. The upper E horizon is a very aggressive weathering system, controlled by organic acids as major proton donors and complexing metals, which leads to dissolution of primary and secondary minerals. The Si isotope composition in the clay fraction (comprised of primary and secondary minerals) document an increasing light 28Si depletion with soil age in eluvial E horizon. The mass balance approach show also a progressive depletion in light 28Si in secondary (pedogenic) clay minerals of the E horizon (from -0.51‰

134

Part 3 Chapter 6 – Mineralogical evolutions to 0.64‰), explained by a preferential release of light Si isotope during their dissolution.

6.3.3.2 Fe, Al and Si mobilization and accumulation

No significant accumulation of secondary minerals is measured in the youngest profiles (P1 to P3). As described in the section 6.1, we observe an accumulation of Al and Fe secondary phases (i.e. organo- metallic complexes, short range-ordered Fe- and Al-(hydr)oxides and ITM), in the Bh, Bhs and Bs horizons of P4 and P5. In the Bs horizon, inorganic acids (carbonic and nitric acid) are the major proton donor and the weathering system is less aggressive.

Surprisingly, no REE accumulation was measured in the illuvial horizons of podzols (neither in the Vancouver chronosequence, nor in other podzols in the literature), despite their known affinity for OM and Fe oxydes. Sorption, adsorption, co-precipitation, surface complexes formation, ion exchange and penetration of the lattice of the secondary Al and Fe phases did not affect significantly the total content of REE in the bulk soil samples of Bh, Bhs and Bs horizons. The main pathway followed by REE during podzolization is a release from the minerals during weathering, and a subsequent leaching with percolating water, resulting to a net loss of REE from the soil profile. Consequently, we concluded that REE can not be used to trace the neoformation of secondary phases in podzols.

The stable Fe and Si isotopic signature of the illuvial horizons confirmed the neoformation of minerals, with material coming from the overlying eluvial horizon. The bulk soil Fe isotopic signature revealed a decrease of δ56Fe values simultaneous to the positive flux of OM-bound and colloidal Fe pool in the Bhs horizon, from 270 to 335 years (Figure A2.2.5 b). Whether this enrichment in light Fe isotopes was due to (i) the preferential input of light Fe coming from the surface, (ii) a preferential incorporation of light Fe in the secondary Fe phases or (iii) a preferential complexation of light Fe by the OM, still needs to be further examined. The Si isotope compositions in the clay fraction also document an increasing light 28Si enrichment in the

135

Part 3 Chapter 6 – Mineralogical evolutions illuvial B horizon. The known preferential incorporation of light 28Si during neoformation of secondary pedogenic minerals is one factor explaining their isotopically lighter signature relative to primary lithogenic minerals. In addition, the mass balance approach revealed that the preferential release of light 28Si during dissolution of secondary minerals in the E horizon, partly accounts for the enrichment in light during re-precipitation of new clay minerals in the Bhs horizon. These data show that the secondary minerals (pedogenic pool) are, with primary minerals, also a source for neoformation of secondary minerals in the clay fraction. Consequently, those new clay minerals can be called “tertiary minerals”.

The evolution of Si isotopic signature in pedogenic clay minerals of the podzolic soil chronosequence therefore corroborates the process of dissolution and re-precipitation of aluminosilicate phases during podzoliaztion. For Fe and Al, the origin and dissolution- reprecipitation cycles, especially in the B horizons, and the implications for OMA, still need to be further examined.

Implications for SOM protection mechanisms:

Our results suggest that the formation of OMA can be a mechanism for OM protection in the Bh, Bhs and Bs of P4 and P5 profiles of the chronosequence, where secondary Al and Fe phases with a large SSA are identified.

136

Part 3 Chapter 7 – SOM evolutions

Chapter 7 - Evolution of SOM, microbial populations and organo-mineral associations5

The objective of this chapter is to describe the evolution of SOM in the chronosequence. We will first describe the accumulation of OM with time, its composition and repartition within functional fractions (including mineral-associated C). Secondly, we will quantify and describe the evolution of the microbial populations within the horizons of different age and depth. Finally, we will characterize C susceptibility to decomposition, and attempt to determine the relevant C protection mechanisms.

5 In preparation: Vermeire M.L., Doetterl S., Delmelle P., VanOost K., Van Ranst E., Delvaux B., Cornélis J.T., Co-dependent evolution of soil microbial populations and organo-mineral associations in a podzolic soil chronosequence. 137

Part 3 Chapter 7 – SOM evolutions

7.1 Development of a C accumulation horizon

With time, a progressive increase of the OC content is observed along soil sequence, as well as the development of an illuvial Bh accumulating OM (Figure 7.1). In P1-120 yrs and P2-175 yrs, the OC content does not vary significantly with depth, and is around 6 g.kg-1 in P1 and 3 g.kg-1 in P2. In P3-270 yrs, the Bh exhibits a C content amounting to 17.8 g.kg-1. Furthermore, the pH decreases starting from the surface in P3 as compared to P1 and P2, whereas visual evidences of bleaching indicate the presence of E horizon. Sitka spruce, the dominant tree species, produces strong organic acids. Its litter is known to acidify soils following land cover transformation (Lindeburg et al., 2013). In P4 and P5, the accumulation of OM in the illuvial horizons (Bh, Bhs, and to a smaller extend in Bs) further increases, and reaches 25 and 43 g.kg-1 in the P4 and P5 Bhs, respectively.

7.2 Evolution of OM composition and quality

7.2.1 Bulk SOM composition

Along the chronosequence, the composition of OM likely changes from P1 to P5, as suggested by changes in bulk SOM C:N ratio (Figure 7.2 a.). This ratio is known to give a rough indication of OM evolution. The value of the C:N ratio of the litter (Oh horizon) overlying the profiles increases from P1 (31) to P4 (43) (Figure 7.2 a.). At the P1- 120 yrs location, shrubs and trees of sitka spruce are present, and the Oh was 6 cm thick. With an increasing distance from the sea, vegetation develops. It is composed, at the P4-330 yrs location, of Douglas fir, Hemlock, Red cedar, Salal and Western sword fern (26 cm- thick Oh). The values of the C:N ratio of the Oh horizons are in the range of the values for plant materials: around 20 - 50 for tree leaves, and between 25 - 80 for herbaceous plants (Kleber et al., 2007). Increasing C:N ratios of the litter in a giant podzol chronosequence was also observed by Jones et al. (2015), and interpreted as a declining quality of the litter.

138

P1 – 120 yrs P2 – 175 yrs P3 – 270 yrs P4 – 330 yrs P5 – 530 yrs

0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50

E E E BC1 Bh Bh Bh Bw Bhs Bhs Bs Bw Bs C-t Bw C-l Bw C-mp BC BC2 BC BC C-r BC

Figure 7.1 – In the 5 profiles of increasing age (P1 to P5): Evolution of the C fractions in the soil samples: Total C (C-t), labile C (C-l), mineral-protected C (C-mp) and recalcitrant C (C-r) [g kg-1].

139

Part 3 Chapter 7 – SOM evolutions

“Low quality litter” are degraded more slowly and composed of a high woody residues content (i.e., coarse roots and wood, with a higher cellulose and lignin concentrations than leaves and herbaceous plant material, and a low N content) (Cotrufo et al., 2013). In P1, the value of the C:N ratio in the mineral BC1 surface horizon (29.7) is similar to the one of the Oh horizon. However, in all other pedons, the value of the C:N ratio (around 20) in the surface mineral horizons (Bw in P2 and E in P3, P4 and P5), is smaller than the one in Oh horizons. In the deeper horizons, a progressive increase of the value of the C:N ratio of the illuvial horizon is observed, from 21.6 in P3-Bh, to 26.7 in P4-Bhs and 30.9 in P5-Bhs.

A decrease of the C/N ratio have been attributed, in the literature, to an increase of the SOM microbial processing and/or a higher proportion of microbial-derived compounds (Schmidt et al., 2000, Rumpel and Kögel-Knabner, 2011, Miltner et al., 2012, Sollins et al., 2009, Kögel-Knabner et al., 2008b). Indeed, the C:N ratio of bacteria is around 5 – 8, and of fungi ~ 10 (Kleber et al., 2007). This seems to be confirmed in the P3, P4 and P5 profiles of the chronosequence, as the surface E horizons presents a higher proportion of microbial-derived Amino-sugars compared to the illuvial B horizons (Figure 7.2 b.). A decrease of SOM content and C:N ratio with increasing depth is generally observed in soils, and approached the C:N of microbes in most subsoils (Schmidt et al., 2011, Rumpel and Kögel-Knabner, 2011). In podzols, the situation is, however, quite different. An increase of the C:N ratio in the Bh, Bhs and Bs horizons is frequently observed, and a C:N value > 25 in illuvial horizon is considered as a pertinent podzolization criterion (Baize, 1993). This might indicate a vegetal signature in the illuvial horizons, either due to the accumulation of DOM leached out from surface horizons, and/or roots (Buurman and Jongmans 2005), or to a low activity of microorganisms (Aran et al., 2001). As mentioned in the section 5.2, P4 and P5 are hydromorphic podzols. According to Buurman and Jongmans (2005), these hydric conditions lead to (1) an inaccessibility of the illuvial horizon for plant roots, making DOM leached out of the upper horizons as the predominant contributor to SOC accumulation, (2) an inhibition of microbial degradation (slow OM dynamics, favoring C accumulation) and lower microbial proportion in the total C.

140

-1 a. C:N ratio b. AS [mg.gC ] P1 P2 P3 P4 P5 0 10 20 30 40 50 0 10 20 30 40 Oh Oh Oh Oh

BC1 Bw E E

Bh Bh

Bhs

P1 Bw Bw P1 P2 P2 P3 P3 BC2 BC BC BC P4 P4

P5 P5

Figure 7.2 – a. Evolution of the C:N ratio with depth in the litter (Oh) and soil horizons for the five soil profiles of increasing age (BC1 and BC2 for P1-120 years, Bw and BC for P2-175 years, E, Bh, Bw, Bc for P3-270 years, and E, Bh, Bhs, Bs, Bw, BC for P4-330 years and P5-530 ears); b. Proportion of AS in the total C content (mg AS. g-1 C).

141

Part 3 Chapter 7 – SOM evolutions

7.2.2 Evolution of the repartition of C within fractions

NaOCl treatment has been shown to oxidize preferentially chemically labile compounds (such as polysaccharides and aromatics), leaving behind older compounds, inherently stable against chemical attack (black C and aliphatic compounds such as n-alkanes and n-fatty acids) and/or protected by association with soil minerals (von Lützow et al., 2007, Mikutta et al., 2006). Here, we use the term “labile” C to refer to an operationally-defined fraction (oxidized by NaOCl). HF treatment is used to release SOM from organo–mineral associations through dissolution of hydrated silicate minerals and the formation of complexes with Fe and Al (von Lützow et al., 2007, Eusterhues et al., 2007). This fraction estimates the amount of OM “protected” through its association with minerals (CMP). OM left after the NaOCl and HF treatment is hypothesized to be “recalcitrant” (Mikutta et al., 2006). Here, we use the term “recalcitrant” C to refer to an operationally- defined fraction (resistant to both oxidation and HF treatment). We suppose that recalcitrant organic substances exhibit specific properties at molecular-level that might lower its degradation.

In P1 and P2, the total C content is low (Figure 7.1), and the majority of the C is in the “labile” fraction (~ 70% of the total C). In P3, the labile C fraction is also the main phase in all horizons, and represents 78% of the total C in the Bh (Figure 7.3). In P4 and P5, the labile C fraction still represents an important proportion of the total C in the different horizons, excepted in P4 Bh (5%) and P5 Bs (7%). Yet the most striking feature is the importance of CMP in the Bh, Bhs and -1 Bs horizons. CMP amounts to 11.3, 13.2 and 3.7 g.kg soil in P4 Bh, Bhs and Bs, respectively, and thus accounts for 78, 53 and 28% of the total C. In P5, it amounts to 4.3, 15.9 and 7.7 g.kg-1 in Bh, Bhs and Bs, respectively, and thus accounts for 14, 37 and 74% of the total C. A strong and positive correlation is observed between the CMP content, and the organo-Fe and -Al complexes (R² = 0.72 and 0.83, for Fep and Alp respectively), Fe oxides (Fed-Fep: R² = 0.77), but not with Alo (R² = 0.36). Oxidation-resistant and HF-soluble OM fraction (CMP) usually correlates with the content of Fe and Al phases in podzols (Eusterhues et al., 2003), and in a wide range of different soils (Mikutta et al., 2006, Kögel-Knabner et al 2008b).

142

C fractions (g.kg-1)

0 10 20 30 40 0 10 20 30 40 0 10 20 30 40

P3 E 291952 P4 E 69 427 P5 E 3012 58 Bh 17 78 5 Bh 28 14 58

Bh 175 78 Bhs 7 53 41 Bhs 18 37 45

Bs 528 67 Bs 19 74 7 BW 2122 57 C-r Bw 3451 Bw BC 1017 72 C-mp C-l BC BC

Figure 7.3 - C fractions in the P3, P4 and P5 profile horizons. The labile OC (C-l) is calculated by subtracting the stable OC fraction (C-s, carbon left after treatment with NaOCl) from the total OC. The recalcitrant OC (C-r) is the non- extractable organic carbon after NaOCl and HF treatments. The mineral-protected OC (C-mp) was calculated by subtracting the C-r fraction from the C-s fraction. The numbers in each bar is the percentage of each fraction in the total C content of the horizon. C-r + C-mp + C-l = C tot.

143

Part 3 Chapter 7 – SOM evolutions

In P1 BC1 and BC2, and P2 Bw and BC horizons, as well as in Bw and BC horizons of P3, P4 and P5 profiles, recalcitrant C accounts to 1 g.kg-1 soil. Depending on total C content in the horizon, the amount of recalcitrant C represents different percentages of total C content (12- 62%). This could be due to the presence of “black carbon” in the sedimentary parent material, as observed in the P1 BC1 Kubiena box (Figure 5.2). Furthermore, recalcitrant C compounds accumulate in the E horizons. It represents 30 and 29 % of the total C in the P3 and P4 E -1 (4.4 and 2.4 g CR.kg soil, respectively), and accounts to 69% of the -1 total C in P5 E (10.2 g CR.kg soil, Figure 7.3). The selective preservation of “recalcitrant” organic compounds seems to represent a (temporary?) protection mechanism in the surface horizons. Interaction between SOM and minerals will be the dominant protection mechanism in the B horizons.

Chemical extractions present some limits, hence their outputs should be treated with caution (Sleutel et al., 2009, Favilli et al., 2008). However, considering the C:N ratio of the C fractions leads to further interpretation (Figure 7.4)

0 20 40 60 80 100 0 20 40 60 80 100

P3 P5 E E Bh Bh Bw Bhs BC Bs

C:N - L C:N - L C:N - R C:N - MP C:N - T C:N - R C:N - T

Figure 7.4 – Evolution of the C:N ratio in the C fractions (total (-T), recalcitrant (-R), labile (-L) and mineral-protected (-MP) carbon fraction), in the first horizons of P3 and P5.

144

Part 3 Chapter 7 – SOM evolutions

In P3 and P5, the bulk C:N ratio in the E horizon combines the values of a labile C fraction (C:N ~6-7, suggesting a microbial composition), and a recalcitrant fraction (C:N = 47.1 in P3 and 81.3 in P5, suggesting vegetal structural components with a high C content). The C:N value of P5 E CMP fraction (6.0) is similar to the one of the P5 E labile C fraction. The values of the C:N ratio of the different fractions in the illuvial B horizons are more similar to each other, hence also to the one of the bulk C:N (~20-30). Such values of the C:N ratio, as well as the relatively small proportion of AS in the illuvial horizons (Figure 7.2 b.), suggest that C compounds in the accumulation horizon are weakly decomposed substances, in all fractions. Buurman et al. (2005) observed similar patterns in a podzol hydrosequence. Therein, SOM in E horizons is invariably moving towards a dominance of (1) recalcitrant plant-derived aliphatic compounds (reflecting residual accumulation) and (2) easily degradable bacterial products (especially polysaccharide). Microbial degradation can even result in a temporary enrichment in polysaccharides. In contrast, transported OM (DOM) dominates the chemistry of OM in root-free, waterlogged B horizons (poorly drained podzols). The aromatics and polysaccharides dominate, whilst aliphatic and lignin compounds play only a minor role.

7.3 Evolution of the microbial populations

The measure of the AS content is a semi-quantitative estimation of the amount of dead microorganisms in the soil samples. They occur in living cells but persist after lysis of the cells. Only < 10% of amino sugars are found in microbial biomass, whereas > 90% is present in microbial residues that form part of the SOM (Pronk et al., 2015). Consequently, they are a useful biomarker for investigating microbial contribution to SOM (Bodé et al., 2009, Bodé et al., 2013, Zhang and Amelung, 1996, Amelung et al., 2001, Glaser et al., 2004). The total AS content in the chronosequence profiles is higher in the surface horizons and decreases with depth (Figure 7.5).

145

AS tot [µg.g-1 soil]

0 200 400 600 0 200 400 600 0 200 400 600 0 200 400 600 0 200 400 600

BC1 Bw E E E Bh Bh Bh Bhs Bhs

Bs Bs AS tot Bw C-L Bw Bw C-T BC BC2 P1 BC P2 P3 BC P4 BC P5 0 20 40 60 0 20 40 60 0 20 40 60 0 20 40 60 0 20 40 60

C-T and C-L [g C.kg-1 soil]

Figure 7.5 – Total amino sugar content (Glucosamine + Galactosamine + Muramic acis content, in µg AS.g-1 soil) and total (-T) and labile (-L) C fractions (g C.kg-1 soil) in the 5 profiles of increasing age (P1-120, P2-175, P3-270, P4-330 and P5-530 yrs).

146

Part 3 Chapter 7 – SOM evolutions

The total AS amount is low in the P2-175 yrs profile (~50 µg.g-1 soil), intermediate in the P1-120 yrs and P4-330 yrs profile (~200 µg.g-1 soil in the surface horizons), and high in P3-270 and P5-530 yrs (~300-400 µg.g-1 soil in the surface horizons).

Microorganisms develop in habitats where a C substrate is available, and in zones where the environmental conditions are favorable (notably adequate water, oxygen, and pH conditions) (Ekschmitt et al., 2008, Chenu and Stotzky, 2002). The differences in these two main factors explains the contrasted AS content in the chronosequence. The low content in P4 compared to P3 and P5 seems to be the combination of (1) a lower C content in P4 compared to P5 (Figure 7.1) (2) a similar C content in P4 compared to P3, but more adequate environmental condition (aerated structure, as revealed by micromorphological observations, Figure 5.2), and a substrate composed by a majority of “labile C” in P3 compared to P4. P2 exhibits the lower C content, explaining the very low AS content.

Chitin, the basis unit of the cell wall of fungi, is composed of N- acetyl glucosamine (Glu), while bacterial cell walls are composed of peptidoglycan, constructed of N-acetyl glucosamine and muramic acid (Mur, only found in bacterial cell walls) (Bodé et al., 2009). Consequently, the Glu/Mur ratio has been used as an indicator of the relative contribution of fungi and bacteria to SOM in soil samples (Liang et al., 2007, Amelung et al., 2001). The Glu/Mur ratio of bacteria range between <2 and 8 (Amelung et al, 2001), and was estimated around 271 for fungi by Glaser et al. (2004). In P1, P2 and P3 profiles, the ratio is high (> 40, and reached 98.8 in the P3 Bh), indicating a mainly fungal-dominated population of microorganisms within the profiles (Figure 7.6). These ratio are similar to the one observed by Amelung et al (2001) in beech litter (89.0), interpreted as a majority of fungal-derived Glu. In P4 and P5, a different microbial population is observed between the E horizon - mainly fungal-dominated population (higher ratio: 54.9 in P4 E and 62.2 in P5 E) - and the underlying horizons, having a ratio around 20 (reflecting an increase of the bacterial contribution). The Glu/mur ratio in the Bh, Bhs, Bs, Bw and BC horizons of P4 and P5 are in the range of values observed in mineral soils by Amelung et al. (2002) (~ 18-26), and Glaser et al. (2004)

147

Part 3 Chapter 7 – SOM evolutions

(~ 15-24). Interestingly, even if the total AS content and the proportion of AS in the total C differ between P4 and P5, the evolution of the microbial population composition is very similar. In podzols, the presence of fungi in the E horizon is already documented (Gadd, 2007, Van Breemen et al., 2000a, Nikonov et al., 2001). Fungi are the primary agents of decomposition (Bardgett and van der Putten, 2014), and play a prominent role in litter degradation (Beare et al., 1995). Furthermore, ectomycorrhizal fungi have been shown to participate actively to mineral weathering. They drill innumerable narrow cylindrical micropores (3–10 µm) into weatherable minerals in podzol E horizons, and produce micro- to millimolar concentrations of organic acids in fungal tips (Jongmans et al., 1997, Van Breemen et al., 2000a, Van Hees et al., 2003, Smits et al., 2005, Hoffland et al., 2002). However, very few studies give information on bacterial populations (Nikonov et al., 2001), and even fewer on their functions in podzols. Vodyanitskii (2003) mentioned a potential role of bacteria on the iron cycle in the B horizons.

Glu/mur 0 50 100 150 P1 P2 P3 P4 P5 BC1 Bw E E

Bh Bh

Bhs

Bs P1 Bw Bw P2 P3 BC2 BC BC BC P4 P5

Figure 7.6 – Evolution of the ratio between Glucosamine (Glu) and muramic acid (mur) content (two amino-sugars) in the 5 profiles of increasing age (P1-120, P2-175, P3-270, P4-330 and P5-530 yrs)..

148

Part 3 Chapter 7 – SOM evolutions

Fungi tolerates acidity and dry events, are all heterotrophs and obligatory aerobes. Bacteria are “aquatic organisms” (they need water), but can switch towards other electron acceptors than oxygen (the most significant is iron) (Chenu and Stotzky, 2002, Baldock and Skjemstad, 2000, Kögel-Knabner et al., 2008b, von Lützow et al., 2006). Fungi are generally specialized towards an assimilation of C directly from the litter (lignin compounds) while bacteria specialize in labile C compounds assimilation in the lower horizons (Poll et al., 2006, Ekschmitt et al., 2008). It is consequently pertinent to observe more fungi in the surface E horizon of podzols, more acidic and aerated, receiving input of C from the litter, and a majority of bacteria in the cemented hydromorphous illuvial horizons, receiving DOC coming from the E horizon.

The environmental constraint in the Bh, Bhs and Bs horizons of the chronosequence led to a change in microbial community. This will have an impact on SOM decomposition dynamics. Illuvial horizons can be considered as “non-preferred soil spaces” (sensu Ekschmitt et al 2008), where microbial activity may be reduced by suboptimal environmental conditions, and nutrient limitation (organic matter may be less accessible because of its association with reactive mineral surfaces) (Schmidt et al., 2011).

7.4 Evolution of total respiration and C decomposability

Incubations are useful for comparative and process-level investigations (Torn et al., 2009). We use them here to compare the C degradation dynamics between samples, and assess the impact of C quality and protection by minerals. We make the assumption that difference between the horizons in terms of environmental constraint are lowered during the incubations (compared to field conditions), as the same moisture and aeration conditions are applied. Due to the very sandy texture of the soil, no aggregates were observed, but differences in microenvironment might come from a different content in micro aggregates. After the pre-incubation of 15 days, the cumulative CO2 emitted with time increased linearly during the whole

149

Part 3 Chapter 7 – SOM evolutions incubation experiment (data not shown). Consequently, the -1 -1 degradation process rate in each horizon (in mg CO2.100g soil. day ) was determined by calculating the slope of the cumulative CO2 -1 emitted with time (all R² were >0.95), and converted in mg CO2-C.g C.day-1 per horizon (Figure 7.7). The total respiration rate per horizon -1 -1 (mg CO2-C.100g soil.day ) was correlated with the total C content of the sample (Figure 7.8 a.). When considered per horizon type, a different C decomposition dynamics was observed between the E and B horizons (Figure 7.8 b.) The lowest total respiration rates were observed in the Bw and BC horizons (C-depleted), and highest respiration rates were observed in the E and B horizons (Figure 7.7). When considering the “decomposability of the C” (mg CO2-C emitted per g C per day) the inverse tendency was observed (Figure 7.7). The horizons with the lower respiration rates per g C were the B horizons, and the highest respiration rates were observed in the Bw-BC horizons. E horizons have an intermediate decomposability. The B horizons are the ones with the highest CMP and secondary minerals content. Interaction of soil OM with Fe and Al phases is considered to be the main reason for stabilization of SOM from microbial attack in podzols (von Lützow et al., 2006, Jones et al., 2015, Eusterhues et al., 2003, Kalbitz et al., 2005, von Lützow et al., 2008, Mikutta et al., 2005a), and other soils (Kleber et al., 2005, Mikutta et al., 2006, Kaiser and Guggenberger, 2007). Such minerals provide large reactive surface areas (Eusterhues et al., 2005), which renders them most suitable for interaction with OM either via sportive interactions or coprecipitation (Kögel-Knabner et al., 2008b). A smaller biological activity due to Fe and Al bound to the organic matter was observed by many studies (Boudot, 1992, Boudot et al., 1989, Jones and Edwards, 1998, Sollins et al., 1996, Baldock and Skjemstad, 2000). Amelung et al. (2001) showed that Al and Fe oxyhydroxide inhibited synthesis of bacterial AS by a factor 2. As proposed by Aran et al. (2001) the high OM accumulation in podzol B horizons might be due to a protection of OM against mineralization.

150

Respiration [mg CO2-C / day]

0 0,2 0,4 0,6 0,8 0 0,2 0,4 0,6 0,8 0 0,2 0,4 0,6 0,8 0 0,2 0,4 0,6 0,8 0 0,2 0,4 0,6 0,8

P1 P2 P3 P5 BC1 Bw E E P4 E Bh Bh Bh Bhs Bhs

Bs Bs Bw Bw Bw BC2 BC BC BC BC

Figure 7.7 – Total respiration in the 5 profiles, in mg CO2 per day and (1) per 100 g of soil (continue line); (2) per gram of C (dotted line, OM “decomposability”)

151

0,6 0,6 0,6 a. b. c. 0,5 0,5 0,5 R² = 0,66 0,4 0,4 0,4

R² = 0,88 R² = 0,6703 C/g C/g C.day]

0,3 0,3 - 0,3

C/100g C/100g soil]

C/100g C/100g soil] -

0,2 0,2 0,2

Respiration rate Respiration rate

R² = 0,7076 [mgCO2

Respiration rate E [mgCO2 0,1 [mgCO2 0,1 B 0,1 R² = 0,64 Bw BC 0 0 0 0 20 40 60 0 20 40 60 0 20 40 C tot [gC. kg soil] C tot [gC. kg soil] Bulk C:N

-1 -1 Figure 7.8 – Correlations between the total C content (g C.kg soil) and the total respiration rates (mg CO2-C.100g soil.day-1) for the horizons of the 5 profiles (a.) and per horizon type (b.: for the P3, P4 and P5 E horizons (“E” serie), for the P3 Bh, P4 and P5 Bh, Bhs and Bs horizons (“B” serie), and P1 BC1 and BC2, and P2, P3, P4 and P5 Bw and BC horizons (“Bw BC” serie)). (c.) Correlation between the bulk soil OM C:N ratio and the OM decomposability (respiration -1 -1 rate in mg CO2-C.g C .day ).

152

Part 3 Chapter 7 – SOM evolutions

However, the interpretation of the role played by mineral phases in the lower OM decomposability in the B horizons is not straightforward. Other parameters impacts the respiration rates. First, a difference in OM “quality” (e.g. molecular weight, solubility, structural complexity, C:N). In the chronosequence, the respiration per g C was correlated with the bulk C:N (Figure 7.8 c). However, the bulk C:N hide the fact that OM is divided in two fractions of different properties in the E horizon (low C:N in the labile and mineral-protected C fractions, and high C:N in the recalcitrant fraction). In contrast, the OM composition in the B horizon is more homogeneous, and have a vegetal signature (Figure 7.4). A decomposing resource with a high C:N compared to that of living microorganisms (C:N ratios of bacteria is 5- 7:1, and of fungi 7-25:1) may limit microbial activity (Lavelle et al., 1993). The OM in the B horizon have a high C:N (20-30). This might partly explain the slower degradation.

Another factor possibly impacting respiration rates is the amount of microorganisms, and microbial community composition (Cotrufo et al 2013). According to Six et al. (2006), there is little to no support for the hypothesis that fungi have a greater Microbial Growth Efficiency than bacteria. Microbial community structure is very sensitive to changes in environmental conditions. Consequently, to investigate the eventual change in microbial population during incubations, and to quantify the living organisms participating to respiration, we analyzed phospholipid fatty acids (PLFAs). PLFAs are widely accepted as biomarkers for viable microbial biomass (Liang et al., 2008b, Zelles, 1999), unlike AS that are more an indicator of microbial necromass. The total PLFA content at the end of the 144 days of incubation (Figure 7.9), revealed the same pattern than the AS: (1) a low total content of microorganisms in P1 and P2 and P4, and higher in P3 and P5; (2) a total content in microorganisms higher in the surface horizons than in the deeper horizon (excepted in P5 Bh); (3) more fungi are present in the E horizon. Especially in the P5 Bh and Bhs, the protection by OMA is confirmed, because an important amount of microorganisms is present, and the respiration /gC is low.

153

A. PLFA nmol/g soil B. PLFA µmol/g C C. Fungal PLFA

0 50 100 150 0 5 10 0 10 20 P1 P2 P3 P4 P5 BC1 Bw E E

Bh Bh

Bhs

Bs Bw Bw P1 P2 BC2 BC BC BC P3 P4 P5

Figure 7.9 – In the 5 soil profiles, PLFA evolution: total PLFA (A. nmol.g-1 soil and B. µmol.g-1 C). C. fungal PLFA.

154

Part 3 Chapter 8 – Synthesis

Chapter 8 - Synthesis: co-evolution of mineral phases, microbial populations and SOM

None of the existing podzolization theories can explain, alone and convincingly, the large variation of podzol morphologies (refer to chapter 3). Also, these theories do not take into account the recent findings on OMA formation and dynamics, and the increasingly recognized preponderant impact of living organisms. As mentioned by Sauer et al. (2007) and Buurman and Jongmans (2005), the variety of environments in which Podzols occur, suggests that the processes of mobilization and immobilization of SOM, Fe, Al, and Si are not identical in all cases, and that the specific combination of soil-forming factors at each location determines which of the possible processes takes place and to what degree.

Here, we will apply the conceptual model developed in part 2 (synthetized in section 2.4) to the Vancouver podzolic soil chronosequence. The objective is to present a mechanistic description of the main pedogenetic processes and SOM stabilization mechanisms per site and horizon, and test whether the model is suitable to explain podzol morphologies.

155

Part 3 Chapter 8 – Synthesis

8.1 P1-120 yrs and P2-175 yrs – colonization

In the cox bay chronosequence, the SOM dynamics develops before any mineral modification. The soil formation starts with an input of SOM from the litter, and microbial processing. The SOM protection mechanisms in the two younger profiles (P1-120 yrs and P2- 175 yrs) are principally associated with inherent properties of the OM (composition of the litter input). No significant accumulation of recalcitrant compounds is measured, nor secondary mineral phases, preventing the formation of OMA. The environmental constraint is relatively low in these profiles, because of the high porosity and important precipitations. The microbial population composition do not vary significantly with depth, and is mainly fungi-dominated, associated with litter degradation (Figure 8.1). This is in the same line of Schulz et al. (2013) and Dümig et al. (2012), who also showed that biological communities drive ecosystem properties and development in soil chronosequences.

Figure 8.1 - Schematic representation of the pedogenetic processes and protection mechanisms relevant in the youngest profiles: P1-120 yrs and P2-175 yrs

156

Part 3 Chapter 8 – Synthesis

P1-120 and P2-175 years

Main processes:

 SOM compartment: Input of OM from the litter and litter degradation.

 Mineral compartment: /

 Physico-chemical compartment: /

Main biota:

Fungi-dominated population, homogeneous in composition through the profile.

Main SOM protection mechanism:

In these young profiles, the main constraints on SOM degradation are the composition of the organic molecules (accumulation of recalcitrant compounds) and an increasing environmental constraint with depth (less organic nutrients and oxygen with increasing depth).

8.2 P3-270 yrs – transition

In P3, microbial population and activity increases, and SOM starts to accumulate in an organic illuvial horizon. The pH decreases in the E compared to deeper horizons, but no weathering or secondary mineral formation is measured (Figure 8.2). As revealed by the micromorphological observations, the environment is relatively homogeneous through the profile, and comparable to the situation in P1 and P2. The microbial population is still mainly fungi-dominated. The very abundant fungal population seems to confirm the important role of fungi in podzol development (Van Breemen et al, 2000), notably for litter degradation and primary minerals weathering. The SOM

157

Part 3 Chapter 8 – Synthesis protection mechanisms in the P3-270 yrs profile are similar to the ones in P1 and P2.

P3-270 yrs

Main processes:

 SOM compartment: Input of OM from the litter and litter degradation. Accumulation of C in the Bh.

 Mineral compartment: /

 Physico-chemical compartment: Decrease of the pH.

Main biota:

Fungi-dominated population, homogeneous in composition through the profile.

Main SOM protection mechanism:

As in P1 and P2, the main constraints on SOM degradation are the composition of the organic molecules (accumulation of recalcitrant compounds) and an increasing environmental constraint with depth (less organic nutrients and oxygen with increasing depth).

158

Part 3 Chapter 8 – Synthesis

Figure 8.2 - Schematic representation of the pedogenetic processes and protection mechanisms relevant in the young podzol profile: P3- 270 yrs.

8.3 P4-330 yrs and P5-530 yrs – developed podzols

In P4-330 yrs and P5-530 yrs, we observe the development of two very distinct horizons in terms of processes, microbial population and C protection mechanisms (Figure 8.3).

8.3.1 Litter degradation and enhanced mineral weathering in the eluvial (E) horizon

The SOM transformations start with OM input from the litter, and degradation. An acidic environment develops, where organic complexing agents (coming from incomplete litter degradation, or from exudates of plant roots, fungi and bacteria) reinforce the action of the H+ ions, and induce a ligand-promoted dissolution (Duchaufour, 1997). The primary minerals are weathered, and release elements in solution. There, Al and Fe interact with DOM. The formation of organo- metallic complexes, and the metal:OM ratio will depend on the

159

Part 3 Chapter 8 – Synthesis transformation rates in the SOM and mineral compartments (degradation vs. weathering). The reviews on podzolization (Sauer et al., 2007, Lundström et al., 2000a, Buurman and Jongmans, 2005) point to the crucial importance of organic acids for (1) the mineral dissolution; (2) the downward transport of solutes, i.e. Al and Fe complexes; 3) the formation of the bleached eluvial E and illuvial B horizons. Given the high permeability of the substrate, and the large annual precipitation, soil solution moves downwards with the water flux, and supplies solutes in the B horizon.

In the E horizon, the main biota are plant roots and fungi. This sandy, acidic and well aerated horizon, meets roots and fungi requirements. As developed in the section 2.2, fungi are all aerobes and heterotrophs, and they are more resistant to acidic conditions and dry events than bacteria (Ekschmitt et al., 2008). Fungi are more abundant in coarse fractions (Kögel-Knabner et al., 2008b). In podzols, primary minerals are usually colonized by fungi (Gadd, 2007, Nikonov et al., 2001). The impacts of biota on the SOM compartment are: (1) litter decomposition (fungi) and (2) input of OM (exudates and dead cells of roots and fungi). Fungi are the primary agents of decomposition (Bardgett and van der Putten, 2014), and play a proeminent role in litter degradation (Beare et al., 1995). Both roots and fungi will also impact the mineral compartment, by participating to the weathering of primary minerals. Fungi exude siderophores and organic acids which promote mineral dissolution (refer to section 2.3.8). In podzol E horizons under coniferous forests, the weathering of primary minerals has been attributed to citric, formic, oxalic, succinic and malic acid excreted by saprotrophic and mycorrhizal fungi (Gadd, 2009). Ectomycorrhizal fungi actively participate in mineral weathering. They drill innumerable narrow cylindrical micropores (3– 10 µm) in the weatherable minerals present in podzol E horizons, and produce micro- to millimolar concentrations of organic acids in fungal tips (Jongmans et al., 1997, Van Breemen et al., 2000a, Van Hees et al., 2003, Smits et al., 2005, Hoffland et al., 2002). Tunnel formation in mineral grains was more intense in nutrient poor sites, indicating a higher contribution of fungi to plant P, Ca, K supply (Gadd, 2007, Van Breemen et al., 2000b). Plant-ectomycorrhizal fungi (rock eating fungi) combination has been proposed as a major driving force in the

160

Part 3 Chapter 8 – Synthesis formation of podzol E horizons (Van Breemen et al., 2000a, Van Breemen et al., 2000b). The E horizon has been considered the « fungal-eaten » part of the soil (Baldock and Broos, 2012).

P4 and P5 - E horizon

Main processes

 SOM compartment: litter degradation.

 Mineral compartment: primary (and secondary) minerals weathering.

 Physico-chemical compartment: acidification and leaching.

The acidification degree (nitric and organic acids) and redox status will control the transformation rate of minerals and OM (through impact on biota), and thereby the release of DOC and ions in soil solution. The climate (precipitations) and structure of the poral system will control the leaching of elements out of the E horizon, and the input for the B horizon

Main biota

Fungi and roots. The combination of an acidic and well aerated environment, and a fresh C input coming from the litter explains the fungi-dominated microbial population.

Main SOM protection mechanism:

As proposed by Von lutzow et al (2008), the principal C protection mechanism in the E horizon is a selective (temporary?) preservation of recalcitrant organic substances.

161

Part 3 Chapter 8 – Synthesis

8.3.2 Translocation of organic and inorganic elements

While migrating downward the profile, the soil solution may undergo several modifications. As described by Kaiser and Kalbitz (2012), soil acts towards DOM as a chromatographic system: the more sorptive (surface-reactive) compounds are retained and the more mobile ones are leached (also refer to section 2.3.6). Downward migration of DOM in the soil profile results from continuous (1) sorption and precipitation, (2) desorption and dissolution, combined with (3) microbial processing. The chromatographic effect also impact the dissolved mineral ions, which can be sorbed and desorbed on mineral surfaces. The passage through the chromatographic system induce changes in the soil solution composition while migrating through the profile. The poral system organization acts on the water transfers, and consequently on the surface exchange processes and the depth of migration. Fast water movement (for example preferential flows, in large pores) might decrease sorption/coprecipitation as well as microbial processing (Kaiser and Kalbitz, 2012). While migrating in the podzol E horizon, the solution experience pH change, and modification of the concentration (and eventually nature) in organic and inorganic constituents. The variations are be complex and depend on the relative importance of weathering-decomposition- OMA formation.

8.3.3 OMA accumulation in the illuvial B horizon

8.3.3.1 Secondary minerals and OM precipitation

As described in the section 2.3.6, and reviewed by kleber et al (2015), the key factors influencing the precipitation of OM by hydrolyzing metals are (1) the molar metal-to-carbon (M/C) ratio present in the soil solution, (2) the pH and concentration in inorganic ions (flocculating cations for example), (3) the type of metal species (Fe3+ or Al3+), and (4) the specific affinity of a given metal cation toward a given organic complexing agent. The impact of these factors on precipitation has also been shown in podzols (Jansen et al., 2004,

162

Part 3 Chapter 8 – Synthesis

Jansen et al., 2005, Nierop et al., 2002, Jansen et al., 2003). The precipitation of organometallic complexes and the formation of the illuvial B horizon thus result from a combination of factors: increase of the inorganic ions concentration, decrease of DOM concentration (due to degradation or sorption), pH change, etc. The dominant mechanism will depend on the biota, soil structure, mineral composition, and climate.

Secondary minerals

The type of minerals formed is difficult to predict, given the high complexity of soil environment (Cornell and Schwertmann, 2003). Podzols are natural systems with DOM being ubiquitously present. OM is known to impacts greatly the crystallinity, SSA, and type of secondary mineral phase formed. In podzols, Fe and Al phases formed are probably a complex mixture of different forms (from low- polymeric metal–organic coprecipitates to well crystalline phases with surface-attached OM) (Kleber et al., 2015). The phases known to occur frequently in podzols are poorly crystalline Fe oxides (ferrihydrite), poorly crystalline aluminosilicates (ITM: proto-imogolite, imogolite, allophane), Al- and Fe- complexes (coprecipitates). In a Haplic Podzol, Eusterhues et al. (2005) measured SSA values of 800 m2g-1 for ferrihydrite, 200 m2g-1 for goethite, 300 m2g-1 for Al-rich allophane, 2 -1 and 500 m g for amorphous silica (Si(OH)4). The respective SSA values were translated into spherical crystallite diameters of 7.5, 2, and 7.5 nm for allophane, ferrihydrite, and goethite, respectively, assuming mass densities of 2.7 g cm-3 for Al-rich allophane and 4 g cm-3 for the Fe oxides (Eusterhues et al., 2005). Likewise, Regelink et al. (2013) reported a size of 2–20 nm for Fe oxides, and of 20–150 nm for Al- and Si-rich particles observed in a Podzol. Consequently, OM impacts its own protection through its effect on mineral formation.

OM composition

In the literature, different OM compositions (plant-derived vs microbial-derived) and ages (younger or older) were measured in OMA

163

Part 3 Chapter 8 – Synthesis within B horizons of podzols (Rumpel et al., 2002, Eusterhues et al., 2007, Spielvogel et al., 2008, Schmidt et al., 2000, Rumpel et al., 2004, Rumpel et al., 2010). In the Vancouver podzols, the accumulating carbon seems to have a more vegetal signature. This is also the case at least for other hydromorphous podzols, where the SOM accumulating in the B horizons originates mainly from the DOM (roots are not present), and the microbial SOM degradation is inhibited (Buurman and Jongmans, 2005). However, in the literature, an increased microbial signature is often observed when moving from the coarse fraction to the fine clay or OMA fraction (Knicker, 2011, Kögel-Knabner et al., 2008b, Cotrufo et al., 2013).

Recent research on OMA brings new keys to explain the differences in mineral-associated-OM composition in different soils and depth. As reviewed by Kleber et al. (2015), a preferential adsorption of more hydrophobic (aromatic, alkyl), less decomposed, high-molecular weight, globular OM over that with contrary properties, is frequently reported in the literature. Due to the preferential adsorption/coprecipitation with minerals of some organic components, soil can be considered as a chromatographic column with the more surface-reactive compounds being retained and the more mobile ones being leached downward the profile with soil solution (Marschner et al., 2008, Neff and Asner, 2001, Kaiser and Kalbitz, 2012). Furthermore, it is well known that “fresh” DOM (high-affinity plant-derived compounds for example) may displace “old,” low- affinity adsorbed OM compounds from mineral surfaces which are subsequently released back into solution by competitive sorbate displacement (Kaiser and Kalbitz, 2012, Gu et al., 1996a, Gu et al., 1996b). Depending on the porosity, and therefore the soil texture and structure, OM can also bypass the chromatographic soil column if in conditions of preferential flow (Kaiser and Kalbitz, 2012).

Consequently, the OM composition in OMA at different depth and in different soil types depends on (i) the C input reaching the concerned horizon, transferred from the surface with soil solution and impacted or not by microbial decomposition and competitive sorbate displacement, depending on soil texture; (ii) the microbial transformation in situ; (iii) the contribution of root inputs. This very

164

Part 3 Chapter 8 – Synthesis complex combination of parameters explain the different OM composition (plant-derived vs microbial-derived) and ages (younger or older) measured in OMA in B horizons of podzols.

8.3.3.2 Dynamics in the illuvial B horizon

The illuvial B horizon is not static. Once formed, the accumulation and dynamics of the different phases depend on a complex equilibrium between: (1) the supply of DOM and aqueous Fe, Al, Si; and (2) the modifications in situ.

Accumulation of OM

The accumulation of organic matter results from the balance between OM input and degradation/dissolution (Buurman and Jongmans, 2005):

 OM input depends on aeration: In poorly aerated B horizons, roots and fungi does not have access, and the supply of OM exclusively comes from the illuvial DOC. In aerated B horizons, root-derived OM is the predominant OM supply.

 Degradation dynamics depends on the aeration and parent material: fast organic matter dynamics occurs on nutrient-rich parent materials (usually in the boreal zone), or well aerated B horizons. In such soils, there will be more accumulation of sesquioxides, and less of organic matter in the B horizon than non- boreal podzols. Slow organic matter dynamics, occurring on nutrient-poor parent materials or under hydromorphic circumstances, favors large organic matter accumulations in the B horizon, and a larger abundance of DOC-derived organic matter coatings. The latter group of podzols is called “non-boreal podzols” by Buurman and Jongmans (2005).

165

Part 3 Chapter 8 – Synthesis

Evolution of mineral phases

The evolution of the mineral phases in the B horizon depends on the input of ions coming from the surface, and the dissolution- reprecipitation cycles. Al and Fe have different properties, and their respective phases will evolve differently.

 Fe oxides are poorly soluble. Their dissolution depends principally on reduction (which is microbially-mediated) and organic complexation. Once in solution after a reducing event, Fe(II) is more soluble and can migrate, until reoxidized. Consequently, the Fe fate in the B horizon depend on the hydromorphic characteristics (and anoxic events). Few studies focus on the impact of redox cycles on the B horizon dynamics. Two types of hydromorphous podzols can be distinguished: 1) the ones with a perched water table. Fritsch et al. (2011) propose that after the accumulation of a first generation of organo-metallic complexes in B horizons of weakly differentiated podzols (first step), a second step follows with the accumulation at greater depth of a second generation of organo-metallic complexes in superimposed Bh and Bs (or BCs) horizons. Pore clogging by organo-metallic complexes reduces the soil permeability at the bottom of these podzols, which in return enables the settling of a perched water table during the rainfall season (Fritsch et al., 2011); 2) the ones with a fluctuating water table coming from below the accumulation horizon. This second type of hydromorphous podzol favor the leaching of elements (Al and Fe) after their release in the soil solution, and the formation of mainly a well-developped Bh illuvial horizon, depleted in Fe and Al (Legros, 2007). In general, hydromorphous podzols are more common in non-boreal podzols (sensu Buurman and Jongmans, 2005). The dissolution-reprecipitation of Fe phases during the presence- absence of the watertable (perched in case of pore clogging, or coming from below in case of the fluctuating water table), or in saturated micro-environments might have a crucial impact on Fe cycle in the B horizon. Redox fluctuations might also have an

166

Part 3 Chapter 8 – Synthesis

impact on podzol genesis in well-drained, seemingly oxic spodosols, because of anoxic microsites (Keiluweit et al., 2016).

 Al phases are more mobile, especially in acidic conditions or in presence of organic substances. The main factor impacting the Al phases seems to be the flow of incoming solution and its composition in complexing organic ligands.

The composition of the soil solution is modified while passing through the B horizons (adsorption/desorption processes, OM decomposition, etc.). The percolating soil solution then impacts minerals beneath the B horizon. In-situ modification have been proposed to be the main mechanism of ITM formation (Ugolini and Dahlgren, 1987) in the illuvial horizon. There, no acido-complexolysis occurs (because of the absence of organic complexing agents), but an ordinary hydrolysis does, due to dissolved bicarbonate linked to a high partial pressure of CO2.

Role of microorganisms

To our knowledge, few studies give information on the microbial population in the B horizon. Some scarce information exists on fungi, but almost none on the bacteria. Hyphal numbers have been shown to decrease abruptly at the interface between the E and the B horizons by Van Breemen et al. (2000a), but Koele et al. (2011) found fungi also in the deeper horizons. It seems that the redox status also impacts the microbial biomass, with only bacteria being capable to develop in B horizons of hydromorphous podzols, but few roots or fungi. Their impact on OM and Fe phases is even less known. Theoretically, they act on SOM by degradation. They might also have an important impact on minerals. In anoxia, bacteria with the ability to switch to other electron acceptors than oxygen (like Fe(III) and Mn(III)) can develop. Among them, the ones susceptible to have the bigger impact are iron reducing bacteria. They would induce a dissolution of secondary iron phases, a release of Fe(II), available for migration and re-precipitation of new phases. This process might have an important impact on illuvial B horizon modifications in podzols. The microbial

167

Part 3 Chapter 8 – Synthesis dynamics might also be different in podzolic Al-accumulation horizons, and Fe-accumulation horizon: (1) Al is toxic for microorganisms (at pH < 4.5-5) and Fe not, and (2) Fe is the final electron acceptor for reducing bacteria.

In the Vancouver chronosequence, the accumulation of OMA induced the formation of a cemented, hydromorphic illuvial horizons (monomorphic OM coating), experiencing episodic water stagnation and consequently anoxic periods. This character strongly impacted the microbial populations and activity. We measured a bacteria- dominated population.

168

Part 3 Chapter 8 – Synthesis

P4 and P5 - B horizons

Main processes:

 SOM compartment: accumulation of OM.

 Mineral compartment: precipitation/redissolution of secondary minerals

 Physico-chemical compartment: redox fluctuations

The accumulation of secondary minerals and (associated-)OM induce the development of an illuvial B horizon, enriched in OMA. In the case of the Vancouver podzols (P4 and P5), the B horizon presents a very different soil structure (cemented, clogged poral system), and experiences episodes of water stagnation. This is not the case in all podzols, and these hydromorphic properties will strongly impact the three compartments, notably through the redox status. The redox status will impact: (1) for OM: the relative importance of DOM and roots as dominant input of OM in the horizon (limited root colonization in hydromorphic conditions), and the degradation dynamics (slower degradation in anoxic conditions); (2) For the minerals: the relative importance of ligand- promoted/reductive dissolution and phase precipitation.

Main biota

Hydromorphic conditions will promote a bacteria-dominated population, and restrict the access to roots and fungi.

Main SOM protection mechanism:

The main constraints on SOM degradation are the adsorption and co-precipitation of OM with pedogenic secondary minerals. The environmental constraint might also be important, especially in hydromorphic conditions. In the Vancouver chronosequence, the degradability of C was the lowest in these horizons.

169

Accumulation of recalcitrant compounds

Inhibition of decomposers activity

Figure 8.3 - Schematic representation of the pedogenetic processes and protection mechanisms relevant in the podzol profiles: P4-330 yrs and P5-530 yrs.

170

Part 4 Chapter 9 – Microbial Fe(III) reduction

Part 4 - OMA dynamics in anoxic conditions

The part 3 revealed that the formation of OMA is a major control on long-term SOM stabilization in the Vancouver podzolic chronosequence. These OMA are localized essentially in the illuvial horizons, and maximal in the Bhs horizon. In chapter 5, we described the progressive development of a cemented horizon, inducing temporary water stagnation in the illuvial horizons of P4 and P5. The objective of this part is to study the impact of a reducing event on organo-mineral associations, in controlled lab conditions.

171

Part 4 Chapter 9 – Microbial Fe(III) reduction

Chapter 9 - Microbial Fe(III) reduction of organo-mineral associations along a podzolic soil chronosequence: impact on C release?6

Abstract

Iron (Fe) oxyhydroxides are known to stabilize organic matter (OM) in soils by forming organo-mineral associations (OMA). Under anoxic conditions, iron reducing bacteria dissolve Fe(III) phases and may potentially destabilize adsorbed or co-precipitated OM. We incubated three soil samples collected in contrasting spodic horizons along a soil chronosequence (from 270 to 550 years old) with Shewanella putrefaciens strain LMG 2279, a well-known iron reducing bacterium. With increasing age, podzolic differenciation and content of poorly crystalline Fe oxyhydroxides, both the rate and extent of microbial Fe(III) reduction increased. Despite the large dissolution of Fe(III) oxyhydroxides, no additional dissolved organic carbon (DOC) was measured as compared to un-inoculated experiments. However, a significant amount of DOC was released in all soil samples upon rewetting preceding the introduction of iron reducing bacteria. OM, in particular the labile organic carbon pool in the oldest soils, likely favors the persistance of poorly crystalline Fe(III) oxyhydroxides (ferrihydrite), and, as such, contributes to increase the extent and rate of microbial Fe(III) reduction. The evolution of the mineral-associated C fraction and its impact on the mobility of Fe (via microbial Fe(III) reduction) as podzolization proceeds is a crucial process in soil development and the resulting OM dynamics.

6 In preparation: Vermeire M.-L., Bonneville S., Stenuit B., Delvaux B., Cornelis J.T., Microbial Fe(III) reduction of organo-mineral associations along a podzolic soil chronosequence: impact on C release? 172

Part 4 Chapter 9 – Microbial Fe(III) reduction

9.1 Introduction

Soil organic carbon (SOC) is the largest terrestrial pool of carbon, hence influences greenhouse gas emissions, soil fertility and plant productivity at the global scale (Lehmann and Kleber, 2015). Yet, the processes controlling the fate of SOC are poorly understood and represent the main source of uncertainty in for climate projections (Friedlingstein et al., 2014). The formation of OMA, in particular with Fe(III) oxyhydroxides, is a major pathway for OM stabilization in soils (Kleber et al., 2015, Kögel-Knabner et al., 2008b, Baldock and Skjemstad, 2000). Due to their large specific surface area and specific chemistry, Fe(III) phases can indeed adsorb and also coprecipitate large amount of OM (Eusterhues et al., 2014b, Bonneville et al., 2009).

When anoxic conditions develop, as in water-logged soils, Fe(III) becomes a major electron acceptor of microbial metabolic pathways. Although Fe(III) oxyhydroxides can undergo abiotic reductive dissolution (Melton et al., 2014), it is recognized that microbial Fe(III) reduction primarily controls iron redox chemistry. The ubiquity of dissimilatory iron-reducing bacteria (DIRB) establishes this metabolic process as one of the most significant biogeochemical process in soils (Weber et al., 2006, Lovley, 2013, Kappler and Straub, 2005, Nealson and Myers, 1992). Few studies investigate the impact of microbial Fe(III) reduction on the fate of the Fe(III) oxide-OM associations (Zachara et al., 2002, Kleber et al., 2015, Keiluweit et al., 2016). To our knowledge, the only four studies investigating this thematic have been focused on fresh, synthetic OMA, but not on soils per se (Pédrot et al., 2011, Shimizu et al., 2013, Eusterhues et al., 2014b, Poggenburg et al., 2016). No clear picture has yet emerged regarding the controls of Fe bioavailability and the fate of OM in OMA in soils. Natural Fe oxyhydroxides differ in terms of crystallinity and chemical reactivity from those synthesized in the laboratory due to interactions with complex OM and/or dissolved inorganic ions during their formation and life cycle (Fritzsche et al., 2012, Perret et al., 2000, Gunnars et al., 2002, Hiemstra et al., 2010a). These interactions may either impede or enhance reductive dissolution, by (i) controlling the structure and reactivity of natural Fe (hydr)oxides (Zachara et al., 2002,

173

Part 4 Chapter 9 – Microbial Fe(III) reduction von Lützow et al., 2006, Eusterhues et al., 2008, Cornell and Schwertmann, 2003), (ii) acting as electron shuttles with quinones and flavins being the most redox active compounds (Roden et al., 2010, Hansel et al., 2004, Jiang and Kappler, 2008, Aeschbacher et al., 2009) (iii) or by surface site passivation (Amstaetter et al., 2012).

Due to the crucial role of OMA in podzolization, podzolic soils have received much attention to understand SOC dynamics and stabilization by mineral phases (Schulze et al., 2009, Schmidt et al., 2000, Grand and Lavkulich, 2013, Jones et al., 2015). Podzol morphology typically involves an ash-grey weathered eluvial horizon (E), overlying a dark illuvial horizon (B) enriched in OM, Al and Fe (Lundström et al., 2000a). If formation and downward transport of soluble OM-Fe and -Al complexes are now accepted as the main mechanisms of eluviation (Lundström et al., 2000a, Lundström et al., 2000b, Jansen et al., 2005), the processes governing the immobilization of Al, Fe and OM in the illuvial horizon are still debated. The precipitation of the organo-metallic complexes due to a decrease of the C/metal ratio might be caused by either (i) adsorption and precipitation of metals or (ii) microbial degradation of the organic moiety of the organo-metallic complexes (Buurman and Jongmans, 2005, Lundström et al., 2000b). Other processes may also contribute to that precipitation depending on environmental conditions, soil composition and soil chemistry (Jansen et al., 2005, Sauer et al., 2007). In this respect, very few studies considered the influence of redox mechanisms and cycles on iron mobility in podzols (Sauer et al., 2007). The occurrence of reducing conditions have been documented in various types of podzols (Vodyanitskii et al., 2006, Kanev, 2011). These conditions are due to pore clogging by organo-metallic complexes which reduces soil permeability and forms perched watertable during rainfall events (Fritsch et al., 2011, Do Nascimento et al., 2004, Montes et al., 2011, Buurman and Jongmans, 2005). The accumulation of OMA and the occurrence of redox cycles in clogged illuvial B horizons (Bhs, Bh, Bs), make the spodic horizon a pertinent model to study the impact of dissimilatory iron reduction (DIR) on natural Fe oxyhydroxides and their associated OM.

174

Part 4 Chapter 9 – Microbial Fe(III) reduction

Our objective is to evaluate how dissimilatory iron reduction (DIR) can affect the Fe(III) oxyhydroxides-organic matter associations and to which degree the dissolution of Fe(III) solid phases results in the release in solution of associated OM. We implemented experimentations on spodic horizons collected in podzolic soils of increasing age and contrasting contents of Fe oxyhydroxides and OM. We characterized their Fe and C pools by using selective extractions. We conducted anoxic incubations in the presence of Shewanella putrefaciens, a facultative anaerobe that substitutes Fe(III) for O2 as the terminal electron acceptor to gain energy for growth and metabolism in anoxic conditions (Glasauer et al., 2003, Nealson and Saffarini, 1994).

9.2 Materials and methods

9.2.1 Soil samples

Soil samples were collected from a chronosequence of podzolic soils under a Sitka spruce (Picea sitchensis) forest, located near Cox Bay, west coast of Vancouver Island, British Columbia (latitude 49° 6'N, longitude 125° 52'W) described elsewhere (Cornelis et al., 2014a, Vermeire et al., 2016). The soils have developed from sandy beach deposits, which are emerging at a rate of 0.26m per year (Singleton and Lavkulich, 1987). The ages of the deposits were determined by dendrochronology and geomorphology; they range from 0 to 530 years (Vermeire et al., 2016). A progressive deepening and differentiation of genetic horizons is observed as podzolization proceeds (Cornelis et al., 2014a). Soil samples used for our incubations were collected from the three oldest pedons (P), i.e. P3 (270 years- old), P4 (330 years-old) and P5 (530 years-old), in the P3-Bh horizon and P4-, P5-Bhs horizons. The soils contain large amounts of quartz with other minerals such as sodic feldspars, amphibole (hornblende), pyroxene (augite), kaolinite, micas (illite) and chlorite (Vermeire et al., 2016). Further soil characteristics are given in Table 9.1.

175

Table 9.1 - Major soil characteristics: WRB classification, pH in water, cationic exchange capacity (CEC), exchangeable cations, base saturation (%BS), particle-size distribution and total organic carbon (TOC) content

Exchangeable cations Particle size Profile Age Horizon Depth WRB* pH CEC %BS (cmolc kg-1) distribution (%) TOC cm (water) cmolc/kg Ca2+ K+ Mg2+ Na+ Sand Silt Clay g/kg P3 270 Bh 7-23 AP 5.1 5.94 0.59 0.04 0.41 0.13 19.76 97.2 1.7 1.1 16.17 P4 330 Bhs 17-17.5 PP 5.5 15.18 0.13 0.04 0.07 0.06 1.96 90.0 6.8 3.2 22.99 P5 530 Bhs 9.5-10 PP 4.5 23.35 0.24 0.08 0.13 0.07 2.23 40.48 * WRB classification: DC stands for Dystric Cambisol, AP for Albic Podzol and PP for Placic Podzol

176

Part 4 Chapter 9 – Microbial Fe(III) reduction

During sampling, we observed a progressive cementation/induration of the spodic illuvial B horizons, particularly in P4 and P5 Bhs, due to the accumulation of Fe and OM. Because of the high mean annual precipitation (~3200 mm), episodic waterlogging alternating with a rapid return to oxic conditions occurs particularly in P4 and P5 Bhs.

9.2.2 Chemical extractions: Fe and C pools in soil samples

Total major elements contents, pH, cationic exchange capacity, particle size distribution, exchangeable cations and total organic carbon were measured as described by Vermeire et al. (2016). The Fe pools in the soil samples were extracted separately with (i) sodium pyrophosphate, (ii) ammonium oxalate-oxalic acid in dark condition and (iii) dithionite-citrate-bicarbonate (DCB) according to Bascomb (1968), Blakemore et al. (1987) and Mehra and Jackson (1960), respectively, and quantified by ICP–AES (Thermo Scientific, iCAP 6000). The pyrophosphate-extractable Fe (Fep) is used as an estimator of organo-Fe complexe content. As Na-pyrophosphate at pH 10 is a dispersing agent, a part of the Fep may include Fe-oxide nanoparticles even though vigorous centrifugation was applied to limit this contribution (Jeanroy and Guillet, 1981). The oxalate extraction (Feo) solubilizes both the organo-Fe complexes and ferrihydrite (and eventually lepidocrocite) (Poulton and Canfield, 2005). The DCB- extractable Fe (Fed) estimates the content of free iron that includes organo-Fe complexes, poorly crystalline (ferrihydrite) and crystalline (goethite and hematite) Fe phases.

The different soil carbon pools were estimated following the protocol of Mikutta et al. (2006). The amount of stable carbon (S-OC) was quantified as the amount of carbon left after oxidation of the labile organic carbon (L-OC) with NaOCl (Kleber et al., 2005). 3 g of air- dried soil sample was reacted three times with 30 mL of 6 wt% NaOCl adjusted to pH 8.0 during 6h at 25°C. Samples were then washed twice with 30 mL 1 M NaCl and then with deionized water. The labile OC (L- OC) is calculated by subtracting the stable OC fraction (S-OC, carbon left after treatment with NaOCl) from the total OC. The NaOCl-treated

177

Part 4 Chapter 9 – Microbial Fe(III) reduction samples were then exposed to HF treatment in order to dissolve minerals and associated organic matter. 2.25 g of NaOCl-treated samples were exposed four times consecutively to 15 mL 10% HF during 2h in order to dissolve minerals and associated organic matter and then washed five times with 15 mL deionized water. The remaining organic carbon after HF extraction was defined as recalcitrant carbon (R-OC). The HF-extractable carbon, i.e., mineral- protected carbon (MP-OC) was calculated by subtracting the R-OC fraction from the stable C fraction (S-OC). The C content in the solid fractions was measured with a FLASH 2000 Analyzer (ThermoFisher Scientific). The soil samples were not sterilized as autoclaving may modify the mineralogy and the OMA.

9.2.3 Microbial incubations

Shewanella putrefaciens (strain LMG 2279) was provided by the Belgian Coordinated Collection of Microorganism (BCCM, Brussels, Belgium) and was routinely grown under aerobic conditions at 25°C on agar Luria–Bertani (LB, 15 g/l tryptone water, 15 g/l agar, 5 g/l NaCl, 5 g/l yeast) plates and in liquid LB medium on a rotary shaker set at 120 rpm. A single colony from plates was used to inoculate precultures grown aerobically to the late exponential phase (~ 24 h) in 250-mL Erlenmeyer containing 50 mL LB medium. Precultures were used to inoculate at 0.6% (v/v) 500-mL flasks containing 150 mL LB medium, and the main cultures were grown to the mid-exponential phase (~ 12 h). Cultures were harvested by centrifugation at 8,500 g for 15 min at 4 °C and cells were then washed three times with a sterile saline solution “S” (composed of 28 mM NH4Cl, 1 mM CaCl2.2H2O, at pH 7.5). Washed cells were concentrated by centrifugation for use in the incubations and cell density was estimated by the optical density (OD) measurements at 600 nm using a conversion factor for Escherichia coli 8 -1 (i.e. one unit of OD600 is equal to 8 x 10 cells mL ) (Volkmer and Heinemann, 2011). Cell density was also determined by measuring dry weight of cell suspensions taking into account a mass of 7 x 10-13 g per S. putrefaciens cell (Claessens et al., 2006).

178

Part 4 Chapter 9 – Microbial Fe(III) reduction

All incubations were conducted during 96h in 160-mL sterile serum bottles sealed with airtight butyl rubber stoppers and aluminum crimp caps, at 25 °C, in the dark, on a rotary shaker (180 rpm). Before starting the inoculation (and sampling - t0), 3g of soil (<2mm) were resuspended in 120 mL sterile saline solution and let to equilibrate for 7h, until pH equilibrium was reached. The suspensions were not buffered because phosphate and carbonate can interact with Fe, and form aqueous complexes or ferrous solids (Zachara et al., 2001, Wang et al., 2015), while organic buffers interfere within dissolved organic + carbon (DOC) measurements. Except for nitrogen (0.003 mol NH4 in saline solution), no other nutrient was added out (e.g., no external organic electron donors, phosphate or vitamins). The bottles were purged with sterile N2 for 1h to reach anoxia. All experiments were carried out in triplicates.

A first series of experiments was conducted to compare microbial Fe(III) reduction activities of S. putrefaciens at three different stages of podzolization (P3-270, P4-330 and P5-530 years) and contrasting Fe and OM contents. Suspensions of 3 g of soil samples in 120 mL sterile saline solution were inoculated with S. putrefaciens (B) under anoxic (A) conditions. This first series of experiments will be referred to as “AB” treatments (i.e. P3AB, P4AB and P5AB).

A second series of experiments was dedicated to control experiments. The following combinations were tested: (i) 3 g of soil samples in 120mL sterile saline solution (S) under anoxic conditions (A) without addition of S. putrefaciens to assess the contribution of autochthonous microorganisms on the release of dissolved Fe(II) and DOC (hereafter referred to as P3AS, P4AS and P5AS). (ii) 3 g of soil samples in 120mL sterile saline solution in oxic conditions (O) with S. putrefaciens (only for P5 sample – hereafter referred to as P5OB), to assess the contribution of anoxic conditions on the release of dissolved Fe(II) and DOC. (iii) S. putrefaciens alone (without soil) in the saline solution in anoxic conditions (hereafter referred to as “BB” treatment), to assess the contribution of S. putrefaciens on the release of DOC.

The initial cell density of each incubations with S. putrefaciens (“B” treatments) was adjusted to 3.25  107 cell mL-1 (1.3 x 109 cell g-1

179

Part 4 Chapter 9 – Microbial Fe(III) reduction soil). Liquid aliquots were retrieved at 9 time steps, to monitor the concentration of dissolved ferrous iron, total dissolved iron and DOC. At each sampling point, three aliquots of 1mL were filtered through a 0.2 µm pore size PTFE membrane. Two aliquots (for total dissolved iron and dissolved Fe(II) measurements, respectively) were filtered directly into 250 µL HCl 0.5 M, while the third aliquot was preserved into a glass vial for DOC measurement. The samples were conserved at 4°C in the dark for maximum two days until further processing.

Dissolved ferrous and total Fe concentrations were measured by spectrophotometry following the ferrozine method (Viollier et al., 2000) using a Genesys 10S VIS spectrophotometer (Thermo Fischer Scientific). DOC concentrations were measured on a Shimadzu TOC-L analyzer. The accuracy of DOC concentration measurements was controlled by using freshly prepared standard solutions of potassium hydrogen phthalate every 6 to 9 samples. At the end of the experiment (96h), pH was measured, and the total concentrations in Al, Fe, K, Mg, Mn, P and S were determined by ICP-AES (Thermo Fischer Scientific, iCAP 6000).

9.2.4 Scanning Electron Microscopy (SEM)

Samples for SEM were prepared according to Schädler et al. (2008) immediately after the end of incubation (96h) in order to avoid artifacts (cell shrinkage during dehydration, surface alteration or formation of secondary minerals). Briefly, suspension were filtered through a 0.2-µm-pore-size PTFE filters. The membranes (with the soil material) were immerged in a 2.5% glutaraldehyde solution for 1h to chemically fix the biological structures, then rinsed three times with a Dulbecco’s phosphate buffered solution, and finally rinsed twice with milliQ water. Subsequently the samples were dehydrated by successive resuspensions in isopropanol at 10%, 30%, 50%, 70%, 80%, 90%, 96%, 100%, vol/vol. Imagery was performed with an Ultra55 FEG- SEM (Zeiss), fitted with a secondary electron detector. Morphologic analysis was performed at 3.00 or 5.00 kV accelerating voltage. The elemental analysis was carried out by Energy Dispersive X-ray

180

Part 4 Chapter 9 – Microbial Fe(III) reduction spectroscopy (EDX) using a silicon drift detector (Quantax system, Bruker) at an acceleration voltage of 3 or 15 kV.

9.3 Results and discussion

9.3.1 Characterization of Fe and C pools

With increasing development of podzolic soil, a gradual accumulation of iron phases and organic matter is observed in the illuvial B horizons from P3 to P5 (Figure 9.1). Total Fe content is lowest in P3-Bh (~21 g Fe kg-1) while it reach ~38 g Fe kg-1 in P4-, P5-Bhs horizons. The content of free iron (Fed) is approximately 10 times smaller in P3-Bh (~2.6 g.kg-1) than in P4- and P5-Bhs, i.e. ~21.6 and ~26.1 g.kg-1, respectively. Though P4- and P5-Bhs horizons exhibit similar total and free Fe contents, they differ in their Fe pool distribution (Figure 9.1). In P4, only 7.8% of the free Fe (Fed) is present in poorly crystalline Fe oxides, while they represent 53.2% in P5. In contrast, the crystalline Fe oxides represents 46% of the free Fe (Fed) in P4-Bhs, but is negligible in P5-Bhs. In both P4- and P5-Bhs horizons, about half of the free Fe (Fed) -46% and 48% respectively-, can be considered as Fe bound to organic ligands to form organo-metallic complexes.

Similarly to Fe pools, a progressive increase of the total OC content is observed along the chronosequence from 16 to 23 and 40 g C kg-1 in P3-Bh, P4- and P5-Bhs horizons, respectively (Figure 9.1). The relative abundance of the C fractions differs between the three soil horizons. In P3-Bh, 76 % of the total OC is labile with only 19% as mineral-protected (MP-OC) and 5% as recalcitrant OC pools. In P4- and P5-Bhs, the MP-OC pools exhibit similar amounts (13 and 16 g C kg-1), but represent 57 % of the total OC in P4-Bhs, for only 39 % in P5-Bhs. There is much more labile OC in P5-Bhs compared to P4-Bhs (17 against 8 g C kg-1, respectively), representing 41% of the total OC in P5.

181

A - Fe pools B - C pools 45 45 40 40 35 35 30 30 25 25 20

20 [g/kg]C Fe [g/kg] Fe 15 15 10 10 5 5 0 0 P3 Bh P4 Bhs P5 Bhs P3 Bh P4 Bhs P5 Bhs Fet Fed Cr-Fe PCr-Fe Fep Total-OC R-OC L-OC MP-OC Figure 9.1 – (A) Fe pools in the three soil samples (see section 9.2.2 for details about chemical extractions). Total Fe (Fet), free Fe (Fed), Fe in crystalline oxyhydroxides (Cr-Fe = Fed - Feo), Fe in poorly crystalline Fe oxyhydroxides (PCr- Fe = Feo - Fep) and the Fe bound to organic complexes (Fep). (B) C pools in the soil samples: total, recalcitrant (R-OC), labile (L-OC) and mineral-protected (MP-OC). 182

A. B.

Figure 9.2 – SEM micrographs of a quartz grain from the P5 soil sample showing a bacteria (plain arrow), and a OM- Fe patch at the surface (dotted arrow) (A.), EDX mapping of the zone (B.) for Fe, C, O and Si.

183

A. B. A.

Figure 9.3 – SEM micrographs performed on the spodic P5-Bhs. (A.) Quartz grain. (B.) Zoom of the crust present on the quartz grain surface.

184

Part 4 Chapter 9 – Microbial Fe(III) reduction

As shown previously (Poggenburg et al., 2016, Fritzsche et al., 2012, Jiang and Kappler, 2008) natural Fe oxyhydroxides occur in association with organic matter in patches on soil mineral particles (in the vast majority – quartz grains) as highlighted by concomitant presence of Fe and C in our elemental maps by EDS (Figure 9.2 and 9.3). In all of our elemental analysis, when present, the Fe is always in close association with organic matter.

9.3.2 Microbial Fe(III) reduction: rate and extent

In the three soil samples, the presence of S. putrefaciens and anoxic conditions (AB treatments) induced an increase in Fe(II) concentration compared to control experiments without added cells (AS treatments, Figure 9.4). At the end of the incubation period, there was 2.8, 5.4 and 1.8 times more Fe(II) released in P3AB, P4AB and P5AB experiments, respectively, compared to the corresponding controls free of added cells (i.e., P3AS, P4AS, P5AS). Moreover, the extent of microbial Fe(III) reduction markedly differed between the three pedons (Figure 9.4): 1.9 and 2.6 µmol.L-1.g-1 of Fe(II) are produced at the end of the incubation period in P3 and P4, while it reaches 16.5 µmol.L-1.g-1 in P5AB (Table 9.2 and Figure 9.4). Overall, there is a strong positive correlation between the content of poorly crystalline Fe oxyhydroxides (Table 9.2) and the amount of dissolved Fe(II) released due to microbial Fe(III) reduction (r2= 0.99). In contrast, the amount of crystalline Fe oxide in soil is not strongly correlated with the extent of microbial Fe(III) reduction (r2= 0.27). Previous studies have shown that poorly crystalline Fe phases (ferrihydrites) are the main source of bioavailable Fe(III) for iron reducing bacteria (Fredrickson et al., 1998, Glasauer et al., 2003, Hansel et al., 2004, Bonneville et al., 2004). Even if these microorganisms can utilize crystalline Fe(III) oxides such as lepidocrocite, goethite and hematite (Bonneville et al., 2009, Kukkadapu et al., 2001) and also structural Fe(III) bound in clay minerals (Kostka et al., 1996, Dong et al., 2009), when grown in nutrient-limited conditions as those typically found in subsurface environments (and as in our experiments), only poorly crystalline Fe(III) phases are reduced in significant amounts (Glasauer et al., 2003).

185

Figure 9.4 - Concentrations of dissolved Fe(II) (D-Fe(II)), total dissolved Fe (D-Fet), and DOC as a function of time during anaerobic (A) and aerobic (O) incubations in presence (B) or not (S) of S. putrefaciens, for P3-Bh, P4- and P5-Bhs soil samples. 186

Part 4 Chapter 9 – Microbial Fe(III) reduction

Figure 9.4 – second part of the legend - No dissolved Fe(II) or Fet were detected in solution in the bacteria control experiment (BB - data not shown). DOC concentrations in the control saline solution without soil and in the BB control experiment were subtracted from the respective values measured in incubations with soil samples. T0 corresponds to the end of the 7h equilibration period. Error bars are standard deviations from the arithmetic mean of triplicate measurements. Note that the dissolved Fe concentrations scale is 10 times larger in P5.

Surprisingly, there is a poor correlation (r2=0.497) between the amount of Fe bound with OM (measured by pyrophosphate extraction) and the extent of microbial Fe(III) reduction (Table 9.2). This could either mean that the Fe present as organo-metal complexes is already in a reduced form, or chelated in organic complexes of relatively high stability and therefore not bioavailable for S. putrefaciens (Haas and Dichristina, 2002). An argument in favor of the first hypothesis is that a substantial amount of DOC and dissolved Fe(II) is measured in all control experiments (without reduction by S. putrefaciens). In all incubations, except for experiment P4AS, total dissolved Fe and Fe(II) concentrations are very similar, meaning that Fe(II) is the main form of dissolved iron (between 80 and 95%, Figure 9.4) probably stabilized by OM and solubilized during wetting of the soil sample (Lovley et al., 1998, Kappler and Straub, 2005, Pédrot et al., 2011). Though Fe(III) represents a large fraction of total dissolved Fe (e.g. in P4AS, ~54% Fe(III)), this does not imply that this dissolved Fe is bioavailable. Haas and Dichristina (2002) have shown that reduction rates of Fe(III)-organic complexes is inversely correlated with the thermodynamic stability constants of these complexes, implying that chemical speciation governs Fe(III) bioavailability of these complexes. As we did not add any electron donors in our experiments, the dissolved organic carbon pool, which is significant (from 67 µmol L-1 in P4AS to 264 µmol L-1 in P5AB, at t0), is the likely source of electron donor fueling microbial Fe(III) reduction. As such, its concentration and bioavailability can influence the extent of dissimilatory iron reduction (Röling et al., 2007).

187

Part 4 Chapter 9 – Microbial Fe(III) reduction

Regarding reduction kinetics, Michaelis-Menten formalism successfully describes the dependence of microbial Fe(III) reduction rates on the initial concentration of Fe(III) substrate, for a range of pure Fe(III) oxyhydroxides (Bonneville et al., 2004, Bonneville et al., 2006) and natural sediments (Hyacinthe et al., 2006). As suggested above, poorly crystalline Fe oxyhydroxides are preferentially used by S. putrefaciens as electron acceptors in our experiments. Assuming that this pool of bioavailable Fe(III) exhibits relatively similar chemical and physical properties accross the chronosequence (Figure 9.1), we could replace the initial microbial Fe(III) reduction rates (Table 9.2) into a Michaelis-Menten kinetic framework. In the three soil samples, the bioavailable Fe(III) pool amounts to 0.11, 0.25 and up to 2.07 mM for P3, P4 and P5 respectively (Table 9.2). These values are smaller than the half-saturation constant values (i.e. Km) reported for 2/6-line ferrihydrite (0.7 to 3 mM in Bonneville et al. (2006) and for natural sediments (0.3 to 11 mM in Hyacinthe et al. (2006)) incubated with S. putrefaciens. As the relative magnitudes of Km and the actual concentration of bioavailable Fe(III) determine whether or not microbial Fe(III) reduction is limited by Fe(III), the iron reduction rate is independent of the availability of bioavailable Fe(III) only when the concentration of substrate is significantly larger than Km. On this account, the iron reducing bacteria in our three soils samples - in particular in P3-Bh and P4-Bhs - should be limited by the availability of reactive Fe(III). Limitation should be less severe in P5-Bhs, which contains the highest concentration of reactive Fe(III).

The rate of microbial Fe(III) reduction in our experiments varies over several orders of magnitude (0.03, 0.22 and 0.96 fmol.L-1.h-1.cell- 1 in P3, P4 and P5, respectively, Table 9.2 and Figure 9.6). They are linearly related to the amount of poorly crystalline Fe(III) phases (r²=0,99). These rates are in the range of values observed by Bonneville et al. (2004) for aqueous Fe(III) citrate reduction (1 fmol h-1 cell-1 for an initial Fe(III) concentration ~2 mM) but much faster than for 2-lines/6- lines ferrihydrite (~2×10-2 fmol h-1 cell-1). Our reduction rates are close to those reported in Fritzsche et al. (2012) with 1.9 fmol.h-1.cell-1 for a similar cell density (of Geobacter sulfurreducens) and initial ferrihydrite colloids concentration than in our incubations.

188

50 700

] 1

- AS

45 ] 1 AB - 600 40 OB 35 500 30 400 25 20 300

15 200 Fe(II) concentration [µmol.l concentration Fe(II) - 10 100

Final DOC concentration [µmol.l DOC Finalconcentration Final D Final 0 0 P3 P4 P5 P3 P4 P5 Figure 9.5 – Final dissolved Fe(II) (D-Fe(II)) and DOC concentrations in the saline solution at the end of the incubations (96h), for the three soils (P3, P4 and P5) with (B) or without adding the bacteria (S) and in anaerobic (A) or oxic (O) conditions. Error bars : standard deviation from the arithmetic mean of triplicate measurements.

189

Table 9.2 – Microbial Fe(III) reduction in the three soil samples: initial reduction rates, cell-normalized rates, coefficients of correlation (R2), final dissolved Fe(II) concentration (96h). Properties of the soil samples: Fe in poorly crystalline (PCr-Fe = Feo-Fep) and crystalline (Cr-Fe = Fed-Feo) oxyhydroxides, and in organo-metal complexes (Fep), mineral-protected C concentrations (MP-OC), and C/Fe (mol:mol) ratios of OMA (MP-OC/PCr-Fe).

Sample Reduction by S. putrefaciens* Properties of the solid phase Initial red. rates Fe(II) released PCr-Fe Cr-Fe Fep MP-OC MP-OC/PCr-Fe µM/h fmol/l.h.cell R² µM mM mM mM mM P3 0,05 0,03 0,82 1,94 0,11 0,07 0,21 0,56 5,05 P4 0,29 0,22 0,96 2,64 0,25 1,49 1,49 9,16 36,39 P5 1,28 0,96 0,98 16,51 2,07 -0,03 1,86 11,03 5,32

*The extent of microbial Fe(III) reduction is the difference at each time step between the dissolved Fe(II) concentration in AS treatment (without bacteria added) and in the AB treatment (with S. putrefaciens added), for each soil sample. Production of Fe(II) by S. putrefaciens increased linearly with time during at least the first 6 h of incubation (figure 9.5), for the three soils. Initial reduction rates were calculated from linear least square regressions of time vs. Fe(II) increase during the first 6 hours (assuming that the reduction of Fe(III) initially followed a pseudo first-order rate kinetics).

190

Part 4 Chapter 9 – Microbial Fe(III) reduction

However, direct comparisons of reduction rates between studies are difficult, as several factors may influence kinetics e.g., cell density (Bonneville et al., 2006), bacteria strain (Lin et al., 2007), electron donor availability and also pH (Fredrickson et al., 1998, Zachara et al., 2002, Glasauer et al., 2003). Many studies on microbial Fe(III) reduction use buffers to constrain the pH ~7. Here, in absence of buffer, the pH of our experiments ranged between 4.5 and 5.8 (the lowest pH being observed in P5AB). Kostka et al. (1996) found that, for S. putrefaciens strain MR-1, the optimum pH range was between 5 to 6, with the bacteria being physiologically active for pH>4. Thus, we feel confident that, in our experiments, pH variations between experiments are not greatly affecting Fe(III) reduction kinetics.

10 9 P3 P4 P5 8

] 1

- 6 S. putrefaciens putrefaciens S. 5

[µmol.l 4 3 2

1 Fe(II) released by by released Fe(II) 0 0 2 4 6 Time [h] Figure 9.6 – Linear increase during the 6 first hours of incubation of the dissolved Fe(II) concentration due to reduction of Fe(III) by S. putrefaciens for the three soils (each point correspond to the difference at each time step between dissolved Fe(II) in the control experiments (AS) and in the treatments with bacteria added (AB)). Initial reduction rates were calculated from linear least square regressions of time vs. production of Fe(II) by S. putrefaciens during the first 6 hours (assuming that the reduction of Fe(III) initially followed a pseudo first-order rate kinetics).

191

Part 4 Chapter 9 – Microbial Fe(III) reduction

9.3.3 Influence of microbial Fe(III) reduction on OM release

The solubilization of OM under reducing/anoxic condition is a matter of debate. Some studies link the release of DOC from soils to anoxia (Grybos et al., 2009, Hagedorn et al., 2000, Buettner et al., 2014, Kalbitz et al., 2000) while others found no evidence of DOC production in organic-rich soil horizons undergoing anoxic episodes (Moore and Dalva, 2001, Fiedler and Kalbitz, 2003, Henneberry et al., 2012). Our results tend to support the latter assertion since no significant additional release of DOC was measured when S. putrefaciens was present (experiments -AB in Figure 9.4) as compared to control experiments, free of added cells (experiments -AS in Figure 9.4). In oxic conditions (only tested for P5 sample, i.e. P5OB in Figure 9.4), the DOC concentrations were slightly lower than under anoxia, with or without cells. However the difference is within the experimental error and thus not significant. In fact, in all experiments, the DOC release started directly after exposure to sterile saline solution, and reached 82, 67 and 265 µmol C L-1 g-1 for P3AB, P4AB and P5AB respectively at the end of 7h equilibration period (i.e. t0, the time of cell addition). Following the introduction of cells, the first 6 hours showed a linear increase of DOC concentration before leveling off at ~170 and 100 µmol C L-1 g-1 for P3AB and P4AB while in P5AB experiment, DOC concentration kept increasing to reach 611 µmol C L- 1 g-1 at the end of experiment. The initial C release rates (calculated by linear regression of [DOC] over the first 6h of incubation) reached 6, 5 and 17 µmol L-1 h-1, respectively for P3AB, P4AB and P5AB, while they were systematically slower without cells (3, 3 and 15 µmol L-1 h-1, for P3AS, P4AS and P5AS respectively), however within the experimental error.

Several hypotheses were proposed to explain the dynamic interactions between microbial Fe(III) reduction and DOC release. A first one proposes that the reductive dissolution of Fe(III) oxyhydroxides induces the release of surface-sorbed OM (Hedges and Keil, 1995, Grybos et al., 2009, Adhikari et al., 2016), or OM encapsulated in colloidal aggregates “cemented” by Fe(III) phases

192

Part 4 Chapter 9 – Microbial Fe(III) reduction

(Henderson et al., 2012, Buettner et al., 2014). Our results do not support this view as active microbial Fe(III) reduction did not lead to any significant additional release of OM compared to controls experiments. Henneberry et al. (2012) also found no evidence that exposure to Fe(III) reducing conditions lead to a release of OM from freshly precipitated OM-Fe(III) flocs. We can assume that the release of OM due to the dissolution of the adsorbent Fe(III) phases is too small to be measured relative to the large DOC concentration already present in solution. Indeed, in our experiments, the DOC concentrations are, at least, one order of magnitude larger than dissolved Fe(II) concentrations (Figure 9.4). However, poorly crystalline Fe(III) phases (i.e. ferrihydrite) are known for their large specific surface area i.e., 600 m2.g-1 (Kleber et al., 2015) up to 800 m2.g- 1 in an haplic podzol (Eusterhues et al., 2005) and for their capacity to trap carbon during co-precipitation (Mikutta et al., 2008). Considering a surface adsorption capacity of 0.52 mg C m-², and a coprecipitation C loading of 1.1 mg C m-2 (Eusterhues et al., 2014b), 16.5 µmol of Fe(II) released by S. putrefaciens in the P5 soil corresponds to 1.76 mg of ferrihydrite (Fe(OH)3) dissolved and 1.06 m² of surface lost. This would represent 46 µmol of adsorbed C, and 97 µmol of coprecipitated C. These values are well above the precision of our DOC measurements. If presently released concomitantly to microbial Fe(III) reduction, a variation of this scale in DOC concentration would have been measured. Overall, our results and the above calculations suggests that S. putrefaciens is able to reduce Fe(III) from poorly crystalline Fe phases, without destabilizing adsorbed or coprecipitated OM.

Some studies have linked the increase of pH due to microbial Fe(III) reduction which consumes protons, and the solubilization of OM (Buettner et al., 2014, Grybos et al., 2007, Grybos et al., 2009). The increase of pH induces the decrease of the positive surface charge of oxides mirrored by an increase of the negative charge of organic molecules (Avena and Koopal, 1999), hence decreasing the adsorption of organic anions on oxide surfaces. In addition, this process of charge inversion the development of “less positive” charge on colloid surfaces also favors dispersion (Bunn et al., 2002, Thompson et al., 2006b, Thompson et al., 2006a).

193

Table 9.3 - Composition and pH of the liquid phase at the end of the incubations, measured by ICP-AES (Thermo Scientific, iCAP 6000). Standard deviation: from the arithmetic mean of triplicate measurements.

Treatment Al Fe K Mg Mn P S Final pH µM µM µM µM µM µM µM µM BB

* < DL: DL below detection limit

194

Part 4 Chapter 9 – Microbial Fe(III) reduction

This hypothesis can be ruled out in our experiments as no increase of pH was observed (Table 9.3), probably due to significant soil buffering capacity (Thompson et al., 2006b, Almeida et al., 2015).

In soils, bacteria can also influence the fate of DOC via a series of biogeochemical processes: (i) production of soluble organic metabolites (Kalbitz et al., 2000, Grybos et al., 2009) or (ii) release of DOC though the degradation of soil organic matter or fermentation (Moore and Dalva, 2001, Röling et al., 2007). These processes could lead to an enrichment of DOC in reducing conditions where OM degradation is slowed compared to oxic environments (Keiluweit et al., 2016, Fiedler and Kalbitz, 2003, Moore and Dalva, 2001). S. putrefaciens cell was also shown to release organic acids, such as succinate or acetate in saline medium (Claessens et al., 2006). Interestingly, in our incubations (with or without S. putrefaciens cells), no significant additional DOC was measured in the anoxic incubations compared to oxic controls. Alternatively, iron reducing bacteria also oxidize DOC into CO2(g), hence inducing a decrease of DOC concentration as microbial Fe(III) reduction proceeds. This process can potentially “mask” a possible DOC release due to reductive dissolution. Let us consider the stoichiometry of the reduction reaction of ferrihydrite (Fe(OH)3) coupled to the oxidation of lactate (Bonneville, 2005):

+ - 2+ - 4 Fe(OH)3 + 7H + CH3CHOHCOO  4 Fe + 10 H2O + CH3COO + - HCO3

- 4 moles of Fe(II) and one mole of HCO3 are produced per mole of lactate consumed. Applying the same stoichiometry in our experiments, the 16 µM of Fe(II) produced by S. putrefaciens in P5AB (our most reactive soil sample) correspond to 4 µM of DOC consumed by oxidation, a minor contribution in comparison to the 605.9 µM of C present in solution.

We believe that the most important process that drives DOC release in our experiments is the direct solubilization of organic matter and plant tissue (Moore and Dalva, 2001). As mentioned above, a significant fraction of DOC in our incubations is produced during the

195

Part 4 Chapter 9 – Microbial Fe(III) reduction

7h equilibration period in the absence of added cells. Several field and experimental studies have shown that DOC concentration in soil solution and export to hydrological system increases upon rewetting of soils (Tipping, 1998), including podzol, following dry periods (Lundquist et al., 1999, Kalbitz et al., 2000, Scott et al., 1998, Christ and David, 1996). In biological active soil, the DOC pattern following rewetting is consistent with the general increase in biological activity with increasing moisture (compared to dry conditions). In our case, the extensive release of DOC immediately after rewetting is more likely due to a “flush” of OM adsorbed onto quartz grains. Beyond the effect of rewetting on DOC, it is important to consider that dry/wet cycles also drive the redox status in those soils, dry period being oxic while submersion quickly (hours) turns soil into an anaerobic environments (see section 8.3.5) (Balakhnina et al., 2010, Husson, 2013).

9.3.4 Influence of OM on microbial Fe(III) reduction

In our experiments with S. putrefaciens (but also without), we observed a strong positive correlation between the dissolved Fe and C concentrations (Figure 9.7). This relationship, already reported across whole catchments (Knorr, 2013), individual horizons (Hagedorn et al., 2000), and soil catenas (Fiedler and Kalbitz, 2003), is the basis to hypothesize that reductive dissolution of Fe(III) phases controls OM release into the liquid phase (Kleber et al., 2015). Our results suggests that it is the other way around, i.e., that microbial Fe(III) reduction is impacted by the OM content and quality in those soils, via an impact on mineral Fe(III) phases crystallinity and potentially by acting as electron shuttle (Hansel et al., 2004, Jiang and Kappler, 2008) or ligands for dissolved Fe(II) (Royer et al., 2002).

Electron shuttling is a well-known “strategy” of S. putrefaciens to reduce Fe(III) phases beyond their near-environment (Marsili et al., 2008, Kotloski and Gralnick, 2013). However, electron shuttling becomes significant at concentrations of DOC above of 5-10 mg C L-1 (Jiang and Kappler, 2008). Thus in P3 and P4, DOC concentrations (1.2 - 2.1 mg C L-1) are probably too low for efficient electron shuttling to occur. However in P5 - where DOC reached 7.3 mg C L-1 - electron

196

Part 4 Chapter 9 – Microbial Fe(III) reduction shuttling may have contributed to microbial Fe(III) reduction. DOC may also have enhanced microbial Fe(III) reduction by complexing released Fe(II), as suggested by the correlation between DOC and dissolved Fe(II) concentrations (Figure 9.7). This process may have limited Fe(II) adsorption onto bacteria and mineral surfaces, and increased thermodynamic gain of dissimilatory iron reduction (Royer et al., 2002, Zachara et al., 2002, Zachara et al., 1998).

Few studies have investigated the impact of OMA on microbial Fe(III) reduction. The reduction of Fe nanoparticles-OM associations is about 8 times faster than for pure nano-lepidocrocite (Pédrot et al., 2011). Faster Fe(III) reduction rates were measured with increasing C load in fresh Fe-OM associations and it was concluded that, at low C/Fe ratio, the inhibitory effect of surface site blocking and aggregation is greater than the enhancement of Fe(III) reduction by electron shuttling via humic acids while at high C/Fe ratios electron shuttling prevails (Shimizu et al., 2013). In contrast, Eusterhues et al. (2014b) observed the opposite, with slower Fe(III) reduction rates for increasing adsorbed or coprecipitated C loadings. In soil samples, it is not easy to determine the C/Fe ratio of the OMA. Following Eusterhues et al. (2003) and Mikutta et al. (2006), we estimated the C/Fe ratio of OMA in our samples using the mineral-protected C pool (MP-OC) and the poorly crystalline Fe content (PCr-Fe) (Figure 9.1, Table 9.2). P4 exhibits the higher MP-OC/PCr-Fe ratio (36.4 compared to 5.0 in P3 and 5.3 in P5) and presents a low microbial reactivity despite a large free Fe (Fed) pool (similar to P5 value – Figure 9.1). Overall, these observations support the results of Eusterhues et al. (2014b): high C/Fe in Fe-OM associations tends to decrease Fe(III) reduction rates. P4 has also the highest proportion of crystalline Fe(III) phases, having low microbial reactivity independently from their organic moieties in OMA. Thus, the slow microbial Fe(III) reduction rate of P4 compared to P3 and P5 could result from the cumulated “inhibitory” effects of (i) surface blocking/aggregation (high C/Fe), (ii) insufficient DOC concentration to insure efficient electron shuttling, (iii) and high proportion of crystalline Fe(III) oxyhydroxides.

197

P3 P4 P5

200 200 700 600 R² = 0.94 160 160

R² = 0.77 ]

1 500 - R² = 0.77 R² = 0.99 120 120 400

80 80 300 200 DOC [µmol.LDOC R² = 0.98 R² = 0.33 40 40 S 100 B 0 0 0 0 1 2 3 4 0 1 2 3 4 0 10 20 30 40

-1 Dissolved Fe(II) [µmol.L ]

Figure 9.7 – DOC concentration against dissolved Fe(II) concentration in the experiments with (“B”) or without bacteria (“S”), for the three soil samples of increasing age.

198

Part 4 Chapter 9 – Microbial Fe(III) reduction

9.3.5 Implications for podzolization processes

Given the ubiquity of DIRB (Lovley, 2013), each anoxic episode occurring in the B horizon is likely to induce the production of dissolved Fe(II) which will re-oxidize and precipitate to form Fe oxyhydroxides. The crystallinity of those neo-formed Fe(III) phases will depend on the accumulation of redox cycles (Thompson et al., 2006b), Fe(II) re-oxidation kinetics, duration of O2 exposure (Thompson, 2016), as well as on the composition of the soil solution, notably the concentrations of DOC or other ions (Zachara et al., 2002). Subsequently, the nature of the resulting Fe(III) phases will have an impact on the evolution of OM stabilization, and on the Fe(III) bioavailability as redox cycles occurs.

Along our chronosequence, the Fe pools and thus the OMA present in these illuvial horizons are quite distinct (Figure 9.1). As illustrated in Figure 9.8, the evolution of Fe(III) phases (using Feo/Fed ratio as a proxy) appears to be linked to the content of labile C, as proposed by Cornell and Schwertmann (2003). In P3, there is a relatively high amount of labile OC combined with a low total Fe content, thus most of the Fe is bound to OM forming complexes. In P4, the relative dominance of crystalline Fe(III) phases (low Feo/Fed ratio) is potentially due to lower content in labile OC which allows the precipitation of ferrihydrite but also its subsequent transformation into crystalline Fe(III) phases. In contrast in P5, the occurrence of high level of Fe allowed for the precipitation of ferrihydrite (as evidenced by a high Feo/Fed in Figure 9.8). However, its transformation towards more stable Fe(III) phases is hindered, or at least slowed down by the large amount of labile OC. Poorly crystalline Fe(III) phases are typically found in podzol B horizons (Zachara et al., 2002, Sauer et al., 2007, Lundström et al., 2000b). Yet, our results indicate that Fe(III) phases crystallinity seems to be controlled by the balance between Fe and C inputs and that more stable Fe(III) oxyhydroxides could predominate in podzolic soils having low labile OC.

199

Part 4 Chapter 9 – Microbial Fe(III) reduction

Fh 1,4 1,2 1 P5 0,8 P3 Feo/Fed 0,6 P4 0,4 0,2 Gt/Hm 0 0 5 10 15 20 25 Labile C [g/kg]

Figure 9.8 – Plot of Feo/Fed ratio against bioavailable C content in P3- Bh (270 years), P4-Bhs (330 years) and P5-Bhs (550 years) soil samples (see Figure 9.1 and section 9.2.2). Here, the Feo/Fed ratio estimates the crystallinity of Fe(III) oxyhydroxides (Fh – Ferrihydrite, Gt – Goethite, Hm – Hematite).

In our chronosequence subjected to high precipitation rate (~3200 mm annually), both P4 and P5 present indurated, cemented B horizons that may limit water drainage and cause waterlogging. As those soils turns anoxic, our results show that Fe(III) phases of B horizons becomes readily bioavailable for iron-reducing bacteria and that significant Fe(II) production can occur. Because of the sandy texture of those soils, anoxic episodes are probably short-lived and re- oxidation of Fe(II) is likely. An in situ monitoring of these redox variations induced by dry/wet cycles coupled with a detailed characterization of the Fe and OM phases are necessary to better understand the formation and evolution of OMA and podzolization processes.

200

Part 5 General Conclusion

Part 5: General conclusion and perspectives

I. General conclusion

The objectives of this work were: (1) to propose a conceptual model of C stabilization processes in soil; (2) to test this model in a podzolic soil chronosequence (0 to 530 years old soil profiles), and evaluate the short and long term (from day to century) evolution of SOM protection mechanisms.

Objective 1: Development of a soil conceptual model

Both the nature and formation pathway of soil organic matter (SOM) have been subject to debate. Recently, the traditional view of “humification” or “secondary synthesis”, has been opposed to an emergent view, “the degradative concept” (Lehmann and Kleber, 2015). In this view, the biogenic macromolecules are progressively degraded into smaller and more oxidized compounds. Besides, many, if not most, of soil organic molecules are of microbial origin. We have integrated this degradative concept in our model as a “SOM compartment” (Figure 2.8).

Also, the knowledge on C stabilization mechanisms evolved recently. C stabilization is seen as a continuum. So it is in the Microbial Efficiency-Matrix Stabilization framework of Cotrufo et al. (2013), or the soil continuum model of Lehmann and Kleber (2015).

201

Part 5 General Conclusion

The first SOM protection mechanism is an intrinsic recalcitrance, due to molecular characteristics, that renders natural OM less susceptible to degradation. This protection mechanism is significant during the initial stages of decomposition, and plays a crucial role in the short to medium term C stabilization (days to yeas). It was associated to the SOM compartment in our model.

The recalcitrance of SOM is now considered as relative, instead of absolute. Indeed, the decomposition of organic molecules depends essentially on microorganism enzymatic activity that relies on soil chemical and physical properties. The importance of the second protection mechanism, “inhibition of decomposers activity”, is consequently increasingly recognized.

The soil void space shelters soil microorganisms. The pore size distribution and connectivity determines the characteristics of their habitats. SOM in pore environments not adequate for microorganisms, or inaccessible to decomposer organisms is protected from degradation. The pore space is thus the second compartment of our model. It determines the second protection mechanism.

The third protection mechanism consists of the formation of organo-mineral associations (OMA). OMA exert a major control on long term stabilization of SOM (decades to millennia). Secondary minerals (i.e. oxides, hydroxides, oxyhydroxides of Al and Fe, phyllosilicates and short range-ordered aluminosilicates) are the most significant phases for OMAs. Secondary minerals may evolve with time. They are impacted by the presence of organic molecules. The “mineral compartment” of our model describes the mineralogical modifications leading to the formation of secondary minerals involved in OMA.

Soil biota play fundamental roles in mineral and organic evolutions. Their activity impacts (1) SOM transformation processes; (2) the weathering of primary minerals (secretion of H+, organic ligands, modification of the soil solution concentration); (2) direct or indirect formation of secondary minerals (3) dissolution of secondary

202

Part 5 General Conclusion minerals. Soil biota represent the last compartment of our conceptual model.

Consequently, SOM transformation and stabilization is strongly interconnected with biological and mineral compartments, leading to specific evolution of C stabilization in space and time during soil formation.

Objective 2: Testing the model on a podzolic soil chronosequence

Here we assess the short and long term (from day to century) evolution of SOM protection mechanisms in a podzolic soil chronosequence.

 Long term evolution of SOM protection mechanisms (centuries)

In the Vancouver chronosequence, soil genesis started with vegetation development, and an input of C from the litter (Figure I.1). As described in Chapter 7, a small accumulation of C started in the P1 and P2 profiles (120 and 175 yrs), followed by an increase of C accumulation in the illuvial horizon of P3 (270 yrs), P4 (330 yrs) and P5 (530 yrs). The C composition also evolved with time. In P1 and P2, most of the C was “labile”, with a vegetal and microbial origin, respectively, as suggested by the C:N ratio. In the E horizon of the three podzols (P3, P4 and P5), OM is characterized by a microbial origin. In the illuvial horizons of P4 and P5 (to a lesser extent in P3-Bh), OM originates from vegetation and is increasingly controlled by association with mineral phases.

The modifications in the organic compartment induced important changes in the physico-chemical compartment (Figure I.1). The biological activity led to a pH decrease in the surface horizon, as first measured in P3 E (270 yrs). This, in turn, impacted the mineral compartment (Figure I.1). Significant mineralogical modifications were measured in P4 and P5 podzol profiles (Chapter 6). Net acidification

203

Part 5 General Conclusion and pH decrease induced weathering in the surface horizons, as revealed by the evolution of the TRB, REE signature and Fe and Si isotopes. Secondary minerals, estimated by selective extractions, accumulated in the illuvial Bh, Bhs and Bs horizons of the two podzols. Consequently, the formation of OMA is a significant protection mechanism of SOM only in these horizons. The precipitation of secondary minerals did not impact the REE signature. In contrast, the Si and Fe isotopic compositions trace the formation of secondary minerals and their successive cycles of dissolution and precipitation. The Si and Fe isotopic signatures revealed an enrichment in aluminosilicates and Fe oxyhydroxides in the illuvial B horizons from the weathering of Si- and Fe-bearing minerals in the topsoil.

As developed in chapter 5, the micro-morphological data revealed an evolution of the physical environment in the chronosequence. In P1, P2 and P3 profiles, as well as in P4 and P5 E horizons, the porosity is high. In the P4 and P5 illuvial horizons (Bh, Bhs and Bs), the precipitation of secondary minerals, associated with OM, induced the clogging of the porosity, and the occurrence of periodic water saturation. This, in addition to the adsorption of OM on the newly formed mineral phases, will strongly impact OM degradation dynamics.

The modifications observed in the 3 compartments induced an evolution of SOM protection mechanisms and microbial populations in the chronosequence (Figure I.2). In P1 and P2 profiles, the C content was small but composed principally of “labile” C. No secondary mineral phases were present. Consequently, SOM protection was limited in these profiles. The microbial population is fungi-dominated, homogeneous in composition through the profiles.

P3 is a young podzol. An accumulation of illuvial C was measured, but not of secondary mineral phases. The soil structure was porous through the profile. An accumulation of recalcitrant compounds in the E may decreases SOM degradation rate. The microbial population is also fungi-dominated in the whole profile.

204

Part 5 General Conclusion

The oldest profiles, P4 and P5, present two contrasted horizons in terms of stabilization mechanisms and microbial population. Our results highlight the relative accumulation of recalcitrant C compounds in the eluvial horizons, mainly dominated by fungi. In the illuvial horizons, the accumulation of secondary minerals allowed the formation of OMA, and induced the clogging of the porosity. Both mechanisms resulted in a stabilization of OM against degradation. The episodic anoxic conditions, due to the cementation, may result in a lower accessibility for microbial activity and root development. The microbial population was bacteria-dominated in these horizons.

 Short term evolution of SOM protection mechanisms (days)

Part 3 revealed that the formation of OMA chiefly controls the long-term stabilization of SOM in the Vancouver podzolic soil chronosequence. These OMA are localized in the illuvial horizons, and maximal in the Bhs horizon of P4 and P5. In chapter 5, we showed that these horizons were periodically water-saturated. In chapter 9, we studied, in controlled conditions, the impact of a reducing event on organo-Fe associations in Bh-Bhs horizons. Under anoxic conditions, iron reducing bacteria dissolve Fe(III) phases and may potentially destabilize adsorbed or co-precipitated OM. We incubated three soil samples collected in three Bh-Bhs horizons along the chronosequence (from 270 to 550 years old) with Shewanella putrefaciens, a well- known iron reducing bacterium. With increasing age, podzolic differenciation and content of poorly crystalline Fe oxyhydroxides, both the rate and extent of microbial Fe(III) reduction increased. Despite the large dissolution of Fe(III) oxyhydroxides, no additional dissolved organic carbon (DOC) was measured as compared to un- inoculated experiments. However, a significant amount of DOC was released in all soil samples upon rewetting preceding the introduction of iron reducing bacteria. The presence of OM likely favors the persistance of bioavailable poorly crystalline Fe(III) oxyhydroxides (ferrihydrite), and, as such, contributes to increase the extent and rate of microbial Fe(III) reduction. The evolution of the mineral-associated C fraction and its impact on the mobility of Fe (via microbial Fe(III)

205

Part 5 General Conclusion reduction) as podzolization proceeds might be a crucial process in soil development and the resulting OM dynamics.

 These results confirmed that minerals, SOM, physico-chemical conditions and microbial populations evolve interdependently with time and pedogenesis. The resulting SOM protection mechanisms are site- and horizon-specific, and change over various time scales (days to centuries) during soil formation.

 Our conceptual model confirms the need of an interdisciplinary approach for better understanding the C dynamics in soils, integrating the drivers of microbial decomposition: 1) pedoclimatic/ environmental; 2) biochemical; 3) pedological.

206

Figure I.1 – Evolutions in the organic, physico-chemical and mineral compartments, for the five Vancouver profiles of increasing age (P1-120 yrs, P2-170 yrs, P3-270 yrs, P4-330 yrs and P5-530 yrs). 207

Figure I.2 – Evolutions of the main SOM protection mechanisms and of the microorganisms compartment, for the five Vancouver profiles of increasing age (P1-120 yrs, P2-170 yrs, P3-270 yrs, P4-330 yrs and P5-530 yrs).

208

Part 5 General Conclusion

II. Perspectives

Perspective 1: Investigations in the Vancouver chronosequence

As developed in our conceptual model, the pore water is the medium for biological and chemical reactions. Its composition vary in space and time, impacted by the water flow. The latter is controlled by the pore architecture, and the precipitation/evapotranspiration relation. The water flux strongly impacts soil processes by transporting organic (substrate for microorganisms) and inorganic (ions for secondary minerals formation) solutes to other horizons. As mentioned by Lucas et al. (1996), the final result of pedogenesis is controlled by the system hydrodynamics.

The understanding of podzol formation in the Vancouver chronosequence would benefit from an analysis of the evolution of soil solution composition, in association with the system hydrodynamics, in the different horizons. Cycles of drying-rehumectation might have a major role in phase formation and dissolution, through the impact on the solution equilibrium (content in dissolved organic carbon (DOC) and inorganic ions).

Perspective 2: Better understanding of the processes in podzols

Podzol turned out to be a very efficient soil model to test our conceptual model and study SOM protection mechanisms. They present two very distinct horizons in terms of processes, separated in space. E, compared to B(h, hs, s), differ in: structure (porous vs. pore- clogged); microbial population (fungi vs. bacteria); OM input (litter vs DOC); mineral composition (residual primary minerals vs increase in secondary minerals); dominant SOM protection mechanism (recalcitrance vs OMA).

The processes occurring in the E horizon are: (1) SOM transport and leaching and (2) primary minerals weathering, mainly fungal- driven (e.g. Van Bremen et al. 2000a). However, the processes 209

Part 5 General Conclusion involved in the B horizon formation and evolution, and explaining the differences in podzol morphologies still need to be deciphered. More specifically (refer to Figure II.1):

Litter Primary Fungi minerals Plant-derived C

Bacteria ④

①

② ③ OMA

Figure ↗II.1 Microbial – Knowledge- gaps in podzol genesis/evolution and research perspectives.derived C Secondary minerals ↗ oxidation state ① OM composition ↘ size As revealed by a literature review (section 8.3), the OM composition and age in the B horizon differ between podzols. The conceptual model, exposed in section 2.4, proposes an explanation for these differences, based on the relative importance of: 1) the OM transformations in this horizon; 2) the composition of the weathering solution, after the passage through the E horizon; 3) the impact of the poral system in the two first mechanisms (well-drained podzol vs hydromorphic; porous vs. pore clogging). However, the main drivers

210

Part 5 General Conclusion explaining the differences in OM composition in the B horizons need to be determined: Environmental (water saturation)? Biotic (type of vegetation and microorganisms)? Mineralogy of the parent material?

② Type of OMA formed

Soils are complex, heterogeneous and in constant change over various scales of time and space. The prediction of the secondary minerals and OMA formed in situ is thus challenging. As mentioned by Kleber et al., (2015), the formation of “organically poisoned” Fe and Al oxides in soil via coprecipitation is likely a widespread mechanism, especially in Podzols. So far, it cannot be confirmed with reliable quantitative measures. The understanding of the illuvial horizon would benefit from: 1) a finer description of these phases (coprecipitates? crystalline/amorphous phases with adsorbed OM?); 2) determining the role of the C/Metal ratio in solution on the type of phase formed; 3) determining the role of the C/Metal ratio of the solid phase on its reactivity; 4) determining the impact of the type of OM on the phases formation and reactivity.

③ OMA dynamics

Secondary minerals constantly evolve through cycles of dissolution and precipitation. This should be included in SOM stabilization models. SOM partly control soil mineralogical evolution (mineral dissolution, complexation of the dissolved ions, impact on the type of precipitated phases). The protective impact of OMA on OM is well documented (reduction of OM bioavailability for microorganisms). However, whether the presence of associated OM protects the mineral counterpart of the OMA from dissolution is still poorly known. Adsorbed OM might either protect mineral surfaces from dissolution (surface passivation), or enhance dissolution (e.g. via electron shuttling, ligand-promoted dissolution). The presence of OM might also have an indirect impact on mineral bioavailability/solubility through the effect on crystallization and type of secondary minerals

211

Part 5 General Conclusion formed. Further studies are needed to determine the main impact, potentially differing from one OM type to another.

The OMA understanding have drastically increased these last years (e.g. Kleber et al 2015). An integration of these knowledge would bring key advances to understand podzol genesis and evolution. The pore space is the place for sorption-desorption processes and OMA precipitation-dissolution. The impact of the pore space, should be taken into account given its impact on: 1) soil solution composition (water flow, redox status, etc.), and more specifically the C/M ratio in solution; 2) the OMA dynamics (OM degradation/ stabilization? Mineral dissolution/precipitation?); 3) the availability of nutrients for microorganisms, and the suitability of the habitat.

A particular event susceptible to have a major impact on OMA dynamics in B(h, hs, s) horizons is the anoxic episode. Indeed, Fe secondary phases (among the more important phases for OMA formation) are redox-sensitive. Few studies consider the impact of anoxic events, and potential Fe reductive dissolution, in podzol theories. However, given the impact of hydromorphic characteristics on podzol morphology (Buurman and Jongmans, 2005), anoxic events might play a crucial role. In situ experiments measuring redox fluctuation and resulting Fe dynamics would be interesting.

④ Soil microorganisms

Given the increasingly recognized role of microorganisms in all aspects of soil development (mineral and SOM evolutions), our understanding of C stablization would be enhanced by a better characterization of the microbial population and diversity in soils. The role played by microorganisms on the organic counterpart of OMA (degradation, direct deposition) is much more documented than their effects on minerals (especially secondary minerals dissolution- formation). Geomicrobiology is a relatively recent branch and should receive more attention. The effects of microbial activity in podzol B horizons are still poorly documented. In the two hydromorphic podzols of the Vancouver chronosequence, we showed that the

212

Part 5 General Conclusion population is bacteria-dominated. Is this also the case for non- hydromorphic podzols? What are the functions of the bacterial community measured? More specifically, iron reducing bacteria are ubiquitous according to Weber et al. (2006) and Lovley (2013), but are they present in podzols? What is their role in the B horizon evolution?

Perspective 3: Application of the conceptual model to other soils and environments

The conceptual model was an interesting tool to study SOM dynamics in podzols. In order to validate it, it would be interesting to apply it on other soil types playing a significant role in C cycle, like Histosols, Andosols and Chernozems. As shown in the table 0.1, these soils have the highest C concentrations (in Mg/ha) and represent an important % of the global SOC content.

Also, this study focuses on the evolutions of SOM protection mechanisms due to variation of one main factor of soil formation (i.e. time). It could be interesting to test the model for variations of other factors, like vegetation (land use change), or climate, especially in the global climate change context.

213

Part 5 General Conclusion

III. Final words

Considering its major role in soil fertility and within the global C cycle, understanding the mechanisms that control SOM formation and evolution within soils is fundamental. This understanding should not be focused only on the development of management strategies to increase C sequestration in soils. “This goal seems counterproductive given that soil organic matter is most beneficial when it decays and releases energy and nutrients” (Lehmann and Kleber, 2015). As said by Janzen (2015) “maximizing carbon ‘stocks’ is less critical than maintaining ‘flows’ to sustain the manifold functions performed by ecosystems. Organic carbon may be best viewed, not as a reservoir entrapped in soil, but as a stream of atoms flowing through”, and the sustainable management of soil organic matter turnover, is more important than the increase of non-productive organic matter deposits. Soils should be considered as open ecosystems, largely interacting with other spheres and cycles (Nitrogen, water, etc.), and as the place where most organisms live and die (Ponge, 2015); not only as substrate for crops and infrastructures.

214

Part 6: Appendix

A1. Related publications and collaborations

A2. Co-authored papers

A3. Carbon stable isotopes

A4. SOMFrac ring trial

215

Part 6: Appendix

216

Part 6: Appendix A1. Related publications and collaborations

A1. Related publications and collaborations

A1.1 Conference proceedings

Vermeire M.-L., Lavkulich L. and Cornélis J.-T., The interdependent relationship between C dynamics and soil-forming processes in a podzolic chronosequence under a temperate rainforest. Geophysical Research Abstracts, Vol. 15, EGU 2013-4661-2, EGU General Assembly 2013. April 7-12, Vienna, Austria. Oral.

Vermeire M.L., Doetterl S., Bodé S., Delmelle P., Van Oost K. and Cornelis J-T., Evolution of C dynamics, microorganism populations and soil-forming processes in a podzolic chronosequence. DGB Workshop, May 4-6, 2014, Freising, Germany. “Soil processes – is the whole system regulated at ‘hot spots’? From micro-scales to the pedon”. Oral.

Vermeire M.L., Doetterl S., Bodé S., Delmelle P., Van Oost K. and Cornelis J-T., Capacity of microorganisms to decompose organic carbon affected by an increasing content of reactive mineral phases in a podzolic soil chronosequence. Soil Science Society of Belgium (SSSB) Thematic Day, December 5, 2014, Brussels. “Soil plant interactions in a changing world”. Oral

Vermeire M.L., Cornu S., Fekiacova Z., Delvaux B., Cornelis J.T., Do rare earth elements (REE) trace pedogenic processes during podzolization ? Goldschmidt 2015, June 26-1 july, Prague, CR. Oral.

Vermeire M.-L., Bonneville S., Stenuit B., Delvaux B., Cornélis J.-T., Bacterial FeIII reduction enhances the dissociation of Fe oxyhydroxides - organic matter associations in podzolic Bhs soil horizons. Goldschmidt 2016, August 16-21, Yokohama, Japan. Oral.

Vermeire M.-L., Doetterl S., Bode S., Delmelle P., Van Oost K. and Cornelis J.-T., Capacity of microorganisms to decompose organic carbon affected by an increasing content of reactive mineral phases in

217

Part 6: Appendix A1. Related publications and collaborations a podzolic soil chronosequence. Geophysical Research Abstracts, Vol. 16, EGU2014-12547-1, EGU General Assembly 2014, Vienna, Austria, April 27-May 2, 2014. Poster.

A1.2 Publications

Cornelis J.T., Weis D., Lavkulich L., Vermeire M.L., Delvaux B., Barling J., 2014. Silicon isotopes record dissolution and re-precipitation of pedogenic clay minerals in a podzolic soil chronosequence, Geoderma, 235-236, 19-29.

Vermeire M.L., Cornu S., Fekiacova Z., Detienne M., Delvaux B., Cornelis J.T., 2016. Rare Earth Elements dynamics along pedogenesis in a chronosequence of podzolic soils, Chemical Geology 446: 163-174.

Fekiacova, Z., Vermeire, M. L., Bechon, L., Cornelis, J. T., Cornu, S., 2017, Can Fe isotope fractionations trace the pedogenetic mechanisms involved in podzolization? Geoderma. DOI: 10.1016/j.geoderma.2017.02.020.

Vermeire M.-L., Bonneville S., Stenuit B., Delvaux B., Cornélis J.-T., Microbial Fe(III) reduction of organo-mineral associations along a podzolic soil chronosequence: impact on C release? - In preparation.

Vermeire M.L., Doetterl S., Delmelle P., VanOost K., Van Ranst E., Delvaux B., Cornélis J.T., Co-dependent evolution of soil microbial populations and organo-mineral associations in a podzolic soil chronosequence. - In preparation.

A1.3 Participation to conferences and workshops

 Soil Carbon Sequestration, for climate, food security and ecosystem services, Reykjavík, Iceland, May 26-29, 2013, hosted by the Soil Conservation Service of Iceland and the Agricultural University of Iceland. Followed by the SoilTrEC Training "Land-use practice and sustainable use of soil", May 30-June 2, 2013, Sólheimar, Iceland.

218

Part 6: Appendix A1. Related publications and collaborations

 Day of Young Soil Scientists 2003, Belgian Society of Soil Science, Bruxelles, Belgium, November 26, 2013.

 Thematic day : 50 years of sharing knowledge in soil science for worldwide sustainable development, Belgian Society of Soil Science, Ghent, Belgium, December 5, 2013

 SSSB annual excursion, “Soils of the Campine, a historical perspective”, Belgian Society of Soil Science, June 5, 2013.

 Visit of the Leibniz Centre for Agricultural Landscape Research (ZALF): February 5, 2014. Meeting with Prof. Dr. Michael Sommer, head of the Institute of Soil Landscape Research. Müncheberg, Germany

 Visit of the CEREGE (Centre Européen de Recherche et d’Enseignement des Geosciences de l’Environnement), Aix en Provence (France): November 19-21, 2013. Meeting with the INRA team of soil and water geochemistry (Gérôme Balesdent, Sophie Cornu, Zuzana Fekiacova)

A1.4 Research stays and field campaign

 Research stay in CEREGE, Aix en Provence (France): February 9 - March 1, 2014. REE analysis on the Vancouver chronosequence samples with Sophie Cornu and Zuzana Fekiacova (INRA, National Research Institute of Agriculture).

 Research stay in Arvorezinha (Brazil), in association with the Federal University of Santa Maria: August 2-30, 2014. Sampling campaign for the SOGLO project

219

Part 6: Appendix A1. Related publications and collaborations

A1.5 Collaboration with other Belgian universities

 Collaboration with Sebastian Doetterl, Isotope Bioscience laboratory – ISOFYS, Ghent University: Amino-sugars and PLFA analysis on the Vancouver samples

 Collaboration with Steeve Bonneville, Département des Sciences de la Terre et de l’Environnement, ULB. Organic matter decomposition through Fe reduction with geobacter along the podzol chronoséquence.

220

Part 6: Appendix A2. Co-authored papers

A2. Co-authored papers

A2.1 - Silicon isotopes record dissolution and re- precipitation of pedogenic clay minerals in a podzolic soil chronosequence7

Abstract

By providing the largest part of the reactive surface area of soils, secondary minerals play a major role in terrestrial biogeochemical processes. The understanding of the mechanisms governing neo(trans-)formation of pedogenic clay minerals in soils is therefore of the utmost importance to learn how soils evolve and impact the chemistry of elements in terrestrial environments. Soil-forming processes governing the evolution of secondary aluminosilicates in Podzols are however still not fully understood. The evolution of silicon (Si) isotope signature in the clay fraction of a podzolic soil chronosequence can provide new insight into these processes, enabling to trace the source of Si in secondary aluminosilicates during podzol-forming processes characterized by the mobilization, transport and precipitation of carbon, metals and Si. The Si isotope compositions in the clay fraction (comprised of primary and secondary minerals) document an increasing light 28Si enrichment and depletion with soil age, respectively in illuvial B horizons and eluvial E horizon. The mass balance approach demonstrates that secondary minerals in the topsoil 30 eluvial E horizons are isotopically heavier with δ Si values increasing from -0.39 to +0.64‰ in c.a. 200 years, while secondary minerals in

7 Adapted from Cornélis, J.T., Weis D., Lavkulich, L., Vermeire M.L., Delvaux B., Barling, J. (2014) Silicon isotopes record dissolution and re-precipitation of pedogenic clay minerals in a podzolic soil chronosequence, Geoderma 235–236, 19– 29.

221

Part 6: Appendix A2. Co-authored papers the illuvial Bhs horizon are isotopically lighter (δ30Si = -2.31‰), compared to the original “unweathered” secondary minerals in BC horizon (δ30Si = -1.40‰). The evolution of Si isotope signatures is explained by the dissolution of pedogenic clay minerals in the topsoil, which is a source of light 28Si for the re-precipitation of new clay minerals in the subsoil. This provides consistent evidence that in strong weathering environment such as encountered in Podzols, Si released from secondary minerals is partially used to form “tertiary clay minerals” over very short time scales (ca. 300 years). Our dataset demonstrates the usefulness to measure Si isotope signatures in the clay fraction to discern clay mineral changes (e.g., neoformation versus solid state transformation) during soil evolution. This offers new opportunity to better understand clay mineral genesis under environmental changes, and the short-term impact of the dissolution and re-precipitation of pedogenic clay minerals on soil fertility, soil carbon budget and elemental cycles in soil-plant systems.

A2.1.1 Introduction

Soil is a precious but threatened resource (Banwart, 2011). In order to protect it for the future we need a better understanding of the soil-forming processes controlling the evolution of newly-formed minerals (secondary minerals). Soil formation progressively modifies parent rock material and controls the pathways of primary mineral weathering and secondary mineral synthesis in the clay fraction (Chadwick and Chorover, 2001). The secondary minerals consist of layer-type aluminosilicates (called pedogenic clay minerals) and Fe-, Al-oxyhydroxides, both of which play a major role not only in soil fertility, but also in the transfer of elements and pollutants from land to ocean given their high surface reactivity (Sposito, 2008). Moreover, the capacity of charged mineral surfaces to form adsorption complexes can stabilize organic carbon (OC) in soils through the formation of organo-mineral associations, partly controlling global C budget (Torn et al., 1997, Parfitt et al., 1997).

The formation of secondary minerals and their evolution during pedogenesis have been studied for over a half century (Wilson, 1999).

222

Part 6: Appendix A2. Co-authored papers

The proportion and the chemistry of minerals in the clay fraction change with soil evolution (Egli et al., 2002, Righi et al., 1999, Turpault et al., 2008). Some environmental changes (vegetation type, agricultural practices, land-use, climate and drainage) can amplify the modification of clay mineralogy on very short time-scales (10-1000 yrs) (Caner et al., 2010a, Collignon et al., 2012, Cornu et al., 2012, Mareschal et al., 2013). These rapid clay modifications occur in chemically reactive soil micro-environments, i.e. the part of the soil influenced by roots and earthworms (Calvaruso et al., 2009, Jouquet et al., 2007), and can play a key role in geochemical balance of several minor and major elements in soils and sediments (Michalopoulos and Aller, 1995, Velde and Meunier, 2008). However, the origin of elements involved in clay neo(trans-)formation is still not well understood.

Podzol, the focus of this study, is a type of soil that covers more than 3% of the Earth’s land surface. The low stock of weatherable minerals, the acidic conditions and complexing capacity of organic acids in the environment where Podzols developed are responsible for mobilization, transport and precipitation of carbon (C), metals (Fe, Al) and silicon (Si) in the soil profile (Lundström et al., 2000a). A fully developed Podzol consists of a leached grey subsurface eluvial E horizon contrasting with the accumulation of elements in the dark illuvial B horizons. The topsoil is characterized by the production of organic acids that form soluble organo-metallic complexes enhancing weathering in the eluvial E horizon. This E horizon overlies the dark C- enriched Bh horizon and reddish Fe-, Si-, Al-enriched Bhs/Bs horizons (Lundström et al., 2000). Given the very acidic conditions in Podzols, besides the weathering of primary minerals, secondary clay minerals can be dissolved in the podzolic weathering front (Ugolini and Dahlgren, 1987, Zabowski and Ugolini, 1992). The weathering front of the primary and secondary minerals describes the soil depth where minerals dissolve faster than they form. A podzolic soil chronosequence, i.e. in which all soil-forming factors remain constant except time; represents an ideal natural system for the study of the effect of time on pedogenic clay minerals behavior in soils.

223

Part 6: Appendix A2. Co-authored papers

Stable Si isotopes fractionate during silicate weathering and the biogeochemical Si cycling (Opfergelt et al., 2010, Ziegler et al., 2005), and as such provide a means of tracing the bio-physico- chemical processes in terrestrial environments (Cornelis et al., 2011). In addition to its incorporation in the mineral structure during the formation of crystalline layer-type aluminosilicates, poorly-crystalline aluminosilicates and pedogenic opal, monosilicic acid (H4SiO4) released into soil solution can also be transferred into the biosphere to produce biogenic opal (phytoliths) or be adsorbed onto secondary Fe oxy-hydroxides. The incorporation of Si in mineral structures through neoformation of secondary pedogenic and biogenic precipitates and its adsorption onto the surfaces of Fe oxides are two processes favoring the retention of light 28Si in soils and contributing to the enrichment of rivers in heavy 30Si (Delstanche et al., 2009, Georg et al., 2007, Opfergelt et al., 2006, Ziegler et al., 2005). Clay minerals can also be unstable in organic and inorganic acidic environments where they dissolve (Sokolova, 2013, Zabowski and Ugolini, 1992), and enrich soil solutions (Cornelis et al., 2010) and rivers (Cardinal et al., 2010) in light 28Si. The naturally occurring mass-dependent Si isotopic fractionation is induced by dissolution, precipitation and adsorption but not by complexation as chemical binding of Si to organic matter are negligible (Pokrovski and Schott, 1998). It has also been demonstrated that the Si isotopic compositions of secondary clay minerals relates to climatic gradient and its control on clay mineralogy (Opfergelt et al., 2012). However Si isotopes have never been used to better understand clay mineral modifications induced by soil-forming processes under identical geo-climatic conditions. The rapid modification of clay mineralogy in Podzol is well documented (Caner et al., 2010a, Egli et al., 2002, Righi et al., 1999), but the fate of Si released in soil solution after clay modification has not yet been studied, even though it is of crucial importance for identifying the sources controlling the formation of pedogenic clay minerals in soils.

In this study, we aim to use Si isotope signatures of the clay fraction in a podzolic soil chronosequence for gaining better insights into the origin of Si in pedogenic clay minerals. To achieve this goal, we analyzed Si isotopes, elemental (Ge/Si, Al/Si, Fe/Si) ratios and determined clay fraction mineralogy for an age sequence of four soil

224

Part 6: Appendix A2. Co-authored papers profiles undergoing podzolization (Cox Bay on Vancouver Island, Canada) (Figure A2.1.1) and for a single Podzol pedon (Gaume, Belgium). The Cox Bay chronosequence offers an opportunity to study the variation of Si isotopic composition and elemental ratios of the clay fraction in the vertical pedogenic scale: E, Bh, Bhs, Bs, Bw and BC horizons, and in the horizontal time-dependent scale: duration of pedogenesis from 0 to 335 years. We used the Belgian Podzol as a “natural duplicate” in temperate climate to corroborate the processes documented in the soil samples from the Cox Bay podzolic soil chronosequence.

A2.1.2 Materials and methods

A2.1.2.1 Sample collection and location

We sampled a soil chronosequence undergoing podzolization in Cox Bay (CB), on the west coast of Vancouver Island (British Columbia, Canada). At the Cox Bay study site, three main vegetative associations are identified in the chronosequence. These correspond to Sitka spruce (Picea sitchensis ) in the younger site (CB-120 yrs), and Sitka spruce (Picea sitchensis ) and salal (Gaultheria Shallon) in the sites of 175 and 270 yrs (CB-175 and -270 yrs). The oldest site (CB-335 yrs) is characterized by Sitka spruce (Picea sitchensis ), Douglas fir (Pseudotsuga menziesii), salal (Gaultheria Shallon) and western sword fern (Polystichum munitum).

Heavy mean annual precipitation (3200 mm) coupled with frequent fogs and sea sprays ensure an abundance of moisture and nutrients year round in this maritime temperate climate (Cfb: without dry season and with warm summer; Peel et al. (2007)). The Tofino Area Greywacke Unit is the source of the beach sand parent material, from which soils have developed in the age sequence (Singleton and Lavkulich, 1987). Sampling sites were located along a transect (0-94 m) perpendicular to the present shoreline (Figure A2.1.1). Dendrochronology and geomorphology established surface duration of pedogenesis ranging from 0 to 335 yrs for the four selected pedons.

225

Figure A2.1.1 - Cross section of the Cox Bay study area showing site locations and soil horizons, depending on their respective age of soil formation: CB-0 yr, CB-120 yrs, CB-175 yrs, CB-270 yrs and CB-335 yrs.

226

Part 6: Appendix A2. Co-authored papers

Tree ages were determined counting the tree rings in the increment bores. Assuming that the beach built towards the ocean in a configuration parallel to the existing shoreline and that a linear deposition rate occurred with time between successive oldest trees, the rate of advance of the beach front was estimated to be 0.26 m per year. At this rate, the 13-m strip of sand containing tree seedlings would have accumulated in approximately 50 years (Singleton and Lavkulich, 1987).

With soil development, there was progressive deepening and differentiation of genetic horizons during podzolization, resulting in soil classification (World Reference Base for Soil Resources – WRB) that ranged from Dystric Cambisol at the youngest sites (CB-120 yrs; CB-175 yrs) to a Placic Podzol at the oldest site (CB-335 yrs) (Figure A2.1.1). The 335-year-old Podzol is characterized by the following soil horizons development: eluvial albic E horizon (strongly weathered horizon) → illuvial spodic Bh horizon (enriched in organic matter) → Bhs horizon (enriched in Fe oxyhydroxides and organic matter) → Bs horizon (enriched in poorly-crystalline aluminosilicates and Fe oxyhydroxides) → Bw horizon (development of colour and structure without illuvial accumulation of materials) → BC horizon (weakly coloured and structured; little affected by pedogenic processes).

The sampling area of the Podzol in Gaume (Belgium), ranging in altitude from 300 to 350 m above sea level, has an annual rainfall of 1100 mm and a mean annual temperature of 7.7°C (Herbauts, 1982), and is also characterized by a maritime temperate climate (Cfb; Peel et al. (2007)). The Podzol is located on the Lower Lias outcrop in Southeast Belgium (Gaume). The bedrock (calcareous sandstone of Lower Lias age) is covered by a two-layered sheet: an autochthonous sandy layer, formed by the dissolution of the calcareous bedrock, is overlaid by a mixture of this sandy material with loessic silt-sized particles. The Belgian Podzol developed under heather (Calluna vulgaris) is characterized by a similar morphological profile as the Podzol in Cox Bay sequence (CB-335 years) with the following horizons: E-Bh-Bhs-Bs-Bw-BC.

227

Part 6: Appendix A2. Co-authored papers

A2.1.2.2 Physico-chemical characterizations

The soil samples were air-dried, then sieved and homogenized. The content of free iron oxides was assessed after selective dissolution of Fe oxides using Na-dithionite-citrate-bicarbonate and ammonium oxalate-oxalic acid (Fedcb = crystalline Fe oxides, Feox = poorly- crystalline Fe oxides). The content of Si bound to poorly crystalline aluminosilicates and weakly-ordered Fe oxyhydroxides was estimated on fine earth by extraction with ammonium oxalate-oxalic acid (Siox). Al complexed with organic ligands was assessed using the complexing agent Na-pyrophosphate at pH 10 (Alp). The total organic carbon (OCtot) content was measured on ground samples using CNS analyzer.

The clay fraction (<2 µm) was separated using a ‘clean procedure’ without any oxidative treatment. Air-dried soil was dispersed in deionized water and sonicated. The suspension was then separated on a 50 µm sieve, re-suspended in deionized water and sonicated and sieved until the supernatant was clear after sonication. The fraction retained in the sieve was collected as the >50 µm sand fraction. Clay (0–2 µm) and silt (2–50 µm) fractions were then collected by gravimetric sedimentation after dispersion using an ultrasonic probe and Na+ as a dispersion agent.

A2.1.2.3 X-ray diffraction patterns

XRD analyses were carried out on the clay-sized fraction (<2µm) of soil horizons sampled in the Cox Bay chronosequence (120, 175, 270 and 335 years), using CuKα radiation in a Bruker Advance diffractometer. After removal of the organic matter by treating the sample with 6% H2O2 at 50°C, and removal of Fe-oxyhydroxides using dithionite-citrate-bicarbonate, eight standard treatments were applied to determine mineralogy of the clay fraction: K-saturation (KCl 1N) followed by drying and heating at 20, 105, 300 and 550°C, and Mg- saturation (MgCl2; 1N) followed by drying at 20°C and saturation with ethylene-glycol (eg). XRD analysis was also performed on powder samples of the clay fraction after removal of organic matter and Fe oxyhydroxides but without any further treatment for quantifying

228

Part 6: Appendix A2. Co-authored papers mineralogy of the clay fraction using the Siroquant software V4.0 (Sietronics Pty Ltd), and for the following horizons: BC horizon (CB-120 years), E horizons (CB-175, 270 and 335 years), Bh horizon (CB-335 years) and Bhs horizon (CB-335 years).

A2.1.2.4 Isotopic and geochemical analyses

Silicon isotope compositions and elemental (Ge, Al, Fe and Si) concentrations were measured on clay-sized fraction (<2µm) extracted from all the horizons of the four soil profiles in Cox Bay (clay- CB 120 yrs; clay-CB 175 yrs; clay-CB 270 yrs; clay-CB 335 yrs) and the undated podzolic soil profile in Gaume (clay-G), and also on parent material of soils in Cox Bay (sand fraction of the beach sand; Beach-CB 0 yr). An alkaline digestion with 99.99% pure NaOH is used to transform solid samples into an aqueous HF-free solution (Georg et al., 2006). All dissolutions and chemical separations were carried out in Class 100 laminar flow hoods in Class 1000 clean labs, mass spectrometric analyses were performed in Class 10,000 laboratories at the Pacific Centre for Isotopic and Geochemical Research (PCIGR) at the University of British Columbia (UBC). Al, Fe and Si contents of the dissolved NaOH fusions were analysed by ICP-OES (Varian 725-ES) with Europium as the internal standard. For Ge measurements, the dissolved NaOH fusions were dried and re-dissolved in 1% v/v HNO3 with 10 ppb Indium (In) for analysis by HR-ICP-MS (Element 2) in medium resolution.

The remaining dissolved NaOH fusion solution was purified for isotopic analyses through cation exchange chromatography (Georg et al., 2006). The Si isotope compositions were measured on a Nu Plasma (Nu 021; Nu Instruments Ltd, UK) MC-ICP-MS in dry plasma mode using type B cones and a Cetac Aridus II desolvating nebulizer system. Instrumental mass bias was corrected by simple sample-standard bracketing of measured Si isotope ratios, i.e. one sample measurement normalized to the average of two bracketing NBS-28 standard measurements. Silicon isotopic compositions are expressed as deviations in 30Si/28Si relative to the NBS-28 reference standard using the delta (δ) per mil (‰) notation: δ30Si =

229

Part 6: Appendix A2. Co-authored papers

30 28 30 28 [( Si/ Sisample)/( Si/ SiNBS28)-1] × 1000. Each sample was measured at least twice during different analytical sessions. Silicon isotopic (δ30Si) values are reported as the mean of replicate isotopic analyses (n>2) ± 2 standard deviations (SD). The NBS-28 (quartz standard) which processed through the full analytical procedure, and analyzed over a period of 7 months during 5 data acquisition sessions gave a value of δ30Si = 0.01±0.18‰ (2SD, n=66). Accuracy and reproducibility were also checked on reference materials (diatomite and BHVO-2) at the beginning and at the end of each sample series. These gave values identical within error to previously published values: 1.24±0.13‰ (2SD, n=15) for diatomite and -0.29±0.19‰ (2SD, n=6) for BHVO-2 (Reynolds et al., 2007, Savage et al., 2012).

A2.1.3 Results

A2.1.3.1 Soil Mineralogy

The parent material of the soil chronosequence (0-335 years) is Cox Bay beach sand (Singleton and Lavkulich, 1987), which is comprised of very well-sorted glacial sands with little wearing off and smoothing sharp edges and corners. The primary minerals present in the beach sand C material identified by X-Ray Diffraction and microscopy are quartz, amphibole, pyroxene, olivine and feldspars, as well kaolinite precipitating in the dissolution pits of feldspars. The parent material does not contain inherited clay minerals, except kaolinite present in the weathered feldspar. We observe an increase of oxalate-extractable Siox in Bhs, Bs and Bw horizons of the Podzol (CB- 335 years) (Table A2.1.1). We also document a strong mobilization of Fe in Podzol after 335 years, characterized by an accumulation of crystalline and amorphous Fe oxides in the Bhs horizon, which is related to an increase of OC content. This co-accumulation of Fe oxides and OC is also observed in the Belgian Podzol in Bh and Bhs horizon. The content of clay-sized minerals is quite constant in the Belgian Podzol while we observe an increase of clay content towards the topsoil in the Canadian podzolic soil chronosequence (Table A2.1.1). The content of clay-sized minerals in the entire soil profiles increase over time in the chronosequence.

230

Table A2.1.1 - Summary of the major soil physical and chemical characteristics (fine earth <2mm) of the investigated soils. a b Horizon Depth pH Soil fractions Siox Sidcb Feox Fedcb Alox Alp OCtot Clay fraction sand silt clay Si Al Fe Ge µg.g- cm % g.kg-1 % 1 Cox Bay 120 yrs (Dystric Cambisol) BC 0-75 5.9 99.2 0.6 0.3 0.1 0.4 1.7 1.8 0.7 0.5 9.5 15.8 8.6 12.0 2.9 Cox Bay 175 yrs (Dystric Cambisol) E 0-3 5.4 90.2 7.1 2.7 0.1 1.1 1.4 2.8 0.4 0.3 35.2 22.0 8.3 8.1 5.7 Bw 3-44 5.8 99.0 0.6 0.4 0.1 0.3 1.7 2.0 0.9 0.7 4.3 13.5 7.4 10.5 2.0 BC 44-75 5.9 99.6 0.2 0.1 0.1 0.2 1.0 1.3 0.6 0.4 2.7 16.1 8.9 9.4 3.1 Cox Bay 270 yrs (Haplic Podzol) E 0-7 4.6 90.8 6.1 3.1 0.1 0.8 1.2 2.5 0.6 0.4 13.3 23.3 9.1 5.6 8.9 Bh 7-23 5.1 97.2 1.7 1.0 0.2 0.4 2.3 2.6 1.3 0.9 16.1 15.5 8.7 12.1 3.0 Bw 23-57 5.3 97.4 1.8 0.8 0.2 0.4 2.1 2.3 1.1 0.8 10.4 16.1 8.8 12.0 2.8 BC 57-75 5.4 98.2 1.1 0.7 0.2 0.3 1.7 2.2 1.4 1.0 7.6 13.1 10.0 11.3 2.4 Cox Bay 335 yrs (Placic Podzol) E 0-16 4.8 82.3 14.4 3.0 0.1 0.4 0.2 0.5 0.6 0.5 10.7 26.5 10.5 1.9 12.9 Bh 16-23 5.6 88.0 8.7 2.8 0.5 1.0 3.3 4.4 8.8 5.0 36.8 16.9 14.3 6.2 5.8 Bhs 23-24 Nd 90.0 6.8 2.9 1.1 1.6 21.5 44.0 5.2 4.4 17.8 6.2 9.4 30.2 4.6 Bs 24-28 5.1 94.9 4.2 0.9 2.9 1.0 3.7 4.0 7.8 1.3 5.2 14.2 19.6 9.3 3.0 Bw 28-60 5.1 96.1 2.4 1.4 2.5 0.9 2.1 2.9 6.4 1.1 3.8 13.2 20.3 7.5 3.8 Gaume (Haplic Podzol) E 19-35 5.0 94.0 3.1 2.9 0.0 0.0 0.10 2.1 0.04 Nd 1.3 12.6 7.6 10.9 2.81 Bh 35-40 4.7 89.0 7.0 4.0 0.1 0.3 4.58 16.7 1.28 Nd 14.4 11.0 7.8 14.5 2.87 Bhs 40-47 4.8 90.0 6.0 4.0 0.3 0.3 5.53 16.8 2.01 Nd 6.1 7.7 9.6 19.7 1.47 Bs 47-58 5.1 91.6 3.4 5.0 0.5 0.4 0.66 6.8 2.16 Nd 3.6 9.9 12.7 12.9 1.87 BC 70-100 4.6 92.9 2.5 4.6 0.2 0.1 0.08 2.8 0.65 Nd 0.7 9.8 13.9 13.8 1.37 a Dithionite- (dcb), oxalate- (ox) and pyrophosphate- (p) extractable contents of Fe, Al and Si b total organic carbon. Nd = not determined 231

Part 6: Appendix A2. Co-authored papers

The mineralogy of the clay fraction in the Cox Bay podzolic chronosequence is dominated by quartz, amphiboles, chlorites, vermiculite, mixed-layers minerals (MLM), smectite, illite, and kaolinite and evolve depending on soil age and the development of soil horizons (Figure A2.1.2). In the youngest soil profile (CB-120 yrs), the clay mineralogy is characterized by the presence of quartz, Na-feldspar and amphiboles as primary minerals and kaolinite, chlorite and illite as pedogenic clay minerals (data not shown). XRD patterns display similar mineral compositions in the E horizons of CB-175 yrs and CB-270 yrs profiles. In those soil horizons, peaks at 1.40, 1.00, 0.83 and 0.70 nm, correspond respectively to chlorite, illite, amphibole and kaolinite (disappearance of the 0.7 nm peak after K 550°C treatment). A band at 1.40 nm (Mg 20°C treatment) that shifts to 1.60-1.70 nm after Mg- eg treatment due to swelling indicates the presence of discrete smectite. In addition, the combination of the peaks at 1.40 nm after Mg-20°C and Mg-eg, and the collapse of the peak from 1.10 to 1.00 nm due to the dehydration after a K-saturation followed by heating correspond to vermiculite. Finally, the presence of a wide peak at 1.20 nm after Mg-20°C treatment that shifts after Mg-eg treatment indicates irregularly mixed-layers minerals (MLM).

In the CB-335 yrs profile, mineralogical differences were observed. In the E horizon, relative to the E horizons of CB-175 yrs and CB-270 yrs profiles, XRD patterns show a strong decrease of the abundance of kaolinite (the 0.70 nm peak has almost disappeared), absence of chlorite (no peak at 1.40 nm after K treatments), increase of the relative abundance of smectite compared to vermiculite (increase of the peak at 1.60-1.70 nm and almost no peak at 1.40 nm after the Mg eg treatment). In the Bh horizon relative to the E horizon of CB-335 yrs profile, XRD patterns show the presence of kaolinite and chlorite, absence of smectite (no swelling after Mg eg treatment), increase in the abundance of vermiculite, and a decrease of the abundance of MLM (smaller peak at 1.20 nm after Mg 20°C treatment).

The mineralogy of the Belgian Podzol (Gaume) is compared to the mineralogy of the Canadian Podzol. The primary minerals of the

232

Part 6: Appendix A2. Co-authored papers loess contain quartz, feldspars, micas and small amounts of trioctahedral chlorites and amphiboles (Van Ranst et al., 1982). The mineralogy of the clay fraction in the Belgian Podzol is comprised of vermiculite, smectite, hydroxyl-interlayered vermiculite, chlorite, MLM and kaolinite (Herbauts, 1982).

As we are not able to precisely quantify each type of 2:1 minerals on the powder of the clay fraction (chlorite, smectite, vermiculite, illite, MLM) with Siroquant Software, we carried out the clay mineralogy quantification in the soil chronosequence by separating the minerals in the clay fraction in 4 groups: quartz, amphiboles, kaolinite and 2:1 minerals (Figure A2.1.3). Compared to the mineralogy of BC horizon (CB-120 years) at the initial stage of soil formation (quartz = 15%, amphiboles = 63%, kaolinite = 5% and 2:1 minerals = 17%), the quantification of clay mineralogy indicates an increase of the relative abundance of kaolinite (+14%) and 2:1 minerals (+6%) in E horizon of the 175-year-old soil. Then we observe a strong decrease of the relative abundance of kaolinite in older and more weathered E horizons: -12% in the 270-year-old soil and -18% in the 335-year-old soil, while the relative abundance of 2:1 minerals is constant between the two oldest soils (= 18%). In the Bh and Bhs horizon of the 335-year-old soil, we note an increase of the relative abundance of kaolinite (+7% and +12%, respectively) compared to the stronger weathered E horizon (= 1%). The evolution of primary clay- sized minerals is characterized by a decrease of the relative abundance of amphiboles in the early stage of soil formation (-21%), then by a relative increase of the abundance (+14 and +19%) in the more weathered E horizons, which is related to the decrease of kaolinite, while quartz content remains quite constant (= 16±2%) during pedogenesis.

233

A. CB-175 yrs : E B. CB-270 yrs : E

0.7 0.70 1.63

1.39

1.39 0.99 0.83

1.62 Mg eg 0.83 1.19 1.20 0.99 Mg eg

Mg 20°C Mg 20°C

K 550°C K 550°C

K 300°C K 300°C

K 150°C K 150°C

K 20°C K 20°C

4 10 4 10 2 - theta [°] 2 - theta [°]

Figure A2.1.2 a - XRD patterns of the clay-sized fraction (<2µm) of soils of the Cox Bay soil chronosequence after six treatments: K-saturation followed by drying at 20, 105, 300 and 550°C, and Mg-saturation followed by drying at 20°C and saturation with ethylene-glycol. (A) CB-175 yrs E horizon, (B) CB-270 yrs E horizon. 234

C. CB-335 yrs : E D. CB-335 yrs : Bh

1.64

0.70 1.39 1.39 0.99 0.83 0.70 Mg eg 1.19 0.83 0.99

Mg eg Mg 20°C 1.20

K 550°C Mg 20°C

K 550°C

K 300°C K 300°C

K 150°C K 150°C

K 20°C K 20°C

4 10 4 10 2 - theta [°] 2 - theta [°]

Figure A2.1.2 b - XRD patterns of the clay-sized fraction (<2µm) of soils of the Cox Bay soil chronosequence after six treatments: K-saturation followed by drying at 20, 105, 300 and 550°C, and Mg-saturation followed by drying at 20°C and saturation with ethylene-glycol. (C) CB-335 yrs E horizon, (D) CB-335 yrs Bh horizon 235

Figure A2.1.3 - Quantitative evolution of the mineralogy in the clay-sized fraction of the Cox Bay soil chronosequence. The clay-size mineralogy is comprised of primary minerals (quartz and amphiboles) and pedogenic clay minerals (kaolinite and 2:1 minerals). Chlorite, vermiculite, smectite, illite and mixed-layers minerals (MLM) are the 2:1 aluminosilicates encountered in the podzolic chronosequence.

236

Part 6: Appendix A2. Co-authored papers

A2.1.3.2 Si isotopic modifications in the clay fraction over time

In our study, pedogenic clay minerals in the clay fraction of BC horizon are considered as “unweathered” secondary minerals compared to pedogenic clay minerals in more weathered horizon (E, Bh, Bhs and Bs) since BC horizon is not yet reached by the podzolic weathering front (Lundström et al., 2000a). In the Cox Bay soil chronosequence, we therefore compare the Si isotopic signatures of the clay fraction in each soil horizons with those in the “unweathered” clay fraction in the slightly weathered BC pedogenic horizon.

In the Cox Bay chronosequence, Si in the “unweathered” clay fraction (BC horizon; δ30Si = -0.52±0.16‰, 2SD, n=3) is isotopically lighter compared to the primary lithogenic minerals in the parent beach sand material (C material; δ30Si = -0.27±0.10‰, 2SD, n=3) (Figure A2.1.4 A). In the early phase of soil formation, the difference of Si isotope signature between the lihtogenic primary minerals in the sand fraction of the C material and the clay fraction in BC material, 30ɛ is -0.25‰ (min-max = -0.12 ˗ -0.37‰). This is not the fractionation factor due to precipitation of pedogenic clay minerals as the clay fraction also comprises lithogenic primary minerals.

Relative to the “unweathered” BC clay fraction (δ30Si = - 0.52±0.16‰, 2SD, n=3), the clay fraction of the topsoil eluvial E horizons shows depletion in light 28Si (i.e., less negative δ30Si values: from -0.33±0.02‰ to -0.10±0.22‰, Figure A2.1.4 B,C). The clay fraction in the subsoil illuvial Bh-Bs horizons is isotopically lighter (i.e., enriched in light 28Si) than “unweathered” BC clay fraction (δ30Si from -0.60±0.06‰ to -0.84±0.08‰ ‰; Figure A2.1.4 B,C). The magnitude of light Si depletion/enrichment in the clay fraction increases with soil 30 age, with ΔSiE-BC varying from +0.20‰ (at t=175 years) to +0.42‰ (at 30 t= 335years); and ΔSiB-BC varying from -0.17‰ (at t=175 years) to - 0.32‰ (at t= 335years).

A comparable depletion/enrichment in light 28Si in the clay 30 fraction during pedogenesis is found in the Belgian Podzol ( ΔSiE-BC = 30 +0.29‰; ΔSiB-BC = -0.27‰) from a similar temperate climate but with different parent material and rainfall conditions (Figure A2.1.4 D).

237

Part 6: Appendix A2. Co-authored papers

Figure A2.1.4 - first part

238

Part 6: Appendix A2. Co-authored papers

Figure A2.1.4 – second part. Silicon isotopic signature (δ30Si ‰; mean values ± standard deviation represented by error bars) in the clay-sized fraction depending on soil ages and in primary lithogenic minerals in the beach sand parental material. (A): 0- and 120-year-old soil fraction (δ30Si of primary minerals in beach sand in black and δ30Si of the clay fraction of the 120-year-old BC horizon in blue), (B): 175- and 270- year-old clay fractions (175 yrs = red Δ; 270 yrs = green ), (C): 335- year-old clay fraction (purple ), and (D): clay fraction in an undated Belgian Podzol (brown ). After only 175 yrs (B), we observe the depletion in light 28Si in the clay fraction of the eluvial E horizon and enrichment in light 28Si in the clay fraction of deeper illuvial soil horizon; respectively, relative depletion in light 28Si (+0.20‰) and relative enrichment in light 28Si (-0.17‰) compared to the original Si isotopic signature of the unweathered clay fraction in the BC horizon. The isotopic fractionation increases over time with an enrichment in heavy 30Si of +0.42‰ in the clay fraction of the E horizon and a concomitant enrichment in light 28Si of -0.32‰ in the clay fraction of the Bhs horizon (after 335 years). We observe exactly the same tendency in the Belgian Podzol with enrichment in light 28Si in the clay fraction of the Bhs horizon of -0.27‰ compared to the unweathered clay fraction in BC horizon.

239

Part 6: Appendix A2. Co-authored papers

A2.1.3.3 Geochemical modifications in the clay fraction over time

As the clay fraction becomes relatively more depleted in Si, the clay fraction becomes more enriched in light 28Si (Figure A2.1.5 A). Our results show that Si isotopic signature of the clay fraction becomes increasingly light with enrichment in Al (higher Al/Si ratio in the clay fraction) (Figure A2.1.5 B).

The enrichment in light 28Si (and the increase of Al/Si ratio) in the clay fraction also relates to an increase in the proportion of poorly- crystalline Si components in the clay fraction (estimated by the Siox/Siclay ratio). As the Si-bearing phases of the clay fraction accumulates poorly-crystalline aluminosilicates, the Si isotopic composition becomes more enriched in light Si isotope (Figure A2.1.5 C). We observe also that the enrichment in light 28Si in the clay fraction is not systematically related to a relative depletion in Ge, i.e. lower Ge/Si ratio (Figure A2.1.5 D).

A2.1.4 Discussion

A2.1.4.1 Evolution of clay-sized mineralogy

Different processes, such as transformation and neoformation, modify the chemical composition of the clay mineral within soil profile and control the clay content and mineralogy during pedogenesis. As water acts to mediate chemical reactions and to transport reactants and products from topsoil (Chadwick and Chorover, 2001), we observe the highest content of pedogenic subproducts (clay-sized minerals) in the top- and subsoil (0-24 cm). The depth where clay-sized minerals concentrate (~3%) increases over time, which highlights the deepening of the weathering front: 0-3 cm after 175 years, 0-7 cm after 270 years, 0-24 cm after 335 years.

240

A. B.

C. D.

Figure A2.1.5 - Evolution of Si isotope composition with elemental composition (Si, Al, Ge) and the proportion of poorly crystalline Si (Siox/Siclay) in the clay fraction for the Cox Bay soil chronosequence (175-year-old soil = red Δ; 270- year-old soil = green ; 335-year-old soil = purple ) and for the Gaume Podzol (brown )

241

Part 6: Appendix A2. Co-authored papers

We show that the chemical modifications of clay mineral structure in the podzolic weathering front mobilize Al (and Fe) and Si from secondary minerals over time. The evolution of Al/Si in the clay fraction substantiates the preferential mobilization of Al, relative to Si, during the dissolution of secondary clay minerals, in particular in the presence of organic acids with high complexing capacities, such as those encountered in Podzols (Sokolova, 2013, Stumm, 1992). The clay mineralogy evolution (Figure A2.1.2) in Podzols studied here under maritime temperate climate is very similar to the ones observed from postglacial moraines (Righi et al., 1999) and tills (Egli et al., 2002) in Switzerland. The aluminization of primary clay minerals, such as chlorites, leads to formation of irregularly-interstratified minerals in the moderately acid B horizons. In the stronger weathering E system, Al-removal from interlayers by organic complexing agents leads to the formation of vermiculite. Further alteration induces the formation of smectite-like minerals in the E eluvial horizon. Finally, the Siox content (Table A2.1.1) confirms that the formation of poorly-crystalline aluminosilicates (ITM) occurs when the concentration of organic acids is sufficiently low to allow the precipitation of Al with Si, as suggested by Ugolini and Dahlgren (1987) in the fulvate bicarbonate theory of podzolization. The clay mineralogy evolves with increasing weathering in the age sequence and formation of typical podzolic soil horizons (E, Bh, Bhs, Bs, Bw), which is in good agreement with the formation of two geochemical E and B compartments during podzolization (Ugolini and Sletten, 1991). The upper E horizon is controlled by organic acids as major proton donors and complexing metals, which leads to dissolution of primary and secondary minerals. In the lower B compartment, inorganic acid (carbonic and nitric acid) is the major proton donor where the weathering system is less aggressive.

Four important mineralogical evolutions are observed in the Cox Bay soil chronosequence, as a result of podzolization: (i) the neoformation of kaolinite, illite and chlorite from dissolution of primary minerals at the very beginning of soil formation, (ii) the disappearance of kaolinite in the strongest weathered E horizon, then (iii) the increase of relative abundance of kaolinite in Bh and Bhs horizons compared to E horizon (Figure A2.1.3), and finally (iv) the

242

Part 6: Appendix A2. Co-authored papers accumulation of imogolite-type materials in Bhs and Bs horizons (Table A2.1.1).

A2.1.4.2 Dissolution and re-precipitation of pedogenic clay minerals during podzolization

Since the clay fraction of soils comprises aluminosilicates and Fe-, Al-oxyhydroxides, Si in the clay fraction includes Si incorporated in primary minerals (quartz and amphiboles), secondary minerals (kaolinite and 2:1 minerals) and Si adsorbed onto Fe oxyhydroxides. In the Bhs horizon of the 335-year-old soil, the high content of free Fe -1 (Fedcb = 44 g.kg ) is in the same order of magnitude than in a weathering sequence in Cameroon (20-85 g.kg-1) (Opfergelt et al., 2009), for which the variations of δ30Si values in the clay fraction due to adsorption onto Fe oxides are known (Opfergelt et al., 2010). We have to take into account the pool of Si adsorbed onto Fe oxides in the clay fraction as this Si pool significantly influence the enrichment in light 28Si in the clay fraction: the difference of the Si isotope signature in the clay fraction of B horizons before and after dithionite-treatment (i.e., after the release of Si from the surface of Fe oxides) in the Cameroon weathering sequence varies between 0.08 and 0.45‰ (Opfergelt et al., 2010). However, all of the Fe in the Cameroon weathering sequence is in the clay fraction, while in the temperate soils of the Cox Bay chronosequence, only 20% of the bulk Fe content is in the clay fraction for Bhs horizon (=8.8 g.kg-1), where we observe the largest enrichment in light Si isotope. In eluvial E horizons, we observe the largest depletion in light 28Si while the Fe content in the clay fraction represent between 70 and 100% of the total Fe concentration in bulk soil (until 2.8 g.kg-1). The ratio of Fe oxides in the clay fraction to Si content in the clay fraction is similar between Bhs (14%) and E (13%) horizons, while the Si isotope composition in the clay fraction follows opposite trends in these two horizons. As a consequence, we assume that the δ30Si values of the clay fraction of Belgian and Canadian temperate soils can be considered representative of the Si isotopic composition of the primary and secondary silicates, and not significantly influenced by the fractionation of Si isotopes through adsorption onto Fe oxides. The

243

Part 6: Appendix A2. Co-authored papers role played by the Si adsorption onto Fe oxides on Si isotope compositions of the clay fraction must however be further investigated for tropical soils.

It is well established that the preferential incorporation of light 28Si during neoformation of secondary pedogenic minerals accounts for their isotopically lighter signature relative to primary lithogenic minerals (Georg et al., 2009, Opfergelt et al., 2010, Ziegler et al., 2005). The Si isotope composition of the soil clay fraction depends on the degree of soil weathering and the evolution of the clay mineralogy (Ziegler et al., 2005, Opfergelt et al., 2010, Opfergelt et al., 2012).

Using the quantification of primary minerals (quartz and amphiboles) and secondary minerals (kaolinite and 2:1 minerals) and the Si isotope signature of lithogenic primary minerals (-0.27‰), we can compute δ30Si value of “unweathered” secondary clay minerals in the clay fraction of BC horizon (-1.40‰; Table A2.1.2). The isotopic fractionation factor between primary lithogenic minerals and 30 30 30 secondary pedogenic minerals ( ε = δ Simin I - δ Simin I) is therefore - 1.13‰. The mass balance approach (Table A2.1.2) show also a progressive depletion in light 28Si in secondary minerals of the E horizon (from -0.51‰ to 0.64‰) and an enrichment in light 28Si in secondary minerals of the illuvial horizons (until -2.31‰). In identical bio-geo-climatic conditions, the Si isotopic fractionation associated with the dissolution of primary lithogenic minerals and neoformation of secondary pedogenic minerals should generate comparable Si isotopic signatures in the clay fraction in the entire soil profile with no evolution over time given identical fractionation factor between the primary and secondary Si pools. Here, we show that that the signature of secondary minerals varies in the soil profile and the relative depletion/enrichment in E and B horizons increases with time in the Cox Bay chronosequence. The dissolution of primary minerals and precipitation of secondary minerals therefore cannot explain the increasing depletion/enrichment in light 28Si in the clay fraction over time and with depth. This highlights that the evolution of δ30Si values in the clay fraction of the soil profiles observed here rules out the weathering of primary minerals (lithogenic Si pool) as the sole source for the neoformation of secondary minerals in the clay fraction.

244

Table A2.1.2 - Quantification of primary and secondary minerals in the clay-sized fraction of the Cox Bay soil chronosequence. The clay-sized quantification is then used for the Si isotopic mass balance approach

Measured data Computed 30 30 30 c Primary minerals Secondary minerals δ Si(‰) δ Si (‰) Δ SiBC-x (‰) (% in the clay (% in the clay in the clay fraction of pedogenic clay fraction)a fraction) mineralsb BC horizon (120 yrs) 78 22 -0.52 -1.40 - E horizon (175 yrs) 59 42 -0.32 -0.39 +1.01 E horizon (270 yrs) 75 25 -0.33 -0.51 +0.89 E horizon (335 yrs) 81 19 -0.10 +0.64 +2.04 Bh horizon (335 yrs) 72 28 -0.45 -0.92 +0.48 Bhs horizon (335 yrs) 70 30 -0.84 -2.31 -0.91 a mineralogy of the clay fraction quantified using the Siroquant software V4.0; primary minerals = quartz + amphiboles; secondary minerals = kaolinite + 2:1 minerals (vermiculite, smectite, illite, chlorite, mixed-layers minerals) b The δ30Si of secondary minerals present in the clay fraction is computed as follows: 30 30 30 δ Simin II = ((δ Siclay fraction – (% min I * δ Simin I)) / % min II)), 30 where min I = primary minerals, min II = secondary minerals and δ Simin I = -0.27‰ c Si isotope discrimination between “unweathered” clay minerals in BC horizon and pedogenic clay minerals in the “x” horizon of interest (x 30 30 30 30 = E, Bh or Bhs horizons): δ SiE - δ SiBC or δ SiBC - δ SiBh/B

245

Part 6: Appendix A2. Co-authored papers

Germanium (Ge), a chemical analog of Si, generally follows similar inorganic geochemical pathways than Si (Froelich and Andreae, 1981). However, secondary pedogenic (clay) and biogenic (phytoliths) minerals display contrasting Ge/Si ratios: neoformed clay minerals are enriched in Ge (higher Ge/Si) while biogenic silica polymerized in plants as phytoliths is depleted in Ge (lower Ge/Si) (Derry et al., 2005, Kurtz et al., 2002). Although there is a negative relationship between Ge/Si ratios and δ30Si in the youngest soils (Cambisols) of the Canadian soil chronosequence, the absence of a relationship between Ge/Si and δ30Si ratios in the oldest soil (Podzol) of the Canadian chronosequence and in the Belgian Podzol (Figure A2.1.5 D) allows us to dismiss the dissolution of phytoliths (biogenic Si pool) as a major source of Si for clay neoformation. This process would be characterized by enrichment in light 28Si and depletion in Ge in secondary clay minerals relative to beach sand parent material, as phytoliths are Ge-depleted (low Ge/Si ratio) relative to primary minerals (Derry et al., 2005).

The mass balance approach (Table A2.1.2) shows that the enrichment in light 28Si of secondary minerals of Bhs horizon (-2.31‰) compared to the “unweathered” secondary minerals in the BC horizon (-1.40‰) partly explains the depletion in light 28Si of secondary minerals in the clay fraction of E horizon (+0.64‰) for the oldest soil (Podzol CB-335 years). Our data highlight that the isotopic fractionation due to preferential release of light 28Si during dissolution 30 of secondary minerals in the E horizon (Δ Si E-BC = +2.04‰) partly accounts for the enrichment in light 28Si during re-precipitation of new 30 clay minerals in Bhs horizon (Δ Si Bhs-BC = -0.91‰). This combined with the fact that kaolinite are progressively dissolved in the E horizon and is almost completely dissolved in the strongly weathered E horizon (CB-335 years) (Figure A2.1.3), highlight that 28Si is redistributed in the soil profile through re-precipitation of new pedogenic clay minerals deeper in the soil profile and leaching. As a part of Si precipitating during the neoformation comes from the dissolution of secondary clay minerals, we name those new clay minerals as “tertiary minerals”. This implies that the preferential lessivage of clay particles enriched in light 28Si and the resulting relative accumulation of primary clay-size minerals in topsoil cannot be responsible for the on-going enrichment

246

Part 6: Appendix A2. Co-authored papers in light 28Si in the clay fraction. Indeed, the increasing enrichment in light 28Si in new tertiary minerals (tertiary kaolinite) in B horizons can only be related to a Si source progressively enriched in light 28Si over time. Kaolinite seems to play a key role in the successive formation of clay minerals as the content of 2:1 clay minerals is quite constant during pedogenesis in the soil chronosequence (Figure A2.1.3).

The preferential release and incorporation of light 28Si during dissolution and re-precipitation of clay minerals in the pedogenic Si pool accounts for the Si isotopic depletion/enrichment in the clay fraction over time in the podzolic chronosequence. The preferential incorporation of light 28Si during precipitation of Si released from the dissolution of pedogenic clay minerals (in E and Bh horizons) explains the increasing enrichment in light 28Si in newly-formed clay minerals (tertiary clay minerals in Bhs horizon) during podzolization. This is confirmed by the fact that pedogenic clay minerals in E horizons are increasing heavier over time (Table A2.1.2), showing that the dissolution of pedogenic clay minerals discriminate against the release of heavy 30Si as already demonstrated for diatoms (Demarest et al., 2009) and crystalline basalt (Ziegler et al., 2005). Besides the lithogenic and biogenic Si pools, we provide evidence that pedogenic Si pool is therefore involved in the neoformation of pedogenic clay minerals and as such in the evolution of their Si isotope signatures (Figure A2.1.6).

247

Figure A2.1.6. Conceptual representation of the contribution of Si released from the dissolution of primary and secondary Si pools (lithogenic, biogenic and pedogenic) to the re-precipitation of new “tertiary” clay minerals during podzolization. Phase I → Phase II (C → BC) = transition from the parent C material to the pedogenic BC horizon with neoformation of secondary clay minerals. Phase II = formation of typical podzolic soil horizons: E, Bh, Bhs, Bs and Bw. Phase II → Phase III on the other hand) = transition from young to older Podzol characterized by (i) the deepening of 30 the E horizon where secondary clay minerals are weathered and enriched in heavy Si (EII → EIII), and (ii) B horizons 28 tertiary clay minerals re-precipitate and are enriched in light Si (BCII → BhsIII). 248

Part 6: Appendix A2. Co-authored papers

A2.1.4.3 Implications for podzolization theory

For the first time, we document enrichment in light 28Si in secondary clay minerals over time in a podzolic soil chronosequence. 28 The highest enrichment in light Si and oxalate-extractable Siox in Bhs/Bs horizons relative to E/Bh horizons (Figure A2.1.5 C; Table A2.1.2) highlight that the dissolution of secondary aluminosilicates in E/Bh horizons act as a Si source for formation of poorly-crystalline aluminosilicates (imogolite-type materials ITM) in Bhs/Bs horizons. The release of Si from the dissolution of primary and secondary clay minerals and precipitation of dissolved Si with Al released by microbial decomposition from the organic ligands (Lundström et al., 1995) can explain the formation of ITM in Bhs/Bs horizons (Ugolini and Dahlgren, 1987). During podzol development, ITM undergo additional dissolution for then re-precipitating Si as crystalline tertiary clay minerals in Bhs horizon. The evolution of Si isotopic signature in pedogenic clay minerals of the podzolic soil chronosequence therefore corroborates the process of dissolution and re-precipitation of aluminosilicate phases during podzoliaztion (fulvate bicarbonate theory; Ugolini and Dahlgren (1987)). We can infer that low contents of poorly-crystalline ITM in the Bhs/Bs horizons play a key role in the evolution of Podzols and the progressive enrichment in light 28Si in pedogenic clay minerals. The absence of ITM in the Bh horizon and the lighter δ30Si in Bhs/Bs indicates their high reactivity during podzolization, dissolving as organic-rich Bh horizon forms and precipitating as Fe-,Si-,Al-enriched Bhs/Bs horizons form. This is confirmed by the high reactivity of ITM also reflected in Ge/Si and δ30Si patterns in soil solutions of the Santa Cruz soil chronosequence, which indicates seasonal precipitation and dissolution of hydroxyaluminosilicates such as allophane (White et al., 2012). The 30 positive correlation between Siox/Siclay and δ Si values in the clay fraction (Figure A2.1.5 C) highlights that during podzolization, pedogenic clay minerals become enriched in light 28Si together with Al in the poorly-crystalline part of the clay fraction. Based on these findings, poorly-crystalline aluminosilicates can be regarded as a temporary reactive reservoir of light 28Si in Bs horizon. This reservoir acts as a source of light 28Si in tertiary crystalline clay minerals, such as

249

Part 6: Appendix A2. Co-authored papers tertiary kaolinite, in Bhs horizon that will develop in the current Bs horizon during podzolization. The dissolution and re-precipitation of pedogenic clay minerals is therefore an important podzol-forming process (Figure A2.1.6).

A2.1.4.4 Implications for tracing the effects of environmental changes on soils

In the Cox bay soil chronosequence, we show that the production of acidity (protons and complexing organic acids) in temperate forests and the subsequent Podzol formation implies heavy 30Si enrichment in pedogenic clay minerals of E horizons relative to the 30 “unweathered” clay minerals in BC horizon; Δ SiE-BC increasing from +0.89 to +2.04‰ in ca. 200 years (Table A2.1.2). The preferential loss of light 28Si in weathered clay minerals in E horizons compared the “unweathered” clay minerals in BC horizon is recorded in the Si isotope signature of pedogenic clay minerals on very short time-scale. Moreover, the Si isotope fractionation between the “unweathered” clay minerals in BC and pedogenic clay minerals precipitating in Bhs 30 28 (Δ Si BC-Bhs) of -0.91‰ highlights that a part of light Si released in topsoil is used for re-precipitation in the subsoil (Table A2.1.2). As a consequence, Si isotope signatures in the clay fraction of soils can be used to trace the modifications of pedogenic clay minerals in other soil-plant systems, e.g. in highly weathered tropical and subtropical environments (Ferralsols, Lixisols, Nitisols, …), in frozen soils (Cryosols), in soils characterized by illuviation of clay minerals (Luvisols), in young soils (Cambisols) and in soils with high biological activity (Chernozems). Si isotope composition of pedogenic clay minerals can be useful to trace and quantify the impact of environmental changes (temperature, rainfall, acid deposition, land use …) on pedogenic clay evolution. This is central to a better understanding of soil development and associated terrestrial biogeochemical processes.

250

Part 6: Appendix A2. Co-authored papers

A2.1.5 Conclusions

The process of dissolution of pedogenic clay minerals during podzolization is confirmed by the Si isotopic signature of the clay fraction in a podzolic soil chronosequence (Cox Bay, Vancouver Island). Our dataset shows Si isotopic, geochemical and mineralogical trends with depth and as a function of pedogenic time, providing an orthogonal dataset which sheds light on the origin and evolution of pedogenic clay minerals in the clay fraction. The depletion in light 28Si in pedogenic clay minerals in topsoil increases over time (from +0.89 to +2.04‰) and a part of light 28Si released accounts for the relative enrichment in light 28Si in pedogenic clay minerals in subsoil (-0.91‰). This highlights that Si released from the partial dissolution of secondary clay minerals in topsoil contributes to the neoformation of tertiary clay minerals in subsoil. Clay mineral dissolution has often been regarded as an irreversible process, while the increase of 28Si enrichment over time in the clay fraction documented in this study indicates successive formation of clay minerals, which depends on the downward movement of the weathering front in the soil. The continuous weathering of pedogenic clay minerals is an important process in the formation of Podzols as we show that the Si released in soil solution contributes to the reformation of clay minerals deeper in soils over very short time scales (ca. 300 years). The recording of Si isotopic ratios in the clay fraction as a function of the age of soil formation is therefore an untapped resource for tracing pedogenic processes controlling the Si incorporation in pedogenic clay minerals during podzolization, and offering new perspectives for unraveling the genesis of pedogenic subproducts in various soil types. This has important implications as the process of dissolution and re- precipitation of pedogenic clay minerals would play a major role in several soil biogeochemical processes such as the retention of plant nutrients, the preservation of organic carbon from microbial decomposition, and the transfer of elements and pollutants from land to ocean. Further investigations are needed for quantifying the contribution of pedogenic Si pool to newly-formed clay minerals (tertiary, quaternary …) compared to the contribution of lithogenic and biogenic Si pools. Our dataset shows that the Si isotope

251

Part 6: Appendix A2. Co-authored papers compositions of soils are not only influenced by biogenic (phytolith formationb/dissolution) and litho-, pedo-genic processes (primary mineral dissolution and secondary mineral precipitation) but also by a more advanced weathering process, i.e. successive formation of pedogenic clay minerals. This should be taken into account when δ30Si values of the bulk soil and soil solutions are used for studying soil weathering degree and tracing dissolved and particulate Si transferred from soil-plant systems to the hydrospehere.

252

Part 6: Appendix A2. Co-authored papers

A2.2 Can Fe isotope fractionations trace the pedogenetic mechanisms involved in podzolization?8

Abstract

Stable Fe isotopes have shown the potential for tracing pedogenetic processes. Large isotopic fractionations were especially observed in Podzols. Nevertheless, a clear link between isotopic fractionation and elementary processes still needs to be established. To spatially distinguish the mechanisms successively involved in pedogenesis, we studied a podzolic chronosequence from Vancouver Island (British Columbia, Canada). We analyzed depth variations in soil properties (pH, particle size fractions, organic carbon, Si/Al ratio, Zr, Ti), Fe concentrations in different Fe pools, and Fe isotopic compositions. The Si/Al ratio, Zr, and Ti demonstrated that the Cox Bay Podzols developed from the same parental material, satisfying the requirements of a chronosequence. We showed that acidification, to a pH of 4.7, was a prerequisite for the start of podzolization. This first phase took place after 270 years. Furthermore, we observed that once the required pH was reached (4.7), podzolization occurred rapidly over 50 years. We found that Fe isotope fractionation in the A/E horizon was linked to mineral dissolution, while in the podzolic B horizons this fractionation was clearly associated with the accumulation of Fe- organic complexes and poorly crystalline Fe oxyhydroxides.

8 Adapted from Fekiacova, Z., Vermeire, M. L., Bechon, L., Cornelis, J. T., Cornu, S., 2017, Can Fe isotope fractionations trace the pedogenetic mechanisms involved in podzolization? Accepted in Geoderma. DOI: 10.1016/j.geoderma.2017.02.020

253

Part 6: Appendix A2. Co-authored papers

A2.2.1 Introduction

Iron is one of the major elements on Earth and is ubiquitous in rocks, minerals and biological samples. Stable Fe isotope analyses provide valuable information on the fate of Fe and are a promising tool for addressing questions on pedogenetic processes and soil evolution (e.g., Beard et al. (2003)), particularly the identification of the mechanisms involved in a pedogenesis. While a range of isotopic variations from -0.62 to +0.72 ‰ have been documented in soils (Wiederhold et al., 2007a, Wiederhold et al., 2007b, Emmanuel et al., 2005, Fantle and DePaolo, 2004, Poitrasson et al., 2008, Thompson et al., 2007, Fekiacova et al., 2013) no clear link between isotopic variations and pedogenetic mechanisms has yet been established. Iron isotopic fractionation related to pedogenesis is complex and is a function of the soil type. Isotopic fractionation during silicate weathering in a young ecosystem showed that a release of preferentially light Fe during silicate dissolution is a dominant mechanism and a source of light Fe for the newly formed Fe(III)- oxyhydroxides (Kiczka et al., 2010). Yet, other processes (e. g., precipitation, adsorption, organic matter complexation) are involved in soil formation and may impact the Fe isotopic signature. In particular, significant isotopic fractionation was documented in soils impacted by redox processes, i.e., Podzols and Gleysols (e.g., (Wiederhold et al., 2007a, Wiederhold et al., 2007b, Fekiacova et al., 2013).

It is generally accepted that podzolization consists of two key processes: (i) mobilization from the surface horizon and translocation of organic matter and Fe, Al, Si and (ii) immobilization and stabilization of organic matter and Fe, Al in the subsoil (Lundström et al., 2000a, Buurman and Jongmans, 2005). Sauer et al. (2007) discussed different hypotheses on the elementary mechanisms behind these two processes. We hypothesized that Fe isotopic fractionation could be used to trace elementary mechanisms in redox processes. To do so, we needed to understand the behavior of stable Fe isotopes during podzolization.

254

Part 6: Appendix A2. Co-authored papers

The aim of this paper was therefore to examine the impact of different mechanisms of podzol formation on the behavior of stable Fe isotopes. To spatially distinguish the mechanisms successively involved in pedogenesis, we studied Fe isotopic fractionation along a podzolic chronosequence from Vancouver Island (British Columbia, Canada).

We first verified that the studied soils satisfied the terms of a chronosequence as defined by Sauer (2015), and demonstrated by Barrett and Schaetzl (1992), VandenBygaart and Protz (1995), notably. Then, we studied the dynamics of podzolization on the basis of classical pedological data and mass balance calculations. Lastly, we linked the observed Fe isotopic variations in the Podzol depth profiles to the different soil forming mechanisms.

A2.2.2 Sampling and methods

A2.2.2.1 Site and sampling

The studied soil sequence was sampled in Cox Bay, on Vancouver Island (British Columbia, Canada) and has been well characterized elsewhere (Singleton and Lavkulich, 1987, Vermeire et al., 2016). The climate in this area is classified as maritime temperate, i.e. Cfb: without a dry season and with a warm summer, according to the Köppen classification (Peel et al., 2007) and is characterized by heavy mean annual precipitation (3200 mm) (Singleton and Lavkulich, 1987). Cox Bay soils developed from the beach sand, along a gentle slope: the elevational difference between any two of the sampled sites was less than 0.5 m (Singleton and Lavkulich, 1987). The vegetation cover was uniform over all sampled pits and was dominated by Sitka spruce (Picea sitchensis), salal (Gaultheria shallon) and Douglas fir (Pseudotsuge menziesii) (Cornelis et al., 2014a). The Cox Bay beach sand is derived from the Tofino Area Graywacke Unit (Singleton and Lavkulich, 1987). The sampled soil sequence was designed to represent different stages of soil formation from the beach sand to a well-developed Podzol aged 530 years (Cornelis et al., 2014a). It consisted of one beach sand sample and 5 soil pits: two Cambisol and

255

Part 6: Appendix A2. Co-authored papers three Podzol profiles. A 94 m long transect was designed perpendicularly to the present day shoreline. A surface age was attributed to each soil profile, using dendrochronology and geomorphology data. Tree ages were obtained using increment bores for the largest trees in the immediate vicinity of each site and by counting tree rings by Singleton and Lavkulich (1987) and by Cornelis et al. (2014a). Ages obtained by Singleton and Lavkulich (1987) and Cornelis et al. (2014a) are in good agreement. Taking the average rate of sand deposition into account, i.e., 0.26 m per year (Singleton and Lavkulich, 1987) and assuming a linear deposition rate with time between the successive oldest trees, a 13-m strip of active beach sand containing tree seedlings would have accumulated in approximately 50 years (Singleton and Lavkulich, 1987). Therefore 50 years were added to the tree age obtained by Cornelis et al. (2014) for each site to determine the site age. These ages range from 120 to 530 years.

The beach sand was collected on the beach and corresponds to a composite sample made of several sampling points collected down to 40 cm depth. For the soil profiles, a continuous sampling of each horizon was performed in pits.

A2.2.2.2 Soil properties

Organic carbon (OC) was analyzed by dry combustion (NF ISO 10694), particle size fractions were determined by wet sieving and sedimentation separation according to Stoke's law (NF X 31-107). These analyses were performed at the Laboratoire d’Analyses des Sols (LAS) in Arras. Soil pH was measured in 5 g : 25 ml soil : water suspension (Vermeire et al., 2016). Total Zr, Ti, Fe, Si and Al in different horizons of the Cox Bay Podzol profiles were analyzed by ICP-AES, after fusion in Li-metaborate + Li-tetraborate at 1000°C (Chao and Sanzolone, 1992) at UCL, Belgium. Total Zr and Ti are considered as immobile elements during weathering (Van Breemen and Buurman, 2002). Furthermore, in the soils, Fe is present in different forms, i.e., Fe-bearing silicate minerals and oxyhydroxides, and organic matter bound-Fe, adsorbed and exchangeable forms, which were determined using partial extractions. Pyrophosphate, oxalate (Tamm in darkness,

256

Part 6: Appendix A2. Co-authored papers

Tamm (1922)) and dithionite-citrate-bicarbonate (DCB, Mehra and Jackson (1960)) Fe extractions were performed on bulk soil samples. We considered that (i) pyrophosphate extracted mainly adsorbed and exchangeable forms (generally negligible), organic matter-bound (OM-bound) and small Fe colloids (Jeanroy and Guillet, 1981); (ii) oxalate also extracted poorly-crystalline Fe oxide while (iii) DCB extracted in addition crystalline Fe oxide pools (Hall et al., 1996). Partial extractions of different Fe pools were performed at LAS Arras (France), with the exception of pyrophosphate extraction. Concentrations of Fe in different pools were calculated by the difference between the total Fe and the different Fe concentrations extracted. Therefore, organic matter-bound (OM-bound) and small Fe colloids correspond to the pyrophosphate fraction, poorly-crystalline Fe oxides to the difference between oxalate and pyrophosphate fractions, crystalline Fe oxides to that between DCB and oxalate fractions and silicate-bound Fe to that between total Fe and DCB fractions.

A2.2.2.3 Iron isotope analyses

Iron isotopic analyses were performed on the bulk soil samples. Sample preparations were carried out in a laminar flow box, at INRA- GSE Aix-en-Provence. Approximately 250 mg of sample powder was first treated with 30% H2O2 in order to eliminate organic matter. After H2O2 treatment, the solution was evaporated to dryness and the residue dissolved using a mixture of concentrated HF-HNO3, followed by concentrated HCl acid, at ~130 °C. We verified that the digestion data were equivalent, within analytical error, to the total Fe concentration data obtained after fusion in Li-metaborate + Li- tetraborate at 1000°C.

The Cu-Fe fraction was isolated using the AG1X8, 200-400 mesh resin (Moynier et al., 2006). Iron was further separated and purified by anion exchange chromatography using the AG MP1, 100– 200 mesh, chloride form (Maréchal et al., 1999). All reagents were ultrapure distilled acids and overall procedural blanks contained negligible quantities of Fe (< 0.1 ‰) compared to the total dissolved

257

Part 6: Appendix A2. Co-authored papers sample Fe content. Iron isotope analyses were performed using MC- ICPMS Neptune at the Pole Spectrometrie Ocean, Ifremer, Brest, in high-resolution mode. We used Ni (NIST 986) as an internal standard and standard-sample-standard bracketing with IRMM-014 to correct for instrumental mass discrimination. Sample solutions were diluted to match the concentration of the IRMM-014 and internal standards within 5 %, i.e. 3 ppm Fe and 4 ppm Ni. Iron isotope results are reported relative to the IRMM-014 standard using the conventional δ notation:

56 56 54 56 54 δ Fe=[( Fe/ Fe)sample/( Fe/ Fe)IRMM-014-1]× 1000 (‰) (1)

The external reproducibility (2σ) calculated on the basis of repeated measurement of the IRMM-014 standard was 0.10 and 0.16 ‰ (N = 64) for δ56Fe and δ57Fe, respectively. In a δ57Fe vs. δ56Fe diagram, all soil sample measurements plot along a line with a slope of 1.44 which is equal, within error margins, to the theoretical value of ln(M57/M54)/ln(M56/M54) = 1.487, indicating mass-dependent fractionation and no influence of isobaric interferences.

A2.2.2.4 Mass balance calculations

In order to investigate the dynamics of the different mechanisms involved in podzolization, i.e., acidification, organic matter accumulation and Fe mobilization (section 4.2), elemental fluxes were calculated on the basis of Brimhall’s equation (Brimhall et al., 1991): mj, flux = 1/1000 × (ρref × Cj, ref × Th × τj, w)/(εi, w +1) (2)

-2 where m represents the flux of element j (g.cm ), ρref the bulk density -3 of the reference material (g.cm ), Cj, ref the concentration of element j -1 in the reference material (g.kg ), Th the horizon thickness (cm), τj, w the mass fraction of element j gained or lost from the weathered product with respect to the mass originally present in the reference material and εi, w the soil volume change over time calculated using an immobile element i.

258

Part 6: Appendix A2. Co-authored papers

We considered Zr (see section 2.1) as the immobile element since Ti can be mobilized through complexation with organic compounds (Cornu et al., 1999, Righi et al., 1997, Dumon and Vigneaux, 1979). The material from which the podzolization developed was taken as a reference. The elemental flux is presented as flux per cm (mj, flux/Th).

A2.2.3 Results

A2.2.3.1 Depth evolution of soil properties along the chronosequence

The evolution of pH through the studied profiles is shown in Table A2.2.1 and Figure A2.2.1. While the beach sand had basic pH (7.7), the Cambisols at pits 1 and 2 showed close to neutral pH (6), which decreased progressively towards clearly acid values, mainly lower than 5 in the upper horizon at pits 3, 4 and 5.

Total Fe concentration in the Cambisol at pit 2 and the Podzol at pit 3 was uniform with depth (Table A2.2.1, Figure A2.2.1). In the Podzols, at pits 4 and 5, we observed that upper E horizons were significantly depleted, while the subsurface Bhs horizons were significantly enriched in Fe, compared to the deep BC-horizon (Figure A2.2.1). Similarly, we observed an accumulation of OC concentration although its maximum was recorded in the Bh while the maximum of Fe was found in the Bhs horizon. In contrast, in the Cambisol (pit 2), the maximum concentration in OC was observed in the surface horizon and decreased exponentially with depth (Figure A2.2.1). Podzol at pit 3 represents an intermediate situation with constant OC concentration in both E and Bh horizons.

In the studied chronosequence, Zr concentrations were uniform along the profiles, with the exception of the two Podzols at pits 4 and 5 that showed a significant enrichment in the upper E horizons (Table A2.2.1). For Ti, enrichment in the Bhs horizon was observed in those pits, confirming possible mobilization by organic compounds while Zr remained immobile.

259

Table A2.2.1: Pedological data, Fe isotopic compositions and concentrations of total Fe in the Cox Bay Podzols. Particle size distribution and organic C (OC) were determined at the Laboratoire d’Analyses des Sols (LAS) Arras. Sampling Fe Pit Age, δ56Fe ±2σ, OC, Clay, Silt, Sand, Zr tot, Ti tot, Soil type Horizon depth, tot, pH Si/Al number years ‰ water g.kg-1 g/kg g.kg-1 g.kg-1 g.kg-1 g.kg-1 cm g.kg-1 0 Beach sand 0 C 0 - 40 0.15 ± 0.06 48.92 7.7 0.94 22 8 970 3.8 0.23 7.14 1 Cambisol 120 BC 0 - 75 0.13 ± 0.01 34.64 5.9 4.65 25 17 958 4.1 0.12 4.39 2 Cambisol 175 E 0 - 3 nd 29.70 5.4 34.10 41 63 896 nd 0.12 4.46 Bw 3 - 44 0.15 ± 0.08 33.62 5.8 4.28 21 20 959 4.4 0.09 4.10 BC 44 - 75 nd 25.00 5.9 2.44 21 9 970 4.8 0.07 2.95 3 Podzol 270 E 0 - 7 0.11 ± 0.07 18.99 4.6 17.00 43 40 917 5.4 0.08 2.87 Bh 7 - 23 0.05 ± 0.11 20.67 5.1 18.60 10* 17* 972* 5.1 0.06 2.33 Bw 23 - 57 0.16 ± 0.08 18.43 5.3 11.60 16 7 977 5.5 0.08 2.20 BC 57 - 75 0.15 ± 0.04 25.87 5.4 8.33 9 8 983 4.7 0.07 3.02 4 Podzol 335 E 0 - 10 0.20 ± 0.04 10.81 4.9 8.73 50 133 817 6.7 0.20 3.39 Bh 10 - 17 0.14 ± 0.14 25.29 5.5 21.30 24 29 947 4.8 0.09 3.50 Bhs 17 - 18 -0.02 ± 0.11 37.76 nd 17.30 nd nd nd 4.5 0.08 3.67 Bs 18 - 23 0.10 ± 0.10 31.18 5.3 5.03 10 19 971 4.4 0.08 3.59 Bw 23 - 63 0.11 ± 0.08 25.27 5.4 3.06 8 11 981 4.8 0.10 3.04 BC 63 - 113 0.13 ± 0.01 24.87 5.3 1.95 7 7 986 5.0 0.07 3.01 5 Podzol 530 E 0 - 8 0.25 ± 0.04 10.66 4.5 21.40 53 179 768 8.9 0.23 3.11 Bh 8 - 9.5 0.14 ± 0.10 19.55 4.5 33.40 94 114 792 6.1 0.19 4.17 Bhs 9.5 - 10 0.00 ± 0.05 39.25 4.5 32.80 nd nd nd 5.4 0.08 4.36 Bs 10 - 15 0.10 ± 0.01 29.29 4.8 12.50 17 35 948 4.8 0.07 3.00 Bw 15 - 40 0.19 ± 0.10 26.79 5.0 1.46 15 7 978 4.8 0.07 3.21 BC 40 - 60 0.09 ± 0.10 26.28 5.1 0.78 3 6 991 4.9 0.07 3.09 Data for Fe isotopic compositions are from this work. Total Fe (the fusion data), Zr, Ti, Si, Al concentrations and pH are from Vermeire et al. (2016). nd = not determined, * data from Cornelis et al. (2014).

260

Figure A2.2.1: Total Fe (a) and organic C (OC) (b) concentrations, pH (c) and Fe isotopic compositions (d) in the Cox Bay soils. Error bars in δ56Fe plots represent the 2σ. Error bars in Fe tot and OC plots correspond to 5%.

261

Part 6: Appendix A2. Co-authored papers

In the Cox Bay soils, at pits 1 to 3, the Si/Al ratio remained stable, with an average value of 4.7. In contrast, in the Podzols at pits 4 and 5, significant enrichment was detected in the upper E horizons (Table A2.2.1).

A2.2.3.2 Fe partial extractions

Significant changes were recorded vertically and laterally, in proportions of different Fe pools (Table A2.2.2, Figure A2.2.2). The Bh, Bhs and Bs horizons of pits 4 and 5 had smaller proportions of silicate- bound Fe and larger proportions of oxides and OM-bound Fe compared to the other horizons. Furthermore, E-Bh-Bhs-Bs horizons showed variations in the proportions of different Fe pools laterally along the chronosequence. Changes from pit 3 to pit 4 were restricted to the Bhs horizon (absent at pit 3), where an increase in the OM- bound and colloidal Fe pool and a respective decrease in the silicate- bound Fe pool were observed. In contrast, major variations were observed from pit 4 to pit 5: (i) in the Bh horizon, the Fe-silicates dropped from 82 to 48% while OM-bound and colloidal, poorly- crystalline and crystalline oxide pools increased from 5 to 17%, from 5 to 22% and from 8 to 13 %, respectively; (ii) similarly, in the Bhs horizon, the Fe-silicates dropped from 61 to 39% while poorly- crystalline oxides increased from 7 to 24% and (iii) in the Bs horizon, the Fe-silicates dropped from 82 to 61% while poorly-crystalline and crystalline oxide pools increased, from 14 to 27% and from 3 to 9 %, respectively. These changes record mineral transformations during podzolization.

A2.2.3.3 Fe isotopes depth evolution along the soil sequence

In the beach sand, the Cambisols at pits 1 and 2 and the Podzol at pit 3, the δ56Fe had an average value of 0.13‰ with no significant variation outside analytical error (Table A2.2.1, Figure A2.2.1). In contrast, δ56Fe values increased in the E horizons and decreased in the B horizons at pits 4 and 5 (Figure A2.2.1). The highest δ56Fe values were observed in the E horizons and correspond to the minimum of Fe

262

Part 6: Appendix A2. Co-authored papers concentration. The lowest δ56Fe values observed in the Bhs horizons correspond to the maximum of Fe concentration and high OC concentrations. These isotopic depth variations are in agreement with previously published data of Fe isotopes in Podzol profiles (Wiederhold et al., 2007b).

A2.2.4 Discussion

In the following section we discuss the evolution of the studied soil in space and time. To ensure that the Cox Bay soils satisfied the requirements of a chronosequence we first identified invariable soil forming factors and determined a reference material for podzolization for each soil pit. We then used the chronosequence to decipher the dynamics of different processes involved in podzolization. Finally, we attempted to assign Fe isotopic fractionation to a specific podzolization mechanism.

A2.2.4.1 Lateral and vertical variations in the soil parent sediment prior to pedogenesis

A chronosequence is a series of soils that developed from the same parental material, vegetation, topography and climate, with pedogenesis duration being the only varying factor (Jenny, 1961, Huggett, 1998, Schaetzl et al., 1994). As described above (section 2.1), distances between pits are short and the slope is gentle, thus both climate and relief can be considered as constant over the soil sequence. Nevertheless, the sediment composition may have varied over the 530 yrs of its deposition, both laterally and vertically, due to variations in proportions among different potential sources.

263

Table A2.2.2: Concentrations of different Fe pools in the Cox Bay Podzols.

Pit Soil type Age, Horizon Sampling Fe Pyro, Fe oxalate, Fe DCB, g.kg-1 number years depth, cm g.kg-1 g.kg-1 0 Beach sand 0 C 0 - 40 0.03 7.61 7 1 Cambisol 120 BC 0 - 75 0.51 3.16 4.29 2 Cambisol 175 E 0 - 3 0.72 2.14 3.87 Bw 3 - 44 0.66 2.96 3.98 BC 44 - 75 0.30 2.54 3.29 3 Podzol 270 E 0 - 7 0.87 1.92 3.15 Bh 7 - 23 1.42 2.86 3.69 Bw 23 - 57 1.17 2.7 3.58 BC 57 - 75 0.85 2.94 3.86 4 Podzol 335 E 0 - 10 0.63 1.07 1.54 Bh 10 - 17 1.33 2.54 4.5 Bhs 17 - 18 9.96 12.5 14.8 Bs 18 - 23 0.29 4.57 5.54 Bw 23 - 63 0.09 3.29 4.08 BC 63 - 113 0.05 2.68 3.29 5 Podzol 530 E 0 - 8 0.47 0.88 1.9 Bh 8 - 9.5 3.29 7.65 10.1 Bhs 9.5 - 10 12.44 21.8 24.1 Bs 10 - 15 0.85 8.89 11.5 Bw 15 - 40 0.03 3.16 3.72 BC 40 - 60 0.02 2.63 3.14 Partial extractions of Fe-oxalate (Fe oxalate) and Fe-DCB (Fe DCB) were performed at LAS Arras. Pyrophosphate extraction (Fe Pyro) data are from Vermeire et al., 2016. nd = not determined.

264

Figure A2.2.2: Variations in the proportions of different Fe pools in pits 2, 3, 4 and 5 of the Cox Bay soils. For convenience, the proportions are reported as numbers over each bar. 265

Figure A2.2.3: Results of the principal component analysis performed on the beach sand (BS), BC and Bw horizons of the studied soil profiles. (a) Variable space, (b) Observation space.

266

Part 6: Appendix A2. Co-authored papers

I. Weathering vs parent material source change

To assess these variations, we performed principal component analysis (PCA) considering the sand, the total Fe, Zr, Ti concentrations and the Si/Al ratios, which can be considered as a proxy for the quartz content (as quartz does not contain Al). We included in the analysis all the horizons of the studied soil profiles that experienced neither weathering nor podzolization, i.e., beach sand (BS), BC and Bw horizons (Figure A2.2.3).

The first two principal components of the PCA explained 94% of the variance of the dataset. The first principal component explained 79% of the variance. This component opposed two groups of variables: the Si/Al ratios representing the quartz fraction with a loading of - 0.931 and total Zr, Ti and Fe with loadings higher than 0.9 (Figure A2.2.3a). This first principal component is classically encountered in soils and is interpreted as a dilution factor by quartz contained in the sand fraction. The higher the proportion of quartz in the horizon, the lower the concentration of most elements except Si, which is the main element of quartz. This can be interpreted either as a sign of weathering or as a change in the sediment composition. Weathering would deplete the material in Fe, and enrich it in Zr and Ti, which we did not observe in the studied profiles. Instead, the Ti, Zr and Fe are strongly correlated. Thus the weathering hypothesis is unlikely. The second principal component of the PCA explained 16 % of the variance. The sand fraction was the only variable with high loading on this principal component (0.842). The two principal components identified in the PCA were interpreted as representing the evolution of the sediment composition in space and in time.

II. Evolution of parent sediment composition in space and time

We identified four different groups of observations in the space defined by the first two principal components (Figure A2.2.3b): (i) the first group formed by the beach sand alone, rich in Fe, Ti, Zr; (ii) a second group formed by the BC horizon of pit 1 and the Bw horizon

267

Part 6: Appendix A2. Co-authored papers of pit 2, relatively rich in Fe, Ti and Zr; (iii) a third group formed by the BC horizon of pit 2 and the BC and Bw horizons of pit 3, poor in Fe, Ti and Zr and (vi) the last group formed by the BC and Bw horizons of pits 4 and 5, and including the BC horizon of pit 3, all poor in Fe, Ti, Zr but rich in sand. On the basis of this distribution, we concluded that the parent sediment was homogeneous in the BC from pit 3 to pit 5 but different from the BC at pits 1 and 2, and from the beach sand.

The sediment from which the soils at pits 4 and 5 developed was vertically homogeneous. At pit 3, a change in the parent material composition was observed between BC and Bw horizons, the latter being closer in composition to the BC horizon at pit 2.

III. Factors influencing the evolution of parent sediment composition

It is generally accepted that the source of eroded material is the main factor controlling the sediment composition, although other parameters such as the transport history, the hydrological regime and the depositional environment can also impact sediment composition (Carter, 1974). The material deposited on the beach can derive from a single source or from multiple sources. Vermeire et al. (2016) examined the C and BC horizons of the Cox Bay Podzols and found identical REE-patterns and unchanging mineralogical compositions. They suggested that the sedimentary material came from the same source. They explained the observed variability of REE concentrations as resulting from changes in sedimentation dynamics yielding a variable content of clay, silt and sand (Vermeire et al. 2016). Such a depositional scheme has already been proposed for Barkley Sound located about 50 km SE of Cox Bay, on Vancouver Island (Carter, 1974).

All in all, we considered that the studied soil sequence fulfilled the conditions for being qualified as a chronosequence and, for the mass balance calculation, we used the BC horizon as reference material for pits 4 and 5, and the Bw horizon for pit 3.

268

Part 6: Appendix A2. Co-authored papers

A2.2.4.2 Dynamics of different processes involved in podsolization

During podzolization, a forest development induces a decrease in pH over time and a change in the organic matter entering the soils (Koerner et al., 1997, Temminghoff et al., 1998, Cornu et al., 2009). These processes - acidification, organic matter accumulation and Fe mobilization - contribute to the differentiation of typical Podzol horizons and their duration has been explored in a few studies. Although Podzols can form over a short period, a great variability in the timespan of podzolization has been revealed (Sauer et al., 2008).

I. Dynamics of acidification

During the incipient stage of soil development, under developing coniferous vegetation, the pH of the surface horizon dropped from 8 in the beach sand to 4.6 at pit 3, over less than 270 years (Figure A2.2.4a). The drop in pH was faster from 0 (beach sand) to 120 years (pit 1) than from 120 years to 270 years (pit 3). Such a fast acidification has already been observed for soil formation from a parental material with low buffer capacity (absence of carbonates) and under coniferous trees (Stützer, 1998, Leth and Breuning-Madsen, 1992, Crocker and Dickson, 1957). In Cox Bay, the parent sediment is devoid of carbonates and the development of coniferous trees (Sitka Spruce) provides a supply of acidifying organic matter (Lindeburg et al., 2013). Similar acidification phases have been observed in Podzol chronosequences published in the literature (Sauer et al., 2008) with variable duration most likely depending on the mineral composition of the parental material.

From 270 to 530 years, the pH of the surface horizon remained stable at around 4.7 ±0.2. This stabilization of the pH corresponds to the first Podzol profile in the chronosequence. Stabilization of pH during podzolization is recorded in the literature (Sauer et al., 2008, Starr and Lindroos, 2006, Righi et al., 1999, Birkeland, 1984, Bain et al., 1993, Bockheim, 1980) with the exception of Caner et al. (2010b). The average pH recorded at the start of podzolization was 4.2 ±0.4 (Caner et al., 2010b, Sauer et al., 2008, Starr and Lindroos, 2006, Righi et al.,

269

Part 6: Appendix A2. Co-authored papers

1999). This pH corresponds to the upper limit of the Al/Fe oxide buffer range according to Stützer (1998). The pH recorded for the start of podzolization in the Vancouver Island chronosequence falls within the average range obtained for Podzol chronosequences.

II. Dynamics of organic matter accumulation

Pedogenetic evolution consists in an accumulation of OC in the upper A/E horizon up to 175 years related to the progressive installation of vegetation followed by a decrease concomitant with the formation of a weak Bh horizon (Figure A2.2.1b). In terms of elementary OC flux per cm, no significant evolution was observed over time in the surface horizon, while this flux increased in the podzolic B horizon (Bh-Bhs-Bs) from 270 years (pit 3; Figure A2.2.4b). Such evolutions have already been observed elsewhere for Podzols: for the A horizon by Schaetzl et al. (1994) for the same duration and for A and B horizons by Sauer et al. (2008) on a longer time scale.

III. Dynamics of Fe mobilization

The major Fe transfers evidenced by decreasing FeDCB concentration in the E horizon and accumulation of FeDCB in the Bhs horizon started at 270 years (pit 3, Figure A2.2.4c). This evolution, synchronized with an increase in OC in the B horizon, marks the beginning of podzolization from 270 to 335 years as demonstrated by the strong correlation between elementary OC and FeDCB fluxes in the B horizons (R2=0.99, n=3, correlation significant at 1% confidence level). This correlation was already observed in other Podzols (Eusterhues et al., 2005). Once the required pH has been reached, podzolization starts within 50 years. Sauer et al. (2008) compiled data on incipient podzolization and showed that it is generally recorded within 200-500 years. However, this duration includes the initial acidification phase. More rapid cases of incipient podzolization were observed, e.g., 60 years (Cornu et al., 2008) and 144 years (Caner et al., 2010), due to faster acidification. In the case studied by Cornu et al. (2008), the chronosequence started from a mature Cambisol,

270

Part 6: Appendix A2. Co-authored papers instead of a parental material, explaining the rapid start of podzolization. The duration, i.e. the timespan between the last non- podzolic and the first podzolic profile, observed by Cornu et al., (2008) is comparable to that obtained in this study, or that observed by Caner et al. (2010).

The removal of FeDCB from the E horizon (Figure A2.2.4c), corresponds to the removal of crystalline Fe oxides but is accompanied by a large removal of silicate-bound Fe (Figure A2.2.5a), from hornblende, augite and chlorite that are weathered in the E horizon (Vermeire et al., 2016). Accumulation of FeDCB in the B podzolic horizons (Figure A2.2.4c) can be explained by a positive flux of OM- bound and colloidal Fe in the Bhs horizon at 335 (pit 4) and 530 years (pit 5) and a small positive flux in crystalline Fe oxides at 530 years (pit 5) in the same horizon (Figure A2.2.5b). Likewise, a loss of Fe is recorded at both 335 and 530 years. These results are in agreement with Caner et al. (2010) who observed changes in silicate mineralogy, in both E and B horizons, and an increase in both oxalate and DCB Fe concentrations in the B horizon on the same time step. Rapid silicate transformations in Podzols were also recorded elsewhere (see the compilation by Cornu et al. (2012); Cornelis et al. (2014b)).

271

Part 6: Appendix A2. Co-authored papers

Figure A2.2.4: Variation with time of (a) pH in the surface A/E horizons, (b) OC flux/cm and (c) DCB extracted Fe flux/cm in the surface A/E and subsurface B horizons of the Cox Bay soils. Grey fields indicate the time span (<50 years) of OC accumulation and Fe mobilization, triggered after the pH threshold was reached

272

Part 6: Appendix A2. Co-authored papers

Figure A2.2.5: Variations with time of the Fe isotopic composition and of the Fe flux/cm of the different individual Fe pools, i.e., OM bound and colloidal Fe, poorly crystalline Fe oxides, crystalline Fe oxides and silicate-bound Fe in (a) the surface A/E and (b) deep Bhs horizons.

273

Part 6: Appendix A2. Co-authored papers

A2.2.4.3 Fe isotopic fractionation as a function of the different podzolization mechanisms

Wiederhold et al. (2007b) interpreted the isotopic variation observed in two Podzol profiles sampled at two different locations in Germany as related to Fe mobilization, yet they offered no information on the timescale of the podzolization process. Our study on the Cox Bay chronosequence provides new input for the understanding of Fe isotope fractionation during podzolization.

No isotopic variations, outside analytical error, were observed during the first 270 years, indicating that acidification and OC accumulation had no impact on the isotopic composition of the Fe- bearing phases (Figure A2.2.5). From 270 to 335 years, significant changes marked Fe isotopic profiles and the observed isotopic patterns were similar to those described by Wiederhold et al. (2007b). The increase in δ56Fe values in the E horizon and a negative flux of silicate bound-Fe and crystalline Fe oxides (Figure A2.2.5a) are consistent with the concept of light Fe removal due to mineral dissolution and down-profile translocation. Furthermore, a decrease in δ56Fe values simultaneous to the positive flux of OM-bound and colloidal Fe pool in the Bhs horizon, from 270 to 335 years (Figure A2.2.5b), indicates preferential accumulation of light Fe in this horizon. These light δ56Fe values were associated with a maximum of OM- bound and colloidal Fe, and of poorly crystalline Fe oxides (Figure A2.2.2), suggesting that light Fe accumulation is predominantly due to precipitation of these Fe pools in the Bhs horizon (pit 4, 5). These data are in agreement with the interpretation of Wiederhold et al. (2007b) who suggested that the total Fe isotopic pattern in the Podzol profile is driven by the isotopic signature of the poorly crystalline Fe oxides. Nevertheless, the observed light Fe enrichment in the German Podzols was significantly larger than that found in our Podzols. This discrepancy could result from an age difference. In the absence of isotopic data on older Cox Bay profiles, however, no firm conclusion can be drawn.

274

Part 6: Appendix A2. Co-authored papers

A2.2.5 Conclusion

Previously published literature indicated that podzolization is a rapid process. We therefore chose a unique chronosequence composed of 5 soil profiles developed within a short duration, from 120 to 530 years, in order to spatially distinguish the mechanisms successively involved in Podzol formation. We showed that:

i. Acidification to a pH of around 4.5 is a prerequisite for the start of podzolization. This initial phase is due to the accumulation of organic matter in the surface horizon and precedes the mobilization of organic matter and Fe. In the Cox Bay chronosequence, this phase took 270 years, which is the duration classically observed in the literature. ii. Once the threshold pH value has been reached, OC and Fe mobilization and accumulation are triggered rapidly, in less than 50 years. Iron is mainly accumulated as organic complexes over the first 150 years and then as pedogenic Fe oxides. As shown in the literature, Fe accumulation can continue over several thousands of years. The absence of soil profiles older than 530 years in Cox Bay prevented further discussion on this point. iii. Iron isotopic fractionation in the upper A/E horizon was linked to mineral dissolution. Preferential release of light Fe isotopes from the silicate minerals resulted in enrichment of heavy Fe isotopes and positive δ56Fe values (0.25 ‰) in this horizon. We hypothesize that the intensity of this fractionation may be related to the nature of the weathered silicates. iv. Iron isotopic fractionation in the podzolic B horizons was clearly associated with the accumulation of Fe-organic complexes and poorly crystalline Fe oxyhydroxides. Light Fe enrichment was measured in the illuvial Bh-Bhs horizons, where the accumulation of secondary Fe phases was maximal. Whether this enrichment was due to (i) the preferential input of light Fe coming from the surface, (ii) a preferential incorporation of light Fe in the secondary

275

Part 6: Appendix A2. Co-authored papers

Fe phases or (iii) a preferential complexation of light Fe by the OM, still remains to be elucidated.

For future investigations, we suggest that Fe isotopic studies should be coupled with mineral separations and characterization. We believe that physical separation of different mineral pools could help unravel the Fe isotope signatures of individual solid phases. In addition, we suggest that to gain greater insight into the dynamics of the subsequent mechanisms, it would be necessary, in future studies, to use a chronosequence that spans over thousands of years, but with detailed sampling over the early stages of podzolization.

276

Part 6: Appendix A3. Carbon stable isotopes

A3. Carbon stable isotopes

δ13C (‰VPDB)

-28 -27 -26 -25 -24 -23

P1 P2 P3 P4 P1 P5 P2 BC1 Bw E E P3 P4 Bh Bh P5

Bhs

Bw Bw

BC2 BC BC BC

Figure A3.1 – Evolution of the δ13C in the horizons of the 5 profiles of increasing age.

277

Part 6: Appendix A4. SOMFrac ring trial

A4. SOMFrac ring trial

The aim of this SOMFrac Ring Trial is to evaluate existing and emerging SOC fractionation schemes in their ability to separate meaningful and as homogeneous as possible SOC fractions with respect to their turnover times. The fractionation scheme should be applicable to a broach range of soil samples without adjusting the method.

Selection criteria of the soil samples

 C3-C4 vegetation change to trace young/labile SOC. Sufficiently strong 13C label in C4 plot as compared to C3 plot (long-term). No young C3 plant-derived C input at C4 plot. Two subplots available: C4 and C3 (reference). Known land use history

 Sufficiently sample material available

 Broad range of texture classes/soil types among the selected sites

Site 1 – BS: Miscanthus long-term field site Braunschweig (permanent vegetation since 1993 at experiemental field site of the Julius Kühn Institute Braunschweig). Reference: Grassland

Site 2 – CL: Closeaux Field Experiment, INRA Experimental Station in Versailles, Île de France. Under C3-type crops since about 1930. A wheat (C3) monoculture was established in the late 1980s. In 1994 several plots were converted to monoculture maize (C4)

278

Part 6: Appendix A4. SOMFrac ring trial

Site 3 – RT: Long-term field trial Rotthalmünster. Long-term agricultural field trial (C3) converted into maize (C4) monoculture in 1979. Bavaria, Germany

Additional measurements/activities

 Basic soil parameters of the six samples (pH, CEC, SOC, Ntot, texture, 13C) (University Braunschweig, Thünen)

 Long-term incubation (6 months) of all six samples recording the 13 CO2 production (Thünen)

 14C content of selected fractions (passive C fractions)(Agroscope, Switzerland)

 Chemical characterisation of samples (NMR, DRIFT)?

 Modelling SOC dynamics using pools derived from different fractionation schemes (Delphine Derrien, Christopher Poeplau, Carlos Sierra)

279

Table A4.1 – SOMFrac participants: 21 participating labs, 21 SOM fractionation schemes tested.

No Name Country Method type Method 1 Trigalet S. Belgium Aggr+Density 2 Griepentrog M. Belgium Density free, macro-, µ-aggregate occluded POM and watersoluble 3 Kaiser M. Germany Aggr+Chem OM and two mineral associated OM fractions 4 Cotrufo F. USA Density or Size 5 Rovira P. Spain Chem+Size 6 Rovira P. Spain Chem sequetial oxid. with H2SO4 7 Grand S. USA Density density fract. 8 Nebbioso A. Italy Chem 9 Soong J. Belgium Density 10 Sleutel S. Belgium Chem oxid with NaOCl and HF 11 Wiesmeier M. Germany Aggr.+Density+Chem wet sieving, density+NaOCl oxid. 12 Gregorich E. Canada Density+Size 13 Yevdokimov I. Russland Density+Size 14 Six J. Switzerland Aggr+Density wet sieving+density 15 Vermeire M.L. Belgium Chem oxid with NaOCl and HF 16 Benbi D. India Density and Size 17 Chenu C. France ? 18 Macdonald L. Australia Density 19 Gunina A. Germany Aggr+Density wet sieving+ 1.85 SPT 20 Barbanzi L. Italy Aggr wet sieving 21 Yeasmin S. Australia Chem H2O2-oxidation

280

References

ADHIKARI, D., POULSON, S. R., SUMAILA, S., DYNES, J. J., MCBETH, J. M. & YANG, Y. 2016. Asynchronous reductive release of iron and organic carbon from hematite–humic acid complexes. Chemical Geology, 430, 13-20. AESCHBACHER, M., SANDER, M. & SCHWARZENBACH, R. P. 2009. Novel electrochemical approach to assess the redox properties of humic substances. Environmental science & technology, 44, 87-93. AIDE, M. & SMITH-AIDE, C. 2003. Assessing soil genesis by rare-earth elemental analysis. Soil Science Society of America Journal, 67, 1470-1476. AIDE, M. T. & AIDE, C. 2012. Rare Earth Elements: Their Importance in Understanding Soil Genesis. ISRN Soil Science, 2012, 1-11. ALMEIDA, V. R., SZPOGANICZA, B. & BONNEVILLE, S. 2015. Potentiometric Titration and Out-Of-Equilibrium pH Response of the Biotite‑Water System. Journal of the Brazilian Chemical Society, 26, 1848-1860. AMELUNG, W., BRODOWSKI, S., SANDHAGE-HOFMANN, A. & BOL, R. 2008. Chapter 6 Combining Biomarker with Stable Isotope Analyses for Assessing the Transformation and Turnover of Soil Organic Matter. In: DONALD, L. S. (ed.) Advances in Agronomy. Academic Press. AMELUNG, W., LOBE, I. & DU PREEZ, C. C. 2002. Fate of microbial residues in sandy soils of the South African Highveld as influenced by prolonged arable cropping. European Journal of Soil Science, 53, 29-35. AMELUNG, W., MILTNER, A., ZHANG, X. & ZECH, W. 2001. Fate of microbial residues during litter decomposition as affected by minerals. Soil Science, 166, 598-606. AMSTAETTER, K., BORCH, T. & KAPPLER, A. 2012. Influence of humic acid imposed changes of ferrihydrite aggregation on microbial Fe (III) reduction. Geochimica et Cosmochimica Acta, 85, 326- 341.

281

References

AMUNDSON, R. 2001. The carbon budget in soils. Annual Review of Earth and Planetary Sciences, 29, 535-62. ARAN, D., GURY, M. & JEANROY, E. 2001. Organo-metallic complexes in an Andosol: a comparative study with a Cambisol and Podzol. Geoderma, 99, 65-79. AUBERT, D., PROBST, A. & STILLE, P. 2004. Distribution and origin of major and trace elements (particularly REE, U and Th) into labile and residual phases in an acid soil profile (Vosges Mountains, France). Applied Geochemistry, 19, 899-916. AUBERT, D., STILLE, P. & PROBST, A. 2001. REE fractionation during granite weathering and removal by waters and suspended loads: Sr and Nd isotopic evidence. Geochimica et Cosmochimica Acta, 65, 387-406. AVENA, M. J. & KOOPAL, L. K. 1999. Kinetics of humic acid adsorption at solid-water interfaces. Environmental science & technology, 33, 2739-2744. BADRI, D. V. & VIVANCO, J. M. 2009. Regulation and function of root exudates. Plant, Cell & Environment, 32, 666-681. BAIN, D. C., MELLOR, A., ROBERTSON-RINTOUL, M. & BUCKLAND, S. 1993. Variations in weathering processes and rates with time in a chronosequence of soils from Glen Feshie, Scotland. Geoderma, 57, 275-293. BAIS, H. P., WEIR, T. L., PERRY, L. G., GILROY, S. & VIVANCO, J. M. 2006. The role of root exudates in rhizosphere interactions with plants and other organisms. Annu. Rev. Plant Biol., 57, 233-266. BAIZE, D. 1993. Soil Science Analyses. A guide to current use., Chichester. BALAKHNINA, T. I., BENNICELLI, R. P., STĘPNIEWSKA, Z., STĘPNIEWSKI, W. & FOMINA, I. R. 2010. Oxidative damage and antioxidant defense system in leaves of Vicia faba major L. cv. Bartom during soil flooding and subsequent drainage. Plant and soil, 327, 293-301. BALDOCK, J. A. & BROOS, K. 2012. Soil organic matter. In: HUANG, P. M., LI, Y. & SUMNER, M. E. (eds.) Handbook of Soil Sciences, second edn, Vol. 1: Properties and Processes. Boca Raton, FL: CRC Press/Taylor & Francis Group.

282

References

BALDOCK, J. A., MASIELLO, C. A., GÉLINAS, Y. & HEDGES, J. I. 2004. Cycling and composition of organic matter in terrestrial and marine ecosystems. Marine Chemistry, 92, 39-64. BALDOCK, J. A. & SKJEMSTAD, J. 2000. Role of the soil matrix and minerals in protecting natural organic materials against biological attack. Organic geochemistry, 31, 697-710. BANWART, S. 2011. Save our soils. Nature, 474, 151-152. BARDGETT, R. D. & VAN DER PUTTEN, W. H. 2014. Belowground biodiversity and ecosystem functioning. Nature, 515, 505-11. BARRÉ, P., PLANTE, A. F., CÉCILLON, L., LUTFALLA, S., BAUDIN, F., BERNARD, S., CHRISTENSEN, B. T., EGLIN, T., FERNANDEZ, J. M. & HOUOT, S. 2016. The energetic and chemical signatures of persistent soil organic matter. Biogeochemistry, 130, 1-12. BARRETT, L. R. & SCHAETZL, R. J. 1992. An examination of podzolization near Lake Michigan using chronofunctions. Canadian Journal of Soil Science, 72, 527-541. BASCOMB, C. L. 1968. DISTRIBUTION OF PYROPHOSPHATE- EXTRACTABLE IRON AND ORGANIC CARBON IN SOILS OF VARIOUS GROUPS. Journal of Soil Science, 19, 251-268. BEA, F. 1996. Residence of REE, Y, Th and U in granites and crustal protoliths; implications for the chemistry of crustal melts. Journal of Petrology, 37, 521-552. BEARD, B. L., JOHNSON, C. M., SKULAN, J. L., NEALSON, K. H., COX, L. & SUN, H. 2003. Application of Fe isotopes to tracing the geochemical and biological cycling of Fe. Chemical Geology, 195, 87-117. BEARE, M., COLEMAN, D., CROSSLEY JR, D., HENDRIX, P. & ODUM, E. 1995. A hierarchical approach to evaluating the significance of soil biodiversity to biogeochemical cycling. The significance and regulation of soil biodiversity. Springer. BERGER, A., JANOTS, E., GNOS, E., FREI, R. & BERNIER, F. 2014. Rare earth element mineralogy and geochemistry in a laterite profile from Madagascar. Applied Geochemistry, 41, 218-228. BIRKELAND, P. W. 1984. Holocene soil chronofunctions, Southern Alps, New Zealand. Geoderma, 34, 115-134. BLAKEMORE, L. C., P.L., S. & DALY, B. K. 1987. Methods for chemical analysis of soils, Lower Hutt, N.Z.

283

References

BLUM, W. E. 2005. Functions of soil for society and the environment. Reviews in Environmental Science and Bio/Technology, 4, 75- 79. BOCKHEIM, J. 1980. Solution and use of chronofunctions in studying soil development. Geoderma, 24, 71-85. BODÉ, S., DENEF, K. & BOECKX, P. 2009. Development and evaluation of a high-performance liquid chromatography/isotope ratio mass spectrometry methodology for delta13C analyses of amino sugars in soil. Rapid Commun Mass Spectrom, 23, 2519- 26. BODÉ, S., FANCY, R. & BOECKX, P. 2013. Stable isotope probing of amino sugars--a promising tool to assess microbial interactions in soils. Rapid Commun Mass Spectrom, 27, 1367-79. BONNEVILLE, S. 2005. Kinetics of microbial Fe (III) oxyhydroxidereduction: The role of mineral properties. Ph.D. Dissertation. Universiteit Utrecht. BONNEVILLE, S., BEHRENDS, T. & VAN CAPPELLEN, P. 2009. Solubility and dissimilatory reduction kinetics of iron (III) oxyhydroxides: a linear free energy relationship. Geochimica et cosmochimica acta, 73, 5273-5282. BONNEVILLE, S., BEHRENDS, T., VAN CAPPELLEN, P., HYACINTHE, C. & RÖLING, W. F. 2006. Reduction of Fe (III) colloids by Shewanella putrefaciens: a kinetic model. Geochimica et cosmochimica acta, 70, 5842-5854. BONNEVILLE, S., MORGAN, D. J., SCHMALENBERGER, A., BRAY, A., BROWN, A., BANWART, S. A. & BENNING, L. G. 2011. Tree- mycorrhiza symbiosis accelerate mineral weathering: Evidences from nanometer-scale elemental fluxes at the hypha–mineral interface. Geochimica et Cosmochimica Acta, 75, 6988-7005. BONNEVILLE, S., VANCAPPELLEN, P. & BEHRENDS, T. 2004. Microbial reduction of iron(III) oxyhydroxides: effects of mineral solubility and availability. Chemical Geology, 212, 255-268. BOSATTA, E. & ÅGREN, G. I. 1999. Soil organic matter quality interpreted thermodynamically. Soil Biology and Biochemistry, 31, 1889-1891. BOUDOT, J.-P. 1992. Relative efficiency of complexed aluminum noncrystalline Al hydroxide, allophane and imogolite in

284

References

retarding the biodegradation of citric acid. Geoderma, 52, 29- 39. BOUDOT, J., BEL HADJ BRAHIM, A., STEIMAN, R. & SEIGLE-MURANDI, F. 1989. Biodegradation of synthetic organo-metallic complexes of iron and aluminium with selected metal to carbon ratios. Soil Biology and Biochemistry, 21, 961-966. BOWEN, N. 1922. The reaction principle in petrogenesis. The Journal of Geology, 30, 177-198. BRADY, N. C. & WEIL, R. R. 2007. The Nature and Properties of Soils, 14th Edition, Pearson. BRANTLEY, S. L., MEGONIGAL, J. P., SCATENA, F. N., BALOGH- BRUNSTAD, Z., BARNES, R. T., BRUNS, M. A., VAN CAPPELLEN, P., DONTSOVA, K., HARTNETT, H. E., HARTSHORN, A. S., HEIMSATH, A., HERNDON, E., JIN, L., KELLER, C. K., LEAKE, J. R., MCDOWELL, W. H., MEINZER, F. C., MOZDZER, T. J., PETSCH, S., PETT-RIDGE, J., PREGITZER, K. S., RAYMOND, P. A., RIEBE, C. S., SHUMAKER, K., SUTTON-GRIER, A., WALTER, R. & YOO, K. 2011. Twelve testable hypotheses on the geobiology of weathering. Geobiology, 9, 140-65. BRAUN, J.-J., PAGEL, M., MULLER, J.-P., BILONG, P., MICHARD, A. & GUILLET, B. 1990. Cerium anomalies in lateritic profiles. Geochimica et Cosmochimica Acta, 54, 781-795. BRAUN, J. J., PAGEL, M., HERBILLN, A. & ROSIN, C. 1993. Mobilization and redistribution of REEs and thorium in a syenitic lateritic profile: A mass balance study. Geochimica et Cosmochimica Acta, 57, 4419-4434. BRAUN, J. J., VIERS, J., DUPRÉ, B., POLVE, M., NDAM, J. & MULLER, J. P. 1998. Solid/liquid REE fractionation in the lateritic system of Goyoum, East Cameroon: The implication for the present dynamics of the soil covers of the humid tropical regions. Geochimica et Cosmochimica Acta, 62, 273-299. BRIMHALL, G. H., FORD, C., BRATT, J., TAYLOR, G. & WARIN, O. 1991. Quantitative geochemical approach to pedogenesis: importance of parent material reduction, volumetric expansion, and eolian influx in lateritization. Geoderma, 51, 51- 91. BROECKLING, C. D., BROZ, A. K., BERGELSON, J., MANTER, D. K. & VIVANCO, J. M. 2008. Root exudates regulate soil fungal

285

References

community composition and diversity. Applied and environmental microbiology, 74, 738-744. BUETTNER, S. W., KRAMER, M. G., CHADWICK, O. A. & THOMPSON, A. 2014. Mobilization of colloidal carbon during iron reduction in basaltic soils. Geoderma, 221-222, 139-145. BUNN, R. A., MAGELKY, R. D., RYAN, J. N. & ELIMELECH, M. 2002. Mobilization of natural colloids from an iron oxide-coated sand aquifer: effect of pH and ionic strength. Environmental Science & Technology, 36, 314-322. BURDON, J. 2001. Are the traditional concepts of the structures of humic substances realistic? Soil Science, 166, 752-769. BUURMAN, P. & JONGMANS, A. G. 2005. Podzolisation and soil organic matter dynamics. Geoderma, 125, 71-83. BUURMAN, P., VAN BERGEN, P., JONGMANS, A., MEIJER, E., DURAN, B. & VAN LAGEN, B. 2005. Spatial and temporal variation in podzol organic matter studied by pyrolysis‐gas chromatography/mass spectrometry and micromorphology. European Journal of Soil Science, 56, 253-270. CALVARUSO, C., MARESCHAL, L., TURPAULT, M.-P. & LECLERC, E. 2009. Rapid clay weathering in the rhizosphere of Norway spruce and oak in an acid forest ecosystem. Soil Science Society of America Journal, 73, 331-338. CANER, L., JOUSSEIN, E., SALVADOR-BLANES, S., HUBERT, F., SCHLICHT, J.-F. & DUIGOU, N. 2010a. Short-time clay-mineral evolution in a soil chronosequence in Oléron Island (France). Journal of Plant Nutrition and Soil Science, 173, 591-600. CANER, L., JOUSSEIN, E., SALVADOR‐BLANES, S., HUBERT, F., SCHLICHT, J. F. & DUIGOU, N. 2010b. Short‐time clay‐mineral evolution in a soil chronosequence in Oléron Island (France). Journal of Plant Nutrition and Soil Science, 173, 591-600. CAO, X., CHEN, Y., WANG, X. & DENG, X. 2001. Effects of redox potential and pH value on the release of rare earth elements from soil. Chemosphere, 44, 655-661. CARDINAL, D., GAILLARDET, J., HUGHES, H., OPFERGELT, S. & ANDRE, L. 2010. Contrasting silicon isotope signatures in rivers from the Congo Basin and the specific behaviour of organic‐rich waters. Geophysical Research Letters, 37.

286

References

CARTER, L. 1974. An evaluation of the provenance of terrigenous sediments from offshore Vancouver Island. Canadian Journal of Earth Sciences, 11, 664-677. CHABAUX, F., DEQUINCEY, O., LÉVÈQUE, J.-J., LEPRUN, J.-C., CLAUER, N., RIOTTE, J. & PAQUET, H. 2003. Tracing and dating recent chemical transfers in weathering profiles by trace-element geochemistry and 238 U 234 U 230 Th disequilibria: the example of the Kaya lateritic toposequence (Burkina-Faso). Comptes Rendus Geoscience, 335, 1219-1231. CHABBI, A., KÖGEL-KNABNER, I. & RUMPEL, C. 2009. Stabilised carbon in subsoil horizons is located in spatially distinct parts of the soil profile. Soil Biology and Biochemistry, 41, 256-261. CHADWICK, O. & CHOROVER, J. 2001. The chemistry of pedogenic thresholds. Geoderma. CHAO, T. T. 1976. The significance of secondary iron and manganese oxides in geochemical exploration. Economic Geology, 71, 1560-1569. CHAO, T. T. & SANZOLONE, R. F. 1992. Decomposition techniques. Journal of Geochemical Exploration, 44, 65-106. CHENU, C. & STOTZKY, G. 2002. Interactions between microorganisms and soil particles: an overview. In: HUANG, P., BOLLAG, J. & SENESI, N. (eds.) Interactions between soil particles and microorganisms.: John Wiley & Sons, Ltd., Manchester, UK. CHRIST, M. J. & DAVID, M. B. 1996. Temperature and moisture effects on the production of dissolved organic carbon in a spodosol. Soil Biology and Biochemistry, 28, 1191-1199. CHURCHMAN, G. J. & LOWE, D. J. 2012. Alteration, Formation, and Occurrence of Minerals in Soils. In: HUANG, P. M., LI, Y. & SUMNER, M. E. (eds.) Handbook of Soil Sciences, second edn, Vol. 1: Properties and Processes. Boca Raton, FL: CRC Press/Taylor & Francis Group. CIDU, R., VITTORI ANTISARI, L., BIDDAU, R., BUSCAROLI, A., CARBONE, S., DA PELO, S., DINELLI, E., VIANELLO, G. & ZANNONI, D. 2013. Dynamics of rare earth elements in water–soil systems: The case study of the Pineta San Vitale (Ravenna, Italy). Geoderma, 193-194, 52-67. CISMASU, A. C., MICHEL, F. M., TCACIUC, A. P., TYLISZCZAK, T. & BROWN JR, G. E. 2011. Composition and structural aspects of

287

References

naturally occurring ferrihydrite. Comptes Rendus Geoscience, 343, 210-218. CLAESSENS, J., VAN LITH, Y., LAVERMAN, A. M. & VAN CAPPELLEN, P. 2006. Acid–base activity of live bacteria: implications for quantifying cell wall charge. Geochimica et cosmochimica acta, 70, 267-276. COLLIGNON, C., RANGER, J. & TURPAULT, M. P. 2012. Seasonal dynamics of Al- and Fe-bearing secondary minerals in an acid forest soil: influence of Norway spruce roots (Picea abies(L.) Karst.). European Journal of Soil Science, 63, 592-602. COMPTON, J. S., WHITE, R. A. & SMITH, M. 2003. Rare earth element behavior in soils and salt pan sediments of a semi-arid granitic terrain in the Western Cape, South Africa. Chemical Geology, 201, 239-255. CONANT, R. T., RYAN, M. G., ÅGREN, G. I., BIRGE, H. E., DAVIDSON, E. A., ELIASSON, P. E., EVANS, S. E., FREY, S. D., GIARDINA, C. P. & HOPKINS, F. M. 2011. Temperature and soil organic matter decomposition rates–synthesis of current knowledge and a way forward. Global Change Biology, 17, 3392-3404. CONDIE, K. C., DENGATE, J. & CULLERS, R. L. 1995. Behavior of rare earth elements in a paleoweathering profile on granodiorite in the Front Range, Colorado, USA. Geochimica et Cosmochimica Acta, 59, 279-294. CORDES, D. L. 1972. An ecological study of the Sitka spruce forest on the west coast of Vancouver Island., University of British Columbia. CORNELIS, J.-T., DELVAUX, B., CARDINAL, D., ANDRÉ, L., RANGER, J. & OPFERGELT, S. 2010. Tracing mechanisms controlling the release of dissolved silicon in forest soil solutions using Si isotopes and Ge/Si ratios. Geochimica et Cosmochimica Acta, 74, 3913-3924. CORNELIS, J.-T., DELVAUX, B., GEORG, R., LUCAS, Y., RANGER, J. & OPFERGELT, S. 2011. Tracing the origin of dissolved silicon transferred from various soil-plant systems towards rivers: a review. Biogeosciences, 8, 89-112. CORNELIS, J.-T., WEIS, D., LAVKULICH, L., VERMEIRE, M.-L., DELVAUX, B. & BARLING, J. 2014a. Silicon isotopes record dissolution and

288

References

re-precipitation of pedogenic clay minerals in a podzolic soil chronosequence. Geoderma, 235-236, 19-29. CORNELIS, J. T., DUMON, M., TOLOSSA, A. R., DELVAUX, B., DECKERS, J. & VAN RANST, E. 2014b. The effect of pedological conditions on the sources and sinks of silicon in the Vertic Planosols in south-western Ethiopia. Catena, 112, 131-138. CORNELL, R. M. & SCHWERTMANN, U. 2003. The iron oxides: structure, properties, reactions, occurrences and uses, John Wiley & Sons. CORNU, S., BESNAULT, A. & BERMOND, A. 2008. Soil podzolization induced by reforestation as shown by sequential and kinetic extractions of Fe and Al. European Journal of Soil Science, 59, 222-232. CORNU, S., LUCAS, Y., LEBON, E., AMBROSI, J. P., LUIZÃO, F., ROUILLER, J., BONNAY, M. & NEAL, C. 1999. Evidence of titanium mobility in soil profiles, Manaus, central Amazonia. Geoderma, 91, 281- 295. CORNU, S., MONTAGNE, D., HUBERT, F., BARRÉ, P. & CANER, L. 2012. Evidence of short-term clay evolution in soils under human impact. Comptes Rendus Geoscience, 344. CORNU, S., MONTAGNE, D. & VASCONCELOS, P. M. 2009. Dating constituent formation in soils to determine rates of soil processes: A review. Geoderma, 153, 293-303. COTRUFO, M. F., WALLENSTEIN, M. D., BOOT, C. M., DENEF, K. & PAUL, E. 2013. The Microbial Efficiency‐Matrix Stabilization (MEMS) framework integrates plant litter decomposition with soil organic matter stabilization: do labile plant inputs form stable soil organic matter? Global Change Biology, 19, 988-995. CROCKER, R. L. & DICKSON, B. 1957. Soil development on the recessional moraines of the Herbert and Mendenhall Glaciers, south-eastern Alaska. The Journal of Ecology, 169-185. DAVIDSON, E. A. & JANSSENS, I. A. 2006. Temperature sensitivity of soil carbon decomposition and feedbacks to climate change. Nature, 440, 165-173. DAVRANCHE, M., GRYBOS, M., GRUAU, G., PÉDROT, M., DIA, A. & MARSAC, R. 2011. Rare earth element patterns: A tool for identifying trace metal sources during wetland soil reduction. Chemical Geology, 284, 127-137.

289

References

DAVYDCHUK, V. S., ZARUDNAYA, R. F. & MIHELI, S. V. 1990. Landscapes of Chernobyl zone and their evaluation by condition of radionuclide migration. , Kiev, Naukova dumka. DE LEEUW, J. & LARGEAU, C. 1993. A review of macromolecular organic compounds that comprise living organisms and their role in kerogen, coal, and petroleum formation. Organic Geochemistry. Springer. DE WIT, R. & BOUVIER, T. 2006. ‘Everything is everywhere, but, the environment selects’; what did Baas Becking and Beijerinck really say? Environmental microbiology, 8, 755-758. DELSTANCHE, S., OPFERGELT, S., CARDINAL, D., ELSASS, F., ANDRÉ, L. & DELVAUX, B. 2009. Silicon isotopic fractionation during adsorption of aqueous monosilicic acid onto iron oxide. Geochimica et Cosmochimica Acta, 73, 923-934. DEMAREST, M. S., BRZEZINSKI, M. A. & BEUCHER, C. P. 2009. Fractionation of silicon isotopes during biogenic silica dissolution. Geochimica et Cosmochimica Acta, 73, 5572-5583. DENEF, K., BUBENHEIM, H., LENHART, K., VERMEULEN, J., VAN CLEEMPUT, O., BOECKX, P. & MÜLLER, C. 2007. Community shifts and carbon translocation within metabolically-active rhizosphere microorganisms in grasslands under elevated CO 2. Biogeosciences, 4, 769-779. DENEF, K., ROOBROECK, D., MANIMEL WADU, M. C. W., LOOTENS, P. & BOECKX, P. 2009. Microbial community composition and rhizodeposit-carbon assimilation in differently managed temperate grassland soils. Soil Biology and Biochemistry, 41, 144-153. DEQUINCEY, O., CHABAUX, F., LEPRUN, J. C., PAQUET, H., CLAUER, N. & LARQUE, P. 2006. Lanthanide and trace element mobilization in a lateritic toposequence: inferences from the Kaya laterite in Burkina Faso. European Journal of Soil Science, 57, 816-830. DERENNE, S. & LARGEAU, C. 2001. A review of some important families of refractory macromolecules: composition, origin, and fate in soils and sediments. Soil Science, 166, 833-847. DERRIEN, D., MAROL, C., BALABANE, M. & BALESDENT, J. 2006. The turnover of carbohydrate carbon in a cultivated soil estimated by 13C natural abundances. European Journal of Soil Science, 57, 547-557.

290

References

DERRY, L. A., KURTZ, A. C., ZIEGLER, K. & CHADWICK, O. A. 2005. Biological control of terrestrial silica cycling and export fluxes to watersheds. Nature, 433, 728-731. DINEL, H., SCHNITZER, M. & MEHUYS, G. 1990. Soil lipids: origin, nature, content, decomposition, and effect on soil physical properties. Soil biochemistry, 6, 397-429. DO NASCIMENTO, N., BUENO, G., FRITSCH, E., HERBILLON, A., ALLARD, T., MELFI, A., ASTOLFO, R., BOUCHER, H. & LI, Y. 2004. Podzolization as a deferralitization process: a study of an Acrisol–Podzol sequence derived from Palaeozoic sandstones in the northern upper Amazon Basin. European journal of soil science, 55, 523-538. DONG, H., JAISI, D. P., KIM, J. & ZHANG, G. 2009. Review Paper. Microbe-clay mineral interactions. American Mineralogist, 94, 1505-1519. DUCHAUFOUR, P. 1997. Abrégé de pédologie: sol, végétation, environnement, Paris, Masson. DÜMIG, A., HÄUSLER, W., STEFFENS, M. & KÖGEL-KNABNER, I. 2012. Clay fractions from a soil chronosequence after glacier retreat reveal the initial evolution of organo–mineral associations. Geochimica et Cosmochimica Acta, 85, 1-18. DUMON, J. & VIGNEAUX, M. 1979. Evidence for some mobility of titanium in podzols and under laboratory conditions as a result of the action of organic agents. Physics and Chemistry of the Earth, 11, 331-337. DUNGAIT, J. A. J., HOPKINS, D. W., GREGORY, A. S. & WHITMORE, A. P. 2012. Soil organic matter turnover is governed by accessibility not recalcitrance. Global Change Biology, 18, 1781-1796. EGLI, M., ZANELLI, R., KAHR, G., MIRABELLA, A. & FITZE, P. 2002. Soil evolution and development of the clay mineral assemblages of a Podzol and a Cambisol in ‘Meggerwald’, Switzerland. Clay Minerals, 37, 351-366. EKSCHMITT, K., KANDELER, E., POLL, C., BRUNE, A., BUSCOT, F., FRIEDRICH, M., GLEIXNER, G., HARTMANN, A., KÄSTNER, M. & MARHAN, S. 2008. Soil‐carbon preservation through habitat constraints and biological limitations on decomposer activity. Journal of Plant Nutrition and Soil Science, 171, 27-35.

291

References

EKSCHMITT, K., LIU, M., VETTER, S., FOX, O. & WOLTERS, V. 2005. Strategies used by soil biota to overcome soil organic matter stability—why is dead organic matter left over in the soil? Geoderma, 128, 167-176. ELLERBROCK, R. & KAISER, M. 2005. Stability and composition of different soluble soil organic matter fractions–evidence from δ 13 C and FTIR signatures. Geoderma, 128, 28-37. EMMANUEL, S., EREL, Y., MATTHEWS, A. & TEUTSCH, N. 2005. A preliminary mixing model for Fe isotopes in soils. Chemical Geology, 222, 23-34. ETTEMA, C. H. & WARDLE, D. A. 2002. Spatial soil ecology. Trends in ecology & evolution, 17, 177-183. EUSTERHUES, K., HÄDRICH, A., NEIDHARDT, J., KÜSEL, K., KELLER, T. F., JANDT, K. D. & TOTSCHE, K. U. 2014b. Reduction of ferrihydrite with adsorbed and coprecipitated organic matter: microbial reduction by Geobacter bremensis vs. abiotic reduction by Na- dithionite. Biogeosciences, 11, 4953-4966. EUSTERHUES, K., NEIDHARDT, J., HÄDRICH, A., KÜSEL, K. & TOTSCHE, K. U. 2014a. Biodegradation of ferrihydrite-associated organic matter. Biogeochemistry, 119, 45-50. EUSTERHUES, K., RENNERT, T., KNICKER, H., KÖGEL-KNABNER, I., TOTSCHE, K. U. & SCHWERTMANN, U. 2010. Fractionation of organic matter due to reaction with ferrihydrite: coprecipitation versus adsorption. Environmental science & technology, 45, 527-533. EUSTERHUES, K., RUMPEL, C., KLEBER, M. & KÖGEL-KNABNER, I. 2003. Stabilisation of soil organic matter by interactions with minerals as revealed by mineral dissolution and oxidative degradation. Organic Geochemistry, 34, 1591-1600. EUSTERHUES, K., RUMPEL, C. & KOGEL-KNABNER, I. 2005. Organo- mineral associations in sandy acid forest soils: importance of specific surface area, iron oxides and micropores. European Journal of Soil Science, 56, 753–763 EUSTERHUES, K., RUMPEL, C. & KÖGEL-KNABNER, I. 2007. Composition and radiocarbon age of HF-resistant soil organic matter in a Podzol and a Cambisol. Organic Geochemistry, 38, 1356-1372. EUSTERHUES, K., WAGNER, F. E., HÄUSLER, W., HANZLIK, M., KNICKER, H., TOTSCHE, K. U., KÖGEL-KNABNER, I. & SCHWERTMANN, U.

292

References

2008. Characterization of ferrihydrite-soil organic matter coprecipitates by X-ray diffraction and Mossbauer spectroscopy. Environmental science & technology, 42, 7891- 7897. FANTLE, M. S. & DEPAOLO, D. J. 2004. Iron isotopic fractionation during continental weathering. Earth and Planetary Science Letters, 228, 547-562. FAO 2015. Status of the World’s Soil Resources (SWSR) – Main Report. Food and Agriculture Organization of the United Nations and Intergovernmental Technical Panel on Soils. Rome, Italy. FAO 2016. Global Soil Biodiversity Atlas, Luxembourg, Publications Office of the European Union. FARMER, V., RUSSELL, J. & BERROW, M. 1980. Imogolite and proto‐ imogolite allophane in spodic horizons: evidence for a mobile aluminium silicate complex in podzol formation. Journal of Soil Science, 31, 673-684. FAVILLI, F., EGLI, M., CHERUBINI, P., SARTORI, G., HAEBERLI, W. & DELBOS, E. 2008. Comparison of different methods of obtaining a resilient organic matter fraction in Alpine soils. Geoderma, 145, 355-369. FEKIACOVA, Z., PICHAT, S., CORNU, S. & BALESDENT, J. 2013. Inferences from the vertical distribution of Fe isotopic compositions on pedogenetic processes in soils. Geoderma, 209-210, 110-118. FENG, W., PLANTE, A. F., AUFDENKAMPE, A. K. & SIX, J. 2014. Soil organic matter stability in organo-mineral complexes as a function of increasing C loading. Soil Biology and Biochemistry, 69, 398-405. FIEDLER, S. & KALBITZ, K. 2003. Concentrations and properties of dissolved organic matter in forest soils as affected by the redox regime. Soil Science, 168, 793-801. FINLAY, R., WALLANDER, H., SMITS, M., HOLMSTROM, S., VAN HEES, P., LIAN, B. & ROSLING, A. 2009. The role of fungi in biogenic weathering in boreal forest soils. Fungal Biology Reviews, 23, 101-106. FLEMMING, H.-C. & WINGENDER, J. 2010. The biofilm matrix. Nature Reviews Microbiology, 8, 623-633.

293

References

FONTAINE, S., BAROT, S., BARRE, P., BDIOUI, N., MARY, B. & RUMPEL, C. 2007. Stability of organic carbon in deep soil layers controlled by fresh carbon supply. Nature, 450, 277-80. FONTAINE, S., MARIOTTI, A. & ABBADIE, L. 2003. The priming effect of organic matter: a question of microbial competition? Soil Biology and Biochemistry, 35, 837-843. FREDRICKSON, J. K., ZACHARA, J. M., KENNEDY, D. W., DONG, H., ONSTOTT, T. C., HINMAN, N. W. & LI, S.-M. 1998. Biogenic iron mineralization accompanying the dissimilatory reduction of hydrous ferric oxide by a groundwater bacterium. Geochimica et Cosmochimica Acta, 62, 3239-3257. FRIEDLINGSTEIN, P., MEINSHAUSEN, M., ARORA, V. K., JONES, C. D., ANAV, A., LIDDICOAT, S. K. & KNUTTI, R. 2014. Uncertainties in CMIP5 climate projections due to carbon cycle feedbacks. Journal of Climate, 27, 511-526. FRIETSCH, R. & PERDAHL, J.-A. 1995. Rare earth elements in apatite and magnetite in Kiruna-type iron ores and some other iron ore types. Ore Geology Reviews, 9, 489-510. FRITSCH, E., BALAN, E., DO NASCIMENTO, N. R., ALLARD, T., BARDY, M., BUENO, G., DERENNE, S., MELFI, A. J. & CALAS, G. 2011. Deciphering the weathering processes using environmental mineralogy and geochemistry: Towards an integrated model of laterite and podzol genesis in the Upper Amazon Basin. Comptes Rendus Geoscience, 343, 188-198. FRITZSCHE, A., BOSCH, J., RENNERT, T., HEISTER, K., BRAUNSCHWEIG, J., MECKENSTOCK, R. U. & TOTSCHE, K. U. 2012. Fast microbial reduction of ferrihydrite colloids from a soil effluent. Geochimica et Cosmochimica Acta, 77, 444-456. FRITZSCHE, A., SCHRÖDER, C., WIECZOREK, A. K., HÄNDEL, M., RITSCHEL, T. & TOTSCHE, K. U. 2015. Structure and composition of Fe–OM co-precipitates that form in soil-derived solutions. Geochimica et Cosmochimica Acta, 169, 167-183. FROELICH, P. N. & ANDREAE, M. O. 1981. The marine geochemistry of germanium: Ekasilicon. Science, 213, 205-207. FROSTEGÅRD, Å. & BÅÅTH, E. 1996. The use of phospholipid fatty acid analysis to estimate bacterial and fungal biomass in soil. Biology and Fertility of Soils, 22, 59-65.

294

References

GADD, G. M. 2007. Geomycology: biogeochemical transformations of rocks, minerals, metals and radionuclides by fungi, bioweathering and bioremediation. Mycological research, 111, 3-49. GALAN, E., FERNANDEZ-CALIANI, J., MIRAS, A., APARICIO, P. & MARQUEZ, M. 2007. Residence and fractionation of rare earth elements during kaolinization of alkaline peraluminous granites in NW Spain. Clay Minerals, 42, 341-352. GANGLOFF, S., STILLE, P., PIERRET, M.-C., WEBER, T. & CHABAUX, F. 2014. Characterization and evolution of dissolved organic matter in acidic forest soil and its impact on the mobility of major and trace elements (case of the Strengbach watershed). Geochimica et Cosmochimica Acta, 130, 21-41. GEORG, R., REYNOLDS, B., FRANK, M. & HALLIDAY, A. 2006. New sample preparation techniques for the determination of Si isotopic compositions using MC-ICPMS. Chemical Geology, 235, 95-104. GEORG, R., REYNOLDS, B., WEST, A., BURTON, K. & HALLIDAY, A. 2007. Silicon isotope variations accompanying basalt weathering in Iceland. Earth and Planetary Science Letters, 261, 476-490. GEORG, R., ZHU, C., REYNOLDS, B. & HALLIDAY, A. 2009. Stable silicon isotopes of groundwater, feldspars, and clay coatings in the Navajo Sandstone aquifer, Black Mesa, Arizona, USA. Geochimica et Cosmochimica Acta, 73, 2229-2241. GLASAUER, S., WEIDLER, P. G., LANGLEY, S. & BEVERIDGE, T. J. 2003. Controls on Fe reduction and mineral formation by a subsurface bacterium. Geochimica et Cosmochimica Acta, 67, 1277-1288. GLASER, B., TURRIÓN, M. A.-B. & ALEF, K. 2004. Amino sugars and muramic acid—biomarkers for soil microbial community structure analysis. Soil Biology and Biochemistry, 36, 399-407. GLEIXNER, G., POIRIER, N., BOL, R. & BALESDENT, J. 2002. Molecular dynamics of organic matter in a cultivated soil. Organic Geochemistry, 33, 357-366. GODFRAY, H. C. J., BEDDINGTON, J. R., CRUTE, I. R., HADDAD, L., LAWRENCE, D., MUIR, J. F., PRETTY, J., ROBINSON, S., THOMAS, S. M. & TOULMIN, C. 2010. Food security: the challenge of feeding 9 billion people. science, 327, 812-818.

295

References

GOLDICH, S. S. 1938. A study in rock-weathering. The Journal of Geology, 17-58. GONG, Q., DENG, J., YANG, L., ZHANG, J., WANG, Q. & ZHANG, G. 2011. Behavior of major and trace elements during weathering of sericite–quartz schist. Journal of Asian Earth Sciences, 42, 1-13. GOYNE, K. W., BRANTLEY, S. L. & CHOROVER, J. 2010. Rare earth element release from phosphate minerals in the presence of organic acids. Chemical Geology, 278, 1-14. GRAMSS, G., VOIGT, K.-D. & KIRSCHE, B. 1999. Degradation of polycyclic aromatic hydrocarbons with three to seven aromatic rings by higher fungi in sterile and unsterile soils. Biodegradation, 10, 51-62. GRAND, S. & LAVKULICH, L. M. 2013. Potential influence of poorly crystalline minerals on soil chemistry in Podzols of southwestern Canada. European Journal of Soil Science, 64, 651-660. GRANDY, A. & NEFF, J. 2008. Molecular C dynamics downstream: the biochemical decomposition sequence and its impact on soil organic matter structure and function. The Science of the total environment, 404, 297-307. GRYBOS, M., DAVRANCHE, M., GRUAU, G. & PETITJEAN, P. 2007. Is trace metal release in wetland soils controlled by organic matter mobility or Fe-oxyhydroxides reduction? Journal of Colloid and Interface Science, 314, 490-501. GRYBOS, M., DAVRANCHE, M., GRUAU, G., PETITJEAN, P. & PÉDROT, M. 2009. Increasing pH drives organic matter solubilization from wetland soils under reducing conditions. Geoderma, 154, 13-19. GU, B., MEHLHORN, T. L., LIANG, L. & MCCARTHY, J. F. 1996a. Competitive adsorption, displacement, and transport of organic matter on iron oxide: I. Competitive adsorption. Geochimica et Cosmochimica Acta, 60, 1943-1950. GU, B., MEHLHORN, T. L., LIANG, L. & MCCARTHY, J. F. 1996b. Competitive adsorption, displacement, and transport of organic matter on iron oxide: II. Displacement and transport. Geochimica et Cosmochimica Acta, 60, 2977-2992.

296

References

GUGGENBERGER, G. & KAISER, K. 2003. Dissolved organic matter in soil: challenging the paradigm of sorptive preservation. Geoderma, 113, 293-310. GUNNARS, A., BLOMQVIST, S., JOHANSSON, P. & ANDERSSON, C. 2002. Formation of Fe (III) oxyhydroxide colloids in freshwater and brackish seawater, with incorporation of phosphate and calcium. Geochimica et Cosmochimica Acta, 66, 745-758. GUSTAFSSON, J., BHATTACHARYA, P., BAIN, D., FRASER, A. & MCHARDY, W. 1995. Podzolisation mechanisms and the synthesis of imogolite in northern Scandinavia. Geoderma, 66, 167-184. GUSTAFSSON, J. P. 2001. The surface chemistry of imogolite. clays and clay Minerals, 49, 73-80. GUSTAFSSON, J. P., BHATTACHARYA, P. & KARLTUN, E. 1999. Mineralogy of poorly crystalline aluminium phases in the B horizon of Podzols in southern Sweden. Applied Geochemistry, 14, 707-718. HAAS, J. R. & DICHRISTINA, T. J. 2002. Effects of Fe (III) chemical speciation on dissimilatory Fe (III) reduction by Shewanella putrefaciens. Environmental science & technology, 36, 373- 380. HAGEDORN, F., KAISER, K., FEYEN, H. & SCHLEPPI, P. 2000. Effects of redox conditions and flow processes on the mobility of dissolved organic carbon and nitrogen in a forest soil. Journal of Environmental Quality, 29, 288-297. HALL, G., VAIVE, J., BEER, R. & HOASHI, M. 1996. Selective leaches revisited, with emphasis on the amorphous Fe oxyhydroxide phase extraction. Journal of Geochemical Exploration, 56, 59- 78. HAMER, U., MARSCHNER, B., BRODOWSKI, S. & AMELUNG, W. 2004. Interactive priming of black carbon and glucose mineralisation. Organic Geochemistry, 35, 823-830. HANSEL, C. M., BENNER, S. G., NICO, P. & FENDORF, S. 2004. Structural constraints of ferric (hydr) oxides on dissimilatory iron reduction and the fate of Fe (II). Geochimica et Cosmochimica Acta, 68, 3217-3229. HARLAVAN, Y., EREL, Y. & BLUM, J. D. 2009. The coupled release of REE and Pb to the soil labile pool with time by weathering of

297

References

accessory phases, Wind River Mountains, WY. Geochimica et Cosmochimica Acta, 73, 320-336. HASSINK, J., BOUWMAN, L., ZWART, K. & BRUSSAARD, L. 1993. Relationships between habitable pore space, soil biota and mineralization rates in grassland soils. Soil Biology and Biochemistry, 25, 47-55. HATCHER, P. G. 2004. The CHNs of organic geochemistry: characterization of molecularly uncharacterized non-living organic matter. Marine chemistry, 92, 5-8. HAYES, M. & SWIFT, R. 1990. Genesis, isolation, composition and structures of soil humic substances. Soil Colloids and Their Associations in Aggregates. Springer. HEDGES, J. I., EGLINTON, G., HATCHER, P. G., KIRCHMAN, D. L., ARNOSTI, C., DERENNE, S., EVERSHED, R. P., KÖGEL-KNABNER, I., DE LEEUW, J. & LITTKE, R. 2000. The molecularly- uncharacterized component of nonliving organic matter in natural environments. Organic Geochemistry, 31, 945-958. HEDGES, J. I. & KEIL, R. G. 1995. Sedimentary organic matter preservation: an assessment and speculative synthesis. Marine Chemistry, 49, 81-115. HEDGES, J. I. & KEIL, R. G. 1999. Organic geochemical perspectives on estuarine processes: sorption reactions and consequences. Marine Chemistry, 65, 55-65. HEIMANN, M. & REICHSTEIN, M. 2008. Terrestrial ecosystem carbon dynamics and climate feedbacks. Nature, 451, 289-292. HENDERSON, P. 1984. General geochemical properties and abundances of the rare earth elements. Rare earth element geochemistry. Elsevier Amsterdam. HENDERSON, R., KABENGI, N., MANTRIPRAGADA, N., CABRERA, M., HASSAN, S. & THOMPSON, A. 2012. Anoxia-Induced Release of Colloid- and Nanoparticle-Bound Phosphorus in Grassland Soils. Environmental Science & Technology, 46, 11727-11734. HENNEBERRY, Y. K., KRAUS, T. E., NICO, P. S. & HORWATH, W. R. 2012. Structural stability of coprecipitated natural organic matter and ferric iron under reducing conditions. Organic geochemistry, 48, 81-89. HENRIET, C., DE JAEGER, N., DOREL, M., OPFERGELT, S. & DELVAUX, B. 2008. The reserve of weatherable primary silicates impacts the

298

References

accumulation of biogenic silicon in volcanic ash soils. Biogeochemistry, 90, 209-223. HERBAUTS, J. 1982. Chemical and mineralogical properties of sandy and loamy‐sandy ochreous brown earths in relation to incipient podzolization in a brown earth—podzol evolutive sequence. Journal of Soil Science, 33, 743-762. HERBILLON, A. J. Chemical estimation of weatherable minerals present in the diagnostic horizons of low activity clay soils. In: BEINROTH, M. N., CAMARGO, M. N. & ESWARAN, H., eds. Proceedings of the 8th International Classification Workshop: Classification, Characterization, and Utilization of Ultisols. Part I. EMBRAPA, 1986 Rio de Janeiro. 39-48. HIEMSTRA, T., ANTELO, J., RAHNEMAIE, R. & VAN RIEMSDIJK, W. H. 2010a. Nanoparticles in natural systems I: the effective reactive surface area of the natural oxide fraction in field samples. Geochimica et Cosmochimica Acta, 74, 41-58. HIEMSTRA, T., ANTELO, J., VAN ROTTERDAM, A. M. D. & VAN RIEMSDIJK, W. H. 2010b. Nanoparticles in natural systems II: The natural oxide fraction at interaction with natural organic matter and phosphate. Geochimica et Cosmochimica Acta, 74, 59-69. HIGUCHI, T. 1990. Lignin biochemistry: biosynthesis and biodegradation. Wood Science and Technology, 24, 23-63. HISSLER, C., STILLE, P., JUILLERET, J., IFFLY, J. F., PERRONE, T. & MORVAN, G. 2015. Elucidating the formation of terra fuscas using Sr–Nd–Pb isotopes and rare earth elements. Applied Geochemistry, 54, 85-99. HOBBIE, S. E. & VITOUSEK, P. M. 2000. Nutrient limitation of decomposition in Hawaiian forests. Ecology, 81, 1867-1877. HOFFLAND, E., GIESLER, R., JONGMANS, T. & VAN BREEMEN, N. 2002. Increasing feldspar tunneling by fungi across a north Sweden podzol chronosequence. Ecosystems, 5, 11-22. HU, Z., HANEKLAUS, S., SPAROVEK, G. & SCHNUG, E. 2006. Rare Earth Elements in Soils. Communications in Soil Science and Plant Analysis, 37, 1381-1420. HUANG, P.-M., WANG, M.-K. & CHIU, C.-Y. 2005. Soil mineral–organic matter–microbe interactions: impacts on biogeochemical processes and biodiversity in soils. Pedobiologia, 49, 609-635.

299

References

HUGGETT, R. 1998. Soil chronosequences, soil development, and soil evolution: a critical review. Catena, 32, 155-172. HUGHES, J. M., CAMERON, M. & MARIANO, A. N. 1991. Rare-earth- element ordering and structural variations in natural rare- earth-bearing apatites. American Mineralogist, 76, 1165-1173. HUSSON, O. 2013. Redox potential (Eh) and pH as drivers of soil/plant/microorganism systems: a transdisciplinary overview pointing to integrative opportunities for agronomy. Plant and Soil, 362, 389-417. HYACINTHE, C., BONNEVILLE, S. & VAN CAPPELLEN, P. 2006. Reactive iron (III) in sediments: chemical versus microbial extractions. Geochimica et Cosmochimica Acta, 70, 4166-4180. IPCC-WG1 2013. Climate Change 2013 - The Physical Science Basis - fifth assessment report. Intergovernmental panel on climate change - Working Group I. Geneva, Switzerland: IPCC. JANOTS, E., BERNIER, F., BRUNET, F., MUÑOZ, M., TRCERA, N., BERGER, A. & LANSON, M. 2015. Ce(III) and Ce(IV) (re)distribution and fractionation in a laterite profile from Madagascar: Insights from in situ XANES spectroscopy at the Ce LIII-edge. Geochimica et Cosmochimica Acta, 153, 134-148. JANSEN, B., NIEROP, K. G. & VERSTRATEN, J. M. 2003. Mobility of Fe (II), Fe (III) and Al in acidic forest soils mediated by dissolved organic matter: influence of solution pH and metal/organic carbon ratios. Geoderma, 113, 323-340. JANSEN, B., NIEROP, K. G. J. & VERSTRATEN, J. M. 2004. Mobilization of dissolved organic matter, aluminium and iron in podzol eluvial horizons as affected by formation of metal-organic complexes and interactions with solid soil material. European Journal of Soil Science, 55, 287-297. JANSEN, B., NIEROP, K. G. J. & VERSTRATEN, J. M. 2005. Mechanisms controlling the mobility of dissolved organic matter, aluminium and iron in podzol B horizons. European Journal of Soil Science, 56, 537-550. JANZEN, H. 2015. Beyond carbon sequestration: soil as conduit of solar energy. European Journal of Soil Science, 66, 19-32. JEANROY, E. & GUILLET, B. 1981. The occurrence of suspended ferruginous particles in pyrophosphate extracts of some soil horizons. Geoderma, 26, 95-105.

300

References

JENNY, H. 1941 Factors of Soil Formation., New York. JENNY, H. 1961. Derivation of state factor equations of soils and ecosystems. Soil Science Society of America Journal, 25, 385- 388. JIANG, J. & KAPPLER, A. 2008. Kinetics of microbial and chemical reduction of humic substances: implications for electron shuttling. Environmental science & technology, 42, 3563-3569. JOHANNESSON, K. H., TANG, J., DANIELS, J. M., BOUNDS, W. J. & BURDIGE, D. J. 2004. Rare earth element concentrations and speciation in organic-rich blackwaters of the Great Dismal Swamp, Virginia, USA. Chemical Geology, 209, 271-294. JONES, A. R., SANDERMAN, J., ALLEN, D., DALAL, R. & SCHMIDT, S. 2015. Subtropical giant podzol chronosequence reveals that soil carbon stabilisation is not governed by litter quality. Biogeochemistry, 124, 205-217. JONES, D. & EDWARDS, A. 1998. Influence of sorption on the biological utilization of two simple carbon substrates. Soil Biology and Biochemistry, 30, 1895-1902. JONES, D. L., NGUYEN, C. & FINLAY, R. D. 2009. Carbon flow in the rhizosphere: carbon trading at the soil–root interface. Plant and Soil, 321, 5-33. JONGMANS, A., VAN BREEMEN, N., LUNDSTRÖM, U., VAN HEES, P., FINLAY, R., SRINIVASAN, M., UNESTAM, T., GIESLER, R., MELKERUD, P.-A. & OLSSON, M. 1997. Rock-eating fungi. Nature, 389, 682-683. JORDENS, A., CHENG, Y. P. & WATERS, K. E. 2013. A review of the beneficiation of rare earth element bearing minerals. Minerals Engineering, 41, 97-114. JOUQUET, P., BOTTINELLI, N., LATA, J.-C., MORA, P. & CAQUINEAU, S. 2007. Role of the fungus-growing termite Pseudacanthotermes spiniger (Isoptera, Macrotermitinae) in the dynamic of clay and soil organic matter content. An experimental analysis. Geoderma, 139, 127-133. KAHLE, M., KLEBER, M. & JAHN, R. 2003. Retention of dissolved organic matter by illitic soils and clay fractions: influence of mineral phase properties. Journal of Plant Nutrition and Soil Science, 166, 737-741.

301

References

KAISER, K. & GUGGENBERGER, G. 2000. The role of DOM sorption to mineral surfaces in the preservation of organic matter in soils. Organic geochemistry, 31, 711-725. KAISER, K. & GUGGENBERGER, G. 2003. Mineral surfaces and soil organic matter. European Journal of Soil Science, 54, 219-236. KAISER, K. & GUGGENBERGER, G. 2007. Sorptive stabilization of organic matter by microporous goethite: sorption into small pores vs. surface complexation. European Journal of Soil Science, 58, 45-59. KAISER, K. & KALBITZ, K. 2012. Cycling downwards – dissolved organic matter in soils. Soil Biology and Biochemistry, 52, 29-32. KAISER, K. & ZECH, W. 1996. Defects in estimation of aluminum in humus complexes of podzolic soils by pyrophosphate extraction. Soil Science, 161, 452-458. KALBITZ, K., SCHWESIG, D., RETHEMEYER, J. & MATZNER, E. 2005. Stabilization of dissolved organic matter by sorption to the mineral soil. Soil Biology and Biochemistry, 37, 1319-1331. KALBITZ, K., SOLINGER, S., PARK, J.-H., MICHALZIK, B. & MATZNER, E. 2000. Controls on the dynamics of dissolved organic matter in soils: a review. Soil science, 165, 277-304. KANDELER, E., TSCHERKO, D., BRUCE, K. D., STEMMER, M., HOBBS, P. J., BARDGETT, R. D. & AMELUNG, W. 2000. Structure and function of the soil microbial community in microhabitats of a heavy metal polluted soil. Biology and Fertility of Soils, 32, 390- 400. KANEV, V. 2011. Dynamics of acid-soluble iron compounds in soddy- podzolic soils of the southern Komi Republic. Eurasian Soil Science, 44, 1201-1214. KAPPLER, A. & STRAUB, K. L. 2005. Geomicrobiological Cycling of Iron. Reviews in Mineralogy and Geochemistry, 59, 85-108. KEILUWEIT, M., NICO, P. S., KLEBER, M. & FENDORF, S. 2016. Are oxygen limitations under recognized regulators of organic carbon turnover in upland soils? Biogeochemistry, 127, 157- 171. KELLEHER, B. P. & SIMPSON, A. J. 2006. Humic substances in soils: are they really chemically distinct? Environmental Science & Technology, 40, 4605-4611.

302

References

KEMMITT, S., LANYON, C., WAITE, I., WEN, Q., ADDISCOTT, T., BIRD, N. R., O’DONNELL, A. & BROOKES, P. 2008. Mineralization of native soil organic matter is not regulated by the size, activity or composition of the soil microbial biomass—a new perspective. Soil Biology and Biochemistry, 40, 61-73. KICZKA, M., WIEDERHOLD, J. G., FROMMER, J., KRAEMER, S. M., BOURDON, B. & KRETZSCHMAR, R. 2010. Iron isotope fractionation during proton-and ligand-promoted dissolution of primary phyllosilicates. Geochimica et Cosmochimica Acta, 74, 3112-3128. KIEM, R. & KÖGEL-KNABNER, I. 2003. Contribution of lignin and polysaccharides to the refractory carbon pool in C-depleted arable soils. Soil Biology and Biochemistry, 35, 101-118. KING, F. H. 1907. The Soil: its Nature, Relations, and Fundamental Principles of Management., New York. KLEBER, M. 2010. What is recalcitrant soil organic matter? Environmental Chemistry, 7, 320. KLEBER, M., EUSTERHUES, K., KEILUWEIT, M., MIKUTTA, C., MIKUTTA, R. & NICO, P. S. 2015. Chapter One-Mineral–Organic Associations: Formation, Properties, and Relevance in Soil Environments. Advances in Agronomy, 130, 1-140. KLEBER, M. & JOHNSON, M. G. 2010. Advances in understanding the molecular structure of soil organic matter: implications for interactions in the environment. Advances in agronomy, 106, 77-142. KLEBER, M., MIKUTTA, R., TORN, M. S. & JAHN, R. 2005. Poorly crystalline mineral phases protect organic matter in acid subsoil horizons. European Journal of Soil Science, 0, 050912034650054. KLEBER, M., NICO, P. S., PLANTE, A., FILLEY, T., KRAMER, M., SWANSTON, C. & SOLLINS, P. 2011. Old and stable soil organic matter is not necessarily chemically recalcitrant: implications for modeling concepts and temperature sensitivity. Global Change Biology, 17, 1097-1107. KLEBER, M., SOLLINS, P. & SUTTON, R. 2007. A conceptual model of organo-mineral interactions in soils: self-assembly of organic molecular fragments into zonal structures on mineral surfaces. Biogeochemistry, 85, 9-24.

303

References

KLOTZBÜCHER, T., KAISER, K., GUGGENBERGER, G., GATZEK, C. & KALBITZ, K. 2011. A new conceptual model for the fate of lignin in decomposing plant litter. Ecology, 92, 1052-1062. KNICKER, H. 2011. Soil organic N-An under-rated player for C sequestration in soils? Soil Biology and Biochemistry, 43, 1118- 1129. KNORR, K.-H. 2013. DOC-dynamics in a small headwater catchment as driven by redox fluctuations and hydrological flow paths–are DOC exports mediated by iron reduction/oxidation cycles? Biogeosciences, 10, 891-904. KOELE, N., STORCH, F. & HILDEBRAND, E. E. 2011. The coarse‐soil fraction is the main living space of fungal hyphae in the BhBs horizon of a Podzol. Journal of Plant Nutrition and Soil Science, 174, 750-753. KOEPPENKASTROP, D. & DE CARLO, E. H. 1992. Sorption of rare-earth elements from seawater onto synthetic mineral particles: An experimental approach. Chemical Geology, 95, 251-263. KOERNER, W., DUPOUEY, J., DAMBRINE, E. & BENOIT, M. 1997. Influence of past land use on the vegetation and soils of present day forest in the Vosges mountains, France. Journal of ecology, 351-358. KÖGEL-KNABNER, I. 2002. The macromolecular organic composition of plant and microbial residues as inputs to soil organic matter. Soil Biology and Biochemistry, 34, 139-162. KÖGEL-KNABNER, I., GUGGENBERGER, G., KLEBER, M., KANDELER, E., KALBITZ, K., SCHEU, S., EUSTERHUES, K. & LEINWEBER, P. 2008b. Organo-mineral associations in temperate soils: Integrating biology, mineralogy, and organic matter chemistry. Journal of Plant Nutrition and Soil Science, 171, 61-82. KOLATTUKUDY, P. E. 2001. Polyesters in Higher Plants. In: BABEL, W. & STEINBÜCHEL, A. (eds.) Biopolyesters. Berlin, Heidelberg: Springer Berlin Heidelberg. KONONOVA, M. 1961. Soil Organic Matter: Its Nature, Its Role in Soil Formation and in Soil Fertility. , New York, NY, Pergamon Press Inc. KOSTKA, J. E., STUCKI, L. J. W., NEALSON, K. H. & WU, J. 1996. Reduction of structural Fe (III) in smectite by a pure culture of

304

References

the Fe-reducing bacterium Shewanella putrifaciens strain MR- 1. Clays and Clay Minerals, 44, 522-529. KOTLOSKI, N. J. & GRALNICK, J. A. 2013. Flavin electron shuttles dominate extracellular electron transfer by Shewanella oneidensis. MBio, 4, e00553-12. KRAEMER, S. M. 2004. Iron oxide dissolution and solubility in the presence of siderophores. Aquatic sciences, 66, 3-18. KRULL, E. S., BALDOCK, J. A., SKJEMSTAD, J. O. & SMERNIK, R. J. 2009. Characteristics of biochar: organo-chemical properties. Biochar for environmental management: Science and technology. Earthscan, London, 53-65. KUKKADAPU, R. K., ZACHARA, J. M., SMITH, S. C., FREDRICKSON, J. K. & LIU, C. 2001. Dissimilatory bacterial reduction of Al-substituted goethite in subsurface sediments. Geochimica et Cosmochimica Acta, 65, 2913-2924. KURTZ, A. C., DERRY, L. A. & CHADWICK, O. A. 2002. Germanium-silicon fractionation in the weathering environment. Geochimica et Cosmochimica Acta, 66, 1525-1537. KUZYAKOV, Y. & BLAGODATSKAYA, E. 2015. Microbial hotspots and hot moments in soil: Concept & review. Soil Biology and Biochemistry, 83, 184-199. KUZYAKOV, Y. & DOMANSKI, G. 2000. Carbon input by plants into the soil. Review. Journal of Plant Nutrition and Soil Science, 163, 421-431. KUZYAKOV, Y., FRIEDEL, J. & STAHR, K. 2000. Review of mechanisms and quantification of priming effects. Soil Biology and Biochemistry, 32, 1485-1498. LAL, R. 2004. Soil carbon sequestration impacts on global climate change and food security. Science, 304, 1623-7. LAL, R. 2009. Challenges and opportunities in soil organic matter research. European Journal of Soil Science, 60, 158-169. LAND, M., ÖHLANDER, B., INGRI, J. & THUNBERG, J. 1999. Solid speciation and fractionation of rare earth elements in a spodosol profile from northern Sweden as revealed by sequential extraction. Chemical Geology, 160, 121-138. LAVELLE, P., BLANCHART, E., MARTIN, A., MARTIN, S. & SPAIN, A. 1993. A hierarchical model for decomposition in terrestrial

305

References

ecosystems: application to soils of the humid tropics. Biotropica, 130-150. LAVEUF, C. & CORNU, S. 2009. A review on the potentiality of Rare Earth Elements to trace pedogenetic processes. Geoderma, 154, 1-12. LAVEUF, C., CORNU, S. & JUILLOT, F. 2008. Rare earth elements as tracers of pedogenetic processes. Comptes Rendus - Geoscience, 340, 523-532. LEGROS, J. P. 2007. Les grands sols du monde, Lausanne, Suisse, Presses polytechniques et universitaires romandes. LEHMANN, J. & KLEBER, M. 2015. The contentious nature of soil organic matter. Nature, 528, 60-68. LEHMANN, J., SOLOMON, D., KINYANGI, J., DATHE, L., WIRICK, S. & JACOBSEN, C. 2008. Spatial complexity of soil organic matter forms at nanometre scales. Nature Geoscience, 1, 238-242. LETH, S. & BREUNING-MADSEN, H. 1992. Changes in soil profile development and nutrient status due to the afforestation of agricultural land. Geografisk Tidsskrift-Danish Journal of Geography, 92, 70-74. LI, F.-M., WANG, X.-L., LI, Y., GUO, S.-H. & ZHONG, A.-P. 2006. Selective extraction and separation of Fe, Mn oxides and organic materials in river surficial sediments. Journal of Environmental Sciences, 18, 1233-1240. LI, Y., BEISSON, F., OHLROGGE, J. & POLLARD, M. 2007. Monoacylglycerols are components of root waxes and can be produced in the aerial cuticle by ectopic expression of a suberin-associated acyltransferase. Plant physiology, 144, 1267-1277. LIANG, B., LEHMANN, J., SOLOMON, D., SOHI, S., THIES, J. E., SKJEMSTAD, J. O., LUIZAO, F. J., ENGELHARD, M. H., NEVES, E. G. & WIRICK, S. 2008a. Stability of biomass-derived black carbon in soils. Geochimica et Cosmochimica Acta, 72, 6069- 6078. LIANG, C. & BALSER, T. C. 2011. Microbial production of recalcitrant organic matter in global soils: implications for productivity and climate policy. Nature Reviews Microbiology, 9, 75-75.

306

References

LIANG, C., FUJINUMA, R. & BALSER, T. C. 2008b. Comparing PLFA and amino sugars for microbial analysis in an Upper Michigan old growth forest. Soil Biology and Biochemistry, 40, 2063-2065. LIANG, C., ZHANG, X. & BALSER, T. C. 2007. Net microbial amino sugar accumulation process in soil as influenced by different plant material inputs. Biology and Fertility of Soils, 44, 1-7. LIN, B., HYACINTHE, C., BONNEVILLE, S., BRASTER, M., VAN CAPPELLEN, P. & RÖLING, W. F. 2007. Phylogenetic and physiological diversity of dissimilatory ferric iron reducers in sediments of the polluted Scheldt estuary, Northwest Europe. Environmental Microbiology, 9, 1956-1968. LINDEBURG, K. S., ALMOND, P., ROERING, J. J. & CHADWICK, O. A. 2013. Pathways of soil genesis in the Coast Range of Oregon, USA. Plant and Soil, 367, 57-75. LORENZ, K. & LAL, R. 2014. Soil organic carbon sequestration in agroforestry systems. A review. Agronomy for Sustainable Development, 34, 443-454. LORENZ, K., LAL, R., PRESTON, C. M. & NIEROP, K. G. J. 2007. Strengthening the soil organic carbon pool by increasing contributions from recalcitrant aliphatic bio(macro)molecules. Geoderma, 142, 1-10. LOVLEY, D. 2013. Dissimilatory Fe(III)- and Mn(IV)-Reducing Prokaryotes. In: ROSENBERG, E., DELONG, E. F., LORY, S., STACKEBRANDT, E. & THOMPSON, F. (eds.) The Prokaryotes: Prokaryotic Physiology and Biochemistry. Berlin, Heidelberg: Springer Berlin Heidelberg. LOVLEY, D., FRAGA, J. L., BLUNT-HARRIS, E. L., HAYES, L., PHILLIPS, E. & COATES, J. D. 1998. Humic substances as a mediator for microbially catalyzed metal reduction. Acta hydrochimica et hydrobiologica, 26, 152-157. LUCAS, Y., NAHON, D., CORNU, S. & EYROLLE, F. 1996. Genèse et fonctionnement des sols en milieu équatorial. Comptes rendus de l'Académie des sciences. Série 2. Sciences de la terre et des planètes, 322, 1-16. LUNDQUIST, E., JACKSON, L. & SCOW, K. 1999. Wet–dry cycles affect dissolved organic carbon in two California agricultural soils. Soil Biology and Biochemistry, 31, 1031-1038.

307

References

LUNDSTRÖM, U., BREEMEN, N. V. & JONGMANS, A. 1995. Evidence for microbial decomposition of organic acids during podzolization. European Journal of Soil Science, 46, 489-496. LUNDSTRÖM, U., VAN, N. & BAIN, D. 2000a. The podzolization process. A review. Geoderma. LUNDSTRÖM, U. V., VAN BREEMEN, N., BAIN, D., VAN HEES, P., GIESLER, R., GUSTAFSSON, J. P., ILVESNIEMI, H., KARLTUN, E., MELKERUD, P.-A. & OLSSON, M. 2000b. Advances in understanding the podzolization process resulting from a multidisciplinary study of three coniferous forest soils in the Nordic Countries. Geoderma, 94, 335-353. LÜTZOW, M. V. & KÖGEL-KNABNER, I. 2010. Response to the Concept paper:'What is recalcitrant soil organic matter?'by Markus Kleber. Environmental Chemistry, 7, 333-335. MA, L., JIN, L. & BRANTLEY, S. L. 2011. How mineralogy and slope aspect affect REE release and fractionation during shale weathering in the Susquehanna/Shale Hills Critical Zone Observatory. Chemical Geology, 290, 31-49. MAJOR, J., LEHMANN, J., RONDON, M. & GOODALE, C. 2010. Fate of soil‐applied black carbon: downward migration, leaching and soil respiration. Global Change Biology, 16, 1366-1379. MANZONI, S. & PORPORATO, A. 2009. Soil carbon and nitrogen mineralization: theory and models across scales. Soil Biology and Biochemistry, 41, 1355-1379. MANZONI, S., TAYLOR, P., RICHTER, A., PORPORATO, A. & ÅGREN, G. I. 2012. Environmental and stoichiometric controls on microbial carbon‐use efficiency in soils. New Phytologist, 196, 79-91. MARÉCHAL, C. N., TÉLOUK, P. & ALBARÈDE, F. 1999. Precise analysis of copper and zinc isotopic compositions by plasma-source mass spectrometry. Chemical Geology, 156, 251-273. MARESCHAL, L., TURPAULT, M.-P., BONNAUD, P. & RANGER, J. 2013. Relationship between the weathering of clay minerals and the nitrification rate: a rapid tree species effect. Biogeochemistry, 112, 293-309. MARSAC, R., DAVRANCHE, M., GRUAU, G., BOUHNIK-LE COZ, M. & DIA, A. 2011. An improved description of the interactions between rare earth elements and humic acids by modeling: PHREEQC-

308

References

Model VI coupling. Geochimica et Cosmochimica Acta, 75, 5625-5637. MARSAC, R., DAVRANCHE, M., GRUAU, G. & DIA, A. 2010. Metal loading effect on rare earth element binding to humic acid: Experimental and modelling evidence. Geochimica et Cosmochimica Acta, 74, 1749-1761. MARSCHNER, B., BRODOWSKI, S., DREVES, A., GLEIXNER, G., GUDE, A., GROOTES, P. M., HAMER, U., HEIM, A., JANDL, G., JI, R., KAISER, K., KALBITZ, K., KRAMER, C., LEINWEBER, P., RETHEMEYER, J., SCHÄFFER, A., SCHMIDT, M. W. I., SCHWARK, L. & WIESENBERG, G. L. B. 2008. How relevant is recalcitrance for the stabilization of organic matter in soils? Journal of Plant Nutrition and Soil Science, 171, 91-110. MARSH, J. S. 1991. REE fractionation and Ce anomalies in weathered Karoo dolerite. Chemical Geology, 90, 189-194. MARSILI, E., BARON, D. B., SHIKHARE, I. D., COURSOLLE, D., GRALNICK, J. A. & BOND, D. R. 2008. Shewanella secretes flavins that mediate extracellular electron transfer. Proceedings of the National Academy of Sciences, 105, 3968-3973. MAYER, L. M. 1994a. Relationships between mineral surfaces and organic carbon concentrations in soils and sediments. Chemical Geology, 114, 347-363. MAYER, L. M. 1994b. Surface area control of organic carbon accumulation in continental shelf sediments. Geochimica et Cosmochimica Acta, 58, 1271-1284. MCKEAGUE, J., ROSS, G., GAMBLE, D. & MAHANEY, W. 1978. Properties, criteria of classification and genesis of podzolic soils in Canada. Quaternary Soils, 27-60. MEHRA, O. P. & JACKSON, M. L. 1960. Iron oxide removal from soils and clays by a dithionite-citrate system buffered with sodium bicarbonate. Clays and Clay Minerals, 7, 317-327. MELTON, E. D., SWANNER, E. D., BEHRENS, S., SCHMIDT, C. & KAPPLER, A. 2014. The interplay of microbially mediated and abiotic reactions in the biogeochemical Fe cycle. Nature Reviews Microbiology, 12, 797-808. MICHALOPOULOS, P. & ALLER, R. C. 1995. Rapid clay mineral formation of Amazon delta sediments: Reverse weathering and oceanic elemental cycles. Science, 270, 614.

309

References

MIKUTTA, C., MIKUTTA, R., BONNEVILLE, S., WAGNER, F., VOEGELIN, A., CHRISTL, I. & KRETZSCHMAR, R. 2008. Synthetic coprecipitates of exopolysaccharides and ferrihydrite. Part I: Characterization. Geochimica et Cosmochimica Acta, 72, 1111- 1127. MIKUTTA, R., KLEBER, M. & JAHN, R. 2005a. Poorly crystalline minerals protect organic carbon in clay subfractions from acid subsoil horizons. Geoderma, 128, 106-115. MIKUTTA, R., KLEBER, M., KAISER, K. & JAHN, R. 2005b. Review: Organic Matter Removal from Soils using Hydrogen Peroxide, Sodium Hypochlorite, and Disodium Peroxodisulfate. Soil Science Society of Amerca, 69, 120-135. MIKUTTA, R., KLEBER, M., TORN, M. S. & JAHN, R. 2006. Stabilization of Soil Organic Matter: Association with Minerals or Chemical Recalcitrance? Biogeochemistry, 77, 25-56. MIKUTTA, R., MIKUTTA, C., KALBITZ, K., SCHEEL, T., KAISER, K. & JAHN, R. 2007. Biodegradation of forest floor organic matter bound to minerals via different binding mechanisms. Geochimica et Cosmochimica Acta, 71, 2569-2590. MILTNER, A., BOMBACH, P., SCHMIDT-BRÜCKEN, B. & KÄSTNER, M. 2012. SOM genesis: microbial biomass as a significant source. Biogeochemistry, 111, 41-55. MINAMI, K. 2009. Soil and humanity: Culture, civilization, livelihood and health. Soil science and plant nutrition, 55, 603-615. MONGELLI, G. 1993. REE and other trace elements in a granitic weathering profile from “Serre”, southern Italy. Chemical Geology, 103, 17-25. MONTES, C. R., LUCAS, Y., PEREIRA, O. J. R., ACHARD, R., GRIMALDI, M. & MELFI, A. J. 2011. Deep plant-derived carbon storage in Amazonian podzols. Biogeosciences, 8, 113-120. MOORE, T. & DALVA, M. 2001. Some controls on the release of dissolved organic carbon by plant tissues and soils. Soil Science, 166, 38-47. MOSSIN, L., MORTENSEN, M. & NØRNBERG, P. 2002. Imogolite related to podzolization processes in Danish podzols. Geoderma, 109, 103-116. MOURIER, B., POULENARD, J., CHAUVEL, C., FAIVRE, P. & CARCAILLET, C. 2008. Distinguishing subalpine soil types using extractible Al

310

References

and Fe fractions and REE geochemistry. Geoderma, 145, 107- 120. MOYNIER, F., ALBARÈDE, F. & HERZOG, G. 2006. Isotopic composition of zinc, copper, and iron in lunar samples. Geochimica et Cosmochimica Acta, 70, 6103-6117. MULLER, J. E. & CARSON, D. J. T. 1969. GELOGY AND MINERAL DEPOSITS OF ALBERNI MAP-AREA, BRITISH COLUMBIA. Geol Surv Can. NAKADA, R., TAKAHASHI, Y. & TANIMIZU, M. 2013. Isotopic and speciation study on cerium during its solid–water distribution with implication for Ce stable isotope as a paleo-redox proxy. Geochimica et Cosmochimica Acta, 103, 49-62. NANNIPIERI, P., ASCHER, J., CECCHERINI, M., LANDI, L., PIETRAMELLARA, G. & RENELLA, G. 2003. Microbial diversity and soil functions. European Journal of Soil Science, 54, 655- 670. NAWRATH, C. 2006. Unraveling the complex network of cuticular structure and function. Current Opinion in Plant Biology, 9, 281- 287. NDJIGUI, P.-D., BILONG, P., BITOM, D. & DIA, A. 2008. Mobilization and redistribution of major and trace elements in two weathering profiles developed on serpentinites in the Lomié ultramafic complex, South-East Cameroon. Journal of African Earth Sciences, 50, 305-328. NEALSON, K. H. & MYERS, C. R. 1992. Microbial reduction of manganese and iron: new approaches to carbon cycling. Applied and Environmental Microbiology, 58, 439. NEALSON, K. H. & SAFFARINI, D. 1994. Iron and manganese in anaerobic respiration: environmental significance, physiology, and regulation. Annual Reviews in Microbiology, 48, 311-343. NEFF, J., TOWNSEND, A., GLEIXNER, G., LEHMAN, S., TURNBULL, J. & BOWMAN, W. 2002. Variable effects of nitrogen additions on the stability and turnover of soil carbon. Nature, 419, 915-917. NEFF, J. C. & ASNER, G. P. 2001. Dissolved organic carbon in terrestrial ecosystems: synthesis and a model. Ecosystems, 4, 29-48. NEILANDS, J. 1995. Siderophores: structure and function of microbial iron transport compounds. Journal of Biological Chemistry, 270, 26723-26726.

311

References

NESBITT, H. W. 1979. Mobility and fractionation of rare earth elements during weathering of a granodiorite. NEWMAN, D. K. 2001. How bacteria respire minerals. Science, 292, 1312-1313. NICKEL, E. 1973. Experimental dissolution of light and heavy minerals in comparison with weathering and intrastratal solution. Contributions to Sedimentology, 1, 1-68. NIEROP, K. G. J. J., JANSEN, B. & VERSTRATEN, J. M. 2002. Dissolved organic matter, aluminium and iron interactions: precipitation induced by metal/carbon ratio, pH and competition. Science of The Total Environment, 300, 201-211. NIKONOV, V., LUKINA, N., POLYANSKAYA, L. & PANIKOVA, A. 2001. Distribution of microorganisms in the Al–Fe–humus podzols of natural and anthropogenically impacted boreal spruce forests. Microbiology, 70, 319-328. NUNAN, N., LERCH, T., POUTEAU, V., MORA, P., CHANGEY, F., KÄTTERER, T., GIUSTI-MILLER, S. & HERRMANN, A. 2015. Metabolising old soil carbon: Simply a matter of simple organic matter? Soil Biology and Biochemistry, 88, 128-136. ÖHLANDER, B., INGRI, J., LAND, M. & SCHÖBERG, H. 2000. Change of Sm-Nd isotope composition during weathering of till. Geochimica et Cosmochimica Acta, 64, 813-820. ÖHLANDER, B., LAND, M., INGRI, J. & WIDERLUND, A. 1996. Mobility of rare earth elements during weathering of till in northern Sweden. Applied Geochemistry, 11, 93-99. OPFERGELT, S., CARDINAL, D., ANDRÉ, L., DELVIGNE, C., BREMOND, L. & DELVAUX, B. 2010. Variations of δ 30 Si and Ge/Si with weathering and biogenic input in tropical basaltic ash soils under monoculture. Geochimica et Cosmochimica Acta, 74, 225-240. OPFERGELT, S., CARDINAL, D., HENRIET, C., DRAYE, X., ANDRÉ, L. & DELVAUX, B. 2006. Silicon isotopic fractionation by banana (Musa spp.) grown in a continuous nutrient flow device. Plant and Soil, 285, 333-345. OPFERGELT, S., DE BOURNONVILLE, G., CARDINAL, D., ANDRÉ, L., DELSTANCHE, S. & DELVAUX, B. 2009. Impact of soil weathering degree on silicon isotopic fractionation during adsorption onto

312

References

iron oxides in basaltic ash soils, Cameroon. Geochimica et Cosmochimica Acta, 73, 7226-7240. OPFERGELT, S., GEORG, R., DELVAUX, B., CABIDOCHE, Y.-M., BURTON, K. & HALLIDAY, A. 2012. Silicon isotopes and the tracing of desilication in volcanic soil weathering sequences, Guadeloupe. Chemical Geology, 326, 113-122. PAGE, J. R., MILLER, R. H., KEENEY, D. H., BAKER, D. E., ROSCOE, J. R. & J.D., R. (eds.) 1982. Methods of soil analysis: part 2, chemical and microbiological properties, 2nd edn, Madison, WI. PALUMBO, B., BELLANCA, A., NERI, R. & ROE, M. J. 2001. Trace metal partitioning in Fe–Mn nodules from Sicilian soils, Italy. Chemical Geology, 173, 257-269. PANAHI, A., YOUNG, G. M. & RAINBIRD, R. H. 2000. Behavior of major and trace elements (including REE) during Paleoproterozoic pedogenesis and diagenetic alteration of an Archean granite near Ville Marie, Quebec, Canada. Geochimica et Cosmochimica Acta, 64, 2199-2220. PARFITT, R., THENG, B., WHITTON, J. & SHEPHERD, T. 1997. Effects of clay minerals and land use on organic matter pools. Geoderma, 75, 1-12. PARFITT, R. L. 2009. Allophane and imogolite: role in soil biogeochemical processes. Clay Minerals, 44, 135-155. PAUL, E. A. & CLARK, F. E. 1989. Soil microbiology and biochemistry., San Diego, California, Academic Press, Inc. PEDROT, M., DIA, A., DAVRANCHE, M., BOUHNIK-LE COZ, M., HENIN, O. & GRUAU, G. 2008. Insights into colloid-mediated trace element release at the soil/water interface. J Colloid Interface Sci, 325, 187-97. PÉDROT, M., LE BOUDEC, A., DAVRANCHE, M., DIA, A. & HENIN, O. 2011. How does organic matter constrain the nature, size and availability of Fe nanoparticles for biological reduction? Journal of colloid and interface science, 359, 75-85. PEEL, M. C., FINLAYSON, B. L. & MCMAHON, T. A. 2007. Updated world map of the Köppen-Geiger climate classification. Hydrology and earth system sciences discussions, 4, 439-473. PERRET, D., GAILLARD, J.-F., DOMINIK, J. & ATTEIA, O. 2000. The diversity of natural hydrous iron oxides. Environmental science & technology, 34, 3540-3546.

313

References

PICCOLO, A. 2001. The supramolecular structure of humic substances. Soil science, 166, 810-832. POGGENBURG, C., MIKUTTA, R., SANDER, M., SCHIPPERS, A., MARCHANKA, A., DOHRMANN, R. & GUGGENBERGER, G. 2016. Microbial reduction of ferrihydrite-organic matter coprecipitates by Shewanella putrefaciens and Geobacter metallireducens in comparison to mediated electrochemical reduction. Chemical Geology. POITRASSON, F., VIERS, J., MARTIN, F. & BRAUN, J.-J. 2008. Limited iron isotope variations in recent lateritic soils from Nsimi, Cameroon: implications for the global Fe geochemical cycle. Chemical Geology, 253, 54-63. POKROVSKI, G. S. & SCHOTT, J. 1998. Experimental study of the complexation of silicon and germanium with aqueous organic species: implications for germanium and silicon transport and Ge/Si ratio in natural waters. Geochimica et Cosmochimica Acta, 62, 3413-3428. POLL, C., INGWERSEN, J., STEMMER, M., GERZABEK, M. H. & KANDELER, E. 2006. Mechanisms of solute transport affect small-scale abundance and function of soil microorganisms in the detritusphere. European Journal of Soil Science, 57, 583- 595. POLLIERER, M. M., LANGEL, R., KÖRNER, C., MARAUN, M. & SCHEU, S. 2007. The underestimated importance of belowground carbon input for forest soil animal food webs. Ecology Letters, 10, 729- 736. PONGE, J.-F. 2015. The soil as an ecosystem. Biology and Fertility of Soils, 51, 645-648. POULTON, S. W. & CANFIELD, D. E. 2005. Development of a sequential extraction procedure for iron: Implications for iron partitioning in continentally derived particulates. Chemical Geology, 214, 209-221. POURRET, O., DAVRANCHE, M., GRUAU, G. & DIA, A. 2007a. Organic complexation of rare earth elements in natural waters: Evaluating model calculations from ultrafiltration data. Geochimica et Cosmochimica Acta, 71, 2718-2735.

314

References

POURRET, O., DAVRANCHE, M., GRUAU, G. & DIA, A. 2007b. Rare earth elements complexation with humic acid. Chemical Geology, 243, 128-141. PRESTON, C. M. & SCHMIDT, M. W. I. 2006. Black (pyrogenic) carbon in boreal forests: a synthesis of current knowledge and uncertainties. Biogeosciences Discussions, 3, 211-271. PRONK, G. J., HEISTER, K. & KÖGEL-KNABNER, I. 2015. Amino sugars reflect microbial residues as affected by clay mineral composition of artificial soils. Organic Geochemistry, 83, 109- 113. PRUDENCIO, M., BRAGA, M. & GOUVEIA, M. 1993. REE mobilization, fractionation and precipitation during weathering of basalts. Chemical Geology, 107, 251-254. QUENEA, K., DERENNE, S., LARGEAU, C., RUMPEL, C. & MARIOTTI, A. 2004. Variation in lipid relative abundance and composition among different particle size fractions of a forest soil. Organic Geochemistry, 35, 1355-1370. RANKIN, P. C. & CHILDS, C. W. 1976. Rare-earth elements in iron- manganese concretions from some New Zealand soils. Chemical Geology, 18, 55-64. RASSE, D. P., RUMPEL, C. & DIGNAC, M.-F. 2005. Is soil carbon mostly root carbon? Mechanisms for a specific stabilisation. Plant and Soil, 269, 341-356. REGELINK, I. C., WENG, L., KOOPMANS, G. F. & VAN RIEMSDIJK, W. H. 2013. Asymmetric flow field-flow fractionation as a new approach to analyse iron-(hydr) oxide nanoparticles in soil extracts. Geoderma, 202, 134-141. REICHSTEIN, M., KÄTTERER, T., ANDRÉN, O., CIAIS, P., SCHULZE, E.-D., CRAMER, W., PAPALE, D. & VALENTINI, R. 2005. Temperature sensitivity of decomposition in relation to soil organic matter pools: critique and outlook. Biogeosciences, 2, 317-321. REYNOLDS, B. C., AGGARWAL, J., ANDRÉ, L., BAXTER, D., BEUCHER, C., BRZEZINSKI, M. A., ENGSTRÖM, E., GEORG, R. B., LAND, M. & LENG, M. J. 2007. An inter-laboratory comparison of Si isotope reference materials. Journal of Analytical Atomic Spectrometry, 22, 561-568.

315

References

RIGHI, D., HUBER, K. & KELLER, C. 1999. Clay formation and podzol development from postglacial moraines in Switzerland. Clay Minerals, 34, 319-332. RIGHI, D., RÄISÄNEN, M. L. & GILLOT, F. 1997. Clay mineral transformations in podzolized tills in central Finland. Clay Minerals, 32, 531-544. RILLIG, M. C., CALDWELL, B. A., WÖSTEN, H. A. & SOLLINS, P. 2007. Role of proteins in soil carbon and nitrogen storage: controls on persistence. Biogeochemistry, 85, 25-44. RITZ, K. & YOUNG, I. M. 2004. Interactions between soil structure and fungi. Mycologist, 18, 52-59. RODELLA, A. & SABOYA, L. 1999. Calibration for conductimetric determination of carbon dioxide. Soil Biology and Biochemistry, 31, 2059-2060. RODEN, E. E., KAPPLER, A., BAUER, I., JIANG, J., PAUL, A., STOESSER, R., KONISHI, H. & XU, H. 2010. Extracellular electron transfer through microbial reduction of solid-phase humic substances. Nature geoscience, 3, 417-421. RÖLING, W. F., VAN BREUKELEN, B. M., BRUGGEMAN, F. J. & WESTERHOFF, H. V. 2007. Ecological control analysis: being (s) in control of mass flux and metabolite concentrations in anaerobic degradation processes. Environmental microbiology, 9, 500-511. ROYER, R. A., BURGOS, W. D., FISHER, A. S., UNZ, R. F. & DEMPSEY, B. A. 2002. Enhancement of biological reduction of hematite by electron shuttling and Fe (II) complexation. Environmental science & technology, 36, 1939-1946. RUDNICK, R. L. & GAO, S. 2003. 3.01 - Composition of the Continental Crust. In: TUREKIAN, H. D. H. K. (ed.) Treatise on Geochemistry. Oxford: Pergamon. RUMPEL, C., EUSTERHUES, K. & KÖGEL-KNABNER, I. 2004. Location and chemical composition of stabilized organic carbon in topsoil and subsoil horizons of two acid forest soils. Soil Biology and Biochemistry, 36, 177-190. RUMPEL, C. & KÖGEL-KNABNER, I. 2011. Deep soil organic matter—a key but poorly understood component of terrestrial C cycle. Plant and Soil, 338, 143-158.

316

References

SAATZ, J., VETTERLEIN, D., MATTUSCH, J., OTTO, M. & DAUS, B. 2015. The influence of gadolinium and yttrium on biomass production and nutrient balance of maize plants. Environmental Pollution, 204, 32-38. SALOME, C., NUNAN, N., POUTEAU, V., LERCH, T. Z. & CHENU, C. 2010. Carbon dynamics in topsoil and in subsoil may be controlled by different regulatory mechanisms. Global Change Biology, 16, 416-426. SANBORN, P., LAMONTAGNE, L. & HENDERSHOT, W. 2011. Podzolic soils of Canada: Genesis, distribution, and classification. Canadian Journal of Soil Science, 91, 843-880. SANEMATSU, K., MORIYAMA, T., SOTOUKY, L. & WATANABE, Y. 2011. Mobility of Rare Earth Elements in Basalt-Derived Laterite at the Bolaven Plateau, Southern Laos. Resource Geology, 61, 140-158. SAUER, D. 2015. Pedological concepts to be considered in soil chronosequence studies. Soil Research, 53, 577-591. SAUER, D., SCHÜLLI-MAURER, I., SPERSTAD, R., SØRENSEN, R. & STAHR, K. 2008. Podzol development with time in sandy beach deposits in southern Norway. Journal of Plant Nutrition and Soil Science, 171, 483-497. SAUER, D., SPONAGEL, H., SOMMER, M., GIANI, L., JAHN, R. & STAHR, K. 2007. Podzol: Soil of the Year 2007. A review on its genesis, occurrence, and functions. Journal of Plant Nutrition and Soil Science, 170, 581-597. SAVAGE, P. S., GEORG, R. B., WILLIAMS, H. M., TURNER, S., HALLIDAY, A. N. & CHAPPELL, B. W. 2012. The silicon isotope composition of granites. Geochimica et Cosmochimica Acta, 92, 184-202. SCHÄDLER, S., BURKHARDT, C. & KAPPLER, A. 2008. Evaluation of Electron Microscopic Sample Preparation Methods and Imaging Techniques for Characterization of Cell-Mineral Aggregates. Geomicrobiology Journal, 25, 228-239. SCHAETZL, R., BARRETT, L. & WINKLER, J. 1994. Choosing models for soil chronofunctions and fitting them to data. European Journal of Soil Science, 45, 219-232. SCHEEL, T., HAUMAIER, L., ELLERBROCK, R. H., RÜHLMANN, J. & KALBITZ, K. 2008. Properties of organic matter precipitated

317

References

from acidic forest soil solutions. Organic Geochemistry, 39, 1439-1453. SCHMIDT, C., ZIMMERMANN, R. & GAUB, H. 1990. Multilayer adsorption of lysozyme on a hydrophobic substrate. Biophysical journal, 57, 577. SCHMIDT, M. W., KNICKER, H. & KOÈGEL-KNABNER, I. 2000. Organic matter accumulating in Aeh and Bh horizons of a Podzol— chemical characterization in primary organo-mineral associations. Organic Geochemistry, 31, 727-734. SCHMIDT, M. W., TORN, M. S., ABIVEN, S., DITTMAR, T., GUGGENBERGER, G., JANSSENS, I. A., KLEBER, M., KOGEL- KNABNER, I., LEHMANN, J., MANNING, D. A., NANNIPIERI, P., RASSE, D. P., WEINER, S. & TRUMBORE, S. E. 2011. Persistence of soil organic matter as an ecosystem property. Nature, 478, 49-56. SCHULTEN, H.-R. & SCHNITZER, M. 1997. Chemical model structures for soil organic matter and soils. Soil Science, 162, 115-130. SCHULZ, S., BRANKATSCHK, R., DÜMIG, A., KÖGEL-KNABNER, I., SCHLOTER, M. & ZEYER, J. 2013. The role of microorganisms at different stages of ecosystem development for soil formation. Biogeosciences, 10, 3983-3996. SCHULZE, K., BORKEN, W., MUHR, J. & MATZNER, E. 2009. Stock, turnover time and accumulation of organic matter in bulk and density fractions of a Podzol soil. European Journal of Soil Science, 60, 567-577. SCHUPPLI, P., ROSS, G. & MCKEAGUE, J. 1983. The effective removal of suspended materials from pyrophosphate extracts of soils from tropical and temperate regions. Soil Science Society of America Journal, 47, 1026-1032. SCHWERTMANN, U. 1966. Inhibitory effect of soil organic matter on the crystallization of amorphous ferric hydroxide. Nature, 212, 645-646. SCOTT, M., JONES, M., WOOF, C. & TIPPING, E. 1998. Concentrations and fluxes of dissolved organic carbon in drainage water from an upland peat system. Environment International, 24, 537- 546. SEXSTONE, A. J., REVSBECH, N. P., PARKIN, T. B. & TIEDJE, J. M. 1985. Direct measurement of oxygen profiles and denitrification

318

References

rates in soil aggregates. Soil science society of America journal, 49, 645-651. SHIMIZU, M., ZHOU, J., SCHRÖDER, C., OBST, M., KAPPLER, A. & BORCH, T. 2013. Dissimilatory reduction and transformation of ferrihydrite-humic acid coprecipitates. Environmental science & technology, 47, 13375-13384. SIMPSON, A. J., SIMPSON, M. J., SMITH, E. & KELLEHER, B. P. 2007. Microbially derived inputs to soil organic matter: are current estimates too low? Environmental Science & Technology, 41, 8070-8076. SINGLETON, G. A. & LAVKULICH, M. 1987. A soil chronosequence on beach sands, Vancouver Island, British Columbia. Canadian Journal of Soil Science, 67, 795-810. SIREGAR, A., KLEBER, M., MIKUTTA, R. & JAHN, R. 2005. Sodium hypochlorite oxidation reduces soil organic matter concentrations without affecting inorganic soil constituents. European Journal of Soil Science, 56, 481-490. SIX, J., BOSSUYT, H., DEGRYZE, S. & DENEF, K. 2004. A history of research on the link between (micro)aggregates, soil biota, and soil organic matter dynamics. Soil and Tillage Research, 79, 7- 31. SIX, J., CONANT, R., PAUL, E. & PAUSTIAN, K. 2002. Stabilization mechanisms of soil organic matter: Implications for C- saturation of soils. Plant and soil. SIX, J., FREY, S. D., THIET, R. K. & BATTEN, K. M. 2006. Bacterial and Fungal Contributions to Carbon Sequestration in Agroecosystems. Soil Science Society of America Journal, 70, 555. SKUBLOV, S. & DRUGOVA, G. 2003. Patterns of trace-element distribution in calcic amphiboles as a function of metamorphic grade. The Canadian Mineralogist, 41, 383-392. SLEUTEL, S., LEINWEBER, P., ARA BEGUM, S., KADER, M. A. & DE NEVE, S. 2009. Shifts in soil organic matter composition following treatment with sodium hypochlorite and hydrofluoric acid. Geoderma, 149, 257-266. SMITH, P., FANG, C., DAWSON, J. J. & MONCRIEFF, J. B. 2008. Impact of global warming on soil organic carbon. Advances in agronomy, 97, 1-43.

319

References

SMITS, M. M., HERRMANN, A. M., DUANE, M., DUCKWORTH, O. W., BONNEVILLE, S., BENNING, L. G. & LUNDSTRÖM, U. 2009. The fungal–mineral interface: challenges and considerations of micro-analytical developments. Fungal Biology Reviews, 23, 122-131. SMITS, M. M., HOFFLAND, E., JONGMANS, A. G. & VAN BREEMEN, N. 2005. Contribution of mineral tunneling to total feldspar weathering. Geoderma, 125, 59-69. SOKOLOVA, T. A. 2013. Decomposition of clay minerals in model experiments and in soils: possible mechanisms, rates, and diagnostics (analysis of literature). Eurasian Soil Science, 46, 182-197. SOLLINS, P., HOMANN, P. & CALDWELL, B. 1996. Stabilization and destabilization of soil organic matter: mechanisms and controls. Geoderma. SOLLINS, P., KRAMER, M. G., SWANSTON, C., LAJTHA, K., FILLEY, T., AUFDENKAMPE, A. K., WAGAI, R. & BOWDEN, R. D. 2009. Sequential density fractionation across soils of contrasting mineralogy: evidence for both microbial- and mineral- controlled soil organic matter stabilization. Biogeochemistry, 96, 209-231. SOLLINS, P., SWANSTON, C., KLEBER, M., FILLEY, T., KRAMER, M., CROW, S., CALDWELL, B. A., LAJTHA, K. & BOWDEN, R. 2006. Organic C and N stabilization in a forest soil: Evidence from sequential density fractionation. Soil Biology and Biochemistry, 38, 3313-3324. SOLLINS, P., SWANSTON, C. & KRAMER, M. 2007. Stabilization and destabilization of soil organic matter—a new focus. Biogeochemistry, 85, 1-7. SONKE, J. E. 2006. Lanthanide-humic substances complexation. II. Calibration of humic ion-binding model V. Environmental science & technology, 40, 7481-7487. SONKE, J. E. & SALTERS, V. J. M. 2006. Lanthanide–humic substances complexation. I. Experimental evidence for a lanthanide contraction effect. Geochimica et Cosmochimica Acta, 70, 1495-1506.

320

References

SPECHT, C. H., KUMKE, M. U. & FRIMMEL, F. H. 2000. Characterization of NOM adsorption to clay minerals by size exclusion chromatography. Water Research, 34, 4063-4069. SPOSITO, G. 2008. The chemistry of soils, New York, NY., Oxford university press, Inc. STARR, M. & LINDROOS, A.-J. 2006. Changes in the rate of release of Ca and Mg and normative mineralogy due to weathering along a 5300-year chronosequence of boreal forest soils. Geoderma, 133, 269-280. STERN, J. C., SONKE, J. E. & SALTERS, V. J. 2007. A capillary electrophoresis-ICP-MS study of rare earth element complexation by humic acids. Chemical Geology, 246, 170-180. STEVENSON, F. J. 1994. Humus Chemistry: Genesis, Composition, Reactions, New York, Wiley. STILLE, P., PIERRET, M.-C., STEINMANN, M., CHABAUX, F., BOUTIN, R., AUBERT, D., POURCELOT, L. & MORVAN, G. 2009. Impact of atmospheric deposition, biogeochemical cycling and water– mineral interaction on REE fractionation in acidic surface soils and soil water (the Strengbach case). Chemical Geology, 264, 173-186. STOOPS, G. 2003. Guidelines for analysis and description of soil and regolith thin sections, Madison, Wisconsin, USA, Soil Science Society of America Inc. STUMM, W. 1992. Chemistry of the solid-water interface., New York, John Wiley & Son Inc. STÜTZER, A. 1998. Early stages of podzolisation in young aeolian sediments, western Jutland. Catena, 32, 115-129. SUTTON, R. & SPOSITO, G. 2005. Molecular structure in soil humic substances: the new view. Environmental Science & Technology, 39, 9009-9015. SWANSTON, C. W., TORN, M. S., HANSON, P. J., SOUTHON, J. R., GARTEN, C. T., HANLON, E. M. & GANIO, L. 2005. Initial characterization of processes of soil carbon stabilization using forest stand-level radiocarbon enrichment. Geoderma, 128, 52-62. TAKAHASHI, Y., SAKASHIMA, T. & SHIMIZU, H. 2003. Observation of tetravalent cerium in zircon and its reduction by radiation effect. Geophysical Research Letters, 30, 37-1.

321

References

TAMM, O. 1922. Eine methode zur bestimmung der anorganischen komponente des gelkomplexes im boden. Meddelanden fran Statens Skogsförsöksanstalt, 19, 385-404. TANG, J. & JOHANNESSON, K. H. 2003. Speciation of rare earth elements in natural terrestrial waters: assessing the role of dissolved organic matter from the modeling approach. Geochimica et Cosmochimica Acta, 67, 2321-2339. TANG, J. & JOHANNESSON, K. H. 2010. Ligand extraction of rare earth elements from aquifer sediments: Implications for rare earth element complexation with organic matter in natural waters. Geochimica et Cosmochimica Acta, 74, 6690-6705. TAUNTON, A. E., WELCH, S. A. & BANFIELD, J. F. 2000. Microbial controls on phosphate and lanthanide distributions during granite weathering and soil formation. Chemical Geology, 169, 371-382. TAYLOR, L., LEAKE, J., QUIRK, J., HARDY, K., BANWART, S. & BEERLING, D. 2009. Biological weathering and the long‐term carbon cycle: integrating mycorrhizal evolution and function into the current paradigm. Geobiology, 7, 171-191. TEMMINGHOFF, E., ZEE, S. & HAAN, F. 1998. Effects of dissolved organic matter on the mobility of copper in a contaminated sandy soil. European Journal of Soil Science, 49, 617-628. TEN HAVE, R. & TEUNISSEN, P. J. 2001. Oxidative mechanisms involved in lignin degradation by white-rot fungi. Chemical Reviews, 101, 3397-3414. THENG, B. K. & YUAN, G. 2008. Nanoparticles in the soil environment. Elements, 4, 395-399. THEVENOT, M., DIGNAC, M.-F. & RUMPEL, C. 2010. Fate of lignins in soils: A review. Soil Biology and Biochemistry, 42, 1200-1211. THOMPSON, A. 2016. The role of redox variability in structuring iron cycling in soils. Goldschmidt 2016. Japan. THOMPSON, A., CHADWICK, O. A., BOMAN, S. & CHOROVER, J. 2006a. Colloid mobilization during soil iron redox oscillations. Environmental science & technology, 40, 5743-5749. THOMPSON, A., CHADWICK, O. A., RANCOURT, D. G. & CHOROVER, J. 2006b. Iron-oxide crystallinity increases during soil redox oscillations. Geochimica et Cosmochimica Acta, 70, 1710-1727.

322

References

THOMPSON, A., RUIZ, J., CHADWICK, O. A., TITUS, M. & CHOROVER, J. 2007. Rayleigh fractionation of iron isotopes during pedogenesis along a climate sequence of Hawaiian basalt. Chemical Geology, 238, 72-83. TIPPING, E. 1998. Humic ion-binding model VI: an improved description of the interactions of protons and metal ions with humic substances. Aquatic geochemistry, 4, 3-47. TISDALL, J. & OADES, J. M. 1982. Organic matter and water‐stable aggregates in soils. Journal of soil science, 33, 141-163. TORN, M., SWANSTON, C., CASTANHA, C. & TRUMBORE, S. 2009. Storage and turnover of organic matter in soil. Biophysico- chemical processes involving natural nonliving organic matter in environmental systems. Wiley, Hoboken, 219-272. TORN, M., TRUMBORE, S., CHADWICK, O. & VITOUSEK…, P. 1997. Mineral control of soil organic carbon storage and turnover. Nature. TORN, M. S., VITOUSEK, P. M. & TRUMBORE, S. E. 2005. The influence of nutrient availability on soil organic matter turnover estimated by incubations and radiocarbon modeling. Ecosystems, 8, 352-372. TOTSCHE, K. U., RENNERT, T., GERZABEK, M. H., KÖGEL-KNABNER, I., SMALLA, K., SPITELLER, M. & VOGEL, H.-J. 2010. Biogeochemical interfaces in soil: The interdisciplinary challenge for soil science. Journal of Plant Nutrition and Soil Science, 173, 88-99. TOWELL, D. G., SPIRN, R. V. & WINCHESTER, J. W. 1969. Europium anomalies and the genesis of basalt: A discussion. Chemical Geology, 4, 461-464. TRIPATHI, J. K. & RAJAMANI, V. 2007. Geochemistry and origin of ferruginous nodules in weathered granodioritic gneisses, Mysore Plateau, Southern India. Geochimica et Cosmochimica Acta, 71, 1674-1688. TRUMBORE, S. 2009. Radiocarbon and Soil Carbon Dynamics. Annual Review of Earth and Planetary Sciences, 37, 47-66. TRUMBORE, S. E. & CZIMCZIK, C. I. 2008. An uncertain future for soil carbon. science, 321.

323

References

TURPAULT, M. P., RIGHI, D. & UTÉRANO, C. 2008. Clay minerals: Precise markers of the spatial and temporal variability of the biogeochemical soil environment. Geoderma, 147, 108-115. TYLER, G. 2004. Vertical distribution of major, minor, and rare elements in a Haplic Podzol. Geoderma, 119, 277-290. UGOLINI, F. C. & DAHLGREN, R. A. 1987. The mechanism of podzolization as revealed by soil solution studies. In: RIGHI, D. & CHAUVEL, A. (eds.) Podzols and podzolization. Plasisier, France: Assoc. Fr. Estude Sol. UGOLINI, F. C. & SLETTEN, R. S. 1991. The role of proton donors in pedogenesis as revealed by soil solution studies. Soil Science, 151, 59-75. UROZ, S., CALVARUSO, C., TURPAULT, M.-P. & FREY-KLETT, P. 2009. Mineral weathering by bacteria: ecology, actors and mechanisms. Trends in Microbiology, 17, 378-387. VAN BREEMEN, N. & BUURMAN, P. 2002. Soil formation, Springer Science & Business Media. VAN BREEMEN, N., FINLAY, R., LUNDSTRÖM, U., JONGMANS, A. G., GIESLER, R. & OLSSON, M. 2000b. Mycorrhizal weathering: A true case of mineral plant nutrition? Biogeochemistry, 49, 53- 67. VAN BREEMEN, N., LUNDSTRÖM, U. S. & JONGMANS, A. G. 2000a. Do plants drive podzolization via rock-eating mycorrhizal fungi? Geoderma, 94, 163-171. VAN HEES, P., GODBOLD, D., JENTSCHKE, G. & JONES, D. 2003. Impact of ectomycorrhizas on the concentration and biodegradation of simple organic acids in a forest soil. European Journal of Soil Science, 54, 697-706. VAN HEES, P. A. W., LUNDSTRÖM, U. S. & GIESLER, R. 2000. Low molecular weight organic acids and their Al-complexes in soil solution—composition, distribution and seasonal variation in three podzolized soils. Geoderma, 94, 173-200. VAN RANST, E., DE CONINCK, F., TAVERNIER, R. & LANGOHR, R. 1982. Mineralogy in silty to loamy soils of central and high Belgium in respect to autochtonous and allochtonous materials. Bulletin de la Société belge de Géologie, 91, 27-44.

324

References

VANDENBYGAART, A. & PROTZ, R. 1995. Soil genesis on a chronosequence, Pinery Provincial Park, Ontario. Canadian Journal of Soil Science, 75, 63-72. VÁZQUEZ-ORTEGA, A., PERDRIAL, J., HARPOLD, A., ZAPATA-RÍOS, X., RASMUSSEN, C., MCINTOSH, J., SCHAAP, M., PELLETIER, J. D., BROOKS, P. D., AMISTADI, M. K. & CHOROVER, J. 2015. Rare earth elements as reactive tracers of biogeochemical weathering in forested rhyolitic terrain. Chemical Geology, 391, 19-32. VELDE, B. B. & MEUNIER, A. 2008. The origin of clay minerals in soils and weathered rocks, Springer Science & Business Media. VERMEIRE, M.-L., CORNU, S., FEKIACOVA, Z., DETIENNE, M., DELVAUX, B. & CORNÉLIS, J.-T. 2016. Rare earth elements dynamics along pedogenesis in a chronosequence of podzolic soils. Chemical Geology, 446, 163-174. VIOLLIER, E., INGLETT, P., HUNTER, K., ROYCHOUDHURY, A. & VAN CAPPELLEN, P. 2000. The ferrozine method revisited: Fe (II)/Fe (III) determination in natural waters. Applied geochemistry, 15, 785-790. VODYANITSKII, Y. N. 2003. Iron hydroxides in biogenic neoformations of forest soils of the Russian plain. EURASIAN SOIL SCIENCE C/C OF POCHVOVEDENIE, 36, 1286-1297. VODYANITSKII, Y. N., GORYACHKIN, S. V. & SAVICHEV, A. T. 2011. Distribution of rare-earth (Y, La, Ce) and other heavy metals in the profiles of the podzolic soil group. Eurasian Soil Science, 44, 500-509. VODYANITSKII, Y. N., SAVICHEV, A. T., VASIL’EV, A. A., LOBANOVA, E. S., CHASHCHIN, A. N. & PROKOPOVICH, E. V. 2010. Contents of heavy alkaline-earth (Sr, Ba) and rare-earth (Y, La, Ce) metals in technogenically contaminated soils. Eurasian Soil Science, 43, 822-832. VODYANITSKII, Y. N., VASIL’EV, A., KOZHEVA, A. & SATAEV, E. 2006. Specific features of iron behavior in soddy-podzolic and alluvial gleyed soils of the Middle Cis-Urals region. Eurasian Soil Science, 39, 354-366. VOLKMER, B. & HEINEMANN, M. 2011. Condition-Dependent Cell Volume and Concentration of Escherichia coli to Facilitate Data

325

References

Conversion for Systems Biology Modeling. PLoS ONE, 6, e23126. VON LÜTZOW, M. & KÖGEL-KNABNER, I. 2009. Temperature sensitivity of soil organic matter decomposition—what do we know? Biology and Fertility of Soils, 46, 1-15. VON LÜTZOW, M., KÖGEL-KNABNER, I., EKSCHMITT, K., FLESSA, H., GUGGENBERGER, G., MATZNER, E. & MARSCHNER, B. 2007. SOM fractionation methods: Relevance to functional pools and to stabilization mechanisms. Soil Biology and Biochemistry, 39, 2183-2207. VON LÜTZOW, M., KÖGEL-KNABNER, I., EKSCHMITT, K., MATZNER, E., GUGGENBERGER, G., MARSCHNER, B. & FLESSA, H. 2006. Stabilization of organic matter in temperate soils: mechanisms and their relevance under different soil conditions - a review. European Journal of Soil Science, 57, 426-455. VON LÜTZOW, M., KÖGEL-KNABNER, I., LUDWIG, B., MATZNER, E., FLESSA, H., EKSCHMITT, K., GUGGENBERGER, G., MARSCHNER, B. & KALBITZ, K. 2008. Stabilization mechanisms of organic matter in four temperate soils: Development and application of a conceptual model. Journal of Plant Nutrition and Soil Science, 171, 111-124. VON WANDRUSZKA, R. 1998. THE MICELLAR MODEL OF HUMIC ACID: EVIDENCE FROM PYRENE FLUORESCENCE MEASUREMENTS. Soil science, 163, 921-930. WALKER, L. R., WARDLE, D. A., BARDGETT, R. D. & CLARKSON, B. D. 2010. The use of chronosequences in studies of ecological succession and soil development. Journal of Ecology, 98, 725- 736. WANG, H., ZHU, J., FU, Q.-L., XIONG, J.-W., HONG, C., HU, H.-Q. & VIOLANTE, A. 2015. Adsorption of Phosphate onto Ferrihydrite and Ferrihydrite-Humic Acid Complexes. Pedosphere, 25, 405- 414. WANG, Q., HUANG, B., GUAN, Z., YANG, L. & LI, B. 2001. Speciation of rare earth elements in soil by sequential extraction then HPLC coupled with visible and ICP-MS detection. Analytical and Bioanalytical Chemistry, 370, 1041-1047.

326

References

WEBER, K. A., ACHENBACH, L. A. & COATES, J. D. 2006. Microorganisms pumping iron: anaerobic microbial iron oxidation and reduction. Nature Reviews Microbiology, 4, 752-764. WEINER, S. & DOVE, P. M. 2003. An overview of biomineralization processes and the problem of the vital effect. Reviews in mineralogy and geochemistry, 54, 1-29. WENG, L., VAN RIEMSDIJK, W. H. & HIEMSTRA, T. 2007. Adsorption of humic acids onto goethite: Effects of molar mass, pH and ionic strength. Journal of Colloid and Interface Science, 314, 107-118. WENG, L. P., KOOPAL, L. K., HIEMSTRA, T., MEEUSSEN, J. C. L. & VAN RIEMSDIJK, W. H. 2005. Interactions of calcium and fulvic acid at the goethite-water interface. Geochimica et Cosmochimica Acta, 69, 325-339. WENGEL, M., KOTHE, E., SCHMIDT, C. M., HEIDE, K. & GLEIXNER, G. 2006. Degradation of organic matter from black shales and charcoal by the wood-rotting fungus Schizophyllum commune and release of DOC and heavy metals in the aqueous phase. Science of the Total Environment, 367, 383-393. WERSHAW, R. L. 1994. Membrane-micelle model for humus in soils and sediments and its relation to humification. US Geological Survey; USGPO. WERSHAW, R. L. 2004. Evaluation of conceptual models of natural organic matter (humus) from a consideration of the chemical and biochemical processes of humification. Scientific Investigations Report No. 2004-5121. Reston, VA.: U.S. Geological Survey. WHITE, A. F., VIVIT, D. V., SCHULZ, M. S., BULLEN, T. D., EVETT, R. R. & AAGARWAL, J. 2012. Biogenic and pedogenic controls on Si distributions and cycling in grasslands of the Santa Cruz soil chronosequence, California. Geochimica et Cosmochimica Acta, 94, 72-94. WIEDERHOLD, J. G., TEUTSCH, N., KRAEMER, S. M., HALLIDAY, A. N. & KRETZSCHMAR, R. 2007a. Iron isotope fractionation during pedogenesis in redoximorphic soils. Soil Science Society of America Journal, 71, 1840-1850. WIEDERHOLD, J. G., TEUTSCH, N., KRAEMER, S. M., HALLIDAY, A. N. & KRETZSCHMAR, R. 2007b. Iron isotope fractionation in oxic

327

References

soils by mineral weathering and podzolization. Geochimica et Cosmochimica Acta, 71, 5821-5833. WIESENBERG, G. L., SCHWARZBAUER, J., SCHMIDT, M. W. & SCHWARK, L. 2004. Source and turnover of organic matter in agricultural soils derived from n-alkane/n-carboxylic acid compositions and C-isotope signatures. Organic Geochemistry, 35, 1371-1393. WILSON, M. 1999. The origin and formation of clay minerals in soils: past, present and future perspectivesd. Clay Minerals, 34, 7-25. WOLTERS, V. 2000. Invertebrate control of soil organic matter stability. Biology and fertility of Soils, 31, 1-19. WÖSTEN, H. A. 2001. Hydrophobins: multipurpose proteins. Annual Reviews in Microbiology, 55, 625-646. WRB 2015. World Reference Base for Soil Resources 2014, update 2015, International soil classification system for naming soils and creating legends for soil maps., Rome, IUSS Working Group, FAO. XU, R. K., HU, Y. F., DYNES, J. J., ZHAO, A. Z., BLYTH, R. I. R., KOZAK, L. M. & HUANG, P. M. 2010. Coordination nature of aluminum (oxy)hydroxides formed under the influence of low molecular weight organic acids and a soil humic acid studied by X-ray absorption spectroscopy. Geochimica et Cosmochimica Acta, 74, 6422-6435. YAMAMOTO, Y., TAKAHASHI, Y. & SHIMIZU, H. 2005. Systematics of stability constants of fulvate complexes with rare earth ions. Chemistry Letters, 34, 880-881. YAN, X.-P., KERRICH, R. & HENDRY, M. J. 1999. Sequential leachates of multiple grain size fractions from a clay-rich till, Saskatchewan, Canada: implications for controls on the rare earth element geochemistry of porewaters in an aquitard. Chemical Geology, 158, 53-79. YUSOFF, Z. M., NGWENYA, B. T. & PARSONS, I. 2013. Mobility and fractionation of REEs during deep weathering of geochemically contrasting granites in a tropical setting, Malaysia. Chemical Geology, 349-350, 71-86. ZABOWSKI, D. & UGOLINI, F. 1992. Seasonality in the mineral stability of a subalpine spodosol. Soil Science, 154, 497-507.

328

References

ZACHARA, J. M., FREDRICKSON, J. K., LI, S.-M., KENNEDY, D. W., SMITH, S. C. & GASSMAN, P. L. 1998. Bacterial reduction of crystalline Fe3+ oxides in single phase suspensions and subsurface materials. American Mineralogist, 83, 1426-1443. ZACHARA, J. M., FREDRICKSON, J. K., SMITH, S. C. & GASSMAN, P. L. 2001. Solubilization of Fe (III) oxide-bound trace metals by a dissimilatory Fe (III) reducing bacterium. Geochimica et Cosmochimica Acta, 65, 75-93. ZACHARA, J. M., KUKKADAPU, R. K., FREDRICKSON, J. K., GORBY, Y. A. & SMITH, S. C. 2002. Biomineralization of Poorly Crystalline Fe(III) Oxides by Dissimilatory Metal Reducing Bacteria (DMRB). Geomicrobiology Journal, 19, 179-207. ZELLES, L. 1997. Phospholipid fatty acid profiles in selected members of soil microbial communities. Chemosphere, 35, 275-294. ZELLES, L. 1999. Fatty acid patterns of phospholipids and lipopolysaccharides in the characterisation of microbial communities in soil: A review. Biology and Fertility of Soils, 29, 111-129. ZHANG, S. & SHAN, X.-Q. 2001. Speciation of rare earth elements in soil and accumulation by wheat with rare earth fertilizer application. Environmental Pollution, 112, 395-405. ZHANG, X. & AMELUNG, W. 1996. Gas chromatograph1c determination of muramic acid, glucosamine, mannosamine, and galactosamine in soils. Soil Biology and Biochemistry, 28, 1201-1206. ZIEGLER, K., CHADWICK, O. A., BRZEZINSKI, M. A. & KELLY, E. F. 2005. Natural variations of δ 30 Si ratios during progressive basalt weathering, Hawaiian Islands. Geochimica et Cosmochimica Acta, 69, 4597-4610. ZIMMERMANN, M., LEIFELD, J., ABIVEN, S., SCHMIDT, M. W. & FUHRER, J. 2007. Sodium hypochlorite separates an older soil organic matter fraction than acid hydrolysis. Geoderma, 139, 171-179.

329